summaryrefslogtreecommitdiff
path: root/lib/libc/regex/grot
diff options
context:
space:
mode:
Diffstat (limited to 'lib/libc/regex/grot')
-rw-r--r--lib/libc/regex/grot/Makefile124
-rw-r--r--lib/libc/regex/grot/debug.c244
-rw-r--r--lib/libc/regex/grot/limits.h6
-rw-r--r--lib/libc/regex/grot/main.c512
-rwxr-xr-xlib/libc/regex/grot/mkh77
-rw-r--r--lib/libc/regex/grot/split.c318
-rw-r--r--lib/libc/regex/grot/stdlib.h4
-rw-r--r--lib/libc/regex/grot/tests450
8 files changed, 1735 insertions, 0 deletions
diff --git a/lib/libc/regex/grot/Makefile b/lib/libc/regex/grot/Makefile
new file mode 100644
index 000000000000..b7eefb956662
--- /dev/null
+++ b/lib/libc/regex/grot/Makefile
@@ -0,0 +1,124 @@
+# $FreeBSD$
+# You probably want to take -DREDEBUG out of CFLAGS, and put something like
+# -O in, *after* testing (-DREDEBUG strengthens testing by enabling a lot of
+# internal assertion checking). Take -Dconst= out for an ANSI compiler.
+# Do not take -DPOSIX_MISTAKE out. REGCFLAGS isn't important to you (it's
+# for my use in some special contexts).
+
+PATHS= ${.CURDIR}/.. ${.CURDIR}/../../locale ${.CURDIR}/../../../../include
+.PATH: ${PATHS}
+
+CFLAGS+= -DPOSIX_MISTAKE -DREDEBUG $(REGCFLAGS)
+.for incpath in ${PATHS}
+CFLAGS+= -I${incpath}
+.endfor
+
+# If you have an ANSI compiler, take -o out of MKHFLAGS. If you want
+# the Berkeley __P macro, put -b in.
+MKHFLAGS =
+
+LDFLAGS =
+
+# If you have an ANSI environment, take limits.h and stdlib.h out of
+# HMISSING and take memmove out of SRCMISSING and OBJMISSING.
+HMISSING =
+SRCMISSING = split.c
+OBJMISSING = split.o
+H = cclass.h cname.h regex2.h utils.h $(HMISSING)
+REGSRC = regcomp.c regerror.c regexec.c regfree.c engine.c
+SRC = $(REGSRC) debug.c main.c $(SRCMISSING)
+
+# Internal stuff, should not need changing.
+OBJPRODN = regcomp.o regexec.o regerror.o regfree.o
+OBJS = $(OBJPRODN) debug.o main.o $(OBJMISSING)
+
+# Stuff that matters only if you're trying to lint the package.
+LINTFLAGS = -I. -Dstatic= -Dconst= -DREDEBUG
+LINTC = regcomp.c regexec.c regerror.c regfree.c debug.c main.c $(SRCMISSING)
+JUNKLINT =possible pointer alignment|null effect
+
+.SUFFIXES: .ih .h
+.c.ih:
+ ./mkh $(MKHFLAGS) -p $< >$@
+
+default: r
+
+re: $(OBJS)
+ $(CC) $(CFLAGS) $(LDFLAGS) $(OBJS) -o $@
+
+o: $(OBJPRODN)
+
+REGEXHSRC = regex2.h reg*.c
+h: $(REGEXHSRC)
+ ./mkh $(MKHFLAGS) -i _REGEX_H_ $(REGEXHSRC) >regex.tmp
+ cmp -s regex.tmp regex.h 2>/dev/null || cp regex.tmp regex.h
+ rm -f regex.tmp
+
+regcomp.o regexec.o regfree.o debug.o: utils.h regex.h regex2.h
+regcomp.o: cclass.h cname.h regcomp.ih
+regexec.o: engine.c engine.ih
+regerror.o: regerror.ih
+regerror.o: utils.h
+debug.o: debug.ih
+main.o: main.ih
+
+r: re tests
+ ./re <tests
+ ./re -el <tests
+ ./re -er <tests
+
+ra: ./re tests
+ -./re <tests
+ -./re -el <tests
+ -./re -er <tests
+
+rx: ./re tests
+ ./re -x <tests
+ ./re -x -el <tests
+ ./re -x -er <tests
+
+t: ./re tests
+ -time ./re <tests
+ -time ./re -cs <tests
+ -time ./re -el <tests
+ -time ./re -cs -el <tests
+
+e: echeck re tests
+ ./re -d <tests | awk -f echeck
+
+l: $(LINTC)
+ lint $(LINTFLAGS) -h $(LINTC) 2>&1 | egrep -v '$(JUNKLINT)' | tee lint
+
+fullprint:
+ ti README WHATSNEW notes todo | hplist
+ ti *.h | hplist
+ hplist *.c
+ hplist regex.3 regex.7
+
+print:
+ ti README WHATSNEW notes todo | hplist
+ ti *.h | hplist
+ hplist reg*.c engine.c
+
+clean: tidy
+ rm -f *.o *.s *.ih re
+
+tidy:
+ rm -f junk* core dtr regex.tmp lint
+
+spotless: clean
+ rm -f mkh regex.h
+
+mkh: /usr/henry/bin/mkh
+ cp $? $@
+
+DTRH = cclass.h cname.h regex2.h utils.h limits.h stdlib.h
+DTR = README WHATSNEW Makefile Makefile.44bsd mkh regex.3 regex.7 tests $(DTRH) $(SRC)
+dtr: $(DTR)
+ makedtr $(DTR) >$@
+
+cio: $(DTR)
+ cio $(DTR)
+
+rdf: $(DTR)
+ rcsdiff -c $(DTR) 2>&1 | p
diff --git a/lib/libc/regex/grot/debug.c b/lib/libc/regex/grot/debug.c
new file mode 100644
index 000000000000..24bad149f684
--- /dev/null
+++ b/lib/libc/regex/grot/debug.c
@@ -0,0 +1,244 @@
+/* $FreeBSD$
+ */
+#include <stdio.h>
+#include <string.h>
+#include <ctype.h>
+#include <limits.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <regex.h>
+
+#include "utils.h"
+#include "regex2.h"
+#include "debug.ih"
+
+/*
+ - regprint - print a regexp for debugging
+ == void regprint(regex_t *r, FILE *d);
+ */
+void
+regprint(r, d)
+regex_t *r;
+FILE *d;
+{
+ struct re_guts *g = r->re_g;
+ int i;
+ int c;
+ int last;
+ int nincat[NC];
+
+ fprintf(d, "%ld states, %d categories", (long)g->nstates,
+ g->ncategories);
+ fprintf(d, ", first %ld last %ld", (long)g->firststate,
+ (long)g->laststate);
+ if (g->iflags&USEBOL)
+ fprintf(d, ", USEBOL");
+ if (g->iflags&USEEOL)
+ fprintf(d, ", USEEOL");
+ if (g->iflags&BAD)
+ fprintf(d, ", BAD");
+ if (g->nsub > 0)
+ fprintf(d, ", nsub=%ld", (long)g->nsub);
+ if (g->must != NULL)
+ fprintf(d, ", must(%ld) `%*s'", (long)g->mlen, (int)g->mlen,
+ g->must);
+ if (g->backrefs)
+ fprintf(d, ", backrefs");
+ if (g->nplus > 0)
+ fprintf(d, ", nplus %ld", (long)g->nplus);
+ fprintf(d, "\n");
+ s_print(g, d);
+ for (i = 0; i < g->ncategories; i++) {
+ nincat[i] = 0;
+ for (c = CHAR_MIN; c <= CHAR_MAX; c++)
+ if (g->categories[c] == i)
+ nincat[i]++;
+ }
+ fprintf(d, "cc0#%d", nincat[0]);
+ for (i = 1; i < g->ncategories; i++)
+ if (nincat[i] == 1) {
+ for (c = CHAR_MIN; c <= CHAR_MAX; c++)
+ if (g->categories[c] == i)
+ break;
+ fprintf(d, ", %d=%s", i, regchar(c));
+ }
+ fprintf(d, "\n");
+ for (i = 1; i < g->ncategories; i++)
+ if (nincat[i] != 1) {
+ fprintf(d, "cc%d\t", i);
+ last = -1;
+ for (c = CHAR_MIN; c <= CHAR_MAX+1; c++) /* +1 does flush */
+ if (c <= CHAR_MAX && g->categories[c] == i) {
+ if (last < 0) {
+ fprintf(d, "%s", regchar(c));
+ last = c;
+ }
+ } else {
+ if (last >= 0) {
+ if (last != c-1)
+ fprintf(d, "-%s",
+ regchar(c-1));
+ last = -1;
+ }
+ }
+ fprintf(d, "\n");
+ }
+}
+
+/*
+ - s_print - print the strip for debugging
+ == static void s_print(struct re_guts *g, FILE *d);
+ */
+static void
+s_print(g, d)
+struct re_guts *g;
+FILE *d;
+{
+ sop *s;
+ cset *cs;
+ int i;
+ int done = 0;
+ sop opnd;
+ int col = 0;
+ int last;
+ sopno offset = 2;
+# define GAP() { if (offset % 5 == 0) { \
+ if (col > 40) { \
+ fprintf(d, "\n\t"); \
+ col = 0; \
+ } else { \
+ fprintf(d, " "); \
+ col++; \
+ } \
+ } else \
+ col++; \
+ offset++; \
+ }
+
+ if (OP(g->strip[0]) != OEND)
+ fprintf(d, "missing initial OEND!\n");
+ for (s = &g->strip[1]; !done; s++) {
+ opnd = OPND(*s);
+ switch (OP(*s)) {
+ case OEND:
+ fprintf(d, "\n");
+ done = 1;
+ break;
+ case OCHAR:
+ if (strchr("\\|()^$.[+*?{}!<> ", (char)opnd) != NULL)
+ fprintf(d, "\\%c", (char)opnd);
+ else
+ fprintf(d, "%s", regchar((char)opnd));
+ break;
+ case OBOL:
+ fprintf(d, "^");
+ break;
+ case OEOL:
+ fprintf(d, "$");
+ break;
+ case OBOW:
+ fprintf(d, "\\{");
+ break;
+ case OEOW:
+ fprintf(d, "\\}");
+ break;
+ case OANY:
+ fprintf(d, ".");
+ break;
+ case OANYOF:
+ fprintf(d, "[(%ld)", (long)opnd);
+ cs = &g->sets[opnd];
+ last = -1;
+ for (i = 0; i < g->csetsize+1; i++) /* +1 flushes */
+ if (CHIN(cs, i) && i < g->csetsize) {
+ if (last < 0) {
+ fprintf(d, "%s", regchar(i));
+ last = i;
+ }
+ } else {
+ if (last >= 0) {
+ if (last != i-1)
+ fprintf(d, "-%s",
+ regchar(i-1));
+ last = -1;
+ }
+ }
+ fprintf(d, "]");
+ break;
+ case OBACK_:
+ fprintf(d, "(\\<%ld>", (long)opnd);
+ break;
+ case O_BACK:
+ fprintf(d, "<%ld>\\)", (long)opnd);
+ break;
+ case OPLUS_:
+ fprintf(d, "(+");
+ if (OP(*(s+opnd)) != O_PLUS)
+ fprintf(d, "<%ld>", (long)opnd);
+ break;
+ case O_PLUS:
+ if (OP(*(s-opnd)) != OPLUS_)
+ fprintf(d, "<%ld>", (long)opnd);
+ fprintf(d, "+)");
+ break;
+ case OQUEST_:
+ fprintf(d, "(?");
+ if (OP(*(s+opnd)) != O_QUEST)
+ fprintf(d, "<%ld>", (long)opnd);
+ break;
+ case O_QUEST:
+ if (OP(*(s-opnd)) != OQUEST_)
+ fprintf(d, "<%ld>", (long)opnd);
+ fprintf(d, "?)");
+ break;
+ case OLPAREN:
+ fprintf(d, "((<%ld>", (long)opnd);
+ break;
+ case ORPAREN:
+ fprintf(d, "<%ld>))", (long)opnd);
+ break;
+ case OCH_:
+ fprintf(d, "<");
+ if (OP(*(s+opnd)) != OOR2)
+ fprintf(d, "<%ld>", (long)opnd);
+ break;
+ case OOR1:
+ if (OP(*(s-opnd)) != OOR1 && OP(*(s-opnd)) != OCH_)
+ fprintf(d, "<%ld>", (long)opnd);
+ fprintf(d, "|");
+ break;
+ case OOR2:
+ fprintf(d, "|");
+ if (OP(*(s+opnd)) != OOR2 && OP(*(s+opnd)) != O_CH)
+ fprintf(d, "<%ld>", (long)opnd);
+ break;
+ case O_CH:
+ if (OP(*(s-opnd)) != OOR1)
+ fprintf(d, "<%ld>", (long)opnd);
+ fprintf(d, ">");
+ break;
+ default:
+ fprintf(d, "!%d(%d)!", OP(*s), opnd);
+ break;
+ }
+ if (!done)
+ GAP();
+ }
+}
+
+/*
+ - regchar - make a character printable
+ == static char *regchar(int ch);
+ */
+static char * /* -> representation */
+regchar(ch)
+int ch;
+{
+ static char buf[10];
+
+ if (isprint(ch) || ch == ' ')
+ sprintf(buf, "%c", ch);
+ else
+ sprintf(buf, "\\%o", ch);
+ return(buf);
+}
diff --git a/lib/libc/regex/grot/limits.h b/lib/libc/regex/grot/limits.h
new file mode 100644
index 000000000000..6049fdf2e1b6
--- /dev/null
+++ b/lib/libc/regex/grot/limits.h
@@ -0,0 +1,6 @@
+/* $FreeBSD$
+ */
+#define _POSIX2_RE_DUP_MAX 255
+#define CHAR_MIN (-128)
+#define CHAR_MAX 127
+#define CHAR_BIT 8
diff --git a/lib/libc/regex/grot/main.c b/lib/libc/regex/grot/main.c
new file mode 100644
index 000000000000..cb56e48433c9
--- /dev/null
+++ b/lib/libc/regex/grot/main.c
@@ -0,0 +1,512 @@
+/* $FreeBSD$
+ */
+#include <stdio.h>
+#include <string.h>
+#include <sys/types.h>
+#include <regex.h>
+#include <assert.h>
+
+#include "main.ih"
+
+char *progname;
+int debug = 0;
+int line = 0;
+int status = 0;
+
+int copts = REG_EXTENDED;
+int eopts = 0;
+regoff_t startoff = 0;
+regoff_t endoff = 0;
+
+
+extern int split();
+extern void regprint();
+
+/*
+ - main - do the simple case, hand off to regress() for regression
+ */
+main(argc, argv)
+int argc;
+char *argv[];
+{
+ regex_t re;
+# define NS 10
+ regmatch_t subs[NS];
+ char erbuf[100];
+ int err;
+ size_t len;
+ int c;
+ int errflg = 0;
+ int i;
+ extern int optind;
+ extern char *optarg;
+
+ progname = argv[0];
+
+ while ((c = getopt(argc, argv, "c:e:S:E:x")) != EOF)
+ switch (c) {
+ case 'c': /* compile options */
+ copts = options('c', optarg);
+ break;
+ case 'e': /* execute options */
+ eopts = options('e', optarg);
+ break;
+ case 'S': /* start offset */
+ startoff = (regoff_t)atoi(optarg);
+ break;
+ case 'E': /* end offset */
+ endoff = (regoff_t)atoi(optarg);
+ break;
+ case 'x': /* Debugging. */
+ debug++;
+ break;
+ case '?':
+ default:
+ errflg++;
+ break;
+ }
+ if (errflg) {
+ fprintf(stderr, "usage: %s ", progname);
+ fprintf(stderr, "[-c copt][-C][-d] [re]\n");
+ exit(2);
+ }
+
+ if (optind >= argc) {
+ regress(stdin);
+ exit(status);
+ }
+
+ err = regcomp(&re, argv[optind++], copts);
+ if (err) {
+ len = regerror(err, &re, erbuf, sizeof(erbuf));
+ fprintf(stderr, "error %s, %d/%d `%s'\n",
+ eprint(err), len, sizeof(erbuf), erbuf);
+ exit(status);
+ }
+ regprint(&re, stdout);
+
+ if (optind >= argc) {
+ regfree(&re);
+ exit(status);
+ }
+
+ if (eopts&REG_STARTEND) {
+ subs[0].rm_so = startoff;
+ subs[0].rm_eo = strlen(argv[optind]) - endoff;
+ }
+ err = regexec(&re, argv[optind], (size_t)NS, subs, eopts);
+ if (err) {
+ len = regerror(err, &re, erbuf, sizeof(erbuf));
+ fprintf(stderr, "error %s, %d/%d `%s'\n",
+ eprint(err), len, sizeof(erbuf), erbuf);
+ exit(status);
+ }
+ if (!(copts&REG_NOSUB)) {
+ len = (int)(subs[0].rm_eo - subs[0].rm_so);
+ if (subs[0].rm_so != -1) {
+ if (len != 0)
+ printf("match `%.*s'\n", len,
+ argv[optind] + subs[0].rm_so);
+ else
+ printf("match `'@%.1s\n",
+ argv[optind] + subs[0].rm_so);
+ }
+ for (i = 1; i < NS; i++)
+ if (subs[i].rm_so != -1)
+ printf("(%d) `%.*s'\n", i,
+ (int)(subs[i].rm_eo - subs[i].rm_so),
+ argv[optind] + subs[i].rm_so);
+ }
+ exit(status);
+}
+
+/*
+ - regress - main loop of regression test
+ == void regress(FILE *in);
+ */
+void
+regress(in)
+FILE *in;
+{
+ char inbuf[1000];
+# define MAXF 10
+ char *f[MAXF];
+ int nf;
+ int i;
+ char erbuf[100];
+ size_t ne;
+ char *badpat = "invalid regular expression";
+# define SHORT 10
+ char *bpname = "REG_BADPAT";
+ regex_t re;
+
+ while (fgets(inbuf, sizeof(inbuf), in) != NULL) {
+ line++;
+ if (inbuf[0] == '#' || inbuf[0] == '\n')
+ continue; /* NOTE CONTINUE */
+ inbuf[strlen(inbuf)-1] = '\0'; /* get rid of stupid \n */
+ if (debug)
+ fprintf(stdout, "%d:\n", line);
+ nf = split(inbuf, f, MAXF, "\t\t");
+ if (nf < 3) {
+ fprintf(stderr, "bad input, line %d\n", line);
+ exit(1);
+ }
+ for (i = 0; i < nf; i++)
+ if (strcmp(f[i], "\"\"") == 0)
+ f[i] = "";
+ if (nf <= 3)
+ f[3] = NULL;
+ if (nf <= 4)
+ f[4] = NULL;
+ try(f[0], f[1], f[2], f[3], f[4], options('c', f[1]));
+ if (opt('&', f[1])) /* try with either type of RE */
+ try(f[0], f[1], f[2], f[3], f[4],
+ options('c', f[1]) &~ REG_EXTENDED);
+ }
+
+ ne = regerror(REG_BADPAT, (regex_t *)NULL, erbuf, sizeof(erbuf));
+ if (strcmp(erbuf, badpat) != 0 || ne != strlen(badpat)+1) {
+ fprintf(stderr, "end: regerror() test gave `%s' not `%s'\n",
+ erbuf, badpat);
+ status = 1;
+ }
+ ne = regerror(REG_BADPAT, (regex_t *)NULL, erbuf, (size_t)SHORT);
+ if (strncmp(erbuf, badpat, SHORT-1) != 0 || erbuf[SHORT-1] != '\0' ||
+ ne != strlen(badpat)+1) {
+ fprintf(stderr, "end: regerror() short test gave `%s' not `%.*s'\n",
+ erbuf, SHORT-1, badpat);
+ status = 1;
+ }
+ ne = regerror(REG_ITOA|REG_BADPAT, (regex_t *)NULL, erbuf, sizeof(erbuf));
+ if (strcmp(erbuf, bpname) != 0 || ne != strlen(bpname)+1) {
+ fprintf(stderr, "end: regerror() ITOA test gave `%s' not `%s'\n",
+ erbuf, bpname);
+ status = 1;
+ }
+ re.re_endp = bpname;
+ ne = regerror(REG_ATOI, &re, erbuf, sizeof(erbuf));
+ if (atoi(erbuf) != (int)REG_BADPAT) {
+ fprintf(stderr, "end: regerror() ATOI test gave `%s' not `%ld'\n",
+ erbuf, (long)REG_BADPAT);
+ status = 1;
+ } else if (ne != strlen(erbuf)+1) {
+ fprintf(stderr, "end: regerror() ATOI test len(`%s') = %ld\n",
+ erbuf, (long)REG_BADPAT);
+ status = 1;
+ }
+}
+
+/*
+ - try - try it, and report on problems
+ == void try(char *f0, char *f1, char *f2, char *f3, char *f4, int opts);
+ */
+void
+try(f0, f1, f2, f3, f4, opts)
+char *f0;
+char *f1;
+char *f2;
+char *f3;
+char *f4;
+int opts; /* may not match f1 */
+{
+ regex_t re;
+# define NSUBS 10
+ regmatch_t subs[NSUBS];
+# define NSHOULD 15
+ char *should[NSHOULD];
+ int nshould;
+ char erbuf[100];
+ int err;
+ int len;
+ char *type = (opts & REG_EXTENDED) ? "ERE" : "BRE";
+ int i;
+ char *grump;
+ char f0copy[1000];
+ char f2copy[1000];
+
+ strcpy(f0copy, f0);
+ re.re_endp = (opts&REG_PEND) ? f0copy + strlen(f0copy) : NULL;
+ fixstr(f0copy);
+ err = regcomp(&re, f0copy, opts);
+ if (err != 0 && (!opt('C', f1) || err != efind(f2))) {
+ /* unexpected error or wrong error */
+ len = regerror(err, &re, erbuf, sizeof(erbuf));
+ fprintf(stderr, "%d: %s error %s, %d/%d `%s'\n",
+ line, type, eprint(err), len,
+ sizeof(erbuf), erbuf);
+ status = 1;
+ } else if (err == 0 && opt('C', f1)) {
+ /* unexpected success */
+ fprintf(stderr, "%d: %s should have given REG_%s\n",
+ line, type, f2);
+ status = 1;
+ err = 1; /* so we won't try regexec */
+ }
+
+ if (err != 0) {
+ regfree(&re);
+ return;
+ }
+
+ strcpy(f2copy, f2);
+ fixstr(f2copy);
+
+ if (options('e', f1)&REG_STARTEND) {
+ if (strchr(f2, '(') == NULL || strchr(f2, ')') == NULL)
+ fprintf(stderr, "%d: bad STARTEND syntax\n", line);
+ subs[0].rm_so = strchr(f2, '(') - f2 + 1;
+ subs[0].rm_eo = strchr(f2, ')') - f2;
+ }
+ err = regexec(&re, f2copy, NSUBS, subs, options('e', f1));
+
+ if (err != 0 && (f3 != NULL || err != REG_NOMATCH)) {
+ /* unexpected error or wrong error */
+ len = regerror(err, &re, erbuf, sizeof(erbuf));
+ fprintf(stderr, "%d: %s exec error %s, %d/%d `%s'\n",
+ line, type, eprint(err), len,
+ sizeof(erbuf), erbuf);
+ status = 1;
+ } else if (err != 0) {
+ /* nothing more to check */
+ } else if (f3 == NULL) {
+ /* unexpected success */
+ fprintf(stderr, "%d: %s exec should have failed\n",
+ line, type);
+ status = 1;
+ err = 1; /* just on principle */
+ } else if (opts&REG_NOSUB) {
+ /* nothing more to check */
+ } else if ((grump = check(f2, subs[0], f3)) != NULL) {
+ fprintf(stderr, "%d: %s %s\n", line, type, grump);
+ status = 1;
+ err = 1;
+ }
+
+ if (err != 0 || f4 == NULL) {
+ regfree(&re);
+ return;
+ }
+
+ for (i = 1; i < NSHOULD; i++)
+ should[i] = NULL;
+ nshould = split(f4, should+1, NSHOULD-1, ",");
+ if (nshould == 0) {
+ nshould = 1;
+ should[1] = "";
+ }
+ for (i = 1; i < NSUBS; i++) {
+ grump = check(f2, subs[i], should[i]);
+ if (grump != NULL) {
+ fprintf(stderr, "%d: %s $%d %s\n", line,
+ type, i, grump);
+ status = 1;
+ err = 1;
+ }
+ }
+
+ regfree(&re);
+}
+
+/*
+ - options - pick options out of a regression-test string
+ == int options(int type, char *s);
+ */
+int
+options(type, s)
+int type; /* 'c' compile, 'e' exec */
+char *s;
+{
+ char *p;
+ int o = (type == 'c') ? copts : eopts;
+ char *legal = (type == 'c') ? "bisnmp" : "^$#tl";
+
+ for (p = s; *p != '\0'; p++)
+ if (strchr(legal, *p) != NULL)
+ switch (*p) {
+ case 'b':
+ o &= ~REG_EXTENDED;
+ break;
+ case 'i':
+ o |= REG_ICASE;
+ break;
+ case 's':
+ o |= REG_NOSUB;
+ break;
+ case 'n':
+ o |= REG_NEWLINE;
+ break;
+ case 'm':
+ o &= ~REG_EXTENDED;
+ o |= REG_NOSPEC;
+ break;
+ case 'p':
+ o |= REG_PEND;
+ break;
+ case '^':
+ o |= REG_NOTBOL;
+ break;
+ case '$':
+ o |= REG_NOTEOL;
+ break;
+ case '#':
+ o |= REG_STARTEND;
+ break;
+ case 't': /* trace */
+ o |= REG_TRACE;
+ break;
+ case 'l': /* force long representation */
+ o |= REG_LARGE;
+ break;
+ case 'r': /* force backref use */
+ o |= REG_BACKR;
+ break;
+ }
+ return(o);
+}
+
+/*
+ - opt - is a particular option in a regression string?
+ == int opt(int c, char *s);
+ */
+int /* predicate */
+opt(c, s)
+int c;
+char *s;
+{
+ return(strchr(s, c) != NULL);
+}
+
+/*
+ - fixstr - transform magic characters in strings
+ == void fixstr(char *p);
+ */
+void
+fixstr(p)
+char *p;
+{
+ if (p == NULL)
+ return;
+
+ for (; *p != '\0'; p++)
+ if (*p == 'N')
+ *p = '\n';
+ else if (*p == 'T')
+ *p = '\t';
+ else if (*p == 'S')
+ *p = ' ';
+ else if (*p == 'Z')
+ *p = '\0';
+}
+
+/*
+ - check - check a substring match
+ == char *check(char *str, regmatch_t sub, char *should);
+ */
+char * /* NULL or complaint */
+check(str, sub, should)
+char *str;
+regmatch_t sub;
+char *should;
+{
+ int len;
+ int shlen;
+ char *p;
+ static char grump[500];
+ char *at = NULL;
+
+ if (should != NULL && strcmp(should, "-") == 0)
+ should = NULL;
+ if (should != NULL && should[0] == '@') {
+ at = should + 1;
+ should = "";
+ }
+
+ /* check rm_so and rm_eo for consistency */
+ if (sub.rm_so > sub.rm_eo || (sub.rm_so == -1 && sub.rm_eo != -1) ||
+ (sub.rm_so != -1 && sub.rm_eo == -1) ||
+ (sub.rm_so != -1 && sub.rm_so < 0) ||
+ (sub.rm_eo != -1 && sub.rm_eo < 0) ) {
+ sprintf(grump, "start %ld end %ld", (long)sub.rm_so,
+ (long)sub.rm_eo);
+ return(grump);
+ }
+
+ /* check for no match */
+ if (sub.rm_so == -1 && should == NULL)
+ return(NULL);
+ if (sub.rm_so == -1)
+ return("did not match");
+
+ /* check for in range */
+ if (sub.rm_eo > strlen(str)) {
+ sprintf(grump, "start %ld end %ld, past end of string",
+ (long)sub.rm_so, (long)sub.rm_eo);
+ return(grump);
+ }
+
+ len = (int)(sub.rm_eo - sub.rm_so);
+ shlen = (int)strlen(should);
+ p = str + sub.rm_so;
+
+ /* check for not supposed to match */
+ if (should == NULL) {
+ sprintf(grump, "matched `%.*s'", len, p);
+ return(grump);
+ }
+
+ /* check for wrong match */
+ if (len != shlen || strncmp(p, should, (size_t)shlen) != 0) {
+ sprintf(grump, "matched `%.*s' instead", len, p);
+ return(grump);
+ }
+ if (shlen > 0)
+ return(NULL);
+
+ /* check null match in right place */
+ if (at == NULL)
+ return(NULL);
+ shlen = strlen(at);
+ if (shlen == 0)
+ shlen = 1; /* force check for end-of-string */
+ if (strncmp(p, at, shlen) != 0) {
+ sprintf(grump, "matched null at `%.20s'", p);
+ return(grump);
+ }
+ return(NULL);
+}
+
+/*
+ - eprint - convert error number to name
+ == static char *eprint(int err);
+ */
+static char *
+eprint(err)
+int err;
+{
+ static char epbuf[100];
+ size_t len;
+
+ len = regerror(REG_ITOA|err, (regex_t *)NULL, epbuf, sizeof(epbuf));
+ assert(len <= sizeof(epbuf));
+ return(epbuf);
+}
+
+/*
+ - efind - convert error name to number
+ == static int efind(char *name);
+ */
+static int
+efind(name)
+char *name;
+{
+ static char efbuf[100];
+ size_t n;
+ regex_t re;
+
+ sprintf(efbuf, "REG_%s", name);
+ assert(strlen(efbuf) < sizeof(efbuf));
+ re.re_endp = efbuf;
+ (void) regerror(REG_ATOI, &re, efbuf, sizeof(efbuf));
+ return(atoi(efbuf));
+}
diff --git a/lib/libc/regex/grot/mkh b/lib/libc/regex/grot/mkh
new file mode 100755
index 000000000000..1deba7983d5c
--- /dev/null
+++ b/lib/libc/regex/grot/mkh
@@ -0,0 +1,77 @@
+#! /bin/sh
+# mkh - pull headers out of C source
+# $FreeBSD$
+PATH=/bin:/usr/bin ; export PATH
+
+# egrep pattern to pick out marked lines
+egrep='^ =([ ]|$)'
+
+# Sed program to process marked lines into lines for the header file.
+# The markers have already been removed. Two things are done here: removal
+# of backslashed newlines, and some fudging of comments. The first is done
+# because -o needs to have prototypes on one line to strip them down.
+# Getting comments into the output is tricky; we turn C++-style // comments
+# into /* */ comments, after altering any existing */'s to avoid trouble.
+peel=' /\\$/N
+ /\\\n[ ]*/s///g
+ /\/\//s;\*/;* /;g
+ /\/\//s;//\(.*\);/*\1 */;'
+
+for a
+do
+ case "$a" in
+ -o) # old (pre-function-prototype) compiler
+ # add code to comment out argument lists
+ peel="$peel
+ "'/^\([^#\/][^\/]*[a-zA-Z0-9_)]\)(\(.*\))/s;;\1(/*\2*/);'
+ shift
+ ;;
+ -b) # funny Berkeley __P macro
+ peel="$peel
+ "'/^\([^#\/][^\/]*[a-zA-Z0-9_)]\)(\(.*\))/s;;\1 __P((\2));'
+ shift
+ ;;
+ -s) # compiler doesn't like `static foo();'
+ # add code to get rid of the `static'
+ peel="$peel
+ "'/^static[ ][^\/]*[a-zA-Z0-9_)](.*)/s;static.;;'
+ shift
+ ;;
+ -p) # private declarations
+ egrep='^ ==([ ]|$)'
+ shift
+ ;;
+ -i) # wrap in #ifndef, argument is name
+ ifndef="$2"
+ shift ; shift
+ ;;
+ *) break
+ ;;
+ esac
+done
+
+if test " $ifndef" != " "
+then
+ echo "#ifndef $ifndef"
+ echo "#define $ifndef /* never again */"
+fi
+echo "/* ========= begin header generated by $0 ========= */"
+echo '#ifdef __cplusplus'
+echo 'extern "C" {'
+echo '#endif'
+for f
+do
+ echo
+ echo "/* === $f === */"
+ egrep "$egrep" $f | sed 's/^ ==*[ ]//;s/^ ==*$//' | sed "$peel"
+ echo
+done
+echo '#ifdef __cplusplus'
+echo '}'
+echo '#endif'
+echo "/* ========= end header generated by $0 ========= */"
+if test " $ifndef" != " "
+then
+ echo "#endif"
+fi
+exit 0
diff --git a/lib/libc/regex/grot/split.c b/lib/libc/regex/grot/split.c
new file mode 100644
index 000000000000..e001f39ed5ec
--- /dev/null
+++ b/lib/libc/regex/grot/split.c
@@ -0,0 +1,318 @@
+/* $FreeBSD$
+ */
+#include <stdio.h>
+#include <string.h>
+
+/*
+ - split - divide a string into fields, like awk split()
+ = int split(char *string, char *fields[], int nfields, char *sep);
+ */
+int /* number of fields, including overflow */
+split(string, fields, nfields, sep)
+char *string;
+char *fields[]; /* list is not NULL-terminated */
+int nfields; /* number of entries available in fields[] */
+char *sep; /* "" white, "c" single char, "ab" [ab]+ */
+{
+ char *p = string;
+ char c; /* latest character */
+ char sepc = sep[0];
+ char sepc2;
+ int fn;
+ char **fp = fields;
+ char *sepp;
+ int trimtrail;
+
+ /* white space */
+ if (sepc == '\0') {
+ while ((c = *p++) == ' ' || c == '\t')
+ continue;
+ p--;
+ trimtrail = 1;
+ sep = " \t"; /* note, code below knows this is 2 long */
+ sepc = ' ';
+ } else
+ trimtrail = 0;
+ sepc2 = sep[1]; /* now we can safely pick this up */
+
+ /* catch empties */
+ if (*p == '\0')
+ return(0);
+
+ /* single separator */
+ if (sepc2 == '\0') {
+ fn = nfields;
+ for (;;) {
+ *fp++ = p;
+ fn--;
+ if (fn == 0)
+ break;
+ while ((c = *p++) != sepc)
+ if (c == '\0')
+ return(nfields - fn);
+ *(p-1) = '\0';
+ }
+ /* we have overflowed the fields vector -- just count them */
+ fn = nfields;
+ for (;;) {
+ while ((c = *p++) != sepc)
+ if (c == '\0')
+ return(fn);
+ fn++;
+ }
+ /* not reached */
+ }
+
+ /* two separators */
+ if (sep[2] == '\0') {
+ fn = nfields;
+ for (;;) {
+ *fp++ = p;
+ fn--;
+ while ((c = *p++) != sepc && c != sepc2)
+ if (c == '\0') {
+ if (trimtrail && **(fp-1) == '\0')
+ fn++;
+ return(nfields - fn);
+ }
+ if (fn == 0)
+ break;
+ *(p-1) = '\0';
+ while ((c = *p++) == sepc || c == sepc2)
+ continue;
+ p--;
+ }
+ /* we have overflowed the fields vector -- just count them */
+ fn = nfields;
+ while (c != '\0') {
+ while ((c = *p++) == sepc || c == sepc2)
+ continue;
+ p--;
+ fn++;
+ while ((c = *p++) != '\0' && c != sepc && c != sepc2)
+ continue;
+ }
+ /* might have to trim trailing white space */
+ if (trimtrail) {
+ p--;
+ while ((c = *--p) == sepc || c == sepc2)
+ continue;
+ p++;
+ if (*p != '\0') {
+ if (fn == nfields+1)
+ *p = '\0';
+ fn--;
+ }
+ }
+ return(fn);
+ }
+
+ /* n separators */
+ fn = 0;
+ for (;;) {
+ if (fn < nfields)
+ *fp++ = p;
+ fn++;
+ for (;;) {
+ c = *p++;
+ if (c == '\0')
+ return(fn);
+ sepp = sep;
+ while ((sepc = *sepp++) != '\0' && sepc != c)
+ continue;
+ if (sepc != '\0') /* it was a separator */
+ break;
+ }
+ if (fn < nfields)
+ *(p-1) = '\0';
+ for (;;) {
+ c = *p++;
+ sepp = sep;
+ while ((sepc = *sepp++) != '\0' && sepc != c)
+ continue;
+ if (sepc == '\0') /* it wasn't a separator */
+ break;
+ }
+ p--;
+ }
+
+ /* not reached */
+}
+
+#ifdef TEST_SPLIT
+
+
+/*
+ * test program
+ * pgm runs regression
+ * pgm sep splits stdin lines by sep
+ * pgm str sep splits str by sep
+ * pgm str sep n splits str by sep n times
+ */
+int
+main(argc, argv)
+int argc;
+char *argv[];
+{
+ char buf[512];
+ int n;
+# define MNF 10
+ char *fields[MNF];
+
+ if (argc > 4)
+ for (n = atoi(argv[3]); n > 0; n--) {
+ (void) strcpy(buf, argv[1]);
+ }
+ else if (argc > 3)
+ for (n = atoi(argv[3]); n > 0; n--) {
+ (void) strcpy(buf, argv[1]);
+ (void) split(buf, fields, MNF, argv[2]);
+ }
+ else if (argc > 2)
+ dosplit(argv[1], argv[2]);
+ else if (argc > 1)
+ while (fgets(buf, sizeof(buf), stdin) != NULL) {
+ buf[strlen(buf)-1] = '\0'; /* stomp newline */
+ dosplit(buf, argv[1]);
+ }
+ else
+ regress();
+
+ exit(0);
+}
+
+dosplit(string, seps)
+char *string;
+char *seps;
+{
+# define NF 5
+ char *fields[NF];
+ int nf;
+
+ nf = split(string, fields, NF, seps);
+ print(nf, NF, fields);
+}
+
+print(nf, nfp, fields)
+int nf;
+int nfp;
+char *fields[];
+{
+ int fn;
+ int bound;
+
+ bound = (nf > nfp) ? nfp : nf;
+ printf("%d:\t", nf);
+ for (fn = 0; fn < bound; fn++)
+ printf("\"%s\"%s", fields[fn], (fn+1 < nf) ? ", " : "\n");
+}
+
+#define RNF 5 /* some table entries know this */
+struct {
+ char *str;
+ char *seps;
+ int nf;
+ char *fi[RNF];
+} tests[] = {
+ "", " ", 0, { "" },
+ " ", " ", 2, { "", "" },
+ "x", " ", 1, { "x" },
+ "xy", " ", 1, { "xy" },
+ "x y", " ", 2, { "x", "y" },
+ "abc def g ", " ", 5, { "abc", "def", "", "g", "" },
+ " a bcd", " ", 4, { "", "", "a", "bcd" },
+ "a b c d e f", " ", 6, { "a", "b", "c", "d", "e f" },
+ " a b c d ", " ", 6, { "", "a", "b", "c", "d " },
+
+ "", " _", 0, { "" },
+ " ", " _", 2, { "", "" },
+ "x", " _", 1, { "x" },
+ "x y", " _", 2, { "x", "y" },
+ "ab _ cd", " _", 2, { "ab", "cd" },
+ " a_b c ", " _", 5, { "", "a", "b", "c", "" },
+ "a b c_d e f", " _", 6, { "a", "b", "c", "d", "e f" },
+ " a b c d ", " _", 6, { "", "a", "b", "c", "d " },
+
+ "", " _~", 0, { "" },
+ " ", " _~", 2, { "", "" },
+ "x", " _~", 1, { "x" },
+ "x y", " _~", 2, { "x", "y" },
+ "ab _~ cd", " _~", 2, { "ab", "cd" },
+ " a_b c~", " _~", 5, { "", "a", "b", "c", "" },
+ "a b_c d~e f", " _~", 6, { "a", "b", "c", "d", "e f" },
+ "~a b c d ", " _~", 6, { "", "a", "b", "c", "d " },
+
+ "", " _~-", 0, { "" },
+ " ", " _~-", 2, { "", "" },
+ "x", " _~-", 1, { "x" },
+ "x y", " _~-", 2, { "x", "y" },
+ "ab _~- cd", " _~-", 2, { "ab", "cd" },
+ " a_b c~", " _~-", 5, { "", "a", "b", "c", "" },
+ "a b_c-d~e f", " _~-", 6, { "a", "b", "c", "d", "e f" },
+ "~a-b c d ", " _~-", 6, { "", "a", "b", "c", "d " },
+
+ "", " ", 0, { "" },
+ " ", " ", 2, { "", "" },
+ "x", " ", 1, { "x" },
+ "xy", " ", 1, { "xy" },
+ "x y", " ", 2, { "x", "y" },
+ "abc def g ", " ", 4, { "abc", "def", "g", "" },
+ " a bcd", " ", 3, { "", "a", "bcd" },
+ "a b c d e f", " ", 6, { "a", "b", "c", "d", "e f" },
+ " a b c d ", " ", 6, { "", "a", "b", "c", "d " },
+
+ "", "", 0, { "" },
+ " ", "", 0, { "" },
+ "x", "", 1, { "x" },
+ "xy", "", 1, { "xy" },
+ "x y", "", 2, { "x", "y" },
+ "abc def g ", "", 3, { "abc", "def", "g" },
+ "\t a bcd", "", 2, { "a", "bcd" },
+ " a \tb\t c ", "", 3, { "a", "b", "c" },
+ "a b c d e ", "", 5, { "a", "b", "c", "d", "e" },
+ "a b\tc d e f", "", 6, { "a", "b", "c", "d", "e f" },
+ " a b c d e f ", "", 6, { "a", "b", "c", "d", "e f " },
+
+ NULL, NULL, 0, { NULL },
+};
+
+regress()
+{
+ char buf[512];
+ int n;
+ char *fields[RNF+1];
+ int nf;
+ int i;
+ int printit;
+ char *f;
+
+ for (n = 0; tests[n].str != NULL; n++) {
+ (void) strcpy(buf, tests[n].str);
+ fields[RNF] = NULL;
+ nf = split(buf, fields, RNF, tests[n].seps);
+ printit = 0;
+ if (nf != tests[n].nf) {
+ printf("split `%s' by `%s' gave %d fields, not %d\n",
+ tests[n].str, tests[n].seps, nf, tests[n].nf);
+ printit = 1;
+ } else if (fields[RNF] != NULL) {
+ printf("split() went beyond array end\n");
+ printit = 1;
+ } else {
+ for (i = 0; i < nf && i < RNF; i++) {
+ f = fields[i];
+ if (f == NULL)
+ f = "(NULL)";
+ if (strcmp(f, tests[n].fi[i]) != 0) {
+ printf("split `%s' by `%s' field %d is `%s', not `%s'\n",
+ tests[n].str, tests[n].seps,
+ i, fields[i], tests[n].fi[i]);
+ printit = 1;
+ }
+ }
+ }
+ if (printit)
+ print(nf, RNF, fields);
+ }
+}
+#endif
diff --git a/lib/libc/regex/grot/stdlib.h b/lib/libc/regex/grot/stdlib.h
new file mode 100644
index 000000000000..7e58a4564423
--- /dev/null
+++ b/lib/libc/regex/grot/stdlib.h
@@ -0,0 +1,4 @@
+/* $FreeBSD$
+ */
+extern char *malloc();
+extern char *realloc();
diff --git a/lib/libc/regex/grot/tests b/lib/libc/regex/grot/tests
new file mode 100644
index 000000000000..07e9dfb8539d
--- /dev/null
+++ b/lib/libc/regex/grot/tests
@@ -0,0 +1,450 @@
+# regular expression test set
+# $FreeBSD$
+# Lines are at least three fields, separated by one or more tabs. "" stands
+# for an empty field. First field is an RE. Second field is flags. If
+# C flag given, regcomp() is expected to fail, and the third field is the
+# error name (minus the leading REG_).
+#
+# Otherwise it is expected to succeed, and the third field is the string to
+# try matching it against. If there is no fourth field, the match is
+# expected to fail. If there is a fourth field, it is the substring that
+# the RE is expected to match. If there is a fifth field, it is a comma-
+# separated list of what the subexpressions should match, with - indicating
+# no match for that one. In both the fourth and fifth fields, a (sub)field
+# starting with @ indicates that the (sub)expression is expected to match
+# a null string followed by the stuff after the @; this provides a way to
+# test where null strings match. The character `N' in REs and strings
+# is newline, `S' is space, `T' is tab, `Z' is NUL.
+#
+# The full list of flags:
+# - placeholder, does nothing
+# b RE is a BRE, not an ERE
+# & try it as both an ERE and a BRE
+# C regcomp() error expected, third field is error name
+# i REG_ICASE
+# m ("mundane") REG_NOSPEC
+# s REG_NOSUB (not really testable)
+# n REG_NEWLINE
+# ^ REG_NOTBOL
+# $ REG_NOTEOL
+# # REG_STARTEND (see below)
+# p REG_PEND
+#
+# For REG_STARTEND, the start/end offsets are those of the substring
+# enclosed in ().
+
+# basics
+a & a a
+abc & abc abc
+abc|de - abc abc
+a|b|c - abc a
+
+# parentheses and perversions thereof
+a(b)c - abc abc
+a\(b\)c b abc abc
+a( C EPAREN
+a( b a( a(
+a\( - a( a(
+a\( bC EPAREN
+a\(b bC EPAREN
+a(b C EPAREN
+a(b b a(b a(b
+# gag me with a right parenthesis -- 1003.2 goofed here (my fault, partly)
+a) - a) a)
+) - ) )
+# end gagging (in a just world, those *should* give EPAREN)
+a) b a) a)
+a\) bC EPAREN
+\) bC EPAREN
+a()b - ab ab
+a\(\)b b ab ab
+
+# anchoring and REG_NEWLINE
+^abc$ & abc abc
+a^b - a^b
+a^b b a^b a^b
+a$b - a$b
+a$b b a$b a$b
+^ & abc @abc
+$ & abc @
+^$ & "" @
+$^ - "" @
+\($\)\(^\) b "" @
+# stop retching, those are legitimate (although disgusting)
+^^ - "" @
+$$ - "" @
+b$ & abNc
+b$ &n abNc b
+^b$ & aNbNc
+^b$ &n aNbNc b
+^$ &n aNNb @Nb
+^$ n abc
+^$ n abcN @
+$^ n aNNb @Nb
+\($\)\(^\) bn aNNb @Nb
+^^ n^ aNNb @Nb
+$$ n aNNb @NN
+^a ^ a
+a$ $ a
+^a ^n aNb
+^b ^n aNb b
+a$ $n bNa
+b$ $n bNa b
+a*(^b$)c* - b b
+a*\(^b$\)c* b b b
+
+# certain syntax errors and non-errors
+| C EMPTY
+| b | |
+* C BADRPT
+* b * *
++ C BADRPT
+? C BADRPT
+"" &C EMPTY
+() - abc @abc
+\(\) b abc @abc
+a||b C EMPTY
+|ab C EMPTY
+ab| C EMPTY
+(|a)b C EMPTY
+(a|)b C EMPTY
+(*a) C BADRPT
+(+a) C BADRPT
+(?a) C BADRPT
+({1}a) C BADRPT
+\(\{1\}a\) bC BADRPT
+(a|*b) C BADRPT
+(a|+b) C BADRPT
+(a|?b) C BADRPT
+(a|{1}b) C BADRPT
+^* C BADRPT
+^* b * *
+^+ C BADRPT
+^? C BADRPT
+^{1} C BADRPT
+^\{1\} bC BADRPT
+
+# metacharacters, backslashes
+a.c & abc abc
+a[bc]d & abd abd
+a\*c & a*c a*c
+a\\b & a\b a\b
+a\\\*b & a\*b a\*b
+a\bc & abc abc
+a\ &C EESCAPE
+a\\bc & a\bc a\bc
+\{ bC BADRPT
+# trailing $ is a peculiar special case for the BRE code
+a$ & a a
+a$ & a$
+a\$ & a
+a\$ & a$ a$
+a\\$ & a
+a\\$ & a$
+a\\$ & a\$
+a\\$ & a\ a\
+
+# back references, ugh
+a\(b\)\2c bC ESUBREG
+a\(b\1\)c bC ESUBREG
+a\(b*\)c\1d b abbcbbd abbcbbd bb
+a\(b*\)c\1d b abbcbd
+a\(b*\)c\1d b abbcbbbd
+^\(.\)\1 b abc
+a\([bc]\)\1d b abcdabbd abbd b
+a\(\([bc]\)\2\)*d b abbccd abbccd
+a\(\([bc]\)\2\)*d b abbcbd
+# actually, this next one probably ought to fail, but the spec is unclear
+a\(\(b\)*\2\)*d b abbbd abbbd
+# here is a case that no NFA implementation does right
+\(ab*\)[ab]*\1 b ababaaa ababaaa a
+# check out normal matching in the presence of back refs
+\(a\)\1bcd b aabcd aabcd
+\(a\)\1bc*d b aabcd aabcd
+\(a\)\1bc*d b aabd aabd
+\(a\)\1bc*d b aabcccd aabcccd
+\(a\)\1bc*[ce]d b aabcccd aabcccd
+^\(a\)\1b\(c\)*cd$ b aabcccd aabcccd
+
+# ordinary repetitions
+ab*c & abc abc
+ab+c - abc abc
+ab?c - abc abc
+a\(*\)b b a*b a*b
+a\(**\)b b ab ab
+a\(***\)b bC BADRPT
+*a b *a *a
+**a b a a
+***a bC BADRPT
+
+# the dreaded bounded repetitions
+{ & { {
+{abc & {abc {abc
+{1 C BADRPT
+{1} C BADRPT
+a{b & a{b a{b
+a{1}b - ab ab
+a\{1\}b b ab ab
+a{1,}b - ab ab
+a\{1,\}b b ab ab
+a{1,2}b - aab aab
+a\{1,2\}b b aab aab
+a{1 C EBRACE
+a\{1 bC EBRACE
+a{1a C EBRACE
+a\{1a bC EBRACE
+a{1a} C BADBR
+a\{1a\} bC BADBR
+a{,2} - a{,2} a{,2}
+a\{,2\} bC BADBR
+a{,} - a{,} a{,}
+a\{,\} bC BADBR
+a{1,x} C BADBR
+a\{1,x\} bC BADBR
+a{1,x C EBRACE
+a\{1,x bC EBRACE
+a{300} C BADBR
+a\{300\} bC BADBR
+a{1,0} C BADBR
+a\{1,0\} bC BADBR
+ab{0,0}c - abcac ac
+ab\{0,0\}c b abcac ac
+ab{0,1}c - abcac abc
+ab\{0,1\}c b abcac abc
+ab{0,3}c - abbcac abbc
+ab\{0,3\}c b abbcac abbc
+ab{1,1}c - acabc abc
+ab\{1,1\}c b acabc abc
+ab{1,3}c - acabc abc
+ab\{1,3\}c b acabc abc
+ab{2,2}c - abcabbc abbc
+ab\{2,2\}c b abcabbc abbc
+ab{2,4}c - abcabbc abbc
+ab\{2,4\}c b abcabbc abbc
+((a{1,10}){1,10}){1,10} - a a a,a
+((a{1,10}){1,10}){1,10}bb - aaaaaaaaaaaaaaaaaaaaaaaaaaaaaabb aaaaaaaaaaaaaaaaaaaaaaaaaaaaaabb
+
+# multiple repetitions
+a** &C BADRPT
+a++ C BADRPT
+a?? C BADRPT
+a*+ C BADRPT
+a*? C BADRPT
+a+* C BADRPT
+a+? C BADRPT
+a?* C BADRPT
+a?+ C BADRPT
+a{1}{1} C BADRPT
+a*{1} C BADRPT
+a+{1} C BADRPT
+a?{1} C BADRPT
+a{1}* C BADRPT
+a{1}+ C BADRPT
+a{1}? C BADRPT
+a*{b} - a{b} a{b}
+a\{1\}\{1\} bC BADRPT
+a*\{1\} bC BADRPT
+a\{1\}* bC BADRPT
+
+# brackets, and numerous perversions thereof
+a[b]c & abc abc
+a[ab]c & abc abc
+a[^ab]c & adc adc
+a[]b]c & a]c a]c
+a[[b]c & a[c a[c
+a[-b]c & a-c a-c
+a[^]b]c & adc adc
+a[^-b]c & adc adc
+a[b-]c & a-c a-c
+a[b &C EBRACK
+a[] &C EBRACK
+a[1-3]c & a2c a2c
+a[3-1]c &C ERANGE
+a[1-3-5]c &C ERANGE
+a[[.-.]--]c & a-c a-c
+a[1- &C ERANGE
+a[[. &C EBRACK
+a[[.x &C EBRACK
+a[[.x. &C EBRACK
+a[[.x.] &C EBRACK
+a[[.x.]] & ax ax
+a[[.x,.]] &C ECOLLATE
+a[[.one.]]b & a1b a1b
+a[[.notdef.]]b &C ECOLLATE
+a[[.].]]b & a]b a]b
+a[[:alpha:]]c & abc abc
+a[[:notdef:]]c &C ECTYPE
+a[[: &C EBRACK
+a[[:alpha &C EBRACK
+a[[:alpha:] &C EBRACK
+a[[:alpha,:] &C ECTYPE
+a[[:]:]]b &C ECTYPE
+a[[:-:]]b &C ECTYPE
+a[[:alph:]] &C ECTYPE
+a[[:alphabet:]] &C ECTYPE
+[[:alnum:]]+ - -%@a0X- a0X
+[[:alpha:]]+ - -%@aX0- aX
+[[:blank:]]+ - aSSTb SST
+[[:cntrl:]]+ - aNTb NT
+[[:digit:]]+ - a019b 019
+[[:graph:]]+ - Sa%bS a%b
+[[:lower:]]+ - AabC ab
+[[:print:]]+ - NaSbN aSb
+[[:punct:]]+ - S%-&T %-&
+[[:space:]]+ - aSNTb SNT
+[[:upper:]]+ - aBCd BC
+[[:xdigit:]]+ - p0f3Cq 0f3C
+a[[=b=]]c & abc abc
+a[[= &C EBRACK
+a[[=b &C EBRACK
+a[[=b= &C EBRACK
+a[[=b=] &C EBRACK
+a[[=b,=]] &C ECOLLATE
+a[[=one=]]b & a1b a1b
+
+# complexities
+a(((b)))c - abc abc
+a(b|(c))d - abd abd
+a(b*|c)d - abbd abbd
+# just gotta have one DFA-buster, of course
+a[ab]{20} - aaaaabaaaabaaaabaaaab aaaaabaaaabaaaabaaaab
+# and an inline expansion in case somebody gets tricky
+a[ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab] - aaaaabaaaabaaaabaaaab aaaaabaaaabaaaabaaaab
+# and in case somebody just slips in an NFA...
+a[ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab][ab](wee|week)(knights|night) - aaaaabaaaabaaaabaaaabweeknights aaaaabaaaabaaaabaaaabweeknights
+# fish for anomalies as the number of states passes 32
+12345678901234567890123456789 - a12345678901234567890123456789b 12345678901234567890123456789
+123456789012345678901234567890 - a123456789012345678901234567890b 123456789012345678901234567890
+1234567890123456789012345678901 - a1234567890123456789012345678901b 1234567890123456789012345678901
+12345678901234567890123456789012 - a12345678901234567890123456789012b 12345678901234567890123456789012
+123456789012345678901234567890123 - a123456789012345678901234567890123b 123456789012345678901234567890123
+# and one really big one, beyond any plausible word width
+1234567890123456789012345678901234567890123456789012345678901234567890 - a1234567890123456789012345678901234567890123456789012345678901234567890b 1234567890123456789012345678901234567890123456789012345678901234567890
+# fish for problems as brackets go past 8
+[ab][cd][ef][gh][ij][kl][mn] - xacegikmoq acegikm
+[ab][cd][ef][gh][ij][kl][mn][op] - xacegikmoq acegikmo
+[ab][cd][ef][gh][ij][kl][mn][op][qr] - xacegikmoqy acegikmoq
+[ab][cd][ef][gh][ij][kl][mn][op][q] - xacegikmoqy acegikmoq
+
+# subtleties of matching
+abc & xabcy abc
+a\(b\)?c\1d b acd
+aBc i Abc Abc
+a[Bc]*d i abBCcd abBCcd
+0[[:upper:]]1 &i 0a1 0a1
+0[[:lower:]]1 &i 0A1 0A1
+a[^b]c &i abc
+a[^b]c &i aBc
+a[^b]c &i adc adc
+[a]b[c] - abc abc
+[a]b[a] - aba aba
+[abc]b[abc] - abc abc
+[abc]b[abd] - abd abd
+a(b?c)+d - accd accd
+(wee|week)(knights|night) - weeknights weeknights
+(we|wee|week|frob)(knights|night|day) - weeknights weeknights
+a[bc]d - xyzaaabcaababdacd abd
+a[ab]c - aaabc abc
+abc s abc abc
+
+# subexpressions
+a(b)(c)d - abcd abcd b,c
+a(((b)))c - abc abc b,b,b
+a(b|(c))d - abd abd b,-
+a(b*|c|e)d - abbd abbd bb
+a(b*|c|e)d - acd acd c
+a(b*|c|e)d - ad ad @d
+a(b?)c - abc abc b
+a(b?)c - ac ac @c
+a(b+)c - abc abc b
+a(b+)c - abbbc abbbc bbb
+a(b*)c - ac ac @c
+(a|ab)(bc([de]+)f|cde) - abcdef abcdef a,bcdef,de
+# the regression tester only asks for 9 subexpressions
+a(b)(c)(d)(e)(f)(g)(h)(i)(j)k - abcdefghijk abcdefghijk b,c,d,e,f,g,h,i,j
+a(b)(c)(d)(e)(f)(g)(h)(i)(j)(k)l - abcdefghijkl abcdefghijkl b,c,d,e,f,g,h,i,j,k
+a([bc]?)c - abc abc b
+a([bc]?)c - ac ac @c
+a([bc]+)c - abc abc b
+a([bc]+)c - abcc abcc bc
+a([bc]+)bc - abcbc abcbc bc
+a(bb+|b)b - abb abb b
+a(bbb+|bb+|b)b - abb abb b
+a(bbb+|bb+|b)b - abbb abbb bb
+a(bbb+|bb+|b)bb - abbb abbb b
+(.*).* - abcdef abcdef abcdef
+(a*)* - bc @b @b
+
+# do we get the right subexpression when it is used more than once?
+a(b|c)*d - ad ad -
+a(b|c)*d - abcd abcd c
+a(b|c)+d - abd abd b
+a(b|c)+d - abcd abcd c
+a(b|c?)+d - ad ad @d
+a(b|c?)+d - abcd abcd @d
+a(b|c){0,0}d - ad ad -
+a(b|c){0,1}d - ad ad -
+a(b|c){0,1}d - abd abd b
+a(b|c){0,2}d - ad ad -
+a(b|c){0,2}d - abcd abcd c
+a(b|c){0,}d - ad ad -
+a(b|c){0,}d - abcd abcd c
+a(b|c){1,1}d - abd abd b
+a(b|c){1,1}d - acd acd c
+a(b|c){1,2}d - abd abd b
+a(b|c){1,2}d - abcd abcd c
+a(b|c){1,}d - abd abd b
+a(b|c){1,}d - abcd abcd c
+a(b|c){2,2}d - acbd acbd b
+a(b|c){2,2}d - abcd abcd c
+a(b|c){2,4}d - abcd abcd c
+a(b|c){2,4}d - abcbd abcbd b
+a(b|c){2,4}d - abcbcd abcbcd c
+a(b|c){2,}d - abcd abcd c
+a(b|c){2,}d - abcbd abcbd b
+a(b+|((c)*))+d - abd abd @d,@d,-
+a(b+|((c)*))+d - abcd abcd @d,@d,-
+
+# check out the STARTEND option
+[abc] &# a(b)c b
+[abc] &# a(d)c
+[abc] &# a(bc)d b
+[abc] &# a(dc)d c
+. &# a()c
+b.*c &# b(bc)c bc
+b.* &# b(bc)c bc
+.*c &# b(bc)c bc
+
+# plain strings, with the NOSPEC flag
+abc m abc abc
+abc m xabcy abc
+abc m xyz
+a*b m aba*b a*b
+a*b m ab
+"" mC EMPTY
+
+# cases involving NULs
+aZb & a a
+aZb &p a
+aZb &p# (aZb) aZb
+aZ*b &p# (ab) ab
+a.b &# (aZb) aZb
+a.* &# (aZb)c aZb
+
+# word boundaries (ick)
+[[:<:]]a & a a
+[[:<:]]a & ba
+[[:<:]]a & -a a
+a[[:>:]] & a a
+a[[:>:]] & ab
+a[[:>:]] & a- a
+[[:<:]]a.c[[:>:]] & axcd-dayc-dazce-abc abc
+[[:<:]]a.c[[:>:]] & axcd-dayc-dazce-abc-q abc
+[[:<:]]a.c[[:>:]] & axc-dayc-dazce-abc axc
+
+# past problems
+(A[1])|(A[2])|(A[3])|(A[4])|(A[5])|(A[6])|(A[7])|(A[8])|(A[9])|(A[A]) - A1 A1
+abcdefghijklmnop i abcdefghijklmnop abcdefghijklmnop
+abcdefghijklmnopqrstuv i abcdefghijklmnopqrstuv abcdefghijklmnopqrstuv
+(ALAK)|(ALT[AB])|(CC[123]1)|(CM[123]1)|(GAMC)|(LC[23][EO ])|(SEM[1234])|(SL[ES][12])|(SLWW)|(SLF )|(SLDT)|(VWH[12])|(WH[34][EW])|(WP1[ESN]) - CC11 CC11
+CC[13]1|a{21}[23][EO][123][Es][12]a{15}aa[34][EW]aaaaaaa[X]a - CC11 CC11
generated by cgit v1.2.3 (git 2.25.1) at 2025-11-04 05:19:16 +0000