diff options
author | Jordan K. Hubbard <jkh@FreeBSD.org> | 1993-06-18 04:22:21 +0000 |
---|---|---|
committer | Jordan K. Hubbard <jkh@FreeBSD.org> | 1993-06-18 04:22:21 +0000 |
commit | b76095a4307cc94ec7cd722853f9b032e45e6ea4 (patch) | |
tree | 890f91d43eec35dc2f71a54410491f6503ca5b38 /gnu/usr.bin/tar | |
parent | 7c434002a4e47486e9a2d7b2f32b1ddf42d37e2a (diff) | |
download | src-test2-b76095a4307cc94ec7cd722853f9b032e45e6ea4.tar.gz src-test2-b76095a4307cc94ec7cd722853f9b032e45e6ea4.zip |
Notes
Diffstat (limited to 'gnu/usr.bin/tar')
35 files changed, 21463 insertions, 0 deletions
diff --git a/gnu/usr.bin/tar/COPYING b/gnu/usr.bin/tar/COPYING new file mode 100644 index 000000000000..a43ea2126fb6 --- /dev/null +++ b/gnu/usr.bin/tar/COPYING @@ -0,0 +1,339 @@ + GNU GENERAL PUBLIC LICENSE + Version 2, June 1991 + + Copyright (C) 1989, 1991 Free Software Foundation, Inc. + 675 Mass Ave, Cambridge, MA 02139, USA + Everyone is permitted to copy and distribute verbatim copies + of this license document, but changing it is not allowed. + + Preamble + + The licenses for most software are designed to take away your +freedom to share and change it. By contrast, the GNU General Public +License is intended to guarantee your freedom to share and change free +software--to make sure the software is free for all its users. This +General Public License applies to most of the Free Software +Foundation's software and to any other program whose authors commit to +using it. (Some other Free Software Foundation software is covered by +the GNU Library General Public License instead.) You can apply it to +your programs, too. + + When we speak of free software, we are referring to freedom, not +price. Our General Public Licenses are designed to make sure that you +have the freedom to distribute copies of free software (and charge for +this service if you wish), that you receive source code or can get it +if you want it, that you can change the software or use pieces of it +in new free programs; and that you know you can do these things. + + To protect your rights, we need to make restrictions that forbid +anyone to deny you these rights or to ask you to surrender the rights. +These restrictions translate to certain responsibilities for you if you +distribute copies of the software, or if you modify it. + + For example, if you distribute copies of such a program, whether +gratis or for a fee, you must give the recipients all the rights that +you have. You must make sure that they, too, receive or can get the +source code. And you must show them these terms so they know their +rights. + + We protect your rights with two steps: (1) copyright the software, and +(2) offer you this license which gives you legal permission to copy, +distribute and/or modify the software. + + Also, for each author's protection and ours, we want to make certain +that everyone understands that there is no warranty for this free +software. If the software is modified by someone else and passed on, we +want its recipients to know that what they have is not the original, so +that any problems introduced by others will not reflect on the original +authors' reputations. + + Finally, any free program is threatened constantly by software +patents. We wish to avoid the danger that redistributors of a free +program will individually obtain patent licenses, in effect making the +program proprietary. To prevent this, we have made it clear that any +patent must be licensed for everyone's free use or not licensed at all. + + The precise terms and conditions for copying, distribution and +modification follow. + + GNU GENERAL PUBLIC LICENSE + TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION + + 0. This License applies to any program or other work which contains +a notice placed by the copyright holder saying it may be distributed +under the terms of this General Public License. The "Program", below, +refers to any such program or work, and a "work based on the Program" +means either the Program or any derivative work under copyright law: +that is to say, a work containing the Program or a portion of it, +either verbatim or with modifications and/or translated into another +language. (Hereinafter, translation is included without limitation in +the term "modification".) Each licensee is addressed as "you". + +Activities other than copying, distribution and modification are not +covered by this License; they are outside its scope. The act of +running the Program is not restricted, and the output from the Program +is covered only if its contents constitute a work based on the +Program (independent of having been made by running the Program). +Whether that is true depends on what the Program does. + + 1. You may copy and distribute verbatim copies of the Program's +source code as you receive it, in any medium, provided that you +conspicuously and appropriately publish on each copy an appropriate +copyright notice and disclaimer of warranty; keep intact all the +notices that refer to this License and to the absence of any warranty; +and give any other recipients of the Program a copy of this License +along with the Program. + +You may charge a fee for the physical act of transferring a copy, and +you may at your option offer warranty protection in exchange for a fee. + + 2. You may modify your copy or copies of the Program or any portion +of it, thus forming a work based on the Program, and copy and +distribute such modifications or work under the terms of Section 1 +above, provided that you also meet all of these conditions: + + a) You must cause the modified files to carry prominent notices + stating that you changed the files and the date of any change. + + b) You must cause any work that you distribute or publish, that in + whole or in part contains or is derived from the Program or any + part thereof, to be licensed as a whole at no charge to all third + parties under the terms of this License. + + c) If the modified program normally reads commands interactively + when run, you must cause it, when started running for such + interactive use in the most ordinary way, to print or display an + announcement including an appropriate copyright notice and a + notice that there is no warranty (or else, saying that you provide + a warranty) and that users may redistribute the program under + these conditions, and telling the user how to view a copy of this + License. (Exception: if the Program itself is interactive but + does not normally print such an announcement, your work based on + the Program is not required to print an announcement.) + +These requirements apply to the modified work as a whole. If +identifiable sections of that work are not derived from the Program, +and can be reasonably considered independent and separate works in +themselves, then this License, and its terms, do not apply to those +sections when you distribute them as separate works. But when you +distribute the same sections as part of a whole which is a work based +on the Program, the distribution of the whole must be on the terms of +this License, whose permissions for other licensees extend to the +entire whole, and thus to each and every part regardless of who wrote it. + +Thus, it is not the intent of this section to claim rights or contest +your rights to work written entirely by you; rather, the intent is to +exercise the right to control the distribution of derivative or +collective works based on the Program. + +In addition, mere aggregation of another work not based on the Program +with the Program (or with a work based on the Program) on a volume of +a storage or distribution medium does not bring the other work under +the scope of this License. + + 3. You may copy and distribute the Program (or a work based on it, +under Section 2) in object code or executable form under the terms of +Sections 1 and 2 above provided that you also do one of the following: + + a) Accompany it with the complete corresponding machine-readable + source code, which must be distributed under the terms of Sections + 1 and 2 above on a medium customarily used for software interchange; or, + + b) Accompany it with a written offer, valid for at least three + years, to give any third party, for a charge no more than your + cost of physically performing source distribution, a complete + machine-readable copy of the corresponding source code, to be + distributed under the terms of Sections 1 and 2 above on a medium + customarily used for software interchange; or, + + c) Accompany it with the information you received as to the offer + to distribute corresponding source code. (This alternative is + allowed only for noncommercial distribution and only if you + received the program in object code or executable form with such + an offer, in accord with Subsection b above.) + +The source code for a work means the preferred form of the work for +making modifications to it. For an executable work, complete source +code means all the source code for all modules it contains, plus any +associated interface definition files, plus the scripts used to +control compilation and installation of the executable. However, as a +special exception, the source code distributed need not include +anything that is normally distributed (in either source or binary +form) with the major components (compiler, kernel, and so on) of the +operating system on which the executable runs, unless that component +itself accompanies the executable. + +If distribution of executable or object code is made by offering +access to copy from a designated place, then offering equivalent +access to copy the source code from the same place counts as +distribution of the source code, even though third parties are not +compelled to copy the source along with the object code. + + 4. You may not copy, modify, sublicense, or distribute the Program +except as expressly provided under this License. Any attempt +otherwise to copy, modify, sublicense or distribute the Program is +void, and will automatically terminate your rights under this License. +However, parties who have received copies, or rights, from you under +this License will not have their licenses terminated so long as such +parties remain in full compliance. + + 5. You are not required to accept this License, since you have not +signed it. However, nothing else grants you permission to modify or +distribute the Program or its derivative works. These actions are +prohibited by law if you do not accept this License. Therefore, by +modifying or distributing the Program (or any work based on the +Program), you indicate your acceptance of this License to do so, and +all its terms and conditions for copying, distributing or modifying +the Program or works based on it. + + 6. Each time you redistribute the Program (or any work based on the +Program), the recipient automatically receives a license from the +original licensor to copy, distribute or modify the Program subject to +these terms and conditions. You may not impose any further +restrictions on the recipients' exercise of the rights granted herein. +You are not responsible for enforcing compliance by third parties to +this License. + + 7. If, as a consequence of a court judgment or allegation of patent +infringement or for any other reason (not limited to patent issues), +conditions are imposed on you (whether by court order, agreement or +otherwise) that contradict the conditions of this License, they do not +excuse you from the conditions of this License. If you cannot +distribute so as to satisfy simultaneously your obligations under this +License and any other pertinent obligations, then as a consequence you +may not distribute the Program at all. For example, if a patent +license would not permit royalty-free redistribution of the Program by +all those who receive copies directly or indirectly through you, then +the only way you could satisfy both it and this License would be to +refrain entirely from distribution of the Program. + +If any portion of this section is held invalid or unenforceable under +any particular circumstance, the balance of the section is intended to +apply and the section as a whole is intended to apply in other +circumstances. + +It is not the purpose of this section to induce you to infringe any +patents or other property right claims or to contest validity of any +such claims; this section has the sole purpose of protecting the +integrity of the free software distribution system, which is +implemented by public license practices. Many people have made +generous contributions to the wide range of software distributed +through that system in reliance on consistent application of that +system; it is up to the author/donor to decide if he or she is willing +to distribute software through any other system and a licensee cannot +impose that choice. + +This section is intended to make thoroughly clear what is believed to +be a consequence of the rest of this License. + + 8. If the distribution and/or use of the Program is restricted in +certain countries either by patents or by copyrighted interfaces, the +original copyright holder who places the Program under this License +may add an explicit geographical distribution limitation excluding +those countries, so that distribution is permitted only in or among +countries not thus excluded. In such case, this License incorporates +the limitation as if written in the body of this License. + + 9. The Free Software Foundation may publish revised and/or new versions +of the General Public License from time to time. Such new versions will +be similar in spirit to the present version, but may differ in detail to +address new problems or concerns. + +Each version is given a distinguishing version number. If the Program +specifies a version number of this License which applies to it and "any +later version", you have the option of following the terms and conditions +either of that version or of any later version published by the Free +Software Foundation. If the Program does not specify a version number of +this License, you may choose any version ever published by the Free Software +Foundation. + + 10. If you wish to incorporate parts of the Program into other free +programs whose distribution conditions are different, write to the author +to ask for permission. For software which is copyrighted by the Free +Software Foundation, write to the Free Software Foundation; we sometimes +make exceptions for this. Our decision will be guided by the two goals +of preserving the free status of all derivatives of our free software and +of promoting the sharing and reuse of software generally. + + NO WARRANTY + + 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY +FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN +OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES +PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED +OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS +TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE +PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, +REPAIR OR CORRECTION. + + 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR +REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, +INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING +OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED +TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY +YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER +PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE +POSSIBILITY OF SUCH DAMAGES. + + END OF TERMS AND CONDITIONS + + Appendix: How to Apply These Terms to Your New Programs + + If you develop a new program, and you want it to be of the greatest +possible use to the public, the best way to achieve this is to make it +free software which everyone can redistribute and change under these terms. + + To do so, attach the following notices to the program. It is safest +to attach them to the start of each source file to most effectively +convey the exclusion of warranty; and each file should have at least +the "copyright" line and a pointer to where the full notice is found. + + <one line to give the program's name and a brief idea of what it does.> + Copyright (C) 19yy <name of author> + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + +Also add information on how to contact you by electronic and paper mail. + +If the program is interactive, make it output a short notice like this +when it starts in an interactive mode: + + Gnomovision version 69, Copyright (C) 19yy name of author + Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. + This is free software, and you are welcome to redistribute it + under certain conditions; type `show c' for details. + +The hypothetical commands `show w' and `show c' should show the appropriate +parts of the General Public License. Of course, the commands you use may +be called something other than `show w' and `show c'; they could even be +mouse-clicks or menu items--whatever suits your program. + +You should also get your employer (if you work as a programmer) or your +school, if any, to sign a "copyright disclaimer" for the program, if +necessary. Here is a sample; alter the names: + + Yoyodyne, Inc., hereby disclaims all copyright interest in the program + `Gnomovision' (which makes passes at compilers) written by James Hacker. + + <signature of Ty Coon>, 1 April 1989 + Ty Coon, President of Vice + +This General Public License does not permit incorporating your program into +proprietary programs. If your program is a subroutine library, you may +consider it more useful to permit linking proprietary applications with the +library. If this is what you want to do, use the GNU Library General +Public License instead of this License. diff --git a/gnu/usr.bin/tar/ChangeLog b/gnu/usr.bin/tar/ChangeLog new file mode 100644 index 000000000000..7934f2b1b2e3 --- /dev/null +++ b/gnu/usr.bin/tar/ChangeLog @@ -0,0 +1,1732 @@ +Thu Mar 25 13:32:40 1993 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * version.c: Released version 1.11.2. + + * Makefile.in (dist): Do the link differently; some of the + files have changed filesystems which makes it more complex. + + * Makefile.in (dist, shar): Use gzip instead of compress. + + * create.c (dump_file): Test for curdev==-1, not curdev<0. + Some losing NFS systems give negative device numbers + sometimes. + +Thu Mar 25 11:55:15 1993 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * level-0, level-1 (TAR_PART1): Use `--block-size', not just + `--block', which is now ambiguous. + +Wed Mar 24 22:12:51 1993 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * backup-specs (TAR): New variable. + + * level-0, level-1 (TAR_PART1): Get path of GNU tar from `TAR' + variable, don't hardcode it. + +Sat Mar 20 00:20:05 1993 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * backup-specs (SLEEP_MESSAGE): put backslashes in front of nested + double quotes. + + * level-0, level-1 (BACKUP_DIRS): Don't put in quotes. + (LOGFILE): Use sed to construct name, not awk. + + * dump-remind (recipients): Replaced inefficient pipeline with a + single, simple sed script. + (volno): Deal with the possibility that VOLNO_FILE may not be + created yet. + +Fri Mar 19 15:05:15 1993 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * backup-specs (VOLNO_FILE): Removed abusive comment by Noah. + + * buffer.c (new_volume): Write the global volume number to the + volno file before running the info script, so that the script + can look at it. + +Thu Mar 18 20:11:54 1993 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * Makefile.in (AUX): Include `dump-remind' in distribution. + + * backup-specs (SLEEP_MESSAGE): New variable. + level-0, level-1: Use it instead of external `dont_touch' file. + + * level-0, level-1: Put most of the script in () and pipe + everything from the subshell through tee -a $LOGFILE. Since you + really want most of the output to go to the logfile anyway, and + since all those pipelines were preventing one from getting the + exit status of most commands, this seems like the right idea. + + * level-0, level-1 (LOGFILE): Use YYYY-MM-DD (all numeric) format + for log file name, since that makes the file names sortable in a + coherent way. Suffix should always be `level-n' where n is the + dump level. level-0 script was just using `-full' instead. + + * level-0, level-1 (DUMP_LEVEL): New variable. Set to `0' or `1' + in each script as appropriate. + + * level-0, level-1 (HOST): Renamed to `localhost' for clarity. + (host): renamed to `remotehost' for clarity. + + * level-0, level-1 (startdate): New variable. Use it in Subject + line of mailed report. + + * level-0, level-1: Fixed all instances where sed is called with a + script on the command line to use `-e' option. + + * level-0, level-1: Don't try to call logfile.sed to filter + LOGFILE. It's not distributed with tar and was never really used + anyway. + + * level-0, level-1: Put quotes around most variable names (barring + those that are known to intentionally contain text that should be + expanded into multiple words, like `TAR_PART1'). + + * level-0, level-1: Got rid of annoying trailing backslashes in awk + scripts. They were gratuitous. Made them a little more readable + by adding some whitespace. + +Wed Mar 17 10:30:58 1993 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * tar.c (describe, long_options): Changed --compress-block to + --block-compress. + (options): Fixed f_compress_block sanity check error message + to give the correct name of the option. + +Tue Mar 16 14:52:40 1993 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * extract.c (extract_archive): case LF_DIR: Do chown when + necessary. Don't bother jumping to set_filestat for + f_modified; repeat the chmod code here. Replace `break', + deleted on 2 September 1992. + + * tar.c (describe, long_options, options): Added gzip options + and use-compress-program option. + * tar.h: Added new compression options. + * buffer.c (child_open, open_archive): Use new compression options. + + * create.c (start_header): Only mask off high bits when + creating old-style archives. + * list.c (decode_header): Mask off potentially misleading + high bits from the mode when reading headers. + +Mon Mar 15 11:34:34 1993 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * extract.c (extract_archive): Put arguments in the right + order for error message. + + * create.c (deal_with_sparse): if the last byte was null, we + didn't write it out. + + * gnu.c, create.c, extract.c, diffarch.c, list.c throughout: + Replace malloc calls with ck_malloc and realloc with ck_realloc. + + * tar.c (describe): Improve doc for -L. + + * tar.c (name_next): Don't apply exclusion to explicitly named + files. + + * tar.c (long_options, describe): Added new-volume-script as + an alias for info-script. + + * extract.c (extract_archive): LF_DUMPDIR case; misplaced paren. + + * extract.c (extract_archive): extract_file case, first if, + include space for null in namelen computation. + + * extract.c (extract_sparse_file): Use value returned by write + to properly create error message. + + * create.c (create_archive): Don't assume we have anything to + dump. + + * buffer.c (open_archive): Set current_file_name for the + volume header so that verbose listings work properly. + + * Makefile.in (realclean): Added getdate.c. + +Thu Jan 14 23:38:44 1993 David J. MacKenzie (djm@kropotkin.gnu.ai.mit.edu) + + * tar.c: Include fnmatch.h after port.h to make sure we get our FNM_* + (e.g. on HPUX 8). + +Tue Nov 24 08:30:54 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * tar.c (addname), gnu.c (read_dir_file): Use HAVE_GETCWD, not USG. + + * port.h, rmt.h: Use HAVE_STRING_H, not USG. + + * port.h: Add dir header decls. + * create.c, gnu.c: Use SYSNDIR, SYSDIR, and NDIR + instead of BSD42 and USG. Rename DP_NAMELEN to NLENGTH. + Use `struct dirent' instead of `struct direct'. + * create.c, gnu.c, tar.c: Remove dir header decls. + +Wed Nov 18 15:31:30 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * tar.c: Change FNM_TARPATH to FNM_LEADING_DIR to match change + in fnmatch.[ch]. + +Wed Oct 21 00:52:24 1992 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * level-0, level-1: put curly braces around variables for clarity. + + * backup-specs (DUMP_REMIND_SCRIPT): define it (but commented out + so that distributed dump scripts won't use it by default). + level-0, level-1 (TAR_PART1): use --info-script if + DUMP_REMIND_SCRIPT is defined. + dump-remind: new file (intended as an example). + +Thu Oct 15 03:33:28 1992 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * level-0, level-1: remove $LOGFILE.tmp files before exiting. + +Fri Oct 2 00:28:01 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * tar.c (describe): Fix some tab alignments. + + * Makefile.in (SRC3): Add getdate.c, for systems without bison/yacc + (like MS-DOS). + + * diffarch.c (diff_sparse_files): Add missing arg to fprintf calls. + + * extract.c (extract_archive, restore_saved_dir_info), + buffer.c (child_open), list.c (decode_header, print_header): + Delete unused vars. + + * port.c [__MSDOS__]: Have strstr, rename, and mkdir. Don't + define ck_pipe. + + * buffer.c, tar.c (init_volume_number, closeout_volume_number), + create.c (write_long): Declare as void, not int, since they + don't return a value. + +Thu Sep 24 00:06:02 1992 Michael I Bushnell (mib@churchy.gnu.ai.mit.edu) + + * level-0, level-1 (TAR_PART1): remove --atime-preserve + because of a total screw. + +Tue Sep 22 14:15:48 1992 Michael I Bushnell (mib@wookumz.gnu.ai.mit.edu) + + * buffer.c (close_archive): Removed leftover `break' from when + this was a switch. + +Tue Sep 22 08:33:16 1992 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * create.c, port.h: indented #pragma directives with 1 space. + +Fri Sep 18 14:15:17 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * All source files: re indented using GNU indent. + + * rtapelib.c (__rmt_read): Only read the amount left in the + buffer; otherwise a broken rmt server (which puts too much + data out) could overwrite past our buffer. + +Thu Sep 17 14:08:58 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * create.c: Throughout, use struct utimbuf rather than array + of longs. + + * configure.in: Check for getpwuid and getgrgid. + + * Makefile.in (SRC3, AUX): Move alloca.c to SRC3. + (OBJ3): Add @ALLOCA@. + + * Makefile.in (getdate.c): Look in srcdir for getdate.y. + + * buffer.c (close_archive): We can't check WTERMSIG + meaningfully unless we already know tha WIFSIGNALED is true. + (There is no guarantee it WTERMSIG will return zero when + WIFSIGNALED is false.) + * port.c (rmdir, mkdir): Check WIFSIGNALED rather than + WTERMSIG. + + * Makefile.in (getdate.c): Use $(YACC) instead of `yacc'. + +Tue Sep 15 14:49:48 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * version.c: Released version 1.11.1. + + * Makefile (AUX): Added NEWS. + + * Makefile.in (rmt): Added $(LIBS). + * configure.in: Added tests for libraries needed on Solaris. + + * mangle.c (extract_mangle): Null terminate link name for + losing archives missing it. + + * Makefile.in: added target and rule for getdate.c: getdate.y; + some makes don't have one built in. + +Mon Sep 14 16:23:15 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * tar.c (options, main): Advise use of --help rather than + +help. + + * create.c (write_long): Using hstat here is a Bad Idea, and + totally unnecessary at that. + + * list.c (read_header): Compute both signed and normal + checksums. + + * configure.in: Define BSD in the presence of /sdmach or + /../../mach. + + * diffarch.c, buffer.c: Declare valloc as void* rather than + char*. + + * Makefile.in: Don't install info files. + + * configure.in: Check for malloc was scrambled. + + * port.h: Undefine index and rindex if necessary; some + string.h's define them for us. + + * tar.c (addname): Missing braces after if. + * gnu.c (read_dir_file): Missing braces after if. + + * names.c: Add include of <stdio.h>, + + * create.c (start_header): Set current_file_name so that + print_header (used for verbose create) works properly. + (dump_file): Set current_link_name when setting up symlink + and hardlink records. + +Fri Sep 11 01:05:52 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * fnmatch.[ch]: New files. + * wildmat.c: File removed. + * tar.c: Include fnmatch.h and use fnmatch instead of wildmat. + * Makefile.in, makefile.pc: Replace wildmat.o(bj) with fnmatch. + +Thu Sep 10 23:19:30 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * buffer.c, tar.c: Remove redundant decls of getenv, rindex. + + * Makefile.in: Add uninstall target. + Define libdir instead of hardcoding /etc for installing rmt. + +Thu Sep 10 13:06:03 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * list.c (read_header): On second thought, that doesn't work + either, so just store the names in malloced areas. Sigh. + + * NEWS: New file. + * README: Removed things that belong in NEWS; point to it. + + * list.c (read_header): current_file_name and + current_link_name need to be set to the arrays in head rather + than header; header is the actual read buffer and will change. + + * extract.c (extract_archive): + * buffer.c (new_volume): `#' directives need to start in + column 1. + +Thu Sep 10 06:09:18 1992 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * level-0, level-1 (TAR_PART1): put --atime-preserve inside quotes. + +Wed Sep 9 13:34:26 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * Makefile.in (AUX): Add getpagesize.h. + (AUX): Comment out manuals. + (all): Comment out dependency on tar.info. + + * version.c: Release of version 1.11. + + * level-0, level-1 (TAR_PART1): Use --atime-preserve. + + * Makefile, configure.in: Arrange to use local malloc on HP-UX. + + * port.h Use the canonical Autoconf chunk for alloca instead + of just looking for gcc. + +Wed Sep 9 03:16:58 1992 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * port.h: If compiling with gcc, use __builtin_alloca. + +Tue Sep 8 16:13:41 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * extract.c: Removed long name support from here. + * list.c (read_header): Understand and skip longname/longlink + headers here. Names for current file are stored in new global + variables. All source files except create.c changed to refer + to current_file_name and current_link_name instead of fields + directly from the current header. + +Thu Sep 3 12:41:08 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * create.c (write_long): New function. + (dump_file): When writing link records or symlink records, use + new write_long function instead of mangling when the link + target is too long. + (start_header): Use write_long instead of mangling for long + names. + * extract.c (saverec): Recognize LF_LONGNAME and LF_LONGLINK. + (saverec): Throughout, use longname and longlink if they are set. + +Wed Sep 2 14:41:13 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * mangle.c: This is now deprecated; retain extract_mangle for + backward compatability. + + * list.c (print_header): patch from Chris Arthur to prevent + printing 0 when the gid or uid is null. + + * list.c (decode_header): patch from Chris Arthur to use the + gid field when the gid is empty, and similarly for uid. + + * extract.c: saved_dir_info, saved_dir_info_head: new type and + var. + (extract_archive): When extracting directories, now save info + in saved_dir_info_head. + (restore_saved_dir_info): New function. + * list.c (read_and): Call restore_saved_dir_info at the end of + the run. + This patch is from Chris Arthur (csa@pennies.sw.stratus.com). + +Mon Aug 31 15:39:55 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * create.c (create_archive): If there are no names specified, + write nothing on the archive instead of dumping ".". + + * buffer.c (open_archive): Useful error message. + + * tar.c, tar.h: Recognize f_atime_preserve. + * create.c (dump_file): Implement f_atime_preserve. + + * rmt.h (_remdev): Don't require /dev/ to be in remote archive + names; obey new force-local flag. + * tar.c, tar.h: Implement new force-local flag. + + * tar.c (describe): same-owner and same-order were confused. + + * create.c (dump_file): Check for toplevel had sense reversed. + + * buffer.c (new_archive): Don't free old_name...when these + come from the command line, they aren't malloced, and it isn't + important to save this trivial amount of memory. + + * tar.h: replace ar_file with ar_files, n_ar_files, + cur_ar_files. + * buffer.c (open_archive): multi-volume compressed archives + never worked; give an appropriate error. Change open of + ar_file to open of ar_files[0]. + (writeerror, readerror, flush_archive): use + ar_files[cur_ar_file] instead of ar_file. + (new_archive): Necessary changes to support ar_files. + * tar.c (options): handle multiple tape drive arguments. + +Fri Aug 28 17:42:13 1992 Michael I Bushnell (mib@wookumz.gnu.ai.mit.edu) + + * list.c (decode_header), create.c (start_header), tar.h (TMAGIC): + Undo djm's changes below; tar does not support the final + Posix.1 format; it's bad to make it look like it does. + +Sun Jul 19 02:13:46 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * port.h: Try to prevent redefining major. + * port.c: HAVE_BZERO -> minix. Fix a typo. + + * list.c (decode_header): Recognize the final POSIX.1 magic as + well as the early draft magic for ustar. + * create.c (start_header): Create a final POSIX.1 magic string + instead of an early draft string for ustar. + * tar.h (TMAGIC): Remove the trailing blanks. + + * rmt.c, rtapelib.c: Use POSIX and STDC headers if available. + * rmt.h: Declare the external functions defined in rtapelib.c. + +Tue Jul 14 00:44:37 1992 David J. MacKenzie (djm@apple-gunkies.gnu.ai.mit.edu) + + * pathmax.h: New file. + * port.h: Include it. + * create.c (create_archive): Allocate PATH_MAX instead of + NAME_MAX for temporary buffer so we don't have to figure out + what NAME_MAX is (portably). + +Fri Jul 10 08:30:42 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * gnu.c (collect_and_sort_names): write_dir_file has no argument. + + * level-0, level-1: Avoid silly Sun awk lossage. + +Mon Jul 6 20:11:32 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * port.c (rename): If unlinking the source at the end fails, + unlink the destination instead to avoid leaving a mess. + +Fri Jul 3 15:16:42 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * buffer.c, diffarch.c, update.c, rtapelib.c: Change NO_MTIO to + HAVE_SYS_MTIO_H. + + * port.c, tar.h: Change FOO_MISSING to HAVE_FOO. + +Tue Jun 23 23:39:02 1992 David J. MacKenzie (djm@goldman.gnu.ai.mit.edu) + + * rmt.c: Add #ifdefs to work on ISC. + +Wed May 20 00:12:27 1992 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) + + * port.h: Define major, minor, makedev if the system doesn't. + +Wed May 13 21:16:38 1992 Michael I Bushnell (mib@apple-gunkies.gnu.ai.mit.edu) + + * gnu.c (add_dir_name): Store legitimate value into + dir_contents when get_dir_contents returns NULL. + +Thu May 7 23:44:35 1992 Michael I Bushnell (mib@apple-gunkies.gnu.ai.mit.edu) + + * gnu.c (add_dir_name): Check for return of NULL from get_dir_contents; + see djm's change of Fri Jul 26 01:12:58 1991. + +Mon May 4 22:50:57 1992 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) + + * tar.h: Make comments for option names say -- instead of +. + +Thu Apr 30 03:09:16 1992 Noah Friedman (friedman@nutrimat.gnu.ai.mit.edu) + + * level-1: Added `$' before VOLNO_FILE in definition of TAR_PART1. + Added line to remove $VOLNO_FILE from any previous dump before + starting. + + * level-0, level-1: Change long options to use `--' instead of `+' + (support for `+' will go away soon) + +Wed Apr 29 14:23:10 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * tar.c, tar.t: Added +volno-file option. + buffer.c: New functions init_volume_number, + closeout_volume_number. + tar.c (main): Call new functions in the right place. + + * buffer.c (fl_write, fl_read): Mod to allow losing tape + drives which use short counts to indicate end of tape + correctly handle the multi-tape stuff. The read half won't + co-exist with f_reblock; there's no way to fix that, of + course. + + * tar.c, tar.h: Added new option +show-omitted-dirs, from + Karl Berry. + list.c (read_and): Implemented show-omitted-dirs. + + * tar.c, tar.h: Added new option +checkpoint. + buffer.c (fl_read, fl_write): Implemented +checkpoint lazily. + + * create.c (dump_file): Added toplevel argument; some devices + can be negative, so the old method was bogus. All callers + changed. + + * tar.c, tar.h: Added new option +ignore-failed-read. + create.c (dump_file): Implemented +ignore-failed-read. + + * create.c (finish_sparse_file): Commented out debugging printf. + + * tar.c, tar.h: Added new option +remove-files to delete files + after they are added to the archive. + create.c (dump_file): Implemented +remove-files for + everything but directories. I don't think they need it. + +Tue Apr 28 13:21:42 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * create.c: (dump_file): save_name needs to be set equal to p, + not something inside the header, because the header changes at + the first buffer flush. + +Fri Apr 24 10:41:13 1992 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * create.c: Djm incorrectly moved the include of port.h to + precede the include of sys/file.h; restored. + + * tar.c (main): Cases CMD_EXTRACT and CMD_LIST: declare error + string with const. + + * gnu.c (collect_and_sort_names): Leave if around + write_dir_file in place. + +Wed Apr 22 02:16:14 1992 David J. MacKenzie (djm@churchy.gnu.ai.mit.edu) + + * rtapelib.c: SIGTYPE -> RETSIGTYPE. + +Mon Mar 9 22:42:05 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * rtapelib.c: Reformat and make comments more complete. + Rename a few variables for clarity. + +Thu Mar 5 14:07:34 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) + + * tar.c (describe): Document long options as starting with --. + +Thu Jan 23 22:54:41 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * tar.c (options): Check get_date return value for error indication. + +Tue Dec 24 00:03:03 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * tar.c, gnu.c, extract.c, create.c, port.h, rmt.h: Change + POSIX ifdefs to HAVE_UNISTD_H and _POSIX_VERSION. + +Fri Dec 20 13:50:38 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * testpad.c (main): flush stderr so perror and fprintf + cooperate right. + +Wed Dec 18 16:52:42 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * port.h: Check MAJOR_IN_MKDEV and MAJOR_IN_SYSMACROS to find + where to get major, minor and makedev. + * create.c, list.c, update.c: Don't check USG to include + sys/sysmacros.h. + +Thu Dec 12 21:57:10 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * mangle.c (extract_mangle): Correctly null terminate name of + link target. + +Thu Nov 21 07:44:18 1991 Michael I Bushnell (mib at nutrimat) + + * create.c (dump_file, at start of ISREG output loop): use + filename from header instead of real name to make sure that we + get the mangled version and not one that is too long and + overflows buffers. + +Sat Nov 16 01:37:45 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * tar.h: Use new criteria for STDC version of msg. + +Sat Nov 2 21:31:57 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * create.c, gnu.c, tar.c: Use DIRENT instead of NDIR to select + between dirent.h and ndir.h for USG. + + * port.c: Rename WANT_FOO to FOO_MISSING to make sharing code + and configure script with other utilities easier. Use + VPRINTF_MISSING and DOPRNT_MISSING instead of FOO_MSG to + select error reporting routines. + +Thu Oct 17 20:19:02 1991 Michael I Bushnell (mib at churchy.gnu.ai.mit.edu) + + * level-0: Repair damage from previous mod: stdin to rsh must + be the terminal or tar's questions lose. + +Sat Aug 31 15:05:27 1991 Noah Friedman (friedman at nutrimat.gnu.ai.mit.edu) + + * level-0: Fixed several syntax errors associated with + stdout/stderr redirection. + Made sure remote host executes commands from sh where redirection + is necessary, since root's shell might be csh in some places and + the redirect syntax differs. + +Thu Aug 29 00:54:01 1991 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * tar.c (long_options). Fixed info-script long option. + +Mon Aug 26 16:53:50 1991 David J. MacKenzie (djm at pogo.gnu.ai.mit.edu) + + * configure, Makefile.in: Only put $< in Makefiles if VPATH + is being used, because older makes don't understand it. + +Mon Aug 19 01:47:57 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * create.c: Indent '#pragma alloca' so non-ANSI compilers + don't choke on it. + +Wed Aug 14 14:10:43 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * list.c (UGSWIDTH): Increase from 11 (sort of like Unix tar) to + 18, so that with normal user and group names of <= 8 chars, + the columns never shift in a tar -t listing. + +Fri Aug 2 00:41:08 1991 David J. MacKenzie (djm at apple-gunkies) + + * Makefile.in (dist): Include texinfo.tex and tar.info*. + (install): Install tar.info*. + * configure: Set INSTALLDATA. + + * configure: Create config.status. Remove it and Makefile if + interrupted while creating them. + + * configure: Check for +srcdir etc. arg and look for + Makefile.in in that directory. Set VPATH if srcdir is not `.'. + * Makefile.in: Add `prefix'. + (tar.info): New target. + +Tue Jul 30 17:08:04 1991 David J. MacKenzie (djm at apple-gunkies) + + * configure: NEED_TZSET has become FTIME_MISSING. + +Mon Jul 29 19:23:10 1991 David J. MacKenzie (djm at wombat.gnu.ai.mit.edu) + + * port.c [F_CHSIZE]: Additional version. + +Sat Jul 27 22:27:47 1991 David J. MacKenzie (djm at wombat.gnu.ai.mit.edu) + + * rmt.h: Clean up ifdefs. + + * makefile.pc: Fix typo. + port.h: Change MSDOS to __MSDOS__. + [__MSDOS__]: Define off_t. Include io.h and not sys/param.h. + [__TURBOC__]: Use void * and don't define const. + +Fri Jul 26 01:12:58 1991 David J. MacKenzie (djm at bleen) + + * buffer.c: Rename `eof' to `hit_eof' to avoid conflict with an + MSDOS function. + * gnu.c (get_dir_contents): Return NULL, not "\0\0\0\0", on error. + * diffarch.c (diff_archive): Open files in binary mode. + Don't use or free a non-malloc'd return value from get_dir_contents. + * msd_dir.c [__TURBOC__]: Include stdlib.h. + * rmt.h: lseek returns off_t, not long. + + * tar.c (describe): -X is +exclude-from, not +exclude. + (names_notfound): Free memory only if amiga, not !unix. + + * tar.h, tar.c: Add +null option to make -T read + null-terminated filenames (such as those produced by GNU find + -print0), and disable -C option. + This guarantees that odd filenames will get archived. + * tar.c (read_name_from_file): New function. + (name_next): Call it instead of fgets. + +Wed Jul 24 11:17:48 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * create.c [_AIX]: Declare alloca. + + * buffer.c (open_archive): Check for successful open before, + not after, fstatting the fd. + +Tue Jul 23 20:51:31 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * configure: Only define BSD42 if sys/file.h exists. + If alloca is missing and /usr/ucblib exists (SVR4), use it + instead of -lPW. + + * port.h [!__STDC__]: #define const. + * gnu.c (dirent_cmp): Fix args to agree with ANSI C prototype. + * create.c: Declare ck_realloc. + * gnu.c, diffarch.c: Move check for symlinks to after port.h include. + +Sat Jul 20 00:03:54 1991 David J. MacKenzie (djm at apple-gunkies) + + * msd_dir.[ch]: Use POSIX-style `struct dirent' instead of + `struct direct'. + * create.c, gnu.c, tar.c: Adjust callers. + +Thu Jul 18 00:05:01 1991 David J. MacKenzie (djm at bleen) + + * port.c (ck_malloc, ck_realloc): Return PTR, not char *. + * gnu.c, create.c, tar.c: Fix decls. + + * port.c: Don't use the preprocessor to guess missing + functions on Unix; let configure do it. + [WANT_GETWD] (getwd): Function removed; not needed because + getcwd is used if needed. + * gnu.c, tar.c: Use getcwd if POSIX. + + * rtapelib.c: Use SIGTYPE instead of testing SIGNAL_VOID. + Default to void (more common these days) instead of int. + + * tar.c, gnu.c, mangle.c: Remove VOIDSTAR defn. Use PTR instead. + * port.h: Define PTR. + + * gnu.c, tar.c [__MSDOS__ || USG]: Remove incorrect getcwd + decl; put correct one in port.h [!POSIX]. + + * tar.c (describe): Print on stdout instead of stderr; it's + not so much a usage message (since you have to ask for it + explicitly) as on-line help, and you really need to be able to + page it because it's more than a screen long. + + * Make #ifdefs for sys/file.h or fcntl.h, directory header, + sys/mtio.h consistent between files. Use NO_MTIO instead of + tricks with USG and HAVE_MTIO and NO_RMTIOCTL. + * Move decls of ANSI C and POSIX functions to port.h and + use standard headers to declare them if available + [STDC_HEADERS or POSIX]. + * Add many missing function declarations and return types. + * Some places used __MSDOS__, some MSDOS; standardize on __MSDOS__. + * Change S_IF macros to S_IS for POSIX. + * port.h: Define appropriate S_IS macros if missing. + * port.h: Rename macros for testing exit status to conform to + POSIX; use the system's versions if available [POSIX]. + * Use POSIX PATH_MAX and NAME_MAX instead of MAXPATHLEN and MAXNAMLEN. + * port.h: Define PATH_MAX and NAME_MAX. + * create.c, gnu.c, tar.c: Use ck_malloc and free instead of + auto arrays of size PATH_MAX or NAME_MAX, since with pathconf + they might not be constants. + * Move all definitions of O_* to port.h to reduce redundancy. + * Make all source files that now need to include port.h do so. + * port.c: Remove #undefs of WANT_* so you can use -DWANT_* + when compiling, instead of having to edit port.c. + [WANT_DUMB_GET_DATE] (get_date): Function removed. + Even systems without bison can get bison output and compile it. + [WANT_STRING] (index, rindex, bcopy, bzero, bcmp): Functions + removed; the translation is now done by macros in port.h. + * wildmat.c (wildmat): Use POSIX.2 '!' instead of '^' to negate + character classes. + +Mon Jul 15 13:47:45 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * testpad.c (main): Return type void. + + * port.c [WANT_STRING]: Don't include memory.h if NO_MEMORY_H. + + * create.c (dump_file) [AIX]: Fix typo, `allocate' for `alloca'. + * gnu.c (collect_and_sort_names): Move misplaced brace out of #ifdef. + From: Minh Tran-Le <TRANLE@intellicorp.com>. + + * configure: Also look in sys/signal.h for signal decl. + +Wed Jul 10 01:42:55 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * Rename rtape_server.c to rmt.c and rtape_lib.c to rtapelib.c. + + * configure, Makefile.in: $(INSTALLPROG) -> $(INSTALL). + +Tue Jul 9 01:38:37 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) + + * Most files: Refer to GPL version 2. + * COPYING: Use version 2. + + * port.c [__TURBOC__] (utime): New function. + + * xmalloc: New function (just calls ck_malloc), for alloca.c + and bison.simple (in getdate.y output). + + * Makefile.in (AUX): Include alloca.c and tcexparg.c, a + command line globber for Turbo C. + +Mon Jul 8 14:30:52 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * testpad.c: Open and write to testpad.h instead of stdout, + because some MS-DOS makes (Borland's at least) can't do + redirection in commands. + * Makefile.in: Don't redirect testpad output. + +Mon Jul 8 12:56:35 1991 Michael I Bushnell (mib at churchy.gnu.ai.mit.edu) + + * buffer.c (fl_read): Missing \n in printf. + +Mon Jul 8 03:40:28 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * create.c, extract.c, gnu.c, diffarch.c, tar.c: Comment out + unused variables. + + * tar.c (options): Cast get_date arg to VOIDSTAR instead of + `struct timeb *', since on some non-BSD systems the latter is + undefined. + +Sat Jul 6 04:53:14 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * Replace Makefile with configure, Makefile.in, and makefile.pc. + Update README with current compilation instructions. + + * port.c [WANT_RENAME] (rename): New function. + +Wed Jul 3 18:10:52 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * testpad.c (main): Avoid warning from some compilers on array + address. + + * rtape_server.c (sys_errlist): Should be declared extern. + +Mon Jul 1 14:14:06 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * Release of version 1.10; appropriate changes to README. + + * create.c: Removed printf's about sparse files. + + * Fix a misplaced quote in level-0 and change some >& into + 2>&1. + +Fri Jun 21 23:04:31 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * list.c (skip_extended_headers): Userec was being called in + the wrong place. + +Thu Jun 20 19:10:35 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) + + * tar.h: Use ANSI prototypes for msg and msg_perror if + STDC_MSG is defined, even if BSD42 is also. + + * Makefile: Replace DESTDIR with bindir. + (install): Don't install tar.texinfo. There's no standard + place for texinfo files, and /usr/local/man is inappropriate. + Add TAGS, distclean, and realclean targets and SHELL= line. + + * version.c: Move old change history to bottom of ChangeLog. + +Wed Jun 12 12:43:58 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * rtape_lib.c (__rmt_write): #ifdef should reference + SIGNAL_VOID, not USG. + +Wed Jun 5 14:57:11 1991 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * tar.c (name_match, addname): Ugly hack to handle -C without + any files specified. + tar.h (struct name): New field for ugly hack. + +Mon Jun 3 14:46:46 1991 Michael I Bushnell (mib@geech.gnu.ai.mit.edu) + + * testpad.c: New file to determine if we need special padding + in struct header in tar.h. + + * tar.h (struct header): include padding if necessary, include + testpad.h. + + * Makefile: rules to create testpad.h, etc. + +Wed May 22 16:02:35 1991 Michael I Bushnell (mib@churchy.gnu.ai.mit.edu) + + * tar.c (options): -L takes an argument. + + * rtape_lib.c (__rmt_open): add /usr/bin/nsh to the list of + remote shell programs. + + * create.c: define MAXPATHLEN if we don't get it from a system + header file. + + * create.c (deal_with_sparse): return a real return value if + we can't open the file. + + * tar.c (long_options): +newer takes an argument. + (describe): fix printing in various trivial ways + +Tue May 21 17:15:19 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * tar.c (long_options): +get and +concatentate don't require arguments + +Mon May 20 15:55:30 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * create.c (write_eot): Don't try and write an EOF if we are + already at one. + + * port.c (strstr): Looking for null string should return zero. + +Sun May 19 22:30:10 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * tar.c (options): -l doesn't take an argument + + * Makefile: minor fix for SGI 4D defines from torda@scum.ethz.ch + + * rtape_server.c (main.c): Suggested mod for 386/AIX from + Minh Tran-Le. I'm suspicious about this one. + + * create.c (dump_file): Mods from Minh Tran-Le for hidden + files on AIX. + gnu.c (collect_and_sort_name, get_dir_contents): AIX hidden file mod. + + * tar.c: (name_next): Mod from David Taylor to allow -C inside + a file list given to -T. + + * Makefile: Comment describing presence of USE_REXEC. + + * extract.c (extract_archive, case LF_SPARSE): zero check for + last element on numbytes needs to look at value after + converted from octal. + + * port.c: Don't always demand strstr, check for HAVE_STRSTR + instead. + Makefile: Comment describing presence of HAVE_STRSTR option. + +Sun May 19 18:39:48 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) + + * port.c (get_date): Renamed from getdate, to avoid SVR4 conflict. + * tar.c: Call get_date instead of getdate. + +Fri May 10 02:58:17 1991 Noah Friedman (friedman at nutrimat) + + * tar.c: added "\n\" to the end of some documentation strings + where they were left off. + +Thu May 9 17:28:54 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * Makefile: added level-0, level-1, and backup-specs to AUX. + * version.c: changed to 1.10 beta. + * README: updated for 1.10 beta release. + +Tue Apr 2 12:04:54 1991 Michael I Bushnell (mib at godwin) + + * create.c (dump_file): HPUX's st_blocks is in 1024 byte units + instead of 512 like the rest of the world, so I special cased + it. + * tar.c: Undo Noah's changes. + +Mon Apr 1 17:49:28 1991 Noah Friedman (friedman at wookumz.gnu.ai.mit.edu) + + (This ought to be temporary until things are fixed properly. ) + + * tar.c: (struct option long_options): flag for "sparse" zero if + compiling under hpux. + tar.c: (functon options): case 'S' is a no-op if compiling under + hpux. + +Sat Mar 30 12:20:41 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * tar.h: new variable tape_length. + + * tar.c (options): add new option +tape-length / -L. + + * buffer.c (fl_write): Turn #ifdef TEST code for limited tape + length on always, for tape-length option. + + * create.c (dump_file): avoid apollo lossage where S_IFIFO == S_IFSOCK. + + * buffer.c: include regex.h + * buffer.c (fl_read, open_archive): Use regex routines for + volume header match. + * xmalloc.c: removed file; wasn't necessary. + * tar.c: (main) use ck_malloc instead of xmalloc. + +Thu Mar 28 04:05:05 1991 Noah Friedman (friedman at goldman) + + * regex.c, regex.o: New links. + * tar.c: include regex.h. + * Makefile (OBJ2): Add regex.o. + (regex.o, tar.o): Depend on regex.h + (SRC2, AUX): Add the new files. + +Sat Mar 23 15:39:42 1991 Noah Friedman (friedman at wookumz.gnu.ai.mit.edu) + + * Makefile: added default flags and options for compiling under + hpux. + + * Added files alloca.c and xmalloc.c + +Sat Mar 23 14:35:31 1991 Michael I Bushnell (mib at geech.gnu.ai.mit.edu) + + * port.c: Define WANT_VALLOC in HPUX. + +Fri Mar 15 06:20:15 1991 David J. MacKenzie (djm at geech.ai.mit.edu) + + * rtape_lib.c: If USG and not HAVE_MTIO, define NO_RMTIOCTL + automatically. + (_rmt_rexec): Temporarily re-open stdin and stdout to + /dev/tty, to guarantee that rexec() can prompt and read the + login name and password from the user. + From pascal@cnam.cnam.fr (Pascal Meheut). + * Makefile: Mention -DUSE_REXEC. + +Fri Mar 8 20:15:11 1991 Michael I Bushnell (mib at wookumz.ai.mit.edu) + + * tar.h, Makefile: Makefile CPP macro HAVE_SIZE_T might be + useful for some people. + + * gnu.c: lstat->stat define where appropriate + + * buffer.c (fl_write): keep track of amount written for +totals. + * tar.c, tar.h: set flag f_totals from +totals option + * tar.h (f_totals, tot_written): new variables + * tar.c (main): print total written with CMD_CREATE + + * tar.c (main): return appropriate exit status + +Thu Jan 17 00:50:21 1991 David J. MacKenzie (djm at apple-gunkies) + + * port.c: Remove a spurious `+' between functions (a remnant + of a context diff, apparently). + +Wed Jan 9 19:43:59 1991 Michael I Bushnell (mib at pogo.ai.mit.edu) + + * create.c (where_is_data): Rewritten to be better, and then + #ifdef-ed out. + (deal_with_sparse): Severly pruned. Now we write or don't + write only complete blocks, not worrying about partial blocks. + This simplifies calculations, removes bugs, and elides the + second scan through the block. The first was zero_record, the + second was where_is_data. + +Mon Jan 7 17:13:29 1991 Michael I Bushnell (mib at wookumz.ai.mit.edu) + + * create.c (deal_with_sparse): Second computation (for short + reads) of numbytes increment had subtraction backwards. + Need to handle calling where_is_data better when we did a + short read (it might go past the end of the read), also, set + sparsearray[...].offset in this case too. + +Fri Jan 4 12:24:38 EST 1991 Jay Fenlason (hack@ai.mit.edu) + + * buffer.c Return a special error code if the archive you're + trying to read starts with a different label than the one specified + on the command line. + +Wed Jan 2 12:05:21 EST 1991 Jay Fenlason (hack@ai.mit.edu) + + * gnu.c Prepend the current directory to the gnu_dumpfile, so that + -C's won't affect where the output goes. (sigh.) + +Tue Dec 18 18:05:59 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * (gnu.c) Don't complain if the gnudumpfile we're reading info + from doesn't exist. + + * create.c Write out gnudumpfile after finishing writing the archive. + + * tar.c Add +exclude FNAME, and make +exclude-from do what +exclude + used to. + + Make +version an operation, not an option. + + add +confirmation alias for +interactive. + +Tue Dec 4 13:28:08 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * tar.c (check_exclude) Don't let MUMBLE match MUMBLE.c or fooMUMBLE + but only foo/MUMBLE + + * Add the name mangler (mangle.c, plus changes to create.c and + extract.c) + + * extract.c Three small patches from Chip Salzenberg + (tct!chip@uunet.uu.net) + + Don't complain when extracting a link, IFF it already exists. + + Don't complain when extracting a directory IFF it already + exists. + + Don't ad u+wx to directories when running as root. + + * gnu.c Some changes from Chip Salzenberg to make + +listed-incremental work. + + * port.c Add the F_FREESP emulation of the ftruncate syscall. + +Wed Nov 21 15:57:07 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + Remove excess \n from lots of msg() calls. + +Mon Nov 19 14:09:43 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * tar.c Rename +volume to +label + +Fri Nov 16 15:43:44 1990 David J. MacKenzie (djm at apple-gunkies) + + * tar.c (describe): Include the default values for -b and -f + (as set in the Makefile) in the message. + +Thu Nov 15 13:36:45 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * extract.c (extract_archive) Do the utime() call before the + chmod() call, 'cuz some versons of utime() trash the file's mode + bits. + + * list.c (read_and) Call do_something on volume headers and + multivol files even if they don't match the names we're looking for, + etc. . . + +Tue Nov 6 13:51:46 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * port.c (un-quote-string) Don't try to write a null + if there's already one there. + +Thu Nov 1 14:58:57 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * buffer.c (new_volume) fflush(msg_file) before reading for + confirmation on new volume. On EOF or error, print error msg and + abort. + +Mon Oct 29 12:06:35 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * getdate.y Use new version of getdate(). + + * tar.c (name_add) Use sizeof(char *) instead of sizeof(int) + + * README give the correct return address. + +Thu Oct 25 16:03:58 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + rtape_lib.c Change RMTIOCTL to NO_RMTIOCTL, so it is on by default. + + rmt.h Add _isrmt() #define for NO_REMOTE case. + + gnu.c Add forward reference for add_dir_name(). + +Tue Oct 16 11:04:52 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + 1.09 New -G file implementation of gnu-dump stuff. + + * tar.c (name_add) Get the calls to ck_realloc and ck_malloc right. + +Thu Oct 11 11:23:38 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * gnu.c Fix A couple of typos. + +Wed Sep 19 13:35:03 1990 David J. MacKenzie (djm at apple-gunkies) + + * getdate.y [USG] (ftime): Use `daylight' unless + DAYLIGHT_MISSING is defined. + +Mon Sep 17 18:04:21 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * gnu.c (gnu_restore) Don't use a passed char* for the + file name, use skipcrud+head->header.name, just like everything + else does. This means that gnu_restore will still work with + small buffers, etc. + +Thu Sep 13 15:01:17 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * tar.c (add_exclude) Don't bus-error if the exclude file doesn't + end with a newline. + +Sun Sep 9 22:35:27 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * Makefile (dist): Remove .fname when done. + +Thu Sep 6 12:48:58 EDT 1990 Jay Fenlason (hack@ai.mti.edu) + + * gnu.c (gnu_restore) Rember to skip_file() over the directory + contents, even if we don't have to do anything with them. + + * create.c extract.c diffarch.c Free sparsearray after we're done + with it. + +Tue Sep 4 10:18:50 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * Makefile Include gnu.c in dist + + * gnu.c move add_dir above read_dir_file so that cc doesn't complain + about add_dir returning void. + +Sun Sep 2 20:46:34 1990 David J. MacKenzie (djm at apple-gunkies) + + * getdate.y: Declare some more functions and add storage + classes where omitted to shut compiler up. + [USG] (ftime): Don't use extern var `daylight'; appears that + some systems don't have it. + +Wed Aug 29 00:05:06 1990 David J. MacKenzie (djm at apple-gunkies) + + * getdate.y (lookup): In the code that allows `Aug.' to be + recognized as `Aug', don't chop off the final `.' from words + like `a.m.', so they can be recognized. + +Thu Aug 16 11:34:07 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * buffer.c (open_archive) If -O, write verbosity to stderr + instead of stdout. + +Fri Aug 10 12:29:28 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * getdate.y Handle an explicit DST in the input string. + A dozen line patch from Per Foreby (perf@efd.lth.se). + +Mon Jul 16 13:05:11 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * tar.c rename -g -G +incremental, +listed-imcremental, etc. + +Fri Jul 13 14:10:33 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * tar.c Make +newer and +newer-mtime work according to their names. + + * gnu.c If +newer or +newer-mtime, use the time specified on the + command line. + + * buffer.c, create.c Add test to see if dimwit is trying to + archive the archive. + + * tar.c (long_options[]) re-ordered, so that groups of similar + options are next to each other. . . I think. + + (describe) Modified to more closely reflect reality. + +Fri Jul 6 13:13:59 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * tar.c add compile-time option for SYS V (?) style + tape-drive names /dev/rmt/{n}[lmh] + + * tar.c Fix getopt-style stuff so that -C always works correctly. + + * gnu.c, tar.c make filename to -G optional. + + * {all over}, replace some fprintf(stderr...) calls with calls + to msg(). + + * port.c Make -Dmumble_MSG option on command line override + internal assumptions. + + * Makefile Mention -Dmumble_MSG options + +Fri Jul 6 02:35:31 1990 David J. MacKenzie (djm at apple-gunkies) + + * tar.c (options): Don't change `c' if it is 0, as getopt now + handles that internally. + +Mon Jul 2 15:21:13 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * gnu.c (new file) Moved all the f_gnudump stuff here where we + can keep track of it easier. Also made -G take a file name where it + stores the inode information about directories so that we can + detect moved directores. + + * create.c (dump_file) Changed slightly to work with the new + f_gnudump. + + * tar.c Moved the f_gnudump stuff to gnu.c + + * tar.c, extract.c added the +do-chown option, which forces tar + to always try to chown the created files to their original owners. + + * version.c New version 1.09 + +Sun Jun 24 14:26:28 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * create.c: Change ifdefs for directory library header + selection to be like the ones in tar.c. + * Makefile [Xenix]: Link with -ldir to get the dirent.h + directory library. + +Thu Jun 7 03:31:51 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * Makefile, buffer.c, diffarch.c: Change MTIO symbol to HAVE_MTIO + because SCO Xenix defines 'MTIO' for an incompatible tape driver + system in a file included by termio.h. + * tar.h: Don't define size_t for Xenix. + +Tue Jun 5 11:38:00 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * create.c (dump_file) Only print the + "... is on a different filesystem..." if f_verbose is on. + also add a case for S_IFSOCK and treat it like a FIFO. + (Not sure if that's the right thing to do or not, but it's better + than all those Unknown File Type msgs.) + +Thu May 31 19:25:36 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * port.c Use #ifdef sparc instead of #ifdef SPARC since + the lowercase version is defined, and the uppercase one isn't. + +Tue May 22 11:49:18 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * port.c (ck_malloc) if size==0 pretend size=1 + (ck_realloc) if(!ptr) call ck_malloc instead. + +Tue May 15 12:05:45 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * diffarch.c (diff_archive) If not f_absolute_paths, and attempt to + open a file listed in the archive fails, try /filename also. This will + allow diff to open the wrong file if both /filename and filename exist, + but there's nothing we can do about that. + +Fri May 11 16:17:43 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * Makefile, Descripbe new -DMTIO option. + + * buffer.c diffarch.c Change ifdefs slightly, so that + -DMTIO will include sys/mtio.h even if USG is defined. + This is for HUPX and similar BSD/USG crossovers. + +Tue May 8 13:14:54 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + * update.c (update_archive) Call reset_eof() when appropriate. + + * buffer.c (reset_eof) New function, that turns of EOF flag, and + re-sets the ar_record and ar_last pointers. This will allow + 'tar rf non-existant-file' to not core-dump. + +Fri May 4 14:05:31 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * tar.c: Recognize the +sparse option. It was documented, but + only the short form (-S) was actually recognized. + +Tue Apr 17 21:34:14 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * create.c Don't access location 0 if ->dir_contents is null. + +Wed Apr 11 17:30:03 EDT 1990 Jay Fenlason (hack@ai.mit.edu) + + * buffer.c (flush_archive, close_archive, new_volume) Always check + the return value of rmtclose(), and only give a warning msg if it is + <0. Some device drivers (including Sun floppy disk, and HP + streaming tape) return -1 after an IO error (or something like that.) + +Fri Mar 23 00:06:30 1990 Jim Kingdon (kingdon at mole.ai.mit.edu) + + * tar.c (long_options): Make it so +append +extract +list +update + +catenate and +delete don't take arguments. + +Mon Mar 12 13:33:53 EST 1990 + + * buffer.c (open_archive, fl_write) Set the mtime of the volume + header to the current time. + +Wed Mar 7 14:10:10 EST 1990 Jay Fenlason (hack@ai.mit.edu) + + * buffer.c Fix +compress-block A two character patch from + Juha Sarlin (juha@tds.kth.se) + Replace #ifdef __GNU__ with #ifdef __STDC__ + (new_volume) If open of new archive fails, ask again + (Is probably user error.) + + * tar.c Replace #ifdef __GNU__ with #ifdef __STDC__ + + * port.c Clean up #ifdef and #defines a bit. + (quote_copy_string) Sometimes the malloc'd buffer + would be up to two characters too short. + + * extract.c (extract_archive) Don't declare ind static. + + * create.c (dump_file) Don't declare index_offset static. + + * diffarch.c Remove diff_name variable, and always use + head->header.name, which will always work, unlike diff_name, which + becomes trash when the next block is read in. + +Thu Mar 1 13:43:30 EST 1990 Jay Fenlason (hack@wookumz.ai.mit.edu) + + * Makefile Mention the -NO_REMOTE option. + * port.c Fix typo, and define WANT_FTRUNCATE on i386 machines. + +Mon Feb 26 17:44:53 1990 Jim Kingdon (kingdon at pogo.ai.mit.edu) + + * getdate.y: Declare yylex and yyerror as static. + #define yyparse to getdate_yyparse. + +Sun Feb 25 20:47:23 1990 David J. MacKenzie (djm at albert.ai.mit.edu) + + * tar.c: Remove +old option, since it is a valid abbreviation of + +old-archive, which does the same thing. + (describe): A few small cleanups in message. + +Mon Feb 5 14:29:21 EST 1990 Jay Fenlason (hack@wookumz) + + * port.c define LOSING_MSG on sparc, since doprnt_msg doesn't work. + Fix typo in #ifdef WANT_GETWD + +Fri Jan 26 16:11:20 EST 1990 Jay Fenlason (hack@wookumz) + + 1.08 Sparse file support added. Also various other features. + + * diffarch.c (compare_chunk) Include correct arguments in + a call to fprintf() for an error msg. + (compare_chunks, compare_dir) First argument is a long, not an int. + + * tar.c (options) Use tar variable (argv[0]) as the name to print + in an error msg, instead of a constant "tar". + (confirm) Use external variable char TTY_NAME[] for name of file + to open for confirmation input. + + * buffer.c (new_volume) Ditto. + + * port.c Add declaration for TTY_NAME[]. + + * rmt.h Add long declarations for lseek() and __rmt_lseek(); + +Tue Jan 23 14:06:21 EST 1990 Jay Fenlason (hack@wookumz) + * tar.c, create.c Create the +newer-mtime option, which is like + +newer, but only looks for files whose mtime is newer than the + given date. + + * rtape_lib.c Make *both* instances of signal-handler stuff use + void (*foo)() on USG systems. + +Thu Jan 11 14:03:45 EST 1990 Jay Fenlason (hack@wookumz) + + * getdate.y Parse European dates of the form YYMMDD. + In ftime() Init timezone by calling localtime(), and remember that + timezone is in seconds, but we want timeb->timezone to be in minutes. + This small patch from Joergen Haegg (jh@aahas.se) + + * rtape_lib.c (__rmt_open) Also look for /usr/bsd/rsh. + Declare signal handler as returning void instead of int if USG is + defined. + + * port.c Declare WANT_GETWD for SGI 4-D IRIS. + + * Makefile Include defines for SGI 4D version. There are a simple + patch from Mike Muuss (mike@brl.mil). + + * buffer.c (fl_read) Work properly on broken Ultrix systems where + read() returns -1 with errno==ENOSPC on end of tape. Correctly go + on to the next volume if f_multivol. + + * list.c (list_archive,print_header) Flush msg_file after printing + messages. + + * port.c Delete unused references to alloca(). + Don't crash if malloc() returns zero in quote_copy_string. + Flush stderr in msg() and msg_perror(). + + * tar.c Flush msg_file after printing confirmation msg. + +Wed Jan 10 01:58:46 1990 David J. MacKenzie (djm at hobbes.ai.mit.edu) + + * tar.c (main): Change -help option and references to it to +help, + and remove suggestion to run info (which is unreleased, so not + likely to be of any help). + +Tue Jan 9 16:16:00 EST 1990 Jay Fenlason (hack @wookumz) + + * create.c (dump_file) Close file descriptor if start_header() + fails. + (dump_file) Change test for ./ ness to not think that + .{any character} is a ./ These are both trivial changes from + Piercarlo "Peter" Grandi pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk + + * diffarch.c (diff_init) Print correct number of bytes in error + message. + +Tue Jan 9 03:19:49 1990 David J. MacKenzie (djm at hobbes.ai.mit.edu) + + * Makefile: Add comment at top noting that two source files also + contain #defines that might need to be changed by hand. + + * create.c, diffarch.c, extract.c: Change L_SET to 0 in lseek + calls, because only BSD defines it. + * create.c (dump_file): Make sparse file checking code conditional + on BSD42 because it uses st_blocks, which the other systems lack. + +Tue Jan 2 13:35:56 EST 1990 Jay Fenlason (hack@gnu) + + * port.c (quote_copy_string) Fix so it doesn't scramble memory if + the last character is non-printable. A trivial fix from Kian-Tat Lim + (ktl@wag240.caltech.edu). + +Tue Dec 19 11:19:37 1989 Jim Kingdon (kingdon at pogo) + + * port.c [BSD42]: Define DOPRNT_MSG. + tar.h [BSD42]: Do not prototype msg{,_perror}. + +Fri Dec 8 11:02:47 EST 1989 Jay Fenlason (hack@gnu) + + * create.c (dump_file) Remove typo in msg. + +Fri Dec 1 19:26:47 1989 David J. MacKenzie (djm at trix) + + * Makefile: Remove comments referring to certain systems lacking + getopt, since it is now provided always and needed by all systems. + + * port.c: Remove copy of getopt.c, as it is now linked in + separately to always get the current version. + + * tar.c: Rename +cat-tars option to +catenate or +concatenate, + and +local-filesystem to +one-file-system (preferred by rms + and used in GNU cp for the same purpose). + (describe): Reflect changes. + +Tue Nov 28 04:28:26 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) + + * port.c: Move declaration of alloca into #else /* sparc */ + so it will compile on sparcs. + +Mon Nov 27 15:17:08 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) + + * tar.c (options): Remove -version option (replaced by +version). + (describe): Mention long options. + +Sat Nov 25 04:25:23 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) + + * getoldopt.c (getoldopt): Make `opt_index' argument a pointer to + an int, not char. + + * tar.c: Modify long options per rms's suggestions: + Make preserve-permissions an alias for same-permissions. + Make preserve-order an alias for same-order. + Define preserve to mean both of those combined. + Make old an alias for old-archive. + Make portability an alias for old-archive, also. + Rename sym-links to dereference. + Rename gnudump to incremental. + Rename filename to file. + Make compare an alias for diff. Leave diff but prefer compare. + Rename blocking-factor to block-size. + Rename chdir to directory. + Make uncompress an alias for compress. + Rename confirm to interactive. + Make get an alias for extract. + Rename volume-header to volume. + + Also make +version an alias for -version. + + (options): Shorten code that interprets long options by using + the equivalent short options' code. This also makes it tons + easier to change the long options. + + (describe): Make usage message more internally consistent + stylistically. + +Mon Nov 20 14:55:39 EST 1989 hack@ai.mit.edu + + * list.c (read_and) Call check_exclude() to see if the files + should be skipped on extract or list. + +Thu Nov 9 18:59:32 1989 Jim Kingdon (kingdon at hobbes.ai.mit.edu) + + * buffer.c (fl_read): Fix typos in error message + "tar EOF not on block boundary". + +Mon Oct 23 13:09:40 EDT 1989 (hack@ai.mit.edu) + + * tar.c (long_options[]) Add an option for blocked compression. + +Thu Oct 19 13:38:16 EDT 1989 (hack@ai.mit.edu) + + * buffer.c (writeerror) Print a more useful error msg. + +Wed Sep 27 18:33:41 EDT 1989 (hack@ai.mit.edu) + + * tar.c (main) Mention "tar -help" if the luser types a non-workable + set of options. + +Mon Sep 11 15:03:29 EDT 1989 (hack@ai.mit.edu) + + * tar.c (options) Have -F correctly set info_script. + +Tue Aug 29 12:58:06 EDT 1989 (hack@ai.mit.edu) + + * Makefile Include ChangeLog in tar.tar and tar.tar.Z + +Mon Aug 28 17:42:24 EDT 1989 (hack@ai.mit.edu) + + * tar.c (options) Made -F imply -M + Also remind tar that the -f option takes an argument! + + * Modified -F option to make it do what (I think) it + should. e.g, if you say -F, tar won't send a msg to + msg_file and wait for a <return> It'll just run the program + it was given, and when the prog returns, the new tape had + *better* be ready. . . + + * buffer.c (open_archive) Give error message and abort if + the luser didn't give an archive name. + +Fri Aug 25 20:05:27 EDT 1989 Joy Kendall (jak at hobbes) + + * Added code to make a new option to run a specified script + at the end of each tape in a multi-volume backup. Changed: + tar.c: made new switch, -F, and new long-named option, + "info-script". Code is where you would expect. + tar.h: added flag f_run_script_at_end, and an extern char * + called info_script, which optarg gets set to. + buffer.c: line 1158 in new_volume(): if f_run_script_at_end + is set, we give info_script to system(), otherwise we do + what we've always done. **FIXME** I'm not sure if that's all + that has to be done here. + +Thu Aug 24 10:09:38 EDT 1989 Joy Kendall (jak at spiff) +(These changes made over the course of 6/89 - 8/89) + + * diffarch.c: diff_archive: Added switches for LF_SPARSE in the + case statements that needed it. Also, skip any extended headers + if we need to when we skip over a file. (need to change + the bit about, if the size doesn't agree AND the file is NOT + sparse, then there's a discrepancy, because I added another + field to the header which should be able to deal with the + sizes) If the file is sparse, call the added routine + "diff_sparse_files" to compare. Also added routine + "fill_in_sparse_array". + + * extract.c: extract_archive: added the switch LF_SPARSE + to the case statement as needed, and code to treat the + sparse file. At label "again_file", modified opening the + file to see if we should have O_APPEND be one of the modes. + Added code at label "extract_file" to call the new routine + "extract_sparse_file" when we have an LF_SPARSE flag. + + Note: really should erase the commented-out code in there, + because it's confusing. + + * update.c: made sure that if a file needed to be "skipped" + over, it would check to see if the linkflag was sparse, and + if so, would then make sure to skip over any "extended + headers" that might come after the header itself. Do so by + calling "skip_extended_headers". + + * create.c: create_archive: added code to detect a sparse + file when in the long case statement. Added ways to detect + extended headers, and label "extend" (ack! should get rid of + that, is atrocious). Call the new routine "finish_sparse_file" + if the linkflag is LF_SPARSE to write the info to the tape. + Also added routines "init_sparsearray", "deal_with_sparse", + "clear_buffer", "where_is_data", "zero_record", and + "find_new_file_size". + + * tar.h: Added the #define's SPARSE_EXT_HDR and + SPARSE_IN_HDR. Added the struct sparse and the struct + sp_array. Added the linkflag LF_SPARSE. Changed the tar + header in several ways: + - added an array of struct sparse's SPARSE_IN_HDR long + - added a char flag isextended + - added a char string realsize to store the true + size of a sparse file + Added another choice to the union record called a + struct extended_header, which is an array of 21 struct + sparse's and a char isextended flag. Added flag + f_sparse_file to list of flags. + + * tar.c: added long-named options to make tar compatible with + getopt_long, changed Makefile. + +... ... .. ..:..:.. ... .... Jay Fenlason (hack@ai.mit.edu) + + 1.07 New version to go on beta tape with GCC 1.35 + Better USG support. Also support for __builtin_alloca + if we're compiling with GCC. + diffarch.c: Include the correct header files so MTIOCTOP + is defined. + tar.c: Don't print the verbose list of options unless + given -help. The list of options is *way* too long. + + 1.06 Use STDC_MSG if __STDC__ defined + ENXIO meand end-of-volume in archive (for the UNIX PC) + Added break after volume-header case (line 440) extract.c + Added patch from arnold@unix.cc.emory.edu to rtape_lib.c + Added f_absolute_paths option. + Deleted refereces to UN*X manual sections (dump(8), etc) + Fixed to not core-dump on illegal options + Modified msg_perror to call perror("") instead of perror(0) + patch so -X - works + Fixed tar.c so 'tar cf - -C dir' doesn't core-dump + tar.c (name_match): Fixed to chdir() to the appropriate + directory if the matching name's change_dir is set. This + makes tar xv -C foo {files} work. + + 1.05 A fix to make confirm() work when the archive is on stdin + include 'extern FILE *msg_file;' in pr_mkdir(), and fix + tar.h to work with __STDC__ + + Added to port.c: mkdir() ftruncate() Removed: lstat() + Fixed -G to work with -X + Another fix to tar.texinfo + Changed tar.c to say argv[0]":you must specify exactly ... + buffer.c: modified child_open() to keep tar from hanging when + it is done reading/writing a compressed archive + added fflush(msg_file) before printing error messages + create.c: fixed to make link_names non-absolute + + 1.04 Added functions msg() and msg_perror() Modified all the + files to call them. Also checked that all (I hope) + calls to msg_perror() have a valid errno value + (modified anno() to leave errno alone), etc + Re-fixed the -X option. This time for sure. . . + re-modified the msg stuff. flushed anno() completely + Modified the directory stuff so it should work on sysV boxes + added ftime() to getdate.y + Fixed un_quote_string() so it won't wedge on \" Also fixed + \ddd (like \123, etc) + More fixes to tar.texinfo + + 1.03 Fixed buffer.c so 'tar tzf NON_EXISTENT_FILE' returns an error + message instead of hanging forever + More fixes to tar.texinfo + + 1.02 Fixed tar.c so 'tar -h' and 'tar -v' don't cause core dump + Also fixed the 'usage' message to be more up-to-date. + Fixed diffarch.c so verify should compile without MTIOCTOP + defined + + 1.01 Fixed typoes in tar.texinfo + Fixed a bug in the #define for rmtcreat() + Fixed the -X option to not call realloc() of 0. + + Version 1.00: version.c added. -version option added + Installed new version of the remote-tape library + Added -help option + +Local Variables: +mode: indented-text +left-margin: 8 +version-control: never +End: diff --git a/gnu/usr.bin/tar/Makefile b/gnu/usr.bin/tar/Makefile new file mode 100644 index 000000000000..810fe3b7a74d --- /dev/null +++ b/gnu/usr.bin/tar/Makefile @@ -0,0 +1,14 @@ +PROG= tar +SRCS= buffer.c create.c diffarch.c extract.c fnmatch.c getdate.y \ + getoldopt.c getopt.c getopt1.c gnu.c list.c mangle.c names.c port.c \ + regex.c rtapelib.c tar.c update.c version.c +CFLAGS+= -DRETSIGTYPE=void -DDIRENT=1 -DHAVE_SYS_MTIO_H=1 -DHAVE_UNISTD_H=1 +CFLAGS+= -DHAVE_GETGRGID=1 -DHAVE_GETPWUID=1 -DHAVE_STRING_H=1 +CFLAGS+= -DHAVE_LIMITS_H=1 -DHAVE_STRSTR=1 -DHAVE_VALLOC=1 -DHAVE_MKDIR=1 +CFLAGS+= -DHAVE_MKNOD=1 -DHAVE_RENAME=1 -DHAVE_FTRUNCATE=1 -DHAVE_GETCWD=1 +CFLAGS+= -DHAVE_VPRINTF=1 -DNEEDPAD -I${.CURDIR} +CFLAGS+= -DDEF_AR_FILE=\"/dev/rst0\" -DDEFBLOCKING=20 +NOMAN=noman + +.include <bsd.prog.mk> +.include "../../usr.bin/Makefile.inc" diff --git a/gnu/usr.bin/tar/Makefile.gnu b/gnu/usr.bin/tar/Makefile.gnu new file mode 100644 index 000000000000..a03617ae3666 --- /dev/null +++ b/gnu/usr.bin/tar/Makefile.gnu @@ -0,0 +1,185 @@ +# Generated automatically from Makefile.in by configure. +# Un*x Makefile for GNU tar program. +# Copyright (C) 1991, 1992, 1993 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2, or (at your option) +# any later version. + +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. + +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + +#### Start of system configuration section. #### + +srcdir = . +VPATH = . + +# If you use gcc, you should either run the fixincludes script that +# comes with it or else use gcc with the -traditional option. Otherwise +# ioctl calls will be compiled incorrectly on some systems. +CC = gcc +YACC = bison -y +INSTALL = /usr/local/bin/install -c +INSTALL_PROGRAM = $(INSTALL) +INSTALL_DATA = $(INSTALL) -m 644 + +# Things you might add to DEFS: +# -DSTDC_HEADERS If you have ANSI C headers and libraries. +# -DHAVE_UNISTD_H If you have unistd.h. +# -DHAVE_STRING_H If you don't have ANSI C headers but have string.h. +# -DHAVE_LIMITS_H If you have limits.h. +# -DBSD42 If you have sys/dir.h (unless you use -DPOSIX), +# sys/file.h, and st_blocks in `struct stat'. +# -DDIRENT If you have dirent.h. +# -DSYSNDIR Old Xenix systems (sys/ndir.h). +# -DSYSDIR Old BSD systems (sys/dir.h). +# -DNDIR Old System V systems (ndir.h). +# -DMAJOR_IN_MKDEV If major, minor, makedev defined in sys/mkdev.h. +# -DMAJOR_IN_SYSMACROS If major, minor, makedev defined in sys/sysmacros.h. +# -DRETSIGTYPE=int If your signal handlers return int, not void. +# -DHAVE_SYS_MTIO_H If you have sys/mtio.h (magtape ioctls). +# -DHAVE_SYS_GENTAPE_H If you have sys/gentape.h (ISC magtape ioctls). +# -DHAVE_NETDB_H To use rexec for remote tape operations +# instead of forking rsh or remsh. +# -DNO_REMOTE If you have neither a remote shell nor rexec. +# -DHAVE_VPRINTF If you have vprintf function. +# -DHAVE_DOPRNT If you have _doprnt function (but lack vprintf). +# -DHAVE_FTIME If you have ftime system call. +# -DHAVE_STRSTR If you have strstr function. +# -DHAVE_VALLOC If you have valloc function. +# -DHAVE_MKDIR If you have mkdir and rmdir system calls. +# -DHAVE_MKNOD If you have mknod system call. +# -DHAVE_RENAME If you have rename system call. +# -DHAVE_GETCWD If not POSIX.1 but have getcwd function. +# -DHAVE_FTRUNCATE If you have ftruncate system call. +# -DV7 On Version 7 Unix (not tested in a long time). +# -DEMUL_OPEN3 If you lack a 3-argument version of open, and want +# to emulate it with system calls you do have. +# -DNO_OPEN3 If you lack the 3-argument open and want to +# disable the tar -k option instead of emulating open. +# -DXENIX If you have sys/inode.h and need it to be included. + +DEF_AR_FILE = /dev/rst0 +DEFBLOCKING = 20 +DEFS = -DRETSIGTYPE=void -DDIRENT=1 -DHAVE_SYS_MTIO_H=1 -DHAVE_UNISTD_H=1 -DHAVE_GETGRGID=1 -DHAVE_GETPWUID=1 -DHAVE_STRING_H=1 -DHAVE_LIMITS_H=1 -DHAVE_STRSTR=1 -DHAVE_VALLOC=1 -DHAVE_MKDIR=1 -DHAVE_MKNOD=1 -DHAVE_RENAME=1 -DHAVE_FTRUNCATE=1 -DHAVE_GETCWD=1 -DHAVE_VPRINTF=1 -DDEF_AR_FILE=\"$(DEF_AR_FILE)\" -DDEFBLOCKING=$(DEFBLOCKING) + +# Set this to rtapelib.o unless you defined NO_REMOTE, in which case +# make it empty. +RTAPELIB = rtapelib.o +LIBS = + +CFLAGS = -g +LDFLAGS = -g + +prefix = /usr/bin +exec_prefix = $(prefix) + +# Prefix for each installed program, normally empty or `g'. +binprefix = + +# The directory to install tar in. +bindir = $(exec_prefix)/bin + +# Where to put the rmt executable. +libdir = /sbin + +# The directory to install the info files in. +infodir = $(prefix)/info + +#### End of system configuration section. #### + +SHELL = /bin/sh + +SRC1 = tar.c create.c extract.c buffer.c getoldopt.c update.c gnu.c mangle.c +SRC2 = version.c list.c names.c diffarch.c port.c fnmatch.c getopt.c malloc.c +SRC3 = getopt1.c regex.c getdate.y getdate.c alloca.c +SRCS = $(SRC1) $(SRC2) $(SRC3) +OBJ1 = tar.o create.o extract.o buffer.o getoldopt.o update.o gnu.o mangle.o +OBJ2 = version.o list.o names.o diffarch.o port.o fnmatch.o getopt.o +OBJ3 = getopt1.o regex.o getdate.o $(RTAPELIB) +OBJS = $(OBJ1) $(OBJ2) $(OBJ3) +AUX = README INSTALL NEWS COPYING ChangeLog Makefile.in makefile.pc \ + configure configure.in \ + tar.h fnmatch.h pathmax.h port.h open3.h getopt.h regex.h \ + rmt.h rmt.c rtapelib.c \ + msd_dir.h msd_dir.c tcexparg.c \ + level-0 level-1 backup-specs dump-remind getpagesize.h +# tar.texinfo tar.info* texinfo.tex \ + +all: tar rmt +# tar.info + +.c.o: + $(CC) -c $(CFLAGS) $(CPPFLAGS) $(DEFS) -I$(srcdir) -I. $< + +tar: $(OBJS) + $(CC) $(LDFLAGS) -o $@ $(OBJS) $(LIBS) + +rmt: rmt.c + $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(srcdir)/rmt.c $(LIBS) + +tar.info: tar.texinfo + makeinfo $(srcdir)/tar.texinfo + +install: all + $(INSTALL_PROGRAM) tar $(bindir)/$(binprefix)tar + -test ! -f rmt || $(INSTALL_PROGRAM) rmt $(libdir)/rmt +# for file in $(srcdir)/tar.info*; \ +# do $(INSTALL_DATA) $$file $(infodir)/$$file; \ +# done + +uninstall: + rm -f $(bindir)/$(binprefix)tar $(infodir)/tar.info* + -rm -f $(libdir)/rmt + +$(OBJS): tar.h pathmax.h port.h +regex.o buffer.o tar.o: regex.h +tar.o fnmatch.o: fnmatch.h + +getdate.c: getdate.y + $(YACC) $(srcdir)/getdate.y + mv y.tab.c getdate.c +# getdate.y has 8 shift/reduce conflicts. + +TAGS: $(SRCS) + etags $(SRCS) + +clean: + rm -f *.o tar rmt core +mostlyclean: clean + +distclean: clean + rm -f Makefile config.status + +realclean: distclean + rm -f TAGS *.info* getdate.c y.tab.c + +shar: $(SRCS) $(AUX) + shar $(SRCS) $(AUX) | gzip > tar-`sed -e '/version_string/!d' -e 's/[^0-9.]*\([0-9.]*\).*/\1/' -e q version.c`.shar.z + +dist: $(SRCS) $(AUX) + echo tar-`sed -e '/version_string/!d' -e 's/[^0-9.]*\([0-9.]*\).*/\1/' -e q version.c` > .fname + -rm -rf `cat .fname` + mkdir `cat .fname` + for file in $(SRCS) $(AUX); do \ + ln $$file `cat .fname` || cp $$file `cat .fname`; done + tar chzf `cat .fname`.tar.z `cat .fname` + -rm -rf `cat .fname` .fname + +tar.zoo: $(SRCS) $(AUX) + -rm -rf tmp.dir + -mkdir tmp.dir + -rm tar.zoo + for X in $(SRCS) $(AUX) ; do echo $$X ; sed 's/$$/
/' $$X > tmp.dir/$$X ; done + cd tmp.dir ; zoo aM ../tar.zoo * + -rm -rf tmp.dir + +# Prevent GNU make v3 from overflowing arg limit on SysV. +.NOEXPORT: diff --git a/gnu/usr.bin/tar/README b/gnu/usr.bin/tar/README new file mode 100644 index 000000000000..4b577e78eb70 --- /dev/null +++ b/gnu/usr.bin/tar/README @@ -0,0 +1,40 @@ +Hey! Emacs! Yo! This is -*- Text -*- !!! + +This GNU tar 1.11.2. Please send bug reports, etc., to +bug-gnu-utils@prep.ai.mit.edu. This is a beta-test release. Please +try it out. There is no manual; the release of version 1.12 will +contain a manual. + +GNU tar is based heavily on John Gilmore's public domain tar, but with +added features. The manual is currently being written. + +This distribution also includes rmt, the remote tape server (which +normally must reside in /etc). The mt tape drive control program is +in the GNU cpio distribution. + +See the file INSTALL for compilation and installation instructions for Unix. +See the file NEWS for information on all that is new in this version +of tar. + +makefile.pc is a makefile for Turbo C 2.0 on MS-DOS. + +Various people have been having problems using floppies on a NeXT. In +order to have them work right, you need to kill the automounting +program which tries to monut floppies as soon as they are added. + +If you want to do incremental dumps, use the distributed backup +scripts. They are what we use at the FSF to do all our backups. Most +importantly, do not use --incremental (-G) or --after-date (-N) or +--newer-mtime to do incremental dumps. The only option that works +correctly for this purpose is --listed-incremental. (When extracting +incremental dumps, use --incremental (-G).) + +If your system needs to link with -lPW to get alloca, but has +rename in the C library (so HAVE_RENAME is defined), -lPW might +give you an incorrect version of rename. On HP-UX this manifests +itself as an undefined data symbol called "Error" when linking cp, ln, +and mv. If this happens, use `ar x' to extract alloca.o from libPW.a +and `ar rc' to put it in a library liballoca.a, and put that in LIBS +instead of -lPW. This problem does not occur when using gcc, which +has alloca built in. + diff --git a/gnu/usr.bin/tar/buffer.c b/gnu/usr.bin/tar/buffer.c new file mode 100644 index 000000000000..e0ffc2d28654 --- /dev/null +++ b/gnu/usr.bin/tar/buffer.c @@ -0,0 +1,1584 @@ +/* Buffer management for tar. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * Buffer management for tar. + * + * Written by John Gilmore, ihnp4!hoptoad!gnu, on 25 August 1985. + */ + +#include <stdio.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif +#include <sys/types.h> /* For non-Berkeley systems */ +#include <signal.h> +#include <time.h> +time_t time (); + +#ifdef HAVE_SYS_MTIO_H +#include <sys/ioctl.h> +#include <sys/mtio.h> +#endif + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#ifdef __MSDOS__ +#include <process.h> +#endif + +#ifdef XENIX +#include <sys/inode.h> +#endif + +#include "tar.h" +#include "port.h" +#include "rmt.h" +#include "regex.h" + +/* Either stdout or stderr: The thing we write messages (standard msgs, not + errors) to. Stdout unless we're writing a pipe, in which case stderr */ +FILE *msg_file = stdout; + +#define STDIN 0 /* Standard input file descriptor */ +#define STDOUT 1 /* Standard output file descriptor */ + +#define PREAD 0 /* Read file descriptor from pipe() */ +#define PWRITE 1 /* Write file descriptor from pipe() */ + +#define MAGIC_STAT 105 /* Magic status returned by child, if + it can't exec. We hope compress/sh + never return this status! */ + +void *valloc (); + +void writeerror (); +void readerror (); + +void ck_pipe (); +void ck_close (); + +int backspace_output (); +extern void finish_header (); +void flush_archive (); +int isfile (); +int new_volume (); +void verify_volume (); +extern void to_oct (); + +#ifndef __MSDOS__ +/* Obnoxious test to see if dimwit is trying to dump the archive */ +dev_t ar_dev; +ino_t ar_ino; +#endif + +/* + * The record pointed to by save_rec should not be overlaid + * when reading in a new tape block. Copy it to record_save_area first, and + * change the pointer in *save_rec to point to record_save_area. + * Saved_recno records the record number at the time of the save. + * This is used by annofile() to print the record number of a file's + * header record. + */ +static union record **save_rec; +union record record_save_area; +static long saved_recno; + +/* + * PID of child program, if f_compress or remote archive access. + */ +static int childpid = 0; + +/* + * Record number of the start of this block of records + */ +long baserec; + +/* + * Error recovery stuff + */ +static int r_error_count; + +/* + * Have we hit EOF yet? + */ +static int hit_eof; + +/* Checkpointing counter */ +static int checkpoint; + +/* JF we're reading, but we just read the last record and its time to update */ +extern time_to_start_writing; +int file_to_switch_to = -1; /* If remote update, close archive, and use + this descriptor to write to */ + +static int volno = 1; /* JF which volume of a multi-volume tape + we're on */ +static int global_volno = 1; /* Volume number to print in external messages. */ + +char *save_name = 0; /* Name of the file we are currently writing */ +long save_totsize; /* total size of file we are writing. Only + valid if save_name is non_zero */ +long save_sizeleft; /* Where we are in the file we are writing. + Only valid if save_name is non-zero */ + +int write_archive_to_stdout; + +/* Used by fl_read and fl_write to store the real info about saved names */ +static char real_s_name[NAMSIZ]; +static long real_s_totsize; +static long real_s_sizeleft; + +/* Reset the EOF flag (if set), and re-set ar_record, etc */ + +void +reset_eof () +{ + if (hit_eof) + { + hit_eof = 0; + ar_record = ar_block; + ar_last = ar_block + blocking; + ar_reading = 0; + } +} + +/* + * Return the location of the next available input or output record. + * Return NULL for EOF. Once we have returned NULL, we just keep returning + * it, to avoid accidentally going on to the next file on the "tape". + */ +union record * +findrec () +{ + if (ar_record == ar_last) + { + if (hit_eof) + return (union record *) NULL; /* EOF */ + flush_archive (); + if (ar_record == ar_last) + { + hit_eof++; + return (union record *) NULL; /* EOF */ + } + } + return ar_record; +} + + +/* + * Indicate that we have used all records up thru the argument. + * (should the arg have an off-by-1? XXX FIXME) + */ +void +userec (rec) + union record *rec; +{ + while (rec >= ar_record) + ar_record++; + /* + * Do NOT flush the archive here. If we do, the same + * argument to userec() could mean the next record (if the + * input block is exactly one record long), which is not what + * is intended. + */ + if (ar_record > ar_last) + abort (); +} + + +/* + * Return a pointer to the end of the current records buffer. + * All the space between findrec() and endofrecs() is available + * for filling with data, or taking data from. + */ +union record * +endofrecs () +{ + return ar_last; +} + + +/* + * Duplicate a file descriptor into a certain slot. + * Equivalent to BSD "dup2" with error reporting. + */ +void +dupto (from, to, msg) + int from, to; + char *msg; +{ + int err; + + if (from != to) + { + err = close (to); + if (err < 0 && errno != EBADF) + { + msg_perror ("Cannot close descriptor %d", to); + exit (EX_SYSTEM); + } + err = dup (from); + if (err != to) + { + msg_perror ("cannot dup %s", msg); + exit (EX_SYSTEM); + } + ck_close (from); + } +} + +#ifdef __MSDOS__ +void +child_open () +{ + fprintf (stderr, "MS-DOS %s can't use compressed or remote archives\n", tar); + exit (EX_ARGSBAD); +} + +#else +void +child_open () +{ + int pipe[2]; + int err = 0; + + int kidpipe[2]; + int kidchildpid; + +#define READ 0 +#define WRITE 1 + + ck_pipe (pipe); + + childpid = fork (); + if (childpid < 0) + { + msg_perror ("cannot fork"); + exit (EX_SYSTEM); + } + if (childpid > 0) + { + /* We're the parent. Clean up and be happy */ + /* This, at least, is easy */ + + if (ar_reading) + { + f_reblock++; + archive = pipe[READ]; + ck_close (pipe[WRITE]); + } + else + { + archive = pipe[WRITE]; + ck_close (pipe[READ]); + } + return; + } + + /* We're the kid */ + if (ar_reading) + { + dupto (pipe[WRITE], STDOUT, "(child) pipe to stdout"); + ck_close (pipe[READ]); + } + else + { + dupto (pipe[READ], STDIN, "(child) pipe to stdin"); + ck_close (pipe[WRITE]); + } + + /* We need a child tar only if + 1: we're reading/writing stdin/out (to force reblocking) + 2: the file is to be accessed by rmt (compress doesn't know how) + 3: the file is not a plain file */ +#ifdef NO_REMOTE + if (!(ar_files[0][0] == '-' && ar_files[0][1] == '\0') && isfile (ar_files[0])) +#else + if (!(ar_files[0][0] == '-' && ar_files[0][1] == '\0') && !_remdev (ar_files[0]) && isfile (ar_files[0])) +#endif + { + /* We don't need a child tar. Open the archive */ + if (ar_reading) + { + archive = open (ar_files[0], O_RDONLY | O_BINARY, 0666); + if (archive < 0) + { + msg_perror ("can't open archive %s", ar_files[0]); + exit (EX_BADARCH); + } + dupto (archive, STDIN, "archive to stdin"); + /* close(archive); */ + } + else + { + archive = creat (ar_files[0], 0666); + if (archive < 0) + { + msg_perror ("can't open archive %s", ar_files[0]); + exit (EX_BADARCH); + } + dupto (archive, STDOUT, "archive to stdout"); + /* close(archive); */ + } + } + else + { + /* We need a child tar */ + ck_pipe (kidpipe); + + kidchildpid = fork (); + if (kidchildpid < 0) + { + msg_perror ("child can't fork"); + exit (EX_SYSTEM); + } + + if (kidchildpid > 0) + { + /* About to exec compress: set up the files */ + if (ar_reading) + { + dupto (kidpipe[READ], STDIN, "((child)) pipe to stdin"); + ck_close (kidpipe[WRITE]); + /* dup2(pipe[WRITE],STDOUT); */ + } + else + { + /* dup2(pipe[READ],STDIN); */ + dupto (kidpipe[WRITE], STDOUT, "((child)) pipe to stdout"); + ck_close (kidpipe[READ]); + } + /* ck_close(pipe[READ]); */ + /* ck_close(pipe[WRITE]); */ + /* ck_close(kidpipe[READ]); + ck_close(kidpipe[WRITE]); */ + } + else + { + /* Grandchild. Do the right thing, namely sit here and + read/write the archive, and feed stuff back to compress */ + tar = "tar (child)"; + if (ar_reading) + { + dupto (kidpipe[WRITE], STDOUT, "[child] pipe to stdout"); + ck_close (kidpipe[READ]); + } + else + { + dupto (kidpipe[READ], STDIN, "[child] pipe to stdin"); + ck_close (kidpipe[WRITE]); + } + + if (ar_files[0][0] == '-' && ar_files[0][1] == '\0') + { + if (ar_reading) + archive = STDIN; + else + archive = STDOUT; + } + else /* This can't happen if (ar_reading==2) + archive = rmtopen(ar_files[0], O_RDWR|O_CREAT|O_BINARY, 0666); + else */ if (ar_reading) + archive = rmtopen (ar_files[0], O_RDONLY | O_BINARY, 0666); + else + archive = rmtcreat (ar_files[0], 0666); + + if (archive < 0) + { + msg_perror ("can't open archive %s", ar_files[0]); + exit (EX_BADARCH); + } + + if (ar_reading) + { + for (;;) + { + char *ptr; + int max, count; + + r_error_count = 0; + error_loop: + err = rmtread (archive, ar_block->charptr, (int) (blocksize)); + if (err < 0) + { + readerror (); + goto error_loop; + } + if (err == 0) + break; + ptr = ar_block->charptr; + max = err; + while (max) + { + count = (max < RECORDSIZE) ? max : RECORDSIZE; + err = write (STDOUT, ptr, count); + if (err != count) + { + if (err < 0) + { + msg_perror ("can't write to compression program"); + exit (EX_SYSTEM); + } + else + msg ("write to compression program short %d bytes", + count - err); + count = (err < 0) ? 0 : err; + } + ptr += count; + max -= count; + } + } + } + else + { + for (;;) + { + int n; + char *ptr; + + n = blocksize; + ptr = ar_block->charptr; + while (n) + { + err = read (STDIN, ptr, (n < RECORDSIZE) ? n : RECORDSIZE); + if (err <= 0) + break; + n -= err; + ptr += err; + } + /* EOF */ + if (err == 0) + { + if (!f_compress_block) + blocksize -= n; + else + bzero (ar_block->charptr + blocksize - n, n); + err = rmtwrite (archive, ar_block->charptr, blocksize); + if (err != (blocksize)) + writeerror (err); + if (!f_compress_block) + blocksize += n; + break; + } + if (n) + { + msg_perror ("can't read from compression program"); + exit (EX_SYSTEM); + } + err = rmtwrite (archive, ar_block->charptr, (int) blocksize); + if (err != blocksize) + writeerror (err); + } + } + + /* close_archive(); */ + exit (0); + } + } + /* So we should exec compress (-d) */ + if (ar_reading) + execlp (f_compressprog, f_compressprog, "-d", (char *) 0); + else + execlp (f_compressprog, f_compressprog, (char *) 0); + msg_perror ("can't exec %s", f_compressprog); + _exit (EX_SYSTEM); +} + + +/* return non-zero if p is the name of a directory */ +int +isfile (p) + char *p; +{ + struct stat stbuf; + + if (stat (p, &stbuf) < 0) + return 1; + if (S_ISREG (stbuf.st_mode)) + return 1; + return 0; +} + +#endif + +/* + * Open an archive file. The argument specifies whether we are + * reading or writing. + */ +/* JF if the arg is 2, open for reading and writing. */ +void +open_archive (reading) + int reading; +{ + msg_file = f_exstdout ? stderr : stdout; + + if (blocksize == 0) + { + msg ("invalid value for blocksize"); + exit (EX_ARGSBAD); + } + + if (n_ar_files == 0) + { + msg ("No archive name given, what should I do?"); + exit (EX_BADARCH); + } + + /*NOSTRICT*/ + if (f_multivol) + { + ar_block = (union record *) valloc ((unsigned) (blocksize + (2 * RECORDSIZE))); + if (ar_block) + ar_block += 2; + } + else + ar_block = (union record *) valloc ((unsigned) blocksize); + if (!ar_block) + { + msg ("could not allocate memory for blocking factor %d", + blocking); + exit (EX_ARGSBAD); + } + + ar_record = ar_block; + ar_last = ar_block + blocking; + ar_reading = reading; + + if (f_multivol && f_verify) + { + msg ("cannot verify multi-volume archives"); + exit (EX_ARGSBAD); + } + + if (f_compressprog) + { + if (reading == 2 || f_verify) + { + msg ("cannot update or verify compressed archives"); + exit (EX_ARGSBAD); + } + if (f_multivol) + { + msg ("cannot use multi-volume compressed archives"); + exit (EX_ARGSBAD); + } + child_open (); + if (!reading && ar_files[0][0] == '-' && ar_files[0][1] == '\0') + msg_file = stderr; + /* child_open(rem_host, rem_file); */ + } + else if (ar_files[0][0] == '-' && ar_files[0][1] == '\0') + { + f_reblock++; /* Could be a pipe, be safe */ + if (f_verify) + { + msg ("can't verify stdin/stdout archive"); + exit (EX_ARGSBAD); + } + if (reading == 2) + { + archive = STDIN; + msg_file = stderr; + write_archive_to_stdout++; + } + else if (reading) + archive = STDIN; + else + { + archive = STDOUT; + msg_file = stderr; + } + } + else if (reading == 2 || f_verify) + { + archive = rmtopen (ar_files[0], O_RDWR | O_CREAT | O_BINARY, 0666); + } + else if (reading) + { + archive = rmtopen (ar_files[0], O_RDONLY | O_BINARY, 0666); + } + else + { + archive = rmtcreat (ar_files[0], 0666); + } + if (archive < 0) + { + msg_perror ("can't open %s", ar_files[0]); + exit (EX_BADARCH); + } +#ifndef __MSDOS__ + if (!_isrmt (archive)) + { + struct stat tmp_stat; + + fstat (archive, &tmp_stat); + if (S_ISREG (tmp_stat.st_mode)) + { + ar_dev = tmp_stat.st_dev; + ar_ino = tmp_stat.st_ino; + } + } +#endif + +#ifdef __MSDOS__ + setmode (archive, O_BINARY); +#endif + + if (reading) + { + ar_last = ar_block; /* Set up for 1st block = # 0 */ + (void) findrec (); /* Read it in, check for EOF */ + + if (f_volhdr) + { + union record *head; +#if 0 + char *ptr; + + if (f_multivol) + { + ptr = malloc (strlen (f_volhdr) + 20); + sprintf (ptr, "%s Volume %d", f_volhdr, 1); + } + else + ptr = f_volhdr; +#endif + head = findrec (); + if (!head) + { + msg ("Archive not labelled to match %s", f_volhdr); + exit (EX_BADVOL); + } + if (re_match (label_pattern, head->header.arch_name, + strlen (head->header.arch_name), 0, 0) < 0) + { + msg ("Volume mismatch! %s!=%s", f_volhdr, + head->header.arch_name); + exit (EX_BADVOL); + } +#if 0 + if (strcmp (ptr, head->header.name)) + { + msg ("Volume mismatch! %s!=%s", ptr, head->header.name); + exit (EX_BADVOL); + } + if (ptr != f_volhdr) + free (ptr); +#endif + } + } + else if (f_volhdr) + { + bzero ((void *) ar_block, RECORDSIZE); + if (f_multivol) + sprintf (ar_block->header.arch_name, "%s Volume 1", f_volhdr); + else + strcpy (ar_block->header.arch_name, f_volhdr); + current_file_name = ar_block->header.arch_name; + ar_block->header.linkflag = LF_VOLHDR; + to_oct (time (0), 1 + 12, ar_block->header.mtime); + finish_header (ar_block); + /* ar_record++; */ + } +} + + +/* + * Remember a union record * as pointing to something that we + * need to keep when reading onward in the file. Only one such + * thing can be remembered at once, and it only works when reading + * an archive. + * + * We calculate "offset" then add it because some compilers end up + * adding (baserec+ar_record), doing a 9-bit shift of baserec, then + * subtracting ar_block from that, shifting it back, losing the top 9 bits. + */ +void +saverec (pointer) + union record **pointer; +{ + long offset; + + save_rec = pointer; + offset = ar_record - ar_block; + saved_recno = baserec + offset; +} + +/* + * Perform a write to flush the buffer. + */ + +/*send_buffer_to_file(); + if(new_volume) { + deal_with_new_volume_stuff(); + send_buffer_to_file(); + } + */ + +void +fl_write () +{ + int err; + int copy_back; + static long bytes_written = 0; + + if (f_checkpoint && !(++checkpoint % 10)) + msg ("Write checkpoint %d\n", checkpoint); + if (tape_length && bytes_written >= tape_length * 1024) + { + errno = ENOSPC; + err = 0; + } + else + err = rmtwrite (archive, ar_block->charptr, (int) blocksize); + if (err != blocksize && !f_multivol) + writeerror (err); + else if (f_totals) + tot_written += blocksize; + + if (err > 0) + bytes_written += err; + if (err == blocksize) + { + if (f_multivol) + { + if (!save_name) + { + real_s_name[0] = '\0'; + real_s_totsize = 0; + real_s_sizeleft = 0; + return; + } +#ifdef __MSDOS__ + if (save_name[1] == ':') + save_name += 2; +#endif + while (*save_name == '/') + save_name++; + + strcpy (real_s_name, save_name); + real_s_totsize = save_totsize; + real_s_sizeleft = save_sizeleft; + } + return; + } + + /* We're multivol Panic if we didn't get the right kind of response */ + /* ENXIO is for the UNIX PC */ + if (err < 0 && errno != ENOSPC && errno != EIO && errno != ENXIO) + writeerror (err); + + /* If error indicates a short write, we just move to the next tape. */ + + if (new_volume (0) < 0) + return; + bytes_written = 0; + if (f_volhdr && real_s_name[0]) + { + copy_back = 2; + ar_block -= 2; + } + else if (f_volhdr || real_s_name[0]) + { + copy_back = 1; + ar_block--; + } + else + copy_back = 0; + if (f_volhdr) + { + bzero ((void *) ar_block, RECORDSIZE); + sprintf (ar_block->header.arch_name, "%s Volume %d", f_volhdr, volno); + to_oct (time (0), 1 + 12, ar_block->header.mtime); + ar_block->header.linkflag = LF_VOLHDR; + finish_header (ar_block); + } + if (real_s_name[0]) + { + int tmp; + + if (f_volhdr) + ar_block++; + bzero ((void *) ar_block, RECORDSIZE); + strcpy (ar_block->header.arch_name, real_s_name); + ar_block->header.linkflag = LF_MULTIVOL; + to_oct ((long) real_s_sizeleft, 1 + 12, + ar_block->header.size); + to_oct ((long) real_s_totsize - real_s_sizeleft, + 1 + 12, ar_block->header.offset); + tmp = f_verbose; + f_verbose = 0; + finish_header (ar_block); + f_verbose = tmp; + if (f_volhdr) + ar_block--; + } + + err = rmtwrite (archive, ar_block->charptr, (int) blocksize); + if (err != blocksize) + writeerror (err); + else if (f_totals) + tot_written += blocksize; + + + bytes_written = blocksize; + if (copy_back) + { + ar_block += copy_back; + bcopy ((void *) (ar_block + blocking - copy_back), + (void *) ar_record, + copy_back * RECORDSIZE); + ar_record += copy_back; + + if (real_s_sizeleft >= copy_back * RECORDSIZE) + real_s_sizeleft -= copy_back * RECORDSIZE; + else if ((real_s_sizeleft + RECORDSIZE - 1) / RECORDSIZE <= copy_back) + real_s_name[0] = '\0'; + else + { +#ifdef __MSDOS__ + if (save_name[1] == ':') + save_name += 2; +#endif + while (*save_name == '/') + save_name++; + + strcpy (real_s_name, save_name); + real_s_sizeleft = save_sizeleft; + real_s_totsize = save_totsize; + } + copy_back = 0; + } +} + +/* Handle write errors on the archive. Write errors are always fatal */ +/* Hitting the end of a volume does not cause a write error unless the write +* was the first block of the volume */ + +void +writeerror (err) + int err; +{ + if (err < 0) + { + msg_perror ("can't write to %s", ar_files[cur_ar_file]); + exit (EX_BADARCH); + } + else + { + msg ("only wrote %u of %u bytes to %s", err, blocksize, ar_files[cur_ar_file]); + exit (EX_BADARCH); + } +} + +/* + * Handle read errors on the archive. + * + * If the read should be retried, readerror() returns to the caller. + */ +void +readerror () +{ +# define READ_ERROR_MAX 10 + + read_error_flag++; /* Tell callers */ + + msg_perror ("read error on %s", ar_files[cur_ar_file]); + + if (baserec == 0) + { + /* First block of tape. Probably stupidity error */ + exit (EX_BADARCH); + } + + /* + * Read error in mid archive. We retry up to READ_ERROR_MAX times + * and then give up on reading the archive. We set read_error_flag + * for our callers, so they can cope if they want. + */ + if (r_error_count++ > READ_ERROR_MAX) + { + msg ("Too many errors, quitting."); + exit (EX_BADARCH); + } + return; +} + + +/* + * Perform a read to flush the buffer. + */ +void +fl_read () +{ + int err; /* Result from system call */ + int left; /* Bytes left */ + char *more; /* Pointer to next byte to read */ + + if (f_checkpoint && !(++checkpoint % 10)) + msg ("Read checkpoint %d\n", checkpoint); + + /* + * Clear the count of errors. This only applies to a single + * call to fl_read. We leave read_error_flag alone; it is + * only turned off by higher level software. + */ + r_error_count = 0; /* Clear error count */ + + /* + * If we are about to wipe out a record that + * somebody needs to keep, copy it out to a holding + * area and adjust somebody's pointer to it. + */ + if (save_rec && + *save_rec >= ar_record && + *save_rec < ar_last) + { + record_save_area = **save_rec; + *save_rec = &record_save_area; + } + if (write_archive_to_stdout && baserec != 0) + { + err = rmtwrite (1, ar_block->charptr, blocksize); + if (err != blocksize) + writeerror (err); + } + if (f_multivol) + { + if (save_name) + { + if (save_name != real_s_name) + { +#ifdef __MSDOS__ + if (save_name[1] == ':') + save_name += 2; +#endif + while (*save_name == '/') + save_name++; + + strcpy (real_s_name, save_name); + save_name = real_s_name; + } + real_s_totsize = save_totsize; + real_s_sizeleft = save_sizeleft; + + } + else + { + real_s_name[0] = '\0'; + real_s_totsize = 0; + real_s_sizeleft = 0; + } + } + +error_loop: + err = rmtread (archive, ar_block->charptr, (int) blocksize); + if (err == blocksize) + return; + + if ((err == 0 || (err < 0 && errno == ENOSPC) || (err > 0 && !f_reblock)) && f_multivol) + { + union record *head; + + try_volume: + if (new_volume ((cmd_mode == CMD_APPEND || cmd_mode == CMD_CAT || cmd_mode == CMD_UPDATE) ? 2 : 1) < 0) + return; + vol_error: + err = rmtread (archive, ar_block->charptr, (int) blocksize); + if (err < 0) + { + readerror (); + goto vol_error; + } + if (err != blocksize) + goto short_read; + + head = ar_block; + + if (head->header.linkflag == LF_VOLHDR) + { + if (f_volhdr) + { +#if 0 + char *ptr; + + ptr = (char *) malloc (strlen (f_volhdr) + 20); + sprintf (ptr, "%s Volume %d", f_volhdr, volno); +#endif + if (re_match (label_pattern, head->header.arch_name, + strlen (head->header.arch_name), + 0, 0) < 0) + { + msg ("Volume mismatch! %s!=%s", f_volhdr, + head->header.arch_name); + --volno; + --global_volno; + goto try_volume; + } + +#if 0 + if (strcmp (ptr, head->header.name)) + { + msg ("Volume mismatch! %s!=%s", ptr, head->header.name); + --volno; + --global_volno; + free (ptr); + goto try_volume; + } + free (ptr); +#endif + } + if (f_verbose) + fprintf (msg_file, "Reading %s\n", head->header.arch_name); + head++; + } + else if (f_volhdr) + { + msg ("Warning: No volume header!"); + } + + if (real_s_name[0]) + { + long from_oct (); + + if (head->header.linkflag != LF_MULTIVOL || strcmp (head->header.arch_name, real_s_name)) + { + msg ("%s is not continued on this volume!", real_s_name); + --volno; + --global_volno; + goto try_volume; + } + if (real_s_totsize != from_oct (1 + 12, head->header.size) + from_oct (1 + 12, head->header.offset)) + { + msg ("%s is the wrong size (%ld!=%ld+%ld)", + head->header.arch_name, save_totsize, + from_oct (1 + 12, head->header.size), + from_oct (1 + 12, head->header.offset)); + --volno; + --global_volno; + goto try_volume; + } + if (real_s_totsize - real_s_sizeleft != from_oct (1 + 12, head->header.offset)) + { + msg ("This volume is out of sequence"); + --volno; + --global_volno; + goto try_volume; + } + head++; + } + ar_record = head; + return; + } + else if (err < 0) + { + readerror (); + goto error_loop; /* Try again */ + } + +short_read: + more = ar_block->charptr + err; + left = blocksize - err; + +again: + if (0 == (((unsigned) left) % RECORDSIZE)) + { + /* FIXME, for size=0, multi vol support */ + /* On the first block, warn about the problem */ + if (!f_reblock && baserec == 0 && f_verbose && err > 0) + { + /* msg("Blocksize = %d record%s", + err / RECORDSIZE, (err > RECORDSIZE)? "s": "");*/ + msg ("Blocksize = %d records", err / RECORDSIZE); + } + ar_last = ar_block + ((unsigned) (blocksize - left)) / RECORDSIZE; + return; + } + if (f_reblock) + { + /* + * User warned us about this. Fix up. + */ + if (left > 0) + { + error2loop: + err = rmtread (archive, more, (int) left); + if (err < 0) + { + readerror (); + goto error2loop; /* Try again */ + } + if (err == 0) + { + msg ("archive %s EOF not on block boundary", ar_files[cur_ar_file]); + exit (EX_BADARCH); + } + left -= err; + more += err; + goto again; + } + } + else + { + msg ("only read %d bytes from archive %s", err, ar_files[cur_ar_file]); + exit (EX_BADARCH); + } +} + + +/* + * Flush the current buffer to/from the archive. + */ +void +flush_archive () +{ + int c; + + baserec += ar_last - ar_block;/* Keep track of block #s */ + ar_record = ar_block; /* Restore pointer to start */ + ar_last = ar_block + blocking;/* Restore pointer to end */ + + if (ar_reading) + { + if (time_to_start_writing) + { + time_to_start_writing = 0; + ar_reading = 0; + + if (file_to_switch_to >= 0) + { + if ((c = rmtclose (archive)) < 0) + msg_perror ("Warning: can't close %s(%d,%d)", ar_files[cur_ar_file], archive, c); + + archive = file_to_switch_to; + } + else + (void) backspace_output (); + fl_write (); + } + else + fl_read (); + } + else + { + fl_write (); + } +} + +/* Backspace the archive descriptor by one blocks worth. + If its a tape, MTIOCTOP will work. If its something else, + we try to seek on it. If we can't seek, we lose! */ +int +backspace_output () +{ + long cur; + /* int er; */ + extern char *output_start; + +#ifdef MTIOCTOP + struct mtop t; + + t.mt_op = MTBSR; + t.mt_count = 1; + if ((rmtioctl (archive, MTIOCTOP, &t)) >= 0) + return 1; + if (errno == EIO && (rmtioctl (archive, MTIOCTOP, &t)) >= 0) + return 1; +#endif + + cur = rmtlseek (archive, 0L, 1); + cur -= blocksize; + /* Seek back to the beginning of this block and + start writing there. */ + + if (rmtlseek (archive, cur, 0) != cur) + { + /* Lseek failed. Try a different method */ + msg ("Couldn't backspace archive file. It may be unreadable without -i."); + /* Replace the first part of the block with nulls */ + if (ar_block->charptr != output_start) + bzero (ar_block->charptr, output_start - ar_block->charptr); + return 2; + } + return 3; +} + + +/* + * Close the archive file. + */ +void +close_archive () +{ + int child; + int status; + int c; + + if (time_to_start_writing || !ar_reading) + flush_archive (); + if (cmd_mode == CMD_DELETE) + { + off_t pos; + + pos = rmtlseek (archive, 0L, 1); +#ifndef __MSDOS__ + (void) ftruncate (archive, pos); +#else + (void) rmtwrite (archive, "", 0); +#endif + } + if (f_verify) + verify_volume (); + + if ((c = rmtclose (archive)) < 0) + msg_perror ("Warning: can't close %s(%d,%d)", ar_files[cur_ar_file], archive, c); + +#ifndef __MSDOS__ + if (childpid) + { + /* + * Loop waiting for the right child to die, or for + * no more kids. + */ + while (((child = wait (&status)) != childpid) && child != -1) + ; + + if (child != -1) + { + if (WIFSIGNALED (status)) + { + /* SIGPIPE is OK, everything else is a problem. */ + if (WTERMSIG (status) != SIGPIPE) + msg ("child died with signal %d%s", WTERMSIG (status), + WIFCOREDUMPED (status) ? " (core dumped)" : ""); + } + else + { + /* Child voluntarily terminated -- but why? */ + if (WEXITSTATUS (status) == MAGIC_STAT) + { + exit (EX_SYSTEM); /* Child had trouble */ + } + if (WEXITSTATUS (status) == (SIGPIPE + 128)) + { + /* + * /bin/sh returns this if its child + * dies with SIGPIPE. 'Sok. + */ + /* Do nothing. */ + } + else if (WEXITSTATUS (status)) + msg ("child returned status %d", + WEXITSTATUS (status)); + } + } + } +#endif /* __MSDOS__ */ +} + + +#ifdef DONTDEF +/* + * Message management. + * + * anno writes a message prefix on stream (eg stdout, stderr). + * + * The specified prefix is normally output followed by a colon and a space. + * However, if other command line options are set, more output can come + * out, such as the record # within the archive. + * + * If the specified prefix is NULL, no output is produced unless the + * command line option(s) are set. + * + * If the third argument is 1, the "saved" record # is used; if 0, the + * "current" record # is used. + */ +void +anno (stream, prefix, savedp) + FILE *stream; + char *prefix; + int savedp; +{ +# define MAXANNO 50 + char buffer[MAXANNO]; /* Holds annorecment */ +# define ANNOWIDTH 13 + int space; + long offset; + int save_e; + + save_e = errno; + /* Make sure previous output gets out in sequence */ + if (stream == stderr) + fflush (stdout); + if (f_sayblock) + { + if (prefix) + { + fputs (prefix, stream); + putc (' ', stream); + } + offset = ar_record - ar_block; + (void) sprintf (buffer, "rec %d: ", + savedp ? saved_recno : + baserec + offset); + fputs (buffer, stream); + space = ANNOWIDTH - strlen (buffer); + if (space > 0) + { + fprintf (stream, "%*s", space, ""); + } + } + else if (prefix) + { + fputs (prefix, stream); + fputs (": ", stream); + } + errno = save_e; +} + +#endif + +/* Called to initialize the global volume number. */ +void +init_volume_number () +{ + FILE *vf; + + vf = fopen (f_volno_file, "r"); + if (!vf && errno != ENOENT) + msg_perror ("%s", f_volno_file); + + if (vf) + { + fscanf (vf, "%d", &global_volno); + fclose (vf); + } +} + +/* Called to write out the closing global volume number. */ +void +closeout_volume_number () +{ + FILE *vf; + + vf = fopen (f_volno_file, "w"); + if (!vf) + msg_perror ("%s", f_volno_file); + else + { + fprintf (vf, "%d\n", global_volno); + fclose (vf); + } +} + +/* We've hit the end of the old volume. Close it and open the next one */ +/* Values for type: 0: writing 1: reading 2: updating */ +int +new_volume (type) + int type; +{ + int c; + char inbuf[80]; + char *p; + static FILE *read_file = 0; + extern int now_verifying; + extern char TTY_NAME[]; + static int looped = 0; + + if (!read_file && !f_run_script_at_end) + read_file = (archive == 0) ? fopen (TTY_NAME, "r") : stdin; + + if (now_verifying) + return -1; + if (f_verify) + verify_volume (); + if ((c = rmtclose (archive)) < 0) + msg_perror ("Warning: can't close %s(%d,%d)", ar_files[cur_ar_file], archive, c); + + global_volno++; + volno++; + cur_ar_file++; + if (cur_ar_file == n_ar_files) + { + cur_ar_file = 0; + looped = 1; + } + +tryagain: + if (looped) + { + /* We have to prompt from now on. */ + if (f_run_script_at_end) + { + closeout_volume_number (); + system (info_script); + } + else + for (;;) + { + fprintf (msg_file, "\007Prepare volume #%d for %s and hit return: ", global_volno, ar_files[cur_ar_file]); + fflush (msg_file); + if (fgets (inbuf, sizeof (inbuf), read_file) == 0) + { + fprintf (msg_file, "EOF? What does that mean?"); + if (cmd_mode != CMD_EXTRACT && cmd_mode != CMD_LIST && cmd_mode != CMD_DIFF) + msg ("Warning: Archive is INCOMPLETE!"); + exit (EX_BADARCH); + } + if (inbuf[0] == '\n' || inbuf[0] == 'y' || inbuf[0] == 'Y') + break; + + switch (inbuf[0]) + { + case '?': + { + fprintf (msg_file, "\ + n [name] Give a new filename for the next (and subsequent) volume(s)\n\ + q Abort tar\n\ + ! Spawn a subshell\n\ + ? Print this list\n"); + } + break; + + case 'q': /* Quit */ + fprintf (msg_file, "No new volume; exiting.\n"); + if (cmd_mode != CMD_EXTRACT && cmd_mode != CMD_LIST && cmd_mode != CMD_DIFF) + msg ("Warning: Archive is INCOMPLETE!"); + exit (EX_BADARCH); + + case 'n': /* Get new file name */ + { + char *q, *r; + static char *old_name; + + for (q = &inbuf[1]; *q == ' ' || *q == '\t'; q++) + ; + for (r = q; *r; r++) + if (*r == '\n') + *r = '\0'; + old_name = p = (char *) malloc ((unsigned) (strlen (q) + 2)); + if (p == 0) + { + msg ("Can't allocate memory for name"); + exit (EX_SYSTEM); + } + (void) strcpy (p, q); + ar_files[cur_ar_file] = p; + } + break; + + case '!': +#ifdef __MSDOS__ + spawnl (P_WAIT, getenv ("COMSPEC"), "-", 0); +#else + /* JF this needs work! */ + switch (fork ()) + { + case -1: + msg_perror ("can't fork!"); + break; + case 0: + p = getenv ("SHELL"); + if (p == 0) + p = "/bin/sh"; + execlp (p, "-sh", "-i", 0); + msg_perror ("can't exec a shell %s", p); + _exit (55); + default: + wait (0); + break; + } +#endif + break; + } + } + } + + + if (type == 2 || f_verify) + archive = rmtopen (ar_files[cur_ar_file], O_RDWR | O_CREAT, 0666); + else if (type == 1) + archive = rmtopen (ar_files[cur_ar_file], O_RDONLY, 0666); + else if (type == 0) + archive = rmtcreat (ar_files[cur_ar_file], 0666); + else + archive = -1; + + if (archive < 0) + { + msg_perror ("can't open %s", ar_files[cur_ar_file]); + goto tryagain; + } +#ifdef __MSDOS__ + setmode (archive, O_BINARY); +#endif + return 0; +} + +/* this is a useless function that takes a buffer returned by wantbytes + and does nothing with it. If the function called by wantbytes returns + an error indicator (non-zero), this function is called for the rest of + the file. + */ +int +no_op (size, data) + int size; + char *data; +{ + return 0; +} + +/* Some other routine wants SIZE bytes in the archive. For each chunk of + the archive, call FUNC with the size of the chunk, and the address of + the chunk it can work with. + */ +int +wantbytes (size, func) + long size; + int (*func) (); +{ + char *data; + long data_size; + + while (size) + { + data = findrec ()->charptr; + if (data == NULL) + { /* Check it... */ + msg ("Unexpected EOF on archive file"); + return -1; + } + data_size = endofrecs ()->charptr - data; + if (data_size > size) + data_size = size; + if ((*func) (data_size, data)) + func = no_op; + userec ((union record *) (data + data_size - 1)); + size -= data_size; + } + return 0; +} diff --git a/gnu/usr.bin/tar/create.c b/gnu/usr.bin/tar/create.c new file mode 100644 index 000000000000..62b9c51178f4 --- /dev/null +++ b/gnu/usr.bin/tar/create.c @@ -0,0 +1,1454 @@ +/* Create a tar archive. + Copyright (C) 1985, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * Create a tar archive. + * + * Written 25 Aug 1985 by John Gilmore, ihnp4!hoptoad!gnu. + */ + +#ifdef _AIX + #pragma alloca +#endif +#include <sys/types.h> +#include <stdio.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#include "tar.h" +#include "port.h" + +#ifndef __MSDOS__ +#include <pwd.h> +#include <grp.h> +#endif + +#if defined (_POSIX_VERSION) +#include <utime.h> +#else +struct utimbuf +{ + long actime; + long modtime; +}; + +#endif + +extern struct stat hstat; /* Stat struct corresponding */ + +#ifndef __MSDOS__ +extern dev_t ar_dev; +extern ino_t ar_ino; +#endif + +/* JF */ +extern struct name *gnu_list_name; + +/* + * If there are no symbolic links, there is no lstat(). Use stat(). + */ +#ifndef S_ISLNK +#define lstat stat +#endif + +extern void print_header (); + +union record *start_header (); +void blank_name_list (); +int check_exclude (); +PTR ck_malloc (); +PTR ck_realloc (); +void clear_buffer (); +void close_archive (); +void collect_and_sort_names (); +int confirm (); +int deal_with_sparse (); +void find_new_file_size (); +void finish_header (); +int finish_sparse_file (); +void finduname (); +void findgname (); +int is_dot_or_dotdot (); +void open_archive (); +char *name_next (); +void name_close (); +void to_oct (); +void dump_file (); +void write_dir_file (); +void write_eot (); +void write_long (); +int zero_record (); + +/* This code moved from tar.h since create.c is the only file that cares + about 'struct link's. This means that other files might not have to + include sys/types.h any more. */ + +struct link + { + struct link *next; + dev_t dev; + ino_t ino; + short linkcount; + char name[1]; + }; + +struct link *linklist; /* Points to first link in list */ + +static nolinks; /* Gets set if we run out of RAM */ + +/* + * "Scratch" space to store the information about a sparse file before + * writing the info into the header or extended header + */ +/* struct sp_array *sparsearray;*/ + +/* number of elts storable in the sparsearray */ +/*int sparse_array_size = 10;*/ + +void +create_archive () +{ + register char *p; + char *name_from_list (); + + open_archive (0); /* Open for writing */ + + if (f_gnudump) + { + char *buf = ck_malloc (PATH_MAX); + char *q, *bufp; + + collect_and_sort_names (); + + while (p = name_from_list ()) + dump_file (p, -1, 1); + /* if(!f_dironly) { */ + blank_name_list (); + while (p = name_from_list ()) + { + strcpy (buf, p); + if (p[strlen (p) - 1] != '/') + strcat (buf, "/"); + bufp = buf + strlen (buf); + for (q = gnu_list_name->dir_contents; q && *q; q += strlen (q) + 1) + { + if (*q == 'Y') + { + strcpy (bufp, q + 1); + dump_file (buf, -1, 1); + } + } + } + /* } */ + free (buf); + } + else + { + while (p = name_next (1)) + dump_file (p, -1, 1); + } + + write_eot (); + close_archive (); + if (f_gnudump) + write_dir_file (); + name_close (); +} + +/* + * Dump a single file. If it's a directory, recurse. + * Result is 1 for success, 0 for failure. + * Sets global "hstat" to stat() output for this file. + */ +void +dump_file (p, curdev, toplevel) + char *p; /* File name to dump */ + int curdev; /* Device our parent dir was on */ + int toplevel; /* Whether we are a toplevel call */ +{ + union record *header; + char type; + extern char *save_name; /* JF for multi-volume support */ + extern long save_totsize; + extern long save_sizeleft; + union record *exhdr; + char save_linkflag; + extern time_t new_time; + int critical_error = 0; + struct utimbuf restore_times; + /* int sparse_ind = 0;*/ + + + if (f_confirm && !confirm ("add", p)) + return; + + /* + * Use stat if following (rather than dumping) 4.2BSD's + * symbolic links. Otherwise, use lstat (which, on non-4.2 + * systems, is #define'd to stat anyway. + */ +#ifdef STX_HIDDEN /* AIX */ + if (0 != f_follow_links ? + statx (p, &hstat, STATSIZE, STX_HIDDEN) : + statx (p, &hstat, STATSIZE, STX_HIDDEN | STX_LINK)) +#else + if (0 != f_follow_links ? stat (p, &hstat) : lstat (p, &hstat)) +#endif + { + badperror: + msg_perror ("can't add file %s", p); + badfile: + if (!f_ignore_failed_read || critical_error) + errors++; + return; + } + + restore_times.actime = hstat.st_atime; + restore_times.modtime = hstat.st_mtime; + +#ifdef S_ISHIDDEN + if (S_ISHIDDEN (hstat.st_mode)) + { + char *new = (char *) alloca (strlen (p) + 2); + if (new) + { + strcpy (new, p); + strcat (new, "@"); + p = new; + } + } +#endif + + /* See if we only want new files, and check if this one is too old to + put in the archive. */ + if (f_new_files + && !f_gnudump + && new_time > hstat.st_mtime + && !S_ISDIR (hstat.st_mode) + && (f_new_files > 1 || new_time > hstat.st_ctime)) + { + if (curdev == -1) + { + msg ("%s: is unchanged; not dumped", p); + } + return; + } + +#ifndef __MSDOS__ + /* See if we are trying to dump the archive */ + if (ar_dev && hstat.st_dev == ar_dev && hstat.st_ino == ar_ino) + { + msg ("%s is the archive; not dumped", p); + return; + } +#endif + /* + * Check for multiple links. + * + * We maintain a list of all such files that we've written so + * far. Any time we see another, we check the list and + * avoid dumping the data again if we've done it once already. + */ + if (hstat.st_nlink > 1 + && (S_ISREG (hstat.st_mode) +#ifdef S_ISCTG + || S_ISCTG (hstat.st_mode) +#endif +#ifdef S_ISCHR + || S_ISCHR (hstat.st_mode) +#endif +#ifdef S_ISBLK + || S_ISBLK (hstat.st_mode) +#endif +#ifdef S_ISFIFO + || S_ISFIFO (hstat.st_mode) +#endif + )) + { + register struct link *lp; + + /* First quick and dirty. Hashing, etc later FIXME */ + for (lp = linklist; lp; lp = lp->next) + { + if (lp->ino == hstat.st_ino && + lp->dev == hstat.st_dev) + { + char *link_name = lp->name; + + /* We found a link. */ + while (!f_absolute_paths && *link_name == '/') + { + static int link_warn = 0; + + if (!link_warn) + { + msg ("Removing leading / from absolute links"); + link_warn++; + } + link_name++; + } + if (link_name - lp->name >= NAMSIZ) + write_long (link_name, LF_LONGLINK); + current_link_name = link_name; + + hstat.st_size = 0; + header = start_header (p, &hstat); + if (header == NULL) + { + critical_error = 1; + goto badfile; + } + strncpy (header->header.arch_linkname, + link_name, NAMSIZ); + + /* Force null truncated */ + header->header.arch_linkname[NAMSIZ - 1] = 0; + + header->header.linkflag = LF_LINK; + finish_header (header); + /* FIXME: Maybe remove from list after all links found? */ + if (f_remove_files) + { + if (unlink (p) == -1) + msg_perror ("cannot remove %s", p); + } + return; /* We dumped it */ + } + } + + /* Not found. Add it to the list of possible links. */ + lp = (struct link *) ck_malloc ((unsigned) (sizeof (struct link) + strlen (p))); + if (!lp) + { + if (!nolinks) + { + msg ( + "no memory for links, they will be dumped as separate files"); + nolinks++; + } + } + lp->ino = hstat.st_ino; + lp->dev = hstat.st_dev; + strcpy (lp->name, p); + lp->next = linklist; + linklist = lp; + } + + /* + * This is not a link to a previously dumped file, so dump it. + */ + if (S_ISREG (hstat.st_mode) +#ifdef S_ISCTG + || S_ISCTG (hstat.st_mode) +#endif + ) + { + int f; /* File descriptor */ + long bufsize, count; + long sizeleft; + register union record *start; + int header_moved; + char isextended = 0; + int upperbound; + /* int end_nulls = 0; */ + + header_moved = 0; + +#ifdef BSD42 + if (f_sparse_files) + { + /* + * JK - This is the test for sparseness: whether the + * "size" of the file matches the number of blocks + * allocated for it. If there is a smaller number + * of blocks that would be necessary to accommodate + * a file of this size, we have a sparse file, i.e., + * at least one of those records in the file is just + * a useless hole. + */ +#ifdef hpux /* Nice of HPUX to gratuitiously change it, huh? - mib */ + if (hstat.st_size - (hstat.st_blocks * 1024) > 1024) +#else + if (hstat.st_size - (hstat.st_blocks * RECORDSIZE) > RECORDSIZE) +#endif + { + int filesize = hstat.st_size; + register int i; + + header = start_header (p, &hstat); + if (header == NULL) + { + critical_error = 1; + goto badfile; + } + header->header.linkflag = LF_SPARSE; + header_moved++; + + /* + * Call the routine that figures out the + * layout of the sparse file in question. + * UPPERBOUND is the index of the last + * element of the "sparsearray," i.e., + * the number of elements it needed to + * describe the file. + */ + + upperbound = deal_with_sparse (p, header); + + /* + * See if we'll need an extended header + * later + */ + if (upperbound > SPARSE_IN_HDR - 1) + header->header.isextended++; + /* + * We store the "real" file size so + * we can show that in case someone wants + * to list the archive, i.e., tar tvf <file>. + * It might be kind of disconcerting if the + * shrunken file size was the one that showed + * up. + */ + to_oct ((long) hstat.st_size, 1 + 12, + header->header.realsize); + + /* + * This will be the new "size" of the + * file, i.e., the size of the file + * minus the records of holes that we're + * skipping over. + */ + + find_new_file_size (&filesize, upperbound); + hstat.st_size = filesize; + to_oct ((long) filesize, 1 + 12, + header->header.size); + /* to_oct((long) end_nulls, 1+12, + header->header.ending_blanks);*/ + + for (i = 0; i < SPARSE_IN_HDR; i++) + { + if (!sparsearray[i].numbytes) + break; + to_oct (sparsearray[i].offset, 1 + 12, + header->header.sp[i].offset); + to_oct (sparsearray[i].numbytes, 1 + 12, + header->header.sp[i].numbytes); + } + + } + } +#else + upperbound = SPARSE_IN_HDR - 1; +#endif + + sizeleft = hstat.st_size; + /* Don't bother opening empty, world readable files. */ + if (sizeleft > 0 || 0444 != (0444 & hstat.st_mode)) + { + f = open (p, O_RDONLY | O_BINARY); + if (f < 0) + goto badperror; + } + else + { + f = -1; + } + + /* If the file is sparse, we've already taken care of this */ + if (!header_moved) + { + header = start_header (p, &hstat); + if (header == NULL) + { + if (f >= 0) + (void) close (f); + critical_error = 1; + goto badfile; + } + } +#ifdef S_ISCTG + /* Mark contiguous files, if we support them */ + if (f_standard && S_ISCTG (hstat.st_mode)) + { + header->header.linkflag = LF_CONTIG; + } +#endif + isextended = header->header.isextended; + save_linkflag = header->header.linkflag; + finish_header (header); + if (isextended) + { + /* int sum = 0;*/ + register int i; + /* register union record *exhdr;*/ + /* int arraybound = SPARSE_EXT_HDR;*/ + /* static */ int index_offset = SPARSE_IN_HDR; + + extend:exhdr = findrec (); + + if (exhdr == NULL) + { + critical_error = 1; + goto badfile; + } + bzero (exhdr->charptr, RECORDSIZE); + for (i = 0; i < SPARSE_EXT_HDR; i++) + { + if (i + index_offset > upperbound) + break; + to_oct ((long) sparsearray[i + index_offset].numbytes, + 1 + 12, + exhdr->ext_hdr.sp[i].numbytes); + to_oct ((long) sparsearray[i + index_offset].offset, + 1 + 12, + exhdr->ext_hdr.sp[i].offset); + } + userec (exhdr); + /* sum += i; + if (sum < upperbound) + goto extend;*/ + if (index_offset + i <= upperbound) + { + index_offset += i; + exhdr->ext_hdr.isextended++; + goto extend; + } + + } + if (save_linkflag == LF_SPARSE) + { + if (finish_sparse_file (f, &sizeleft, hstat.st_size, p)) + goto padit; + } + else + while (sizeleft > 0) + { + + if (f_multivol) + { + save_name = p; + save_sizeleft = sizeleft; + save_totsize = hstat.st_size; + } + start = findrec (); + + bufsize = endofrecs ()->charptr - start->charptr; + + if (sizeleft < bufsize) + { + /* Last read -- zero out area beyond */ + bufsize = (int) sizeleft; + count = bufsize % RECORDSIZE; + if (count) + bzero (start->charptr + sizeleft, + (int) (RECORDSIZE - count)); + } + count = read (f, start->charptr, bufsize); + if (count < 0) + { + msg_perror ("read error at byte %ld, reading\ + %d bytes, in file %s", hstat.st_size - sizeleft, bufsize, p); + goto padit; + } + sizeleft -= count; + + /* This is nonportable (the type of userec's arg). */ + userec (start + (count - 1) / RECORDSIZE); + + if (count == bufsize) + continue; + msg ("file %s shrunk by %d bytes, padding with zeros.", p, sizeleft); + goto padit; /* Short read */ + } + + if (f_multivol) + save_name = 0; + + if (f >= 0) + (void) close (f); + + if (f_remove_files) + { + if (unlink (p) == -1) + msg_perror ("cannot remove %s", p); + } + if (f_atime_preserve) + utime (p, &restore_times); + return; + + /* + * File shrunk or gave error, pad out tape to match + * the size we specified in the header. + */ + padit: + while (sizeleft > 0) + { + save_sizeleft = sizeleft; + start = findrec (); + bzero (start->charptr, RECORDSIZE); + userec (start); + sizeleft -= RECORDSIZE; + } + if (f_multivol) + save_name = 0; + if (f >= 0) + (void) close (f); + if (f_atime_preserve) + utime (p, &restore_times); + return; + } + +#ifdef S_ISLNK + else if (S_ISLNK (hstat.st_mode)) + { + int size; + char *buf = alloca (PATH_MAX + 1); + + size = readlink (p, buf, PATH_MAX + 1); + if (size < 0) + goto badperror; + buf[size] = '\0'; + if (size >= NAMSIZ) + write_long (buf, LF_LONGLINK); + current_link_name = buf; + + hstat.st_size = 0; /* Force 0 size on symlink */ + header = start_header (p, &hstat); + if (header == NULL) + { + critical_error = 1; + goto badfile; + } + strncpy (header->header.arch_linkname, buf, NAMSIZ); + header->header.arch_linkname[NAMSIZ - 1] = '\0'; + header->header.linkflag = LF_SYMLINK; + finish_header (header); /* Nothing more to do to it */ + if (f_remove_files) + { + if (unlink (p) == -1) + msg_perror ("cannot remove %s", p); + } + return; + } +#endif + + else if (S_ISDIR (hstat.st_mode)) + { + register DIR *dirp; + register struct dirent *d; + char *namebuf; + int buflen; + register int len; + int our_device = hstat.st_dev; + + /* Build new prototype name */ + len = strlen (p); + buflen = len + NAMSIZ; + namebuf = ck_malloc (buflen + 1); + strncpy (namebuf, p, buflen); + while (len >= 1 && '/' == namebuf[len - 1]) + len--; /* Delete trailing slashes */ + namebuf[len++] = '/'; /* Now add exactly one back */ + namebuf[len] = '\0'; /* Make sure null-terminated */ + + /* + * Output directory header record with permissions + * FIXME, do this AFTER files, to avoid R/O dir problems? + * If old archive format, don't write record at all. + */ + if (!f_oldarch) + { + hstat.st_size = 0; /* Force 0 size on dir */ + /* + * If people could really read standard archives, + * this should be: (FIXME) + header = start_header(f_standard? p: namebuf, &hstat); + * but since they'd interpret LF_DIR records as + * regular files, we'd better put the / on the name. + */ + header = start_header (namebuf, &hstat); + if (header == NULL) + { + critical_error = 1; + goto badfile; /* eg name too long */ + } + + if (f_gnudump) + header->header.linkflag = LF_DUMPDIR; + else if (f_standard) + header->header.linkflag = LF_DIR; + + /* If we're gnudumping, we aren't done yet so don't close it. */ + if (!f_gnudump) + finish_header (header); /* Done with directory header */ + } + + if (f_gnudump) + { + int sizeleft; + int totsize; + int bufsize; + union record *start; + int count; + char *buf, *p_buf; + + buf = gnu_list_name->dir_contents; /* FOO */ + totsize = 0; + for (p_buf = buf; p_buf && *p_buf;) + { + int tmp; + + tmp = strlen (p_buf) + 1; + totsize += tmp; + p_buf += tmp; + } + totsize++; + to_oct ((long) totsize, 1 + 12, header->header.size); + finish_header (header); + p_buf = buf; + sizeleft = totsize; + while (sizeleft > 0) + { + if (f_multivol) + { + save_name = p; + save_sizeleft = sizeleft; + save_totsize = totsize; + } + start = findrec (); + bufsize = endofrecs ()->charptr - start->charptr; + if (sizeleft < bufsize) + { + bufsize = sizeleft; + count = bufsize % RECORDSIZE; + if (count) + bzero (start->charptr + sizeleft, RECORDSIZE - count); + } + bcopy (p_buf, start->charptr, bufsize); + sizeleft -= bufsize; + p_buf += bufsize; + userec (start + (bufsize - 1) / RECORDSIZE); + } + if (f_multivol) + save_name = 0; + if (f_atime_preserve) + utime (p, &restore_times); + return; + } + + /* Now output all the files in the directory */ +#if 0 + if (f_dironly) + return; /* Unless the cmdline said not to */ +#endif + /* + * See if we are crossing from one file system to another, + * and avoid doing so if the user only wants to dump one file system. + */ + if (f_local_filesys && !toplevel && curdev != hstat.st_dev) + { + if (f_verbose) + msg ("%s: is on a different filesystem; not dumped", p); + return; + } + + + errno = 0; + dirp = opendir (p); + if (!dirp) + { + if (errno) + { + msg_perror ("can't open directory %s", p); + } + else + { + msg ("error opening directory %s", + p); + } + return; + } + + /* Hack to remove "./" from the front of all the file names */ + if (len == 2 && namebuf[0] == '.' && namebuf[1] == '/') + len = 0; + + /* Should speed this up by cd-ing into the dir, FIXME */ + while (NULL != (d = readdir (dirp))) + { + /* Skip . and .. */ + if (is_dot_or_dotdot (d->d_name)) + continue; + + if (NLENGTH (d) + len >= buflen) + { + buflen = len + NLENGTH (d); + namebuf = ck_realloc (namebuf, buflen + 1); + /* namebuf[len]='\0'; + msg("file name %s%s too long", + namebuf, d->d_name); + continue; */ + } + strcpy (namebuf + len, d->d_name); + if (f_exclude && check_exclude (namebuf)) + continue; + dump_file (namebuf, our_device, 0); + } + + closedir (dirp); + free (namebuf); + if (f_atime_preserve) + utime (p, &restore_times); + return; + } + +#ifdef S_ISCHR + else if (S_ISCHR (hstat.st_mode)) + { + type = LF_CHR; + } +#endif + +#ifdef S_ISBLK + else if (S_ISBLK (hstat.st_mode)) + { + type = LF_BLK; + } +#endif + + /* Avoid screwy apollo lossage where S_IFIFO == S_IFSOCK */ +#if (_ISP__M68K == 0) && (_ISP__A88K == 0) && defined(S_ISFIFO) + else if (S_ISFIFO (hstat.st_mode)) + { + type = LF_FIFO; + } +#endif + +#ifdef S_ISSOCK + else if (S_ISSOCK (hstat.st_mode)) + { + type = LF_FIFO; + } +#endif + else + goto unknown; + + if (!f_standard) + goto unknown; + + hstat.st_size = 0; /* Force 0 size */ + header = start_header (p, &hstat); + if (header == NULL) + { + critical_error = 1; + goto badfile; /* eg name too long */ + } + + header->header.linkflag = type; +#if defined(S_IFBLK) || defined(S_IFCHR) + if (type != LF_FIFO) + { + to_oct ((long) major (hstat.st_rdev), 8, + header->header.devmajor); + to_oct ((long) minor (hstat.st_rdev), 8, + header->header.devminor); + } +#endif + + finish_header (header); + if (f_remove_files) + { + if (unlink (p) == -1) + msg_perror ("cannot remove %s", p); + } + return; + +unknown: + msg ("%s: Unknown file type; file ignored.", p); +} + +int +finish_sparse_file (fd, sizeleft, fullsize, name) + int fd; + long *sizeleft, fullsize; + char *name; +{ + union record *start; + char tempbuf[RECORDSIZE]; + int bufsize, sparse_ind = 0, count; + long pos; + long nwritten = 0; + + + while (*sizeleft > 0) + { + start = findrec (); + bzero (start->charptr, RECORDSIZE); + bufsize = sparsearray[sparse_ind].numbytes; + if (!bufsize) + { /* we blew it, maybe */ + msg ("Wrote %ld of %ld bytes to file %s", + fullsize - *sizeleft, fullsize, name); + break; + } + pos = lseek (fd, sparsearray[sparse_ind++].offset, 0); + /* + * If the number of bytes to be written here exceeds + * the size of the temporary buffer, do it in steps. + */ + while (bufsize > RECORDSIZE) + { + /* if (amt_read) { + count = read(fd, start->charptr+amt_read, RECORDSIZE-amt_read); + bufsize -= RECORDSIZE - amt_read; + amt_read = 0; + userec(start); + start = findrec(); + bzero(start->charptr, RECORDSIZE); + }*/ + /* store the data */ + count = read (fd, start->charptr, RECORDSIZE); + if (count < 0) + { + msg_perror ("read error at byte %ld, reading %d bytes, in file %s", + fullsize - *sizeleft, bufsize, name); + return 1; + } + bufsize -= count; + *sizeleft -= count; + userec (start); + nwritten += RECORDSIZE; /* XXX */ + start = findrec (); + bzero (start->charptr, RECORDSIZE); + } + + + clear_buffer (tempbuf); + count = read (fd, tempbuf, bufsize); + bcopy (tempbuf, start->charptr, RECORDSIZE); + if (count < 0) + { + msg_perror ("read error at byte %ld, reading %d bytes, in file %s", + fullsize - *sizeleft, bufsize, name); + return 1; + } + /* if (amt_read >= RECORDSIZE) { + amt_read = 0; + userec(start+(count-1)/RECORDSIZE); + if (count != bufsize) { + msg("file %s shrunk by %d bytes, padding with zeros.", name, sizeleft); + return 1; + } + start = findrec(); + } else + amt_read += bufsize;*/ + nwritten += count; /* XXX */ + *sizeleft -= count; + userec (start); + + } + free (sparsearray); + /* printf ("Amount actually written is (I hope) %d.\n", nwritten); */ + /* userec(start+(count-1)/RECORDSIZE);*/ + return 0; + +} + +void +init_sparsearray () +{ + register int i; + + sp_array_size = 10; + /* + * Make room for our scratch space -- initially is 10 elts long + */ + sparsearray = (struct sp_array *) ck_malloc (sp_array_size * sizeof (struct sp_array)); + for (i = 0; i < sp_array_size; i++) + { + sparsearray[i].offset = 0; + sparsearray[i].numbytes = 0; + } +} + + + +/* + * Okay, we've got a sparse file on our hands -- now, what we need to do is + * make a pass through the file and carefully note where any data is, i.e., + * we want to find how far into the file each instance of data is, and how + * many bytes are there. We store this information in the sparsearray, + * which will later be translated into header information. For now, we use + * the sparsearray as convenient storage. + * + * As a side note, this routine is a mess. If I could have found a cleaner + * way to do it, I would have. If anyone wants to find a nicer way to do + * this, feel free. + */ + +/* There is little point in trimming small amounts of null data at the */ +/* head and tail of blocks -- it's ok if we only avoid dumping blocks */ +/* of complete null data */ +int +deal_with_sparse (name, header, nulls_at_end) + char *name; + union record *header; + int nulls_at_end; +{ + long numbytes = 0; + long offset = 0; + /* long save_offset;*/ + int fd; + /* int current_size = hstat.st_size;*/ + int sparse_ind = 0, cc; + char buf[RECORDSIZE]; +#if 0 + int read_last_data = 0; /* did we just read the last record? */ +#endif + int amidst_data = 0; + + header->header.isextended = 0; + /* + * Can't open the file -- this problem will be caught later on, + * so just return. + */ + if ((fd = open (name, O_RDONLY)) < 0) + return 0; + + init_sparsearray (); + clear_buffer (buf); + + while ((cc = read (fd, buf, sizeof buf)) != 0) + { + + if (sparse_ind > sp_array_size - 1) + { + + /* + * realloc the scratch area, since we've run out of room -- + */ + sparsearray = (struct sp_array *) + ck_realloc (sparsearray, + 2 * sp_array_size * (sizeof (struct sp_array))); + sp_array_size *= 2; + } + if (cc == sizeof buf) + { + if (zero_record (buf)) + { + if (amidst_data) + { + sparsearray[sparse_ind++].numbytes + = numbytes; + amidst_data = 0; + } + } + else + { /* !zero_record(buf) */ + if (amidst_data) + numbytes += cc; + else + { + amidst_data = 1; + numbytes = cc; + sparsearray[sparse_ind].offset + = offset; + } + } + } + else if (cc < sizeof buf) + { + /* This has to be the last bit of the file, so this */ + /* is somewhat shorter than the above. */ + if (!zero_record (buf)) + { + if (!amidst_data) + { + amidst_data = 1; + numbytes = cc; + sparsearray[sparse_ind].offset + = offset; + } + else + numbytes += cc; + } + } + offset += cc; + clear_buffer (buf); + } + if (amidst_data) + sparsearray[sparse_ind++].numbytes = numbytes; + else + { + sparsearray[sparse_ind].offset = offset-1; + sparsearray[sparse_ind++].numbytes = 1; + } + close (fd); + + return sparse_ind - 1; +} + +/* + * Just zeroes out the buffer so we don't confuse ourselves with leftover + * data. + */ +void +clear_buffer (buf) + char *buf; +{ + register int i; + + for (i = 0; i < RECORDSIZE; i++) + buf[i] = '\0'; +} + +#if 0 /* I'm leaving this as a monument to Joy Kendall, who wrote it -mib */ +/* + * JK - + * This routine takes a character array, and tells where within that array + * the data can be found. It skips over any zeros, and sets the first + * non-zero point in the array to be the "start", and continues until it + * finds non-data again, which is marked as the "end." This routine is + * mainly for 1) seeing how far into a file we must lseek to data, given + * that we have a sparse file, and 2) determining the "real size" of the + * file, i.e., the number of bytes in the sparse file that are data, as + * opposed to the zeros we are trying to skip. + */ +where_is_data (from, to, buffer) + int *from, *to; + char *buffer; +{ + register int i = 0; + register int save_to = *to; + int amidst_data = 0; + + + while (!buffer[i]) + i++; + *from = i; + + if (*from < 16) /* don't bother */ + *from = 0; + /* keep going to make sure there isn't more real + data in this record */ + while (i < RECORDSIZE) + { + if (!buffer[i]) + { + if (amidst_data) + { + save_to = i; + amidst_data = 0; + } + i++; + } + else if (buffer[i]) + { + if (!amidst_data) + amidst_data = 1; + i++; + } + } + if (i == RECORDSIZE) + *to = i; + else + *to = save_to; + +} + +#endif + +/* Note that this routine is only called if zero_record returned true */ +#if 0 /* But we actually don't need it at all. */ +where_is_data (from, to, buffer) + int *from, *to; + char *buffer; +{ + char *fp, *tp; + + for (fp = buffer; !*fp; fp++) + ; + for (tp = buffer + RECORDSIZE - 1; !*tp; tp--) + ; + *from = fp - buffer; + *to = tp - buffer + 1; +} + +#endif + + + +/* + * Takes a recordful of data and basically cruises through it to see if + * it's made *entirely* of zeros, returning a 0 the instant it finds + * something that is a non-zero, i.e., useful data. + */ +int +zero_record (buffer) + char *buffer; +{ + register int i; + + for (i = 0; i < RECORDSIZE; i++) + if (buffer[i] != '\000') + return 0; + return 1; +} + +void +find_new_file_size (filesize, highest_index) + int *filesize; + int highest_index; +{ + register int i; + + *filesize = 0; + for (i = 0; sparsearray[i].numbytes && i <= highest_index; i++) + *filesize += sparsearray[i].numbytes; +} + +/* + * Make a header block for the file name whose stat info is st . + * Return header pointer for success, NULL if the name is too long. + */ +union record * +start_header (name, st) + char *name; + register struct stat *st; +{ + register union record *header; + + if (strlen (name) >= NAMSIZ) + write_long (name, LF_LONGNAME); + + header = (union record *) findrec (); + bzero (header->charptr, sizeof (*header)); /* XXX speed up */ + + /* + * Check the file name and put it in the record. + */ + if (!f_absolute_paths) + { + static int warned_once = 0; +#ifdef __MSDOS__ + if (name[1] == ':') + { + name += 2; + if (!warned_once++) + msg ("Removing drive spec from names in the archive"); + } +#endif + while ('/' == *name) + { + name++; /* Force relative path */ + if (!warned_once++) + msg ("Removing leading / from absolute path names in the archive."); + } + } + current_file_name = name; + strncpy (header->header.arch_name, name, NAMSIZ); + header->header.arch_name[NAMSIZ - 1] = '\0'; + + to_oct ((long) (f_oldarch ? (st->st_mode & 07777) : st->st_mode), + 8, header->header.mode); + to_oct ((long) st->st_uid, 8, header->header.uid); + to_oct ((long) st->st_gid, 8, header->header.gid); + to_oct ((long) st->st_size, 1 + 12, header->header.size); + to_oct ((long) st->st_mtime, 1 + 12, header->header.mtime); + /* header->header.linkflag is left as null */ + if (f_gnudump) + { + to_oct ((long) st->st_atime, 1 + 12, header->header.atime); + to_oct ((long) st->st_ctime, 1 + 12, header->header.ctime); + } + +#ifndef NONAMES + /* Fill in new Unix Standard fields if desired. */ + if (f_standard) + { + header->header.linkflag = LF_NORMAL; /* New default */ + strcpy (header->header.magic, TMAGIC); /* Mark as Unix Std */ + finduname (header->header.uname, st->st_uid); + findgname (header->header.gname, st->st_gid); + } +#endif + return header; +} + +/* + * Finish off a filled-in header block and write it out. + * We also print the file name and/or full info if verbose is on. + */ +void +finish_header (header) + register union record *header; +{ + register int i, sum; + register char *p; + + bcopy (CHKBLANKS, header->header.chksum, sizeof (header->header.chksum)); + + sum = 0; + p = header->charptr; + for (i = sizeof (*header); --i >= 0;) + { + /* + * We can't use unsigned char here because of old compilers, + * e.g. V7. + */ + sum += 0xFF & *p++; + } + + /* + * Fill in the checksum field. It's formatted differently + * from the other fields: it has [6] digits, a null, then a + * space -- rather than digits, a space, then a null. + * We use to_oct then write the null in over to_oct's space. + * The final space is already there, from checksumming, and + * to_oct doesn't modify it. + * + * This is a fast way to do: + * (void) sprintf(header->header.chksum, "%6o", sum); + */ + to_oct ((long) sum, 8, header->header.chksum); + header->header.chksum[6] = '\0'; /* Zap the space */ + + userec (header); + + if (f_verbose) + { + extern union record *head;/* Points to current tape header */ + extern int head_standard; /* Tape header is in ANSI format */ + + /* These globals are parameters to print_header, sigh */ + head = header; + /* hstat is already set up */ + head_standard = f_standard; + print_header (); + } + + return; +} + + +/* + * Quick and dirty octal conversion. + * Converts long "value" into a "digs"-digit field at "where", + * including a trailing space and room for a null. "digs"==3 means + * 1 digit, a space, and room for a null. + * + * We assume the trailing null is already there and don't fill it in. + * This fact is used by start_header and finish_header, so don't change it! + * + * This should be equivalent to: + * (void) sprintf(where, "%*lo ", digs-2, value); + * except that sprintf fills in the trailing null and we don't. + */ +void +to_oct (value, digs, where) + register long value; + register int digs; + register char *where; +{ + + --digs; /* Trailing null slot is left alone */ + where[--digs] = ' '; /* Put in the space, though */ + + /* Produce the digits -- at least one */ + do + { + where[--digs] = '0' + (char) (value & 7); /* one octal digit */ + value >>= 3; + } + while (digs > 0 && value != 0); + + /* Leading spaces, if necessary */ + while (digs > 0) + where[--digs] = ' '; + +} + + +/* + * Write the EOT record(s). + * We actually zero at least one record, through the end of the block. + * Old tar writes garbage after two zeroed records -- and PDtar used to. + */ +void +write_eot () +{ + union record *p; + int bufsize; + + p = findrec (); + if (p) + { + bufsize = endofrecs ()->charptr - p->charptr; + bzero (p->charptr, bufsize); + userec (p); + } +} + +/* Write a LF_LONGLINK or LF_LONGNAME record. */ +void +write_long (p, type) + char *p; + char type; +{ + int size = strlen (p) + 1; + int bufsize; + union record *header; + struct stat foo; + + + bzero (&foo, sizeof foo); + foo.st_size = size; + + header = start_header ("././@LongLink", &foo); + header->header.linkflag = type; + finish_header (header); + + header = findrec (); + + bufsize = endofrecs ()->charptr - header->charptr; + + while (bufsize < size) + { + bcopy (p, header->charptr, bufsize); + p += bufsize; + size -= bufsize; + userec (header + (bufsize - 1) / RECORDSIZE); + header = findrec (); + bufsize = endofrecs ()->charptr - header->charptr; + } + bcopy (p, header->charptr, size); + bzero (header->charptr + size, bufsize - size); + userec (header + (size - 1) / RECORDSIZE); +} diff --git a/gnu/usr.bin/tar/diffarch.c b/gnu/usr.bin/tar/diffarch.c new file mode 100644 index 000000000000..ce47d9d6cab5 --- /dev/null +++ b/gnu/usr.bin/tar/diffarch.c @@ -0,0 +1,759 @@ +/* Diff files from a tar archive. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * Diff files from a tar archive. + * + * Written 30 April 1987 by John Gilmore, ihnp4!hoptoad!gnu. + */ + +#include <stdio.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif +#include <sys/types.h> + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#ifdef HAVE_SYS_MTIO_H +#include <sys/ioctl.h> +#include <sys/mtio.h> +#endif + +#include "tar.h" +#include "port.h" +#include "rmt.h" + +#ifndef S_ISLNK +#define lstat stat +#endif + +extern void *valloc (); + +extern union record *head; /* Points to current tape header */ +extern struct stat hstat; /* Stat struct corresponding */ +extern int head_standard; /* Tape header is in ANSI format */ + +void decode_header (); +void diff_sparse_files (); +void fill_in_sparse_array (); +void fl_read (); +long from_oct (); +int do_stat (); +extern void print_header (); +int read_header (); +void saverec (); +void sigh (); +extern void skip_file (); +extern void skip_extended_headers (); +int wantbytes (); + +extern FILE *msg_file; + +int now_verifying = 0; /* Are we verifying at the moment? */ + +int diff_fd; /* Descriptor of file we're diffing */ + +char *diff_buf = 0; /* Pointer to area for reading + file contents into */ + +char *diff_dir; /* Directory contents for LF_DUMPDIR */ + +int different = 0; + +/*struct sp_array *sparsearray; +int sp_ar_size = 10;*/ +/* + * Initialize for a diff operation + */ +void +diff_init () +{ + /*NOSTRICT*/ + diff_buf = (char *) valloc ((unsigned) blocksize); + if (!diff_buf) + { + msg ("could not allocate memory for diff buffer of %d bytes", + blocksize); + exit (EX_ARGSBAD); + } +} + +/* + * Diff a file against the archive. + */ +void +diff_archive () +{ + register char *data; + int check, namelen; + int err; + long offset; + struct stat filestat; + int compare_chunk (); + int compare_dir (); + int no_op (); +#ifndef __MSDOS__ + dev_t dev; + ino_t ino; +#endif + char *get_dir_contents (); + long from_oct (); + + errno = EPIPE; /* FIXME, remove perrors */ + + saverec (&head); /* Make sure it sticks around */ + userec (head); /* And go past it in the archive */ + decode_header (head, &hstat, &head_standard, 1); /* Snarf fields */ + + /* Print the record from 'head' and 'hstat' */ + if (f_verbose) + { + if (now_verifying) + fprintf (msg_file, "Verify "); + print_header (); + } + + switch (head->header.linkflag) + { + + default: + msg ("Unknown file type '%c' for %s, diffed as normal file", + head->header.linkflag, current_file_name); + /* FALL THRU */ + + case LF_OLDNORMAL: + case LF_NORMAL: + case LF_SPARSE: + case LF_CONTIG: + /* + * Appears to be a file. + * See if it's really a directory. + */ + namelen = strlen (current_file_name) - 1; + if (current_file_name[namelen] == '/') + goto really_dir; + + + if (do_stat (&filestat)) + { + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) hstat.st_size); + different++; + goto quit; + } + + if (!S_ISREG (filestat.st_mode)) + { + fprintf (msg_file, "%s: not a regular file\n", + current_file_name); + skip_file ((long) hstat.st_size); + different++; + goto quit; + } + + filestat.st_mode &= 07777; + if (filestat.st_mode != hstat.st_mode) + sigh ("mode"); + if (filestat.st_uid != hstat.st_uid) + sigh ("uid"); + if (filestat.st_gid != hstat.st_gid) + sigh ("gid"); + if (filestat.st_mtime != hstat.st_mtime) + sigh ("mod time"); + if (head->header.linkflag != LF_SPARSE && + filestat.st_size != hstat.st_size) + { + sigh ("size"); + skip_file ((long) hstat.st_size); + goto quit; + } + + diff_fd = open (current_file_name, O_NDELAY | O_RDONLY | O_BINARY); + + if (diff_fd < 0 && !f_absolute_paths) + { + char tmpbuf[NAMSIZ + 2]; + + tmpbuf[0] = '/'; + strcpy (&tmpbuf[1], current_file_name); + diff_fd = open (tmpbuf, O_NDELAY | O_RDONLY); + } + if (diff_fd < 0) + { + msg_perror ("cannot open %s", current_file_name); + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) hstat.st_size); + different++; + goto quit; + } + /* + * Need to treat sparse files completely differently here. + */ + if (head->header.linkflag == LF_SPARSE) + diff_sparse_files (hstat.st_size); + else + wantbytes ((long) (hstat.st_size), compare_chunk); + + check = close (diff_fd); + if (check < 0) + msg_perror ("Error while closing %s", current_file_name); + + quit: + break; + +#ifndef __MSDOS__ + case LF_LINK: + if (do_stat (&filestat)) + break; + dev = filestat.st_dev; + ino = filestat.st_ino; + err = stat (current_link_name, &filestat); + if (err < 0) + { + if (errno == ENOENT) + { + fprintf (msg_file, "%s: does not exist\n", current_file_name); + } + else + { + msg_perror ("cannot stat file %s", current_file_name); + } + different++; + break; + } + if (filestat.st_dev != dev || filestat.st_ino != ino) + { + fprintf (msg_file, "%s not linked to %s\n", current_file_name, current_link_name); + break; + } + break; +#endif + +#ifdef S_ISLNK + case LF_SYMLINK: + { + char linkbuf[NAMSIZ + 3]; + check = readlink (current_file_name, linkbuf, + (sizeof linkbuf) - 1); + + if (check < 0) + { + if (errno == ENOENT) + { + fprintf (msg_file, + "%s: no such file or directory\n", + current_file_name); + } + else + { + msg_perror ("cannot read link %s", current_file_name); + } + different++; + break; + } + + linkbuf[check] = '\0'; /* Null-terminate it */ + if (strncmp (current_link_name, linkbuf, check) != 0) + { + fprintf (msg_file, "%s: symlink differs\n", + current_link_name); + different++; + } + } + break; +#endif + +#ifdef S_IFCHR + case LF_CHR: + hstat.st_mode |= S_IFCHR; + goto check_node; +#endif + +#ifdef S_IFBLK + /* If local system doesn't support block devices, use default case */ + case LF_BLK: + hstat.st_mode |= S_IFBLK; + goto check_node; +#endif + +#ifdef S_ISFIFO + /* If local system doesn't support FIFOs, use default case */ + case LF_FIFO: +#ifdef S_IFIFO + hstat.st_mode |= S_IFIFO; +#endif + hstat.st_rdev = 0; /* FIXME, do we need this? */ + goto check_node; +#endif + + check_node: + /* FIXME, deal with umask */ + if (do_stat (&filestat)) + break; + if (hstat.st_rdev != filestat.st_rdev) + { + fprintf (msg_file, "%s: device numbers changed\n", current_file_name); + different++; + break; + } +#ifdef S_IFMT + if (hstat.st_mode != filestat.st_mode) +#else /* POSIX lossage */ + if ((hstat.st_mode & 07777) != (filestat.st_mode & 07777)) +#endif + { + fprintf (msg_file, "%s: mode or device-type changed\n", current_file_name); + different++; + break; + } + break; + + case LF_DUMPDIR: + data = diff_dir = get_dir_contents (current_file_name, 0); + if (data) + { + wantbytes ((long) (hstat.st_size), compare_dir); + free (data); + } + else + wantbytes ((long) (hstat.st_size), no_op); + /* FALL THROUGH */ + + case LF_DIR: + /* Check for trailing / */ + namelen = strlen (current_file_name) - 1; + really_dir: + while (namelen && current_file_name[namelen] == '/') + current_file_name[namelen--] = '\0'; /* Zap / */ + + if (do_stat (&filestat)) + break; + if (!S_ISDIR (filestat.st_mode)) + { + fprintf (msg_file, "%s is no longer a directory\n", current_file_name); + different++; + break; + } + if ((filestat.st_mode & 07777) != (hstat.st_mode & 07777)) + sigh ("mode"); + break; + + case LF_VOLHDR: + break; + + case LF_MULTIVOL: + namelen = strlen (current_file_name) - 1; + if (current_file_name[namelen] == '/') + goto really_dir; + + if (do_stat (&filestat)) + break; + + if (!S_ISREG (filestat.st_mode)) + { + fprintf (msg_file, "%s: not a regular file\n", + current_file_name); + skip_file ((long) hstat.st_size); + different++; + break; + } + + filestat.st_mode &= 07777; + offset = from_oct (1 + 12, head->header.offset); + if (filestat.st_size != hstat.st_size + offset) + { + sigh ("size"); + skip_file ((long) hstat.st_size); + different++; + break; + } + + diff_fd = open (current_file_name, O_NDELAY | O_RDONLY | O_BINARY); + + if (diff_fd < 0) + { + msg_perror ("cannot open file %s", current_file_name); + skip_file ((long) hstat.st_size); + different++; + break; + } + err = lseek (diff_fd, offset, 0); + if (err != offset) + { + msg_perror ("cannot seek to %ld in file %s", offset, current_file_name); + different++; + break; + } + + wantbytes ((long) (hstat.st_size), compare_chunk); + + check = close (diff_fd); + if (check < 0) + { + msg_perror ("Error while closing %s", current_file_name); + } + break; + + } + + /* We don't need to save it any longer. */ + saverec ((union record **) 0);/* Unsave it */ +} + +int +compare_chunk (bytes, buffer) + long bytes; + char *buffer; +{ + int err; + + err = read (diff_fd, diff_buf, bytes); + if (err != bytes) + { + if (err < 0) + { + msg_perror ("can't read %s", current_file_name); + } + else + { + fprintf (msg_file, "%s: could only read %d of %d bytes\n", current_file_name, err, bytes); + } + different++; + return -1; + } + if (bcmp (buffer, diff_buf, bytes)) + { + fprintf (msg_file, "%s: data differs\n", current_file_name); + different++; + return -1; + } + return 0; +} + +int +compare_dir (bytes, buffer) + long bytes; + char *buffer; +{ + if (bcmp (buffer, diff_dir, bytes)) + { + fprintf (msg_file, "%s: data differs\n", current_file_name); + different++; + return -1; + } + diff_dir += bytes; + return 0; +} + +/* + * Sigh about something that differs. + */ +void +sigh (what) + char *what; +{ + + fprintf (msg_file, "%s: %s differs\n", + current_file_name, what); +} + +void +verify_volume () +{ + int status; +#ifdef MTIOCTOP + struct mtop t; + int er; +#endif + + if (!diff_buf) + diff_init (); +#ifdef MTIOCTOP + t.mt_op = MTBSF; + t.mt_count = 1; + if ((er = rmtioctl (archive, MTIOCTOP, &t)) < 0) + { + if (errno != EIO || (er = rmtioctl (archive, MTIOCTOP, &t)) < 0) + { +#endif + if (rmtlseek (archive, 0L, 0) != 0) + { + /* Lseek failed. Try a different method */ + msg_perror ("Couldn't rewind archive file for verify"); + return; + } +#ifdef MTIOCTOP + } + } +#endif + ar_reading = 1; + now_verifying = 1; + fl_read (); + for (;;) + { + status = read_header (); + if (status == 0) + { + unsigned n; + + n = 0; + do + { + n++; + status = read_header (); + } + while (status == 0); + msg ("VERIFY FAILURE: %d invalid header%s detected!", n, n == 1 ? "" : "s"); + } + if (status == 2 || status == EOF) + break; + diff_archive (); + } + ar_reading = 0; + now_verifying = 0; + +} + +int +do_stat (statp) + struct stat *statp; +{ + int err; + + err = f_follow_links ? stat (current_file_name, statp) : lstat (current_file_name, statp); + if (err < 0) + { + if (errno == ENOENT) + { + fprintf (msg_file, "%s: does not exist\n", current_file_name); + } + else + msg_perror ("can't stat file %s", current_file_name); + /* skip_file((long)hstat.st_size); + different++;*/ + return 1; + } + else + return 0; +} + +/* + * JK + * Diff'ing a sparse file with its counterpart on the tar file is a + * bit of a different story than a normal file. First, we must know + * what areas of the file to skip through, i.e., we need to contruct + * a sparsearray, which will hold all the information we need. We must + * compare small amounts of data at a time as we find it. + */ + +void +diff_sparse_files (filesize) + int filesize; + +{ + int sparse_ind = 0; + char *buf; + int buf_size = RECORDSIZE; + union record *datarec; + int err; + long numbytes; + /* int amt_read = 0;*/ + int size = filesize; + + buf = (char *) ck_malloc (buf_size * sizeof (char)); + + fill_in_sparse_array (); + + + while (size > 0) + { + datarec = findrec (); + if (!sparsearray[sparse_ind].numbytes) + break; + + /* + * 'numbytes' is nicer to write than + * 'sparsearray[sparse_ind].numbytes' all the time ... + */ + numbytes = sparsearray[sparse_ind].numbytes; + + lseek (diff_fd, sparsearray[sparse_ind].offset, 0); + /* + * take care to not run out of room in our buffer + */ + while (buf_size < numbytes) + { + buf = (char *) ck_realloc (buf, buf_size * 2 * sizeof (char)); + buf_size *= 2; + } + while (numbytes > RECORDSIZE) + { + if ((err = read (diff_fd, buf, RECORDSIZE)) != RECORDSIZE) + { + if (err < 0) + msg_perror ("can't read %s", current_file_name); + else + fprintf (msg_file, "%s: could only read %d of %d bytes\n", + current_file_name, err, numbytes); + break; + } + if (bcmp (buf, datarec->charptr, RECORDSIZE)) + { + different++; + break; + } + numbytes -= err; + size -= err; + userec (datarec); + datarec = findrec (); + } + if ((err = read (diff_fd, buf, numbytes)) != numbytes) + { + if (err < 0) + msg_perror ("can't read %s", current_file_name); + else + fprintf (msg_file, "%s: could only read %d of %d bytes\n", + current_file_name, err, numbytes); + break; + } + + if (bcmp (buf, datarec->charptr, numbytes)) + { + different++; + break; + } + /* amt_read += numbytes; + if (amt_read >= RECORDSIZE) { + amt_read = 0; + userec(datarec); + datarec = findrec(); + }*/ + userec (datarec); + sparse_ind++; + size -= numbytes; + } + /* + * if the number of bytes read isn't the + * number of bytes supposedly in the file, + * they're different + */ + /* if (amt_read != filesize) + different++;*/ + userec (datarec); + free (sparsearray); + if (different) + fprintf (msg_file, "%s: data differs\n", current_file_name); + +} + +/* + * JK + * This routine should be used more often than it is ... look into + * that. Anyhow, what it does is translate the sparse information + * on the header, and in any subsequent extended headers, into an + * array of structures with true numbers, as opposed to character + * strings. It simply makes our life much easier, doing so many + * comparisong and such. + */ +void +fill_in_sparse_array () +{ + int ind; + + /* + * allocate space for our scratch space; it's initially + * 10 elements long, but can change in this routine if + * necessary + */ + sp_array_size = 10; + sparsearray = (struct sp_array *) ck_malloc (sp_array_size * sizeof (struct sp_array)); + + /* + * there are at most five of these structures in the header + * itself; read these in first + */ + for (ind = 0; ind < SPARSE_IN_HDR; ind++) + { + if (!head->header.sp[ind].numbytes) + break; + sparsearray[ind].offset = + from_oct (1 + 12, head->header.sp[ind].offset); + sparsearray[ind].numbytes = + from_oct (1 + 12, head->header.sp[ind].numbytes); + } + /* + * if the header's extended, we gotta read in exhdr's till + * we're done + */ + if (head->header.isextended) + { + /* how far into the sparsearray we are 'so far' */ + static int so_far_ind = SPARSE_IN_HDR; + union record *exhdr; + + for (;;) + { + exhdr = findrec (); + for (ind = 0; ind < SPARSE_EXT_HDR; ind++) + { + if (ind + so_far_ind > sp_array_size - 1) + { + /* + * we just ran out of room in our + * scratch area - realloc it + */ + sparsearray = (struct sp_array *) + ck_realloc (sparsearray, + sp_array_size * 2 * sizeof (struct sp_array)); + sp_array_size *= 2; + } + /* + * convert the character strings into longs + */ + sparsearray[ind + so_far_ind].offset = + from_oct (1 + 12, exhdr->ext_hdr.sp[ind].offset); + sparsearray[ind + so_far_ind].numbytes = + from_oct (1 + 12, exhdr->ext_hdr.sp[ind].numbytes); + } + /* + * if this is the last extended header for this + * file, we can stop + */ + if (!exhdr->ext_hdr.isextended) + break; + else + { + so_far_ind += SPARSE_EXT_HDR; + userec (exhdr); + } + } + /* be sure to skip past the last one */ + userec (exhdr); + } +} diff --git a/gnu/usr.bin/tar/extract.c b/gnu/usr.bin/tar/extract.c new file mode 100644 index 000000000000..d162cab04ec6 --- /dev/null +++ b/gnu/usr.bin/tar/extract.c @@ -0,0 +1,907 @@ +/* Extract files from a tar archive. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * Extract files from a tar archive. + * + * Written 19 Nov 1985 by John Gilmore, ihnp4!hoptoad!gnu. + */ + +#include <stdio.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif +#include <sys/types.h> +#include <time.h> +time_t time (); + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#ifdef NO_OPEN3 +/* We need the #define's even though we don't use them. */ +#include "open3.h" +#endif + +#ifdef EMUL_OPEN3 +/* Simulated 3-argument open for systems that don't have it */ +#include "open3.h" +#endif + +#include "tar.h" +#include "port.h" + +#if defined(_POSIX_VERSION) +#include <utime.h> +#else +struct utimbuf +{ + long actime; + long modtime; +}; + +#endif + +extern FILE *msg_file; + +extern union record *head; /* Points to current tape header */ +extern struct stat hstat; /* Stat struct corresponding */ +extern int head_standard; /* Tape header is in ANSI format */ + +extern char *save_name; +extern long save_totsize; +extern long save_sizeleft; + +int confirm (); +void decode_header (); +void extract_mangle (); +void extract_sparse_file (); +long from_oct (); +void gnu_restore (); +extern void print_header (); +extern void skip_file (); +extern void skip_extended_headers (); +extern void pr_mkdir (); +void saverec (); + +int make_dirs (); /* Makes required directories */ + +static time_t now = 0; /* Current time */ +static we_are_root = 0; /* True if our effective uid == 0 */ +static int notumask = ~0; /* Masks out bits user doesn't want */ + +/* + * "Scratch" space to store the information about a sparse file before + * writing the info into the header or extended header + */ +/*struct sp_array *sparsearray;*/ + +/* number of elts storable in the sparsearray */ +/*int sp_array_size = 10;*/ + +struct saved_dir_info +{ + char *path; + int mode; + int atime; + int mtime; + struct saved_dir_info *next; +}; + +struct saved_dir_info *saved_dir_info_head; + +/* + * Set up to extract files. + */ +void +extr_init () +{ + int ourmask; + + now = time ((time_t *) 0); + if (geteuid () == 0) + we_are_root = 1; + + /* + * We need to know our umask. But if f_use_protection is set, + * leave our kernel umask at 0, and our "notumask" at ~0. + */ + ourmask = umask (0); /* Read it */ + if (!f_use_protection) + { + (void) umask (ourmask); /* Set it back how it was */ + notumask = ~ourmask; /* Make umask override permissions */ + } +} + + +/* + * Extract a file from the archive. + */ +void +extract_archive () +{ + register char *data; + int fd, check, namelen, written, openflag; + long size; + struct utimbuf acc_upd_times; + register int skipcrud; + register int i; + /* int sparse_ind = 0;*/ + union record *exhdr; + struct saved_dir_info *tmp; + /* int end_nulls; */ + + saverec (&head); /* Make sure it sticks around */ + userec (head); /* And go past it in the archive */ + decode_header (head, &hstat, &head_standard, 1); /* Snarf fields */ + + if (f_confirm && !confirm ("extract", current_file_name)) + { + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) hstat.st_size); + saverec ((union record **) 0); + return; + } + + /* Print the record from 'head' and 'hstat' */ + if (f_verbose) + print_header (); + + /* + * Check for fully specified pathnames and other atrocities. + * + * Note, we can't just make a pointer to the new file name, + * since saverec() might move the header and adjust "head". + * We have to start from "head" every time we want to touch + * the header record. + */ + skipcrud = 0; + while (!f_absolute_paths + && '/' == current_file_name[skipcrud]) + { + static int warned_once = 0; + + skipcrud++; /* Force relative path */ + if (!warned_once++) + { + msg ("Removing leading / from absolute path names in the archive."); + } + } + + switch (head->header.linkflag) + { + + default: + msg ("Unknown file type '%c' for %s, extracted as normal file", + head->header.linkflag, skipcrud + current_file_name); + /* FALL THRU */ + + /* + * JK - What we want to do if the file is sparse is loop through + * the array of sparse structures in the header and read in + * and translate the character strings representing 1) the offset + * at which to write and 2) how many bytes to write into numbers, + * which we store into the scratch array, "sparsearray". This + * array makes our life easier the same way it did in creating + * the tar file that had to deal with a sparse file. + * + * After we read in the first five (at most) sparse structures, + * we check to see if the file has an extended header, i.e., + * if more sparse structures are needed to describe the contents + * of the new file. If so, we read in the extended headers + * and continue to store their contents into the sparsearray. + */ + case LF_SPARSE: + sp_array_size = 10; + sparsearray = (struct sp_array *) ck_malloc (sp_array_size * sizeof (struct sp_array)); + for (i = 0; i < SPARSE_IN_HDR; i++) + { + sparsearray[i].offset = + from_oct (1 + 12, head->header.sp[i].offset); + sparsearray[i].numbytes = + from_oct (1 + 12, head->header.sp[i].numbytes); + if (!sparsearray[i].numbytes) + break; + } + + /* end_nulls = from_oct(1+12, head->header.ending_blanks);*/ + + if (head->header.isextended) + { + /* read in the list of extended headers + and translate them into the sparsearray + as before */ + + /* static */ int ind = SPARSE_IN_HDR; + + for (;;) + { + + exhdr = findrec (); + for (i = 0; i < SPARSE_EXT_HDR; i++) + { + + if (i + ind > sp_array_size - 1) + { + /* + * realloc the scratch area + * since we've run out of room -- + */ + sparsearray = (struct sp_array *) + ck_realloc (sparsearray, + 2 * sp_array_size * (sizeof (struct sp_array))); + sp_array_size *= 2; + } + if (!exhdr->ext_hdr.sp[i].numbytes) + break; + sparsearray[i + ind].offset = + from_oct (1 + 12, exhdr->ext_hdr.sp[i].offset); + sparsearray[i + ind].numbytes = + from_oct (1 + 12, exhdr->ext_hdr.sp[i].numbytes); + } + if (!exhdr->ext_hdr.isextended) + break; + else + { + ind += SPARSE_EXT_HDR; + userec (exhdr); + } + } + userec (exhdr); + } + + /* FALL THRU */ + case LF_OLDNORMAL: + case LF_NORMAL: + case LF_CONTIG: + /* + * Appears to be a file. + * See if it's really a directory. + */ + namelen = strlen (skipcrud + current_file_name) - 1; + if (current_file_name[skipcrud + namelen] == '/') + goto really_dir; + + /* FIXME, deal with protection issues */ + again_file: + openflag = (f_keep ? + O_BINARY | O_NDELAY | O_WRONLY | O_CREAT | O_EXCL : + O_BINARY | O_NDELAY | O_WRONLY | O_CREAT | O_TRUNC) + | ((head->header.linkflag == LF_SPARSE) ? 0 : O_APPEND); + /* + * JK - The last | is a kludge to solve the problem + * the O_APPEND flag causes with files we are + * trying to make sparse: when a file is opened + * with O_APPEND, it writes to the last place + * that something was written, thereby ignoring + * any lseeks that we have done. We add this + * extra condition to make it able to lseek when + * a file is sparse, i.e., we don't open the new + * file with this flag. (Grump -- this bug caused + * me to waste a good deal of time, I might add) + */ + + if (f_exstdout) + { + fd = 1; + goto extract_file; + } +#ifdef O_CTG + /* + * Contiguous files (on the Masscomp) have to specify + * the size in the open call that creates them. + */ + if (head->header.linkflag == LF_CONTIG) + fd = open ((longname ? longname : head->header.name) + + skipcrud, + openflag | O_CTG, + hstat.st_mode, hstat.st_size); + else +#endif + { +#ifdef NO_OPEN3 + /* + * On raw V7 we won't let them specify -k (f_keep), but + * we just bull ahead and create the files. + */ + fd = creat ((longname + ? longname + : head->header.name) + skipcrud, + hstat.st_mode); +#else + /* + * With 3-arg open(), we can do this up right. + */ + fd = open (skipcrud + current_file_name, + openflag, hstat.st_mode); +#endif + } + + if (fd < 0) + { + if (make_dirs (skipcrud + current_file_name)) + goto again_file; + msg_perror ("Could not create file %s", + skipcrud + current_file_name); + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) hstat.st_size); + goto quit; + } + + extract_file: + if (head->header.linkflag == LF_SPARSE) + { + char *name; + int namelen; + + /* + * Kludge alert. NAME is assigned to header.name + * because during the extraction, the space that + * contains the header will get scribbled on, and + * the name will get munged, so any error messages + * that happen to contain the filename will look + * REAL interesting unless we do this. + */ + namelen = strlen (skipcrud + current_file_name) + 1; + name = (char *) ck_malloc ((sizeof (char)) * namelen); + bcopy (skipcrud + current_file_name, name, namelen); + size = hstat.st_size; + extract_sparse_file (fd, &size, hstat.st_size, name); + } + else + for (size = hstat.st_size; + size > 0; + size -= written) + { + + /* long offset, + numbytes;*/ + + if (f_multivol) + { + save_name = current_file_name; + save_totsize = hstat.st_size; + save_sizeleft = size; + } + + /* + * Locate data, determine max length + * writeable, write it, record that + * we have used the data, then check + * if the write worked. + */ + data = findrec ()->charptr; + if (data == NULL) + { /* Check it... */ + msg ("Unexpected EOF on archive file"); + break; + } + /* + * JK - If the file is sparse, use the sparsearray + * that we created before to lseek into the new + * file the proper amount, and to see how many + * bytes we want to write at that position. + */ + /* if (head->header.linkflag == LF_SPARSE) { + off_t pos; + + pos = lseek(fd, (off_t) sparsearray[sparse_ind].offset, 0); + printf("%d at %d\n", (int) pos, sparse_ind); + written = sparsearray[sparse_ind++].numbytes; + } else*/ + written = endofrecs ()->charptr - data; + if (written > size) + written = size; + errno = 0; + check = write (fd, data, written); + /* + * The following is in violation of strict + * typing, since the arg to userec + * should be a struct rec *. FIXME. + */ + userec ((union record *) (data + written - 1)); + if (check == written) + continue; + /* + * Error in writing to file. + * Print it, skip to next file in archive. + */ + if (check < 0) + msg_perror ("couldn't write to file %s", + skipcrud + current_file_name); + else + msg ("could only write %d of %d bytes to file %s", + check, written, skipcrud + current_file_name); + skip_file ((long) (size - written)); + break; /* Still do the close, mod time, chmod, etc */ + } + + if (f_multivol) + save_name = 0; + + /* If writing to stdout, don't try to do anything + to the filename; it doesn't exist, or we don't + want to touch it anyway */ + if (f_exstdout) + break; + + /* if (head->header.isextended) { + register union record *exhdr; + register int i; + + for (i = 0; i < 21; i++) { + long offset; + + if (!exhdr->ext_hdr.sp[i].numbytes) + break; + offset = from_oct(1+12, + exhdr->ext_hdr.sp[i].offset); + written = from_oct(1+12, + exhdr->ext_hdr.sp[i].numbytes); + lseek(fd, offset, 0); + check = write(fd, data, written); + if (check == written) continue; + + } + + + }*/ + check = close (fd); + if (check < 0) + { + msg_perror ("Error while closing %s", + skipcrud + current_file_name); + } + + + set_filestat: + + /* + * If we are root, set the owner and group of the extracted + * file. This does what is wanted both on real Unix and on + * System V. If we are running as a user, we extract as that + * user; if running as root, we extract as the original owner. + */ + if (we_are_root || f_do_chown) + { + if (chown (skipcrud + current_file_name, + hstat.st_uid, hstat.st_gid) < 0) + { + msg_perror ("cannot chown file %s to uid %d gid %d", + skipcrud + current_file_name, + hstat.st_uid, hstat.st_gid); + } + } + + /* + * Set the modified time of the file. + * + * Note that we set the accessed time to "now", which + * is really "the time we started extracting files". + * unless f_gnudump is used, in which case .st_atime is used + */ + if (!f_modified) + { + /* fixme if f_gnudump should set ctime too, but how? */ + if (f_gnudump) + acc_upd_times.actime = hstat.st_atime; + else + acc_upd_times.actime = now; /* Accessed now */ + acc_upd_times.modtime = hstat.st_mtime; /* Mod'd */ + if (utime (skipcrud + current_file_name, + &acc_upd_times) < 0) + { + msg_perror ("couldn't change access and modification times of %s", skipcrud + current_file_name); + } + } + /* We do the utime before the chmod because some versions of + utime are broken and trash the modes of the file. Since + we then change the mode anyway, we don't care. . . */ + + /* + * If '-k' is not set, open() or creat() could have saved + * the permission bits from a previously created file, + * ignoring the ones we specified. + * Even if -k is set, if the file has abnormal + * mode bits, we must chmod since writing or chown() has + * probably reset them. + * + * If -k is set, we know *we* created this file, so the mode + * bits were set by our open(). If the file is "normal", we + * skip the chmod. This works because we did umask(0) if -p + * is set, so umask will have left the specified mode alone. + */ + if ((!f_keep) + || (hstat.st_mode & (S_ISUID | S_ISGID | S_ISVTX))) + { + if (chmod (skipcrud + current_file_name, + notumask & (int) hstat.st_mode) < 0) + { + msg_perror ("cannot change mode of file %s to %ld", + skipcrud + current_file_name, + notumask & (int) hstat.st_mode); + } + } + + quit: + break; + + case LF_LINK: + again_link: + { + struct stat st1, st2; + + check = link (current_link_name, skipcrud + current_file_name); + + if (check == 0) + break; + if (make_dirs (skipcrud + current_file_name)) + goto again_link; + if (f_gnudump && errno == EEXIST) + break; + if (stat (current_link_name, &st1) == 0 + && stat (current_file_name + skipcrud, &st2) == 0 + && st1.st_dev == st2.st_dev + && st1.st_ino == st2.st_ino) + break; + msg_perror ("Could not link %s to %s", + skipcrud + current_file_name, + current_link_name); + } + break; + +#ifdef S_ISLNK + case LF_SYMLINK: + again_symlink: + check = symlink (current_link_name, + skipcrud + current_file_name); + /* FIXME, don't worry uid, gid, etc... */ + if (check == 0) + break; + if (make_dirs (current_file_name + skipcrud)) + goto again_symlink; + msg_perror ("Could not create symlink to %s", + current_link_name); + break; +#endif + +#ifdef S_IFCHR + case LF_CHR: + hstat.st_mode |= S_IFCHR; + goto make_node; +#endif + +#ifdef S_IFBLK + case LF_BLK: + hstat.st_mode |= S_IFBLK; +#endif +#if defined(S_IFCHR) || defined(S_IFBLK) + make_node: + check = mknod (current_file_name + skipcrud, + (int) hstat.st_mode, (int) hstat.st_rdev); + if (check != 0) + { + if (make_dirs (skipcrud + current_file_name)) + goto make_node; + msg_perror ("Could not make %s", + current_file_name + skipcrud); + break; + }; + goto set_filestat; +#endif + +#ifdef S_ISFIFO + /* If local system doesn't support FIFOs, use default case */ + case LF_FIFO: + make_fifo: + check = mkfifo (current_file_name + skipcrud, + (int) hstat.st_mode); + if (check != 0) + { + if (make_dirs (current_file_name + skipcrud)) + goto make_fifo; + msg_perror ("Could not make %s", + skipcrud + current_file_name); + break; + }; + goto set_filestat; +#endif + + case LF_DIR: + case LF_DUMPDIR: + namelen = strlen (current_file_name + skipcrud) - 1; + really_dir: + /* Check for trailing /, and zap as many as we find. */ + while (namelen + && current_file_name[skipcrud + namelen] == '/') + current_file_name[skipcrud + namelen--] = '\0'; + if (f_gnudump) + { /* Read the entry and delete files + that aren't listed in the archive */ + gnu_restore (skipcrud); + + } + else if (head->header.linkflag == LF_DUMPDIR) + skip_file ((long) (hstat.st_size)); + + + again_dir: + check = mkdir (skipcrud + current_file_name, + (we_are_root ? 0 : 0300) | (int) hstat.st_mode); + if (check != 0) + { + struct stat st1; + + if (make_dirs (skipcrud + current_file_name)) + goto again_dir; + /* If we're trying to create '.', let it be. */ + if (current_file_name[skipcrud + namelen] == '.' && + (namelen == 0 || + current_file_name[skipcrud + namelen - 1] == '/')) + goto check_perms; + if (errno == EEXIST + && stat (skipcrud + current_file_name, &st1) == 0 + && (S_ISDIR (st1.st_mode))) + break; + msg_perror ("Could not create directory %s", skipcrud + current_file_name); + break; + } + + check_perms: + if (!we_are_root && 0300 != (0300 & (int) hstat.st_mode)) + { + hstat.st_mode |= 0300; + msg ("Added write and execute permission to directory %s", + skipcrud + current_file_name); + } + + /* + * If we are root, set the owner and group of the extracted + * file. This does what is wanted both on real Unix and on + * System V. If we are running as a user, we extract as that + * user; if running as root, we extract as the original owner. + */ + if (we_are_root || f_do_chown) + { + if (chown (skipcrud + current_file_name, + hstat.st_uid, hstat.st_gid) < 0) + { + msg_perror ("cannot chown file %s to uid %d gid %d", + skipcrud + current_file_name, + hstat.st_uid, hstat.st_gid); + } + } + + if (!f_modified) + { + tmp = ((struct saved_dir_info *) + ck_malloc (sizeof (struct saved_dir_info))); + tmp->path = (char *) ck_malloc (strlen (skipcrud + + current_file_name) + 1); + strcpy (tmp->path, skipcrud + current_file_name); + tmp->mode = hstat.st_mode; + tmp->atime = hstat.st_atime; + tmp->mtime = hstat.st_mtime; + tmp->next = saved_dir_info_head; + saved_dir_info_head = tmp; + } + else + { + /* This functions exactly as the code for set_filestat above. */ + if ((!f_keep) + || (hstat.st_mode & (S_ISUID | S_ISGID | S_ISVTX))) + { + if (chmod (skipcrud + current_file_name, + notumask & (int) hstat.st_mode) < 0) + { + msg_perror ("cannot change mode of file %s to %ld", + skipcrud + current_file_name, + notumask & (int) hstat.st_mode); + } + } + } + break; + + case LF_VOLHDR: + if (f_verbose) + { + printf ("Reading %s\n", current_file_name); + } + break; + + case LF_NAMES: + extract_mangle (head); + break; + + case LF_MULTIVOL: + msg ("Can't extract '%s'--file is continued from another volume\n", current_file_name); + skip_file ((long) hstat.st_size); + break; + + case LF_LONGNAME: + case LF_LONGLINK: + msg ("Visible long name error\n"); + skip_file ((long) hstat.st_size); + break; + } + + /* We don't need to save it any longer. */ + saverec ((union record **) 0);/* Unsave it */ +} + +/* + * After a file/link/symlink/dir creation has failed, see if + * it's because some required directory was not present, and if + * so, create all required dirs. + */ +int +make_dirs (pathname) + char *pathname; +{ + char *p; /* Points into path */ + int madeone = 0; /* Did we do anything yet? */ + int save_errno = errno; /* Remember caller's errno */ + int check; + + if (errno != ENOENT) + return 0; /* Not our problem */ + + for (p = index (pathname, '/'); p != NULL; p = index (p + 1, '/')) + { + /* Avoid mkdir of empty string, if leading or double '/' */ + if (p == pathname || p[-1] == '/') + continue; + /* Avoid mkdir where last part of path is '.' */ + if (p[-1] == '.' && (p == pathname + 1 || p[-2] == '/')) + continue; + *p = 0; /* Truncate the path there */ + check = mkdir (pathname, 0777); /* Try to create it as a dir */ + if (check == 0) + { + /* Fix ownership */ + if (we_are_root) + { + if (chown (pathname, hstat.st_uid, + hstat.st_gid) < 0) + { + msg_perror ("cannot change owner of %s to uid %d gid %d", pathname, hstat.st_uid, hstat.st_gid); + } + } + pr_mkdir (pathname, p - pathname, notumask & 0777); + madeone++; /* Remember if we made one */ + *p = '/'; + continue; + } + *p = '/'; + if (errno == EEXIST) /* Directory already exists */ + continue; + /* + * Some other error in the mkdir. We return to the caller. + */ + break; + } + + errno = save_errno; /* Restore caller's errno */ + return madeone; /* Tell them to retry if we made one */ +} + +void +extract_sparse_file (fd, sizeleft, totalsize, name) + int fd; + long *sizeleft, totalsize; + char *name; +{ + /* register char *data;*/ + union record *datarec; + int sparse_ind = 0; + int written, count; + + /* assuming sizeleft is initially totalsize */ + + + while (*sizeleft > 0) + { + datarec = findrec (); + if (datarec == NULL) + { + msg ("Unexpected EOF on archive file"); + return; + } + lseek (fd, sparsearray[sparse_ind].offset, 0); + written = sparsearray[sparse_ind++].numbytes; + while (written > RECORDSIZE) + { + count = write (fd, datarec->charptr, RECORDSIZE); + if (count < 0) + msg_perror ("couldn't write to file %s", name); + written -= count; + *sizeleft -= count; + userec (datarec); + datarec = findrec (); + } + + count = write (fd, datarec->charptr, written); + + if (count < 0) + { + msg_perror ("couldn't write to file %s", name); + } + else if (count != written) + { + msg ("could only write %d of %d bytes to file %s", count, + totalsize, name); + skip_file ((long) (*sizeleft)); + } + + written -= count; + *sizeleft -= count; + userec (datarec); + } + free (sparsearray); + /* if (end_nulls) { + register int i; + + printf("%d\n", (int) end_nulls); + for (i = 0; i < end_nulls; i++) + write(fd, "\000", 1); + }*/ + userec (datarec); +} + +/* Set back the utime and mode for all the extracted directories. */ +void +restore_saved_dir_info () +{ + struct utimbuf acc_upd_times; + + while (saved_dir_info_head != NULL) + { + /* fixme if f_gnudump should set ctime too, but how? */ + if (f_gnudump) + acc_upd_times.actime = saved_dir_info_head->atime; + else + acc_upd_times.actime = now; /* Accessed now */ + acc_upd_times.modtime = saved_dir_info_head->mtime; /* Mod'd */ + if (utime (saved_dir_info_head->path, &acc_upd_times) < 0) + { + msg_perror ("couldn't change access and modification times of %s", + saved_dir_info_head->path); + } + if ((!f_keep) || (saved_dir_info_head->mode & (S_ISUID | S_ISGID | S_ISVTX))) + { + if (chmod (saved_dir_info_head->path, + notumask & saved_dir_info_head->mode) < 0) + { + msg_perror ("cannot change mode of file %s to %ld", + saved_dir_info_head->path, + notumask & saved_dir_info_head->mode); + } + } + saved_dir_info_head = saved_dir_info_head->next; + } +} diff --git a/gnu/usr.bin/tar/fnmatch.c b/gnu/usr.bin/tar/fnmatch.c new file mode 100644 index 000000000000..ed8c9eea7e35 --- /dev/null +++ b/gnu/usr.bin/tar/fnmatch.c @@ -0,0 +1,173 @@ +/* Copyright (C) 1991, 1992 Free Software Foundation, Inc. + +This library is free software; you can redistribute it and/or +modify it under the terms of the GNU Library General Public License as +published by the Free Software Foundation; either version 2 of the +License, or (at your option) any later version. + +This library is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +Library General Public License for more details. + +You should have received a copy of the GNU Library General Public +License along with this library; see the file COPYING.LIB. If +not, write to the Free Software Foundation, Inc., 675 Mass Ave, +Cambridge, MA 02139, USA. */ + +#include <errno.h> +#include <fnmatch.h> + +#if !defined(__GNU_LIBRARY__) && !defined(STDC_HEADERS) +extern int errno; +#endif + +/* Match STRING against the filename pattern PATTERN, returning zero if + it matches, nonzero if not. */ +int +fnmatch (pattern, string, flags) + const char *pattern; + const char *string; + int flags; +{ + register const char *p = pattern, *n = string; + register char c; + + if ((flags & ~__FNM_FLAGS) != 0) + { + errno = EINVAL; + return -1; + } + + while ((c = *p++) != '\0') + { + switch (c) + { + case '?': + if (*n == '\0') + return FNM_NOMATCH; + else if ((flags & FNM_PATHNAME) && *n == '/') + return FNM_NOMATCH; + else if ((flags & FNM_PERIOD) && *n == '.' && + (n == string || ((flags & FNM_PATHNAME) && n[-1] == '/'))) + return FNM_NOMATCH; + break; + + case '\\': + if (!(flags & FNM_NOESCAPE)) + c = *p++; + if (*n != c) + return FNM_NOMATCH; + break; + + case '*': + if ((flags & FNM_PERIOD) && *n == '.' && + (n == string || ((flags & FNM_PATHNAME) && n[-1] == '/'))) + return FNM_NOMATCH; + + for (c = *p++; c == '?' || c == '*'; c = *p++, ++n) + if (((flags & FNM_PATHNAME) && *n == '/') || + (c == '?' && *n == '\0')) + return FNM_NOMATCH; + + if (c == '\0') + return 0; + + { + char c1 = (!(flags & FNM_NOESCAPE) && c == '\\') ? *p : c; + for (--p; *n != '\0'; ++n) + if ((c == '[' || *n == c1) && + fnmatch (p, n, flags & ~FNM_PERIOD) == 0) + return 0; + return FNM_NOMATCH; + } + + case '[': + { + /* Nonzero if the sense of the character class is inverted. */ + register int not; + + if (*n == '\0') + return FNM_NOMATCH; + + if ((flags & FNM_PERIOD) && *n == '.' && + (n == string || ((flags & FNM_PATHNAME) && n[-1] == '/'))) + return FNM_NOMATCH; + + not = (*p == '!' || *p == '^'); + if (not) + ++p; + + c = *p++; + for (;;) + { + register char cstart = c, cend = c; + + if (!(flags & FNM_NOESCAPE) && c == '\\') + cstart = cend = *p++; + + if (c == '\0') + /* [ (unterminated) loses. */ + return FNM_NOMATCH; + + c = *p++; + + if ((flags & FNM_PATHNAME) && c == '/') + /* [/] can never match. */ + return FNM_NOMATCH; + + if (c == '-' && *p != ']') + { + cend = *p++; + if (!(flags & FNM_NOESCAPE) && cend == '\\') + cend = *p++; + if (cend == '\0') + return FNM_NOMATCH; + c = *p++; + } + + if (*n >= cstart && *n <= cend) + goto matched; + + if (c == ']') + break; + } + if (!not) + return FNM_NOMATCH; + break; + + matched:; + /* Skip the rest of the [...] that already matched. */ + while (c != ']') + { + if (c == '\0') + /* [... (unterminated) loses. */ + return FNM_NOMATCH; + + c = *p++; + if (!(flags & FNM_NOESCAPE) && c == '\\') + /* 1003.2d11 is unclear if this is right. %%% */ + ++p; + } + if (not) + return FNM_NOMATCH; + } + break; + + default: + if (c != *n) + return FNM_NOMATCH; + } + + ++n; + } + + if (*n == '\0') + return 0; + + if ((flags & FNM_LEADING_DIR) && *n == '/') + /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */ + return 0; + + return FNM_NOMATCH; +} diff --git a/gnu/usr.bin/tar/fnmatch.h b/gnu/usr.bin/tar/fnmatch.h new file mode 100644 index 000000000000..d4150a9a9969 --- /dev/null +++ b/gnu/usr.bin/tar/fnmatch.h @@ -0,0 +1,62 @@ +/* Copyright (C) 1991, 1992 Free Software Foundation, Inc. + +This library is free software; you can redistribute it and/or +modify it under the terms of the GNU Library General Public License as +published by the Free Software Foundation; either version 2 of the +License, or (at your option) any later version. + +This library is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +Library General Public License for more details. + +You should have received a copy of the GNU Library General Public +License along with this library; see the file COPYING.LIB. If +not, write to the Free Software Foundation, Inc., 675 Mass Ave, +Cambridge, MA 02139, USA. */ + +#ifndef _FNMATCH_H + +#define _FNMATCH_H 1 + +#ifdef __cplusplus +extern "C" { +#endif + +#if defined (__cplusplus) || (defined (__STDC__) && __STDC__) +#undef __P +#define __P(args) args +#else /* Not C++ or ANSI C. */ +#undef __P +#define __P(args) () +#undef const +#define const +#endif /* C++ or ANSI C. */ + +/* Bits set in the FLAGS argument to `fnmatch'. */ +#ifdef FNM_PATHNAME /* Because it is already defined in <unistd.h> */ +#undef FNM_PATHNAME +#endif +#define FNM_PATHNAME (1 << 0)/* No wildcard can ever match `/'. */ +#define FNM_NOESCAPE (1 << 1)/* Backslashes don't quote special chars. */ +#define FNM_PERIOD (1 << 2)/* Leading `.' is matched only explicitly. */ +#define __FNM_FLAGS (FNM_PATHNAME|FNM_NOESCAPE|FNM_PERIOD|FNM_LEADING_DIR) + +#if !defined (_POSIX_C_SOURCE) || _POSIX_C_SOURCE < 2 || defined (_BSD_SOURCE) +#define FNM_LEADING_DIR (1 << 3)/* Ignore `/...' after a match. */ +#define FNM_FILE_NAME FNM_PATHNAME +#endif + +/* Value returned by `fnmatch' if STRING does not match PATTERN. */ +#define FNM_NOMATCH 1 + +/* Match STRING against the filename pattern PATTERN, + returning zero if it matches, FNM_NOMATCH if not. */ +extern int fnmatch __P ((const char *__pattern, const char *__string, + int __flags)); + +#ifdef __cplusplus +} +#endif + +#endif /* fnmatch.h */ diff --git a/gnu/usr.bin/tar/getdate.y b/gnu/usr.bin/tar/getdate.y new file mode 100644 index 000000000000..7b0ac79924f4 --- /dev/null +++ b/gnu/usr.bin/tar/getdate.y @@ -0,0 +1,969 @@ +%{ +/* $Revision: 2.1 $ +** +** Originally written by Steven M. Bellovin <smb@research.att.com> while +** at the University of North Carolina at Chapel Hill. Later tweaked by +** a couple of people on Usenet. Completely overhauled by Rich $alz +** <rsalz@bbn.com> and Jim Berets <jberets@bbn.com> in August, 1990; +** send any email to Rich. +** +** This grammar has eight shift/reduce conflicts. +** +** This code is in the public domain and has no copyright. +*/ +/* SUPPRESS 287 on yaccpar_sccsid *//* Unusd static variable */ +/* SUPPRESS 288 on yyerrlab *//* Label unused */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#ifdef __GNUC__ +#define alloca __builtin_alloca +#else +#ifdef HAVE_ALLOCA_H +#include <alloca.h> +#else +#ifdef _AIX /* for Bison */ + #pragma alloca +#else +char *alloca (); +#endif +#endif +#endif + +#include <stdio.h> +#include <ctype.h> + +/* The code at the top of get_date which figures out the offset of the + current time zone checks various CPP symbols to see if special + tricks are need, but defaults to using the gettimeofday system call. + Include <sys/time.h> if that will be used. */ + +#if !defined (USG) && !defined (sgi) && !defined (__386BSD__) +#include <sys/time.h> +#endif + +#if defined(vms) + +#include <types.h> +#include <time.h> + +#else + +#include <sys/types.h> + +#if defined(USG) || !defined(HAVE_FTIME) +/* +** If you need to do a tzset() call to set the +** timezone, and don't have ftime(). +*/ +struct timeb { + time_t time; /* Seconds since the epoch */ + unsigned short millitm; /* Field not used */ + short timezone; + short dstflag; /* Field not used */ +}; + +#else + +#include <sys/timeb.h> + +#endif /* defined(USG) && !defined(HAVE_FTIME) */ + +#if defined(BSD4_2) || defined(BSD4_1C) || (defined (hp9000) && !defined (hpux)) +#include <sys/time.h> +#else +#if defined(_AIX) +#include <sys/time.h> +#endif +#include <time.h> +#endif /* defined(BSD4_2) */ + +#endif /* defined(vms) */ + +#if defined (STDC_HEADERS) || defined (USG) +#include <string.h> +#endif + +#if sgi +#undef timezone +#endif + +extern struct tm *localtime(); + +#define yyparse getdate_yyparse +#define yylex getdate_yylex +#define yyerror getdate_yyerror + +#if !defined(lint) && !defined(SABER) +static char RCS[] = + "$Header: str2date.y,v 2.1 90/09/06 08:15:06 cronan Exp $"; +#endif /* !defined(lint) && !defined(SABER) */ + + +#define EPOCH 1970 +#define HOUR(x) ((time_t)(x) * 60) +#define SECSPERDAY (24L * 60L * 60L) + + +/* +** An entry in the lexical lookup table. +*/ +typedef struct _TABLE { + char *name; + int type; + time_t value; +} TABLE; + + +/* +** Daylight-savings mode: on, off, or not yet known. +*/ +typedef enum _DSTMODE { + DSTon, DSToff, DSTmaybe +} DSTMODE; + +/* +** Meridian: am, pm, or 24-hour style. +*/ +typedef enum _MERIDIAN { + MERam, MERpm, MER24 +} MERIDIAN; + + +/* +** Global variables. We could get rid of most of these by using a good +** union as the yacc stack. (This routine was originally written before +** yacc had the %union construct.) Maybe someday; right now we only use +** the %union very rarely. +*/ +static char *yyInput; +static DSTMODE yyDSTmode; +static time_t yyDayOrdinal; +static time_t yyDayNumber; +static int yyHaveDate; +static int yyHaveDay; +static int yyHaveRel; +static int yyHaveTime; +static int yyHaveZone; +static time_t yyTimezone; +static time_t yyDay; +static time_t yyHour; +static time_t yyMinutes; +static time_t yyMonth; +static time_t yySeconds; +static time_t yyYear; +static MERIDIAN yyMeridian; +static time_t yyRelMonth; +static time_t yyRelSeconds; + +%} + +%union { + time_t Number; + enum _MERIDIAN Meridian; +} + +%token tAGO tDAY tDAYZONE tID tMERIDIAN tMINUTE_UNIT tMONTH tMONTH_UNIT +%token tSEC_UNIT tSNUMBER tUNUMBER tZONE tDST + +%type <Number> tDAY tDAYZONE tMINUTE_UNIT tMONTH tMONTH_UNIT +%type <Number> tSEC_UNIT tSNUMBER tUNUMBER tZONE +%type <Meridian> tMERIDIAN o_merid + +%% + +spec : /* NULL */ + | spec item + ; + +item : time { + yyHaveTime++; + } + | zone { + yyHaveZone++; + } + | date { + yyHaveDate++; + } + | day { + yyHaveDay++; + } + | rel { + yyHaveRel++; + } + | number + ; + +time : tUNUMBER tMERIDIAN { + yyHour = $1; + yyMinutes = 0; + yySeconds = 0; + yyMeridian = $2; + } + | tUNUMBER ':' tUNUMBER o_merid { + yyHour = $1; + yyMinutes = $3; + yySeconds = 0; + yyMeridian = $4; + } + | tUNUMBER ':' tUNUMBER tSNUMBER { + yyHour = $1; + yyMinutes = $3; + yyMeridian = MER24; + yyDSTmode = DSToff; + yyTimezone = - ($4 % 100 + ($4 / 100) * 60); + } + | tUNUMBER ':' tUNUMBER ':' tUNUMBER o_merid { + yyHour = $1; + yyMinutes = $3; + yySeconds = $5; + yyMeridian = $6; + } + | tUNUMBER ':' tUNUMBER ':' tUNUMBER tSNUMBER { + yyHour = $1; + yyMinutes = $3; + yySeconds = $5; + yyMeridian = MER24; + yyDSTmode = DSToff; + yyTimezone = - ($6 % 100 + ($6 / 100) * 60); + } + ; + +zone : tZONE { + yyTimezone = $1; + yyDSTmode = DSToff; + } + | tDAYZONE { + yyTimezone = $1; + yyDSTmode = DSTon; + } + | + tZONE tDST { + yyTimezone = $1; + yyDSTmode = DSTon; + } + ; + +day : tDAY { + yyDayOrdinal = 1; + yyDayNumber = $1; + } + | tDAY ',' { + yyDayOrdinal = 1; + yyDayNumber = $1; + } + | tUNUMBER tDAY { + yyDayOrdinal = $1; + yyDayNumber = $2; + } + ; + +date : tUNUMBER '/' tUNUMBER { + yyMonth = $1; + yyDay = $3; + } + | tUNUMBER '/' tUNUMBER '/' tUNUMBER { + yyMonth = $1; + yyDay = $3; + yyYear = $5; + } + | tUNUMBER tSNUMBER tSNUMBER { + /* ISO 8601 format. yyyy-mm-dd. */ + yyYear = $1; + yyMonth = -$2; + yyDay = -$3; + } + | tMONTH tUNUMBER { + yyMonth = $1; + yyDay = $2; + } + | tMONTH tUNUMBER ',' tUNUMBER { + yyMonth = $1; + yyDay = $2; + yyYear = $4; + } + | tUNUMBER tMONTH { + yyMonth = $2; + yyDay = $1; + } + | tUNUMBER tMONTH tUNUMBER { + yyMonth = $2; + yyDay = $1; + yyYear = $3; + } + ; + +rel : relunit tAGO { + yyRelSeconds = -yyRelSeconds; + yyRelMonth = -yyRelMonth; + } + | relunit + ; + +relunit : tUNUMBER tMINUTE_UNIT { + yyRelSeconds += $1 * $2 * 60L; + } + | tSNUMBER tMINUTE_UNIT { + yyRelSeconds += $1 * $2 * 60L; + } + | tMINUTE_UNIT { + yyRelSeconds += $1 * 60L; + } + | tSNUMBER tSEC_UNIT { + yyRelSeconds += $1; + } + | tUNUMBER tSEC_UNIT { + yyRelSeconds += $1; + } + | tSEC_UNIT { + yyRelSeconds++; + } + | tSNUMBER tMONTH_UNIT { + yyRelMonth += $1 * $2; + } + | tUNUMBER tMONTH_UNIT { + yyRelMonth += $1 * $2; + } + | tMONTH_UNIT { + yyRelMonth += $1; + } + ; + +number : tUNUMBER { + if (yyHaveTime && yyHaveDate && !yyHaveRel) + yyYear = $1; + else { + if($1>10000) { + time_t date_part; + + date_part= $1/10000; + yyHaveDate++; + yyDay= (date_part)%100; + yyMonth= (date_part/100)%100; + yyYear = date_part/10000; + } + yyHaveTime++; + if ($1 < 100) { + yyHour = $1; + yyMinutes = 0; + } + else { + yyHour = $1 / 100; + yyMinutes = $1 % 100; + } + yySeconds = 0; + yyMeridian = MER24; + } + } + ; + +o_merid : /* NULL */ { + $$ = MER24; + } + | tMERIDIAN { + $$ = $1; + } + ; + +%% + +/* Month and day table. */ +static TABLE const MonthDayTable[] = { + { "january", tMONTH, 1 }, + { "february", tMONTH, 2 }, + { "march", tMONTH, 3 }, + { "april", tMONTH, 4 }, + { "may", tMONTH, 5 }, + { "june", tMONTH, 6 }, + { "july", tMONTH, 7 }, + { "august", tMONTH, 8 }, + { "september", tMONTH, 9 }, + { "sept", tMONTH, 9 }, + { "october", tMONTH, 10 }, + { "november", tMONTH, 11 }, + { "december", tMONTH, 12 }, + { "sunday", tDAY, 0 }, + { "monday", tDAY, 1 }, + { "tuesday", tDAY, 2 }, + { "tues", tDAY, 2 }, + { "wednesday", tDAY, 3 }, + { "wednes", tDAY, 3 }, + { "thursday", tDAY, 4 }, + { "thur", tDAY, 4 }, + { "thurs", tDAY, 4 }, + { "friday", tDAY, 5 }, + { "saturday", tDAY, 6 }, + { NULL } +}; + +/* Time units table. */ +static TABLE const UnitsTable[] = { + { "year", tMONTH_UNIT, 12 }, + { "month", tMONTH_UNIT, 1 }, + { "fortnight", tMINUTE_UNIT, 14 * 24 * 60 }, + { "week", tMINUTE_UNIT, 7 * 24 * 60 }, + { "day", tMINUTE_UNIT, 1 * 24 * 60 }, + { "hour", tMINUTE_UNIT, 60 }, + { "minute", tMINUTE_UNIT, 1 }, + { "min", tMINUTE_UNIT, 1 }, + { "second", tSEC_UNIT, 1 }, + { "sec", tSEC_UNIT, 1 }, + { NULL } +}; + +/* Assorted relative-time words. */ +static TABLE const OtherTable[] = { + { "tomorrow", tMINUTE_UNIT, 1 * 24 * 60 }, + { "yesterday", tMINUTE_UNIT, -1 * 24 * 60 }, + { "today", tMINUTE_UNIT, 0 }, + { "now", tMINUTE_UNIT, 0 }, + { "last", tUNUMBER, -1 }, + { "this", tMINUTE_UNIT, 0 }, + { "next", tUNUMBER, 2 }, + { "first", tUNUMBER, 1 }, +/* { "second", tUNUMBER, 2 }, */ + { "third", tUNUMBER, 3 }, + { "fourth", tUNUMBER, 4 }, + { "fifth", tUNUMBER, 5 }, + { "sixth", tUNUMBER, 6 }, + { "seventh", tUNUMBER, 7 }, + { "eighth", tUNUMBER, 8 }, + { "ninth", tUNUMBER, 9 }, + { "tenth", tUNUMBER, 10 }, + { "eleventh", tUNUMBER, 11 }, + { "twelfth", tUNUMBER, 12 }, + { "ago", tAGO, 1 }, + { NULL } +}; + +/* The timezone table. */ +/* Some of these are commented out because a time_t can't store a float. */ +static TABLE const TimezoneTable[] = { + { "gmt", tZONE, HOUR( 0) }, /* Greenwich Mean */ + { "ut", tZONE, HOUR( 0) }, /* Universal (Coordinated) */ + { "utc", tZONE, HOUR( 0) }, + { "wet", tZONE, HOUR( 0) }, /* Western European */ + { "bst", tDAYZONE, HOUR( 0) }, /* British Summer */ + { "wat", tZONE, HOUR( 1) }, /* West Africa */ + { "at", tZONE, HOUR( 2) }, /* Azores */ +#if 0 + /* For completeness. BST is also British Summer, and GST is + * also Guam Standard. */ + { "bst", tZONE, HOUR( 3) }, /* Brazil Standard */ + { "gst", tZONE, HOUR( 3) }, /* Greenland Standard */ +#endif +#if 0 + { "nft", tZONE, HOUR(3.5) }, /* Newfoundland */ + { "nst", tZONE, HOUR(3.5) }, /* Newfoundland Standard */ + { "ndt", tDAYZONE, HOUR(3.5) }, /* Newfoundland Daylight */ +#endif + { "ast", tZONE, HOUR( 4) }, /* Atlantic Standard */ + { "adt", tDAYZONE, HOUR( 4) }, /* Atlantic Daylight */ + { "est", tZONE, HOUR( 5) }, /* Eastern Standard */ + { "edt", tDAYZONE, HOUR( 5) }, /* Eastern Daylight */ + { "cst", tZONE, HOUR( 6) }, /* Central Standard */ + { "cdt", tDAYZONE, HOUR( 6) }, /* Central Daylight */ + { "mst", tZONE, HOUR( 7) }, /* Mountain Standard */ + { "mdt", tDAYZONE, HOUR( 7) }, /* Mountain Daylight */ + { "pst", tZONE, HOUR( 8) }, /* Pacific Standard */ + { "pdt", tDAYZONE, HOUR( 8) }, /* Pacific Daylight */ + { "yst", tZONE, HOUR( 9) }, /* Yukon Standard */ + { "ydt", tDAYZONE, HOUR( 9) }, /* Yukon Daylight */ + { "hst", tZONE, HOUR(10) }, /* Hawaii Standard */ + { "hdt", tDAYZONE, HOUR(10) }, /* Hawaii Daylight */ + { "cat", tZONE, HOUR(10) }, /* Central Alaska */ + { "ahst", tZONE, HOUR(10) }, /* Alaska-Hawaii Standard */ + { "nt", tZONE, HOUR(11) }, /* Nome */ + { "idlw", tZONE, HOUR(12) }, /* International Date Line West */ + { "cet", tZONE, -HOUR(1) }, /* Central European */ + { "met", tZONE, -HOUR(1) }, /* Middle European */ + { "mewt", tZONE, -HOUR(1) }, /* Middle European Winter */ + { "mest", tDAYZONE, -HOUR(1) }, /* Middle European Summer */ + { "swt", tZONE, -HOUR(1) }, /* Swedish Winter */ + { "sst", tDAYZONE, -HOUR(1) }, /* Swedish Summer */ + { "fwt", tZONE, -HOUR(1) }, /* French Winter */ + { "fst", tDAYZONE, -HOUR(1) }, /* French Summer */ + { "eet", tZONE, -HOUR(2) }, /* Eastern Europe, USSR Zone 1 */ + { "bt", tZONE, -HOUR(3) }, /* Baghdad, USSR Zone 2 */ +#if 0 + { "it", tZONE, -HOUR(3.5) },/* Iran */ +#endif + { "zp4", tZONE, -HOUR(4) }, /* USSR Zone 3 */ + { "zp5", tZONE, -HOUR(5) }, /* USSR Zone 4 */ +#if 0 + { "ist", tZONE, -HOUR(5.5) },/* Indian Standard */ +#endif + { "zp6", tZONE, -HOUR(6) }, /* USSR Zone 5 */ +#if 0 + /* For completeness. NST is also Newfoundland Stanard, and SST is + * also Swedish Summer. */ + { "nst", tZONE, -HOUR(6.5) },/* North Sumatra */ + { "sst", tZONE, -HOUR(7) }, /* South Sumatra, USSR Zone 6 */ +#endif /* 0 */ + { "wast", tZONE, -HOUR(7) }, /* West Australian Standard */ + { "wadt", tDAYZONE, -HOUR(7) }, /* West Australian Daylight */ +#if 0 + { "jt", tZONE, -HOUR(7.5) },/* Java (3pm in Cronusland!) */ +#endif + { "cct", tZONE, -HOUR(8) }, /* China Coast, USSR Zone 7 */ + { "jst", tZONE, -HOUR(9) }, /* Japan Standard, USSR Zone 8 */ +#if 0 + { "cast", tZONE, -HOUR(9.5) },/* Central Australian Standard */ + { "cadt", tDAYZONE, -HOUR(9.5) },/* Central Australian Daylight */ +#endif + { "east", tZONE, -HOUR(10) }, /* Eastern Australian Standard */ + { "eadt", tDAYZONE, -HOUR(10) }, /* Eastern Australian Daylight */ + { "gst", tZONE, -HOUR(10) }, /* Guam Standard, USSR Zone 9 */ + { "nzt", tZONE, -HOUR(12) }, /* New Zealand */ + { "nzst", tZONE, -HOUR(12) }, /* New Zealand Standard */ + { "nzdt", tDAYZONE, -HOUR(12) }, /* New Zealand Daylight */ + { "idle", tZONE, -HOUR(12) }, /* International Date Line East */ + { NULL } +}; + +/* Military timezone table. */ +static TABLE const MilitaryTable[] = { + { "a", tZONE, HOUR( 1) }, + { "b", tZONE, HOUR( 2) }, + { "c", tZONE, HOUR( 3) }, + { "d", tZONE, HOUR( 4) }, + { "e", tZONE, HOUR( 5) }, + { "f", tZONE, HOUR( 6) }, + { "g", tZONE, HOUR( 7) }, + { "h", tZONE, HOUR( 8) }, + { "i", tZONE, HOUR( 9) }, + { "k", tZONE, HOUR( 10) }, + { "l", tZONE, HOUR( 11) }, + { "m", tZONE, HOUR( 12) }, + { "n", tZONE, HOUR(- 1) }, + { "o", tZONE, HOUR(- 2) }, + { "p", tZONE, HOUR(- 3) }, + { "q", tZONE, HOUR(- 4) }, + { "r", tZONE, HOUR(- 5) }, + { "s", tZONE, HOUR(- 6) }, + { "t", tZONE, HOUR(- 7) }, + { "u", tZONE, HOUR(- 8) }, + { "v", tZONE, HOUR(- 9) }, + { "w", tZONE, HOUR(-10) }, + { "x", tZONE, HOUR(-11) }, + { "y", tZONE, HOUR(-12) }, + { "z", tZONE, HOUR( 0) }, + { NULL } +}; + + + + +/* ARGSUSED */ +static int +yyerror(s) + char *s; +{ + return 0; +} + + +static time_t +ToSeconds(Hours, Minutes, Seconds, Meridian) + time_t Hours; + time_t Minutes; + time_t Seconds; + MERIDIAN Meridian; +{ + if (Minutes < 0 || Minutes > 59 || Seconds < 0 || Seconds > 59) + return -1; + switch (Meridian) { + case MER24: + if (Hours < 0 || Hours > 23) + return -1; + return (Hours * 60L + Minutes) * 60L + Seconds; + case MERam: + if (Hours < 1 || Hours > 12) + return -1; + return (Hours * 60L + Minutes) * 60L + Seconds; + case MERpm: + if (Hours < 1 || Hours > 12) + return -1; + return ((Hours + 12) * 60L + Minutes) * 60L + Seconds; + } + /* NOTREACHED */ +} + + +static time_t +Convert(Month, Day, Year, Hours, Minutes, Seconds, Meridian, DSTmode) + time_t Month; + time_t Day; + time_t Year; + time_t Hours; + time_t Minutes; + time_t Seconds; + MERIDIAN Meridian; + DSTMODE DSTmode; +{ + static int DaysInMonth[12] = { + 31, 0, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 + }; + time_t tod; + time_t Julian; + int i; + + if (Year < 0) + Year = -Year; + if (Year < 100) + Year += 1900; + DaysInMonth[1] = Year % 4 == 0 && (Year % 100 != 0 || Year % 400 == 0) + ? 29 : 28; + if (Year < EPOCH || Year > 1999 + || Month < 1 || Month > 12 + /* Lint fluff: "conversion from long may lose accuracy" */ + || Day < 1 || Day > DaysInMonth[(int)--Month]) + return -1; + + for (Julian = Day - 1, i = 0; i < Month; i++) + Julian += DaysInMonth[i]; + for (i = EPOCH; i < Year; i++) + Julian += 365 + (i % 4 == 0); + Julian *= SECSPERDAY; + Julian += yyTimezone * 60L; + if ((tod = ToSeconds(Hours, Minutes, Seconds, Meridian)) < 0) + return -1; + Julian += tod; + if (DSTmode == DSTon + || (DSTmode == DSTmaybe && localtime(&Julian)->tm_isdst)) + Julian -= 60 * 60; + return Julian; +} + + +static time_t +DSTcorrect(Start, Future) + time_t Start; + time_t Future; +{ + time_t StartDay; + time_t FutureDay; + + StartDay = (localtime(&Start)->tm_hour + 1) % 24; + FutureDay = (localtime(&Future)->tm_hour + 1) % 24; + return (Future - Start) + (StartDay - FutureDay) * 60L * 60L; +} + + +static time_t +RelativeDate(Start, DayOrdinal, DayNumber) + time_t Start; + time_t DayOrdinal; + time_t DayNumber; +{ + struct tm *tm; + time_t now; + + now = Start; + tm = localtime(&now); + now += SECSPERDAY * ((DayNumber - tm->tm_wday + 7) % 7); + now += 7 * SECSPERDAY * (DayOrdinal <= 0 ? DayOrdinal : DayOrdinal - 1); + return DSTcorrect(Start, now); +} + + +static time_t +RelativeMonth(Start, RelMonth) + time_t Start; + time_t RelMonth; +{ + struct tm *tm; + time_t Month; + time_t Year; + + if (RelMonth == 0) + return 0; + tm = localtime(&Start); + Month = 12 * tm->tm_year + tm->tm_mon + RelMonth; + Year = Month / 12; + Month = Month % 12 + 1; + return DSTcorrect(Start, + Convert(Month, (time_t)tm->tm_mday, Year, + (time_t)tm->tm_hour, (time_t)tm->tm_min, (time_t)tm->tm_sec, + MER24, DSTmaybe)); +} + + +static int +LookupWord(buff) + char *buff; +{ + register char *p; + register char *q; + register const TABLE *tp; + int i; + int abbrev; + + /* Make it lowercase. */ + for (p = buff; *p; p++) + if (isupper(*p)) + *p = tolower(*p); + + if (strcmp(buff, "am") == 0 || strcmp(buff, "a.m.") == 0) { + yylval.Meridian = MERam; + return tMERIDIAN; + } + if (strcmp(buff, "pm") == 0 || strcmp(buff, "p.m.") == 0) { + yylval.Meridian = MERpm; + return tMERIDIAN; + } + + /* See if we have an abbreviation for a month. */ + if (strlen(buff) == 3) + abbrev = 1; + else if (strlen(buff) == 4 && buff[3] == '.') { + abbrev = 1; + buff[3] = '\0'; + } + else + abbrev = 0; + + for (tp = MonthDayTable; tp->name; tp++) { + if (abbrev) { + if (strncmp(buff, tp->name, 3) == 0) { + yylval.Number = tp->value; + return tp->type; + } + } + else if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + } + + for (tp = TimezoneTable; tp->name; tp++) + if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + + if (strcmp(buff, "dst") == 0) + return tDST; + + for (tp = UnitsTable; tp->name; tp++) + if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + + /* Strip off any plural and try the units table again. */ + i = strlen(buff) - 1; + if (buff[i] == 's') { + buff[i] = '\0'; + for (tp = UnitsTable; tp->name; tp++) + if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + buff[i] = 's'; /* Put back for "this" in OtherTable. */ + } + + for (tp = OtherTable; tp->name; tp++) + if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + + /* Military timezones. */ + if (buff[1] == '\0' && isalpha(*buff)) { + for (tp = MilitaryTable; tp->name; tp++) + if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + } + + /* Drop out any periods and try the timezone table again. */ + for (i = 0, p = q = buff; *q; q++) + if (*q != '.') + *p++ = *q; + else + i++; + *p = '\0'; + if (i) + for (tp = TimezoneTable; tp->name; tp++) + if (strcmp(buff, tp->name) == 0) { + yylval.Number = tp->value; + return tp->type; + } + + return tID; +} + + +static int +yylex() +{ + register char c; + register char *p; + char buff[20]; + int Count; + int sign; + + for ( ; ; ) { + while (isspace(*yyInput)) + yyInput++; + + if (isdigit(c = *yyInput) || c == '-' || c == '+') { + if (c == '-' || c == '+') { + sign = c == '-' ? -1 : 1; + if (!isdigit(*++yyInput)) + /* skip the '-' sign */ + continue; + } + else + sign = 0; + for (yylval.Number = 0; isdigit(c = *yyInput++); ) + yylval.Number = 10 * yylval.Number + c - '0'; + yyInput--; + if (sign < 0) + yylval.Number = -yylval.Number; + return sign ? tSNUMBER : tUNUMBER; + } + if (isalpha(c)) { + for (p = buff; isalpha(c = *yyInput++) || c == '.'; ) + if (p < &buff[sizeof buff - 1]) + *p++ = c; + *p = '\0'; + yyInput--; + return LookupWord(buff); + } + if (c != '(') + return *yyInput++; + Count = 0; + do { + c = *yyInput++; + if (c == '\0') + return c; + if (c == '(') + Count++; + else if (c == ')') + Count--; + } while (Count > 0); + } +} + + +time_t +get_date(p, now) + char *p; + struct timeb *now; +{ + struct tm *tm; + struct timeb ftz; + time_t Start; + time_t tod; + + yyInput = p; + if (now == NULL) { + now = &ftz; +#if !defined(HAVE_FTIME) + (void)time(&ftz.time); + /* Set the timezone global. */ + tzset(); + { +#if sgi + ftz.timezone = (int) _timezone / 60; +#else /* not sgi */ +#ifdef __386BSD__ + ftz.timezone = 0; +#else /* neither sgi nor 386BSD */ +#if defined (USG) + extern time_t timezone; + + ftz.timezone = (int) timezone / 60; +#else /* neither sgi nor 386BSD nor USG */ + struct timeval tv; + struct timezone tz; + + gettimeofday (&tv, &tz); + ftz.timezone = (int) tz.tz_minuteswest; +#endif /* neither sgi nor 386BSD nor USG */ +#endif /* neither sgi nor 386BSD */ +#endif /* not sgi */ + } +#else /* HAVE_FTIME */ + (void)ftime(&ftz); +#endif /* HAVE_FTIME */ + } + + tm = localtime(&now->time); + yyYear = tm->tm_year; + yyMonth = tm->tm_mon + 1; + yyDay = tm->tm_mday; + yyTimezone = now->timezone; + yyDSTmode = DSTmaybe; + yyHour = 0; + yyMinutes = 0; + yySeconds = 0; + yyMeridian = MER24; + yyRelSeconds = 0; + yyRelMonth = 0; + yyHaveDate = 0; + yyHaveDay = 0; + yyHaveRel = 0; + yyHaveTime = 0; + yyHaveZone = 0; + + if (yyparse() + || yyHaveTime > 1 || yyHaveZone > 1 || yyHaveDate > 1 || yyHaveDay > 1) + return -1; + + if (yyHaveDate || yyHaveTime || yyHaveDay) { + Start = Convert(yyMonth, yyDay, yyYear, yyHour, yyMinutes, yySeconds, + yyMeridian, yyDSTmode); + if (Start < 0) + return -1; + } + else { + Start = now->time; + if (!yyHaveRel) + Start -= ((tm->tm_hour * 60L + tm->tm_min) * 60L) + tm->tm_sec; + } + + Start += yyRelSeconds; + Start += RelativeMonth(Start, yyRelMonth); + + if (yyHaveDay && !yyHaveDate) { + tod = RelativeDate(Start, yyDayOrdinal, yyDayNumber); + Start += tod; + } + + /* Have to do *something* with a legitimate -1 so it's distinguishable + * from the error return value. (Alternately could set errno on error.) */ + return Start == -1 ? 0 : Start; +} + + +#if defined(TEST) + +/* ARGSUSED */ +main(ac, av) + int ac; + char *av[]; +{ + char buff[128]; + time_t d; + + (void)printf("Enter date, or blank line to exit.\n\t> "); + (void)fflush(stdout); + while (gets(buff) && buff[0]) { + d = get_date(buff, (struct timeb *)NULL); + if (d == -1) + (void)printf("Bad format - couldn't convert.\n"); + else + (void)printf("%s", ctime(&d)); + (void)printf("\t> "); + (void)fflush(stdout); + } + exit(0); + /* NOTREACHED */ +} +#endif /* defined(TEST) */ diff --git a/gnu/usr.bin/tar/getoldopt.c b/gnu/usr.bin/tar/getoldopt.c new file mode 100644 index 000000000000..27511b94b342 --- /dev/null +++ b/gnu/usr.bin/tar/getoldopt.c @@ -0,0 +1,96 @@ +/* Replacement for getopt() that can be used by tar. + Copyright (C) 1988 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * Plug-compatible replacement for getopt() for parsing tar-like + * arguments. If the first argument begins with "-", it uses getopt; + * otherwise, it uses the old rules used by tar, dump, and ps. + * + * Written 25 August 1985 by John Gilmore (ihnp4!hoptoad!gnu) + */ + +#include <stdio.h> +#include "getopt.h" +#include "tar.h" /* For msg() declaration if STDC_MSG. */ +#include <sys/types.h> +#include "port.h" + +int +getoldopt (argc, argv, optstring, long_options, opt_index) + int argc; + char **argv; + char *optstring; + struct option *long_options; + int *opt_index; +{ + extern char *optarg; /* Points to next arg */ + extern int optind; /* Global argv index */ + static char *key; /* Points to next keyletter */ + static char use_getopt; /* !=0 if argv[1][0] was '-' */ + char c; + char *place; + + optarg = NULL; + + if (key == NULL) + { /* First time */ + if (argc < 2) + return EOF; + key = argv[1]; + if ((*key == '-') || (*key == '+')) + use_getopt++; + else + optind = 2; + } + + if (use_getopt) + return getopt_long (argc, argv, optstring, + long_options, opt_index); + + c = *key++; + if (c == '\0') + { + key--; + return EOF; + } + place = index (optstring, c); + + if (place == NULL || c == ':') + { + msg ("unknown option %c", c); + return ('?'); + } + + place++; + if (*place == ':') + { + if (optind < argc) + { + optarg = argv[optind]; + optind++; + } + else + { + msg ("%c argument missing", c); + return ('?'); + } + } + + return (c); +} diff --git a/gnu/usr.bin/tar/getopt.c b/gnu/usr.bin/tar/getopt.c new file mode 100644 index 000000000000..3db9abf121b9 --- /dev/null +++ b/gnu/usr.bin/tar/getopt.c @@ -0,0 +1,712 @@ +/* Getopt for GNU. + NOTE: getopt is now part of the C library, so if you don't know what + "Keep this file name-space clean" means, talk to roland@gnu.ai.mit.edu + before changing it! + + Copyright (C) 1987, 88, 89, 90, 91, 92, 1993 + Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* NOTE!!! AIX requires this to be the first thing in the file. + Do not put ANYTHING before it! */ +#if !defined (__GNUC__) && defined (_AIX) + #pragma alloca +#endif + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#ifdef __GNUC__ +#define alloca __builtin_alloca +#else /* not __GNUC__ */ +#if defined (HAVE_ALLOCA_H) || (defined(sparc) && (defined(sun) || (!defined(USG) && !defined(SVR4) && !defined(__svr4__)))) +#include <alloca.h> +#else +#ifndef _AIX +char *alloca (); +#endif +#endif /* alloca.h */ +#endif /* not __GNUC__ */ + +#if !__STDC__ && !defined(const) && IN_GCC +#define const +#endif + +#include <stdio.h> + +/* This needs to come after some library #include + to get __GNU_LIBRARY__ defined. */ +#ifdef __GNU_LIBRARY__ +#undef alloca +/* Don't include stdlib.h for non-GNU C libraries because some of them + contain conflicting prototypes for getopt. */ +#include <stdlib.h> +#else /* Not GNU C library. */ +#define __alloca alloca +#endif /* GNU C library. */ + +/* If GETOPT_COMPAT is defined, `+' as well as `--' can introduce a + long-named option. Because this is not POSIX.2 compliant, it is + being phased out. */ +/* #define GETOPT_COMPAT */ + +/* This version of `getopt' appears to the caller like standard Unix `getopt' + but it behaves differently for the user, since it allows the user + to intersperse the options with the other arguments. + + As `getopt' works, it permutes the elements of ARGV so that, + when it is done, all the options precede everything else. Thus + all application programs are extended to handle flexible argument order. + + Setting the environment variable POSIXLY_CORRECT disables permutation. + Then the behavior is completely standard. + + GNU application programs can use a third alternative mode in which + they can distinguish the relative order of options and other arguments. */ + +#include "getopt.h" + +/* For communication from `getopt' to the caller. + When `getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when `ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +char *optarg = 0; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to `getopt'. + + On entry to `getopt', zero means this is the first call; initialize. + + When `getopt' returns EOF, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, `optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +/* XXX 1003.2 says this must be 1 before any call. */ +int optind = 0; + +/* The next char to be scanned in the option-element + in which the last option character we returned was found. + This allows us to pick up the scan where we left off. + + If this is zero, or a null string, it means resume the scan + by advancing to the next ARGV-element. */ + +static char *nextchar; + +/* Callers store zero here to inhibit the error message + for unrecognized options. */ + +int opterr = 1; + +/* Set to an option character which was unrecognized. + This must be initialized on some systems to avoid linking in the + system's own getopt implementation. */ + +int optopt = '?'; + +/* Describe how to deal with options that follow non-option ARGV-elements. + + If the caller did not specify anything, + the default is REQUIRE_ORDER if the environment variable + POSIXLY_CORRECT is defined, PERMUTE otherwise. + + REQUIRE_ORDER means don't recognize them as options; + stop option processing when the first non-option is seen. + This is what Unix does. + This mode of operation is selected by either setting the environment + variable POSIXLY_CORRECT, or using `+' as the first character + of the list of option characters. + + PERMUTE is the default. We permute the contents of ARGV as we scan, + so that eventually all the non-options are at the end. This allows options + to be given in any order, even with programs that were not written to + expect this. + + RETURN_IN_ORDER is an option available to programs that were written + to expect options and other ARGV-elements in any order and that care about + the ordering of the two. We describe each non-option ARGV-element + as if it were the argument of an option with character code 1. + Using `-' as the first character of the list of option characters + selects this mode of operation. + + The special argument `--' forces an end of option-scanning regardless + of the value of `ordering'. In the case of RETURN_IN_ORDER, only + `--' can cause `getopt' to return EOF with `optind' != ARGC. */ + +static enum +{ + REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER +} ordering; + +#ifdef __GNU_LIBRARY__ +/* We want to avoid inclusion of string.h with non-GNU libraries + because there are many ways it can cause trouble. + On some systems, it contains special magic macros that don't work + in GCC. */ +#include <string.h> +#define my_index strchr +#define my_bcopy(src, dst, n) memcpy ((dst), (src), (n)) +#else + +/* Avoid depending on library functions or files + whose names are inconsistent. */ + +char *getenv (); + +static char * +my_index (string, chr) + char *string; + int chr; +{ + while (*string) + { + if (*string == chr) + return string; + string++; + } + return 0; +} + +static void +my_bcopy (from, to, size) + char *from, *to; + int size; +{ + int i; + for (i = 0; i < size; i++) + to[i] = from[i]; +} +#endif /* GNU C library. */ + +/* Handle permutation of arguments. */ + +/* Describe the part of ARGV that contains non-options that have + been skipped. `first_nonopt' is the index in ARGV of the first of them; + `last_nonopt' is the index after the last of them. */ + +static int first_nonopt; +static int last_nonopt; + +/* Exchange two adjacent subsequences of ARGV. + One subsequence is elements [first_nonopt,last_nonopt) + which contains all the non-options that have been skipped so far. + The other is elements [last_nonopt,optind), which contains all + the options processed since those non-options were skipped. + + `first_nonopt' and `last_nonopt' are relocated so that they describe + the new indices of the non-options in ARGV after they are moved. */ + +static void +exchange (argv) + char **argv; +{ + int nonopts_size = (last_nonopt - first_nonopt) * sizeof (char *); + char **temp = (char **) __alloca (nonopts_size); + + /* Interchange the two blocks of data in ARGV. */ + + my_bcopy ((char *) &argv[first_nonopt], (char *) temp, nonopts_size); + my_bcopy ((char *) &argv[last_nonopt], (char *) &argv[first_nonopt], + (optind - last_nonopt) * sizeof (char *)); + my_bcopy ((char *) temp, + (char *) &argv[first_nonopt + optind - last_nonopt], + nonopts_size); + + /* Update records for the slots the non-options now occupy. */ + + first_nonopt += (optind - last_nonopt); + last_nonopt = optind; +} + +/* Scan elements of ARGV (whose length is ARGC) for option characters + given in OPTSTRING. + + If an element of ARGV starts with '-', and is not exactly "-" or "--", + then it is an option element. The characters of this element + (aside from the initial '-') are option characters. If `getopt' + is called repeatedly, it returns successively each of the option characters + from each of the option elements. + + If `getopt' finds another option character, it returns that character, + updating `optind' and `nextchar' so that the next call to `getopt' can + resume the scan with the following option character or ARGV-element. + + If there are no more option characters, `getopt' returns `EOF'. + Then `optind' is the index in ARGV of the first ARGV-element + that is not an option. (The ARGV-elements have been permuted + so that those that are not options now come last.) + + OPTSTRING is a string containing the legitimate option characters. + If an option character is seen that is not listed in OPTSTRING, + return '?' after printing an error message. If you set `opterr' to + zero, the error message is suppressed but we still return '?'. + + If a char in OPTSTRING is followed by a colon, that means it wants an arg, + so the following text in the same ARGV-element, or the text of the following + ARGV-element, is returned in `optarg'. Two colons mean an option that + wants an optional arg; if there is text in the current ARGV-element, + it is returned in `optarg', otherwise `optarg' is set to zero. + + If OPTSTRING starts with `-' or `+', it requests different methods of + handling the non-option ARGV-elements. + See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above. + + Long-named options begin with `--' instead of `-'. + Their names may be abbreviated as long as the abbreviation is unique + or is an exact match for some defined option. If they have an + argument, it follows the option name in the same ARGV-element, separated + from the option name by a `=', or else the in next ARGV-element. + When `getopt' finds a long-named option, it returns 0 if that option's + `flag' field is nonzero, the value of the option's `val' field + if the `flag' field is zero. + + The elements of ARGV aren't really const, because we permute them. + But we pretend they're const in the prototype to be compatible + with other systems. + + LONGOPTS is a vector of `struct option' terminated by an + element containing a name which is zero. + + LONGIND returns the index in LONGOPT of the long-named option found. + It is only valid when a long-named option has been found by the most + recent call. + + If LONG_ONLY is nonzero, '-' as well as '--' can introduce + long-named options. */ + +int +_getopt_internal (argc, argv, optstring, longopts, longind, long_only) + int argc; + char *const *argv; + const char *optstring; + const struct option *longopts; + int *longind; + int long_only; +{ + int option_index; + + optarg = 0; + + /* Initialize the internal data when the first call is made. + Start processing options with ARGV-element 1 (since ARGV-element 0 + is the program name); the sequence of previously skipped + non-option ARGV-elements is empty. */ + + if (optind == 0) + { + first_nonopt = last_nonopt = optind = 1; + + nextchar = NULL; + + /* Determine how to handle the ordering of options and nonoptions. */ + + if (optstring[0] == '-') + { + ordering = RETURN_IN_ORDER; + ++optstring; + } + else if (optstring[0] == '+') + { + ordering = REQUIRE_ORDER; + ++optstring; + } + else if (getenv ("POSIXLY_CORRECT") != NULL) + ordering = REQUIRE_ORDER; + else + ordering = PERMUTE; + } + + if (nextchar == NULL || *nextchar == '\0') + { + if (ordering == PERMUTE) + { + /* If we have just processed some options following some non-options, + exchange them so that the options come first. */ + + if (first_nonopt != last_nonopt && last_nonopt != optind) + exchange ((char **) argv); + else if (last_nonopt != optind) + first_nonopt = optind; + + /* Now skip any additional non-options + and extend the range of non-options previously skipped. */ + + while (optind < argc + && (argv[optind][0] != '-' || argv[optind][1] == '\0') +#ifdef GETOPT_COMPAT + && (longopts == NULL + || argv[optind][0] != '+' || argv[optind][1] == '\0') +#endif /* GETOPT_COMPAT */ + ) + optind++; + last_nonopt = optind; + } + + /* Special ARGV-element `--' means premature end of options. + Skip it like a null option, + then exchange with previous non-options as if it were an option, + then skip everything else like a non-option. */ + + if (optind != argc && !strcmp (argv[optind], "--")) + { + optind++; + + if (first_nonopt != last_nonopt && last_nonopt != optind) + exchange ((char **) argv); + else if (first_nonopt == last_nonopt) + first_nonopt = optind; + last_nonopt = argc; + + optind = argc; + } + + /* If we have done all the ARGV-elements, stop the scan + and back over any non-options that we skipped and permuted. */ + + if (optind == argc) + { + /* Set the next-arg-index to point at the non-options + that we previously skipped, so the caller will digest them. */ + if (first_nonopt != last_nonopt) + optind = first_nonopt; + return EOF; + } + + /* If we have come to a non-option and did not permute it, + either stop the scan or describe it to the caller and pass it by. */ + + if ((argv[optind][0] != '-' || argv[optind][1] == '\0') +#ifdef GETOPT_COMPAT + && (longopts == NULL + || argv[optind][0] != '+' || argv[optind][1] == '\0') +#endif /* GETOPT_COMPAT */ + ) + { + if (ordering == REQUIRE_ORDER) + return EOF; + optarg = argv[optind++]; + return 1; + } + + /* We have found another option-ARGV-element. + Start decoding its characters. */ + + nextchar = (argv[optind] + 1 + + (longopts != NULL && argv[optind][1] == '-')); + } + + if (longopts != NULL + && ((argv[optind][0] == '-' + && (argv[optind][1] == '-' || long_only)) +#ifdef GETOPT_COMPAT + || argv[optind][0] == '+' +#endif /* GETOPT_COMPAT */ + )) + { + const struct option *p; + char *s = nextchar; + int exact = 0; + int ambig = 0; + const struct option *pfound = NULL; + int indfound; + + while (*s && *s != '=') + s++; + + /* Test all options for either exact match or abbreviated matches. */ + for (p = longopts, option_index = 0; p->name; + p++, option_index++) + if (!strncmp (p->name, nextchar, s - nextchar)) + { + if (s - nextchar == strlen (p->name)) + { + /* Exact match found. */ + pfound = p; + indfound = option_index; + exact = 1; + break; + } + else if (pfound == NULL) + { + /* First nonexact match found. */ + pfound = p; + indfound = option_index; + } + else + /* Second nonexact match found. */ + ambig = 1; + } + + if (ambig && !exact) + { + if (opterr) + fprintf (stderr, "%s: option `%s' is ambiguous\n", + argv[0], argv[optind]); + nextchar += strlen (nextchar); + optind++; + return '?'; + } + + if (pfound != NULL) + { + option_index = indfound; + optind++; + if (*s) + { + /* Don't test has_arg with >, because some C compilers don't + allow it to be used on enums. */ + if (pfound->has_arg) + optarg = s + 1; + else + { + if (opterr) + { + if (argv[optind - 1][1] == '-') + /* --option */ + fprintf (stderr, + "%s: option `--%s' doesn't allow an argument\n", + argv[0], pfound->name); + else + /* +option or -option */ + fprintf (stderr, + "%s: option `%c%s' doesn't allow an argument\n", + argv[0], argv[optind - 1][0], pfound->name); + } + nextchar += strlen (nextchar); + return '?'; + } + } + else if (pfound->has_arg == 1) + { + if (optind < argc) + optarg = argv[optind++]; + else + { + if (opterr) + fprintf (stderr, "%s: option `%s' requires an argument\n", + argv[0], argv[optind - 1]); + nextchar += strlen (nextchar); + return optstring[0] == ':' ? ':' : '?'; + } + } + nextchar += strlen (nextchar); + if (longind != NULL) + *longind = option_index; + if (pfound->flag) + { + *(pfound->flag) = pfound->val; + return 0; + } + return pfound->val; + } + /* Can't find it as a long option. If this is not getopt_long_only, + or the option starts with '--' or is not a valid short + option, then it's an error. + Otherwise interpret it as a short option. */ + if (!long_only || argv[optind][1] == '-' +#ifdef GETOPT_COMPAT + || argv[optind][0] == '+' +#endif /* GETOPT_COMPAT */ + || my_index (optstring, *nextchar) == NULL) + { + if (opterr) + { + if (argv[optind][1] == '-') + /* --option */ + fprintf (stderr, "%s: unrecognized option `--%s'\n", + argv[0], nextchar); + else + /* +option or -option */ + fprintf (stderr, "%s: unrecognized option `%c%s'\n", + argv[0], argv[optind][0], nextchar); + } + nextchar = (char *) ""; + optind++; + return '?'; + } + } + + /* Look at and handle the next option-character. */ + + { + char c = *nextchar++; + char *temp = my_index (optstring, c); + + /* Increment `optind' when we start to process its last character. */ + if (*nextchar == '\0') + ++optind; + + if (temp == NULL || c == ':') + { + if (opterr) + { +#if 0 + if (c < 040 || c >= 0177) + fprintf (stderr, "%s: unrecognized option, character code 0%o\n", + argv[0], c); + else + fprintf (stderr, "%s: unrecognized option `-%c'\n", argv[0], c); +#else + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, "%s: illegal option -- %c\n", argv[0], c); +#endif + } + optopt = c; + return '?'; + } + if (temp[1] == ':') + { + if (temp[2] == ':') + { + /* This is an option that accepts an argument optionally. */ + if (*nextchar != '\0') + { + optarg = nextchar; + optind++; + } + else + optarg = 0; + nextchar = NULL; + } + else + { + /* This is an option that requires an argument. */ + if (*nextchar != '\0') + { + optarg = nextchar; + /* If we end this ARGV-element by taking the rest as an arg, + we must advance to the next element now. */ + optind++; + } + else if (optind == argc) + { + if (opterr) + { +#if 0 + fprintf (stderr, "%s: option `-%c' requires an argument\n", + argv[0], c); +#else + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, "%s: option requires an argument -- %c\n", + argv[0], c); +#endif + } + optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + } + else + /* We already incremented `optind' once; + increment it again when taking next ARGV-elt as argument. */ + optarg = argv[optind++]; + nextchar = NULL; + } + } + return c; + } +} + +int +getopt (argc, argv, optstring) + int argc; + char *const *argv; + const char *optstring; +{ + return _getopt_internal (argc, argv, optstring, + (const struct option *) 0, + (int *) 0, + 0); +} + +#ifdef TEST + +/* Compile with -DTEST to make an executable for use in testing + the above definition of `getopt'. */ + +int +main (argc, argv) + int argc; + char **argv; +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + + c = getopt (argc, argv, "abc:d:0123456789"); + if (c == EOF) + break; + + switch (c) + { + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value `%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/gnu/usr.bin/tar/getopt.h b/gnu/usr.bin/tar/getopt.h new file mode 100644 index 000000000000..93a5cf77816e --- /dev/null +++ b/gnu/usr.bin/tar/getopt.h @@ -0,0 +1,125 @@ +/* Declarations for getopt. + Copyright (C) 1989, 1990, 1991, 1992 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifndef _GETOPT_H +#define _GETOPT_H 1 + +#ifdef __cplusplus +extern "C" { +#endif + +/* For communication from `getopt' to the caller. + When `getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when `ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +extern char *optarg; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to `getopt'. + + On entry to `getopt', zero means this is the first call; initialize. + + When `getopt' returns EOF, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, `optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +extern int optind; + +/* Callers store zero here to inhibit the error message `getopt' prints + for unrecognized options. */ + +extern int opterr; + +/* Describe the long-named options requested by the application. + The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector + of `struct option' terminated by an element containing a name which is + zero. + + The field `has_arg' is: + no_argument (or 0) if the option does not take an argument, + required_argument (or 1) if the option requires an argument, + optional_argument (or 2) if the option takes an optional argument. + + If the field `flag' is not NULL, it points to a variable that is set + to the value given in the field `val' when the option is found, but + left unchanged if the option is not found. + + To have a long-named option do something other than set an `int' to + a compiled-in constant, such as set a value from `optarg', set the + option's `flag' field to zero and its `val' field to a nonzero + value (the equivalent single-letter option character, if there is + one). For long options that have a zero `flag' field, `getopt' + returns the contents of the `val' field. */ + +struct option +{ +#if __STDC__ + const char *name; +#else + char *name; +#endif + /* has_arg can't be an enum because some compilers complain about + type mismatches in all the code that assumes it is an int. */ + int has_arg; + int *flag; + int val; +}; + +/* Names for the values of the `has_arg' field of `struct option'. */ + +#define no_argument 0 +#define required_argument 1 +#define optional_argument 2 + +#if __STDC__ +#if defined(__GNU_LIBRARY__) +/* Many other libraries have conflicting prototypes for getopt, with + differences in the consts, in stdlib.h. To avoid compilation + errors, only prototype getopt for the GNU C library. */ +extern int getopt (int argc, char *const *argv, const char *shortopts); +#else /* not __GNU_LIBRARY__ */ +extern int getopt (); +#endif /* not __GNU_LIBRARY__ */ +extern int getopt_long (int argc, char *const *argv, const char *shortopts, + const struct option *longopts, int *longind); +extern int getopt_long_only (int argc, char *const *argv, + const char *shortopts, + const struct option *longopts, int *longind); + +/* Internal only. Users should not call this directly. */ +extern int _getopt_internal (int argc, char *const *argv, + const char *shortopts, + const struct option *longopts, int *longind, + int long_only); +#else /* not __STDC__ */ +extern int getopt (); +extern int getopt_long (); +extern int getopt_long_only (); + +extern int _getopt_internal (); +#endif /* not __STDC__ */ + +#ifdef __cplusplus +} +#endif + +#endif /* _GETOPT_H */ diff --git a/gnu/usr.bin/tar/getopt1.c b/gnu/usr.bin/tar/getopt1.c new file mode 100644 index 000000000000..c3582cfa5caf --- /dev/null +++ b/gnu/usr.bin/tar/getopt1.c @@ -0,0 +1,161 @@ +/* Getopt for GNU. + Copyright (C) 1987, 88, 89, 90, 91, 1992 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#include "getopt.h" + +#if !__STDC__ && !defined(const) && IN_GCC +#define const +#endif + +#include <stdio.h> + +/* This needs to come after some library #include + to get __GNU_LIBRARY__ defined. */ +#ifdef __GNU_LIBRARY__ +#include <stdlib.h> +#else +char *getenv (); +#endif + +#ifndef NULL +#define NULL 0 +#endif + +int +getopt_long (argc, argv, options, long_options, opt_index) + int argc; + char *const *argv; + const char *options; + const struct option *long_options; + int *opt_index; +{ + return _getopt_internal (argc, argv, options, long_options, opt_index, 0); +} + +/* Like getopt_long, but '-' as well as '--' can indicate a long option. + If an option that starts with '-' (not '--') doesn't match a long option, + but does match a short option, it is parsed as a short option + instead. */ + +int +getopt_long_only (argc, argv, options, long_options, opt_index) + int argc; + char *const *argv; + const char *options; + const struct option *long_options; + int *opt_index; +{ + return _getopt_internal (argc, argv, options, long_options, opt_index, 1); +} + +#ifdef TEST + +#include <stdio.h> + +int +main (argc, argv) + int argc; + char **argv; +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + int option_index = 0; + static struct option long_options[] = + { + {"add", 1, 0, 0}, + {"append", 0, 0, 0}, + {"delete", 1, 0, 0}, + {"verbose", 0, 0, 0}, + {"create", 0, 0, 0}, + {"file", 1, 0, 0}, + {0, 0, 0, 0} + }; + + c = getopt_long (argc, argv, "abc:d:0123456789", + long_options, &option_index); + if (c == EOF) + break; + + switch (c) + { + case 0: + printf ("option %s", long_options[option_index].name); + if (optarg) + printf (" with arg %s", optarg); + printf ("\n"); + break; + + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value `%s'\n", optarg); + break; + + case 'd': + printf ("option d with value `%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/gnu/usr.bin/tar/getpagesize.h b/gnu/usr.bin/tar/getpagesize.h new file mode 100644 index 000000000000..2d43f262c79d --- /dev/null +++ b/gnu/usr.bin/tar/getpagesize.h @@ -0,0 +1,38 @@ +#ifdef BSD +#ifndef BSD4_1 +#define HAVE_GETPAGESIZE +#endif +#endif + +#ifndef HAVE_GETPAGESIZE + +#ifdef VMS +#define getpagesize() 512 +#endif + +#ifdef HAVE_UNISTD_H +#include <unistd.h> +#endif + +#ifdef _SC_PAGESIZE +#define getpagesize() sysconf(_SC_PAGESIZE) +#else + +#include <sys/param.h> + +#ifdef EXEC_PAGESIZE +#define getpagesize() EXEC_PAGESIZE +#else +#ifdef NBPG +#define getpagesize() NBPG * CLSIZE +#ifndef CLSIZE +#define CLSIZE 1 +#endif /* no CLSIZE */ +#else /* no NBPG */ +#define getpagesize() NBPC +#endif /* no NBPG */ +#endif /* no EXEC_PAGESIZE */ +#endif /* no _SC_PAGESIZE */ + +#endif /* not HAVE_GETPAGESIZE */ + diff --git a/gnu/usr.bin/tar/gnu.c b/gnu/usr.bin/tar/gnu.c new file mode 100644 index 000000000000..ef51f2b5fedd --- /dev/null +++ b/gnu/usr.bin/tar/gnu.c @@ -0,0 +1,677 @@ +/* GNU dump extensions to tar. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include <stdio.h> +#include <sys/types.h> +#include <ctype.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif +#include <time.h> +time_t time (); + +#include "tar.h" +#include "port.h" + +#ifndef S_ISLNK +#define lstat stat +#endif + +extern time_t new_time; +extern FILE *msg_file; + +void addname (); +int check_exclude (); +extern PTR ck_malloc (); +extern PTR ck_realloc (); +int confirm (); +extern PTR init_buffer (); +extern char *get_buffer (); +int is_dot_or_dotdot (); +extern void add_buffer (); +extern void flush_buffer (); +void name_gather (); +int recursively_delete (); +void skip_file (); +char *un_quote_string (); + +extern char *new_name (); + +static void add_dir_name (); + +struct dirname + { + struct dirname *next; + char *name; + char *dir_text; + int dev; + int ino; + int allnew; + }; +static struct dirname *dir_list; +static time_t this_time; + +void +add_dir (name, dev, ino, text) + char *name; + char *text; + dev_t dev; + ino_t ino; +{ + struct dirname *dp; + + dp = (struct dirname *) ck_malloc (sizeof (struct dirname)); + if (!dp) + abort (); + dp->next = dir_list; + dir_list = dp; + dp->dev = dev; + dp->ino = ino; + dp->name = ck_malloc (strlen (name) + 1); + strcpy (dp->name, name); + dp->dir_text = text; + dp->allnew = 0; +} + +void +read_dir_file () +{ + int dev; + int ino; + char *strp; + FILE *fp; + char buf[512]; + static char *path = 0; + + if (path == 0) + path = ck_malloc (PATH_MAX); + time (&this_time); + if (gnu_dumpfile[0] != '/') + { +#if defined(__MSDOS__) || defined(HAVE_GETCWD) || defined(_POSIX_VERSION) + if (!getcwd (path, PATH_MAX)) + { + msg ("Couldn't get current directory."); + exit (EX_SYSTEM); + } +#else + char *getwd (); + + if (!getwd (path)) + { + msg ("Couldn't get current directory: %s", path); + exit (EX_SYSTEM); + } +#endif + /* If this doesn't fit, we're in serious trouble */ + strcat (path, "/"); + strcat (path, gnu_dumpfile); + gnu_dumpfile = path; + } + fp = fopen (gnu_dumpfile, "r"); + if (fp == 0 && errno != ENOENT) + { + msg_perror ("Can't open %s", gnu_dumpfile); + return; + } + if (!fp) + return; + fgets (buf, sizeof (buf), fp); + if (!f_new_files) + { + f_new_files++; + new_time = atol (buf); + } + while (fgets (buf, sizeof (buf), fp)) + { + strp = &buf[strlen (buf)]; + if (strp[-1] == '\n') + strp[-1] = '\0'; + strp = buf; + dev = atol (strp); + while (isdigit (*strp)) + strp++; + ino = atol (strp); + while (isspace (*strp)) + strp++; + while (isdigit (*strp)) + strp++; + strp++; + add_dir (un_quote_string (strp), dev, ino, (char *) 0); + } + fclose (fp); +} + +void +write_dir_file () +{ + FILE *fp; + struct dirname *dp; + char *str; + extern char *quote_copy_string (); + + fp = fopen (gnu_dumpfile, "w"); + if (fp == 0) + { + msg_perror ("Can't write to %s", gnu_dumpfile); + return; + } + fprintf (fp, "%lu\n", this_time); + for (dp = dir_list; dp; dp = dp->next) + { + if (!dp->dir_text) + continue; + str = quote_copy_string (dp->name); + if (str) + { + fprintf (fp, "%u %u %s\n", dp->dev, dp->ino, str); + free (str); + } + else + fprintf (fp, "%u %u %s\n", dp->dev, dp->ino, dp->name); + } + fclose (fp); +} + +struct dirname * +get_dir (name) + char *name; +{ + struct dirname *dp; + + for (dp = dir_list; dp; dp = dp->next) + { + if (!strcmp (dp->name, name)) + return dp; + } + return 0; +} + + +/* Collect all the names from argv[] (or whatever), then expand them into + a directory tree, and put all the directories at the beginning. */ +void +collect_and_sort_names () +{ + struct name *n, *n_next; + int num_names; + struct stat statbuf; + int name_cmp (); + char *merge_sort (); + + name_gather (); + + if (gnu_dumpfile) + read_dir_file (); + if (!namelist) + addname ("."); + for (n = namelist; n; n = n_next) + { + n_next = n->next; + if (n->found || n->dir_contents) + continue; + if (n->regexp) /* FIXME just skip regexps for now */ + continue; + if (n->change_dir) + if (chdir (n->change_dir) < 0) + { + msg_perror ("can't chdir to %s", n->change_dir); + continue; + } + +#ifdef AIX + if (statx (n->name, &statbuf, STATSIZE, STX_HIDDEN | STX_LINK)) +#else + if (lstat (n->name, &statbuf) < 0) +#endif /* AIX */ + { + msg_perror ("can't stat %s", n->name); + continue; + } + if (S_ISDIR (statbuf.st_mode)) + { + n->found++; + add_dir_name (n->name, statbuf.st_dev); + } + } + + num_names = 0; + for (n = namelist; n; n = n->next) + num_names++; + namelist = (struct name *) merge_sort ((PTR) namelist, num_names, (char *) (&(namelist->next)) - (char *) namelist, name_cmp); + + for (n = namelist; n; n = n->next) + { + n->found = 0; + } + if (gnu_dumpfile) + write_dir_file (); +} + +int +name_cmp (n1, n2) + struct name *n1, *n2; +{ + if (n1->found) + { + if (n2->found) + return strcmp (n1->name, n2->name); + else + return -1; + } + else if (n2->found) + return 1; + else + return strcmp (n1->name, n2->name); +} + +int +dirent_cmp (p1, p2) + const PTR p1; + const PTR p2; +{ + char *frst, *scnd; + + frst = (*(char **) p1) + 1; + scnd = (*(char **) p2) + 1; + + return strcmp (frst, scnd); +} + +char * +get_dir_contents (p, device) + char *p; + int device; +{ + DIR *dirp; + register struct dirent *d; + char *new_buf; + char *namebuf; + int bufsiz; + int len; + PTR the_buffer; + char *buf; + size_t n_strs; + /* int n_size;*/ + char *p_buf; + char **vec, **p_vec; + + extern int errno; + + errno = 0; + dirp = opendir (p); + bufsiz = strlen (p) + NAMSIZ; + namebuf = ck_malloc (bufsiz + 2); + if (!dirp) + { + if (errno) + msg_perror ("can't open directory %s", p); + else + msg ("error opening directory %s", p); + new_buf = NULL; + } + else + { + struct dirname *dp; + int all_children; + + dp = get_dir (p); + all_children = dp ? dp->allnew : 0; + (void) strcpy (namebuf, p); + if (p[strlen (p) - 1] != '/') + (void) strcat (namebuf, "/"); + len = strlen (namebuf); + + the_buffer = init_buffer (); + while (d = readdir (dirp)) + { + struct stat hs; + + /* Skip . and .. */ + if (is_dot_or_dotdot (d->d_name)) + continue; + if (NLENGTH (d) + len >= bufsiz) + { + bufsiz += NAMSIZ; + namebuf = ck_realloc (namebuf, bufsiz + 2); + } + (void) strcpy (namebuf + len, d->d_name); +#ifdef AIX + if (0 != f_follow_links ? + statx (namebuf, &hs, STATSIZE, STX_HIDDEN) : + statx (namebuf, &hs, STATSIZE, STX_HIDDEN | STX_LINK)) +#else + if (0 != f_follow_links ? stat (namebuf, &hs) : lstat (namebuf, &hs)) +#endif + { + msg_perror ("can't stat %s", namebuf); + continue; + } + if ((f_local_filesys && device != hs.st_dev) + || (f_exclude && check_exclude (namebuf))) + add_buffer (the_buffer, "N", 1); +#ifdef AIX + else if (S_ISHIDDEN (hs.st_mode)) + { + add_buffer (the_buffer, "D", 1); + strcat (d->d_name, "A"); + d->d_namlen++; + } +#endif /* AIX */ + else if (S_ISDIR (hs.st_mode)) + { + if (dp = get_dir (namebuf)) + { + if (dp->dev != hs.st_dev + || dp->ino != hs.st_ino) + { + if (f_verbose) + msg ("directory %s has been renamed.", namebuf); + dp->allnew = 1; + dp->dev = hs.st_dev; + dp->ino = hs.st_ino; + } + dp->dir_text = ""; + } + else + { + if (f_verbose) + msg ("Directory %s is new", namebuf); + add_dir (namebuf, hs.st_dev, hs.st_ino, ""); + dp = get_dir (namebuf); + dp->allnew = 1; + } + if (all_children) + dp->allnew = 1; + + add_buffer (the_buffer, "D", 1); + } + else if (!all_children + && f_new_files + && new_time > hs.st_mtime + && (f_new_files > 1 + || new_time > hs.st_ctime)) + add_buffer (the_buffer, "N", 1); + else + add_buffer (the_buffer, "Y", 1); + add_buffer (the_buffer, d->d_name, (int) (NLENGTH (d) + 1)); + } + add_buffer (the_buffer, "\000\000", 2); + closedir (dirp); + + /* Well, we've read in the contents of the dir, now sort them */ + buf = get_buffer (the_buffer); + if (buf[0] == '\0') + { + flush_buffer (the_buffer); + new_buf = NULL; + } + else + { + n_strs = 0; + for (p_buf = buf; *p_buf;) + { + int tmp; + + tmp = strlen (p_buf) + 1; + n_strs++; + p_buf += tmp; + } + vec = (char **) ck_malloc (sizeof (char *) * (n_strs + 1)); + for (p_vec = vec, p_buf = buf; *p_buf; p_buf += strlen (p_buf) + 1) + *p_vec++ = p_buf; + *p_vec = 0; + qsort ((PTR) vec, n_strs, sizeof (char *), dirent_cmp); + new_buf = (char *) ck_malloc (p_buf - buf + 2); + for (p_vec = vec, p_buf = new_buf; *p_vec; p_vec++) + { + char *p_tmp; + + for (p_tmp = *p_vec; *p_buf++ = *p_tmp++;) + ; + } + *p_buf++ = '\0'; + free (vec); + flush_buffer (the_buffer); + } + } + free (namebuf); + return new_buf; +} + +/* p is a directory. Add all the files in P to the namelist. If any of the + files is a directory, recurse on the subdirectory. . . */ +static void +add_dir_name (p, device) + char *p; + int device; +{ + char *new_buf; + char *p_buf; + + char *namebuf; + int buflen; + register int len; + int sublen; + + /* PTR the_buffer;*/ + + /* char *buf;*/ + /* char **vec,**p_vec;*/ + /* int n_strs,n_size;*/ + + struct name *n; + + int dirent_cmp (); + + new_buf = get_dir_contents (p, device); + + for (n = namelist; n; n = n->next) + { + if (!strcmp (n->name, p)) + { + n->dir_contents = new_buf ? new_buf : "\0\0\0\0"; + break; + } + } + + if (new_buf) + { + len = strlen (p); + buflen = NAMSIZ <= len ? len + NAMSIZ : NAMSIZ; + namebuf = ck_malloc (buflen + 1); + + (void) strcpy (namebuf, p); + if (namebuf[len - 1] != '/') + { + namebuf[len++] = '/'; + namebuf[len] = '\0'; + } + for (p_buf = new_buf; *p_buf; p_buf += sublen + 1) + { + sublen = strlen (p_buf); + if (*p_buf == 'D') + { + if (len + sublen >= buflen) + { + buflen += NAMSIZ; + namebuf = ck_realloc (namebuf, buflen + 1); + } + (void) strcpy (namebuf + len, p_buf + 1); + addname (namebuf); + add_dir_name (namebuf, device); + } + } + free (namebuf); + } +} + +/* Returns non-zero if p is . or .. This could be a macro for speed. */ +int +is_dot_or_dotdot (p) + char *p; +{ + return (p[0] == '.' && (p[1] == '\0' || (p[1] == '.' && p[2] == '\0'))); +} + + + + + + +void +gnu_restore (skipcrud) + int skipcrud; +{ + char *current_dir; + /* int current_dir_length; */ + + char *archive_dir; + /* int archive_dir_length; */ + PTR the_buffer; + char *p; + DIR *dirp; + struct dirent *d; + char *cur, *arc; + extern struct stat hstat; /* Stat struct corresponding */ + long size, copied; + char *from, *to; + extern union record *head; + + dirp = opendir (skipcrud + current_file_name); + + if (!dirp) + { + /* The directory doesn't exist now. It'll be created. + In any case, we don't have to delete any files out + of it */ + skip_file ((long) hstat.st_size); + return; + } + + the_buffer = init_buffer (); + while (d = readdir (dirp)) + { + if (is_dot_or_dotdot (d->d_name)) + continue; + + add_buffer (the_buffer, d->d_name, (int) (NLENGTH (d) + 1)); + } + closedir (dirp); + add_buffer (the_buffer, "", 1); + + current_dir = get_buffer (the_buffer); + archive_dir = (char *) ck_malloc (hstat.st_size); + if (archive_dir == 0) + { + msg ("Can't allocate %d bytes for restore", hstat.st_size); + skip_file ((long) hstat.st_size); + return; + } + to = archive_dir; + for (size = hstat.st_size; size > 0; size -= copied) + { + from = findrec ()->charptr; + if (!from) + { + msg ("Unexpected EOF in archive\n"); + break; + } + copied = endofrecs ()->charptr - from; + if (copied > size) + copied = size; + bcopy ((PTR) from, (PTR) to, (int) copied); + to += copied; + userec ((union record *) (from + copied - 1)); + } + + for (cur = current_dir; *cur; cur += strlen (cur) + 1) + { + for (arc = archive_dir; *arc; arc += strlen (arc) + 1) + { + arc++; + if (!strcmp (arc, cur)) + break; + } + if (*arc == '\0') + { + p = new_name (skipcrud + current_file_name, cur); + if (f_confirm && !confirm ("delete", p)) + { + free (p); + continue; + } + if (f_verbose) + fprintf (msg_file, "%s: deleting %s\n", tar, p); + if (recursively_delete (p)) + { + msg ("%s: Error while deleting %s\n", tar, p); + } + free (p); + } + + } + flush_buffer (the_buffer); + free (archive_dir); +} + +int +recursively_delete (path) + char *path; +{ + struct stat sbuf; + DIR *dirp; + struct dirent *dp; + char *path_buf; + /* int path_len; */ + + + if (lstat (path, &sbuf) < 0) + return 1; + if (S_ISDIR (sbuf.st_mode)) + { + + /* path_len=strlen(path); */ + dirp = opendir (path); + if (dirp == 0) + return 1; + while (dp = readdir (dirp)) + { + if (is_dot_or_dotdot (dp->d_name)) + continue; + path_buf = new_name (path, dp->d_name); + if (recursively_delete (path_buf)) + { + free (path_buf); + closedir (dirp); + return 1; + } + free (path_buf); + } + closedir (dirp); + + if (rmdir (path) < 0) + return 1; + return 0; + } + if (unlink (path) < 0) + return 1; + return 0; +} diff --git a/gnu/usr.bin/tar/list.c b/gnu/usr.bin/tar/list.c new file mode 100644 index 000000000000..a0c65a334bf2 --- /dev/null +++ b/gnu/usr.bin/tar/list.c @@ -0,0 +1,881 @@ +/* List a tar archive. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * List a tar archive. + * + * Also includes support routines for reading a tar archive. + * + * this version written 26 Aug 1985 by John Gilmore (ihnp4!hoptoad!gnu). + */ + +#include <stdio.h> +#include <ctype.h> +#include <sys/types.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif +#include <time.h> + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#define isodigit(c) ( ((c) >= '0') && ((c) <= '7') ) + +#include "tar.h" +#include "port.h" + +extern FILE *msg_file; + +long from_oct (); /* Decode octal number */ +void demode (); /* Print file mode */ +void restore_saved_dir_info (); +PTR ck_malloc (); + +union record *head; /* Points to current archive header */ +struct stat hstat; /* Stat struct corresponding */ +int head_standard; /* Tape header is in ANSI format */ + +int check_exclude (); +void close_archive (); +void decode_header (); +int findgid (); +int finduid (); +void name_gather (); +int name_match (); +void names_notfound (); +void open_archive (); +void print_header (); +int read_header (); +void saverec (); +void skip_file (); +void skip_extended_headers (); + +extern char *quote_copy_string (); + + +/* + * Main loop for reading an archive. + */ +void +read_and (do_something) + void (*do_something) (); +{ + int status = 3; /* Initial status at start of archive */ + int prev_status; + extern time_t new_time; + char save_linkflag; + + name_gather (); /* Gather all the names */ + open_archive (1); /* Open for reading */ + + for (;;) + { + prev_status = status; + status = read_header (); + switch (status) + { + + case 1: /* Valid header */ + /* We should decode next field (mode) first... */ + /* Ensure incoming names are null terminated. */ + + if (!name_match (current_file_name) + || (f_new_files && hstat.st_mtime < new_time) + || (f_exclude && check_exclude (current_file_name))) + { + + int isextended = 0; + + if (head->header.linkflag == LF_VOLHDR + || head->header.linkflag == LF_MULTIVOL + || head->header.linkflag == LF_NAMES) + { + (*do_something) (); + continue; + } + if (f_show_omitted_dirs + && head->header.linkflag == LF_DIR) + msg ("Omitting %s\n", current_file_name); + /* Skip past it in the archive */ + if (head->header.isextended) + isextended = 1; + save_linkflag = head->header.linkflag; + userec (head); + if (isextended) + { + /* register union record *exhdr; + + for (;;) { + exhdr = findrec(); + if (!exhdr->ext_hdr.isextended) { + userec(exhdr); + break; + } + } + userec(exhdr);*/ + skip_extended_headers (); + } + /* Skip to the next header on the archive */ + if (save_linkflag != LF_DIR) + skip_file ((long) hstat.st_size); + continue; + + } + + (*do_something) (); + continue; + + /* + * If the previous header was good, tell them + * that we are skipping bad ones. + */ + case 0: /* Invalid header */ + userec (head); + switch (prev_status) + { + case 3: /* Error on first record */ + msg ("Hmm, this doesn't look like a tar archive."); + /* FALL THRU */ + case 2: /* Error after record of zeroes */ + case 1: /* Error after header rec */ + msg ("Skipping to next file header..."); + case 0: /* Error after error */ + break; + } + continue; + + case 2: /* Record of zeroes */ + userec (head); + status = prev_status; /* If error after 0's */ + if (f_ignorez) + continue; + /* FALL THRU */ + case EOF: /* End of archive */ + break; + } + break; + }; + + restore_saved_dir_info (); + close_archive (); + names_notfound (); /* Print names not found */ +} + + +/* + * Print a header record, based on tar options. + */ +void +list_archive () +{ + extern char *save_name; + int isextended = 0; /* Flag to remember if head is extended */ + + /* Save the record */ + saverec (&head); + + /* Print the header record */ + if (f_verbose) + { + if (f_verbose > 1) + decode_header (head, &hstat, &head_standard, 0); + print_header (); + } + + if (f_gnudump && head->header.linkflag == LF_DUMPDIR) + { + size_t size, written, check; + char *data; + extern long save_totsize; + extern long save_sizeleft; + + userec (head); + if (f_multivol) + { + save_name = current_file_name; + save_totsize = hstat.st_size; + } + for (size = hstat.st_size; size > 0; size -= written) + { + if (f_multivol) + save_sizeleft = size; + data = findrec ()->charptr; + if (data == NULL) + { + msg ("EOF in archive file?"); + break; + } + written = endofrecs ()->charptr - data; + if (written > size) + written = size; + errno = 0; + check = fwrite (data, sizeof (char), written, msg_file); + userec ((union record *) (data + written - 1)); + if (check != written) + { + msg_perror ("only wrote %ld of %ld bytes to file %s", check, written, current_file_name); + skip_file ((long) (size) - written); + break; + } + } + if (f_multivol) + save_name = 0; + saverec ((union record **) 0); /* Unsave it */ + fputc ('\n', msg_file); + fflush (msg_file); + return; + + } + saverec ((union record **) 0);/* Unsave it */ + /* Check to see if we have an extended header to skip over also */ + if (head->header.isextended) + isextended = 1; + + /* Skip past the header in the archive */ + userec (head); + + /* + * If we needed to skip any extended headers, do so now, by + * reading extended headers and skipping past them in the + * archive. + */ + if (isextended) + { + /* register union record *exhdr; + + for (;;) { + exhdr = findrec(); + + if (!exhdr->ext_hdr.isextended) { + userec(exhdr); + break; + } + userec(exhdr); + }*/ + skip_extended_headers (); + } + + if (f_multivol) + save_name = current_file_name; + /* Skip to the next header on the archive */ + + skip_file ((long) hstat.st_size); + + if (f_multivol) + save_name = 0; +} + + +/* + * Read a record that's supposed to be a header record. + * Return its address in "head", and if it is good, the file's + * size in hstat.st_size. + * + * Return 1 for success, 0 if the checksum is bad, EOF on eof, + * 2 for a record full of zeros (EOF marker). + * + * You must always userec(head) to skip past the header which this + * routine reads. + */ +int +read_header () +{ + register int i; + register long sum, signed_sum, recsum; + register char *p; + register union record *header; + long from_oct (); + char **longp; + char *bp, *data; + int size, written; + static char *next_long_name, *next_long_link; + char *name; + +recurse: + + header = findrec (); + head = header; /* This is our current header */ + if (NULL == header) + return EOF; + + recsum = from_oct (8, header->header.chksum); + + sum = 0; + p = header->charptr; + for (i = sizeof (*header); --i >= 0;) + { + /* + * We can't use unsigned char here because of old compilers, + * e.g. V7. + */ + signed_sum += *p; + sum += 0xFF & *p++; + } + + /* Adjust checksum to count the "chksum" field as blanks. */ + for (i = sizeof (header->header.chksum); --i >= 0;) + { + sum -= 0xFF & header->header.chksum[i]; + signed_sum -= (char) header->header.chksum[i]; + } + sum += ' ' * sizeof header->header.chksum; + signed_sum += ' ' * sizeof header->header.chksum; + + if (sum == 8 * ' ') + { + /* + * This is a zeroed record...whole record is 0's except + * for the 8 blanks we faked for the checksum field. + */ + return 2; + } + + if (sum != recsum && signed_sum != recsum) + return 0; + + /* + * Good record. Decode file size and return. + */ + if (header->header.linkflag == LF_LINK) + hstat.st_size = 0; /* Links 0 size on tape */ + else + hstat.st_size = from_oct (1 + 12, header->header.size); + + header->header.arch_name[NAMSIZ - 1] = '\0'; + if (header->header.linkflag == LF_LONGNAME + || header->header.linkflag == LF_LONGLINK) + { + longp = ((header->header.linkflag == LF_LONGNAME) + ? &next_long_name + : &next_long_link); + + userec (header); + if (*longp) + free (*longp); + bp = *longp = (char *) ck_malloc (hstat.st_size); + + for (size = hstat.st_size; + size > 0; + size -= written) + { + data = findrec ()->charptr; + if (data == NULL) + { + msg ("Unexpected EOF on archive file"); + break; + } + written = endofrecs ()->charptr - data; + if (written > size) + written = size; + + bcopy (data, bp, written); + bp += written; + userec ((union record *) (data + written - 1)); + } + goto recurse; + } + else + { + name = (next_long_name + ? next_long_name + : head->header.arch_name); + if (current_file_name) + free (current_file_name); + current_file_name = ck_malloc (strlen (name) + 1); + strcpy (current_file_name, name); + + name = (next_long_link + ? next_long_link + : head->header.arch_linkname); + if (current_link_name) + free (current_link_name); + current_link_name = ck_malloc (strlen (name) + 1); + strcpy (current_link_name, name); + + next_long_link = next_long_name = 0; + return 1; + } +} + + +/* + * Decode things from a file header record into a "struct stat". + * Also set "*stdp" to !=0 or ==0 depending whether header record is "Unix + * Standard" tar format or regular old tar format. + * + * read_header() has already decoded the checksum and length, so we don't. + * + * If wantug != 0, we want the uid/group info decoded from Unix Standard + * tapes (for extraction). If == 0, we are just printing anyway, so save time. + * + * decode_header should NOT be called twice for the same record, since the + * two calls might use different "wantug" values and thus might end up with + * different uid/gid for the two calls. If anybody wants the uid/gid they + * should decode it first, and other callers should decode it without uid/gid + * before calling a routine, e.g. print_header, that assumes decoded data. + */ +void +decode_header (header, st, stdp, wantug) + register union record *header; + register struct stat *st; + int *stdp; + int wantug; +{ + long from_oct (); + + st->st_mode = from_oct (8, header->header.mode); + st->st_mode &= 07777; + st->st_mtime = from_oct (1 + 12, header->header.mtime); + if (f_gnudump) + { + st->st_atime = from_oct (1 + 12, header->header.atime); + st->st_ctime = from_oct (1 + 12, header->header.ctime); + } + + if (0 == strcmp (header->header.magic, TMAGIC)) + { + /* Unix Standard tar archive */ + *stdp = 1; + if (wantug) + { +#ifdef NONAMES + st->st_uid = from_oct (8, header->header.uid); + st->st_gid = from_oct (8, header->header.gid); +#else + st->st_uid = + (*header->header.uname + ? finduid (header->header.uname) + : from_oct (8, header->header.uid)); + st->st_gid = + (*header->header.gname + ? findgid (header->header.gname) + : from_oct (8, header->header.gid)); +#endif + } +#if defined(S_IFBLK) || defined(S_IFCHR) + switch (header->header.linkflag) + { + case LF_BLK: + case LF_CHR: + st->st_rdev = makedev (from_oct (8, header->header.devmajor), + from_oct (8, header->header.devminor)); + } +#endif + } + else + { + /* Old fashioned tar archive */ + *stdp = 0; + st->st_uid = from_oct (8, header->header.uid); + st->st_gid = from_oct (8, header->header.gid); + st->st_rdev = 0; + } +} + + +/* + * Quick and dirty octal conversion. + * + * Result is -1 if the field is invalid (all blank, or nonoctal). + */ +long +from_oct (digs, where) + register int digs; + register char *where; +{ + register long value; + + while (isspace (*where)) + { /* Skip spaces */ + where++; + if (--digs <= 0) + return -1; /* All blank field */ + } + value = 0; + while (digs > 0 && isodigit (*where)) + { /* Scan til nonoctal */ + value = (value << 3) | (*where++ - '0'); + --digs; + } + + if (digs > 0 && *where && !isspace (*where)) + return -1; /* Ended on non-space/nul */ + + return value; +} + + +/* + * Actually print it. + * + * Plain and fancy file header block logging. + * Non-verbose just prints the name, e.g. for "tar t" or "tar x". + * This should just contain file names, so it can be fed back into tar + * with xargs or the "-T" option. The verbose option can give a bunch + * of info, one line per file. I doubt anybody tries to parse its + * format, or if they do, they shouldn't. Unix tar is pretty random here + * anyway. + * + * Note that print_header uses the globals <head>, <hstat>, and + * <head_standard>, which must be set up in advance. This is not very clean + * and should be cleaned up. FIXME. + */ +#define UGSWIDTH 18 /* min width of User, group, size */ +/* UGSWIDTH of 18 means that with user and group names <= 8 chars the columns + never shift during the listing. */ +#define DATEWIDTH 19 /* Last mod date */ +static int ugswidth = UGSWIDTH; /* Max width encountered so far */ + +void +print_header () +{ + char modes[11]; + char *timestamp; + char uform[11], gform[11]; /* These hold formatted ints */ + char *user, *group; + char size[24]; /* Holds a formatted long or maj, min */ + time_t longie; /* To make ctime() call portable */ + int pad; + char *name; + extern long baserec; + + if (f_sayblock) + fprintf (msg_file, "rec %10d: ", baserec + (ar_record - ar_block)); + /* annofile(msg_file, (char *)NULL); */ + + if (f_verbose <= 1) + { + /* Just the fax, mam. */ + char *name; + + name = quote_copy_string (current_file_name); + if (name == 0) + name = current_file_name; + fprintf (msg_file, "%s\n", name); + if (name != current_file_name) + free (name); + } + else + { + /* File type and modes */ + modes[0] = '?'; + switch (head->header.linkflag) + { + case LF_VOLHDR: + modes[0] = 'V'; + break; + + case LF_MULTIVOL: + modes[0] = 'M'; + break; + + case LF_NAMES: + modes[0] = 'N'; + break; + + case LF_LONGNAME: + case LF_LONGLINK: + msg ("Visible longname error\n"); + break; + + case LF_SPARSE: + case LF_NORMAL: + case LF_OLDNORMAL: + case LF_LINK: + modes[0] = '-'; + if ('/' == current_file_name[strlen (current_file_name) - 1]) + modes[0] = 'd'; + break; + case LF_DUMPDIR: + modes[0] = 'd'; + break; + case LF_DIR: + modes[0] = 'd'; + break; + case LF_SYMLINK: + modes[0] = 'l'; + break; + case LF_BLK: + modes[0] = 'b'; + break; + case LF_CHR: + modes[0] = 'c'; + break; + case LF_FIFO: + modes[0] = 'p'; + break; + case LF_CONTIG: + modes[0] = 'C'; + break; + } + + demode ((unsigned) hstat.st_mode, modes + 1); + + /* Timestamp */ + longie = hstat.st_mtime; + timestamp = ctime (&longie); + timestamp[16] = '\0'; + timestamp[24] = '\0'; + + /* User and group names */ + if (*head->header.uname && head_standard) + { + user = head->header.uname; + } + else + { + user = uform; + (void) sprintf (uform, "%d", + from_oct (8, head->header.uid)); + } + if (*head->header.gname && head_standard) + { + group = head->header.gname; + } + else + { + group = gform; + (void) sprintf (gform, "%d", + from_oct (8, head->header.gid)); + } + + /* Format the file size or major/minor device numbers */ + switch (head->header.linkflag) + { +#if defined(S_IFBLK) || defined(S_IFCHR) + case LF_CHR: + case LF_BLK: + (void) sprintf (size, "%d,%d", + major (hstat.st_rdev), + minor (hstat.st_rdev)); + break; +#endif + case LF_SPARSE: + (void) sprintf (size, "%ld", + from_oct (1 + 12, head->header.realsize)); + break; + default: + (void) sprintf (size, "%ld", (long) hstat.st_size); + } + + /* Figure out padding and print the whole line. */ + pad = strlen (user) + strlen (group) + strlen (size) + 1; + if (pad > ugswidth) + ugswidth = pad; + + name = quote_copy_string (current_file_name); + if (!name) + name = current_file_name; + fprintf (msg_file, "%s %s/%s %*s%s %s %s %s", + modes, + user, + group, + ugswidth - pad, + "", + size, + timestamp + 4, timestamp + 20, + name); + + if (name != current_file_name) + free (name); + switch (head->header.linkflag) + { + case LF_SYMLINK: + name = quote_copy_string (current_link_name); + if (!name) + name = current_link_name; + fprintf (msg_file, " -> %s\n", name); + if (name != current_link_name) + free (name); + break; + + case LF_LINK: + name = quote_copy_string (current_link_name); + if (!name) + name = current_link_name; + fprintf (msg_file, " link to %s\n", current_link_name); + if (name != current_link_name) + free (name); + break; + + default: + fprintf (msg_file, " unknown file type '%c'\n", + head->header.linkflag); + break; + + case LF_OLDNORMAL: + case LF_NORMAL: + case LF_SPARSE: + case LF_CHR: + case LF_BLK: + case LF_DIR: + case LF_FIFO: + case LF_CONTIG: + case LF_DUMPDIR: + putc ('\n', msg_file); + break; + + case LF_VOLHDR: + fprintf (msg_file, "--Volume Header--\n"); + break; + + case LF_MULTIVOL: + fprintf (msg_file, "--Continued at byte %ld--\n", from_oct (1 + 12, head->header.offset)); + break; + + case LF_NAMES: + fprintf (msg_file, "--Mangled file names--\n"); + break; + } + } + fflush (msg_file); +} + +/* + * Print a similar line when we make a directory automatically. + */ +void +pr_mkdir (pathname, length, mode) + char *pathname; + int length; + int mode; +{ + char modes[11]; + char *name; + extern long baserec; + + if (f_verbose > 1) + { + /* File type and modes */ + modes[0] = 'd'; + demode ((unsigned) mode, modes + 1); + + if (f_sayblock) + fprintf (msg_file, "rec %10d: ", baserec + (ar_record - ar_block)); + /* annofile(msg_file, (char *)NULL); */ + name = quote_copy_string (pathname); + if (!name) + name = pathname; + fprintf (msg_file, "%s %*s %.*s\n", + modes, + ugswidth + DATEWIDTH, + "Creating directory:", + length, + pathname); + if (name != pathname) + free (name); + } +} + + +/* + * Skip over <size> bytes of data in records in the archive. + */ +void +skip_file (size) + register long size; +{ + union record *x; + extern long save_totsize; + extern long save_sizeleft; + + if (f_multivol) + { + save_totsize = size; + save_sizeleft = size; + } + + while (size > 0) + { + x = findrec (); + if (x == NULL) + { /* Check it... */ + msg ("Unexpected EOF on archive file"); + exit (EX_BADARCH); + } + userec (x); + size -= RECORDSIZE; + if (f_multivol) + save_sizeleft -= RECORDSIZE; + } +} + +void +skip_extended_headers () +{ + register union record *exhdr; + + for (;;) + { + exhdr = findrec (); + if (!exhdr->ext_hdr.isextended) + { + userec (exhdr); + break; + } + userec (exhdr); + } +} + +/* + * Decode the mode string from a stat entry into a 9-char string and a null. + */ +void +demode (mode, string) + register unsigned mode; + register char *string; +{ + register unsigned mask; + register char *rwx = "rwxrwxrwx"; + + for (mask = 0400; mask != 0; mask >>= 1) + { + if (mode & mask) + *string++ = *rwx++; + else + { + *string++ = '-'; + rwx++; + } + } + + if (mode & S_ISUID) + if (string[-7] == 'x') + string[-7] = 's'; + else + string[-7] = 'S'; + if (mode & S_ISGID) + if (string[-4] == 'x') + string[-4] = 's'; + else + string[-4] = 'S'; + if (mode & S_ISVTX) + if (string[-1] == 'x') + string[-1] = 't'; + else + string[-1] = 'T'; + *string = '\0'; +} diff --git a/gnu/usr.bin/tar/mangle.c b/gnu/usr.bin/tar/mangle.c new file mode 100644 index 000000000000..628168473a1b --- /dev/null +++ b/gnu/usr.bin/tar/mangle.c @@ -0,0 +1,270 @@ +/* mangle.c -- encode long filenames + Copyright (C) 1988, 1992 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include <stdio.h> +#include <sys/types.h> +#include <time.h> +time_t time (); + +#include "tar.h" +#include "port.h" + +void add_buffer (); +extern PTR ck_malloc (); +void finish_header (); +extern PTR init_buffer (); +extern char *quote_copy_string (); +extern char *get_buffer (); +char *un_quote_string (); + +extern union record *start_header (); + +extern struct stat hstat; /* Stat struct corresponding */ + +struct mangled + { + struct mangled *next; + int type; + char mangled[NAMSIZ]; + char *linked_to; + char normal[1]; + }; + + +/* Should use a hash table, etc. . */ +struct mangled *first_mangle; +int mangled_num = 0; + +#if 0 /* Deleted because there is now a better way to do all this */ + +char * +find_mangled (name) + char *name; +{ + struct mangled *munge; + + for (munge = first_mangle; munge; munge = munge->next) + if (!strcmp (name, munge->normal)) + return munge->mangled; + return 0; +} + + +#ifdef S_ISLNK +void +add_symlink_mangle (symlink, linkto, buffer) + char *symlink; + char *linkto; + char *buffer; +{ + struct mangled *munge, *kludge; + + munge = (struct mangled *) ck_malloc (sizeof (struct mangled) + strlen (symlink) + strlen (linkto) + 2); + if (!first_mangle) + first_mangle = munge; + else + { + for (kludge = first_mangle; kludge->next; kludge = kludge->next) + ; + kludge->next = munge; + } + munge->type = 1; + munge->next = 0; + strcpy (munge->normal, symlink); + munge->linked_to = munge->normal + strlen (symlink) + 1; + strcpy (munge->linked_to, linkto); + sprintf (munge->mangled, "@@MaNgLeD.%d", mangled_num++); + strncpy (buffer, munge->mangled, NAMSIZ); +} + +#endif + +void +add_mangle (name, buffer) + char *name; + char *buffer; +{ + struct mangled *munge, *kludge; + + munge = (struct mangled *) ck_malloc (sizeof (struct mangled) + strlen (name)); + if (!first_mangle) + first_mangle = munge; + else + { + for (kludge = first_mangle; kludge->next; kludge = kludge->next) + ; + kludge->next = munge; + } + munge->next = 0; + munge->type = 0; + strcpy (munge->normal, name); + sprintf (munge->mangled, "@@MaNgLeD.%d", mangled_num++); + strncpy (buffer, munge->mangled, NAMSIZ); +} + +void +write_mangled () +{ + struct mangled *munge; + struct stat hstat; + union record *header; + char *ptr1, *ptr2; + PTR the_buffer; + int size; + int bufsize; + + if (!first_mangle) + return; + the_buffer = init_buffer (); + for (munge = first_mangle, size = 0; munge; munge = munge->next) + { + ptr1 = quote_copy_string (munge->normal); + if (!ptr1) + ptr1 = munge->normal; + if (munge->type) + { + add_buffer (the_buffer, "Symlink ", 8); + add_buffer (the_buffer, ptr1, strlen (ptr1)); + add_buffer (the_buffer, " to ", 4); + + if (ptr2 = quote_copy_string (munge->linked_to)) + { + add_buffer (the_buffer, ptr2, strlen (ptr2)); + free (ptr2); + } + else + add_buffer (the_buffer, munge->linked_to, strlen (munge->linked_to)); + } + else + { + add_buffer (the_buffer, "Rename ", 7); + add_buffer (the_buffer, munge->mangled, strlen (munge->mangled)); + add_buffer (the_buffer, " to ", 4); + add_buffer (the_buffer, ptr1, strlen (ptr1)); + } + add_buffer (the_buffer, "\n", 1); + if (ptr1 != munge->normal) + free (ptr1); + } + + bzero (&hstat, sizeof (struct stat)); + hstat.st_atime = hstat.st_mtime = hstat.st_ctime = time (0); + ptr1 = get_buffer (the_buffer); + hstat.st_size = strlen (ptr1); + + header = start_header ("././@MaNgLeD_NaMeS", &hstat); + header->header.linkflag = LF_NAMES; + finish_header (header); + size = hstat.st_size; + header = findrec (); + bufsize = endofrecs ()->charptr - header->charptr; + + while (bufsize < size) + { + bcopy (ptr1, header->charptr, bufsize); + ptr1 += bufsize; + size -= bufsize; + userec (header + (bufsize - 1) / RECORDSIZE); + header = findrec (); + bufsize = endofrecs ()->charptr - header->charptr; + } + bcopy (ptr1, header->charptr, size); + bzero (header->charptr + size, bufsize - size); + userec (header + (size - 1) / RECORDSIZE); +} + +#endif + +void +extract_mangle (head) + union record *head; +{ + char *buf; + char *fromtape; + char *to; + char *ptr, *ptrend; + char *nam1, *nam1end; + int size; + int copied; + + size = hstat.st_size; + buf = to = ck_malloc (size + 1); + buf[size] = '\0'; + while (size > 0) + { + fromtape = findrec ()->charptr; + if (fromtape == 0) + { + msg ("Unexpected EOF in mangled names!"); + return; + } + copied = endofrecs ()->charptr - fromtape; + if (copied > size) + copied = size; + bcopy (fromtape, to, copied); + to += copied; + size -= copied; + userec ((union record *) (fromtape + copied - 1)); + } + for (ptr = buf; *ptr; ptr = ptrend) + { + ptrend = index (ptr, '\n'); + *ptrend++ = '\0'; + + if (!strncmp (ptr, "Rename ", 7)) + { + nam1 = ptr + 7; + nam1end = index (nam1, ' '); + while (strncmp (nam1end, " to ", 4)) + { + nam1end++; + nam1end = index (nam1end, ' '); + } + *nam1end = '\0'; + if (ptrend[-2] == '/') + ptrend[-2] = '\0'; + un_quote_string (nam1end + 4); + if (rename (nam1, nam1end + 4)) + msg_perror ("Can't rename %s to %s", nam1, nam1end + 4); + else if (f_verbose) + msg ("Renamed %s to %s", nam1, nam1end + 4); + } +#ifdef S_ISLNK + else if (!strncmp (ptr, "Symlink ", 8)) + { + nam1 = ptr + 8; + nam1end = index (nam1, ' '); + while (strncmp (nam1end, " to ", 4)) + { + nam1end++; + nam1end = index (nam1end, ' '); + } + *nam1end = '\0'; + un_quote_string (nam1); + un_quote_string (nam1end + 4); + if (symlink (nam1, nam1end + 4) && (unlink (nam1end + 4) || symlink (nam1, nam1end + 4))) + msg_perror ("Can't symlink %s to %s", nam1, nam1end + 4); + else if (f_verbose) + msg ("Symlinkd %s to %s", nam1, nam1end + 4); + } +#endif + else + msg ("Unknown demangling command %s", ptr); + } +} diff --git a/gnu/usr.bin/tar/msd_dir.h b/gnu/usr.bin/tar/msd_dir.h new file mode 100644 index 000000000000..06c7a644b42d --- /dev/null +++ b/gnu/usr.bin/tar/msd_dir.h @@ -0,0 +1,44 @@ +/* + * @(#)msd_dir.h 1.4 87/11/06 Public Domain. + * + * A public domain implementation of BSD directory routines for + * MS-DOS. Written by Michael Rendell ({uunet,utai}michael@garfield), + * August 1897 + */ + +#define rewinddir(dirp) seekdir(dirp, 0L) + +#define MAXNAMLEN 12 + +#ifdef __TURBOC__ +typedef int ino_t; +typedef int dev_t; +#endif + +struct dirent + { + ino_t d_ino; /* a bit of a farce */ + int d_reclen; /* more farce */ + int d_namlen; /* length of d_name */ + char d_name[MAXNAMLEN + 1]; /* garentee null termination */ + }; + +struct _dircontents + { + char *_d_entry; + struct _dircontents *_d_next; + }; + +typedef struct _dirdesc + { + int dd_id; /* uniquely identify each open directory */ + long dd_loc; /* where we are in directory entry is this */ + struct _dircontents *dd_contents; /* pointer to contents of dir */ + struct _dircontents *dd_cp; /* pointer to current position */ + } DIR; + +extern DIR *opendir (); +extern struct dirent *readdir (); +extern void seekdir (); +extern long telldir (); +extern void closedir (); diff --git a/gnu/usr.bin/tar/names.c b/gnu/usr.bin/tar/names.c new file mode 100644 index 000000000000..0de6a8898a30 --- /dev/null +++ b/gnu/usr.bin/tar/names.c @@ -0,0 +1,149 @@ +/* Look up user and/or group names. + Copyright (C) 1988, 1992 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * Look up user and/or group names. + * + * This file should be modified for non-unix systems to do something + * reasonable. + */ + +#include <sys/types.h> +#include "tar.h" +#include "port.h" + +#ifndef NONAMES +/* Whole module goes away if NONAMES defined. Otherwise... */ +#include <stdio.h> +#include <pwd.h> +#include <grp.h> + +static int saveuid = -993; +static char saveuname[TUNMLEN]; +static int my_uid = -993; + +static int savegid = -993; +static char savegname[TGNMLEN]; +static int my_gid = -993; + +#define myuid ( my_uid < 0? (my_uid = getuid()): my_uid ) +#define mygid ( my_gid < 0? (my_gid = getgid()): my_gid ) + +/* + * Look up a user or group name from a uid/gid, maintaining a cache. + * FIXME, for now it's a one-entry cache. + * FIXME2, the "-993" is to reduce the chance of a hit on the first lookup. + * + * This is ifdef'd because on Suns, it drags in about 38K of "yellow + * pages" code, roughly doubling the program size. Thanks guys. + */ +void +finduname (uname, uid) + char uname[TUNMLEN]; + int uid; +{ + struct passwd *pw; +#ifndef HAVE_GETPWUID + extern struct passwd *getpwuid (); +#endif + + if (uid != saveuid) + { + saveuid = uid; + saveuname[0] = '\0'; + pw = getpwuid (uid); + if (pw) + strncpy (saveuname, pw->pw_name, TUNMLEN); + } + strncpy (uname, saveuname, TUNMLEN); +} + +int +finduid (uname) + char uname[TUNMLEN]; +{ + struct passwd *pw; + extern struct passwd *getpwnam (); + + if (uname[0] != saveuname[0] /* Quick test w/o proc call */ + || 0 != strncmp (uname, saveuname, TUNMLEN)) + { + strncpy (saveuname, uname, TUNMLEN); + pw = getpwnam (uname); + if (pw) + { + saveuid = pw->pw_uid; + } + else + { + saveuid = myuid; + } + } + return saveuid; +} + + +void +findgname (gname, gid) + char gname[TGNMLEN]; + int gid; +{ + struct group *gr; +#ifndef HAVE_GETGRGID + extern struct group *getgrgid (); +#endif + + if (gid != savegid) + { + savegid = gid; + savegname[0] = '\0'; + (void) setgrent (); + gr = getgrgid (gid); + if (gr) + strncpy (savegname, gr->gr_name, TGNMLEN); + } + (void) strncpy (gname, savegname, TGNMLEN); +} + + +int +findgid (gname) + char gname[TUNMLEN]; +{ + struct group *gr; + extern struct group *getgrnam (); + + if (gname[0] != savegname[0] /* Quick test w/o proc call */ + || 0 != strncmp (gname, savegname, TUNMLEN)) + { + strncpy (savegname, gname, TUNMLEN); + gr = getgrnam (gname); + if (gr) + { + savegid = gr->gr_gid; + } + else + { + savegid = mygid; + } + } + return savegid; +} + +#endif diff --git a/gnu/usr.bin/tar/open3.h b/gnu/usr.bin/tar/open3.h new file mode 100644 index 000000000000..c1c0e59b6761 --- /dev/null +++ b/gnu/usr.bin/tar/open3.h @@ -0,0 +1,67 @@ +/* Defines for Sys V style 3-argument open call. + Copyright (C) 1988 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * open3.h -- #defines for the various flags for the Sys V style 3-argument + * open() call. On BSD or System 5, the system already has this in an + * include file. This file is needed for V7 and MINIX systems for the + * benefit of open3() in port.c, a routine that emulates the 3-argument + * call using system calls available on V7/MINIX. + * + * This file is needed by PD tar even if we aren't using the + * emulator, since the #defines for O_WRONLY, etc. are used in + * a couple of places besides the open() calls, (e.g. in the assignment + * to openflag in extract.c). We just #include this rather than + * #ifdef them out. + * + * Written 6/10/87 by rmtodd@uokmax (Richard Todd). + * + * The names have been changed by John Gilmore, 31 July 1987, since + * Richard called it "bsdopen", and really this change was introduced in + * AT&T Unix systems before BSD picked it up. + */ + +/* Only one of the next three should be specified */ +#define O_RDONLY 0 /* only allow read */ +#define O_WRONLY 1 /* only allow write */ +#define O_RDWR 2 /* both are allowed */ + +/* The rest of these can be OR-ed in to the above. */ +/* + * O_NDELAY isn't implemented by the emulator. It's only useful (to tar) on + * systems that have named pipes anyway; it prevents tar's hanging by + * opening a named pipe. We #ifndef it because some systems already have + * it defined. + */ +#ifndef O_NDELAY +#define O_NDELAY 4 /* don't block on opening devices that would + * block on open -- ignored by emulator. */ +#endif +#define O_CREAT 8 /* create file if needed */ +#define O_EXCL 16 /* file cannot already exist */ +#define O_TRUNC 32 /* truncate file on open */ +#define O_APPEND 64 /* always write at end of file -- ignored by emul */ + +#ifdef EMUL_OPEN3 +/* + * make emulation transparent to rest of file -- redirect all open() calls + * to our routine + */ +#define open open3 +#endif diff --git a/gnu/usr.bin/tar/pathmax.h b/gnu/usr.bin/tar/pathmax.h new file mode 100644 index 000000000000..aeba9f7d2186 --- /dev/null +++ b/gnu/usr.bin/tar/pathmax.h @@ -0,0 +1,53 @@ +/* Define PATH_MAX somehow. Requires sys/types.h. + Copyright (C) 1992 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifndef _PATHMAX_H +#define _PATHMAX_H + +#ifdef HAVE_UNISTD_H +#include <unistd.h> +#endif + +/* Non-POSIX BSD systems might have gcc's limits.h, which doesn't define + PATH_MAX but might cause redefinition warnings when sys/param.h is + later included (as on MORE/BSD 4.3). */ +#if defined(_POSIX_VERSION) || (defined(HAVE_LIMITS_H) && defined(USG)) +#include <limits.h> +#endif + +#ifndef _POSIX_PATH_MAX +#define _POSIX_PATH_MAX 255 +#endif + +#if !defined(PATH_MAX) && defined(_PC_PATH_MAX) +#define PATH_MAX (pathconf ("/", _PC_PATH_MAX) < 1 ? 1024 : pathconf ("/", _PC_PATH_MAX)) +#endif + +/* Don't include sys/param.h if it already has been. */ +#if !defined(PATH_MAX) && !defined(MAXPATHLEN) && !defined(__MSDOS__) +#include <sys/param.h> +#endif + +#if !defined(PATH_MAX) && defined(MAXPATHLEN) +#define PATH_MAX MAXPATHLEN +#endif + +#ifndef PATH_MAX +#define PATH_MAX _POSIX_PATH_MAX +#endif + +#endif /* _PATHMAX_H */ diff --git a/gnu/usr.bin/tar/port.c b/gnu/usr.bin/tar/port.c new file mode 100644 index 000000000000..10ec32ed7288 --- /dev/null +++ b/gnu/usr.bin/tar/port.c @@ -0,0 +1,1256 @@ +/* Supporting routines which may sometimes be missing. + Copyright (C) 1988, 1992 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#include <stdio.h> +#include <sys/types.h> +#include <signal.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#include "tar.h" +#include "port.h" + +extern long baserec; + +/* All machine-dependent #ifdefs should appear here, instead of + being scattered through the file. For UN*X systems, it is better to + figure out what is needed in the configure script, for most of the + features. */ + +#ifdef __MSDOS__ +char TTY_NAME[] = "con"; +#define HAVE_STRSTR +#define HAVE_RENAME +#define HAVE_MKDIR +#else +char TTY_NAME[] = "/dev/tty"; +#endif + +/* End of system-dependent #ifdefs */ + + +#ifndef HAVE_VALLOC +/* + * valloc() does a malloc() on a page boundary. On some systems, + * this can make large block I/O more efficient. + */ +char * +valloc (size) + unsigned size; +{ + return (malloc (size)); +} + +#endif /* !HAVE_VALLOC */ + +#ifndef HAVE_MKDIR +/* + * Written by Robert Rother, Mariah Corporation, August 1985. + * + * If you want it, it's yours. All I ask in return is that if you + * figure out how to do this in a Bourne Shell script you send me + * a copy. + * sdcsvax!rmr or rmr@uscd + * + * Severely hacked over by John Gilmore to make a 4.2BSD compatible + * subroutine. 11Mar86; hoptoad!gnu + * + * Modified by rmtodd@uokmax 6-28-87 -- when making an already existing dir, + * subroutine didn't return EEXIST. It does now. + */ + +/* + * Make a directory. + */ +int +mkdir (dpath, dmode) + char *dpath; + int dmode; +{ + int cpid, status; + struct stat statbuf; + + if (stat (dpath, &statbuf) == 0) + { + errno = EEXIST; /* Stat worked, so it already exists */ + return -1; + } + + /* If stat fails for a reason other than non-existence, return error */ + if (errno != ENOENT) + return -1; + + switch (cpid = fork ()) + { + + case -1: /* Error in fork() */ + return (-1); /* Errno is set already */ + + case 0: /* Child process */ + /* + * Cheap hack to set mode of new directory. Since this + * child process is going away anyway, we zap its umask. + * FIXME, this won't suffice to set SUID, SGID, etc. on this + * directory. Does anybody care? + */ + status = umask (0); /* Get current umask */ + status = umask (status | (0777 & ~dmode)); /* Set for mkdir */ + execl ("/bin/mkdir", "mkdir", dpath, (char *) 0); + _exit (-1); /* Can't exec /bin/mkdir */ + + default: /* Parent process */ + while (cpid != wait (&status)); /* Wait for kid to finish */ + } + + if (WIFSIGNALED (status) || WEXITSTATUS (status) != 0) + { + errno = EIO; /* We don't know why, but */ + return -1; /* /bin/mkdir failed */ + } + + return 0; +} + +int +rmdir (dpath) + char *dpath; +{ + int cpid, status; + struct stat statbuf; + + if (stat (dpath, &statbuf) != 0) + { + /* Stat just set errno. We don't have to */ + return -1; + } + + switch (cpid = fork ()) + { + + case -1: /* Error in fork() */ + return (-1); /* Errno is set already */ + + case 0: /* Child process */ + execl ("/bin/rmdir", "rmdir", dpath, (char *) 0); + _exit (-1); /* Can't exec /bin/mkdir */ + + default: /* Parent process */ + while (cpid != wait (&status)); /* Wait for kid to finish */ + } + + if (WIFSIGNALED (status) || WEXITSTATUS (status) != 0) + { + errno = EIO; /* We don't know why, but */ + return -1; /* /bin/mkdir failed */ + } + + return 0; +} + +#endif /* !HAVE_MKDIR */ + +#ifndef HAVE_RENAME +/* Rename file FROM to file TO. + Return 0 if successful, -1 if not. */ + +int +rename (from, to) + char *from; + char *to; +{ + struct stat from_stats; + + if (stat (from, &from_stats)) + return -1; + + if (unlink (to) && errno != ENOENT) + return -1; + + if (link (from, to)) + return -1; + + if (unlink (from) && errno != ENOENT) + { + unlink (to); + return -1; + } + + return 0; +} + +#endif /* !HAVE_RENAME */ + +#ifdef minix +/* Minix has bcopy but not bzero, and no memset. Thanks, Andy. */ +void +bzero (s1, n) + register char *s1; + register int n; +{ + while (n--) + *s1++ = '\0'; +} + +/* It also has no bcmp() */ +int +bcmp (s1, s2, n) + register char *s1, *s2; + register int n; +{ + for (; n--; ++s1, ++s2) + { + if (*s1 != *s2) + return *s1 - *s2; + } + return 0; +} + +/* + * Groan, Minix doesn't have execlp either! + * + * execlp(file,arg0,arg1...argn,(char *)NULL) + * exec a program, automatically searching for the program through + * all the directories on the PATH. + * + * This version is naive about variable argument lists, it assumes + * a straightforward C calling sequence. If your system has odd stacks + * *and* doesn't have execlp, YOU get to fix it. + */ +int +execlp (filename, arg0) + char *filename, *arg0; +{ + register char *p, *path; + register char *fnbuffer; + char **argstart = &arg0; + struct stat statbuf; + extern char **environ; + + if ((p = getenv ("PATH")) == NULL) + { + /* couldn't find path variable -- try to exec given filename */ + return execve (filename, argstart, environ); + } + + /* + * make a place to build the filename. We malloc larger than we + * need, but we know it will fit in this. + */ + fnbuffer = malloc (strlen (p) + 1 + strlen (filename)); + if (fnbuffer == NULL) + { + errno = ENOMEM; + return -1; + } + + /* + * try each component of the path to see if the file's there + * and executable. + */ + for (path = p; path; path = p) + { + /* construct full path name to try */ + if ((p = index (path, ':')) == NULL) + { + strcpy (fnbuffer, path); + } + else + { + strncpy (fnbuffer, path, p - path); + fnbuffer[p - path] = '\0'; + p++; /* Skip : for next time */ + } + if (strlen (fnbuffer) != 0) + strcat (fnbuffer, "/"); + strcat (fnbuffer, filename); + + /* check to see if file is there and is a normal file */ + if (stat (fnbuffer, &statbuf) < 0) + { + if (errno == ENOENT) + continue; /* file not there,keep on looking */ + else + goto fail; /* failed for some reason, return */ + } + if (!S_ISREG (statbuf.st_mode)) + continue; + + if (execve (fnbuffer, argstart, environ) < 0 + && errno != ENOENT + && errno != ENOEXEC) + { + /* failed, for some other reason besides "file + * not found" or "not a.out format" + */ + goto fail; + } + + /* + * If we got error ENOEXEC, the file is executable but is + * not an object file. Try to execute it as a shell script, + * returning error if we can't execute /bin/sh. + * + * FIXME, this code is broken in several ways. Shell + * scripts should not in general be executed by the user's + * SHELL variable program. On more mature systems, the + * script can specify with #!/bin/whatever. Also, this + * code clobbers argstart[-1] if the exec of the shell + * fails. + */ + if (errno == ENOEXEC) + { + char *shell; + + /* Try to execute command "sh arg0 arg1 ..." */ + if ((shell = getenv ("SHELL")) == NULL) + shell = "/bin/sh"; + argstart[-1] = shell; + argstart[0] = fnbuffer; + execve (shell, &argstart[-1], environ); + goto fail; /* Exec didn't work */ + } + + /* + * If we succeeded, the execve() doesn't return, so we + * can only be here is if the file hasn't been found yet. + * Try the next place on the path. + */ + } + + /* all attempts failed to locate the file. Give up. */ + errno = ENOENT; + +fail: + free (fnbuffer); + return -1; +} + +#endif /* minix */ + + +#ifdef EMUL_OPEN3 +#include "open3.h" +/* + * open3 -- routine to emulate the 3-argument open system + * call that is present in most modern Unix systems. + * This version attempts to support all the flag bits except for O_NDELAY + * and O_APPEND, which are silently ignored. The emulation is not as efficient + * as the real thing (at worst, 4 system calls instead of one), but there's + * not much I can do about that. + * + * Written 6/10/87 by rmtodd@uokmax + * + * open3(path, flag, mode) + * Attempts to open the file specified by + * the given pathname. The following flag bits (#defined in tar.h) + * specify options to the routine: + * O_RDONLY file open for read only + * O_WRONLY file open for write only + * O_RDWR file open for both read & write + * (Needless to say, you should only specify one of the above). + * O_CREAT file is created with specified mode if it needs to be. + * O_TRUNC if file exists, it is truncated to 0 bytes + * O_EXCL used with O_CREAT--routine returns error if file exists + * Function returns file descriptor if successful, -1 and errno if not. + */ + +/* + * array to give arguments to access for various modes + * FIXME, this table depends on the specific integer values of O_XXX, + * and also contains integers (args to 'access') that should be #define's. + */ +static int modes[] = +{ + 04, /* O_RDONLY */ + 02, /* O_WRONLY */ + 06, /* O_RDWR */ + 06, /* invalid but we'd better cope -- O_WRONLY+O_RDWR */ +}; + +/* Shut off the automatic emulation of open(), we'll need it. */ +#undef open + +int +open3 (path, flags, mode) + char *path; + int flags, mode; +{ + int exists = 1; + int call_creat = 0; + int fd; + /* + * We actually do the work by calling the open() or creat() system + * call, depending on the flags. Call_creat is true if we will use + * creat(), false if we will use open(). + */ + + /* + * See if the file exists and is accessible in the requested mode. + * + * Strictly speaking we shouldn't be using access, since access checks + * against real uid, and the open call should check against euid. + * Most cases real uid == euid, so it won't matter. FIXME. + * FIXME, the construction "flags & 3" and the modes table depends + * on the specific integer values of the O_XXX #define's. Foo! + */ + if (access (path, modes[flags & 3]) < 0) + { + if (errno == ENOENT) + { + /* the file does not exist */ + exists = 0; + } + else + { + /* probably permission violation */ + if (flags & O_EXCL) + { + /* Oops, the file exists, we didn't want it. */ + /* No matter what the error, claim EEXIST. */ + errno = EEXIST; + } + return -1; + } + } + + /* if we have the O_CREAT bit set, check for O_EXCL */ + if (flags & O_CREAT) + { + if ((flags & O_EXCL) && exists) + { + /* Oops, the file exists and we didn't want it to. */ + errno = EEXIST; + return -1; + } + /* + * If the file doesn't exist, be sure to call creat() so that + * it will be created with the proper mode. + */ + if (!exists) + call_creat = 1; + } + else + { + /* If O_CREAT isn't set and the file doesn't exist, error. */ + if (!exists) + { + errno = ENOENT; + return -1; + } + } + + /* + * If the O_TRUNC flag is set and the file exists, we want to call + * creat() anyway, since creat() guarantees that the file will be + * truncated and open()-for-writing doesn't. + * (If the file doesn't exist, we're calling creat() anyway and the + * file will be created with zero length.) + */ + if ((flags & O_TRUNC) && exists) + call_creat = 1; + /* actually do the call */ + if (call_creat) + { + /* + * call creat. May have to close and reopen the file if we + * want O_RDONLY or O_RDWR access -- creat() only gives + * O_WRONLY. + */ + fd = creat (path, mode); + if (fd < 0 || (flags & O_WRONLY)) + return fd; + if (close (fd) < 0) + return -1; + /* Fall out to reopen the file we've created */ + } + + /* + * calling old open, we strip most of the new flags just in case. + */ + return open (path, flags & (O_RDONLY | O_WRONLY | O_RDWR | O_BINARY)); +} + +#endif /* EMUL_OPEN3 */ + +#ifndef HAVE_MKNOD +#ifdef __MSDOS__ +typedef int dev_t; +#endif +/* Fake mknod by complaining */ +int +mknod (path, mode, dev) + char *path; + unsigned short mode; + dev_t dev; +{ + int fd; + + errno = ENXIO; /* No such device or address */ + return -1; /* Just give an error */ +} + +/* Fake links by copying */ +int +link (path1, path2) + char *path1; + char *path2; +{ + char buf[256]; + int ifd, ofd; + int nrbytes; + int nwbytes; + + fprintf (stderr, "%s: %s: cannot link to %s, copying instead\n", + tar, path1, path2); + if ((ifd = open (path1, O_RDONLY | O_BINARY)) < 0) + return -1; + if ((ofd = creat (path2, 0666)) < 0) + return -1; + setmode (ofd, O_BINARY); + while ((nrbytes = read (ifd, buf, sizeof (buf))) > 0) + { + if ((nwbytes = write (ofd, buf, nrbytes)) != nrbytes) + { + nrbytes = -1; + break; + } + } + /* Note use of "|" rather than "||" below: we want to close + * the files even if an error occurs. + */ + if ((nrbytes < 0) | (0 != close (ifd)) | (0 != close (ofd))) + { + unlink (path2); + return -1; + } + return 0; +} + +/* everyone owns everything on MS-DOS (or is it no one owns anything?) */ +int +chown (path, uid, gid) + char *path; + int uid; + int gid; +{ + return 0; +} + +int +geteuid () +{ + return 0; +} + +#endif /* !HAVE_MKNOD */ + +#ifdef __TURBOC__ +#include <time.h> +#include <fcntl.h> +#include <io.h> + +struct utimbuf +{ + time_t actime; /* Access time. */ + time_t modtime; /* Modification time. */ +}; + +int +utime (char *filename, struct utimbuf *utb) +{ + struct tm *tm; + struct ftime filetime; + time_t when; + int fd; + int status; + + if (utb == 0) + when = time (0); + else + when = utb->modtime; + + fd = _open (filename, O_RDWR); + if (fd == -1) + return -1; + + tm = localtime (&when); + if (tm->tm_year < 80) + filetime.ft_year = 0; + else + filetime.ft_year = tm->tm_year - 80; + filetime.ft_month = tm->tm_mon + 1; + filetime.ft_day = tm->tm_mday; + if (tm->tm_hour < 0) + filetime.ft_hour = 0; + else + filetime.ft_hour = tm->tm_hour; + filetime.ft_min = tm->tm_min; + filetime.ft_tsec = tm->tm_sec / 2; + + status = setftime (fd, &filetime); + _close (fd); + return status; +} + +#endif /* __TURBOC__ */ + +/* Stash argv[0] here so panic will know what the program is called */ +char *myname = 0; + +void +panic (s) + char *s; +{ + if (myname) + fprintf (stderr, "%s:", myname); + fprintf (stderr, s); + putc ('\n', stderr); + exit (12); +} + + +PTR +ck_malloc (size) + size_t size; +{ + PTR ret; + + if (!size) + size++; + ret = malloc (size); + if (ret == 0) + panic ("Couldn't allocate memory"); + return ret; +} + +/* Used by alloca.c and bison.simple. */ +char * +xmalloc (size) + size_t size; +{ + return (char *) ck_malloc (size); +} + +PTR +ck_realloc (ptr, size) + PTR ptr; + size_t size; +{ + PTR ret; + + if (!ptr) + ret = ck_malloc (size); + else + ret = realloc (ptr, size); + if (ret == 0) + panic ("Couldn't re-allocate memory"); + return ret; +} + +/* Implement a variable sized buffer of 'stuff'. We don't know what it is, + nor do we care, as long as it doesn't mind being aligned on a char boundry. + */ + +struct buffer + { + int allocated; + int length; + char *b; + }; + +#define MIN_ALLOCATE 50 + +char * +init_buffer () +{ + struct buffer *b; + + b = (struct buffer *) ck_malloc (sizeof (struct buffer)); + b->allocated = MIN_ALLOCATE; + b->b = (char *) ck_malloc (MIN_ALLOCATE); + b->length = 0; + return (char *) b; +} + +void +flush_buffer (bb) + char *bb; +{ + struct buffer *b; + + b = (struct buffer *) bb; + free (b->b); + b->b = 0; + b->allocated = 0; + b->length = 0; + free ((void *) b); +} + +void +add_buffer (bb, p, n) + char *bb; + char *p; + int n; +{ + struct buffer *b; + + b = (struct buffer *) bb; + if (b->length + n > b->allocated) + { + b->allocated = b->length + n + MIN_ALLOCATE; + b->b = (char *) ck_realloc (b->b, b->allocated); + } + bcopy (p, b->b + b->length, n); + b->length += n; +} + +char * +get_buffer (bb) + char *bb; +{ + struct buffer *b; + + b = (struct buffer *) bb; + return b->b; +} + +char * +merge_sort (list, n, off, cmp) + char *list; + int (*cmp) (); + unsigned n; + int off; +{ + char *ret; + + char *alist, *blist; + unsigned alength, blength; + + char *tptr; + int tmp; + char **prev; +#define NEXTOF(ptr) (* ((char **)(((char *)(ptr))+off) ) ) + if (n == 1) + return list; + if (n == 2) + { + if ((*cmp) (list, NEXTOF (list)) > 0) + { + ret = NEXTOF (list); + NEXTOF (ret) = list; + NEXTOF (list) = 0; + return ret; + } + return list; + } + alist = list; + alength = (n + 1) / 2; + blength = n / 2; + for (tptr = list, tmp = (n - 1) / 2; tmp; tptr = NEXTOF (tptr), tmp--) + ; + blist = NEXTOF (tptr); + NEXTOF (tptr) = 0; + + alist = merge_sort (alist, alength, off, cmp); + blist = merge_sort (blist, blength, off, cmp); + prev = &ret; + for (; alist && blist;) + { + if ((*cmp) (alist, blist) < 0) + { + tptr = NEXTOF (alist); + *prev = alist; + prev = &(NEXTOF (alist)); + alist = tptr; + } + else + { + tptr = NEXTOF (blist); + *prev = blist; + prev = &(NEXTOF (blist)); + blist = tptr; + } + } + if (alist) + *prev = alist; + else + *prev = blist; + + return ret; +} + +void +ck_close (fd) + int fd; +{ + if (close (fd) < 0) + { + msg_perror ("can't close a file #%d", fd); + exit (EX_SYSTEM); + } +} + +#include <ctype.h> + +/* Quote_copy_string is like quote_string, but instead of modifying the + string in place, it malloc-s a copy of the string, and returns that. + If the string does not have to be quoted, it returns the NULL string. + The allocated copy can, of course, be freed with free() after the + caller is done with it. + */ +char * +quote_copy_string (string) + char *string; +{ + char *from_here; + char *to_there = 0; + char *copy_buf = 0; + int c; + int copying = 0; + + from_here = string; + while (*from_here) + { + c = *from_here++; + if (c == '\\') + { + if (!copying) + { + int n; + + n = (from_here - string) - 1; + copying++; + copy_buf = (char *) malloc (n + 5 + strlen (from_here) * 4); + if (!copy_buf) + return 0; + bcopy (string, copy_buf, n); + to_there = copy_buf + n; + } + *to_there++ = '\\'; + *to_there++ = '\\'; + } + else if (isprint (c)) + { + if (copying) + *to_there++ = c; + } + else + { + if (!copying) + { + int n; + + n = (from_here - string) - 1; + copying++; + copy_buf = (char *) malloc (n + 5 + strlen (from_here) * 4); + if (!copy_buf) + return 0; + bcopy (string, copy_buf, n); + to_there = copy_buf + n; + } + *to_there++ = '\\'; + if (c == '\n') + *to_there++ = 'n'; + else if (c == '\t') + *to_there++ = 't'; + else if (c == '\f') + *to_there++ = 'f'; + else if (c == '\b') + *to_there++ = 'b'; + else if (c == '\r') + *to_there++ = 'r'; + else if (c == '\177') + *to_there++ = '?'; + else + { + to_there[0] = (c >> 6) + '0'; + to_there[1] = ((c >> 3) & 07) + '0'; + to_there[2] = (c & 07) + '0'; + to_there += 3; + } + } + } + if (copying) + { + *to_there = '\0'; + return copy_buf; + } + return (char *) 0; +} + + +/* Un_quote_string takes a quoted c-string (like those produced by + quote_string or quote_copy_string and turns it back into the + un-quoted original. This is done in place. + */ + +/* There is no un-quote-copy-string. Write it yourself */ + +char * +un_quote_string (string) + char *string; +{ + char *ret; + char *from_here; + char *to_there; + int tmp; + + ret = string; + to_there = string; + from_here = string; + while (*from_here) + { + if (*from_here != '\\') + { + if (from_here != to_there) + *to_there++ = *from_here++; + else + from_here++, to_there++; + continue; + } + switch (*++from_here) + { + case '\\': + *to_there++ = *from_here++; + break; + case 'n': + *to_there++ = '\n'; + from_here++; + break; + case 't': + *to_there++ = '\t'; + from_here++; + break; + case 'f': + *to_there++ = '\f'; + from_here++; + break; + case 'b': + *to_there++ = '\b'; + from_here++; + break; + case 'r': + *to_there++ = '\r'; + from_here++; + break; + case '?': + *to_there++ = 0177; + from_here++; + break; + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + tmp = *from_here - '0'; + from_here++; + if (*from_here < '0' || *from_here > '7') + { + *to_there++ = tmp; + break; + } + tmp = tmp * 8 + *from_here - '0'; + from_here++; + if (*from_here < '0' || *from_here > '7') + { + *to_there++ = tmp; + break; + } + tmp = tmp * 8 + *from_here - '0'; + from_here++; + *to_there = tmp; + break; + default: + ret = 0; + *to_there++ = '\\'; + *to_there++ = *from_here++; + break; + } + } + if (*to_there) + *to_there++ = '\0'; + return ret; +} + +#ifndef __MSDOS__ +void +ck_pipe (pipes) + int *pipes; +{ + if (pipe (pipes) < 0) + { + msg_perror ("can't open a pipe"); + exit (EX_SYSTEM); + } +} +#endif /* !__MSDOS__ */ + +#ifndef HAVE_STRSTR +/* + * strstr - find first occurrence of wanted in s + */ + +char * /* found string, or NULL if none */ +strstr (s, wanted) + char *s; + char *wanted; +{ + register char *scan; + register size_t len; + register char firstc; + + if (*wanted == '\0') + return (char *) 0; + /* + * The odd placement of the two tests is so "" is findable. + * Also, we inline the first char for speed. + * The ++ on scan has been moved down for optimization. + */ + firstc = *wanted; + len = strlen (wanted); + for (scan = s; *scan != firstc || strncmp (scan, wanted, len) != 0;) + if (*scan++ == '\0') + return (char *) 0; + return scan; +} + +#endif /* !HAVE_STRSTR */ + +#ifndef HAVE_FTRUNCATE + +#ifdef F_CHSIZE +int +ftruncate (fd, length) + int fd; + off_t length; +{ + return fcntl (fd, F_CHSIZE, length); +} + +#else /* !F_CHSIZE */ +#ifdef F_FREESP +/* code courtesy of William Kucharski, kucharsk@Solbourne.com */ + +int +ftruncate (fd, length) + int fd; /* file descriptor */ + off_t length; /* length to set file to */ +{ + struct flock fl; + + fl.l_whence = 0; + fl.l_len = 0; + fl.l_start = length; + fl.l_type = F_WRLCK; /* write lock on file space */ + + /* + * This relies on the UNDOCUMENTED F_FREESP argument to + * fcntl(2), which truncates the file so that it ends at the + * position indicated by fl.l_start. + * + * Will minor miracles never cease? + */ + + if (fcntl (fd, F_FREESP, &fl) < 0) + return -1; + + return 0; +} + +#else /* !F_FREESP */ + +int +ftruncate (fd, length) + int fd; + off_t length; +{ + errno = EIO; + return -1; +} + +#endif /* !F_FREESP */ +#endif /* !F_CHSIZE */ +#endif /* !HAVE_FTRUNCATE */ + + +extern FILE *msg_file; + +#if defined (HAVE_VPRINTF) && __STDC__ +#include <stdarg.h> + +void +msg (char *str,...) +{ + va_list args; + + va_start (args, str); + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + vfprintf (stderr, str, args); + va_end (args); + putc ('\n', stderr); + fflush (stderr); +} + +void +msg_perror (char *str,...) +{ + va_list args; + int save_e; + + save_e = errno; + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + va_start (args, str); + vfprintf (stderr, str, args); + va_end (args); + errno = save_e; + perror (" "); + fflush (stderr); +} + +#endif /* HAVE_VPRINTF and __STDC__ */ + +#if defined(HAVE_VPRINTF) && !__STDC__ +#include <varargs.h> +void +msg (str, va_alist) + char *str; + va_dcl +{ + va_list args; + + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + va_start (args); + vfprintf (stderr, str, args); + va_end (args); + putc ('\n', stderr); + fflush (stderr); +} + +void +msg_perror (str, va_alist) + char *str; + va_dcl +{ + va_list args; + int save_e; + + save_e = errno; + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + va_start (args); + vfprintf (stderr, str, args); + va_end (args); + errno = save_e; + perror (" "); + fflush (stderr); +} + +#endif /* HAVE_VPRINTF and not __STDC__ */ + +#if !defined(HAVE_VPRINTF) && defined(HAVE_DOPRNT) +void +msg (str, args) + char *str; + int args; +{ + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + _doprnt (str, &args, stderr); + putc ('\n', stderr); + fflush (stderr); +} + +void +msg_perror (str, args) + char *str; + int args; +{ + int save_e; + + save_e = errno; + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + _doprnt (str, &args, stderr); + errno = save_e; + perror (" "); + fflush (stderr); +} + +#endif /* !HAVE_VPRINTF and HAVE_DOPRNT */ + +#if !defined(HAVE_VPRINTF) && !defined(HAVE_DOPRNT) +void +msg (str, a1, a2, a3, a4, a5, a6) + char *str; +{ + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + fprintf (stderr, str, a1, a2, a3, a4, a5, a6); + putc ('\n', stderr); + fflush (stderr); +} + +void +msg_perror (str, a1, a2, a3, a4, a5, a6) + char *str; +{ + int save_e; + + save_e = errno; + fflush (msg_file); + fprintf (stderr, "%s: ", tar); + if (f_sayblock) + fprintf (stderr, "rec %d: ", baserec + (ar_record - ar_block)); + fprintf (stderr, str, a1, a2, a3, a4, a5, a6); + fprintf (stderr, ": "); + errno = save_e; + perror (" "); +} + +#endif /* !HAVE_VPRINTF and !HAVE_DOPRNT */ diff --git a/gnu/usr.bin/tar/port.h b/gnu/usr.bin/tar/port.h new file mode 100644 index 000000000000..4e65a9ace886 --- /dev/null +++ b/gnu/usr.bin/tar/port.h @@ -0,0 +1,215 @@ +/* Portability declarations. Requires sys/types.h. + Copyright (C) 1988, 1992 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* AIX requires this to be the first thing in the file. */ +#ifdef __GNUC__ +#define alloca __builtin_alloca +#else /* not __GNUC__ */ +#if HAVE_ALLOCA_H +#include <alloca.h> +#else /* not HAVE_ALLOCA_H */ +#ifdef _AIX + #pragma alloca +#else /* not _AIX */ +char *alloca (); +#endif /* not _AIX */ +#endif /* not HAVE_ALLOCA_H */ +#endif /* not __GNUC__ */ + +#include "pathmax.h" + +#ifdef _POSIX_VERSION +#include <sys/wait.h> +#else /* !_POSIX_VERSION */ +#define WIFSTOPPED(w) (((w) & 0xff) == 0x7f) +#define WIFSIGNALED(w) (((w) & 0xff) != 0x7f && ((w) & 0xff) != 0) +#define WIFEXITED(w) (((w) & 0xff) == 0) + +#define WSTOPSIG(w) (((w) >> 8) & 0xff) +#define WTERMSIG(w) ((w) & 0x7f) +#define WEXITSTATUS(w) (((w) >> 8) & 0xff) +#endif /* _POSIX_VERSION */ + +/* nonstandard */ +#ifndef WIFCOREDUMPED +#define WIFCOREDUMPED(w) (((w) & 0x80) != 0) +#endif + +#ifdef __MSDOS__ +/* missing things from sys/stat.h */ +#define S_ISUID 0 +#define S_ISGID 0 +#define S_ISVTX 0 + +/* device stuff */ +#define makedev(ma, mi) ((ma << 8) | mi) +#define major(dev) (dev) +#define minor(dev) (dev) +typedef long off_t; +#endif /* __MSDOS__ */ + +#if defined(__STDC__) || defined(__TURBOC__) +#define PTR void * +#else +#define PTR char * +#define const +#endif + +/* Since major is a function on SVR4, we can't just use `ifndef major'. */ +#ifdef major /* Might be defined in sys/types.h. */ +#define HAVE_MAJOR +#endif + +#if !defined(HAVE_MAJOR) && defined(MAJOR_IN_MKDEV) +#include <sys/mkdev.h> +#define HAVE_MAJOR +#endif + +#if !defined(HAVE_MAJOR) && defined(MAJOR_IN_SYSMACROS) +#include <sys/sysmacros.h> +#define HAVE_MAJOR +#endif + +#ifndef HAVE_MAJOR +#define major(dev) (((dev) >> 8) & 0xff) +#define minor(dev) ((dev) & 0xff) +#define makedev(maj, min) (((maj) << 8) | (min)) +#endif +#undef HAVE_MAJOR + +#if defined(STDC_HEADERS) || defined(HAVE_STRING_H) +#include <string.h> +#if !defined(__MSDOS__) && !defined(STDC_HEADERS) +#include <memory.h> +#endif +#ifdef index +#undef index +#endif +#ifdef rindex +#undef rindex +#endif +#define index strchr +#define rindex strrchr +#define bcopy(s, d, n) memcpy(d, s, n) +#define bzero(s, n) memset(s, 0, n) +#define bcmp memcmp +#else +#include <strings.h> +#endif + +#if defined(STDC_HEADERS) +#include <stdlib.h> +#else +char *malloc (), *realloc (); +char *getenv (); +#endif + +#ifndef _POSIX_VERSION +#ifdef __MSDOS__ +#include <io.h> +#else /* !__MSDOS__ */ +off_t lseek (); +#endif /* !__MSDOS__ */ +char *getcwd (); +#endif /* !_POSIX_VERSION */ + +#ifndef NULL +#define NULL 0 +#endif + +#ifndef O_BINARY +#define O_BINARY 0 +#endif +#ifndef O_CREAT +#define O_CREAT 0 +#endif +#ifndef O_NDELAY +#define O_NDELAY 0 +#endif +#ifndef O_RDONLY +#define O_RDONLY 0 +#endif +#ifndef O_RDWR +#define O_RDWR 2 +#endif + +#include <sys/stat.h> +#ifndef S_ISREG /* Doesn't have POSIX.1 stat stuff. */ +#define mode_t unsigned short +#endif +#if !defined(S_ISBLK) && defined(S_IFBLK) +#define S_ISBLK(m) (((m) & S_IFMT) == S_IFBLK) +#endif +#if !defined(S_ISCHR) && defined(S_IFCHR) +#define S_ISCHR(m) (((m) & S_IFMT) == S_IFCHR) +#endif +#if !defined(S_ISDIR) && defined(S_IFDIR) +#define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR) +#endif +#if !defined(S_ISREG) && defined(S_IFREG) +#define S_ISREG(m) (((m) & S_IFMT) == S_IFREG) +#endif +#if !defined(S_ISFIFO) && defined(S_IFIFO) +#define S_ISFIFO(m) (((m) & S_IFMT) == S_IFIFO) +#define mkfifo(path, mode) (mknod ((path), (mode) | S_IFIFO, 0)) +#endif +#if !defined(S_ISLNK) && defined(S_IFLNK) +#define S_ISLNK(m) (((m) & S_IFMT) == S_IFLNK) +#endif +#if !defined(S_ISSOCK) && defined(S_IFSOCK) +#define S_ISSOCK(m) (((m) & S_IFMT) == S_IFSOCK) +#endif +#if !defined(S_ISMPB) && defined(S_IFMPB) /* V7 */ +#define S_ISMPB(m) (((m) & S_IFMT) == S_IFMPB) +#define S_ISMPC(m) (((m) & S_IFMT) == S_IFMPC) +#endif +#if !defined(S_ISNWK) && defined(S_IFNWK) /* HP/UX */ +#define S_ISNWK(m) (((m) & S_IFMT) == S_IFNWK) +#endif +#if !defined(S_ISCTG) && defined(S_IFCTG) /* contiguous file */ +#define S_ISCTG(m) (((m) & S_IFMT) == S_IFCTG) +#endif +#if !defined(S_ISVTX) +#define S_ISVTX 0001000 +#endif + +#ifdef __MSDOS__ +#include "msd_dir.h" +#define NLENGTH(direct) ((direct)->d_namlen) + +#else /* not __MSDOS__ */ + +#if defined(DIRENT) || defined(_POSIX_VERSION) +#include <dirent.h> +#define NLENGTH(direct) (strlen((direct)->d_name)) +#else /* not (DIRENT or _POSIX_VERSION) */ +#define dirent direct +#define NLENGTH(direct) ((direct)->d_namlen) +#ifdef SYSNDIR +#include <sys/ndir.h> +#endif /* SYSNDIR */ +#ifdef SYSDIR +#include <sys/dir.h> +#endif /* SYSDIR */ +#ifdef NDIR +#include <ndir.h> +#endif /* NDIR */ +#endif /* DIRENT or _POSIX_VERSION */ + +#endif /* not __MSDOS__ */ diff --git a/gnu/usr.bin/tar/regex.c b/gnu/usr.bin/tar/regex.c new file mode 100644 index 000000000000..cb94d597c6f7 --- /dev/null +++ b/gnu/usr.bin/tar/regex.c @@ -0,0 +1,4932 @@ +/* Extended regular expression matching and search library, + version 0.11. + (Implements POSIX draft P10003.2/D11.2, except for + internationalization features.) + + Copyright (C) 1993 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* AIX requires this to be the first thing in the file. */ +#if defined (_AIX) && !defined (REGEX_MALLOC) + #pragma alloca +#endif + +#define _GNU_SOURCE + +/* We need this for `regex.h', and perhaps for the Emacs include files. */ +#include <sys/types.h> + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +/* The `emacs' switch turns on certain matching commands + that make sense only in Emacs. */ +#ifdef emacs + +#include "lisp.h" +#include "buffer.h" +#include "syntax.h" + +/* Emacs uses `NULL' as a predicate. */ +#undef NULL + +#else /* not emacs */ + +/* We used to test for `BSTRING' here, but only GCC and Emacs define + `BSTRING', as far as I know, and neither of them use this code. */ +#if HAVE_STRING_H || STDC_HEADERS +#include <string.h> +#ifndef bcmp +#define bcmp(s1, s2, n) memcmp ((s1), (s2), (n)) +#endif +#ifndef bcopy +#define bcopy(s, d, n) memcpy ((d), (s), (n)) +#endif +#ifndef bzero +#define bzero(s, n) memset ((s), 0, (n)) +#endif +#else +#include <strings.h> +#endif + +#ifdef STDC_HEADERS +#include <stdlib.h> +#else +char *malloc (); +char *realloc (); +#endif + + +/* Define the syntax stuff for \<, \>, etc. */ + +/* This must be nonzero for the wordchar and notwordchar pattern + commands in re_match_2. */ +#ifndef Sword +#define Sword 1 +#endif + +#ifdef SYNTAX_TABLE + +extern char *re_syntax_table; + +#else /* not SYNTAX_TABLE */ + +/* How many characters in the character set. */ +#define CHAR_SET_SIZE 256 + +static char re_syntax_table[CHAR_SET_SIZE]; + +static void +init_syntax_once () +{ + register int c; + static int done = 0; + + if (done) + return; + + bzero (re_syntax_table, sizeof re_syntax_table); + + for (c = 'a'; c <= 'z'; c++) + re_syntax_table[c] = Sword; + + for (c = 'A'; c <= 'Z'; c++) + re_syntax_table[c] = Sword; + + for (c = '0'; c <= '9'; c++) + re_syntax_table[c] = Sword; + + re_syntax_table['_'] = Sword; + + done = 1; +} + +#endif /* not SYNTAX_TABLE */ + +#define SYNTAX(c) re_syntax_table[c] + +#endif /* not emacs */ + +/* Get the interface, including the syntax bits. */ +#include "regex.h" + +/* isalpha etc. are used for the character classes. */ +#include <ctype.h> + +#ifndef isascii +#define isascii(c) 1 +#endif + +#ifdef isblank +#define ISBLANK(c) (isascii (c) && isblank (c)) +#else +#define ISBLANK(c) ((c) == ' ' || (c) == '\t') +#endif +#ifdef isgraph +#define ISGRAPH(c) (isascii (c) && isgraph (c)) +#else +#define ISGRAPH(c) (isascii (c) && isprint (c) && !isspace (c)) +#endif + +#define ISPRINT(c) (isascii (c) && isprint (c)) +#define ISDIGIT(c) (isascii (c) && isdigit (c)) +#define ISALNUM(c) (isascii (c) && isalnum (c)) +#define ISALPHA(c) (isascii (c) && isalpha (c)) +#define ISCNTRL(c) (isascii (c) && iscntrl (c)) +#define ISLOWER(c) (isascii (c) && islower (c)) +#define ISPUNCT(c) (isascii (c) && ispunct (c)) +#define ISSPACE(c) (isascii (c) && isspace (c)) +#define ISUPPER(c) (isascii (c) && isupper (c)) +#define ISXDIGIT(c) (isascii (c) && isxdigit (c)) + +#ifndef NULL +#define NULL 0 +#endif + +/* We remove any previous definition of `SIGN_EXTEND_CHAR', + since ours (we hope) works properly with all combinations of + machines, compilers, `char' and `unsigned char' argument types. + (Per Bothner suggested the basic approach.) */ +#undef SIGN_EXTEND_CHAR +#if __STDC__ +#define SIGN_EXTEND_CHAR(c) ((signed char) (c)) +#else /* not __STDC__ */ +/* As in Harbison and Steele. */ +#define SIGN_EXTEND_CHAR(c) ((((unsigned char) (c)) ^ 128) - 128) +#endif + +/* Should we use malloc or alloca? If REGEX_MALLOC is not defined, we + use `alloca' instead of `malloc'. This is because using malloc in + re_search* or re_match* could cause memory leaks when C-g is used in + Emacs; also, malloc is slower and causes storage fragmentation. On + the other hand, malloc is more portable, and easier to debug. + + Because we sometimes use alloca, some routines have to be macros, + not functions -- `alloca'-allocated space disappears at the end of the + function it is called in. */ + +#ifdef REGEX_MALLOC + +#define REGEX_ALLOCATE malloc +#define REGEX_REALLOCATE(source, osize, nsize) realloc (source, nsize) + +#else /* not REGEX_MALLOC */ + +/* Emacs already defines alloca, sometimes. */ +#ifndef alloca + +/* Make alloca work the best possible way. */ +#ifdef __GNUC__ +#define alloca __builtin_alloca +#else /* not __GNUC__ */ +#if HAVE_ALLOCA_H +#include <alloca.h> +#else /* not __GNUC__ or HAVE_ALLOCA_H */ +#ifndef _AIX /* Already did AIX, up at the top. */ +char *alloca (); +#endif /* not _AIX */ +#endif /* not HAVE_ALLOCA_H */ +#endif /* not __GNUC__ */ + +#endif /* not alloca */ + +#define REGEX_ALLOCATE alloca + +/* Assumes a `char *destination' variable. */ +#define REGEX_REALLOCATE(source, osize, nsize) \ + (destination = (char *) alloca (nsize), \ + bcopy (source, destination, osize), \ + destination) + +#endif /* not REGEX_MALLOC */ + + +/* True if `size1' is non-NULL and PTR is pointing anywhere inside + `string1' or just past its end. This works if PTR is NULL, which is + a good thing. */ +#define FIRST_STRING_P(ptr) \ + (size1 && string1 <= (ptr) && (ptr) <= string1 + size1) + +/* (Re)Allocate N items of type T using malloc, or fail. */ +#define TALLOC(n, t) ((t *) malloc ((n) * sizeof (t))) +#define RETALLOC(addr, n, t) ((addr) = (t *) realloc (addr, (n) * sizeof (t))) +#define REGEX_TALLOC(n, t) ((t *) REGEX_ALLOCATE ((n) * sizeof (t))) + +#define BYTEWIDTH 8 /* In bits. */ + +#define STREQ(s1, s2) ((strcmp (s1, s2) == 0)) + +#define MAX(a, b) ((a) > (b) ? (a) : (b)) +#define MIN(a, b) ((a) < (b) ? (a) : (b)) + +typedef char boolean; +#define false 0 +#define true 1 + +/* These are the command codes that appear in compiled regular + expressions. Some opcodes are followed by argument bytes. A + command code can specify any interpretation whatsoever for its + arguments. Zero bytes may appear in the compiled regular expression. + + The value of `exactn' is needed in search.c (search_buffer) in Emacs. + So regex.h defines a symbol `RE_EXACTN_VALUE' to be 1; the value of + `exactn' we use here must also be 1. */ + +typedef enum +{ + no_op = 0, + + /* Followed by one byte giving n, then by n literal bytes. */ + exactn = 1, + + /* Matches any (more or less) character. */ + anychar, + + /* Matches any one char belonging to specified set. First + following byte is number of bitmap bytes. Then come bytes + for a bitmap saying which chars are in. Bits in each byte + are ordered low-bit-first. A character is in the set if its + bit is 1. A character too large to have a bit in the map is + automatically not in the set. */ + charset, + + /* Same parameters as charset, but match any character that is + not one of those specified. */ + charset_not, + + /* Start remembering the text that is matched, for storing in a + register. Followed by one byte with the register number, in + the range 0 to one less than the pattern buffer's re_nsub + field. Then followed by one byte with the number of groups + inner to this one. (This last has to be part of the + start_memory only because we need it in the on_failure_jump + of re_match_2.) */ + start_memory, + + /* Stop remembering the text that is matched and store it in a + memory register. Followed by one byte with the register + number, in the range 0 to one less than `re_nsub' in the + pattern buffer, and one byte with the number of inner groups, + just like `start_memory'. (We need the number of inner + groups here because we don't have any easy way of finding the + corresponding start_memory when we're at a stop_memory.) */ + stop_memory, + + /* Match a duplicate of something remembered. Followed by one + byte containing the register number. */ + duplicate, + + /* Fail unless at beginning of line. */ + begline, + + /* Fail unless at end of line. */ + endline, + + /* Succeeds if at beginning of buffer (if emacs) or at beginning + of string to be matched (if not). */ + begbuf, + + /* Analogously, for end of buffer/string. */ + endbuf, + + /* Followed by two byte relative address to which to jump. */ + jump, + + /* Same as jump, but marks the end of an alternative. */ + jump_past_alt, + + /* Followed by two-byte relative address of place to resume at + in case of failure. */ + on_failure_jump, + + /* Like on_failure_jump, but pushes a placeholder instead of the + current string position when executed. */ + on_failure_keep_string_jump, + + /* Throw away latest failure point and then jump to following + two-byte relative address. */ + pop_failure_jump, + + /* Change to pop_failure_jump if know won't have to backtrack to + match; otherwise change to jump. This is used to jump + back to the beginning of a repeat. If what follows this jump + clearly won't match what the repeat does, such that we can be + sure that there is no use backtracking out of repetitions + already matched, then we change it to a pop_failure_jump. + Followed by two-byte address. */ + maybe_pop_jump, + + /* Jump to following two-byte address, and push a dummy failure + point. This failure point will be thrown away if an attempt + is made to use it for a failure. A `+' construct makes this + before the first repeat. Also used as an intermediary kind + of jump when compiling an alternative. */ + dummy_failure_jump, + + /* Push a dummy failure point and continue. Used at the end of + alternatives. */ + push_dummy_failure, + + /* Followed by two-byte relative address and two-byte number n. + After matching N times, jump to the address upon failure. */ + succeed_n, + + /* Followed by two-byte relative address, and two-byte number n. + Jump to the address N times, then fail. */ + jump_n, + + /* Set the following two-byte relative address to the + subsequent two-byte number. The address *includes* the two + bytes of number. */ + set_number_at, + + wordchar, /* Matches any word-constituent character. */ + notwordchar, /* Matches any char that is not a word-constituent. */ + + wordbeg, /* Succeeds if at word beginning. */ + wordend, /* Succeeds if at word end. */ + + wordbound, /* Succeeds if at a word boundary. */ + notwordbound /* Succeeds if not at a word boundary. */ + +#ifdef emacs + ,before_dot, /* Succeeds if before point. */ + at_dot, /* Succeeds if at point. */ + after_dot, /* Succeeds if after point. */ + + /* Matches any character whose syntax is specified. Followed by + a byte which contains a syntax code, e.g., Sword. */ + syntaxspec, + + /* Matches any character whose syntax is not that specified. */ + notsyntaxspec +#endif /* emacs */ +} re_opcode_t; + +/* Common operations on the compiled pattern. */ + +/* Store NUMBER in two contiguous bytes starting at DESTINATION. */ + +#define STORE_NUMBER(destination, number) \ + do { \ + (destination)[0] = (number) & 0377; \ + (destination)[1] = (number) >> 8; \ + } while (0) + +/* Same as STORE_NUMBER, except increment DESTINATION to + the byte after where the number is stored. Therefore, DESTINATION + must be an lvalue. */ + +#define STORE_NUMBER_AND_INCR(destination, number) \ + do { \ + STORE_NUMBER (destination, number); \ + (destination) += 2; \ + } while (0) + +/* Put into DESTINATION a number stored in two contiguous bytes starting + at SOURCE. */ + +#define EXTRACT_NUMBER(destination, source) \ + do { \ + (destination) = *(source) & 0377; \ + (destination) += SIGN_EXTEND_CHAR (*((source) + 1)) << 8; \ + } while (0) + +#ifdef DEBUG +static void +extract_number (dest, source) + int *dest; + unsigned char *source; +{ + int temp = SIGN_EXTEND_CHAR (*(source + 1)); + *dest = *source & 0377; + *dest += temp << 8; +} + +#ifndef EXTRACT_MACROS /* To debug the macros. */ +#undef EXTRACT_NUMBER +#define EXTRACT_NUMBER(dest, src) extract_number (&dest, src) +#endif /* not EXTRACT_MACROS */ + +#endif /* DEBUG */ + +/* Same as EXTRACT_NUMBER, except increment SOURCE to after the number. + SOURCE must be an lvalue. */ + +#define EXTRACT_NUMBER_AND_INCR(destination, source) \ + do { \ + EXTRACT_NUMBER (destination, source); \ + (source) += 2; \ + } while (0) + +#ifdef DEBUG +static void +extract_number_and_incr (destination, source) + int *destination; + unsigned char **source; +{ + extract_number (destination, *source); + *source += 2; +} + +#ifndef EXTRACT_MACROS +#undef EXTRACT_NUMBER_AND_INCR +#define EXTRACT_NUMBER_AND_INCR(dest, src) \ + extract_number_and_incr (&dest, &src) +#endif /* not EXTRACT_MACROS */ + +#endif /* DEBUG */ + +/* If DEBUG is defined, Regex prints many voluminous messages about what + it is doing (if the variable `debug' is nonzero). If linked with the + main program in `iregex.c', you can enter patterns and strings + interactively. And if linked with the main program in `main.c' and + the other test files, you can run the already-written tests. */ + +#ifdef DEBUG + +/* We use standard I/O for debugging. */ +#include <stdio.h> + +/* It is useful to test things that ``must'' be true when debugging. */ +#include <assert.h> + +static int debug = 0; + +#define DEBUG_STATEMENT(e) e +#define DEBUG_PRINT1(x) if (debug) printf (x) +#define DEBUG_PRINT2(x1, x2) if (debug) printf (x1, x2) +#define DEBUG_PRINT3(x1, x2, x3) if (debug) printf (x1, x2, x3) +#define DEBUG_PRINT4(x1, x2, x3, x4) if (debug) printf (x1, x2, x3, x4) +#define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) \ + if (debug) print_partial_compiled_pattern (s, e) +#define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) \ + if (debug) print_double_string (w, s1, sz1, s2, sz2) + + +extern void printchar (); + +/* Print the fastmap in human-readable form. */ + +void +print_fastmap (fastmap) + char *fastmap; +{ + unsigned was_a_range = 0; + unsigned i = 0; + + while (i < (1 << BYTEWIDTH)) + { + if (fastmap[i++]) + { + was_a_range = 0; + printchar (i - 1); + while (i < (1 << BYTEWIDTH) && fastmap[i]) + { + was_a_range = 1; + i++; + } + if (was_a_range) + { + printf ("-"); + printchar (i - 1); + } + } + } + putchar ('\n'); +} + + +/* Print a compiled pattern string in human-readable form, starting at + the START pointer into it and ending just before the pointer END. */ + +void +print_partial_compiled_pattern (start, end) + unsigned char *start; + unsigned char *end; +{ + int mcnt, mcnt2; + unsigned char *p = start; + unsigned char *pend = end; + + if (start == NULL) + { + printf ("(null)\n"); + return; + } + + /* Loop over pattern commands. */ + while (p < pend) + { + switch ((re_opcode_t) *p++) + { + case no_op: + printf ("/no_op"); + break; + + case exactn: + mcnt = *p++; + printf ("/exactn/%d", mcnt); + do + { + putchar ('/'); + printchar (*p++); + } + while (--mcnt); + break; + + case start_memory: + mcnt = *p++; + printf ("/start_memory/%d/%d", mcnt, *p++); + break; + + case stop_memory: + mcnt = *p++; + printf ("/stop_memory/%d/%d", mcnt, *p++); + break; + + case duplicate: + printf ("/duplicate/%d", *p++); + break; + + case anychar: + printf ("/anychar"); + break; + + case charset: + case charset_not: + { + register int c; + + printf ("/charset%s", + (re_opcode_t) *(p - 1) == charset_not ? "_not" : ""); + + assert (p + *p < pend); + + for (c = 0; c < *p; c++) + { + unsigned bit; + unsigned char map_byte = p[1 + c]; + + putchar ('/'); + + for (bit = 0; bit < BYTEWIDTH; bit++) + if (map_byte & (1 << bit)) + printchar (c * BYTEWIDTH + bit); + } + p += 1 + *p; + break; + } + + case begline: + printf ("/begline"); + break; + + case endline: + printf ("/endline"); + break; + + case on_failure_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/on_failure_jump/0/%d", mcnt); + break; + + case on_failure_keep_string_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/on_failure_keep_string_jump/0/%d", mcnt); + break; + + case dummy_failure_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/dummy_failure_jump/0/%d", mcnt); + break; + + case push_dummy_failure: + printf ("/push_dummy_failure"); + break; + + case maybe_pop_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/maybe_pop_jump/0/%d", mcnt); + break; + + case pop_failure_jump: + extract_number_and_incr (&mcnt, &p); + printf ("/pop_failure_jump/0/%d", mcnt); + break; + + case jump_past_alt: + extract_number_and_incr (&mcnt, &p); + printf ("/jump_past_alt/0/%d", mcnt); + break; + + case jump: + extract_number_and_incr (&mcnt, &p); + printf ("/jump/0/%d", mcnt); + break; + + case succeed_n: + extract_number_and_incr (&mcnt, &p); + extract_number_and_incr (&mcnt2, &p); + printf ("/succeed_n/0/%d/0/%d", mcnt, mcnt2); + break; + + case jump_n: + extract_number_and_incr (&mcnt, &p); + extract_number_and_incr (&mcnt2, &p); + printf ("/jump_n/0/%d/0/%d", mcnt, mcnt2); + break; + + case set_number_at: + extract_number_and_incr (&mcnt, &p); + extract_number_and_incr (&mcnt2, &p); + printf ("/set_number_at/0/%d/0/%d", mcnt, mcnt2); + break; + + case wordbound: + printf ("/wordbound"); + break; + + case notwordbound: + printf ("/notwordbound"); + break; + + case wordbeg: + printf ("/wordbeg"); + break; + + case wordend: + printf ("/wordend"); + +#ifdef emacs + case before_dot: + printf ("/before_dot"); + break; + + case at_dot: + printf ("/at_dot"); + break; + + case after_dot: + printf ("/after_dot"); + break; + + case syntaxspec: + printf ("/syntaxspec"); + mcnt = *p++; + printf ("/%d", mcnt); + break; + + case notsyntaxspec: + printf ("/notsyntaxspec"); + mcnt = *p++; + printf ("/%d", mcnt); + break; +#endif /* emacs */ + + case wordchar: + printf ("/wordchar"); + break; + + case notwordchar: + printf ("/notwordchar"); + break; + + case begbuf: + printf ("/begbuf"); + break; + + case endbuf: + printf ("/endbuf"); + break; + + default: + printf ("?%d", *(p-1)); + } + } + printf ("/\n"); +} + + +void +print_compiled_pattern (bufp) + struct re_pattern_buffer *bufp; +{ + unsigned char *buffer = bufp->buffer; + + print_partial_compiled_pattern (buffer, buffer + bufp->used); + printf ("%d bytes used/%d bytes allocated.\n", bufp->used, bufp->allocated); + + if (bufp->fastmap_accurate && bufp->fastmap) + { + printf ("fastmap: "); + print_fastmap (bufp->fastmap); + } + + printf ("re_nsub: %d\t", bufp->re_nsub); + printf ("regs_alloc: %d\t", bufp->regs_allocated); + printf ("can_be_null: %d\t", bufp->can_be_null); + printf ("newline_anchor: %d\n", bufp->newline_anchor); + printf ("no_sub: %d\t", bufp->no_sub); + printf ("not_bol: %d\t", bufp->not_bol); + printf ("not_eol: %d\t", bufp->not_eol); + printf ("syntax: %d\n", bufp->syntax); + /* Perhaps we should print the translate table? */ +} + + +void +print_double_string (where, string1, size1, string2, size2) + const char *where; + const char *string1; + const char *string2; + int size1; + int size2; +{ + unsigned this_char; + + if (where == NULL) + printf ("(null)"); + else + { + if (FIRST_STRING_P (where)) + { + for (this_char = where - string1; this_char < size1; this_char++) + printchar (string1[this_char]); + + where = string2; + } + + for (this_char = where - string2; this_char < size2; this_char++) + printchar (string2[this_char]); + } +} + +#else /* not DEBUG */ + +#undef assert +#define assert(e) + +#define DEBUG_STATEMENT(e) +#define DEBUG_PRINT1(x) +#define DEBUG_PRINT2(x1, x2) +#define DEBUG_PRINT3(x1, x2, x3) +#define DEBUG_PRINT4(x1, x2, x3, x4) +#define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) +#define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) + +#endif /* not DEBUG */ + +/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can + also be assigned to arbitrarily: each pattern buffer stores its own + syntax, so it can be changed between regex compilations. */ +reg_syntax_t re_syntax_options = RE_SYNTAX_EMACS; + + +/* Specify the precise syntax of regexps for compilation. This provides + for compatibility for various utilities which historically have + different, incompatible syntaxes. + + The argument SYNTAX is a bit mask comprised of the various bits + defined in regex.h. We return the old syntax. */ + +reg_syntax_t +re_set_syntax (syntax) + reg_syntax_t syntax; +{ + reg_syntax_t ret = re_syntax_options; + + re_syntax_options = syntax; + return ret; +} + +/* This table gives an error message for each of the error codes listed + in regex.h. Obviously the order here has to be same as there. */ + +static const char *re_error_msg[] = + { NULL, /* REG_NOERROR */ + "No match", /* REG_NOMATCH */ + "Invalid regular expression", /* REG_BADPAT */ + "Invalid collation character", /* REG_ECOLLATE */ + "Invalid character class name", /* REG_ECTYPE */ + "Trailing backslash", /* REG_EESCAPE */ + "Invalid back reference", /* REG_ESUBREG */ + "Unmatched [ or [^", /* REG_EBRACK */ + "Unmatched ( or \\(", /* REG_EPAREN */ + "Unmatched \\{", /* REG_EBRACE */ + "Invalid content of \\{\\}", /* REG_BADBR */ + "Invalid range end", /* REG_ERANGE */ + "Memory exhausted", /* REG_ESPACE */ + "Invalid preceding regular expression", /* REG_BADRPT */ + "Premature end of regular expression", /* REG_EEND */ + "Regular expression too big", /* REG_ESIZE */ + "Unmatched ) or \\)", /* REG_ERPAREN */ + }; + +/* Subroutine declarations and macros for regex_compile. */ + +static void store_op1 (), store_op2 (); +static void insert_op1 (), insert_op2 (); +static boolean at_begline_loc_p (), at_endline_loc_p (); +static boolean group_in_compile_stack (); +static reg_errcode_t compile_range (); + +/* Fetch the next character in the uncompiled pattern---translating it + if necessary. Also cast from a signed character in the constant + string passed to us by the user to an unsigned char that we can use + as an array index (in, e.g., `translate'). */ +#define PATFETCH(c) \ + do {if (p == pend) return REG_EEND; \ + c = (unsigned char) *p++; \ + if (translate) c = translate[c]; \ + } while (0) + +/* Fetch the next character in the uncompiled pattern, with no + translation. */ +#define PATFETCH_RAW(c) \ + do {if (p == pend) return REG_EEND; \ + c = (unsigned char) *p++; \ + } while (0) + +/* Go backwards one character in the pattern. */ +#define PATUNFETCH p-- + + +/* If `translate' is non-null, return translate[D], else just D. We + cast the subscript to translate because some data is declared as + `char *', to avoid warnings when a string constant is passed. But + when we use a character as a subscript we must make it unsigned. */ +#define TRANSLATE(d) (translate ? translate[(unsigned char) (d)] : (d)) + + +/* Macros for outputting the compiled pattern into `buffer'. */ + +/* If the buffer isn't allocated when it comes in, use this. */ +#define INIT_BUF_SIZE 32 + +/* Make sure we have at least N more bytes of space in buffer. */ +#define GET_BUFFER_SPACE(n) \ + while (b - bufp->buffer + (n) > bufp->allocated) \ + EXTEND_BUFFER () + +/* Make sure we have one more byte of buffer space and then add C to it. */ +#define BUF_PUSH(c) \ + do { \ + GET_BUFFER_SPACE (1); \ + *b++ = (unsigned char) (c); \ + } while (0) + + +/* Ensure we have two more bytes of buffer space and then append C1 and C2. */ +#define BUF_PUSH_2(c1, c2) \ + do { \ + GET_BUFFER_SPACE (2); \ + *b++ = (unsigned char) (c1); \ + *b++ = (unsigned char) (c2); \ + } while (0) + + +/* As with BUF_PUSH_2, except for three bytes. */ +#define BUF_PUSH_3(c1, c2, c3) \ + do { \ + GET_BUFFER_SPACE (3); \ + *b++ = (unsigned char) (c1); \ + *b++ = (unsigned char) (c2); \ + *b++ = (unsigned char) (c3); \ + } while (0) + + +/* Store a jump with opcode OP at LOC to location TO. We store a + relative address offset by the three bytes the jump itself occupies. */ +#define STORE_JUMP(op, loc, to) \ + store_op1 (op, loc, (to) - (loc) - 3) + +/* Likewise, for a two-argument jump. */ +#define STORE_JUMP2(op, loc, to, arg) \ + store_op2 (op, loc, (to) - (loc) - 3, arg) + +/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */ +#define INSERT_JUMP(op, loc, to) \ + insert_op1 (op, loc, (to) - (loc) - 3, b) + +/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */ +#define INSERT_JUMP2(op, loc, to, arg) \ + insert_op2 (op, loc, (to) - (loc) - 3, arg, b) + + +/* This is not an arbitrary limit: the arguments which represent offsets + into the pattern are two bytes long. So if 2^16 bytes turns out to + be too small, many things would have to change. */ +#define MAX_BUF_SIZE (1L << 16) + + +/* Extend the buffer by twice its current size via realloc and + reset the pointers that pointed into the old block to point to the + correct places in the new one. If extending the buffer results in it + being larger than MAX_BUF_SIZE, then flag memory exhausted. */ +#define EXTEND_BUFFER() \ + do { \ + unsigned char *old_buffer = bufp->buffer; \ + if (bufp->allocated == MAX_BUF_SIZE) \ + return REG_ESIZE; \ + bufp->allocated <<= 1; \ + if (bufp->allocated > MAX_BUF_SIZE) \ + bufp->allocated = MAX_BUF_SIZE; \ + bufp->buffer = (unsigned char *) realloc (bufp->buffer, bufp->allocated);\ + if (bufp->buffer == NULL) \ + return REG_ESPACE; \ + /* If the buffer moved, move all the pointers into it. */ \ + if (old_buffer != bufp->buffer) \ + { \ + b = (b - old_buffer) + bufp->buffer; \ + begalt = (begalt - old_buffer) + bufp->buffer; \ + if (fixup_alt_jump) \ + fixup_alt_jump = (fixup_alt_jump - old_buffer) + bufp->buffer;\ + if (laststart) \ + laststart = (laststart - old_buffer) + bufp->buffer; \ + if (pending_exact) \ + pending_exact = (pending_exact - old_buffer) + bufp->buffer; \ + } \ + } while (0) + + +/* Since we have one byte reserved for the register number argument to + {start,stop}_memory, the maximum number of groups we can report + things about is what fits in that byte. */ +#define MAX_REGNUM 255 + +/* But patterns can have more than `MAX_REGNUM' registers. We just + ignore the excess. */ +typedef unsigned regnum_t; + + +/* Macros for the compile stack. */ + +/* Since offsets can go either forwards or backwards, this type needs to + be able to hold values from -(MAX_BUF_SIZE - 1) to MAX_BUF_SIZE - 1. */ +typedef int pattern_offset_t; + +typedef struct +{ + pattern_offset_t begalt_offset; + pattern_offset_t fixup_alt_jump; + pattern_offset_t inner_group_offset; + pattern_offset_t laststart_offset; + regnum_t regnum; +} compile_stack_elt_t; + + +typedef struct +{ + compile_stack_elt_t *stack; + unsigned size; + unsigned avail; /* Offset of next open position. */ +} compile_stack_type; + + +#define INIT_COMPILE_STACK_SIZE 32 + +#define COMPILE_STACK_EMPTY (compile_stack.avail == 0) +#define COMPILE_STACK_FULL (compile_stack.avail == compile_stack.size) + +/* The next available element. */ +#define COMPILE_STACK_TOP (compile_stack.stack[compile_stack.avail]) + + +/* Set the bit for character C in a list. */ +#define SET_LIST_BIT(c) \ + (b[((unsigned char) (c)) / BYTEWIDTH] \ + |= 1 << (((unsigned char) c) % BYTEWIDTH)) + + +/* Get the next unsigned number in the uncompiled pattern. */ +#define GET_UNSIGNED_NUMBER(num) \ + { if (p != pend) \ + { \ + PATFETCH (c); \ + while (ISDIGIT (c)) \ + { \ + if (num < 0) \ + num = 0; \ + num = num * 10 + c - '0'; \ + if (p == pend) \ + break; \ + PATFETCH (c); \ + } \ + } \ + } + +#define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */ + +#define IS_CHAR_CLASS(string) \ + (STREQ (string, "alpha") || STREQ (string, "upper") \ + || STREQ (string, "lower") || STREQ (string, "digit") \ + || STREQ (string, "alnum") || STREQ (string, "xdigit") \ + || STREQ (string, "space") || STREQ (string, "print") \ + || STREQ (string, "punct") || STREQ (string, "graph") \ + || STREQ (string, "cntrl") || STREQ (string, "blank")) + +/* `regex_compile' compiles PATTERN (of length SIZE) according to SYNTAX. + Returns one of error codes defined in `regex.h', or zero for success. + + Assumes the `allocated' (and perhaps `buffer') and `translate' + fields are set in BUFP on entry. + + If it succeeds, results are put in BUFP (if it returns an error, the + contents of BUFP are undefined): + `buffer' is the compiled pattern; + `syntax' is set to SYNTAX; + `used' is set to the length of the compiled pattern; + `fastmap_accurate' is zero; + `re_nsub' is the number of subexpressions in PATTERN; + `not_bol' and `not_eol' are zero; + + The `fastmap' and `newline_anchor' fields are neither + examined nor set. */ + +static reg_errcode_t +regex_compile (pattern, size, syntax, bufp) + const char *pattern; + int size; + reg_syntax_t syntax; + struct re_pattern_buffer *bufp; +{ + /* We fetch characters from PATTERN here. Even though PATTERN is + `char *' (i.e., signed), we declare these variables as unsigned, so + they can be reliably used as array indices. */ + register unsigned char c, c1; + + /* A random tempory spot in PATTERN. */ + const char *p1; + + /* Points to the end of the buffer, where we should append. */ + register unsigned char *b; + + /* Keeps track of unclosed groups. */ + compile_stack_type compile_stack; + + /* Points to the current (ending) position in the pattern. */ + const char *p = pattern; + const char *pend = pattern + size; + + /* How to translate the characters in the pattern. */ + char *translate = bufp->translate; + + /* Address of the count-byte of the most recently inserted `exactn' + command. This makes it possible to tell if a new exact-match + character can be added to that command or if the character requires + a new `exactn' command. */ + unsigned char *pending_exact = 0; + + /* Address of start of the most recently finished expression. + This tells, e.g., postfix * where to find the start of its + operand. Reset at the beginning of groups and alternatives. */ + unsigned char *laststart = 0; + + /* Address of beginning of regexp, or inside of last group. */ + unsigned char *begalt; + + /* Place in the uncompiled pattern (i.e., the {) to + which to go back if the interval is invalid. */ + const char *beg_interval; + + /* Address of the place where a forward jump should go to the end of + the containing expression. Each alternative of an `or' -- except the + last -- ends with a forward jump of this sort. */ + unsigned char *fixup_alt_jump = 0; + + /* Counts open-groups as they are encountered. Remembered for the + matching close-group on the compile stack, so the same register + number is put in the stop_memory as the start_memory. */ + regnum_t regnum = 0; + +#ifdef DEBUG + DEBUG_PRINT1 ("\nCompiling pattern: "); + if (debug) + { + unsigned debug_count; + + for (debug_count = 0; debug_count < size; debug_count++) + printchar (pattern[debug_count]); + putchar ('\n'); + } +#endif /* DEBUG */ + + /* Initialize the compile stack. */ + compile_stack.stack = TALLOC (INIT_COMPILE_STACK_SIZE, compile_stack_elt_t); + if (compile_stack.stack == NULL) + return REG_ESPACE; + + compile_stack.size = INIT_COMPILE_STACK_SIZE; + compile_stack.avail = 0; + + /* Initialize the pattern buffer. */ + bufp->syntax = syntax; + bufp->fastmap_accurate = 0; + bufp->not_bol = bufp->not_eol = 0; + + /* Set `used' to zero, so that if we return an error, the pattern + printer (for debugging) will think there's no pattern. We reset it + at the end. */ + bufp->used = 0; + + /* Always count groups, whether or not bufp->no_sub is set. */ + bufp->re_nsub = 0; + +#if !defined (emacs) && !defined (SYNTAX_TABLE) + /* Initialize the syntax table. */ + init_syntax_once (); +#endif + + if (bufp->allocated == 0) + { + if (bufp->buffer) + { /* If zero allocated, but buffer is non-null, try to realloc + enough space. This loses if buffer's address is bogus, but + that is the user's responsibility. */ + RETALLOC (bufp->buffer, INIT_BUF_SIZE, unsigned char); + } + else + { /* Caller did not allocate a buffer. Do it for them. */ + bufp->buffer = TALLOC (INIT_BUF_SIZE, unsigned char); + } + if (!bufp->buffer) return REG_ESPACE; + + bufp->allocated = INIT_BUF_SIZE; + } + + begalt = b = bufp->buffer; + + /* Loop through the uncompiled pattern until we're at the end. */ + while (p != pend) + { + PATFETCH (c); + + switch (c) + { + case '^': + { + if ( /* If at start of pattern, it's an operator. */ + p == pattern + 1 + /* If context independent, it's an operator. */ + || syntax & RE_CONTEXT_INDEP_ANCHORS + /* Otherwise, depends on what's come before. */ + || at_begline_loc_p (pattern, p, syntax)) + BUF_PUSH (begline); + else + goto normal_char; + } + break; + + + case '$': + { + if ( /* If at end of pattern, it's an operator. */ + p == pend + /* If context independent, it's an operator. */ + || syntax & RE_CONTEXT_INDEP_ANCHORS + /* Otherwise, depends on what's next. */ + || at_endline_loc_p (p, pend, syntax)) + BUF_PUSH (endline); + else + goto normal_char; + } + break; + + + case '+': + case '?': + if ((syntax & RE_BK_PLUS_QM) + || (syntax & RE_LIMITED_OPS)) + goto normal_char; + handle_plus: + case '*': + /* If there is no previous pattern... */ + if (!laststart) + { + if (syntax & RE_CONTEXT_INVALID_OPS) + return REG_BADRPT; + else if (!(syntax & RE_CONTEXT_INDEP_OPS)) + goto normal_char; + } + + { + /* Are we optimizing this jump? */ + boolean keep_string_p = false; + + /* 1 means zero (many) matches is allowed. */ + char zero_times_ok = 0, many_times_ok = 0; + + /* If there is a sequence of repetition chars, collapse it + down to just one (the right one). We can't combine + interval operators with these because of, e.g., `a{2}*', + which should only match an even number of `a's. */ + + for (;;) + { + zero_times_ok |= c != '+'; + many_times_ok |= c != '?'; + + if (p == pend) + break; + + PATFETCH (c); + + if (c == '*' + || (!(syntax & RE_BK_PLUS_QM) && (c == '+' || c == '?'))) + ; + + else if (syntax & RE_BK_PLUS_QM && c == '\\') + { + if (p == pend) return REG_EESCAPE; + + PATFETCH (c1); + if (!(c1 == '+' || c1 == '?')) + { + PATUNFETCH; + PATUNFETCH; + break; + } + + c = c1; + } + else + { + PATUNFETCH; + break; + } + + /* If we get here, we found another repeat character. */ + } + + /* Star, etc. applied to an empty pattern is equivalent + to an empty pattern. */ + if (!laststart) + break; + + /* Now we know whether or not zero matches is allowed + and also whether or not two or more matches is allowed. */ + if (many_times_ok) + { /* More than one repetition is allowed, so put in at the + end a backward relative jump from `b' to before the next + jump we're going to put in below (which jumps from + laststart to after this jump). + + But if we are at the `*' in the exact sequence `.*\n', + insert an unconditional jump backwards to the ., + instead of the beginning of the loop. This way we only + push a failure point once, instead of every time + through the loop. */ + assert (p - 1 > pattern); + + /* Allocate the space for the jump. */ + GET_BUFFER_SPACE (3); + + /* We know we are not at the first character of the pattern, + because laststart was nonzero. And we've already + incremented `p', by the way, to be the character after + the `*'. Do we have to do something analogous here + for null bytes, because of RE_DOT_NOT_NULL? */ + if (TRANSLATE (*(p - 2)) == TRANSLATE ('.') + && p < pend && TRANSLATE (*p) == TRANSLATE ('\n') + && !(syntax & RE_DOT_NEWLINE)) + { /* We have .*\n. */ + STORE_JUMP (jump, b, laststart); + keep_string_p = true; + } + else + /* Anything else. */ + STORE_JUMP (maybe_pop_jump, b, laststart - 3); + + /* We've added more stuff to the buffer. */ + b += 3; + } + + /* On failure, jump from laststart to b + 3, which will be the + end of the buffer after this jump is inserted. */ + GET_BUFFER_SPACE (3); + INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump + : on_failure_jump, + laststart, b + 3); + pending_exact = 0; + b += 3; + + if (!zero_times_ok) + { + /* At least one repetition is required, so insert a + `dummy_failure_jump' before the initial + `on_failure_jump' instruction of the loop. This + effects a skip over that instruction the first time + we hit that loop. */ + GET_BUFFER_SPACE (3); + INSERT_JUMP (dummy_failure_jump, laststart, laststart + 6); + b += 3; + } + } + break; + + + case '.': + laststart = b; + BUF_PUSH (anychar); + break; + + + case '[': + { + boolean had_char_class = false; + + if (p == pend) return REG_EBRACK; + + /* Ensure that we have enough space to push a charset: the + opcode, the length count, and the bitset; 34 bytes in all. */ + GET_BUFFER_SPACE (34); + + laststart = b; + + /* We test `*p == '^' twice, instead of using an if + statement, so we only need one BUF_PUSH. */ + BUF_PUSH (*p == '^' ? charset_not : charset); + if (*p == '^') + p++; + + /* Remember the first position in the bracket expression. */ + p1 = p; + + /* Push the number of bytes in the bitmap. */ + BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH); + + /* Clear the whole map. */ + bzero (b, (1 << BYTEWIDTH) / BYTEWIDTH); + + /* charset_not matches newline according to a syntax bit. */ + if ((re_opcode_t) b[-2] == charset_not + && (syntax & RE_HAT_LISTS_NOT_NEWLINE)) + SET_LIST_BIT ('\n'); + + /* Read in characters and ranges, setting map bits. */ + for (;;) + { + if (p == pend) return REG_EBRACK; + + PATFETCH (c); + + /* \ might escape characters inside [...] and [^...]. */ + if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\') + { + if (p == pend) return REG_EESCAPE; + + PATFETCH (c1); + SET_LIST_BIT (c1); + continue; + } + + /* Could be the end of the bracket expression. If it's + not (i.e., when the bracket expression is `[]' so + far), the ']' character bit gets set way below. */ + if (c == ']' && p != p1 + 1) + break; + + /* Look ahead to see if it's a range when the last thing + was a character class. */ + if (had_char_class && c == '-' && *p != ']') + return REG_ERANGE; + + /* Look ahead to see if it's a range when the last thing + was a character: if this is a hyphen not at the + beginning or the end of a list, then it's the range + operator. */ + if (c == '-' + && !(p - 2 >= pattern && p[-2] == '[') + && !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^') + && *p != ']') + { + reg_errcode_t ret + = compile_range (&p, pend, translate, syntax, b); + if (ret != REG_NOERROR) return ret; + } + + else if (p[0] == '-' && p[1] != ']') + { /* This handles ranges made up of characters only. */ + reg_errcode_t ret; + + /* Move past the `-'. */ + PATFETCH (c1); + + ret = compile_range (&p, pend, translate, syntax, b); + if (ret != REG_NOERROR) return ret; + } + + /* See if we're at the beginning of a possible character + class. */ + + else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == ':') + { /* Leave room for the null. */ + char str[CHAR_CLASS_MAX_LENGTH + 1]; + + PATFETCH (c); + c1 = 0; + + /* If pattern is `[[:'. */ + if (p == pend) return REG_EBRACK; + + for (;;) + { + PATFETCH (c); + if (c == ':' || c == ']' || p == pend + || c1 == CHAR_CLASS_MAX_LENGTH) + break; + str[c1++] = c; + } + str[c1] = '\0'; + + /* If isn't a word bracketed by `[:' and:`]': + undo the ending character, the letters, and leave + the leading `:' and `[' (but set bits for them). */ + if (c == ':' && *p == ']') + { + int ch; + boolean is_alnum = STREQ (str, "alnum"); + boolean is_alpha = STREQ (str, "alpha"); + boolean is_blank = STREQ (str, "blank"); + boolean is_cntrl = STREQ (str, "cntrl"); + boolean is_digit = STREQ (str, "digit"); + boolean is_graph = STREQ (str, "graph"); + boolean is_lower = STREQ (str, "lower"); + boolean is_print = STREQ (str, "print"); + boolean is_punct = STREQ (str, "punct"); + boolean is_space = STREQ (str, "space"); + boolean is_upper = STREQ (str, "upper"); + boolean is_xdigit = STREQ (str, "xdigit"); + + if (!IS_CHAR_CLASS (str)) return REG_ECTYPE; + + /* Throw away the ] at the end of the character + class. */ + PATFETCH (c); + + if (p == pend) return REG_EBRACK; + + for (ch = 0; ch < 1 << BYTEWIDTH; ch++) + { + if ( (is_alnum && ISALNUM (ch)) + || (is_alpha && ISALPHA (ch)) + || (is_blank && ISBLANK (ch)) + || (is_cntrl && ISCNTRL (ch)) + || (is_digit && ISDIGIT (ch)) + || (is_graph && ISGRAPH (ch)) + || (is_lower && ISLOWER (ch)) + || (is_print && ISPRINT (ch)) + || (is_punct && ISPUNCT (ch)) + || (is_space && ISSPACE (ch)) + || (is_upper && ISUPPER (ch)) + || (is_xdigit && ISXDIGIT (ch))) + SET_LIST_BIT (ch); + } + had_char_class = true; + } + else + { + c1++; + while (c1--) + PATUNFETCH; + SET_LIST_BIT ('['); + SET_LIST_BIT (':'); + had_char_class = false; + } + } + else + { + had_char_class = false; + SET_LIST_BIT (c); + } + } + + /* Discard any (non)matching list bytes that are all 0 at the + end of the map. Decrease the map-length byte too. */ + while ((int) b[-1] > 0 && b[b[-1] - 1] == 0) + b[-1]--; + b += b[-1]; + } + break; + + + case '(': + if (syntax & RE_NO_BK_PARENS) + goto handle_open; + else + goto normal_char; + + + case ')': + if (syntax & RE_NO_BK_PARENS) + goto handle_close; + else + goto normal_char; + + + case '\n': + if (syntax & RE_NEWLINE_ALT) + goto handle_alt; + else + goto normal_char; + + + case '|': + if (syntax & RE_NO_BK_VBAR) + goto handle_alt; + else + goto normal_char; + + + case '{': + if (syntax & RE_INTERVALS && syntax & RE_NO_BK_BRACES) + goto handle_interval; + else + goto normal_char; + + + case '\\': + if (p == pend) return REG_EESCAPE; + + /* Do not translate the character after the \, so that we can + distinguish, e.g., \B from \b, even if we normally would + translate, e.g., B to b. */ + PATFETCH_RAW (c); + + switch (c) + { + case '(': + if (syntax & RE_NO_BK_PARENS) + goto normal_backslash; + + handle_open: + bufp->re_nsub++; + regnum++; + + if (COMPILE_STACK_FULL) + { + RETALLOC (compile_stack.stack, compile_stack.size << 1, + compile_stack_elt_t); + if (compile_stack.stack == NULL) return REG_ESPACE; + + compile_stack.size <<= 1; + } + + /* These are the values to restore when we hit end of this + group. They are all relative offsets, so that if the + whole pattern moves because of realloc, they will still + be valid. */ + COMPILE_STACK_TOP.begalt_offset = begalt - bufp->buffer; + COMPILE_STACK_TOP.fixup_alt_jump + = fixup_alt_jump ? fixup_alt_jump - bufp->buffer + 1 : 0; + COMPILE_STACK_TOP.laststart_offset = b - bufp->buffer; + COMPILE_STACK_TOP.regnum = regnum; + + /* We will eventually replace the 0 with the number of + groups inner to this one. But do not push a + start_memory for groups beyond the last one we can + represent in the compiled pattern. */ + if (regnum <= MAX_REGNUM) + { + COMPILE_STACK_TOP.inner_group_offset = b - bufp->buffer + 2; + BUF_PUSH_3 (start_memory, regnum, 0); + } + + compile_stack.avail++; + + fixup_alt_jump = 0; + laststart = 0; + begalt = b; + break; + + + case ')': + if (syntax & RE_NO_BK_PARENS) goto normal_backslash; + + if (COMPILE_STACK_EMPTY) + if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD) + goto normal_backslash; + else + return REG_ERPAREN; + + handle_close: + if (fixup_alt_jump) + { /* Push a dummy failure point at the end of the + alternative for a possible future + `pop_failure_jump' to pop. See comments at + `push_dummy_failure' in `re_match_2'. */ + BUF_PUSH (push_dummy_failure); + + /* We allocated space for this jump when we assigned + to `fixup_alt_jump', in the `handle_alt' case below. */ + STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1); + } + + /* See similar code for backslashed left paren above. */ + if (COMPILE_STACK_EMPTY) + if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD) + goto normal_char; + else + return REG_ERPAREN; + + /* Since we just checked for an empty stack above, this + ``can't happen''. */ + assert (compile_stack.avail != 0); + { + /* We don't just want to restore into `regnum', because + later groups should continue to be numbered higher, + as in `(ab)c(de)' -- the second group is #2. */ + regnum_t this_group_regnum; + + compile_stack.avail--; + begalt = bufp->buffer + COMPILE_STACK_TOP.begalt_offset; + fixup_alt_jump + = COMPILE_STACK_TOP.fixup_alt_jump + ? bufp->buffer + COMPILE_STACK_TOP.fixup_alt_jump - 1 + : 0; + laststart = bufp->buffer + COMPILE_STACK_TOP.laststart_offset; + this_group_regnum = COMPILE_STACK_TOP.regnum; + + /* We're at the end of the group, so now we know how many + groups were inside this one. */ + if (this_group_regnum <= MAX_REGNUM) + { + unsigned char *inner_group_loc + = bufp->buffer + COMPILE_STACK_TOP.inner_group_offset; + + *inner_group_loc = regnum - this_group_regnum; + BUF_PUSH_3 (stop_memory, this_group_regnum, + regnum - this_group_regnum); + } + } + break; + + + case '|': /* `\|'. */ + if (syntax & RE_LIMITED_OPS || syntax & RE_NO_BK_VBAR) + goto normal_backslash; + handle_alt: + if (syntax & RE_LIMITED_OPS) + goto normal_char; + + /* Insert before the previous alternative a jump which + jumps to this alternative if the former fails. */ + GET_BUFFER_SPACE (3); + INSERT_JUMP (on_failure_jump, begalt, b + 6); + pending_exact = 0; + b += 3; + + /* The alternative before this one has a jump after it + which gets executed if it gets matched. Adjust that + jump so it will jump to this alternative's analogous + jump (put in below, which in turn will jump to the next + (if any) alternative's such jump, etc.). The last such + jump jumps to the correct final destination. A picture: + _____ _____ + | | | | + | v | v + a | b | c + + If we are at `b', then fixup_alt_jump right now points to a + three-byte space after `a'. We'll put in the jump, set + fixup_alt_jump to right after `b', and leave behind three + bytes which we'll fill in when we get to after `c'. */ + + if (fixup_alt_jump) + STORE_JUMP (jump_past_alt, fixup_alt_jump, b); + + /* Mark and leave space for a jump after this alternative, + to be filled in later either by next alternative or + when know we're at the end of a series of alternatives. */ + fixup_alt_jump = b; + GET_BUFFER_SPACE (3); + b += 3; + + laststart = 0; + begalt = b; + break; + + + case '{': + /* If \{ is a literal. */ + if (!(syntax & RE_INTERVALS) + /* If we're at `\{' and it's not the open-interval + operator. */ + || ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) + || (p - 2 == pattern && p == pend)) + goto normal_backslash; + + handle_interval: + { + /* If got here, then the syntax allows intervals. */ + + /* At least (most) this many matches must be made. */ + int lower_bound = -1, upper_bound = -1; + + beg_interval = p - 1; + + if (p == pend) + { + if (syntax & RE_NO_BK_BRACES) + goto unfetch_interval; + else + return REG_EBRACE; + } + + GET_UNSIGNED_NUMBER (lower_bound); + + if (c == ',') + { + GET_UNSIGNED_NUMBER (upper_bound); + if (upper_bound < 0) upper_bound = RE_DUP_MAX; + } + else + /* Interval such as `{1}' => match exactly once. */ + upper_bound = lower_bound; + + if (lower_bound < 0 || upper_bound > RE_DUP_MAX + || lower_bound > upper_bound) + { + if (syntax & RE_NO_BK_BRACES) + goto unfetch_interval; + else + return REG_BADBR; + } + + if (!(syntax & RE_NO_BK_BRACES)) + { + if (c != '\\') return REG_EBRACE; + + PATFETCH (c); + } + + if (c != '}') + { + if (syntax & RE_NO_BK_BRACES) + goto unfetch_interval; + else + return REG_BADBR; + } + + /* We just parsed a valid interval. */ + + /* If it's invalid to have no preceding re. */ + if (!laststart) + { + if (syntax & RE_CONTEXT_INVALID_OPS) + return REG_BADRPT; + else if (syntax & RE_CONTEXT_INDEP_OPS) + laststart = b; + else + goto unfetch_interval; + } + + /* If the upper bound is zero, don't want to succeed at + all; jump from `laststart' to `b + 3', which will be + the end of the buffer after we insert the jump. */ + if (upper_bound == 0) + { + GET_BUFFER_SPACE (3); + INSERT_JUMP (jump, laststart, b + 3); + b += 3; + } + + /* Otherwise, we have a nontrivial interval. When + we're all done, the pattern will look like: + set_number_at <jump count> <upper bound> + set_number_at <succeed_n count> <lower bound> + succeed_n <after jump addr> <succed_n count> + <body of loop> + jump_n <succeed_n addr> <jump count> + (The upper bound and `jump_n' are omitted if + `upper_bound' is 1, though.) */ + else + { /* If the upper bound is > 1, we need to insert + more at the end of the loop. */ + unsigned nbytes = 10 + (upper_bound > 1) * 10; + + GET_BUFFER_SPACE (nbytes); + + /* Initialize lower bound of the `succeed_n', even + though it will be set during matching by its + attendant `set_number_at' (inserted next), + because `re_compile_fastmap' needs to know. + Jump to the `jump_n' we might insert below. */ + INSERT_JUMP2 (succeed_n, laststart, + b + 5 + (upper_bound > 1) * 5, + lower_bound); + b += 5; + + /* Code to initialize the lower bound. Insert + before the `succeed_n'. The `5' is the last two + bytes of this `set_number_at', plus 3 bytes of + the following `succeed_n'. */ + insert_op2 (set_number_at, laststart, 5, lower_bound, b); + b += 5; + + if (upper_bound > 1) + { /* More than one repetition is allowed, so + append a backward jump to the `succeed_n' + that starts this interval. + + When we've reached this during matching, + we'll have matched the interval once, so + jump back only `upper_bound - 1' times. */ + STORE_JUMP2 (jump_n, b, laststart + 5, + upper_bound - 1); + b += 5; + + /* The location we want to set is the second + parameter of the `jump_n'; that is `b-2' as + an absolute address. `laststart' will be + the `set_number_at' we're about to insert; + `laststart+3' the number to set, the source + for the relative address. But we are + inserting into the middle of the pattern -- + so everything is getting moved up by 5. + Conclusion: (b - 2) - (laststart + 3) + 5, + i.e., b - laststart. + + We insert this at the beginning of the loop + so that if we fail during matching, we'll + reinitialize the bounds. */ + insert_op2 (set_number_at, laststart, b - laststart, + upper_bound - 1, b); + b += 5; + } + } + pending_exact = 0; + beg_interval = NULL; + } + break; + + unfetch_interval: + /* If an invalid interval, match the characters as literals. */ + assert (beg_interval); + p = beg_interval; + beg_interval = NULL; + + /* normal_char and normal_backslash need `c'. */ + PATFETCH (c); + + if (!(syntax & RE_NO_BK_BRACES)) + { + if (p > pattern && p[-1] == '\\') + goto normal_backslash; + } + goto normal_char; + +#ifdef emacs + /* There is no way to specify the before_dot and after_dot + operators. rms says this is ok. --karl */ + case '=': + BUF_PUSH (at_dot); + break; + + case 's': + laststart = b; + PATFETCH (c); + BUF_PUSH_2 (syntaxspec, syntax_spec_code[c]); + break; + + case 'S': + laststart = b; + PATFETCH (c); + BUF_PUSH_2 (notsyntaxspec, syntax_spec_code[c]); + break; +#endif /* emacs */ + + + case 'w': + laststart = b; + BUF_PUSH (wordchar); + break; + + + case 'W': + laststart = b; + BUF_PUSH (notwordchar); + break; + + + case '<': + BUF_PUSH (wordbeg); + break; + + case '>': + BUF_PUSH (wordend); + break; + + case 'b': + BUF_PUSH (wordbound); + break; + + case 'B': + BUF_PUSH (notwordbound); + break; + + case '`': + BUF_PUSH (begbuf); + break; + + case '\'': + BUF_PUSH (endbuf); + break; + + case '1': case '2': case '3': case '4': case '5': + case '6': case '7': case '8': case '9': + if (syntax & RE_NO_BK_REFS) + goto normal_char; + + c1 = c - '0'; + + if (c1 > regnum) + return REG_ESUBREG; + + /* Can't back reference to a subexpression if inside of it. */ + if (group_in_compile_stack (compile_stack, c1)) + goto normal_char; + + laststart = b; + BUF_PUSH_2 (duplicate, c1); + break; + + + case '+': + case '?': + if (syntax & RE_BK_PLUS_QM) + goto handle_plus; + else + goto normal_backslash; + + default: + normal_backslash: + /* You might think it would be useful for \ to mean + not to translate; but if we don't translate it + it will never match anything. */ + c = TRANSLATE (c); + goto normal_char; + } + break; + + + default: + /* Expects the character in `c'. */ + normal_char: + /* If no exactn currently being built. */ + if (!pending_exact + + /* If last exactn not at current position. */ + || pending_exact + *pending_exact + 1 != b + + /* We have only one byte following the exactn for the count. */ + || *pending_exact == (1 << BYTEWIDTH) - 1 + + /* If followed by a repetition operator. */ + || *p == '*' || *p == '^' + || ((syntax & RE_BK_PLUS_QM) + ? *p == '\\' && (p[1] == '+' || p[1] == '?') + : (*p == '+' || *p == '?')) + || ((syntax & RE_INTERVALS) + && ((syntax & RE_NO_BK_BRACES) + ? *p == '{' + : (p[0] == '\\' && p[1] == '{')))) + { + /* Start building a new exactn. */ + + laststart = b; + + BUF_PUSH_2 (exactn, 0); + pending_exact = b - 1; + } + + BUF_PUSH (c); + (*pending_exact)++; + break; + } /* switch (c) */ + } /* while p != pend */ + + + /* Through the pattern now. */ + + if (fixup_alt_jump) + STORE_JUMP (jump_past_alt, fixup_alt_jump, b); + + if (!COMPILE_STACK_EMPTY) + return REG_EPAREN; + + free (compile_stack.stack); + + /* We have succeeded; set the length of the buffer. */ + bufp->used = b - bufp->buffer; + +#ifdef DEBUG + if (debug) + { + DEBUG_PRINT1 ("\nCompiled pattern: "); + print_compiled_pattern (bufp); + } +#endif /* DEBUG */ + + return REG_NOERROR; +} /* regex_compile */ + +/* Subroutines for `regex_compile'. */ + +/* Store OP at LOC followed by two-byte integer parameter ARG. */ + +static void +store_op1 (op, loc, arg) + re_opcode_t op; + unsigned char *loc; + int arg; +{ + *loc = (unsigned char) op; + STORE_NUMBER (loc + 1, arg); +} + + +/* Like `store_op1', but for two two-byte parameters ARG1 and ARG2. */ + +static void +store_op2 (op, loc, arg1, arg2) + re_opcode_t op; + unsigned char *loc; + int arg1, arg2; +{ + *loc = (unsigned char) op; + STORE_NUMBER (loc + 1, arg1); + STORE_NUMBER (loc + 3, arg2); +} + + +/* Copy the bytes from LOC to END to open up three bytes of space at LOC + for OP followed by two-byte integer parameter ARG. */ + +static void +insert_op1 (op, loc, arg, end) + re_opcode_t op; + unsigned char *loc; + int arg; + unsigned char *end; +{ + register unsigned char *pfrom = end; + register unsigned char *pto = end + 3; + + while (pfrom != loc) + *--pto = *--pfrom; + + store_op1 (op, loc, arg); +} + + +/* Like `insert_op1', but for two two-byte parameters ARG1 and ARG2. */ + +static void +insert_op2 (op, loc, arg1, arg2, end) + re_opcode_t op; + unsigned char *loc; + int arg1, arg2; + unsigned char *end; +{ + register unsigned char *pfrom = end; + register unsigned char *pto = end + 5; + + while (pfrom != loc) + *--pto = *--pfrom; + + store_op2 (op, loc, arg1, arg2); +} + + +/* P points to just after a ^ in PATTERN. Return true if that ^ comes + after an alternative or a begin-subexpression. We assume there is at + least one character before the ^. */ + +static boolean +at_begline_loc_p (pattern, p, syntax) + const char *pattern, *p; + reg_syntax_t syntax; +{ + const char *prev = p - 2; + boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\'; + + return + /* After a subexpression? */ + (*prev == '(' && (syntax & RE_NO_BK_PARENS || prev_prev_backslash)) + /* After an alternative? */ + || (*prev == '|' && (syntax & RE_NO_BK_VBAR || prev_prev_backslash)); +} + + +/* The dual of at_begline_loc_p. This one is for $. We assume there is + at least one character after the $, i.e., `P < PEND'. */ + +static boolean +at_endline_loc_p (p, pend, syntax) + const char *p, *pend; + int syntax; +{ + const char *next = p; + boolean next_backslash = *next == '\\'; + const char *next_next = p + 1 < pend ? p + 1 : NULL; + + return + /* Before a subexpression? */ + (syntax & RE_NO_BK_PARENS ? *next == ')' + : next_backslash && next_next && *next_next == ')') + /* Before an alternative? */ + || (syntax & RE_NO_BK_VBAR ? *next == '|' + : next_backslash && next_next && *next_next == '|'); +} + + +/* Returns true if REGNUM is in one of COMPILE_STACK's elements and + false if it's not. */ + +static boolean +group_in_compile_stack (compile_stack, regnum) + compile_stack_type compile_stack; + regnum_t regnum; +{ + int this_element; + + for (this_element = compile_stack.avail - 1; + this_element >= 0; + this_element--) + if (compile_stack.stack[this_element].regnum == regnum) + return true; + + return false; +} + + +/* Read the ending character of a range (in a bracket expression) from the + uncompiled pattern *P_PTR (which ends at PEND). We assume the + starting character is in `P[-2]'. (`P[-1]' is the character `-'.) + Then we set the translation of all bits between the starting and + ending characters (inclusive) in the compiled pattern B. + + Return an error code. + + We use these short variable names so we can use the same macros as + `regex_compile' itself. */ + +static reg_errcode_t +compile_range (p_ptr, pend, translate, syntax, b) + const char **p_ptr, *pend; + char *translate; + reg_syntax_t syntax; + unsigned char *b; +{ + unsigned this_char; + + const char *p = *p_ptr; + int range_start, range_end; + + if (p == pend) + return REG_ERANGE; + + /* Even though the pattern is a signed `char *', we need to fetch + with unsigned char *'s; if the high bit of the pattern character + is set, the range endpoints will be negative if we fetch using a + signed char *. + + We also want to fetch the endpoints without translating them; the + appropriate translation is done in the bit-setting loop below. */ + range_start = ((unsigned char *) p)[-2]; + range_end = ((unsigned char *) p)[0]; + + /* Have to increment the pointer into the pattern string, so the + caller isn't still at the ending character. */ + (*p_ptr)++; + + /* If the start is after the end, the range is empty. */ + if (range_start > range_end) + return syntax & RE_NO_EMPTY_RANGES ? REG_ERANGE : REG_NOERROR; + + /* Here we see why `this_char' has to be larger than an `unsigned + char' -- the range is inclusive, so if `range_end' == 0xff + (assuming 8-bit characters), we would otherwise go into an infinite + loop, since all characters <= 0xff. */ + for (this_char = range_start; this_char <= range_end; this_char++) + { + SET_LIST_BIT (TRANSLATE (this_char)); + } + + return REG_NOERROR; +} + +/* Failure stack declarations and macros; both re_compile_fastmap and + re_match_2 use a failure stack. These have to be macros because of + REGEX_ALLOCATE. */ + + +/* Number of failure points for which to initially allocate space + when matching. If this number is exceeded, we allocate more + space, so it is not a hard limit. */ +#ifndef INIT_FAILURE_ALLOC +#define INIT_FAILURE_ALLOC 5 +#endif + +/* Roughly the maximum number of failure points on the stack. Would be + exactly that if always used MAX_FAILURE_SPACE each time we failed. + This is a variable only so users of regex can assign to it; we never + change it ourselves. */ +int re_max_failures = 2000; + +typedef const unsigned char *fail_stack_elt_t; + +typedef struct +{ + fail_stack_elt_t *stack; + unsigned size; + unsigned avail; /* Offset of next open position. */ +} fail_stack_type; + +#define FAIL_STACK_EMPTY() (fail_stack.avail == 0) +#define FAIL_STACK_PTR_EMPTY() (fail_stack_ptr->avail == 0) +#define FAIL_STACK_FULL() (fail_stack.avail == fail_stack.size) +#define FAIL_STACK_TOP() (fail_stack.stack[fail_stack.avail]) + + +/* Initialize `fail_stack'. Do `return -2' if the alloc fails. */ + +#define INIT_FAIL_STACK() \ + do { \ + fail_stack.stack = (fail_stack_elt_t *) \ + REGEX_ALLOCATE (INIT_FAILURE_ALLOC * sizeof (fail_stack_elt_t)); \ + \ + if (fail_stack.stack == NULL) \ + return -2; \ + \ + fail_stack.size = INIT_FAILURE_ALLOC; \ + fail_stack.avail = 0; \ + } while (0) + + +/* Double the size of FAIL_STACK, up to approximately `re_max_failures' items. + + Return 1 if succeeds, and 0 if either ran out of memory + allocating space for it or it was already too large. + + REGEX_REALLOCATE requires `destination' be declared. */ + +#define DOUBLE_FAIL_STACK(fail_stack) \ + ((fail_stack).size > re_max_failures * MAX_FAILURE_ITEMS \ + ? 0 \ + : ((fail_stack).stack = (fail_stack_elt_t *) \ + REGEX_REALLOCATE ((fail_stack).stack, \ + (fail_stack).size * sizeof (fail_stack_elt_t), \ + ((fail_stack).size << 1) * sizeof (fail_stack_elt_t)), \ + \ + (fail_stack).stack == NULL \ + ? 0 \ + : ((fail_stack).size <<= 1, \ + 1))) + + +/* Push PATTERN_OP on FAIL_STACK. + + Return 1 if was able to do so and 0 if ran out of memory allocating + space to do so. */ +#define PUSH_PATTERN_OP(pattern_op, fail_stack) \ + ((FAIL_STACK_FULL () \ + && !DOUBLE_FAIL_STACK (fail_stack)) \ + ? 0 \ + : ((fail_stack).stack[(fail_stack).avail++] = pattern_op, \ + 1)) + +/* This pushes an item onto the failure stack. Must be a four-byte + value. Assumes the variable `fail_stack'. Probably should only + be called from within `PUSH_FAILURE_POINT'. */ +#define PUSH_FAILURE_ITEM(item) \ + fail_stack.stack[fail_stack.avail++] = (fail_stack_elt_t) item + +/* The complement operation. Assumes `fail_stack' is nonempty. */ +#define POP_FAILURE_ITEM() fail_stack.stack[--fail_stack.avail] + +/* Used to omit pushing failure point id's when we're not debugging. */ +#ifdef DEBUG +#define DEBUG_PUSH PUSH_FAILURE_ITEM +#define DEBUG_POP(item_addr) *(item_addr) = POP_FAILURE_ITEM () +#else +#define DEBUG_PUSH(item) +#define DEBUG_POP(item_addr) +#endif + + +/* Push the information about the state we will need + if we ever fail back to it. + + Requires variables fail_stack, regstart, regend, reg_info, and + num_regs be declared. DOUBLE_FAIL_STACK requires `destination' be + declared. + + Does `return FAILURE_CODE' if runs out of memory. */ + +#define PUSH_FAILURE_POINT(pattern_place, string_place, failure_code) \ + do { \ + char *destination; \ + /* Must be int, so when we don't save any registers, the arithmetic \ + of 0 + -1 isn't done as unsigned. */ \ + int this_reg; \ + \ + DEBUG_STATEMENT (failure_id++); \ + DEBUG_STATEMENT (nfailure_points_pushed++); \ + DEBUG_PRINT2 ("\nPUSH_FAILURE_POINT #%u:\n", failure_id); \ + DEBUG_PRINT2 (" Before push, next avail: %d\n", (fail_stack).avail);\ + DEBUG_PRINT2 (" size: %d\n", (fail_stack).size);\ + \ + DEBUG_PRINT2 (" slots needed: %d\n", NUM_FAILURE_ITEMS); \ + DEBUG_PRINT2 (" available: %d\n", REMAINING_AVAIL_SLOTS); \ + \ + /* Ensure we have enough space allocated for what we will push. */ \ + while (REMAINING_AVAIL_SLOTS < NUM_FAILURE_ITEMS) \ + { \ + if (!DOUBLE_FAIL_STACK (fail_stack)) \ + return failure_code; \ + \ + DEBUG_PRINT2 ("\n Doubled stack; size now: %d\n", \ + (fail_stack).size); \ + DEBUG_PRINT2 (" slots available: %d\n", REMAINING_AVAIL_SLOTS);\ + } \ + \ + /* Push the info, starting with the registers. */ \ + DEBUG_PRINT1 ("\n"); \ + \ + for (this_reg = lowest_active_reg; this_reg <= highest_active_reg; \ + this_reg++) \ + { \ + DEBUG_PRINT2 (" Pushing reg: %d\n", this_reg); \ + DEBUG_STATEMENT (num_regs_pushed++); \ + \ + DEBUG_PRINT2 (" start: 0x%x\n", regstart[this_reg]); \ + PUSH_FAILURE_ITEM (regstart[this_reg]); \ + \ + DEBUG_PRINT2 (" end: 0x%x\n", regend[this_reg]); \ + PUSH_FAILURE_ITEM (regend[this_reg]); \ + \ + DEBUG_PRINT2 (" info: 0x%x\n ", reg_info[this_reg]); \ + DEBUG_PRINT2 (" match_null=%d", \ + REG_MATCH_NULL_STRING_P (reg_info[this_reg])); \ + DEBUG_PRINT2 (" active=%d", IS_ACTIVE (reg_info[this_reg])); \ + DEBUG_PRINT2 (" matched_something=%d", \ + MATCHED_SOMETHING (reg_info[this_reg])); \ + DEBUG_PRINT2 (" ever_matched=%d", \ + EVER_MATCHED_SOMETHING (reg_info[this_reg])); \ + DEBUG_PRINT1 ("\n"); \ + PUSH_FAILURE_ITEM (reg_info[this_reg].word); \ + } \ + \ + DEBUG_PRINT2 (" Pushing low active reg: %d\n", lowest_active_reg);\ + PUSH_FAILURE_ITEM (lowest_active_reg); \ + \ + DEBUG_PRINT2 (" Pushing high active reg: %d\n", highest_active_reg);\ + PUSH_FAILURE_ITEM (highest_active_reg); \ + \ + DEBUG_PRINT2 (" Pushing pattern 0x%x: ", pattern_place); \ + DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \ + PUSH_FAILURE_ITEM (pattern_place); \ + \ + DEBUG_PRINT2 (" Pushing string 0x%x: `", string_place); \ + DEBUG_PRINT_DOUBLE_STRING (string_place, string1, size1, string2, \ + size2); \ + DEBUG_PRINT1 ("'\n"); \ + PUSH_FAILURE_ITEM (string_place); \ + \ + DEBUG_PRINT2 (" Pushing failure id: %u\n", failure_id); \ + DEBUG_PUSH (failure_id); \ + } while (0) + +/* This is the number of items that are pushed and popped on the stack + for each register. */ +#define NUM_REG_ITEMS 3 + +/* Individual items aside from the registers. */ +#ifdef DEBUG +#define NUM_NONREG_ITEMS 5 /* Includes failure point id. */ +#else +#define NUM_NONREG_ITEMS 4 +#endif + +/* We push at most this many items on the stack. */ +#define MAX_FAILURE_ITEMS ((num_regs - 1) * NUM_REG_ITEMS + NUM_NONREG_ITEMS) + +/* We actually push this many items. */ +#define NUM_FAILURE_ITEMS \ + ((highest_active_reg - lowest_active_reg + 1) * NUM_REG_ITEMS \ + + NUM_NONREG_ITEMS) + +/* How many items can still be added to the stack without overflowing it. */ +#define REMAINING_AVAIL_SLOTS ((fail_stack).size - (fail_stack).avail) + + +/* Pops what PUSH_FAIL_STACK pushes. + + We restore into the parameters, all of which should be lvalues: + STR -- the saved data position. + PAT -- the saved pattern position. + LOW_REG, HIGH_REG -- the highest and lowest active registers. + REGSTART, REGEND -- arrays of string positions. + REG_INFO -- array of information about each subexpression. + + Also assumes the variables `fail_stack' and (if debugging), `bufp', + `pend', `string1', `size1', `string2', and `size2'. */ + +#define POP_FAILURE_POINT(str, pat, low_reg, high_reg, regstart, regend, reg_info)\ +{ \ + DEBUG_STATEMENT (fail_stack_elt_t failure_id;) \ + int this_reg; \ + const unsigned char *string_temp; \ + \ + assert (!FAIL_STACK_EMPTY ()); \ + \ + /* Remove failure points and point to how many regs pushed. */ \ + DEBUG_PRINT1 ("POP_FAILURE_POINT:\n"); \ + DEBUG_PRINT2 (" Before pop, next avail: %d\n", fail_stack.avail); \ + DEBUG_PRINT2 (" size: %d\n", fail_stack.size); \ + \ + assert (fail_stack.avail >= NUM_NONREG_ITEMS); \ + \ + DEBUG_POP (&failure_id); \ + DEBUG_PRINT2 (" Popping failure id: %u\n", failure_id); \ + \ + /* If the saved string location is NULL, it came from an \ + on_failure_keep_string_jump opcode, and we want to throw away the \ + saved NULL, thus retaining our current position in the string. */ \ + string_temp = POP_FAILURE_ITEM (); \ + if (string_temp != NULL) \ + str = (const char *) string_temp; \ + \ + DEBUG_PRINT2 (" Popping string 0x%x: `", str); \ + DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \ + DEBUG_PRINT1 ("'\n"); \ + \ + pat = (unsigned char *) POP_FAILURE_ITEM (); \ + DEBUG_PRINT2 (" Popping pattern 0x%x: ", pat); \ + DEBUG_PRINT_COMPILED_PATTERN (bufp, pat, pend); \ + \ + /* Restore register info. */ \ + high_reg = (unsigned) POP_FAILURE_ITEM (); \ + DEBUG_PRINT2 (" Popping high active reg: %d\n", high_reg); \ + \ + low_reg = (unsigned) POP_FAILURE_ITEM (); \ + DEBUG_PRINT2 (" Popping low active reg: %d\n", low_reg); \ + \ + for (this_reg = high_reg; this_reg >= low_reg; this_reg--) \ + { \ + DEBUG_PRINT2 (" Popping reg: %d\n", this_reg); \ + \ + reg_info[this_reg].word = POP_FAILURE_ITEM (); \ + DEBUG_PRINT2 (" info: 0x%x\n", reg_info[this_reg]); \ + \ + regend[this_reg] = (const char *) POP_FAILURE_ITEM (); \ + DEBUG_PRINT2 (" end: 0x%x\n", regend[this_reg]); \ + \ + regstart[this_reg] = (const char *) POP_FAILURE_ITEM (); \ + DEBUG_PRINT2 (" start: 0x%x\n", regstart[this_reg]); \ + } \ + \ + DEBUG_STATEMENT (nfailure_points_popped++); \ +} /* POP_FAILURE_POINT */ + +/* re_compile_fastmap computes a ``fastmap'' for the compiled pattern in + BUFP. A fastmap records which of the (1 << BYTEWIDTH) possible + characters can start a string that matches the pattern. This fastmap + is used by re_search to skip quickly over impossible starting points. + + The caller must supply the address of a (1 << BYTEWIDTH)-byte data + area as BUFP->fastmap. + + We set the `fastmap', `fastmap_accurate', and `can_be_null' fields in + the pattern buffer. + + Returns 0 if we succeed, -2 if an internal error. */ + +int +re_compile_fastmap (bufp) + struct re_pattern_buffer *bufp; +{ + int j, k; + fail_stack_type fail_stack; +#ifndef REGEX_MALLOC + char *destination; +#endif + /* We don't push any register information onto the failure stack. */ + unsigned num_regs = 0; + + register char *fastmap = bufp->fastmap; + unsigned char *pattern = bufp->buffer; + unsigned long size = bufp->used; + const unsigned char *p = pattern; + register unsigned char *pend = pattern + size; + + /* Assume that each path through the pattern can be null until + proven otherwise. We set this false at the bottom of switch + statement, to which we get only if a particular path doesn't + match the empty string. */ + boolean path_can_be_null = true; + + /* We aren't doing a `succeed_n' to begin with. */ + boolean succeed_n_p = false; + + assert (fastmap != NULL && p != NULL); + + INIT_FAIL_STACK (); + bzero (fastmap, 1 << BYTEWIDTH); /* Assume nothing's valid. */ + bufp->fastmap_accurate = 1; /* It will be when we're done. */ + bufp->can_be_null = 0; + + while (p != pend || !FAIL_STACK_EMPTY ()) + { + if (p == pend) + { + bufp->can_be_null |= path_can_be_null; + + /* Reset for next path. */ + path_can_be_null = true; + + p = fail_stack.stack[--fail_stack.avail]; + } + + /* We should never be about to go beyond the end of the pattern. */ + assert (p < pend); + +#ifdef SWITCH_ENUM_BUG + switch ((int) ((re_opcode_t) *p++)) +#else + switch ((re_opcode_t) *p++) +#endif + { + + /* I guess the idea here is to simply not bother with a fastmap + if a backreference is used, since it's too hard to figure out + the fastmap for the corresponding group. Setting + `can_be_null' stops `re_search_2' from using the fastmap, so + that is all we do. */ + case duplicate: + bufp->can_be_null = 1; + return 0; + + + /* Following are the cases which match a character. These end + with `break'. */ + + case exactn: + fastmap[p[1]] = 1; + break; + + + case charset: + for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) + if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))) + fastmap[j] = 1; + break; + + + case charset_not: + /* Chars beyond end of map must be allowed. */ + for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++) + fastmap[j] = 1; + + for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--) + if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))) + fastmap[j] = 1; + break; + + + case wordchar: + for (j = 0; j < (1 << BYTEWIDTH); j++) + if (SYNTAX (j) == Sword) + fastmap[j] = 1; + break; + + + case notwordchar: + for (j = 0; j < (1 << BYTEWIDTH); j++) + if (SYNTAX (j) != Sword) + fastmap[j] = 1; + break; + + + case anychar: + /* `.' matches anything ... */ + for (j = 0; j < (1 << BYTEWIDTH); j++) + fastmap[j] = 1; + + /* ... except perhaps newline. */ + if (!(bufp->syntax & RE_DOT_NEWLINE)) + fastmap['\n'] = 0; + + /* Return if we have already set `can_be_null'; if we have, + then the fastmap is irrelevant. Something's wrong here. */ + else if (bufp->can_be_null) + return 0; + + /* Otherwise, have to check alternative paths. */ + break; + + +#ifdef emacs + case syntaxspec: + k = *p++; + for (j = 0; j < (1 << BYTEWIDTH); j++) + if (SYNTAX (j) == (enum syntaxcode) k) + fastmap[j] = 1; + break; + + + case notsyntaxspec: + k = *p++; + for (j = 0; j < (1 << BYTEWIDTH); j++) + if (SYNTAX (j) != (enum syntaxcode) k) + fastmap[j] = 1; + break; + + + /* All cases after this match the empty string. These end with + `continue'. */ + + + case before_dot: + case at_dot: + case after_dot: + continue; +#endif /* not emacs */ + + + case no_op: + case begline: + case endline: + case begbuf: + case endbuf: + case wordbound: + case notwordbound: + case wordbeg: + case wordend: + case push_dummy_failure: + continue; + + + case jump_n: + case pop_failure_jump: + case maybe_pop_jump: + case jump: + case jump_past_alt: + case dummy_failure_jump: + EXTRACT_NUMBER_AND_INCR (j, p); + p += j; + if (j > 0) + continue; + + /* Jump backward implies we just went through the body of a + loop and matched nothing. Opcode jumped to should be + `on_failure_jump' or `succeed_n'. Just treat it like an + ordinary jump. For a * loop, it has pushed its failure + point already; if so, discard that as redundant. */ + if ((re_opcode_t) *p != on_failure_jump + && (re_opcode_t) *p != succeed_n) + continue; + + p++; + EXTRACT_NUMBER_AND_INCR (j, p); + p += j; + + /* If what's on the stack is where we are now, pop it. */ + if (!FAIL_STACK_EMPTY () + && fail_stack.stack[fail_stack.avail - 1] == p) + fail_stack.avail--; + + continue; + + + case on_failure_jump: + case on_failure_keep_string_jump: + handle_on_failure_jump: + EXTRACT_NUMBER_AND_INCR (j, p); + + /* For some patterns, e.g., `(a?)?', `p+j' here points to the + end of the pattern. We don't want to push such a point, + since when we restore it above, entering the switch will + increment `p' past the end of the pattern. We don't need + to push such a point since we obviously won't find any more + fastmap entries beyond `pend'. Such a pattern can match + the null string, though. */ + if (p + j < pend) + { + if (!PUSH_PATTERN_OP (p + j, fail_stack)) + return -2; + } + else + bufp->can_be_null = 1; + + if (succeed_n_p) + { + EXTRACT_NUMBER_AND_INCR (k, p); /* Skip the n. */ + succeed_n_p = false; + } + + continue; + + + case succeed_n: + /* Get to the number of times to succeed. */ + p += 2; + + /* Increment p past the n for when k != 0. */ + EXTRACT_NUMBER_AND_INCR (k, p); + if (k == 0) + { + p -= 4; + succeed_n_p = true; /* Spaghetti code alert. */ + goto handle_on_failure_jump; + } + continue; + + + case set_number_at: + p += 4; + continue; + + + case start_memory: + case stop_memory: + p += 2; + continue; + + + default: + abort (); /* We have listed all the cases. */ + } /* switch *p++ */ + + /* Getting here means we have found the possible starting + characters for one path of the pattern -- and that the empty + string does not match. We need not follow this path further. + Instead, look at the next alternative (remembered on the + stack), or quit if no more. The test at the top of the loop + does these things. */ + path_can_be_null = false; + p = pend; + } /* while p */ + + /* Set `can_be_null' for the last path (also the first path, if the + pattern is empty). */ + bufp->can_be_null |= path_can_be_null; + return 0; +} /* re_compile_fastmap */ + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use + this memory for recording register information. STARTS and ENDS + must be allocated using the malloc library routine, and must each + be at least NUM_REGS * sizeof (regoff_t) bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ + +void +re_set_registers (bufp, regs, num_regs, starts, ends) + struct re_pattern_buffer *bufp; + struct re_registers *regs; + unsigned num_regs; + regoff_t *starts, *ends; +{ + if (num_regs) + { + bufp->regs_allocated = REGS_REALLOCATE; + regs->num_regs = num_regs; + regs->start = starts; + regs->end = ends; + } + else + { + bufp->regs_allocated = REGS_UNALLOCATED; + regs->num_regs = 0; + regs->start = regs->end = (regoff_t) 0; + } +} + +/* Searching routines. */ + +/* Like re_search_2, below, but only one string is specified, and + doesn't let you say where to stop matching. */ + +int +re_search (bufp, string, size, startpos, range, regs) + struct re_pattern_buffer *bufp; + const char *string; + int size, startpos, range; + struct re_registers *regs; +{ + return re_search_2 (bufp, NULL, 0, string, size, startpos, range, + regs, size); +} + + +/* Using the compiled pattern in BUFP->buffer, first tries to match the + virtual concatenation of STRING1 and STRING2, starting first at index + STARTPOS, then at STARTPOS + 1, and so on. + + STRING1 and STRING2 have length SIZE1 and SIZE2, respectively. + + RANGE is how far to scan while trying to match. RANGE = 0 means try + only at STARTPOS; in general, the last start tried is STARTPOS + + RANGE. + + In REGS, return the indices of the virtual concatenation of STRING1 + and STRING2 that matched the entire BUFP->buffer and its contained + subexpressions. + + Do not consider matching one past the index STOP in the virtual + concatenation of STRING1 and STRING2. + + We return either the position in the strings at which the match was + found, -1 if no match, or -2 if error (such as failure + stack overflow). */ + +int +re_search_2 (bufp, string1, size1, string2, size2, startpos, range, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int size1, size2; + int startpos; + int range; + struct re_registers *regs; + int stop; +{ + int val; + register char *fastmap = bufp->fastmap; + register char *translate = bufp->translate; + int total_size = size1 + size2; + int endpos = startpos + range; + + /* Check for out-of-range STARTPOS. */ + if (startpos < 0 || startpos > total_size) + return -1; + + /* Fix up RANGE if it might eventually take us outside + the virtual concatenation of STRING1 and STRING2. */ + if (endpos < -1) + range = -1 - startpos; + else if (endpos > total_size) + range = total_size - startpos; + + /* If the search isn't to be a backwards one, don't waste time in a + search for a pattern that must be anchored. */ + if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == begbuf && range > 0) + { + if (startpos > 0) + return -1; + else + range = 1; + } + + /* Update the fastmap now if not correct already. */ + if (fastmap && !bufp->fastmap_accurate) + if (re_compile_fastmap (bufp) == -2) + return -2; + + /* Loop through the string, looking for a place to start matching. */ + for (;;) + { + /* If a fastmap is supplied, skip quickly over characters that + cannot be the start of a match. If the pattern can match the + null string, however, we don't need to skip characters; we want + the first null string. */ + if (fastmap && startpos < total_size && !bufp->can_be_null) + { + if (range > 0) /* Searching forwards. */ + { + register const char *d; + register int lim = 0; + int irange = range; + + if (startpos < size1 && startpos + range >= size1) + lim = range - (size1 - startpos); + + d = (startpos >= size1 ? string2 - size1 : string1) + startpos; + + /* Written out as an if-else to avoid testing `translate' + inside the loop. */ + if (translate) + while (range > lim + && !fastmap[(unsigned char) + translate[(unsigned char) *d++]]) + range--; + else + while (range > lim && !fastmap[(unsigned char) *d++]) + range--; + + startpos += irange - range; + } + else /* Searching backwards. */ + { + register char c = (size1 == 0 || startpos >= size1 + ? string2[startpos - size1] + : string1[startpos]); + + if (!fastmap[(unsigned char) TRANSLATE (c)]) + goto advance; + } + } + + /* If can't match the null string, and that's all we have left, fail. */ + if (range >= 0 && startpos == total_size && fastmap + && !bufp->can_be_null) + return -1; + + val = re_match_2 (bufp, string1, size1, string2, size2, + startpos, regs, stop); + if (val >= 0) + return startpos; + + if (val == -2) + return -2; + + advance: + if (!range) + break; + else if (range > 0) + { + range--; + startpos++; + } + else + { + range++; + startpos--; + } + } + return -1; +} /* re_search_2 */ + +/* Declarations and macros for re_match_2. */ + +static int bcmp_translate (); +static boolean alt_match_null_string_p (), + common_op_match_null_string_p (), + group_match_null_string_p (); + +/* Structure for per-register (a.k.a. per-group) information. + This must not be longer than one word, because we push this value + onto the failure stack. Other register information, such as the + starting and ending positions (which are addresses), and the list of + inner groups (which is a bits list) are maintained in separate + variables. + + We are making a (strictly speaking) nonportable assumption here: that + the compiler will pack our bit fields into something that fits into + the type of `word', i.e., is something that fits into one item on the + failure stack. */ +typedef union +{ + fail_stack_elt_t word; + struct + { + /* This field is one if this group can match the empty string, + zero if not. If not yet determined, `MATCH_NULL_UNSET_VALUE'. */ +#define MATCH_NULL_UNSET_VALUE 3 + unsigned match_null_string_p : 2; + unsigned is_active : 1; + unsigned matched_something : 1; + unsigned ever_matched_something : 1; + } bits; +} register_info_type; + +#define REG_MATCH_NULL_STRING_P(R) ((R).bits.match_null_string_p) +#define IS_ACTIVE(R) ((R).bits.is_active) +#define MATCHED_SOMETHING(R) ((R).bits.matched_something) +#define EVER_MATCHED_SOMETHING(R) ((R).bits.ever_matched_something) + + +/* Call this when have matched a real character; it sets `matched' flags + for the subexpressions which we are currently inside. Also records + that those subexprs have matched. */ +#define SET_REGS_MATCHED() \ + do \ + { \ + unsigned r; \ + for (r = lowest_active_reg; r <= highest_active_reg; r++) \ + { \ + MATCHED_SOMETHING (reg_info[r]) \ + = EVER_MATCHED_SOMETHING (reg_info[r]) \ + = 1; \ + } \ + } \ + while (0) + + +/* This converts PTR, a pointer into one of the search strings `string1' + and `string2' into an offset from the beginning of that string. */ +#define POINTER_TO_OFFSET(ptr) \ + (FIRST_STRING_P (ptr) ? (ptr) - string1 : (ptr) - string2 + size1) + +/* Registers are set to a sentinel when they haven't yet matched. */ +#define REG_UNSET_VALUE ((char *) -1) +#define REG_UNSET(e) ((e) == REG_UNSET_VALUE) + + +/* Macros for dealing with the split strings in re_match_2. */ + +#define MATCHING_IN_FIRST_STRING (dend == end_match_1) + +/* Call before fetching a character with *d. This switches over to + string2 if necessary. */ +#define PREFETCH() \ + while (d == dend) \ + { \ + /* End of string2 => fail. */ \ + if (dend == end_match_2) \ + goto fail; \ + /* End of string1 => advance to string2. */ \ + d = string2; \ + dend = end_match_2; \ + } + + +/* Test if at very beginning or at very end of the virtual concatenation + of `string1' and `string2'. If only one string, it's `string2'. */ +#define AT_STRINGS_BEG(d) ((d) == (size1 ? string1 : string2) || !size2) +#define AT_STRINGS_END(d) ((d) == end2) + + +/* Test if D points to a character which is word-constituent. We have + two special cases to check for: if past the end of string1, look at + the first character in string2; and if before the beginning of + string2, look at the last character in string1. */ +#define WORDCHAR_P(d) \ + (SYNTAX ((d) == end1 ? *string2 \ + : (d) == string2 - 1 ? *(end1 - 1) : *(d)) \ + == Sword) + +/* Test if the character before D and the one at D differ with respect + to being word-constituent. */ +#define AT_WORD_BOUNDARY(d) \ + (AT_STRINGS_BEG (d) || AT_STRINGS_END (d) \ + || WORDCHAR_P (d - 1) != WORDCHAR_P (d)) + + +/* Free everything we malloc. */ +#ifdef REGEX_MALLOC +#define FREE_VAR(var) if (var) free (var); var = NULL +#define FREE_VARIABLES() \ + do { \ + FREE_VAR (fail_stack.stack); \ + FREE_VAR (regstart); \ + FREE_VAR (regend); \ + FREE_VAR (old_regstart); \ + FREE_VAR (old_regend); \ + FREE_VAR (best_regstart); \ + FREE_VAR (best_regend); \ + FREE_VAR (reg_info); \ + FREE_VAR (reg_dummy); \ + FREE_VAR (reg_info_dummy); \ + } while (0) +#else /* not REGEX_MALLOC */ +/* Some MIPS systems (at least) want this to free alloca'd storage. */ +#define FREE_VARIABLES() alloca (0) +#endif /* not REGEX_MALLOC */ + + +/* These values must meet several constraints. They must not be valid + register values; since we have a limit of 255 registers (because + we use only one byte in the pattern for the register number), we can + use numbers larger than 255. They must differ by 1, because of + NUM_FAILURE_ITEMS above. And the value for the lowest register must + be larger than the value for the highest register, so we do not try + to actually save any registers when none are active. */ +#define NO_HIGHEST_ACTIVE_REG (1 << BYTEWIDTH) +#define NO_LOWEST_ACTIVE_REG (NO_HIGHEST_ACTIVE_REG + 1) + +/* Matching routines. */ + +#ifndef emacs /* Emacs never uses this. */ +/* re_match is like re_match_2 except it takes only a single string. */ + +int +re_match (bufp, string, size, pos, regs) + struct re_pattern_buffer *bufp; + const char *string; + int size, pos; + struct re_registers *regs; + { + return re_match_2 (bufp, NULL, 0, string, size, pos, regs, size); +} +#endif /* not emacs */ + + +/* re_match_2 matches the compiled pattern in BUFP against the + the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1 + and SIZE2, respectively). We start matching at POS, and stop + matching at STOP. + + If REGS is non-null and the `no_sub' field of BUFP is nonzero, we + store offsets for the substring each group matched in REGS. See the + documentation for exactly how many groups we fill. + + We return -1 if no match, -2 if an internal error (such as the + failure stack overflowing). Otherwise, we return the length of the + matched substring. */ + +int +re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int size1, size2; + int pos; + struct re_registers *regs; + int stop; +{ + /* General temporaries. */ + int mcnt; + unsigned char *p1; + + /* Just past the end of the corresponding string. */ + const char *end1, *end2; + + /* Pointers into string1 and string2, just past the last characters in + each to consider matching. */ + const char *end_match_1, *end_match_2; + + /* Where we are in the data, and the end of the current string. */ + const char *d, *dend; + + /* Where we are in the pattern, and the end of the pattern. */ + unsigned char *p = bufp->buffer; + register unsigned char *pend = p + bufp->used; + + /* We use this to map every character in the string. */ + char *translate = bufp->translate; + + /* Failure point stack. Each place that can handle a failure further + down the line pushes a failure point on this stack. It consists of + restart, regend, and reg_info for all registers corresponding to + the subexpressions we're currently inside, plus the number of such + registers, and, finally, two char *'s. The first char * is where + to resume scanning the pattern; the second one is where to resume + scanning the strings. If the latter is zero, the failure point is + a ``dummy''; if a failure happens and the failure point is a dummy, + it gets discarded and the next next one is tried. */ + fail_stack_type fail_stack; +#ifdef DEBUG + static unsigned failure_id = 0; + unsigned nfailure_points_pushed = 0, nfailure_points_popped = 0; +#endif + + /* We fill all the registers internally, independent of what we + return, for use in backreferences. The number here includes + an element for register zero. */ + unsigned num_regs = bufp->re_nsub + 1; + + /* The currently active registers. */ + unsigned lowest_active_reg = NO_LOWEST_ACTIVE_REG; + unsigned highest_active_reg = NO_HIGHEST_ACTIVE_REG; + + /* Information on the contents of registers. These are pointers into + the input strings; they record just what was matched (on this + attempt) by a subexpression part of the pattern, that is, the + regnum-th regstart pointer points to where in the pattern we began + matching and the regnum-th regend points to right after where we + stopped matching the regnum-th subexpression. (The zeroth register + keeps track of what the whole pattern matches.) */ + const char **regstart, **regend; + + /* If a group that's operated upon by a repetition operator fails to + match anything, then the register for its start will need to be + restored because it will have been set to wherever in the string we + are when we last see its open-group operator. Similarly for a + register's end. */ + const char **old_regstart, **old_regend; + + /* The is_active field of reg_info helps us keep track of which (possibly + nested) subexpressions we are currently in. The matched_something + field of reg_info[reg_num] helps us tell whether or not we have + matched any of the pattern so far this time through the reg_num-th + subexpression. These two fields get reset each time through any + loop their register is in. */ + register_info_type *reg_info; + + /* The following record the register info as found in the above + variables when we find a match better than any we've seen before. + This happens as we backtrack through the failure points, which in + turn happens only if we have not yet matched the entire string. */ + unsigned best_regs_set = false; + const char **best_regstart, **best_regend; + + /* Logically, this is `best_regend[0]'. But we don't want to have to + allocate space for that if we're not allocating space for anything + else (see below). Also, we never need info about register 0 for + any of the other register vectors, and it seems rather a kludge to + treat `best_regend' differently than the rest. So we keep track of + the end of the best match so far in a separate variable. We + initialize this to NULL so that when we backtrack the first time + and need to test it, it's not garbage. */ + const char *match_end = NULL; + + /* Used when we pop values we don't care about. */ + const char **reg_dummy; + register_info_type *reg_info_dummy; + +#ifdef DEBUG + /* Counts the total number of registers pushed. */ + unsigned num_regs_pushed = 0; +#endif + + DEBUG_PRINT1 ("\n\nEntering re_match_2.\n"); + + INIT_FAIL_STACK (); + + /* Do not bother to initialize all the register variables if there are + no groups in the pattern, as it takes a fair amount of time. If + there are groups, we include space for register 0 (the whole + pattern), even though we never use it, since it simplifies the + array indexing. We should fix this. */ + if (bufp->re_nsub) + { + regstart = REGEX_TALLOC (num_regs, const char *); + regend = REGEX_TALLOC (num_regs, const char *); + old_regstart = REGEX_TALLOC (num_regs, const char *); + old_regend = REGEX_TALLOC (num_regs, const char *); + best_regstart = REGEX_TALLOC (num_regs, const char *); + best_regend = REGEX_TALLOC (num_regs, const char *); + reg_info = REGEX_TALLOC (num_regs, register_info_type); + reg_dummy = REGEX_TALLOC (num_regs, const char *); + reg_info_dummy = REGEX_TALLOC (num_regs, register_info_type); + + if (!(regstart && regend && old_regstart && old_regend && reg_info + && best_regstart && best_regend && reg_dummy && reg_info_dummy)) + { + FREE_VARIABLES (); + return -2; + } + } +#ifdef REGEX_MALLOC + else + { + /* We must initialize all our variables to NULL, so that + `FREE_VARIABLES' doesn't try to free them. */ + regstart = regend = old_regstart = old_regend = best_regstart + = best_regend = reg_dummy = NULL; + reg_info = reg_info_dummy = (register_info_type *) NULL; + } +#endif /* REGEX_MALLOC */ + + /* The starting position is bogus. */ + if (pos < 0 || pos > size1 + size2) + { + FREE_VARIABLES (); + return -1; + } + + /* Initialize subexpression text positions to -1 to mark ones that no + start_memory/stop_memory has been seen for. Also initialize the + register information struct. */ + for (mcnt = 1; mcnt < num_regs; mcnt++) + { + regstart[mcnt] = regend[mcnt] + = old_regstart[mcnt] = old_regend[mcnt] = REG_UNSET_VALUE; + + REG_MATCH_NULL_STRING_P (reg_info[mcnt]) = MATCH_NULL_UNSET_VALUE; + IS_ACTIVE (reg_info[mcnt]) = 0; + MATCHED_SOMETHING (reg_info[mcnt]) = 0; + EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0; + } + + /* We move `string1' into `string2' if the latter's empty -- but not if + `string1' is null. */ + if (size2 == 0 && string1 != NULL) + { + string2 = string1; + size2 = size1; + string1 = 0; + size1 = 0; + } + end1 = string1 + size1; + end2 = string2 + size2; + + /* Compute where to stop matching, within the two strings. */ + if (stop <= size1) + { + end_match_1 = string1 + stop; + end_match_2 = string2; + } + else + { + end_match_1 = end1; + end_match_2 = string2 + stop - size1; + } + + /* `p' scans through the pattern as `d' scans through the data. + `dend' is the end of the input string that `d' points within. `d' + is advanced into the following input string whenever necessary, but + this happens before fetching; therefore, at the beginning of the + loop, `d' can be pointing at the end of a string, but it cannot + equal `string2'. */ + if (size1 > 0 && pos <= size1) + { + d = string1 + pos; + dend = end_match_1; + } + else + { + d = string2 + pos - size1; + dend = end_match_2; + } + + DEBUG_PRINT1 ("The compiled pattern is: "); + DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend); + DEBUG_PRINT1 ("The string to match is: `"); + DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2); + DEBUG_PRINT1 ("'\n"); + + /* This loops over pattern commands. It exits by returning from the + function if the match is complete, or it drops through if the match + fails at this starting point in the input data. */ + for (;;) + { + DEBUG_PRINT2 ("\n0x%x: ", p); + + if (p == pend) + { /* End of pattern means we might have succeeded. */ + DEBUG_PRINT1 ("end of pattern ... "); + + /* If we haven't matched the entire string, and we want the + longest match, try backtracking. */ + if (d != end_match_2) + { + DEBUG_PRINT1 ("backtracking.\n"); + + if (!FAIL_STACK_EMPTY ()) + { /* More failure points to try. */ + boolean same_str_p = (FIRST_STRING_P (match_end) + == MATCHING_IN_FIRST_STRING); + + /* If exceeds best match so far, save it. */ + if (!best_regs_set + || (same_str_p && d > match_end) + || (!same_str_p && !MATCHING_IN_FIRST_STRING)) + { + best_regs_set = true; + match_end = d; + + DEBUG_PRINT1 ("\nSAVING match as best so far.\n"); + + for (mcnt = 1; mcnt < num_regs; mcnt++) + { + best_regstart[mcnt] = regstart[mcnt]; + best_regend[mcnt] = regend[mcnt]; + } + } + goto fail; + } + + /* If no failure points, don't restore garbage. */ + else if (best_regs_set) + { + restore_best_regs: + /* Restore best match. It may happen that `dend == + end_match_1' while the restored d is in string2. + For example, the pattern `x.*y.*z' against the + strings `x-' and `y-z-', if the two strings are + not consecutive in memory. */ + DEBUG_PRINT1 ("Restoring best registers.\n"); + + d = match_end; + dend = ((d >= string1 && d <= end1) + ? end_match_1 : end_match_2); + + for (mcnt = 1; mcnt < num_regs; mcnt++) + { + regstart[mcnt] = best_regstart[mcnt]; + regend[mcnt] = best_regend[mcnt]; + } + } + } /* d != end_match_2 */ + + DEBUG_PRINT1 ("Accepting match.\n"); + + /* If caller wants register contents data back, do it. */ + if (regs && !bufp->no_sub) + { + /* Have the register data arrays been allocated? */ + if (bufp->regs_allocated == REGS_UNALLOCATED) + { /* No. So allocate them with malloc. We need one + extra element beyond `num_regs' for the `-1' marker + GNU code uses. */ + regs->num_regs = MAX (RE_NREGS, num_regs + 1); + regs->start = TALLOC (regs->num_regs, regoff_t); + regs->end = TALLOC (regs->num_regs, regoff_t); + if (regs->start == NULL || regs->end == NULL) + return -2; + bufp->regs_allocated = REGS_REALLOCATE; + } + else if (bufp->regs_allocated == REGS_REALLOCATE) + { /* Yes. If we need more elements than were already + allocated, reallocate them. If we need fewer, just + leave it alone. */ + if (regs->num_regs < num_regs + 1) + { + regs->num_regs = num_regs + 1; + RETALLOC (regs->start, regs->num_regs, regoff_t); + RETALLOC (regs->end, regs->num_regs, regoff_t); + if (regs->start == NULL || regs->end == NULL) + return -2; + } + } + else + assert (bufp->regs_allocated == REGS_FIXED); + + /* Convert the pointer data in `regstart' and `regend' to + indices. Register zero has to be set differently, + since we haven't kept track of any info for it. */ + if (regs->num_regs > 0) + { + regs->start[0] = pos; + regs->end[0] = (MATCHING_IN_FIRST_STRING ? d - string1 + : d - string2 + size1); + } + + /* Go through the first `min (num_regs, regs->num_regs)' + registers, since that is all we initialized. */ + for (mcnt = 1; mcnt < MIN (num_regs, regs->num_regs); mcnt++) + { + if (REG_UNSET (regstart[mcnt]) || REG_UNSET (regend[mcnt])) + regs->start[mcnt] = regs->end[mcnt] = -1; + else + { + regs->start[mcnt] = POINTER_TO_OFFSET (regstart[mcnt]); + regs->end[mcnt] = POINTER_TO_OFFSET (regend[mcnt]); + } + } + + /* If the regs structure we return has more elements than + were in the pattern, set the extra elements to -1. If + we (re)allocated the registers, this is the case, + because we always allocate enough to have at least one + -1 at the end. */ + for (mcnt = num_regs; mcnt < regs->num_regs; mcnt++) + regs->start[mcnt] = regs->end[mcnt] = -1; + } /* regs && !bufp->no_sub */ + + FREE_VARIABLES (); + DEBUG_PRINT4 ("%u failure points pushed, %u popped (%u remain).\n", + nfailure_points_pushed, nfailure_points_popped, + nfailure_points_pushed - nfailure_points_popped); + DEBUG_PRINT2 ("%u registers pushed.\n", num_regs_pushed); + + mcnt = d - pos - (MATCHING_IN_FIRST_STRING + ? string1 + : string2 - size1); + + DEBUG_PRINT2 ("Returning %d from re_match_2.\n", mcnt); + + return mcnt; + } + + /* Otherwise match next pattern command. */ +#ifdef SWITCH_ENUM_BUG + switch ((int) ((re_opcode_t) *p++)) +#else + switch ((re_opcode_t) *p++) +#endif + { + /* Ignore these. Used to ignore the n of succeed_n's which + currently have n == 0. */ + case no_op: + DEBUG_PRINT1 ("EXECUTING no_op.\n"); + break; + + + /* Match the next n pattern characters exactly. The following + byte in the pattern defines n, and the n bytes after that + are the characters to match. */ + case exactn: + mcnt = *p++; + DEBUG_PRINT2 ("EXECUTING exactn %d.\n", mcnt); + + /* This is written out as an if-else so we don't waste time + testing `translate' inside the loop. */ + if (translate) + { + do + { + PREFETCH (); + if (translate[(unsigned char) *d++] != (char) *p++) + goto fail; + } + while (--mcnt); + } + else + { + do + { + PREFETCH (); + if (*d++ != (char) *p++) goto fail; + } + while (--mcnt); + } + SET_REGS_MATCHED (); + break; + + + /* Match any character except possibly a newline or a null. */ + case anychar: + DEBUG_PRINT1 ("EXECUTING anychar.\n"); + + PREFETCH (); + + if ((!(bufp->syntax & RE_DOT_NEWLINE) && TRANSLATE (*d) == '\n') + || (bufp->syntax & RE_DOT_NOT_NULL && TRANSLATE (*d) == '\000')) + goto fail; + + SET_REGS_MATCHED (); + DEBUG_PRINT2 (" Matched `%d'.\n", *d); + d++; + break; + + + case charset: + case charset_not: + { + register unsigned char c; + boolean not = (re_opcode_t) *(p - 1) == charset_not; + + DEBUG_PRINT2 ("EXECUTING charset%s.\n", not ? "_not" : ""); + + PREFETCH (); + c = TRANSLATE (*d); /* The character to match. */ + + /* Cast to `unsigned' instead of `unsigned char' in case the + bit list is a full 32 bytes long. */ + if (c < (unsigned) (*p * BYTEWIDTH) + && p[1 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH))) + not = !not; + + p += 1 + *p; + + if (!not) goto fail; + + SET_REGS_MATCHED (); + d++; + break; + } + + + /* The beginning of a group is represented by start_memory. + The arguments are the register number in the next byte, and the + number of groups inner to this one in the next. The text + matched within the group is recorded (in the internal + registers data structure) under the register number. */ + case start_memory: + DEBUG_PRINT3 ("EXECUTING start_memory %d (%d):\n", *p, p[1]); + + /* Find out if this group can match the empty string. */ + p1 = p; /* To send to group_match_null_string_p. */ + + if (REG_MATCH_NULL_STRING_P (reg_info[*p]) == MATCH_NULL_UNSET_VALUE) + REG_MATCH_NULL_STRING_P (reg_info[*p]) + = group_match_null_string_p (&p1, pend, reg_info); + + /* Save the position in the string where we were the last time + we were at this open-group operator in case the group is + operated upon by a repetition operator, e.g., with `(a*)*b' + against `ab'; then we want to ignore where we are now in + the string in case this attempt to match fails. */ + old_regstart[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p]) + ? REG_UNSET (regstart[*p]) ? d : regstart[*p] + : regstart[*p]; + DEBUG_PRINT2 (" old_regstart: %d\n", + POINTER_TO_OFFSET (old_regstart[*p])); + + regstart[*p] = d; + DEBUG_PRINT2 (" regstart: %d\n", POINTER_TO_OFFSET (regstart[*p])); + + IS_ACTIVE (reg_info[*p]) = 1; + MATCHED_SOMETHING (reg_info[*p]) = 0; + + /* This is the new highest active register. */ + highest_active_reg = *p; + + /* If nothing was active before, this is the new lowest active + register. */ + if (lowest_active_reg == NO_LOWEST_ACTIVE_REG) + lowest_active_reg = *p; + + /* Move past the register number and inner group count. */ + p += 2; + break; + + + /* The stop_memory opcode represents the end of a group. Its + arguments are the same as start_memory's: the register + number, and the number of inner groups. */ + case stop_memory: + DEBUG_PRINT3 ("EXECUTING stop_memory %d (%d):\n", *p, p[1]); + + /* We need to save the string position the last time we were at + this close-group operator in case the group is operated + upon by a repetition operator, e.g., with `((a*)*(b*)*)*' + against `aba'; then we want to ignore where we are now in + the string in case this attempt to match fails. */ + old_regend[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p]) + ? REG_UNSET (regend[*p]) ? d : regend[*p] + : regend[*p]; + DEBUG_PRINT2 (" old_regend: %d\n", + POINTER_TO_OFFSET (old_regend[*p])); + + regend[*p] = d; + DEBUG_PRINT2 (" regend: %d\n", POINTER_TO_OFFSET (regend[*p])); + + /* This register isn't active anymore. */ + IS_ACTIVE (reg_info[*p]) = 0; + + /* If this was the only register active, nothing is active + anymore. */ + if (lowest_active_reg == highest_active_reg) + { + lowest_active_reg = NO_LOWEST_ACTIVE_REG; + highest_active_reg = NO_HIGHEST_ACTIVE_REG; + } + else + { /* We must scan for the new highest active register, since + it isn't necessarily one less than now: consider + (a(b)c(d(e)f)g). When group 3 ends, after the f), the + new highest active register is 1. */ + unsigned char r = *p - 1; + while (r > 0 && !IS_ACTIVE (reg_info[r])) + r--; + + /* If we end up at register zero, that means that we saved + the registers as the result of an `on_failure_jump', not + a `start_memory', and we jumped to past the innermost + `stop_memory'. For example, in ((.)*) we save + registers 1 and 2 as a result of the *, but when we pop + back to the second ), we are at the stop_memory 1. + Thus, nothing is active. */ + if (r == 0) + { + lowest_active_reg = NO_LOWEST_ACTIVE_REG; + highest_active_reg = NO_HIGHEST_ACTIVE_REG; + } + else + highest_active_reg = r; + } + + /* If just failed to match something this time around with a + group that's operated on by a repetition operator, try to + force exit from the ``loop'', and restore the register + information for this group that we had before trying this + last match. */ + if ((!MATCHED_SOMETHING (reg_info[*p]) + || (re_opcode_t) p[-3] == start_memory) + && (p + 2) < pend) + { + boolean is_a_jump_n = false; + + p1 = p + 2; + mcnt = 0; + switch ((re_opcode_t) *p1++) + { + case jump_n: + is_a_jump_n = true; + case pop_failure_jump: + case maybe_pop_jump: + case jump: + case dummy_failure_jump: + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + if (is_a_jump_n) + p1 += 2; + break; + + default: + /* do nothing */ ; + } + p1 += mcnt; + + /* If the next operation is a jump backwards in the pattern + to an on_failure_jump right before the start_memory + corresponding to this stop_memory, exit from the loop + by forcing a failure after pushing on the stack the + on_failure_jump's jump in the pattern, and d. */ + if (mcnt < 0 && (re_opcode_t) *p1 == on_failure_jump + && (re_opcode_t) p1[3] == start_memory && p1[4] == *p) + { + /* If this group ever matched anything, then restore + what its registers were before trying this last + failed match, e.g., with `(a*)*b' against `ab' for + regstart[1], and, e.g., with `((a*)*(b*)*)*' + against `aba' for regend[3]. + + Also restore the registers for inner groups for, + e.g., `((a*)(b*))*' against `aba' (register 3 would + otherwise get trashed). */ + + if (EVER_MATCHED_SOMETHING (reg_info[*p])) + { + unsigned r; + + EVER_MATCHED_SOMETHING (reg_info[*p]) = 0; + + /* Restore this and inner groups' (if any) registers. */ + for (r = *p; r < *p + *(p + 1); r++) + { + regstart[r] = old_regstart[r]; + + /* xx why this test? */ + if ((int) old_regend[r] >= (int) regstart[r]) + regend[r] = old_regend[r]; + } + } + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + PUSH_FAILURE_POINT (p1 + mcnt, d, -2); + + goto fail; + } + } + + /* Move past the register number and the inner group count. */ + p += 2; + break; + + + /* \<digit> has been turned into a `duplicate' command which is + followed by the numeric value of <digit> as the register number. */ + case duplicate: + { + register const char *d2, *dend2; + int regno = *p++; /* Get which register to match against. */ + DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno); + + /* Can't back reference a group which we've never matched. */ + if (REG_UNSET (regstart[regno]) || REG_UNSET (regend[regno])) + goto fail; + + /* Where in input to try to start matching. */ + d2 = regstart[regno]; + + /* Where to stop matching; if both the place to start and + the place to stop matching are in the same string, then + set to the place to stop, otherwise, for now have to use + the end of the first string. */ + + dend2 = ((FIRST_STRING_P (regstart[regno]) + == FIRST_STRING_P (regend[regno])) + ? regend[regno] : end_match_1); + for (;;) + { + /* If necessary, advance to next segment in register + contents. */ + while (d2 == dend2) + { + if (dend2 == end_match_2) break; + if (dend2 == regend[regno]) break; + + /* End of string1 => advance to string2. */ + d2 = string2; + dend2 = regend[regno]; + } + /* At end of register contents => success */ + if (d2 == dend2) break; + + /* If necessary, advance to next segment in data. */ + PREFETCH (); + + /* How many characters left in this segment to match. */ + mcnt = dend - d; + + /* Want how many consecutive characters we can match in + one shot, so, if necessary, adjust the count. */ + if (mcnt > dend2 - d2) + mcnt = dend2 - d2; + + /* Compare that many; failure if mismatch, else move + past them. */ + if (translate + ? bcmp_translate (d, d2, mcnt, translate) + : bcmp (d, d2, mcnt)) + goto fail; + d += mcnt, d2 += mcnt; + } + } + break; + + + /* begline matches the empty string at the beginning of the string + (unless `not_bol' is set in `bufp'), and, if + `newline_anchor' is set, after newlines. */ + case begline: + DEBUG_PRINT1 ("EXECUTING begline.\n"); + + if (AT_STRINGS_BEG (d)) + { + if (!bufp->not_bol) break; + } + else if (d[-1] == '\n' && bufp->newline_anchor) + { + break; + } + /* In all other cases, we fail. */ + goto fail; + + + /* endline is the dual of begline. */ + case endline: + DEBUG_PRINT1 ("EXECUTING endline.\n"); + + if (AT_STRINGS_END (d)) + { + if (!bufp->not_eol) break; + } + + /* We have to ``prefetch'' the next character. */ + else if ((d == end1 ? *string2 : *d) == '\n' + && bufp->newline_anchor) + { + break; + } + goto fail; + + + /* Match at the very beginning of the data. */ + case begbuf: + DEBUG_PRINT1 ("EXECUTING begbuf.\n"); + if (AT_STRINGS_BEG (d)) + break; + goto fail; + + + /* Match at the very end of the data. */ + case endbuf: + DEBUG_PRINT1 ("EXECUTING endbuf.\n"); + if (AT_STRINGS_END (d)) + break; + goto fail; + + + /* on_failure_keep_string_jump is used to optimize `.*\n'. It + pushes NULL as the value for the string on the stack. Then + `pop_failure_point' will keep the current value for the + string, instead of restoring it. To see why, consider + matching `foo\nbar' against `.*\n'. The .* matches the foo; + then the . fails against the \n. But the next thing we want + to do is match the \n against the \n; if we restored the + string value, we would be back at the foo. + + Because this is used only in specific cases, we don't need to + check all the things that `on_failure_jump' does, to make + sure the right things get saved on the stack. Hence we don't + share its code. The only reason to push anything on the + stack at all is that otherwise we would have to change + `anychar's code to do something besides goto fail in this + case; that seems worse than this. */ + case on_failure_keep_string_jump: + DEBUG_PRINT1 ("EXECUTING on_failure_keep_string_jump"); + + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT3 (" %d (to 0x%x):\n", mcnt, p + mcnt); + + PUSH_FAILURE_POINT (p + mcnt, NULL, -2); + break; + + + /* Uses of on_failure_jump: + + Each alternative starts with an on_failure_jump that points + to the beginning of the next alternative. Each alternative + except the last ends with a jump that in effect jumps past + the rest of the alternatives. (They really jump to the + ending jump of the following alternative, because tensioning + these jumps is a hassle.) + + Repeats start with an on_failure_jump that points past both + the repetition text and either the following jump or + pop_failure_jump back to this on_failure_jump. */ + case on_failure_jump: + on_failure: + DEBUG_PRINT1 ("EXECUTING on_failure_jump"); + + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT3 (" %d (to 0x%x)", mcnt, p + mcnt); + + /* If this on_failure_jump comes right before a group (i.e., + the original * applied to a group), save the information + for that group and all inner ones, so that if we fail back + to this point, the group's information will be correct. + For example, in \(a*\)*\1, we need the preceding group, + and in \(\(a*\)b*\)\2, we need the inner group. */ + + /* We can't use `p' to check ahead because we push + a failure point to `p + mcnt' after we do this. */ + p1 = p; + + /* We need to skip no_op's before we look for the + start_memory in case this on_failure_jump is happening as + the result of a completed succeed_n, as in \(a\)\{1,3\}b\1 + against aba. */ + while (p1 < pend && (re_opcode_t) *p1 == no_op) + p1++; + + if (p1 < pend && (re_opcode_t) *p1 == start_memory) + { + /* We have a new highest active register now. This will + get reset at the start_memory we are about to get to, + but we will have saved all the registers relevant to + this repetition op, as described above. */ + highest_active_reg = *(p1 + 1) + *(p1 + 2); + if (lowest_active_reg == NO_LOWEST_ACTIVE_REG) + lowest_active_reg = *(p1 + 1); + } + + DEBUG_PRINT1 (":\n"); + PUSH_FAILURE_POINT (p + mcnt, d, -2); + break; + + + /* A smart repeat ends with `maybe_pop_jump'. + We change it to either `pop_failure_jump' or `jump'. */ + case maybe_pop_jump: + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT2 ("EXECUTING maybe_pop_jump %d.\n", mcnt); + { + register unsigned char *p2 = p; + + /* Compare the beginning of the repeat with what in the + pattern follows its end. If we can establish that there + is nothing that they would both match, i.e., that we + would have to backtrack because of (as in, e.g., `a*a') + then we can change to pop_failure_jump, because we'll + never have to backtrack. + + This is not true in the case of alternatives: in + `(a|ab)*' we do need to backtrack to the `ab' alternative + (e.g., if the string was `ab'). But instead of trying to + detect that here, the alternative has put on a dummy + failure point which is what we will end up popping. */ + + /* Skip over open/close-group commands. */ + while (p2 + 2 < pend + && ((re_opcode_t) *p2 == stop_memory + || (re_opcode_t) *p2 == start_memory)) + p2 += 3; /* Skip over args, too. */ + + /* If we're at the end of the pattern, we can change. */ + if (p2 == pend) + { + /* Consider what happens when matching ":\(.*\)" + against ":/". I don't really understand this code + yet. */ + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 + (" End of pattern: change to `pop_failure_jump'.\n"); + } + + else if ((re_opcode_t) *p2 == exactn + || (bufp->newline_anchor && (re_opcode_t) *p2 == endline)) + { + register unsigned char c + = *p2 == (unsigned char) endline ? '\n' : p2[2]; + p1 = p + mcnt; + + /* p1[0] ... p1[2] are the `on_failure_jump' corresponding + to the `maybe_finalize_jump' of this case. Examine what + follows. */ + if ((re_opcode_t) p1[3] == exactn && p1[5] != c) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n", + c, p1[5]); + } + + else if ((re_opcode_t) p1[3] == charset + || (re_opcode_t) p1[3] == charset_not) + { + int not = (re_opcode_t) p1[3] == charset_not; + + if (c < (unsigned char) (p1[4] * BYTEWIDTH) + && p1[5 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH))) + not = !not; + + /* `not' is equal to 1 if c would match, which means + that we can't change to pop_failure_jump. */ + if (!not) + { + p[-3] = (unsigned char) pop_failure_jump; + DEBUG_PRINT1 (" No match => pop_failure_jump.\n"); + } + } + } + } + p -= 2; /* Point at relative address again. */ + if ((re_opcode_t) p[-1] != pop_failure_jump) + { + p[-1] = (unsigned char) jump; + DEBUG_PRINT1 (" Match => jump.\n"); + goto unconditional_jump; + } + /* Note fall through. */ + + + /* The end of a simple repeat has a pop_failure_jump back to + its matching on_failure_jump, where the latter will push a + failure point. The pop_failure_jump takes off failure + points put on by this pop_failure_jump's matching + on_failure_jump; we got through the pattern to here from the + matching on_failure_jump, so didn't fail. */ + case pop_failure_jump: + { + /* We need to pass separate storage for the lowest and + highest registers, even though we don't care about the + actual values. Otherwise, we will restore only one + register from the stack, since lowest will == highest in + `pop_failure_point'. */ + unsigned dummy_low_reg, dummy_high_reg; + unsigned char *pdummy; + const char *sdummy; + + DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n"); + POP_FAILURE_POINT (sdummy, pdummy, + dummy_low_reg, dummy_high_reg, + reg_dummy, reg_dummy, reg_info_dummy); + } + /* Note fall through. */ + + + /* Unconditionally jump (without popping any failure points). */ + case jump: + unconditional_jump: + EXTRACT_NUMBER_AND_INCR (mcnt, p); /* Get the amount to jump. */ + DEBUG_PRINT2 ("EXECUTING jump %d ", mcnt); + p += mcnt; /* Do the jump. */ + DEBUG_PRINT2 ("(to 0x%x).\n", p); + break; + + + /* We need this opcode so we can detect where alternatives end + in `group_match_null_string_p' et al. */ + case jump_past_alt: + DEBUG_PRINT1 ("EXECUTING jump_past_alt.\n"); + goto unconditional_jump; + + + /* Normally, the on_failure_jump pushes a failure point, which + then gets popped at pop_failure_jump. We will end up at + pop_failure_jump, also, and with a pattern of, say, `a+', we + are skipping over the on_failure_jump, so we have to push + something meaningless for pop_failure_jump to pop. */ + case dummy_failure_jump: + DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n"); + /* It doesn't matter what we push for the string here. What + the code at `fail' tests is the value for the pattern. */ + PUSH_FAILURE_POINT (0, 0, -2); + goto unconditional_jump; + + + /* At the end of an alternative, we need to push a dummy failure + point in case we are followed by a `pop_failure_jump', because + we don't want the failure point for the alternative to be + popped. For example, matching `(a|ab)*' against `aab' + requires that we match the `ab' alternative. */ + case push_dummy_failure: + DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n"); + /* See comments just above at `dummy_failure_jump' about the + two zeroes. */ + PUSH_FAILURE_POINT (0, 0, -2); + break; + + /* Have to succeed matching what follows at least n times. + After that, handle like `on_failure_jump'. */ + case succeed_n: + EXTRACT_NUMBER (mcnt, p + 2); + DEBUG_PRINT2 ("EXECUTING succeed_n %d.\n", mcnt); + + assert (mcnt >= 0); + /* Originally, this is how many times we HAVE to succeed. */ + if (mcnt > 0) + { + mcnt--; + p += 2; + STORE_NUMBER_AND_INCR (p, mcnt); + DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p, mcnt); + } + else if (mcnt == 0) + { + DEBUG_PRINT2 (" Setting two bytes from 0x%x to no_op.\n", p+2); + p[2] = (unsigned char) no_op; + p[3] = (unsigned char) no_op; + goto on_failure; + } + break; + + case jump_n: + EXTRACT_NUMBER (mcnt, p + 2); + DEBUG_PRINT2 ("EXECUTING jump_n %d.\n", mcnt); + + /* Originally, this is how many times we CAN jump. */ + if (mcnt) + { + mcnt--; + STORE_NUMBER (p + 2, mcnt); + goto unconditional_jump; + } + /* If don't have to jump any more, skip over the rest of command. */ + else + p += 4; + break; + + case set_number_at: + { + DEBUG_PRINT1 ("EXECUTING set_number_at.\n"); + + EXTRACT_NUMBER_AND_INCR (mcnt, p); + p1 = p + mcnt; + EXTRACT_NUMBER_AND_INCR (mcnt, p); + DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p1, mcnt); + STORE_NUMBER (p1, mcnt); + break; + } + + case wordbound: + DEBUG_PRINT1 ("EXECUTING wordbound.\n"); + if (AT_WORD_BOUNDARY (d)) + break; + goto fail; + + case notwordbound: + DEBUG_PRINT1 ("EXECUTING notwordbound.\n"); + if (AT_WORD_BOUNDARY (d)) + goto fail; + break; + + case wordbeg: + DEBUG_PRINT1 ("EXECUTING wordbeg.\n"); + if (WORDCHAR_P (d) && (AT_STRINGS_BEG (d) || !WORDCHAR_P (d - 1))) + break; + goto fail; + + case wordend: + DEBUG_PRINT1 ("EXECUTING wordend.\n"); + if (!AT_STRINGS_BEG (d) && WORDCHAR_P (d - 1) + && (!WORDCHAR_P (d) || AT_STRINGS_END (d))) + break; + goto fail; + +#ifdef emacs +#ifdef emacs19 + case before_dot: + DEBUG_PRINT1 ("EXECUTING before_dot.\n"); + if (PTR_CHAR_POS ((unsigned char *) d) >= point) + goto fail; + break; + + case at_dot: + DEBUG_PRINT1 ("EXECUTING at_dot.\n"); + if (PTR_CHAR_POS ((unsigned char *) d) != point) + goto fail; + break; + + case after_dot: + DEBUG_PRINT1 ("EXECUTING after_dot.\n"); + if (PTR_CHAR_POS ((unsigned char *) d) <= point) + goto fail; + break; +#else /* not emacs19 */ + case at_dot: + DEBUG_PRINT1 ("EXECUTING at_dot.\n"); + if (PTR_CHAR_POS ((unsigned char *) d) + 1 != point) + goto fail; + break; +#endif /* not emacs19 */ + + case syntaxspec: + DEBUG_PRINT2 ("EXECUTING syntaxspec %d.\n", mcnt); + mcnt = *p++; + goto matchsyntax; + + case wordchar: + DEBUG_PRINT1 ("EXECUTING Emacs wordchar.\n"); + mcnt = (int) Sword; + matchsyntax: + PREFETCH (); + if (SYNTAX (*d++) != (enum syntaxcode) mcnt) + goto fail; + SET_REGS_MATCHED (); + break; + + case notsyntaxspec: + DEBUG_PRINT2 ("EXECUTING notsyntaxspec %d.\n", mcnt); + mcnt = *p++; + goto matchnotsyntax; + + case notwordchar: + DEBUG_PRINT1 ("EXECUTING Emacs notwordchar.\n"); + mcnt = (int) Sword; + matchnotsyntax: + PREFETCH (); + if (SYNTAX (*d++) == (enum syntaxcode) mcnt) + goto fail; + SET_REGS_MATCHED (); + break; + +#else /* not emacs */ + case wordchar: + DEBUG_PRINT1 ("EXECUTING non-Emacs wordchar.\n"); + PREFETCH (); + if (!WORDCHAR_P (d)) + goto fail; + SET_REGS_MATCHED (); + d++; + break; + + case notwordchar: + DEBUG_PRINT1 ("EXECUTING non-Emacs notwordchar.\n"); + PREFETCH (); + if (WORDCHAR_P (d)) + goto fail; + SET_REGS_MATCHED (); + d++; + break; +#endif /* not emacs */ + + default: + abort (); + } + continue; /* Successfully executed one pattern command; keep going. */ + + + /* We goto here if a matching operation fails. */ + fail: + if (!FAIL_STACK_EMPTY ()) + { /* A restart point is known. Restore to that state. */ + DEBUG_PRINT1 ("\nFAIL:\n"); + POP_FAILURE_POINT (d, p, + lowest_active_reg, highest_active_reg, + regstart, regend, reg_info); + + /* If this failure point is a dummy, try the next one. */ + if (!p) + goto fail; + + /* If we failed to the end of the pattern, don't examine *p. */ + assert (p <= pend); + if (p < pend) + { + boolean is_a_jump_n = false; + + /* If failed to a backwards jump that's part of a repetition + loop, need to pop this failure point and use the next one. */ + switch ((re_opcode_t) *p) + { + case jump_n: + is_a_jump_n = true; + case maybe_pop_jump: + case pop_failure_jump: + case jump: + p1 = p + 1; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + p1 += mcnt; + + if ((is_a_jump_n && (re_opcode_t) *p1 == succeed_n) + || (!is_a_jump_n + && (re_opcode_t) *p1 == on_failure_jump)) + goto fail; + break; + default: + /* do nothing */ ; + } + } + + if (d >= string1 && d <= end1) + dend = end_match_1; + } + else + break; /* Matching at this starting point really fails. */ + } /* for (;;) */ + + if (best_regs_set) + goto restore_best_regs; + + FREE_VARIABLES (); + + return -1; /* Failure to match. */ +} /* re_match_2 */ + +/* Subroutine definitions for re_match_2. */ + + +/* We are passed P pointing to a register number after a start_memory. + + Return true if the pattern up to the corresponding stop_memory can + match the empty string, and false otherwise. + + If we find the matching stop_memory, sets P to point to one past its number. + Otherwise, sets P to an undefined byte less than or equal to END. + + We don't handle duplicates properly (yet). */ + +static boolean +group_match_null_string_p (p, end, reg_info) + unsigned char **p, *end; + register_info_type *reg_info; +{ + int mcnt; + /* Point to after the args to the start_memory. */ + unsigned char *p1 = *p + 2; + + while (p1 < end) + { + /* Skip over opcodes that can match nothing, and return true or + false, as appropriate, when we get to one that can't, or to the + matching stop_memory. */ + + switch ((re_opcode_t) *p1) + { + /* Could be either a loop or a series of alternatives. */ + case on_failure_jump: + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + + /* If the next operation is not a jump backwards in the + pattern. */ + + if (mcnt >= 0) + { + /* Go through the on_failure_jumps of the alternatives, + seeing if any of the alternatives cannot match nothing. + The last alternative starts with only a jump, + whereas the rest start with on_failure_jump and end + with a jump, e.g., here is the pattern for `a|b|c': + + /on_failure_jump/0/6/exactn/1/a/jump_past_alt/0/6 + /on_failure_jump/0/6/exactn/1/b/jump_past_alt/0/3 + /exactn/1/c + + So, we have to first go through the first (n-1) + alternatives and then deal with the last one separately. */ + + + /* Deal with the first (n-1) alternatives, which start + with an on_failure_jump (see above) that jumps to right + past a jump_past_alt. */ + + while ((re_opcode_t) p1[mcnt-3] == jump_past_alt) + { + /* `mcnt' holds how many bytes long the alternative + is, including the ending `jump_past_alt' and + its number. */ + + if (!alt_match_null_string_p (p1, p1 + mcnt - 3, + reg_info)) + return false; + + /* Move to right after this alternative, including the + jump_past_alt. */ + p1 += mcnt; + + /* Break if it's the beginning of an n-th alternative + that doesn't begin with an on_failure_jump. */ + if ((re_opcode_t) *p1 != on_failure_jump) + break; + + /* Still have to check that it's not an n-th + alternative that starts with an on_failure_jump. */ + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + if ((re_opcode_t) p1[mcnt-3] != jump_past_alt) + { + /* Get to the beginning of the n-th alternative. */ + p1 -= 3; + break; + } + } + + /* Deal with the last alternative: go back and get number + of the `jump_past_alt' just before it. `mcnt' contains + the length of the alternative. */ + EXTRACT_NUMBER (mcnt, p1 - 2); + + if (!alt_match_null_string_p (p1, p1 + mcnt, reg_info)) + return false; + + p1 += mcnt; /* Get past the n-th alternative. */ + } /* if mcnt > 0 */ + break; + + + case stop_memory: + assert (p1[1] == **p); + *p = p1 + 2; + return true; + + + default: + if (!common_op_match_null_string_p (&p1, end, reg_info)) + return false; + } + } /* while p1 < end */ + + return false; +} /* group_match_null_string_p */ + + +/* Similar to group_match_null_string_p, but doesn't deal with alternatives: + It expects P to be the first byte of a single alternative and END one + byte past the last. The alternative can contain groups. */ + +static boolean +alt_match_null_string_p (p, end, reg_info) + unsigned char *p, *end; + register_info_type *reg_info; +{ + int mcnt; + unsigned char *p1 = p; + + while (p1 < end) + { + /* Skip over opcodes that can match nothing, and break when we get + to one that can't. */ + + switch ((re_opcode_t) *p1) + { + /* It's a loop. */ + case on_failure_jump: + p1++; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + p1 += mcnt; + break; + + default: + if (!common_op_match_null_string_p (&p1, end, reg_info)) + return false; + } + } /* while p1 < end */ + + return true; +} /* alt_match_null_string_p */ + + +/* Deals with the ops common to group_match_null_string_p and + alt_match_null_string_p. + + Sets P to one after the op and its arguments, if any. */ + +static boolean +common_op_match_null_string_p (p, end, reg_info) + unsigned char **p, *end; + register_info_type *reg_info; +{ + int mcnt; + boolean ret; + int reg_no; + unsigned char *p1 = *p; + + switch ((re_opcode_t) *p1++) + { + case no_op: + case begline: + case endline: + case begbuf: + case endbuf: + case wordbeg: + case wordend: + case wordbound: + case notwordbound: +#ifdef emacs + case before_dot: + case at_dot: + case after_dot: +#endif + break; + + case start_memory: + reg_no = *p1; + assert (reg_no > 0 && reg_no <= MAX_REGNUM); + ret = group_match_null_string_p (&p1, end, reg_info); + + /* Have to set this here in case we're checking a group which + contains a group and a back reference to it. */ + + if (REG_MATCH_NULL_STRING_P (reg_info[reg_no]) == MATCH_NULL_UNSET_VALUE) + REG_MATCH_NULL_STRING_P (reg_info[reg_no]) = ret; + + if (!ret) + return false; + break; + + /* If this is an optimized succeed_n for zero times, make the jump. */ + case jump: + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + if (mcnt >= 0) + p1 += mcnt; + else + return false; + break; + + case succeed_n: + /* Get to the number of times to succeed. */ + p1 += 2; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + + if (mcnt == 0) + { + p1 -= 4; + EXTRACT_NUMBER_AND_INCR (mcnt, p1); + p1 += mcnt; + } + else + return false; + break; + + case duplicate: + if (!REG_MATCH_NULL_STRING_P (reg_info[*p1])) + return false; + break; + + case set_number_at: + p1 += 4; + + default: + /* All other opcodes mean we cannot match the empty string. */ + return false; + } + + *p = p1; + return true; +} /* common_op_match_null_string_p */ + + +/* Return zero if TRANSLATE[S1] and TRANSLATE[S2] are identical for LEN + bytes; nonzero otherwise. */ + +static int +bcmp_translate (s1, s2, len, translate) + unsigned char *s1, *s2; + register int len; + char *translate; +{ + register unsigned char *p1 = s1, *p2 = s2; + while (len) + { + if (translate[*p1++] != translate[*p2++]) return 1; + len--; + } + return 0; +} + +/* Entry points for GNU code. */ + +/* re_compile_pattern is the GNU regular expression compiler: it + compiles PATTERN (of length SIZE) and puts the result in BUFP. + Returns 0 if the pattern was valid, otherwise an error string. + + Assumes the `allocated' (and perhaps `buffer') and `translate' fields + are set in BUFP on entry. + + We call regex_compile to do the actual compilation. */ + +const char * +re_compile_pattern (pattern, length, bufp) + const char *pattern; + int length; + struct re_pattern_buffer *bufp; +{ + reg_errcode_t ret; + + /* GNU code is written to assume at least RE_NREGS registers will be set + (and at least one extra will be -1). */ + bufp->regs_allocated = REGS_UNALLOCATED; + + /* And GNU code determines whether or not to get register information + by passing null for the REGS argument to re_match, etc., not by + setting no_sub. */ + bufp->no_sub = 0; + + /* Match anchors at newline. */ + bufp->newline_anchor = 1; + + ret = regex_compile (pattern, length, re_syntax_options, bufp); + + return re_error_msg[(int) ret]; +} + +/* Entry points compatible with 4.2 BSD regex library. We don't define + them if this is an Emacs or POSIX compilation. */ + +#if !defined (emacs) && !defined (_POSIX_SOURCE) + +/* BSD has one and only one pattern buffer. */ +static struct re_pattern_buffer re_comp_buf; + +char * +re_comp (s) + const char *s; +{ + reg_errcode_t ret; + + if (!s) + { + if (!re_comp_buf.buffer) + return "No previous regular expression"; + return 0; + } + + if (!re_comp_buf.buffer) + { + re_comp_buf.buffer = (unsigned char *) malloc (200); + if (re_comp_buf.buffer == NULL) + return "Memory exhausted"; + re_comp_buf.allocated = 200; + + re_comp_buf.fastmap = (char *) malloc (1 << BYTEWIDTH); + if (re_comp_buf.fastmap == NULL) + return "Memory exhausted"; + } + + /* Since `re_exec' always passes NULL for the `regs' argument, we + don't need to initialize the pattern buffer fields which affect it. */ + + /* Match anchors at newlines. */ + re_comp_buf.newline_anchor = 1; + + ret = regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf); + + /* Yes, we're discarding `const' here. */ + return (char *) re_error_msg[(int) ret]; +} + + +int +re_exec (s) + const char *s; +{ + const int len = strlen (s); + return + 0 <= re_search (&re_comp_buf, s, len, 0, len, (struct re_registers *) 0); +} +#endif /* not emacs and not _POSIX_SOURCE */ + +/* POSIX.2 functions. Don't define these for Emacs. */ + +#ifndef emacs + +/* regcomp takes a regular expression as a string and compiles it. + + PREG is a regex_t *. We do not expect any fields to be initialized, + since POSIX says we shouldn't. Thus, we set + + `buffer' to the compiled pattern; + `used' to the length of the compiled pattern; + `syntax' to RE_SYNTAX_POSIX_EXTENDED if the + REG_EXTENDED bit in CFLAGS is set; otherwise, to + RE_SYNTAX_POSIX_BASIC; + `newline_anchor' to REG_NEWLINE being set in CFLAGS; + `fastmap' and `fastmap_accurate' to zero; + `re_nsub' to the number of subexpressions in PATTERN. + + PATTERN is the address of the pattern string. + + CFLAGS is a series of bits which affect compilation. + + If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we + use POSIX basic syntax. + + If REG_NEWLINE is set, then . and [^...] don't match newline. + Also, regexec will try a match beginning after every newline. + + If REG_ICASE is set, then we considers upper- and lowercase + versions of letters to be equivalent when matching. + + If REG_NOSUB is set, then when PREG is passed to regexec, that + routine will report only success or failure, and nothing about the + registers. + + It returns 0 if it succeeds, nonzero if it doesn't. (See regex.h for + the return codes and their meanings.) */ + +int +regcomp (preg, pattern, cflags) + regex_t *preg; + const char *pattern; + int cflags; +{ + reg_errcode_t ret; + unsigned syntax + = (cflags & REG_EXTENDED) ? + RE_SYNTAX_POSIX_EXTENDED : RE_SYNTAX_POSIX_BASIC; + + /* regex_compile will allocate the space for the compiled pattern. */ + preg->buffer = 0; + preg->allocated = 0; + + /* Don't bother to use a fastmap when searching. This simplifies the + REG_NEWLINE case: if we used a fastmap, we'd have to put all the + characters after newlines into the fastmap. This way, we just try + every character. */ + preg->fastmap = 0; + + if (cflags & REG_ICASE) + { + unsigned i; + + preg->translate = (char *) malloc (CHAR_SET_SIZE); + if (preg->translate == NULL) + return (int) REG_ESPACE; + + /* Map uppercase characters to corresponding lowercase ones. */ + for (i = 0; i < CHAR_SET_SIZE; i++) + preg->translate[i] = ISUPPER (i) ? tolower (i) : i; + } + else + preg->translate = NULL; + + /* If REG_NEWLINE is set, newlines are treated differently. */ + if (cflags & REG_NEWLINE) + { /* REG_NEWLINE implies neither . nor [^...] match newline. */ + syntax &= ~RE_DOT_NEWLINE; + syntax |= RE_HAT_LISTS_NOT_NEWLINE; + /* It also changes the matching behavior. */ + preg->newline_anchor = 1; + } + else + preg->newline_anchor = 0; + + preg->no_sub = !!(cflags & REG_NOSUB); + + /* POSIX says a null character in the pattern terminates it, so we + can use strlen here in compiling the pattern. */ + ret = regex_compile (pattern, strlen (pattern), syntax, preg); + + /* POSIX doesn't distinguish between an unmatched open-group and an + unmatched close-group: both are REG_EPAREN. */ + if (ret == REG_ERPAREN) ret = REG_EPAREN; + + return (int) ret; +} + + +/* regexec searches for a given pattern, specified by PREG, in the + string STRING. + + If NMATCH is zero or REG_NOSUB was set in the cflags argument to + `regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at + least NMATCH elements, and we set them to the offsets of the + corresponding matched substrings. + + EFLAGS specifies `execution flags' which affect matching: if + REG_NOTBOL is set, then ^ does not match at the beginning of the + string; if REG_NOTEOL is set, then $ does not match at the end. + + We return 0 if we find a match and REG_NOMATCH if not. */ + +int +regexec (preg, string, nmatch, pmatch, eflags) + const regex_t *preg; + const char *string; + size_t nmatch; + regmatch_t pmatch[]; + int eflags; +{ + int ret; + struct re_registers regs; + regex_t private_preg; + int len = strlen (string); + boolean want_reg_info = !preg->no_sub && nmatch > 0; + + private_preg = *preg; + + private_preg.not_bol = !!(eflags & REG_NOTBOL); + private_preg.not_eol = !!(eflags & REG_NOTEOL); + + /* The user has told us exactly how many registers to return + information about, via `nmatch'. We have to pass that on to the + matching routines. */ + private_preg.regs_allocated = REGS_FIXED; + + if (want_reg_info) + { + regs.num_regs = nmatch; + regs.start = TALLOC (nmatch, regoff_t); + regs.end = TALLOC (nmatch, regoff_t); + if (regs.start == NULL || regs.end == NULL) + return (int) REG_NOMATCH; + } + + /* Perform the searching operation. */ + ret = re_search (&private_preg, string, len, + /* start: */ 0, /* range: */ len, + want_reg_info ? ®s : (struct re_registers *) 0); + + /* Copy the register information to the POSIX structure. */ + if (want_reg_info) + { + if (ret >= 0) + { + unsigned r; + + for (r = 0; r < nmatch; r++) + { + pmatch[r].rm_so = regs.start[r]; + pmatch[r].rm_eo = regs.end[r]; + } + } + + /* If we needed the temporary register info, free the space now. */ + free (regs.start); + free (regs.end); + } + + /* We want zero return to mean success, unlike `re_search'. */ + return ret >= 0 ? (int) REG_NOERROR : (int) REG_NOMATCH; +} + + +/* Returns a message corresponding to an error code, ERRCODE, returned + from either regcomp or regexec. We don't use PREG here. */ + +size_t +regerror (errcode, preg, errbuf, errbuf_size) + int errcode; + const regex_t *preg; + char *errbuf; + size_t errbuf_size; +{ + const char *msg; + size_t msg_size; + + if (errcode < 0 + || errcode >= (sizeof (re_error_msg) / sizeof (re_error_msg[0]))) + /* Only error codes returned by the rest of the code should be passed + to this routine. If we are given anything else, or if other regex + code generates an invalid error code, then the program has a bug. + Dump core so we can fix it. */ + abort (); + + msg_size = strlen (msg) + 1; /* Includes the null. */ + + if (errbuf_size != 0) + { + if (msg_size > errbuf_size) + { + strncpy (errbuf, msg, errbuf_size - 1); + errbuf[errbuf_size - 1] = 0; + } + else + strcpy (errbuf, msg); + } + + return msg_size; +} + + +/* Free dynamically allocated space used by PREG. */ + +void +regfree (preg) + regex_t *preg; +{ + if (preg->buffer != NULL) + free (preg->buffer); + preg->buffer = NULL; + + preg->allocated = 0; + preg->used = 0; + + if (preg->fastmap != NULL) + free (preg->fastmap); + preg->fastmap = NULL; + preg->fastmap_accurate = 0; + + if (preg->translate != NULL) + free (preg->translate); + preg->translate = NULL; +} + +#endif /* not emacs */ + +/* +Local variables: +make-backup-files: t +version-control: t +trim-versions-without-asking: nil +End: +*/ diff --git a/gnu/usr.bin/tar/regex.h b/gnu/usr.bin/tar/regex.h new file mode 100644 index 000000000000..0840861da369 --- /dev/null +++ b/gnu/usr.bin/tar/regex.h @@ -0,0 +1,490 @@ +/* Definitions for data structures and routines for the regular + expression library, version 0.11. + + Copyright (C) 1985, 89, 90, 91, 92 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifndef __REGEXP_LIBRARY_H__ +#define __REGEXP_LIBRARY_H__ + +/* POSIX says that <sys/types.h> must be included (by the caller) before + <regex.h>. */ + +#ifdef VMS +/* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it + should be there. */ +#include <stddef.h> +#endif + + +/* The following bits are used to determine the regexp syntax we + recognize. The set/not-set meanings are chosen so that Emacs syntax + remains the value 0. The bits are given in alphabetical order, and + the definitions shifted by one from the previous bit; thus, when we + add or remove a bit, only one other definition need change. */ +typedef unsigned reg_syntax_t; + +/* If this bit is not set, then \ inside a bracket expression is literal. + If set, then such a \ quotes the following character. */ +#define RE_BACKSLASH_ESCAPE_IN_LISTS (1) + +/* If this bit is not set, then + and ? are operators, and \+ and \? are + literals. + If set, then \+ and \? are operators and + and ? are literals. */ +#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) + +/* If this bit is set, then character classes are supported. They are: + [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], + [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. + If not set, then character classes are not supported. */ +#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) + +/* If this bit is set, then ^ and $ are always anchors (outside bracket + expressions, of course). + If this bit is not set, then it depends: + ^ is an anchor if it is at the beginning of a regular + expression or after an open-group or an alternation operator; + $ is an anchor if it is at the end of a regular expression, or + before a close-group or an alternation operator. + + This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because + POSIX draft 11.2 says that * etc. in leading positions is undefined. + We already implemented a previous draft which made those constructs + invalid, though, so we haven't changed the code back. */ +#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) + +/* If this bit is set, then special characters are always special + regardless of where they are in the pattern. + If this bit is not set, then special characters are special only in + some contexts; otherwise they are ordinary. Specifically, + * + ? and intervals are only special when not after the beginning, + open-group, or alternation operator. */ +#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) + +/* If this bit is set, then *, +, ?, and { cannot be first in an re or + immediately after an alternation or begin-group operator. */ +#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) + +/* If this bit is set, then . matches newline. + If not set, then it doesn't. */ +#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) + +/* If this bit is set, then . doesn't match NUL. + If not set, then it does. */ +#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) + +/* If this bit is set, nonmatching lists [^...] do not match newline. + If not set, they do. */ +#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) + +/* If this bit is set, either \{...\} or {...} defines an + interval, depending on RE_NO_BK_BRACES. + If not set, \{, \}, {, and } are literals. */ +#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) + +/* If this bit is set, +, ? and | aren't recognized as operators. + If not set, they are. */ +#define RE_LIMITED_OPS (RE_INTERVALS << 1) + +/* If this bit is set, newline is an alternation operator. + If not set, newline is literal. */ +#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) + +/* If this bit is set, then `{...}' defines an interval, and \{ and \} + are literals. + If not set, then `\{...\}' defines an interval. */ +#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) + +/* If this bit is set, (...) defines a group, and \( and \) are literals. + If not set, \(...\) defines a group, and ( and ) are literals. */ +#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) + +/* If this bit is set, then \<digit> matches <digit>. + If not set, then \<digit> is a back-reference. */ +#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) + +/* If this bit is set, then | is an alternation operator, and \| is literal. + If not set, then \| is an alternation operator, and | is literal. */ +#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) + +/* If this bit is set, then an ending range point collating higher + than the starting range point, as in [z-a], is invalid. + If not set, then when ending range point collates higher than the + starting range point, the range is ignored. */ +#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) + +/* If this bit is set, then an unmatched ) is ordinary. + If not set, then an unmatched ) is invalid. */ +#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) + +/* This global variable defines the particular regexp syntax to use (for + some interfaces). When a regexp is compiled, the syntax used is + stored in the pattern buffer, so changing this does not affect + already-compiled regexps. */ +extern reg_syntax_t re_syntax_options; + +/* Define combinations of the above bits for the standard possibilities. + (The [[[ comments delimit what gets put into the Texinfo file, so + don't delete them!) */ +/* [[[begin syntaxes]]] */ +#define RE_SYNTAX_EMACS 0 + +#define RE_SYNTAX_AWK \ + (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ + | RE_UNMATCHED_RIGHT_PAREN_ORD) + +#define RE_SYNTAX_POSIX_AWK \ + (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS) + +#define RE_SYNTAX_GREP \ + (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ + | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ + | RE_NEWLINE_ALT) + +#define RE_SYNTAX_EGREP \ + (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ + | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ + | RE_NO_BK_VBAR) + +#define RE_SYNTAX_POSIX_EGREP \ + (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES) + +/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ +#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC + +#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC + +/* Syntax bits common to both basic and extended POSIX regex syntax. */ +#define _RE_SYNTAX_POSIX_COMMON \ + (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ + | RE_INTERVALS | RE_NO_EMPTY_RANGES) + +#define RE_SYNTAX_POSIX_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM) + +/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes + RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this + isn't minimal, since other operators, such as \`, aren't disabled. */ +#define RE_SYNTAX_POSIX_MINIMAL_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) + +#define RE_SYNTAX_POSIX_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ + | RE_UNMATCHED_RIGHT_PAREN_ORD) + +/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INVALID_OPS + replaces RE_CONTEXT_INDEP_OPS and RE_NO_BK_REFS is added. */ +#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) +/* [[[end syntaxes]]] */ + +/* Maximum number of duplicates an interval can allow. Some systems + (erroneously) define this in other header files, but we want our + value, so remove any previous define. */ +#ifdef RE_DUP_MAX +#undef RE_DUP_MAX +#endif +#define RE_DUP_MAX ((1 << 15) - 1) + + +/* POSIX `cflags' bits (i.e., information for `regcomp'). */ + +/* If this bit is set, then use extended regular expression syntax. + If not set, then use basic regular expression syntax. */ +#define REG_EXTENDED 1 + +/* If this bit is set, then ignore case when matching. + If not set, then case is significant. */ +#define REG_ICASE (REG_EXTENDED << 1) + +/* If this bit is set, then anchors do not match at newline + characters in the string. + If not set, then anchors do match at newlines. */ +#define REG_NEWLINE (REG_ICASE << 1) + +/* If this bit is set, then report only success or fail in regexec. + If not set, then returns differ between not matching and errors. */ +#define REG_NOSUB (REG_NEWLINE << 1) + + +/* POSIX `eflags' bits (i.e., information for regexec). */ + +/* If this bit is set, then the beginning-of-line operator doesn't match + the beginning of the string (presumably because it's not the + beginning of a line). + If not set, then the beginning-of-line operator does match the + beginning of the string. */ +#define REG_NOTBOL 1 + +/* Like REG_NOTBOL, except for the end-of-line. */ +#define REG_NOTEOL (1 << 1) + + +/* If any error codes are removed, changed, or added, update the + `re_error_msg' table in regex.c. */ +typedef enum +{ + REG_NOERROR = 0, /* Success. */ + REG_NOMATCH, /* Didn't find a match (for regexec). */ + + /* POSIX regcomp return error codes. (In the order listed in the + standard.) */ + REG_BADPAT, /* Invalid pattern. */ + REG_ECOLLATE, /* Not implemented. */ + REG_ECTYPE, /* Invalid character class name. */ + REG_EESCAPE, /* Trailing backslash. */ + REG_ESUBREG, /* Invalid back reference. */ + REG_EBRACK, /* Unmatched left bracket. */ + REG_EPAREN, /* Parenthesis imbalance. */ + REG_EBRACE, /* Unmatched \{. */ + REG_BADBR, /* Invalid contents of \{\}. */ + REG_ERANGE, /* Invalid range end. */ + REG_ESPACE, /* Ran out of memory. */ + REG_BADRPT, /* No preceding re for repetition op. */ + + /* Error codes we've added. */ + REG_EEND, /* Premature end. */ + REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ + REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ +} reg_errcode_t; + +/* This data structure represents a compiled pattern. Before calling + the pattern compiler, the fields `buffer', `allocated', `fastmap', + `translate', and `no_sub' can be set. After the pattern has been + compiled, the `re_nsub' field is available. All other fields are + private to the regex routines. */ + +struct re_pattern_buffer +{ +/* [[[begin pattern_buffer]]] */ + /* Space that holds the compiled pattern. It is declared as + `unsigned char *' because its elements are + sometimes used as array indexes. */ + unsigned char *buffer; + + /* Number of bytes to which `buffer' points. */ + unsigned long allocated; + + /* Number of bytes actually used in `buffer'. */ + unsigned long used; + + /* Syntax setting with which the pattern was compiled. */ + reg_syntax_t syntax; + + /* Pointer to a fastmap, if any, otherwise zero. re_search uses + the fastmap, if there is one, to skip over impossible + starting points for matches. */ + char *fastmap; + + /* Either a translate table to apply to all characters before + comparing them, or zero for no translation. The translation + is applied to a pattern when it is compiled and to a string + when it is matched. */ + char *translate; + + /* Number of subexpressions found by the compiler. */ + size_t re_nsub; + + /* Zero if this pattern cannot match the empty string, one else. + Well, in truth it's used only in `re_search_2', to see + whether or not we should use the fastmap, so we don't set + this absolutely perfectly; see `re_compile_fastmap' (the + `duplicate' case). */ + unsigned can_be_null : 1; + + /* If REGS_UNALLOCATED, allocate space in the `regs' structure + for `max (RE_NREGS, re_nsub + 1)' groups. + If REGS_REALLOCATE, reallocate space if necessary. + If REGS_FIXED, use what's there. */ +#define REGS_UNALLOCATED 0 +#define REGS_REALLOCATE 1 +#define REGS_FIXED 2 + unsigned regs_allocated : 2; + + /* Set to zero when `regex_compile' compiles a pattern; set to one + by `re_compile_fastmap' if it updates the fastmap. */ + unsigned fastmap_accurate : 1; + + /* If set, `re_match_2' does not return information about + subexpressions. */ + unsigned no_sub : 1; + + /* If set, a beginning-of-line anchor doesn't match at the + beginning of the string. */ + unsigned not_bol : 1; + + /* Similarly for an end-of-line anchor. */ + unsigned not_eol : 1; + + /* If true, an anchor at a newline matches. */ + unsigned newline_anchor : 1; + +/* [[[end pattern_buffer]]] */ +}; + +typedef struct re_pattern_buffer regex_t; + + +/* search.c (search_buffer) in Emacs needs this one opcode value. It is + defined both in `regex.c' and here. */ +#define RE_EXACTN_VALUE 1 + +/* Type for byte offsets within the string. POSIX mandates this. */ +typedef int regoff_t; + + +/* This is the structure we store register match data in. See + regex.texinfo for a full description of what registers match. */ +struct re_registers +{ + unsigned num_regs; + regoff_t *start; + regoff_t *end; +}; + + +/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, + `re_match_2' returns information about at least this many registers + the first time a `regs' structure is passed. */ +#ifndef RE_NREGS +#define RE_NREGS 30 +#endif + + +/* POSIX specification for registers. Aside from the different names than + `re_registers', POSIX uses an array of structures, instead of a + structure of arrays. */ +typedef struct +{ + regoff_t rm_so; /* Byte offset from string's start to substring's start. */ + regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ +} regmatch_t; + +/* Declarations for routines. */ + +/* To avoid duplicating every routine declaration -- once with a + prototype (if we are ANSI), and once without (if we aren't) -- we + use the following macro to declare argument types. This + unfortunately clutters up the declarations a bit, but I think it's + worth it. */ + +#if __STDC__ + +#define _RE_ARGS(args) args + +#else /* not __STDC__ */ + +#define _RE_ARGS(args) () + +#endif /* not __STDC__ */ + +/* Sets the current default syntax to SYNTAX, and return the old syntax. + You can also simply assign to the `re_syntax_options' variable. */ +extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); + +/* Compile the regular expression PATTERN, with length LENGTH + and syntax given by the global `re_syntax_options', into the buffer + BUFFER. Return NULL if successful, and an error string if not. */ +extern const char *re_compile_pattern + _RE_ARGS ((const char *pattern, int length, + struct re_pattern_buffer *buffer)); + + +/* Compile a fastmap for the compiled pattern in BUFFER; used to + accelerate searches. Return 0 if successful and -2 if was an + internal error. */ +extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); + + +/* Search in the string STRING (with length LENGTH) for the pattern + compiled into BUFFER. Start searching at position START, for RANGE + characters. Return the starting position of the match, -1 for no + match, or -2 for an internal error. Also return register + information in REGS (if REGS and BUFFER->no_sub are nonzero). */ +extern int re_search + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, + int length, int start, int range, struct re_registers *regs)); + + +/* Like `re_search', but search in the concatenation of STRING1 and + STRING2. Also, stop searching at index START + STOP. */ +extern int re_search_2 + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, + int length1, const char *string2, int length2, + int start, int range, struct re_registers *regs, int stop)); + + +/* Like `re_search', but return how many characters in STRING the regexp + in BUFFER matched, starting at position START. */ +extern int re_match + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, + int length, int start, struct re_registers *regs)); + + +/* Relates to `re_match' as `re_search_2' relates to `re_search'. */ +extern int re_match_2 + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, + int length1, const char *string2, int length2, + int start, struct re_registers *regs, int stop)); + + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using BUFFER and REGS will use this memory + for recording register information. STARTS and ENDS must be + allocated with malloc, and must each be at least `NUM_REGS * sizeof + (regoff_t)' bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ +extern void re_set_registers + _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, + unsigned num_regs, regoff_t *starts, regoff_t *ends)); + +/* 4.2 bsd compatibility. */ +extern char *re_comp _RE_ARGS ((const char *)); +extern int re_exec _RE_ARGS ((const char *)); + +/* POSIX compatibility. */ +extern int regcomp _RE_ARGS ((regex_t *preg, const char *pattern, int cflags)); +extern int regexec + _RE_ARGS ((const regex_t *preg, const char *string, size_t nmatch, + regmatch_t pmatch[], int eflags)); +extern size_t regerror + _RE_ARGS ((int errcode, const regex_t *preg, char *errbuf, + size_t errbuf_size)); +extern void regfree _RE_ARGS ((regex_t *preg)); + +#endif /* not __REGEXP_LIBRARY_H__ */ + +/* +Local variables: +make-backup-files: t +version-control: t +trim-versions-without-asking: nil +End: +*/ diff --git a/gnu/usr.bin/tar/rmt.h b/gnu/usr.bin/tar/rmt.h new file mode 100644 index 000000000000..2155223954c3 --- /dev/null +++ b/gnu/usr.bin/tar/rmt.h @@ -0,0 +1,98 @@ +/* Definitions for communicating with a remote tape drive. + Copyright (C) 1988, 1992 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +#ifdef HAVE_UNISTD_H +#include <unistd.h> +#endif + +#if !defined(_POSIX_VERSION) +#ifdef __MSDOS__ +#include <io.h> +#else /* !__MSDOS__ */ +extern off_t lseek (); +#endif /* __MSDOS__ */ +#endif /* _POSIX_VERSION */ + +#ifdef NO_REMOTE +#define _isrmt(f) 0 +#define rmtopen open +#define rmtaccess access +#define rmtstat stat +#define rmtcreat creat +#define rmtlstat lstat +#define rmtread read +#define rmtwrite write +#define rmtlseek lseek +#define rmtclose close +#define rmtioctl ioctl +#define rmtdup dup +#define rmtfstat fstat +#define rmtfcntl fcntl +#define rmtisatty isatty + +#else /* !NO_REMOTE */ + +#define __REM_BIAS 128 +#define RMTIOCTL + +#ifndef O_CREAT +#define O_CREAT 01000 +#endif + +extern char *__rmt_path; + +#if defined(STDC_HEADERS) || defined(HAVE_STRING_H) +#include <string.h> +#ifndef index +#define index strchr +#endif +#else +extern char *index (); +#endif + +#define _remdev(path) (!f_force_local && (__rmt_path=index(path, ':'))) +#define _isrmt(fd) ((fd) >= __REM_BIAS) + +#define rmtopen(path,oflag,mode) (_remdev(path) ? __rmt_open(path, oflag, mode, __REM_BIAS) : open(path, oflag, mode)) +#define rmtaccess(path, amode) (_remdev(path) ? 0 : access(path, amode)) +#define rmtstat(path, buf) (_remdev(path) ? (errno = EOPNOTSUPP), -1 : stat(path, buf)) +#define rmtcreat(path, mode) (_remdev(path) ? __rmt_open (path, 1 | O_CREAT, mode, __REM_BIAS) : creat(path, mode)) +#define rmtlstat(path,buf) (_remdev(path) ? (errno = EOPNOTSUPP), -1 : lstat(path,buf)) + +#define rmtread(fd, buf, n) (_isrmt(fd) ? __rmt_read(fd - __REM_BIAS, buf, n) : read(fd, buf, n)) +#define rmtwrite(fd, buf, n) (_isrmt(fd) ? __rmt_write(fd - __REM_BIAS, buf, n) : write(fd, buf, n)) +#define rmtlseek(fd, off, wh) (_isrmt(fd) ? __rmt_lseek(fd - __REM_BIAS, off, wh) : lseek(fd, off, wh)) +#define rmtclose(fd) (_isrmt(fd) ? __rmt_close(fd - __REM_BIAS) : close(fd)) +#ifdef RMTIOCTL +#define rmtioctl(fd,req,arg) (_isrmt(fd) ? __rmt_ioctl(fd - __REM_BIAS, req, arg) : ioctl(fd, req, arg)) +#else +#define rmtioctl(fd,req,arg) (_isrmt(fd) ? (errno = EOPNOTSUPP), -1 : ioctl(fd, req, arg)) +#endif +#define rmtdup(fd) (_isrmt(fd) ? (errno = EOPNOTSUPP), -1 : dup(fd)) +#define rmtfstat(fd, buf) (_isrmt(fd) ? (errno = EOPNOTSUPP), -1 : fstat(fd, buf)) +#define rmtfcntl(fd,cmd,arg) (_isrmt(fd) ? (errno = EOPNOTSUPP), -1 : fcntl (fd, cmd, arg)) +#define rmtisatty(fd) (_isrmt(fd) ? 0 : isatty(fd)) + +#undef RMTIOCTL + +int __rmt_open (); +int __rmt_close (); +int __rmt_read (); +int __rmt_write (); +long __rmt_lseek (); +int __rmt_ioctl (); +#endif /* !NO_REMOTE */ diff --git a/gnu/usr.bin/tar/rtapelib.c b/gnu/usr.bin/tar/rtapelib.c new file mode 100644 index 000000000000..eece76ffcd0f --- /dev/null +++ b/gnu/usr.bin/tar/rtapelib.c @@ -0,0 +1,582 @@ +/* Functions for communicating with a remote tape drive. + Copyright (C) 1988, 1992 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* The man page rmt(8) for /etc/rmt documents the remote mag tape + protocol which rdump and rrestore use. Unfortunately, the man + page is *WRONG*. The author of the routines I'm including originally + wrote his code just based on the man page, and it didn't work, so he + went to the rdump source to figure out why. The only thing he had to + change was to check for the 'F' return code in addition to the 'E', + and to separate the various arguments with \n instead of a space. I + personally don't think that this is much of a problem, but I wanted to + point it out. -- Arnold Robbins + + Originally written by Jeff Lee, modified some by Arnold Robbins. + Redone as a library that can replace open, read, write, etc., by + Fred Fish, with some additional work by Arnold Robbins. + Modified to make all rmtXXX calls into macros for speed by Jay Fenlason. + Use -DHAVE_NETDB_H for rexec code, courtesy of Dan Kegel, srs!dan. */ + +#include <stdio.h> +#include <sys/types.h> +#include <signal.h> + +#ifdef HAVE_SYS_MTIO_H +#include <sys/ioctl.h> +#include <sys/mtio.h> +#endif + +#ifdef HAVE_NETDB_H +#include <netdb.h> +#endif + +#include <errno.h> +#include <setjmp.h> +#include <sys/stat.h> + +#ifndef errno +extern int errno; +#endif + +#ifdef HAVE_UNISTD_H +#include <unistd.h> +#endif +#ifdef STDC_HEADERS +#include <string.h> +#include <stdlib.h> +#endif + +/* Maximum size of a fully qualified host name. */ +#define MAXHOSTLEN 257 + +/* Size of buffers for reading and writing commands to rmt. + (An arbitrary limit.) */ +#define CMDBUFSIZE 64 + +#ifndef RETSIGTYPE +#define RETSIGTYPE void +#endif + +/* Maximum number of simultaneous remote tape connections. + (Another arbitrary limit.) */ +#define MAXUNIT 4 + +/* Return the parent's read side of remote tape connection FILDES. */ +#define READ(fildes) (from_rmt[fildes][0]) + +/* Return the parent's write side of remote tape connection FILDES. */ +#define WRITE(fildes) (to_rmt[fildes][1]) + +/* The pipes for receiving data from remote tape drives. */ +static int from_rmt[MAXUNIT][2] = +{-1, -1, -1, -1, -1, -1, -1, -1}; + +/* The pipes for sending data to remote tape drives. */ +static int to_rmt[MAXUNIT][2] = +{-1, -1, -1, -1, -1, -1, -1, -1}; + +/* Temporary variable used by macros in rmt.h. */ +char *__rmt_path; + +/* Close remote tape connection FILDES. */ + +static void +_rmt_shutdown (fildes) + int fildes; +{ + close (READ (fildes)); + close (WRITE (fildes)); + READ (fildes) = -1; + WRITE (fildes) = -1; +} + +/* Attempt to perform the remote tape command specified in BUF + on remote tape connection FILDES. + Return 0 if successful, -1 on error. */ + +static int +command (fildes, buf) + int fildes; + char *buf; +{ + register int buflen; + RETSIGTYPE (*pipe_handler) (); + + /* Save the current pipe handler and try to make the request. */ + + pipe_handler = signal (SIGPIPE, SIG_IGN); + buflen = strlen (buf); + if (write (WRITE (fildes), buf, buflen) == buflen) + { + signal (SIGPIPE, pipe_handler); + return 0; + } + + /* Something went wrong. Close down and go home. */ + + signal (SIGPIPE, pipe_handler); + _rmt_shutdown (fildes); + errno = EIO; + return -1; +} + +/* Read and return the status from remote tape connection FILDES. + If an error occurred, return -1 and set errno. */ + +static int +status (fildes) + int fildes; +{ + int i; + char c, *cp; + char buffer[CMDBUFSIZE]; + + /* Read the reply command line. */ + + for (i = 0, cp = buffer; i < CMDBUFSIZE; i++, cp++) + { + if (read (READ (fildes), cp, 1) != 1) + { + _rmt_shutdown (fildes); + errno = EIO; + return -1; + } + if (*cp == '\n') + { + *cp = '\0'; + break; + } + } + + if (i == CMDBUFSIZE) + { + _rmt_shutdown (fildes); + errno = EIO; + return -1; + } + + /* Check the return status. */ + + for (cp = buffer; *cp; cp++) + if (*cp != ' ') + break; + + if (*cp == 'E' || *cp == 'F') + { + errno = atoi (cp + 1); + /* Skip the error message line. */ + while (read (READ (fildes), &c, 1) == 1) + if (c == '\n') + break; + + if (*cp == 'F') + _rmt_shutdown (fildes); + + return -1; + } + + /* Check for mis-synced pipes. */ + + if (*cp != 'A') + { + _rmt_shutdown (fildes); + errno = EIO; + return -1; + } + + /* Got an `A' (success) response. */ + return atoi (cp + 1); +} + +#ifdef HAVE_NETDB_H +/* Execute /etc/rmt as user USER on remote system HOST using rexec. + Return a file descriptor of a bidirectional socket for stdin and stdout. + If USER is NULL, or an empty string, use the current username. + + By default, this code is not used, since it requires that + the user have a .netrc file in his/her home directory, or that the + application designer be willing to have rexec prompt for login and + password info. This may be unacceptable, and .rhosts files for use + with rsh are much more common on BSD systems. */ + +static int +_rmt_rexec (host, user) + char *host; + char *user; +{ + struct servent *rexecserv; + int save_stdin = dup (fileno (stdin)); + int save_stdout = dup (fileno (stdout)); + int tape_fd; /* Return value. */ + + /* When using cpio -o < filename, stdin is no longer the tty. + But the rexec subroutine reads the login and the passwd on stdin, + to allow remote execution of the command. + So, reopen stdin and stdout on /dev/tty before the rexec and + give them back their original value after. */ + if (freopen ("/dev/tty", "r", stdin) == NULL) + freopen ("/dev/null", "r", stdin); + if (freopen ("/dev/tty", "w", stdout) == NULL) + freopen ("/dev/null", "w", stdout); + + rexecserv = getservbyname ("exec", "tcp"); + if (NULL == rexecserv) + { + fprintf (stderr, "exec/tcp: service not available"); + exit (1); + } + if (user != NULL && *user == '\0') + user = NULL; + tape_fd = rexec (&host, rexecserv->s_port, user, NULL, + "/etc/rmt", (int *) NULL); + fclose (stdin); + fdopen (save_stdin, "r"); + fclose (stdout); + fdopen (save_stdout, "w"); + + return tape_fd; +} + +#endif /* HAVE_NETDB_H */ + +/* Open a magtape device on the system specified in PATH, as the given user. + PATH has the form `[user@]system:/dev/????'. + If COMPAT is defined, it can also have the form `system[.user]:/dev/????'. + + OFLAG is O_RDONLY, O_WRONLY, etc. + MODE is ignored; 0666 is always used. + + If successful, return the remote tape pipe number plus BIAS. + On error, return -1. */ + +int +__rmt_open (path, oflag, mode, bias) + char *path; + int oflag; + int mode; + int bias; +{ + int i, rc; + char buffer[CMDBUFSIZE]; /* Command buffer. */ + char system[MAXHOSTLEN]; /* The remote host name. */ + char device[CMDBUFSIZE]; /* The remote device name. */ + char login[CMDBUFSIZE]; /* The remote user name. */ + char *sys, *dev, *user; /* For copying into the above buffers. */ + + sys = system; + dev = device; + user = login; + + /* Find an unused pair of file descriptors. */ + + for (i = 0; i < MAXUNIT; i++) + if (READ (i) == -1 && WRITE (i) == -1) + break; + + if (i == MAXUNIT) + { + errno = EMFILE; + return -1; + } + + /* Pull apart the system and device, and optional user. + Don't munge the original string. */ + + while (*path != '@' +#ifdef COMPAT + && *path != '.' +#endif + && *path != ':') + { + *sys++ = *path++; + } + *sys = '\0'; + path++; + + if (*(path - 1) == '@') + { + /* Saw user part of user@host. Start over. */ + strcpy (user, system); + sys = system; + while (*path != ':') + { + *sys++ = *path++; + } + *sys = '\0'; + path++; + } +#ifdef COMPAT + else if (*(path - 1) == '.') + { + while (*path != ':') + { + *user++ = *path++; + } + *user = '\0'; + path++; + } +#endif + else + *user = '\0'; + + while (*path) + { + *dev++ = *path++; + } + *dev = '\0'; + +#ifdef HAVE_NETDB_H + /* Execute the remote command using rexec. */ + READ (i) = WRITE (i) = _rmt_rexec (system, login); + if (READ (i) < 0) + return -1; +#else /* !HAVE_NETDB_H */ + /* Set up the pipes for the `rsh' command, and fork. */ + + if (pipe (to_rmt[i]) == -1 || pipe (from_rmt[i]) == -1) + return -1; + + rc = fork (); + if (rc == -1) + return -1; + + if (rc == 0) + { + /* Child. */ + close (0); + dup (to_rmt[i][0]); + close (to_rmt[i][0]); + close (to_rmt[i][1]); + + close (1); + dup (from_rmt[i][1]); + close (from_rmt[i][0]); + close (from_rmt[i][1]); + + setuid (getuid ()); + setgid (getgid ()); + + if (*login) + { + execl ("/usr/ucb/rsh", "rsh", system, "-l", login, + "/etc/rmt", (char *) 0); + execl ("/usr/bin/remsh", "remsh", system, "-l", login, + "/etc/rmt", (char *) 0); + execl ("/usr/bin/rsh", "rsh", system, "-l", login, + "/etc/rmt", (char *) 0); + execl ("/usr/bsd/rsh", "rsh", system, "-l", login, + "/etc/rmt", (char *) 0); + execl ("/usr/bin/nsh", "nsh", system, "-l", login, + "/etc/rmt", (char *) 0); + } + else + { + execl ("/usr/ucb/rsh", "rsh", system, + "/etc/rmt", (char *) 0); + execl ("/usr/bin/remsh", "remsh", system, + "/etc/rmt", (char *) 0); + execl ("/usr/bin/rsh", "rsh", system, + "/etc/rmt", (char *) 0); + execl ("/usr/bsd/rsh", "rsh", system, + "/etc/rmt", (char *) 0); + execl ("/usr/bin/nsh", "nsh", system, + "/etc/rmt", (char *) 0); + } + + /* Bad problems if we get here. */ + + perror ("cannot execute remote shell"); + _exit (1); + } + + /* Parent. */ + close (to_rmt[i][0]); + close (from_rmt[i][1]); +#endif /* !HAVE_NETDB_H */ + + /* Attempt to open the tape device. */ + + sprintf (buffer, "O%s\n%d\n", device, oflag); + if (command (i, buffer) == -1 || status (i) == -1) + return -1; + + return i + bias; +} + +/* Close remote tape connection FILDES and shut down. + Return 0 if successful, -1 on error. */ + +int +__rmt_close (fildes) + int fildes; +{ + int rc; + + if (command (fildes, "C\n") == -1) + return -1; + + rc = status (fildes); + _rmt_shutdown (fildes); + return rc; +} + +/* Read up to NBYTE bytes into BUF from remote tape connection FILDES. + Return the number of bytes read on success, -1 on error. */ + +int +__rmt_read (fildes, buf, nbyte) + int fildes; + char *buf; + unsigned int nbyte; +{ + int rc, i; + char buffer[CMDBUFSIZE]; + + sprintf (buffer, "R%d\n", nbyte); + if (command (fildes, buffer) == -1 || (rc = status (fildes)) == -1) + return -1; + + for (i = 0; i < rc; i += nbyte, buf += nbyte) + { + nbyte = read (READ (fildes), buf, rc - i); + if (nbyte <= 0) + { + _rmt_shutdown (fildes); + errno = EIO; + return -1; + } + } + + return rc; +} + +/* Write NBYTE bytes from BUF to remote tape connection FILDES. + Return the number of bytes written on success, -1 on error. */ + +int +__rmt_write (fildes, buf, nbyte) + int fildes; + char *buf; + unsigned int nbyte; +{ + char buffer[CMDBUFSIZE]; + RETSIGTYPE (*pipe_handler) (); + + sprintf (buffer, "W%d\n", nbyte); + if (command (fildes, buffer) == -1) + return -1; + + pipe_handler = signal (SIGPIPE, SIG_IGN); + if (write (WRITE (fildes), buf, nbyte) == nbyte) + { + signal (SIGPIPE, pipe_handler); + return status (fildes); + } + + /* Write error. */ + signal (SIGPIPE, pipe_handler); + _rmt_shutdown (fildes); + errno = EIO; + return -1; +} + +/* Perform an imitation lseek operation on remote tape connection FILDES. + Return the new file offset if successful, -1 if on error. */ + +long +__rmt_lseek (fildes, offset, whence) + int fildes; + long offset; + int whence; +{ + char buffer[CMDBUFSIZE]; + + sprintf (buffer, "L%ld\n%d\n", offset, whence); + if (command (fildes, buffer) == -1) + return -1; + + return status (fildes); +} + +/* Perform a raw tape operation on remote tape connection FILDES. + Return the results of the ioctl, or -1 on error. */ + +#ifdef MTIOCTOP +int +__rmt_ioctl (fildes, op, arg) + int fildes, op; + char *arg; +{ + char c; + int rc, cnt; + char buffer[CMDBUFSIZE]; + + switch (op) + { + default: + errno = EINVAL; + return -1; + + case MTIOCTOP: + /* MTIOCTOP is the easy one. Nothing is transfered in binary. */ + sprintf (buffer, "I%d\n%d\n", ((struct mtop *) arg)->mt_op, + ((struct mtop *) arg)->mt_count); + if (command (fildes, buffer) == -1) + return -1; + return status (fildes); /* Return the count. */ + + case MTIOCGET: + /* Grab the status and read it directly into the structure. + This assumes that the status buffer is not padded + and that 2 shorts fit in a long without any word + alignment problems; i.e., the whole struct is contiguous. + NOTE - this is probably NOT a good assumption. */ + + if (command (fildes, "S") == -1 || (rc = status (fildes)) == -1) + return -1; + + for (; rc > 0; rc -= cnt, arg += cnt) + { + cnt = read (READ (fildes), arg, rc); + if (cnt <= 0) + { + _rmt_shutdown (fildes); + errno = EIO; + return -1; + } + } + + /* Check for byte position. mt_type is a small integer field + (normally) so we will check its magnitude. If it is larger than + 256, we will assume that the bytes are swapped and go through + and reverse all the bytes. */ + + if (((struct mtget *) arg)->mt_type < 256) + return 0; + + for (cnt = 0; cnt < rc; cnt += 2) + { + c = arg[cnt]; + arg[cnt] = arg[cnt + 1]; + arg[cnt + 1] = c; + } + + return 0; + } +} + +#endif diff --git a/gnu/usr.bin/tar/tar.c b/gnu/usr.bin/tar/tar.c new file mode 100644 index 000000000000..938258233980 --- /dev/null +++ b/gnu/usr.bin/tar/tar.c @@ -0,0 +1,1504 @@ +/* Tar -- a tape archiver. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* + * A tar (tape archiver) program. + * + * Written by John Gilmore, ihnp4!hoptoad!gnu, starting 25 Aug 85. + */ + +#include <stdio.h> +#include <sys/types.h> /* Needed for typedefs in tar.h */ +#include "getopt.h" + +/* + * The following causes "tar.h" to produce definitions of all the + * global variables, rather than just "extern" declarations of them. + */ +#define TAR_EXTERN /**/ +#include "tar.h" + +#include "port.h" +#include "regex.h" +#include "fnmatch.h" + +/* + * We should use a conversion routine that does reasonable error + * checking -- atoi doesn't. For now, punt. FIXME. + */ +#define intconv atoi +PTR ck_malloc (); +PTR ck_realloc (); +extern int getoldopt (); +extern void read_and (); +extern void list_archive (); +extern void extract_archive (); +extern void diff_archive (); +extern void create_archive (); +extern void update_archive (); +extern void junk_archive (); +extern void init_volume_number (); +extern void closeout_volume_number (); + +/* JF */ +extern time_t get_date (); + +time_t new_time; + +static FILE *namef; /* File to read names from */ +static char **n_argv; /* Argv used by name routines */ +static int n_argc; /* Argc used by name routines */ +static char **n_ind; /* Store an array of names */ +static int n_indalloc; /* How big is the array? */ +static int n_indused; /* How many entries does it have? */ +static int n_indscan; /* How many of the entries have we scanned? */ + + +extern FILE *msg_file; + +int check_exclude (); +void add_exclude (); +void add_exclude_file (); +void addname (); +void describe (); +void diff_init (); +void extr_init (); +int is_regex (); +void name_add (); +void name_init (); +void options (); +char *un_quote_string (); + +#ifndef S_ISLNK +#define lstat stat +#endif + +#ifndef DEFBLOCKING +#define DEFBLOCKING 20 +#endif + +#ifndef DEF_AR_FILE +#define DEF_AR_FILE "tar.out" +#endif + +/* For long options that unconditionally set a single flag, we have getopt + do it. For the others, we share the code for the equivalent short + named option, the name of which is stored in the otherwise-unused `val' + field of the `struct option'; for long options that have no equivalent + short option, we use nongraphic characters as pseudo short option + characters, starting (for no particular reason) with character 10. */ + +struct option long_options[] = +{ + {"create", 0, 0, 'c'}, + {"append", 0, 0, 'r'}, + {"extract", 0, 0, 'x'}, + {"get", 0, 0, 'x'}, + {"list", 0, 0, 't'}, + {"update", 0, 0, 'u'}, + {"catenate", 0, 0, 'A'}, + {"concatenate", 0, 0, 'A'}, + {"compare", 0, 0, 'd'}, + {"diff", 0, 0, 'd'}, + {"delete", 0, 0, 14}, + {"help", 0, 0, 12}, + + {"null", 0, 0, 16}, + {"directory", 1, 0, 'C'}, + {"record-number", 0, &f_sayblock, 1}, + {"files-from", 1, 0, 'T'}, + {"label", 1, 0, 'V'}, + {"exclude-from", 1, 0, 'X'}, + {"exclude", 1, 0, 15}, + {"file", 1, 0, 'f'}, + {"block-size", 1, 0, 'b'}, + {"version", 0, 0, 11}, + {"verbose", 0, 0, 'v'}, + {"totals", 0, &f_totals, 1}, + + {"read-full-blocks", 0, &f_reblock, 1}, + {"starting-file", 1, 0, 'K'}, + {"to-stdout", 0, &f_exstdout, 1}, + {"ignore-zeros", 0, &f_ignorez, 1}, + {"keep-old-files", 0, 0, 'k'}, + {"same-permissions", 0, &f_use_protection, 1}, + {"preserve-permissions", 0, &f_use_protection, 1}, + {"modification-time", 0, &f_modified, 1}, + {"preserve", 0, 0, 10}, + {"same-order", 0, &f_sorted_names, 1}, + {"same-owner", 0, &f_do_chown, 1}, + {"preserve-order", 0, &f_sorted_names, 1}, + + {"newer", 1, 0, 'N'}, + {"after-date", 1, 0, 'N'}, + {"newer-mtime", 1, 0, 13}, + {"incremental", 0, 0, 'G'}, + {"listed-incremental", 1, 0, 'g'}, + {"multi-volume", 0, &f_multivol, 1}, + {"info-script", 1, 0, 'F'}, + {"new-volume-script", 1, 0, 'F'}, + {"absolute-paths", 0, &f_absolute_paths, 1}, + {"interactive", 0, &f_confirm, 1}, + {"confirmation", 0, &f_confirm, 1}, + + {"verify", 0, &f_verify, 1}, + {"dereference", 0, &f_follow_links, 1}, + {"one-file-system", 0, &f_local_filesys, 1}, + {"old-archive", 0, 0, 'o'}, + {"portability", 0, 0, 'o'}, + {"compress", 0, 0, 'Z'}, + {"uncompress", 0, 0, 'Z'}, + {"block-compress", 0, &f_compress_block, 1}, + {"gzip", 0, 0, 'z'}, + {"ungzip", 0, 0, 'z'}, + {"use-compress-program", 1, 0, 18}, + + + {"same-permissions", 0, &f_use_protection, 1}, + {"sparse", 0, &f_sparse_files, 1}, + {"tape-length", 1, 0, 'L'}, + {"remove-files", 0, &f_remove_files, 1}, + {"ignore-failed-read", 0, &f_ignore_failed_read, 1}, + {"checkpoint", 0, &f_checkpoint, 1}, + {"show-omitted-dirs", 0, &f_show_omitted_dirs, 1}, + {"volno-file", 1, 0, 17}, + {"force-local", 0, &f_force_local, 1}, + {"atime-preserve", 0, &f_atime_preserve, 1}, + + {0, 0, 0, 0} +}; + +/* + * Main routine for tar. + */ +void +main (argc, argv) + int argc; + char **argv; +{ + extern char version_string[]; + + tar = argv[0]; /* JF: was "tar" Set program name */ + filename_terminator = '\n'; + errors = 0; + + options (argc, argv); + + if (!n_argv) + name_init (argc, argv); + + if (f_volno_file) + init_volume_number (); + + switch (cmd_mode) + { + case CMD_CAT: + case CMD_UPDATE: + case CMD_APPEND: + update_archive (); + break; + case CMD_DELETE: + junk_archive (); + break; + case CMD_CREATE: + create_archive (); + if (f_totals) + fprintf (stderr, "Total bytes written: %d\n", tot_written); + break; + case CMD_EXTRACT: + if (f_volhdr) + { + const char *err; + label_pattern = (struct re_pattern_buffer *) + ck_malloc (sizeof *label_pattern); + err = re_compile_pattern (f_volhdr, strlen (f_volhdr), + label_pattern); + if (err) + { + fprintf (stderr, "Bad regular expression: %s\n", + err); + errors++; + break; + } + + } + extr_init (); + read_and (extract_archive); + break; + case CMD_LIST: + if (f_volhdr) + { + const char *err; + label_pattern = (struct re_pattern_buffer *) + ck_malloc (sizeof *label_pattern); + err = re_compile_pattern (f_volhdr, strlen (f_volhdr), + label_pattern); + if (err) + { + fprintf (stderr, "Bad regular expression: %s\n", + err); + errors++; + break; + } + } + read_and (list_archive); +#if 0 + if (!errors) + errors = different; +#endif + break; + case CMD_DIFF: + diff_init (); + read_and (diff_archive); + break; + case CMD_VERSION: + fprintf (stderr, "%s\n", version_string); + break; + case CMD_NONE: + msg ("you must specify exactly one of the r, c, t, x, or d options\n"); + fprintf (stderr, "For more information, type ``%s --help''.\n", tar); + exit (EX_ARGSBAD); + } + if (f_volno_file) + closeout_volume_number (); + exit (errors); + /* NOTREACHED */ +} + + +/* + * Parse the options for tar. + */ +void +options (argc, argv) + int argc; + char **argv; +{ + register int c; /* Option letter */ + int ind = -1; + + /* Set default option values */ + blocking = DEFBLOCKING; /* From Makefile */ + ar_files = (char **) ck_malloc (sizeof (char *) * 10); + ar_files_len = 10; + n_ar_files = 0; + cur_ar_file = 0; + + /* Parse options */ + while ((c = getoldopt (argc, argv, + "-01234567Ab:BcC:df:F:g:GhikK:lL:mMN:oOpPrRsStT:uvV:wWxX:zZ", + long_options, &ind)) != EOF) + { + switch (c) + { + case 0: /* long options that set a single flag */ + break; + case 1: + /* File name or non-parsed option */ + name_add (optarg); + break; + case 'C': + name_add ("-C"); + name_add (optarg); + break; + case 10: /* preserve */ + f_use_protection = f_sorted_names = 1; + break; + case 11: + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_VERSION; + break; + case 12: /* help */ + printf ("This is GNU tar, the tape archiving program.\n"); + describe (); + exit (1); + case 13: + f_new_files++; + goto get_newer; + + case 14: /* Delete in the archive */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_DELETE; + break; + + case 15: + f_exclude++; + add_exclude (optarg); + break; + + case 16: /* -T reads null terminated filenames. */ + filename_terminator = '\0'; + break; + + case 17: + f_volno_file = optarg; + break; + + case 18: + if (f_compressprog) + { + msg ("Only one compression option permitted\n"); + exit (EX_ARGSBAD); + } + f_compressprog = optarg; + break; + + case 'g': /* We are making a GNU dump; save + directories at the beginning of + the archive, and include in each + directory its contents */ + if (f_oldarch) + goto badopt; + f_gnudump++; + gnu_dumpfile = optarg; + break; + + + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + { + /* JF this'll have to be modified for other + systems, of course! */ + int d, add; + static char buf[50]; + + d = getoldopt (argc, argv, "lmh"); +#ifdef MAYBEDEF + sprintf (buf, "/dev/rmt/%d%c", c, d); +#else +#ifndef LOW_NUM +#define LOW_NUM 0 +#define MID_NUM 8 +#define HGH_NUM 16 +#endif + if (d == 'l') + add = LOW_NUM; + else if (d == 'm') + add = MID_NUM; + else if (d == 'h') + add = HGH_NUM; + else + goto badopt; + + sprintf (buf, "/dev/rmt%d", add + c - '0'); +#endif + if (n_ar_files == ar_files_len) + ar_files + = (char **) + ck_malloc (sizeof (char *) + * (ar_files_len *= 2)); + ar_files[n_ar_files++] = buf; + } + break; + + case 'A': /* Arguments are tar files, + just cat them onto the end + of the archive. */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_CAT; + break; + + case 'b': /* Set blocking factor */ + blocking = intconv (optarg); + break; + + case 'B': /* Try to reblock input */ + f_reblock++; /* For reading 4.2BSD pipes */ + break; + + case 'c': /* Create an archive */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_CREATE; + break; + +#if 0 + case 'C': + if (chdir (optarg) < 0) + msg_perror ("Can't change directory to %d", optarg); + break; +#endif + + case 'd': /* Find difference tape/disk */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_DIFF; + break; + + case 'f': /* Use ar_file for the archive */ + if (n_ar_files == ar_files_len) + ar_files + = (char **) ck_malloc (sizeof (char *) + * (ar_files_len *= 2)); + + ar_files[n_ar_files++] = optarg; + break; + + case 'F': + /* Since -F is only useful with -M , make it implied */ + f_run_script_at_end++;/* run this script at the end */ + info_script = optarg; /* of each tape */ + f_multivol++; + break; + + case 'G': /* We are making a GNU dump; save + directories at the beginning of + the archive, and include in each + directory its contents */ + if (f_oldarch) + goto badopt; + f_gnudump++; + gnu_dumpfile = 0; + break; + + case 'h': + f_follow_links++; /* follow symbolic links */ + break; + + case 'i': + f_ignorez++; /* Ignore zero records (eofs) */ + /* + * This can't be the default, because Unix tar + * writes two records of zeros, then pads out the + * block with garbage. + */ + break; + + case 'k': /* Don't overwrite files */ +#ifdef NO_OPEN3 + msg ("can't keep old files on this system"); + exit (EX_ARGSBAD); +#else + f_keep++; +#endif + break; + + case 'K': + f_startfile++; + addname (optarg); + break; + + case 'l': /* When dumping directories, don't + dump files/subdirectories that are + on other filesystems. */ + f_local_filesys++; + break; + + case 'L': + tape_length = intconv (optarg); + f_multivol++; + break; + case 'm': + f_modified++; + break; + + case 'M': /* Make Multivolume archive: + When we can't write any more + into the archive, re-open it, + and continue writing */ + f_multivol++; + break; + + case 'N': /* Only write files newer than X */ + get_newer: + f_new_files++; + new_time = get_date (optarg, (PTR) 0); + if (new_time == (time_t) - 1) + { + msg ("invalid date format `%s'", optarg); + exit (EX_ARGSBAD); + } + break; + + case 'o': /* Generate old archive */ + if (f_gnudump /* || f_dironly */ ) + goto badopt; + f_oldarch++; + break; + + case 'O': + f_exstdout++; + break; + + case 'p': + f_use_protection++; + break; + + case 'P': + f_absolute_paths++; + break; + + case 'r': /* Append files to the archive */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_APPEND; + break; + + case 'R': + f_sayblock++; /* Print block #s for debug */ + break; /* of bad tar archives */ + + case 's': + f_sorted_names++; /* Names to extr are sorted */ + break; + + case 'S': /* deal with sparse files */ + f_sparse_files++; + break; + case 't': + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_LIST; + f_verbose++; /* "t" output == "cv" or "xv" */ + break; + + case 'T': + name_file = optarg; + f_namefile++; + break; + + case 'u': /* Append files to the archive that + aren't there, or are newer than the + copy in the archive */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_UPDATE; + break; + + case 'v': + f_verbose++; + break; + + case 'V': + f_volhdr = optarg; + break; + + case 'w': + f_confirm++; + break; + + case 'W': + f_verify++; + break; + + case 'x': /* Extract files from the archive */ + if (cmd_mode != CMD_NONE) + goto badopt; + cmd_mode = CMD_EXTRACT; + break; + + case 'X': + f_exclude++; + add_exclude_file (optarg); + break; + + case 'z': + if (f_compressprog) + { + msg ("Only one compression option permitted\n"); + exit (EX_ARGSBAD); + } + f_compressprog = "gzip"; + break; + + case 'Z': + if (f_compressprog) + { + msg ("Only one compression option permitted\n"); + exit (EX_ARGSBAD); + } + f_compressprog = "compress"; + break; + + case '?': + badopt: + msg ("Unknown option. Use '%s --help' for a complete list of options.", tar); + exit (EX_ARGSBAD); + + } + } + + blocksize = blocking * RECORDSIZE; + if (n_ar_files == 0) + { + n_ar_files = 1; + ar_files[0] = getenv ("TAPE"); /* From environment, or */ + if (ar_files[0] == 0) + ar_files[0] = DEF_AR_FILE; /* From Makefile */ + } + if (n_ar_files > 1 && !f_multivol) + { + msg ("Multiple archive files requires --multi-volume\n"); + exit (EX_ARGSBAD); + } + if (f_compress_block && !f_compressprog) + { + msg ("You must use a compression option (--gzip, --compress\n\ +or --use-compress-program) with --block-compress.\n"); + exit (EX_ARGSBAD); + } +} + + +/* + * Print as much help as the user's gonna get. + * + * We have to sprinkle in the KLUDGE lines because too many compilers + * cannot handle character strings longer than about 512 bytes. Yuk! + * In particular, MS-DOS and Xenix MSC and PDP-11 V7 Unix have this + * problem. + */ +void +describe () +{ + puts ("choose one of the following:"); + fputs ("\ +-A, --catenate,\n\ + --concatenate append tar files to an archive\n\ +-c, --create create a new archive\n\ +-d, --diff,\n\ + --compare find differences between archive and file system\n\ +--delete delete from the archive (not for use on mag tapes!)\n\ +-r, --append append files to the end of an archive\n\ +-t, --list list the contents of an archive\n\ +-u, --update only append files that are newer than copy in archive\n\ +-x, --extract,\n\ + --get extract files from an archive\n", stdout); + + fprintf (stdout, "\ +Other options:\n\ +--atime-preserve don't change access times on dumped files\n\ +-b, --block-size N block size of Nx512 bytes (default N=%d)\n", DEFBLOCKING); + fputs ("\ +-B, --read-full-blocks reblock as we read (for reading 4.2BSD pipes)\n\ +-C, --directory DIR change to directory DIR\n\ +--checkpoint print directory names while reading the archive\n\ +", stdout); /* KLUDGE */ + fprintf (stdout, "\ +-f, --file [HOSTNAME:]F use archive file or device F (default %s)\n", + DEF_AR_FILE); + fputs ("\ +--force-local archive file is local even if has a colon\n\ +-F, --info-script F\n\ + --new-volume-script F run script at end of each tape (implies -M)\n\ +-G, --incremental create/list/extract old GNU-format incremental backup\n\ +-g, --listed-incremental F create/list/extract new GNU-format incremental backup\n\ +-h, --dereference don't dump symlinks; dump the files they point to\n\ +-i, --ignore-zeros ignore blocks of zeros in archive (normally mean EOF)\n\ +--ignore-failed-read don't exit with non-zero status on unreadable files\n\ +-k, --keep-old-files keep existing files; don't overwrite them from archive\n\ +-K, --starting-file F begin at file F in the archive\n\ +-l, --one-file-system stay in local file system when creating an archive\n\ +-L, --tape-length N change tapes after writing N*1024 bytes\n\ +", stdout); /* KLUDGE */ + fputs ("\ +-m, --modification-time don't extract file modified time\n\ +-M, --multi-volume create/list/extract multi-volume archive\n\ +-N, --after-date DATE,\n\ + --newer DATE only store files newer than DATE\n\ +-o, --old-archive,\n\ + --portability write a V7 format archive, rather than ANSI format\n\ +-O, --to-stdout extract files to standard output\n\ +-p, --same-permissions,\n\ + --preserve-permissions extract all protection information\n\ +-P, --absolute-paths don't strip leading `/'s from file names\n\ +--preserve like -p -s\n\ +", stdout); /* KLUDGE */ + fputs ("\ +-R, --record-number show record number within archive with each message\n\ +--remove-files remove files after adding them to the archive\n\ +-s, --same-order,\n\ + --preserve-order list of names to extract is sorted to match archive\n\ +--same-owner create extracted files with the same ownership \n\ +-S, --sparse handle sparse files efficiently\n\ +-T, --files-from F get names to extract or create from file F\n\ +--null -T reads null-terminated names, disable -C\n\ +--totals print total bytes written with --create\n\ +-v, --verbose verbosely list files processed\n\ +-V, --label NAME create archive with volume name NAME\n\ +--version print tar program version number\n\ +-w, --interactive,\n\ + --confirmation ask for confirmation for every action\n\ +", stdout); /* KLUDGE */ + fputs ("\ +-W, --verify attempt to verify the archive after writing it\n\ +--exclude FILE exclude file FILE\n\ +-X, --exclude-from FILE exclude files listed in FILE\n\ +-Z, --compress,\n\ + --uncompress filter the archive through compress\n\ +-z, --gzip,\n\ + --ungzip filter the archive through gzip\n\ +--use-compress-program PROG\n\ + filter the archive through PROG (which must accept -d)\n\ +--block-compress block the output of compression program for tapes\n\ +-[0-7][lmh] specify drive and density\n\ +", stdout); +} + +void +name_add (name) + char *name; +{ + if (n_indalloc == n_indused) + { + n_indalloc += 10; + n_ind = (char **) (n_indused ? ck_realloc (n_ind, n_indalloc * sizeof (char *)): ck_malloc (n_indalloc * sizeof (char *))); + } + n_ind[n_indused++] = name; +} + +/* + * Set up to gather file names for tar. + * + * They can either come from stdin or from argv. + */ +void +name_init (argc, argv) + int argc; + char **argv; +{ + + if (f_namefile) + { + if (optind < argc) + { + msg ("too many args with -T option"); + exit (EX_ARGSBAD); + } + if (!strcmp (name_file, "-")) + { + namef = stdin; + } + else + { + namef = fopen (name_file, "r"); + if (namef == NULL) + { + msg_perror ("can't open file %s", name_file); + exit (EX_BADFILE); + } + } + } + else + { + /* Get file names from argv, after options. */ + n_argc = argc; + n_argv = argv; + } +} + +/* Read the next filename read from STREAM and null-terminate it. + Put it into BUFFER, reallocating and adjusting *PBUFFER_SIZE if necessary. + Return the new value for BUFFER, or NULL at end of file. */ + +char * +read_name_from_file (buffer, pbuffer_size, stream) + char *buffer; + size_t *pbuffer_size; + FILE *stream; +{ + register int c; + register int indx = 0; + register size_t buffer_size = *pbuffer_size; + + while ((c = getc (stream)) != EOF && c != filename_terminator) + { + if (indx == buffer_size) + { + buffer_size += NAMSIZ; + buffer = ck_realloc (buffer, buffer_size + 2); + } + buffer[indx++] = c; + } + if (indx == 0 && c == EOF) + return NULL; + if (indx == buffer_size) + { + buffer_size += NAMSIZ; + buffer = ck_realloc (buffer, buffer_size + 2); + } + buffer[indx] = '\0'; + *pbuffer_size = buffer_size; + return buffer; +} + +/* + * Get the next name from argv or the name file. + * + * Result is in static storage and can't be relied upon across two calls. + * + * If CHANGE_DIRS is non-zero, treat a filename of the form "-C" as + * meaning that the next filename is the name of a directory to change to. + * If `filename_terminator' is '\0', CHANGE_DIRS is effectively always 0. + */ + +char * +name_next (change_dirs) + int change_dirs; +{ + static char *buffer; /* Holding pattern */ + static int buffer_siz; + register char *p; + register char *q = 0; + register int next_name_is_dir = 0; + extern char *un_quote_string (); + + if (buffer_siz == 0) + { + buffer = ck_malloc (NAMSIZ + 2); + buffer_siz = NAMSIZ; + } + if (filename_terminator == '\0') + change_dirs = 0; +tryagain: + if (namef == NULL) + { + if (n_indscan < n_indused) + p = n_ind[n_indscan++]; + else if (optind < n_argc) + /* Names come from argv, after options */ + p = n_argv[optind++]; + else + { + if (q) + msg ("Missing filename after -C"); + return NULL; + } + + /* JF trivial support for -C option. I don't know if + chdir'ing at this point is dangerous or not. + It seems to work, which is all I ask. */ + if (change_dirs && !q && p[0] == '-' && p[1] == 'C' && p[2] == '\0') + { + q = p; + goto tryagain; + } + if (q) + { + if (chdir (p) < 0) + msg_perror ("Can't chdir to %s", p); + q = 0; + goto tryagain; + } + /* End of JF quick -C hack */ + +#if 0 + if (f_exclude && check_exclude (p)) + goto tryagain; +#endif + return un_quote_string (p); + } + while (p = read_name_from_file (buffer, &buffer_siz, namef)) + { + buffer = p; + if (*p == '\0') + continue; /* Ignore empty lines. */ + q = p + strlen (p) - 1; + while (q > p && *q == '/')/* Zap trailing "/"s. */ + *q-- = '\0'; + if (change_dirs && next_name_is_dir == 0 + && p[0] == '-' && p[1] == 'C' && p[2] == '\0') + { + next_name_is_dir = 1; + goto tryagain; + } + if (next_name_is_dir) + { + if (chdir (p) < 0) + msg_perror ("Can't change to directory %s", p); + next_name_is_dir = 0; + goto tryagain; + } +#if 0 + if (f_exclude && check_exclude (p)) + goto tryagain; +#endif + return un_quote_string (p); + } + return NULL; +} + + +/* + * Close the name file, if any. + */ +void +name_close () +{ + + if (namef != NULL && namef != stdin) + fclose (namef); +} + + +/* + * Gather names in a list for scanning. + * Could hash them later if we really care. + * + * If the names are already sorted to match the archive, we just + * read them one by one. name_gather reads the first one, and it + * is called by name_match as appropriate to read the next ones. + * At EOF, the last name read is just left in the buffer. + * This option lets users of small machines extract an arbitrary + * number of files by doing "tar t" and editing down the list of files. + */ +void +name_gather () +{ + register char *p; + static struct name *namebuf; /* One-name buffer */ + static namelen; + static char *chdir_name; + + if (f_sorted_names) + { + if (!namelen) + { + namelen = NAMSIZ; + namebuf = (struct name *) ck_malloc (sizeof (struct name) + NAMSIZ); + } + p = name_next (0); + if (p) + { + if (*p == '-' && p[1] == 'C' && p[2] == '\0') + { + chdir_name = name_next (0); + p = name_next (0); + if (!p) + { + msg ("Missing file name after -C"); + exit (EX_ARGSBAD); + } + namebuf->change_dir = chdir_name; + } + namebuf->length = strlen (p); + if (namebuf->length >= namelen) + { + namebuf = (struct name *) ck_realloc (namebuf, sizeof (struct name) + namebuf->length); + namelen = namebuf->length; + } + strncpy (namebuf->name, p, namebuf->length); + namebuf->name[namebuf->length] = 0; + namebuf->next = (struct name *) NULL; + namebuf->found = 0; + namelist = namebuf; + namelast = namelist; + } + return; + } + + /* Non sorted names -- read them all in */ + while (p = name_next (0)) + addname (p); +} + +/* + * Add a name to the namelist. + */ +void +addname (name) + char *name; /* pointer to name */ +{ + register int i; /* Length of string */ + register struct name *p; /* Current struct pointer */ + static char *chdir_name; + char *new_name (); + + if (name[0] == '-' && name[1] == 'C' && name[2] == '\0') + { + chdir_name = name_next (0); + name = name_next (0); + if (!chdir_name) + { + msg ("Missing file name after -C"); + exit (EX_ARGSBAD); + } + if (chdir_name[0] != '/') + { + char *path = ck_malloc (PATH_MAX); +#if defined(__MSDOS__) || defined(HAVE_GETCWD) || defined(_POSIX_VERSION) + if (!getcwd (path, PATH_MAX)) + { + msg ("Couldn't get current directory."); + exit (EX_SYSTEM); + } +#else + char *getwd (); + + if (!getwd (path)) + { + msg ("Couldn't get current directory: %s", path); + exit (EX_SYSTEM); + } +#endif + chdir_name = new_name (path, chdir_name); + free (path); + } + } + + if (name) + { + i = strlen (name); + /*NOSTRICT*/ + p = (struct name *) malloc ((unsigned) (sizeof (struct name) + i)); + } + else + p = (struct name *) malloc ((unsigned) (sizeof (struct name))); + if (!p) + { + if (name) + msg ("cannot allocate mem for name '%s'.", name); + else + msg ("cannot allocate mem for chdir record."); + exit (EX_SYSTEM); + } + p->next = (struct name *) NULL; + if (name) + { + p->fake = 0; + p->length = i; + strncpy (p->name, name, i); + p->name[i] = '\0'; /* Null term */ + } + else + p->fake = 1; + p->found = 0; + p->regexp = 0; /* Assume not a regular expression */ + p->firstch = 1; /* Assume first char is literal */ + p->change_dir = chdir_name; + p->dir_contents = 0; /* JF */ + if (name) + { + if (index (name, '*') || index (name, '[') || index (name, '?')) + { + p->regexp = 1; /* No, it's a regexp */ + if (name[0] == '*' || name[0] == '[' || name[0] == '?') + p->firstch = 0; /* Not even 1st char literal */ + } + } + + if (namelast) + namelast->next = p; + namelast = p; + if (!namelist) + namelist = p; +} + +/* + * Return nonzero if name P (from an archive) matches any name from + * the namelist, zero if not. + */ +int +name_match (p) + register char *p; +{ + register struct name *nlp; + register int len; + +again: + if (0 == (nlp = namelist)) /* Empty namelist is easy */ + return 1; + if (nlp->fake) + { + if (nlp->change_dir && chdir (nlp->change_dir)) + msg_perror ("Can't change to directory %d", nlp->change_dir); + namelist = 0; + return 1; + } + len = strlen (p); + for (; nlp != 0; nlp = nlp->next) + { + /* If first chars don't match, quick skip */ + if (nlp->firstch && nlp->name[0] != p[0]) + continue; + + /* Regular expressions (shell globbing, actually). */ + if (nlp->regexp) + { + if (fnmatch (nlp->name, p, FNM_LEADING_DIR) == 0) + { + nlp->found = 1; /* Remember it matched */ + if (f_startfile) + { + free ((void *) namelist); + namelist = 0; + } + if (nlp->change_dir && chdir (nlp->change_dir)) + msg_perror ("Can't change to directory %s", nlp->change_dir); + return 1; /* We got a match */ + } + continue; + } + + /* Plain Old Strings */ + if (nlp->length <= len /* Archive len >= specified */ + && (p[nlp->length] == '\0' || p[nlp->length] == '/') + /* Full match on file/dirname */ + && strncmp (p, nlp->name, nlp->length) == 0) /* Name compare */ + { + nlp->found = 1; /* Remember it matched */ + if (f_startfile) + { + free ((void *) namelist); + namelist = 0; + } + if (nlp->change_dir && chdir (nlp->change_dir)) + msg_perror ("Can't change to directory %s", nlp->change_dir); + return 1; /* We got a match */ + } + } + + /* + * Filename from archive not found in namelist. + * If we have the whole namelist here, just return 0. + * Otherwise, read the next name in and compare it. + * If this was the last name, namelist->found will remain on. + * If not, we loop to compare the newly read name. + */ + if (f_sorted_names && namelist->found) + { + name_gather (); /* Read one more */ + if (!namelist->found) + goto again; + } + return 0; +} + + +/* + * Print the names of things in the namelist that were not matched. + */ +void +names_notfound () +{ + register struct name *nlp, *next; + register char *p; + + for (nlp = namelist; nlp != 0; nlp = next) + { + next = nlp->next; + if (!nlp->found) + msg ("%s not found in archive", nlp->name); + + /* + * We could free() the list, but the process is about + * to die anyway, so save some CPU time. Amigas and + * other similarly broken software will need to waste + * the time, though. + */ +#ifdef amiga + if (!f_sorted_names) + free (nlp); +#endif + } + namelist = (struct name *) NULL; + namelast = (struct name *) NULL; + + if (f_sorted_names) + { + while (0 != (p = name_next (1))) + msg ("%s not found in archive", p); + } +} + +/* These next routines were created by JF */ + +void +name_expand () +{ + ; +} + +/* This is like name_match(), except that it returns a pointer to the name + it matched, and doesn't set ->found The caller will have to do that + if it wants to. Oh, and if the namelist is empty, it returns 0, unlike + name_match(), which returns TRUE */ + +struct name * +name_scan (p) + register char *p; +{ + register struct name *nlp; + register int len; + +again: + if (0 == (nlp = namelist)) /* Empty namelist is easy */ + return 0; + len = strlen (p); + for (; nlp != 0; nlp = nlp->next) + { + /* If first chars don't match, quick skip */ + if (nlp->firstch && nlp->name[0] != p[0]) + continue; + + /* Regular expressions */ + if (nlp->regexp) + { + if (fnmatch (nlp->name, p, FNM_LEADING_DIR) == 0) + return nlp; /* We got a match */ + continue; + } + + /* Plain Old Strings */ + if (nlp->length <= len /* Archive len >= specified */ + && (p[nlp->length] == '\0' || p[nlp->length] == '/') + /* Full match on file/dirname */ + && strncmp (p, nlp->name, nlp->length) == 0) /* Name compare */ + return nlp; /* We got a match */ + } + + /* + * Filename from archive not found in namelist. + * If we have the whole namelist here, just return 0. + * Otherwise, read the next name in and compare it. + * If this was the last name, namelist->found will remain on. + * If not, we loop to compare the newly read name. + */ + if (f_sorted_names && namelist->found) + { + name_gather (); /* Read one more */ + if (!namelist->found) + goto again; + } + return (struct name *) 0; +} + +/* This returns a name from the namelist which doesn't have ->found set. + It sets ->found before returning, so successive calls will find and return + all the non-found names in the namelist */ + +struct name *gnu_list_name; + +char * +name_from_list () +{ + if (!gnu_list_name) + gnu_list_name = namelist; + while (gnu_list_name && gnu_list_name->found) + gnu_list_name = gnu_list_name->next; + if (gnu_list_name) + { + gnu_list_name->found++; + if (gnu_list_name->change_dir) + if (chdir (gnu_list_name->change_dir) < 0) + msg_perror ("can't chdir to %s", gnu_list_name->change_dir); + return gnu_list_name->name; + } + return (char *) 0; +} + +void +blank_name_list () +{ + struct name *n; + + gnu_list_name = 0; + for (n = namelist; n; n = n->next) + n->found = 0; +} + +char * +new_name (path, name) + char *path, *name; +{ + char *path_buf; + + path_buf = (char *) malloc (strlen (path) + strlen (name) + 2); + if (path_buf == 0) + { + msg ("Can't allocate memory for name '%s/%s", path, name); + exit (EX_SYSTEM); + } + (void) sprintf (path_buf, "%s/%s", path, name); + return path_buf; +} + +/* returns non-zero if the luser typed 'y' or 'Y', zero otherwise. */ + +int +confirm (action, file) + char *action, *file; +{ + int c, nl; + static FILE *confirm_file = 0; + extern FILE *msg_file; + extern char TTY_NAME[]; + + fprintf (msg_file, "%s %s?", action, file); + fflush (msg_file); + if (!confirm_file) + { + confirm_file = (archive == 0) ? fopen (TTY_NAME, "r") : stdin; + if (!confirm_file) + { + msg ("Can't read confirmation from user"); + exit (EX_SYSTEM); + } + } + c = getc (confirm_file); + for (nl = c; nl != '\n' && nl != EOF; nl = getc (confirm_file)) + ; + return (c == 'y' || c == 'Y'); +} + +char *x_buffer = 0; +int size_x_buffer; +int free_x_buffer; + +char **exclude = 0; +int size_exclude = 0; +int free_exclude = 0; + +char **re_exclude = 0; +int size_re_exclude = 0; +int free_re_exclude = 0; + +void +add_exclude (name) + char *name; +{ + /* char *rname;*/ + /* char **tmp_ptr;*/ + int size_buf; + + un_quote_string (name); + size_buf = strlen (name); + + if (x_buffer == 0) + { + x_buffer = (char *) ck_malloc (size_buf + 1024); + free_x_buffer = 1024; + } + else if (free_x_buffer <= size_buf) + { + char *old_x_buffer; + char **tmp_ptr; + + old_x_buffer = x_buffer; + x_buffer = (char *) ck_realloc (x_buffer, size_x_buffer + 1024); + free_x_buffer = 1024; + for (tmp_ptr = exclude; tmp_ptr < exclude + size_exclude; tmp_ptr++) + *tmp_ptr = x_buffer + ((*tmp_ptr) - old_x_buffer); + for (tmp_ptr = re_exclude; tmp_ptr < re_exclude + size_re_exclude; tmp_ptr++) + *tmp_ptr = x_buffer + ((*tmp_ptr) - old_x_buffer); + } + + if (is_regex (name)) + { + if (free_re_exclude == 0) + { + re_exclude = (char **) (re_exclude ? ck_realloc (re_exclude, (size_re_exclude + 32) * sizeof (char *)): ck_malloc (sizeof (char *) * 32)); + free_re_exclude += 32; + } + re_exclude[size_re_exclude] = x_buffer + size_x_buffer; + size_re_exclude++; + free_re_exclude--; + } + else + { + if (free_exclude == 0) + { + exclude = (char **) (exclude ? ck_realloc (exclude, (size_exclude + 32) * sizeof (char *)): ck_malloc (sizeof (char *) * 32)); + free_exclude += 32; + } + exclude[size_exclude] = x_buffer + size_x_buffer; + size_exclude++; + free_exclude--; + } + strcpy (x_buffer + size_x_buffer, name); + size_x_buffer += size_buf + 1; + free_x_buffer -= size_buf + 1; +} + +void +add_exclude_file (file) + char *file; +{ + FILE *fp; + char buf[1024]; + + if (strcmp (file, "-")) + fp = fopen (file, "r"); + else + /* Let's hope the person knows what they're doing. */ + /* Using -X - -T - -f - will get you *REALLY* strange + results. . . */ + fp = stdin; + + if (!fp) + { + msg_perror ("can't open %s", file); + exit (2); + } + while (fgets (buf, 1024, fp)) + { + /* int size_buf;*/ + char *end_str; + + end_str = rindex (buf, '\n'); + if (end_str) + *end_str = '\0'; + add_exclude (buf); + + } + fclose (fp); +} + +int +is_regex (str) + char *str; +{ + return index (str, '*') || index (str, '[') || index (str, '?'); +} + +/* Returns non-zero if the file 'name' should not be added/extracted */ +int +check_exclude (name) + char *name; +{ + int n; + char *str; + extern char *strstr (); + + for (n = 0; n < size_re_exclude; n++) + { + if (fnmatch (re_exclude[n], name, FNM_LEADING_DIR) == 0) + return 1; + } + for (n = 0; n < size_exclude; n++) + { + /* Accept the output from strstr only if it is the last + part of the string. There is certainly a faster way to + do this. . . */ + if ((str = strstr (name, exclude[n])) + && (str == name || str[-1] == '/') + && str[strlen (exclude[n])] == '\0') + return 1; + } + return 0; +} diff --git a/gnu/usr.bin/tar/tar.h b/gnu/usr.bin/tar/tar.h new file mode 100644 index 000000000000..c3fec78743bb --- /dev/null +++ b/gnu/usr.bin/tar/tar.h @@ -0,0 +1,291 @@ +/* Declarations for tar archives. + Copyright (C) 1988, 1992, 1993 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* major() and minor() macros (among other things) defined here for hpux */ +#ifdef hpux +#include <sys/mknod.h> +#endif + +/* + * Kludge for handling systems that can't cope with multiple + * external definitions of a variable. In ONE routine (tar.c), + * we #define TAR_EXTERN to null; here, we set it to "extern" if + * it is not already set. + */ +#ifndef TAR_EXTERN +#define TAR_EXTERN extern +#endif + +/* + * Header block on tape. + * + * I'm going to use traditional DP naming conventions here. + * A "block" is a big chunk of stuff that we do I/O on. + * A "record" is a piece of info that we care about. + * Typically many "record"s fit into a "block". + */ +#define RECORDSIZE 512 +#define NAMSIZ 100 +#define TUNMLEN 32 +#define TGNMLEN 32 +#define SPARSE_EXT_HDR 21 +#define SPARSE_IN_HDR 4 + +struct sparse + { + char offset[12]; + char numbytes[12]; + }; + +struct sp_array + { + int offset; + int numbytes; + }; + +union record + { + char charptr[RECORDSIZE]; + struct header + { + char arch_name[NAMSIZ]; + char mode[8]; + char uid[8]; + char gid[8]; + char size[12]; + char mtime[12]; + char chksum[8]; + char linkflag; + char arch_linkname[NAMSIZ]; + char magic[8]; + char uname[TUNMLEN]; + char gname[TGNMLEN]; + char devmajor[8]; + char devminor[8]; + /* these following fields were added by JF for gnu */ + /* and are NOT standard */ + char atime[12]; + char ctime[12]; + char offset[12]; + char longnames[4]; +#ifdef NEEDPAD + char pad; +#endif + struct sparse sp[SPARSE_IN_HDR]; + char isextended; + char realsize[12]; /* true size of the sparse file */ + /* char ending_blanks[12];*//* number of nulls at the + end of the file, if any */ + } + header; + struct extended_header + { + struct sparse sp[21]; + char isextended; + } + ext_hdr; + }; + +/* The checksum field is filled with this while the checksum is computed. */ +#define CHKBLANKS " " /* 8 blanks, no null */ + +/* The magic field is filled with this if uname and gname are valid. */ +#define TMAGIC "ustar " /* 7 chars and a null */ + +/* The linkflag defines the type of file */ +#define LF_OLDNORMAL '\0' /* Normal disk file, Unix compat */ +#define LF_NORMAL '0' /* Normal disk file */ +#define LF_LINK '1' /* Link to previously dumped file */ +#define LF_SYMLINK '2' /* Symbolic link */ +#define LF_CHR '3' /* Character special file */ +#define LF_BLK '4' /* Block special file */ +#define LF_DIR '5' /* Directory */ +#define LF_FIFO '6' /* FIFO special file */ +#define LF_CONTIG '7' /* Contiguous file */ +/* Further link types may be defined later. */ + +/* Note that the standards committee allows only capital A through + capital Z for user-defined expansion. This means that defining something + as, say '8' is a *bad* idea. */ +#define LF_DUMPDIR 'D' /* This is a dir entry that contains + the names of files that were in + the dir at the time the dump + was made */ +#define LF_LONGLINK 'K' /* Identifies the NEXT file on the tape + as having a long linkname */ +#define LF_LONGNAME 'L' /* Identifies the NEXT file on the tape + as having a long name. */ +#define LF_MULTIVOL 'M' /* This is the continuation + of a file that began on another + volume */ +#define LF_NAMES 'N' /* For storing filenames that didn't + fit in 100 characters */ +#define LF_SPARSE 'S' /* This is for sparse files */ +#define LF_VOLHDR 'V' /* This file is a tape/volume header */ +/* Ignore it on extraction */ + +/* + * Exit codes from the "tar" program + */ +#define EX_SUCCESS 0 /* success! */ +#define EX_ARGSBAD 1 /* invalid args */ +#define EX_BADFILE 2 /* invalid filename */ +#define EX_BADARCH 3 /* bad archive */ +#define EX_SYSTEM 4 /* system gave unexpected error */ +#define EX_BADVOL 5 /* Special error code means + Tape volume doesn't match the one + specified on the command line */ + +/* + * Global variables + */ +TAR_EXTERN union record *ar_block; /* Start of block of archive */ +TAR_EXTERN union record *ar_record; /* Current record of archive */ +TAR_EXTERN union record *ar_last; /* Last+1 record of archive block */ +TAR_EXTERN char ar_reading; /* 0 writing, !0 reading archive */ +TAR_EXTERN int blocking; /* Size of each block, in records */ +TAR_EXTERN int blocksize; /* Size of each block, in bytes */ +TAR_EXTERN char *info_script; /* Script to run at end of each tape change */ +TAR_EXTERN char *name_file; /* File containing names to work on */ +TAR_EXTERN char filename_terminator; /* \n or \0. */ +TAR_EXTERN char *tar; /* Name of this program */ +TAR_EXTERN struct sp_array *sparsearray; /* Pointer to the start of the scratch space */ +TAR_EXTERN int sp_array_size; /* Initial size of the sparsearray */ +TAR_EXTERN int tot_written; /* Total written to output */ +TAR_EXTERN struct re_pattern_buffer + *label_pattern; /* compiled regex for extract label */ +TAR_EXTERN char **ar_files; /* list of tape drive names */ +TAR_EXTERN int n_ar_files; /* number of tape drive names */ +TAR_EXTERN int cur_ar_file; /* tape drive currently being used */ +TAR_EXTERN int ar_files_len; /* malloced size of ar_files */ +TAR_EXTERN char *current_file_name, *current_link_name; + +/* + * Flags from the command line + */ +TAR_EXTERN int cmd_mode; +#define CMD_NONE 0 +#define CMD_CAT 1 /* -A */ +#define CMD_CREATE 2 /* -c */ +#define CMD_DIFF 3 /* -d */ +#define CMD_APPEND 4 /* -r */ +#define CMD_LIST 5 /* -t */ +#define CMD_UPDATE 6 /* -u */ +#define CMD_EXTRACT 7 /* -x */ +#define CMD_DELETE 8 /* -D */ +#define CMD_VERSION 9 /* --version */ + + +TAR_EXTERN int f_reblock; /* -B */ +#if 0 +TAR_EXTERN char f_dironly; /* -D */ +#endif +TAR_EXTERN int f_run_script_at_end; /* -F */ +TAR_EXTERN int f_gnudump; /* -G */ +TAR_EXTERN int f_follow_links; /* -h */ +TAR_EXTERN int f_ignorez; /* -i */ +TAR_EXTERN int f_keep; /* -k */ +TAR_EXTERN int f_startfile; /* -K */ +TAR_EXTERN int f_local_filesys; /* -l */ +TAR_EXTERN int tape_length; /* -L */ +TAR_EXTERN int f_modified; /* -m */ +TAR_EXTERN int f_multivol; /* -M */ +TAR_EXTERN int f_new_files; /* -N */ +TAR_EXTERN int f_oldarch; /* -o */ +TAR_EXTERN int f_exstdout; /* -O */ +TAR_EXTERN int f_use_protection;/* -p */ +TAR_EXTERN int f_absolute_paths;/* -P */ +TAR_EXTERN int f_sayblock; /* -R */ +TAR_EXTERN int f_sorted_names; /* -s */ +TAR_EXTERN int f_sparse_files; /* -S ... JK */ +TAR_EXTERN int f_namefile; /* -T */ +TAR_EXTERN int f_verbose; /* -v */ +TAR_EXTERN char *f_volhdr; /* -V */ +TAR_EXTERN int f_confirm; /* -w */ +TAR_EXTERN int f_verify; /* -W */ +TAR_EXTERN int f_exclude; /* -X */ +TAR_EXTERN char *f_compressprog; /* -z and -Z */ +TAR_EXTERN int f_do_chown; /* --do-chown */ +TAR_EXTERN int f_totals; /* --totals */ +TAR_EXTERN int f_remove_files; /* --remove-files */ +TAR_EXTERN int f_ignore_failed_read; /* --ignore-failed-read */ +TAR_EXTERN int f_checkpoint; /* --checkpoint */ +TAR_EXTERN int f_show_omitted_dirs; /* --show-omitted-dirs */ +TAR_EXTERN char *f_volno_file; /* --volno-file */ +TAR_EXTERN int f_force_local; /* --force-local */ +TAR_EXTERN int f_atime_preserve;/* --atime-preserve */ +TAR_EXTERN int f_compress_block; /* --compress-block */ + +/* + * We default to Unix Standard format rather than 4.2BSD tar format. + * The code can actually produce all three: + * f_standard ANSI standard + * f_oldarch V7 + * neither 4.2BSD + * but we don't bother, since 4.2BSD can read ANSI standard format anyway. + * The only advantage to the "neither" option is that we can cmp our + * output to the output of 4.2BSD tar, for debugging. + */ +#define f_standard (!f_oldarch) + +/* + * Structure for keeping track of filenames and lists thereof. + */ +struct name + { + struct name *next; + short length; /* cached strlen(name) */ + char found; /* A matching file has been found */ + char firstch; /* First char is literally matched */ + char regexp; /* This name is a regexp, not literal */ + char *change_dir; /* JF set with the -C option */ + char *dir_contents; /* JF for f_gnudump */ + char fake; /* dummy entry */ + char name[1]; + }; + +TAR_EXTERN struct name *namelist; /* Points to first name in list */ +TAR_EXTERN struct name *namelast; /* Points to last name in list */ + +TAR_EXTERN int archive; /* File descriptor for archive file */ +TAR_EXTERN int errors; /* # of files in error */ + +TAR_EXTERN char *gnu_dumpfile; + +/* + * Error recovery stuff + */ +TAR_EXTERN char read_error_flag; + + +/* + * Declarations of functions available to the world. + */ +union record *findrec (); +void userec (); +union record *endofrecs (); +void anno (); + +#if defined (HAVE_VPRINTF) && __STDC__ +void msg (char *,...); +void msg_perror (char *,...); +#else +void msg (); +void msg_perror (); +#endif diff --git a/gnu/usr.bin/tar/update.c b/gnu/usr.bin/tar/update.c new file mode 100644 index 000000000000..a64317c666dc --- /dev/null +++ b/gnu/usr.bin/tar/update.c @@ -0,0 +1,585 @@ +/* Update a tar archive. + Copyright (C) 1988, 1992 Free Software Foundation + +This file is part of GNU Tar. + +GNU Tar is free software; you can redistribute it and/or modify +it under the terms of the GNU General Public License as published by +the Free Software Foundation; either version 2, or (at your option) +any later version. + +GNU Tar is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with GNU Tar; see the file COPYING. If not, write to +the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */ + +/* JF implement the 'r' 'u' and 'A' options for tar. */ +/* The 'A' option is my own invention: It means that the file-names are + tar files, and they should simply be appended to the end of the archive. + No attempt is made to block the reads from the args; if they're on raw + tape or something like that, it'll probably lose. . . */ + +#include <sys/types.h> +#include <stdio.h> +#include <errno.h> +#ifndef STDC_HEADERS +extern int errno; +#endif + +#ifdef HAVE_SYS_MTIO_H +#include <sys/ioctl.h> +#include <sys/mtio.h> +#endif + +#ifdef BSD42 +#include <sys/file.h> +#else +#ifndef V7 +#include <fcntl.h> +#endif +#endif + +#ifndef __MSDOS__ +#include <pwd.h> +#include <grp.h> +#endif + +#define STDIN 0 +#define STDOUT 1 + +#include "tar.h" +#include "port.h" +#include "rmt.h" + +int time_to_start_writing = 0; /* We've hit the end of the old stuff, + and its time to start writing new stuff + to the tape. This involves seeking + back one block and re-writing the current + block (which has been changed). */ + +char *output_start; /* Pointer to where we started to write in + the first block we write out. This is used + if we can't backspace the output and have + to null out the first part of the block */ + +extern void skip_file (); +extern void skip_extended_headers (); + +extern union record *head; +extern struct stat hstat; + +void append_file (); +void close_archive (); +int confirm (); +void decode_header (); +void fl_read (); +void fl_write (); +void flush_archive (); +int move_arch (); +struct name *name_scan (); +char *name_from_list (); +void name_expand (); +void name_gather (); +void names_notfound (); +void open_archive (); +int read_header (); +void reset_eof (); +void write_block (); +void write_eot (); + +/* Implement the 'r' (add files to end of archive), and 'u' (add files to + end of archive if they arent there, or are more up to date than the + version in the archive.) commands.*/ +void +update_archive () +{ + int found_end = 0; + int status = 3; + int prev_status; + char *p; + struct name *name; + extern void dump_file (); + + name_gather (); + if (cmd_mode == CMD_UPDATE) + name_expand (); + open_archive (2); /* Open for updating */ + + do + { + prev_status = status; + status = read_header (); + switch (status) + { + case EOF: + found_end = 1; + break; + + case 0: /* A bad record */ + userec (head); + switch (prev_status) + { + case 3: + msg ("This doesn't look like a tar archive."); + /* FALL THROUGH */ + case 2: + case 1: + msg ("Skipping to next header"); + case 0: + break; + } + break; + + /* A good record */ + case 1: + /* printf("File %s\n",head->header.name); */ + /* head->header.name[NAMSIZ-1]='\0'; */ + if (cmd_mode == CMD_UPDATE && (name = name_scan (current_file_name))) + { + + /* struct stat hstat; */ + struct stat nstat; + int head_standard; + + decode_header (head, &hstat, &head_standard, 0); + if (stat (current_file_name, &nstat) < 0) + { + msg_perror ("can't stat %s:", current_file_name); + } + else + { + if (hstat.st_mtime >= nstat.st_mtime) + name->found++; + } + } + userec (head); + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) hstat.st_size); + break; + + case 2: + ar_record = head; + found_end = 1; + break; + } + } + while (!found_end); + + reset_eof (); + time_to_start_writing = 1; + output_start = ar_record->charptr; + + while (p = name_from_list ()) + { + if (f_confirm && !confirm ("add", p)) + continue; + if (cmd_mode == CMD_CAT) + append_file (p); + else + dump_file (p, -1, 1); + } + + write_eot (); + close_archive (); + names_notfound (); +} + +/* Catenate file p to the archive without creating a header for it. It had + better be a tar file or the archive is screwed */ + +void +append_file (p) + char *p; +{ + int fd; + struct stat statbuf; + long bytes_left; + union record *start; + long bufsiz, count; + + if (0 != stat (p, &statbuf) || (fd = open (p, O_RDONLY | O_BINARY)) < 0) + { + msg_perror ("can't open file %s", p); + errors++; + return; + } + + bytes_left = statbuf.st_size; + + while (bytes_left > 0) + { + start = findrec (); + bufsiz = endofrecs ()->charptr - start->charptr; + if (bytes_left < bufsiz) + { + bufsiz = bytes_left; + count = bufsiz % RECORDSIZE; + if (count) + bzero (start->charptr + bytes_left, (int) (RECORDSIZE - count)); + } + count = read (fd, start->charptr, bufsiz); + if (count < 0) + { + msg_perror ("read error at byte %ld reading %d bytes in file %s", statbuf.st_size - bytes_left, bufsiz, p); + exit (EX_ARGSBAD); /* FOO */ + } + bytes_left -= count; + userec (start + (count - 1) / RECORDSIZE); + if (count != bufsiz) + { + msg ("%s: file shrunk by %d bytes, yark!", p, bytes_left); + abort (); + } + } + (void) close (fd); +} + +#ifdef DONTDEF +bprint (fp, buf, num) + FILE *fp; + char *buf; +{ + int c; + + if (num == 0 || num == -1) + return; + fputs (" '", fp); + while (num--) + { + c = *buf++; + if (c == '\\') + fputs ("\\\\", fp); + else if (c >= ' ' && c <= '~') + putc (c, fp); + else + switch (c) + { + case '\n': + fputs ("\\n", fp); + break; + case '\r': + fputs ("\\r", fp); + break; + case '\b': + fputs ("\\b", fp); + break; + case '\0': + /* fputs("\\-",fp); */ + break; + default: + fprintf (fp, "\\%03o", c); + break; + } + } + fputs ("'\n", fp); +} + +#endif + +int number_of_blocks_read = 0; + +int number_of_new_records = 0; +int number_of_records_needed = 0; + +union record *new_block = 0; +union record *save_block = 0; + +void +junk_archive () +{ + int found_stuff = 0; + int status = 3; + int prev_status; + struct name *name; + + /* int dummy_head; */ + int number_of_records_to_skip = 0; + int number_of_records_to_keep = 0; + int number_of_kept_records_in_block; + int sub_status; + extern int write_archive_to_stdout; + + /* fprintf(stderr,"Junk files\n"); */ + name_gather (); + open_archive (2); + + while (!found_stuff) + { + prev_status = status; + status = read_header (); + switch (status) + { + case EOF: + found_stuff = 1; + break; + + case 0: + userec (head); + switch (prev_status) + { + case 3: + msg ("This doesn't look like a tar archive."); + /* FALL THROUGH */ + case 2: + case 1: + msg ("Skipping to next header"); + /* FALL THROUGH */ + case 0: + break; + } + break; + + case 1: + /* head->header.name[NAMSIZ-1] = '\0'; */ + /* fprintf(stderr,"file %s\n",head->header.name); */ + if ((name = name_scan (current_file_name)) == (struct name *) 0) + { + userec (head); + /* fprintf(stderr,"Skip %ld\n",(long)(hstat.st_size)); */ + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) (hstat.st_size)); + break; + } + name->found = 1; + found_stuff = 2; + break; + + case 2: + found_stuff = 1; + break; + } + } + /* fprintf(stderr,"Out of first loop\n"); */ + + if (found_stuff != 2) + { + write_eot (); + close_archive (); + names_notfound (); + return; + } + + if (write_archive_to_stdout) + write_archive_to_stdout = 0; + new_block = (union record *) malloc (blocksize); + if (new_block == 0) + { + msg ("Can't allocate secondary block of %d bytes", blocksize); + exit (EX_SYSTEM); + } + + /* Save away records before this one in this block */ + number_of_new_records = ar_record - ar_block; + number_of_records_needed = blocking - number_of_new_records; + if (number_of_new_records) + bcopy ((void *) ar_block, (void *) new_block, (number_of_new_records) * RECORDSIZE); + + /* fprintf(stderr,"Saved %d recs, need %d more\n",number_of_new_records,number_of_records_needed); */ + userec (head); + if (head->header.isextended) + skip_extended_headers (); + skip_file ((long) (hstat.st_size)); + found_stuff = 0; + /* goto flush_file; */ + + for (;;) + { + /* Fill in a block */ + /* another_file: */ + if (ar_record == ar_last) + { + /* fprintf(stderr,"New block\n"); */ + flush_archive (); + number_of_blocks_read++; + } + sub_status = read_header (); + /* fprintf(stderr,"Header type %d\n",sub_status); */ + + if (sub_status == 2 && f_ignorez) + { + userec (head); + continue; + } + if (sub_status == EOF || sub_status == 2) + { + found_stuff = 1; + bzero (new_block[number_of_new_records].charptr, RECORDSIZE * number_of_records_needed); + number_of_new_records += number_of_records_needed; + number_of_records_needed = 0; + write_block (0); + break; + } + + if (sub_status == 0) + { + msg ("Deleting non-header from archive."); + userec (head); + continue; + } + + /* Found another header. Yipee! */ + /* head->header.name[NAMSIZ-1] = '\0'; */ + /* fprintf(stderr,"File %s ",head->header.name); */ + if (name = name_scan (current_file_name)) + { + name->found = 1; + /* fprintf(stderr,"Flush it\n"); */ + /* flush_file: */ + /* decode_header(head,&hstat,&dummy_head,0); */ + userec (head); + number_of_records_to_skip = (hstat.st_size + RECORDSIZE - 1) / RECORDSIZE; + /* fprintf(stderr,"Flushing %d recs from %s\n",number_of_records_to_skip,head->header.name); */ + + while (ar_last - ar_record <= number_of_records_to_skip) + { + + /* fprintf(stderr,"Block: %d <= %d ",ar_last-ar_record,number_of_records_to_skip); */ + number_of_records_to_skip -= (ar_last - ar_record); + flush_archive (); + number_of_blocks_read++; + /* fprintf(stderr,"Block %d left\n",number_of_records_to_skip); */ + } + ar_record += number_of_records_to_skip; + /* fprintf(stderr,"Final %d\n",number_of_records_to_skip); */ + number_of_records_to_skip = 0; + continue; + } + + /* copy_header: */ + new_block[number_of_new_records] = *head; + number_of_new_records++; + number_of_records_needed--; + number_of_records_to_keep = (hstat.st_size + RECORDSIZE - 1) / RECORDSIZE; + userec (head); + if (number_of_records_needed == 0) + write_block (1); + /* copy_data: */ + number_of_kept_records_in_block = ar_last - ar_record; + if (number_of_kept_records_in_block > number_of_records_to_keep) + number_of_kept_records_in_block = number_of_records_to_keep; + + /* fprintf(stderr,"Need %d kept_in %d keep %d\n",blocking,number_of_kept_records_in_block,number_of_records_to_keep); */ + + while (number_of_records_to_keep) + { + int n; + + if (ar_record == ar_last) + { + /* fprintf(stderr,"Flush. . .\n"); */ + fl_read (); + number_of_blocks_read++; + ar_record = ar_block; + number_of_kept_records_in_block = blocking; + if (number_of_kept_records_in_block > number_of_records_to_keep) + number_of_kept_records_in_block = number_of_records_to_keep; + } + n = number_of_kept_records_in_block; + if (n > number_of_records_needed) + n = number_of_records_needed; + + /* fprintf(stderr,"Copying %d\n",n); */ + bcopy ((void *) ar_record, (void *) (new_block + number_of_new_records), n * RECORDSIZE); + number_of_new_records += n; + number_of_records_needed -= n; + ar_record += n; + number_of_records_to_keep -= n; + number_of_kept_records_in_block -= n; + /* fprintf(stderr,"Now new %d need %d keep %d keep_in %d rec %d/%d\n", + number_of_new_records,number_of_records_needed,number_of_records_to_keep, + number_of_kept_records_in_block,ar_record-ar_block,ar_last-ar_block); */ + + if (number_of_records_needed == 0) + { + write_block (1); + } + } + } + + write_eot (); + close_archive (); + names_notfound (); +} + +void +write_block (f) + int f; +{ + /* fprintf(stderr,"Write block\n"); */ + /* We've filled out a block. Write it out. */ + + /* Backspace back to where we started. . . */ + if (archive != STDIN) + (void) move_arch (-(number_of_blocks_read + 1)); + + save_block = ar_block; + ar_block = new_block; + + if (archive == STDIN) + archive = STDOUT; + fl_write (); + + if (archive == STDOUT) + archive = STDIN; + ar_block = save_block; + + if (f) + { + /* Move the tape head back to where we were */ + if (archive != STDIN) + (void) move_arch (number_of_blocks_read); + number_of_blocks_read--; + } + + number_of_records_needed = blocking; + number_of_new_records = 0; +} + +/* Move archive descriptor by n blocks worth. If n is positive we move + forward, else we move negative. If its a tape, MTIOCTOP had better + work. If its something else, we try to seek on it. If we can't + seek, we lose! */ +int +move_arch (n) + int n; +{ + long cur; + +#ifdef MTIOCTOP + struct mtop t; + int er; + + if (n > 0) + { + t.mt_op = MTFSR; + t.mt_count = n; + } + else + { + t.mt_op = MTBSR; + t.mt_count = -n; + } + if ((er = rmtioctl (archive, MTIOCTOP, &t)) >= 0) + return 1; + if (errno == EIO && (er = rmtioctl (archive, MTIOCTOP, &t)) >= 0) + return 1; +#endif + + cur = rmtlseek (archive, 0L, 1); + cur += blocksize * n; + + /* fprintf(stderr,"Fore to %x\n",cur); */ + if (rmtlseek (archive, cur, 0) != cur) + { + /* Lseek failed. Try a different method */ + msg ("Couldn't re-position archive file."); + exit (EX_BADARCH); + } + return 3; +} diff --git a/gnu/usr.bin/tar/version.c b/gnu/usr.bin/tar/version.c new file mode 100644 index 000000000000..4454f62c8e98 --- /dev/null +++ b/gnu/usr.bin/tar/version.c @@ -0,0 +1 @@ +char version_string[] = "GNU tar version 1.11.2"; diff --git a/gnu/usr.bin/tar/y.tab.h b/gnu/usr.bin/tar/y.tab.h new file mode 100644 index 000000000000..4a541d2c97f7 --- /dev/null +++ b/gnu/usr.bin/tar/y.tab.h @@ -0,0 +1,18 @@ +#define tAGO 257 +#define tDAY 258 +#define tDAYZONE 259 +#define tID 260 +#define tMERIDIAN 261 +#define tMINUTE_UNIT 262 +#define tMONTH 263 +#define tMONTH_UNIT 264 +#define tSEC_UNIT 265 +#define tSNUMBER 266 +#define tUNUMBER 267 +#define tZONE 268 +#define tDST 269 +typedef union { + time_t Number; + enum _MERIDIAN Meridian; +} YYSTYPE; +extern YYSTYPE yylval; |