Page MenuHomeFreeBSD
Files F166718164

D53978.diff
No OneTemporary
Actions

D53978.diff
View Options

diff --git a/usr/src/lib/libc/regex/engine.c b/usr/src/lib/libc/regex/engine.c
--- a/usr/src/lib/libc/regex/engine.c
+++ b/usr/src/lib/libc/regex/engine.c
@@ -146,6 +146,8 @@
#define NOTE(s) /* nothing */
#endif
+
+
/*
* Given a multibyte string pointed to by start, step back nchar characters
* from current position pointed to by cur.
@@ -241,9 +243,8 @@
/* We depend on not being used for
* for strings of length 1
*/
- while (*--dp == *--pp && pp != mustfirst);
-
- if (*dp == *pp)
+ while (!XCOMP(g, *--dp, *--pp) && pp != mustfirst);
+ if (!XCOMP(g, *dp, *pp))
break;
/* Jump to next possible match */
@@ -256,9 +257,9 @@
return(REG_NOMATCH);
} else {
for (dp = start; dp < stop; dp++)
- if (*dp == g->must[0] &&
+ if (!XCOMP(g, *dp, g->must[0]) &&
stop - dp >= g->mlen &&
- memcmp(dp, g->must, (size_t)g->mlen) == 0)
+ XXMEMCMP(g, dp, g->must, (size_t)g->mlen) == 0)
break;
if (dp == stop) /* we didn't find g->must */
return(REG_NOMATCH);
@@ -452,6 +453,7 @@
case OEND:
assert(nope);
break;
+ case OICHAR:
case OCHAR:
sp += XMBRTOWC(NULL, sp, stop - start, &m->mbs, 0);
break;
@@ -642,10 +644,11 @@
for (ss = startst; !hard && ss < stopst; ss++)
switch (OP(s = m->g->strip[ss])) {
case OCHAR:
+ case OICHAR:
if (sp == stop)
return(NULL);
sp += XMBRTOWC(&wc, sp, stop - sp, &m->mbs, BADCHAR);
- if (wc != OPND(s))
+ if (XCOMP(m->g, wc, OPND(s)))
return(NULL);
break;
case OANY:
@@ -757,7 +760,7 @@
if (sp > stop - len)
return(NULL); /* not enough left to match */
ssp = m->offp + m->pmatch[i].rm_so;
- if (memcmp(sp, ssp, len) != 0)
+ if (XXMEMCMP(m->g, sp, ssp, len) != 0)
return(NULL);
while (m->g->strip[ss] != (sop)SOP(O_BACK, i))
ss++;
@@ -1014,9 +1017,10 @@
assert(pc == stop-1);
break;
case OCHAR:
+ case OICHAR:
/* only characters can match */
assert(!NONCHAR(ch) || ch != OPND(s));
- if (ch == OPND(s))
+ if (!XCOMP(g, ch, OPND(s)))
FWD(aft, bef, 1);
break;
case OBOS:
@@ -1121,6 +1125,8 @@
return(aft);
}
+
+
#ifdef REDEBUG
/*
- print - print a set of states
diff --git a/usr/src/lib/libc/regex/regcomp.c b/usr/src/lib/libc/regex/regcomp.c
--- a/usr/src/lib/libc/regex/regcomp.c
+++ b/usr/src/lib/libc/regex/regcomp.c
@@ -1394,10 +1394,21 @@
{
cset *cs;
- if ((p->g->cflags&REG_ICASE) && iswalpha(ch) && othercase(ch) != ch)
- bothcases(p, ch);
- else if ((ch & OPDMASK) == ch)
- EMIT(OCHAR, ch);
+ int isb;
+ isb = ((p->g->cflags&REG_ICASE) && iswalpha(ch) && othercase(ch) != ch);
+ if ((ch & OPDMASK) == ch) {
+ if (!isb)
+ EMIT(OCHAR, ch);
+ else {
+ /* is icase */
+ if ((ch & 0x7f) == ch)
+ EMIT(OICHAR, tolower(ch));
+ else {
+ /* wide char case insensitive */
+ bothcases(p, ch);
+ }
+ }
+ }
else {
/*
* Kludge: character is too big to fit into an OCHAR operand.
@@ -1877,6 +1888,7 @@
do {
s = *scan++;
switch (OP(s)) {
+ case OICHAR:
case OCHAR: /* sequence member */
if (newlen == 0) { /* new sequence */
memset(&mbs, 0, sizeof(mbs));
@@ -2003,8 +2015,12 @@
scan = start;
memset(&mbs, 0, sizeof(mbs));
while (cp < g->must + g->mlen) {
- while (OP(s = *scan++) != OCHAR)
- continue;
+
+ s = *scan++;
+ while (OP(s) != OCHAR && OP(s) != OICHAR) {
+ s = *scan++;
+ }
+
clen = wcrtomb(cp, OPND(s), &mbs);
assert(clen != (size_t)-1);
cp += clen;
@@ -2124,8 +2140,11 @@
* (notice that we match right to left, so that last character
* is the first one that would be matched).
*/
- for (mindex = 0; mindex < g->mlen; mindex++)
+ for (mindex = 0; mindex < g->mlen; mindex++) {
g->charjump[(int)g->must[mindex]] = g->mlen - mindex - 1;
+ if (g->cflags & REG_ICASE)
+ g->charjump[othercase((int)g->must[mindex])] = g->mlen - mindex - 1;
+ }
}
/*
@@ -2141,6 +2160,7 @@
* Notice that all values here are minus (g->mlen-1), because of the way
* the search algorithm works.
*/
+
static void
computematchjumps(struct parse *p, struct re_guts *g)
{
@@ -2184,7 +2204,7 @@
* substring.
*/
while (suffix < g->mlen
- && g->must[mindex] != g->must[suffix]) {
+ && XCOMP(g, g->must[mindex], g->must[suffix])) {
g->matchjump[suffix] = MIN(g->matchjump[suffix],
g->mlen - mindex - 1);
suffix = pmatches[suffix];
@@ -2246,3 +2266,27 @@
g->iflags |= BAD;
return(maxnest);
}
+
+/**
+ * Compare two UTF-8 strings case-insensitively
+ *
+ * @param s1 First UTF-8 string
+ * @param s2 Second UTF-8 string
+ * @param len Length of second string in bytes
+ * @return 0 if equal, <0 if s1 < s2, >0 if s1 > s2, -2 on error
+ */
+int icasecmp(const char *s1, const char *s2, size_t len) {
+
+ int i = 0;
+
+ while (i < len) {
+
+ if (tolower(s1[i]) != tolower(s2[i])) {
+ return (s1[i] < s2[i]) ? -1 : 1;
+ }
+ i++;
+ }
+
+ // If we've reached the end of both strings, they're equal
+ return 0;
+}
\ No newline at end of file
diff --git a/usr/src/lib/libc/regex/regex2.h b/usr/src/lib/libc/regex/regex2.h
--- a/usr/src/lib/libc/regex/regex2.h
+++ b/usr/src/lib/libc/regex/regex2.h
@@ -107,7 +107,7 @@
#define OEOS (22L<<OPSHIFT) /* end subj. - */
#define OWBND (23L<<OPSHIFT) /* word bound - */
#define ONWBND (24L<<OPSHIFT) /* not bound - */
-
+#define OICHAR (25L<<OPSHIFT) /* both cases ochar */
/*
* Structures for [] character-set representation.
*/
@@ -201,3 +201,12 @@
#define OUT (CHAR_MIN - 1) /* a non-character value */
#define IGN (CHAR_MIN - 2)
#define ISWORD(c) (iswalnum((uch)(c)) || (c) == '_')
+
+int icasecmp(const char *s1, const char *s2, size_t len);
+
+#define XCOMP(g, c1 , c2) \
+ ((g->cflags & REG_ICASE) ? (towlower(c1) != towlower(c2)) : \
+ ((c1) != (c2)))
+#define XXMEMCMP(g, s1, s2, len) \
+ ((g->cflags & REG_ICASE) ? icasecmp(s1, s2, len) : \
+ memcmp(s1, s2, len))

File Metadata

Mime Type
text/plain
Expires
Sun, Aug 16, 8:44 PM (3 h, 8 m)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
36836745
Default Alt Text
D53978.diff (5 KB)

Event Timeline