src-test2/usr.bin/script/script.c, branch release/3.3.0_cvs FreeBSD source tree https://cgit-dev.freebsd.org/src-test2/atom?h=release%2F3.3.0_cvs 1999-09-16T08:26:13Z This commit was manufactured by cvs2svn to create tag 1999-09-16T08:26:13Z cvs2svn cvs2svn@FreeBSD.org 1999-09-16T08:26:13Z urn:sha1:5be28542797ac26e17de490f50e0830003596d38 'RELENG_3_3_0_RELEASE'. This commit was manufactured to restore the state of the 3.3-RELEASE image. Releases prior to 5.3-RELEASE are omitting the secure/ and crypto/ subdirs. $Id$ -> $FreeBSD$ 1999-08-29T15:57:45Z Peter Wemm peter@FreeBSD.org 1999-08-29T15:57:45Z urn:sha1:a92bee191d618605d749ddc8e410d0ae66d03752 MFC: Pass correct name of shell as argv[0]. 1999-08-11T23:56:38Z Sheldon Hearn sheldonh@FreeBSD.org 1999-08-11T23:56:38Z urn:sha1:a3a03755c0c459d2ec62ab6e196708835531c7c1 Fix the following bugs: 1998-09-19T09:45:42Z Dag-Erling Smørgrav des@FreeBSD.org 1998-09-19T09:45:42Z urn:sha1:b1c669700c0f52b1c3c886958fe6d75b8ac389f7 - if a command was specified and script(1) failed to execute it, it would print the name of your shell in the error message instead of that of the command that failed. - since finish() was installed as a SIGCHLD handler, it would often run before the main loop had had time to process the last few bytes of output. This resulted in very strange truncated error messages. - script(1) would almost always return with an exit status of 0, even if the command returned a non-zero exit status. This broke my 'build world, install it and rebuild the kernel' scripts because 'make installworld' would run even if 'make buildworld' had failed. Rewrite script as a select loop rather than as a twin reader/write process. 1998-03-08T14:19:18Z Peter Wemm peter@FreeBSD.org 1998-03-08T14:19:18Z urn:sha1:e8eb82a81cc4c0aa5731030989a8647bdc655cc0 This allows simple logging of keys sent to a session (turned on with -k). Also allow specifying the script file flush interval. Fix typo while copying patch from the 2.2 system that it came from 1997-12-30T01:20:08Z Peter Wemm peter@FreeBSD.org 1997-12-30T01:20:08Z urn:sha1:0e604e984ac44b6279909e5444afee66cea13435 Use execvp rather than execv so that full paths are not needed. script(1) has annoyed me with it's inflexable command argument parsing 1997-12-29T13:31:46Z Peter Wemm peter@FreeBSD.org 1997-12-29T13:31:46Z urn:sha1:51afb8df2cd90c6887e3177ee5756c8b53f78660 since I first saw it. I finally needed to pass arguments through to the spawned command badly enough (and urgently) that I threw this together. Add usage(). Use err(3) instead of local redefinition. 1997-08-08T12:24:49Z Philippe Charnier charnier@FreeBSD.org 1997-08-08T12:24:49Z urn:sha1:236d2f558312262d86504309292a2358283f626c compare return value from getopt against -1 rather than EOF, per the final 1997-03-29T04:34:07Z Warner Losh imp@FreeBSD.org 1997-03-29T04:34:07Z urn:sha1:1c8af8787354e20c2b38cab5801698133ff8b403 posix standard on the topic. Don't use __dead or __pure in user code. They were obfuscations 1996-09-14T03:00:32Z Bruce Evans bde@FreeBSD.org 1996-09-14T03:00:32Z urn:sha1:eaa86f9d7fcbe6b82b1defa62f31071324c836ed for gcc >= 2.5 and no-ops for gcc >= 2.6. Converted to use __dead2 or __pure2 where it wasn't already done, except in math.h where use of __pure was mostly wrong.