diff -ur src/dfa.c ../../../new/grep/grep/src/dfa.c
--- src/dfa.c 2004-12-08 19:06:42.000000000 -0800
+++ ../../../new/grep/grep/src/dfa.c 2004-12-10 15:07:47.000000000 -0800
@@ -707,6 +707,38 @@
return strncmp(s, lexptr, len) == 0;
}
+extern int __collate_load_error;
+extern char *__collate_substitute(char *);
+
+char *try_collating_thing(char const *ct, int ct_len, int *chars_used) {
+ char *term;
+ if (*ct == '.') {
+ term = strnstr(++ct, ".]", ct_len);
+ } else if (*ct == '=') {
+ term = strnstr(++ct, "=]", ct_len);
+ } else {
+ term = NULL;
+ }
+ if (!term) {
+ *chars_used = 0;
+ return NULL;
+ }
+ int slen = 1 + (term - ct);
+ *chars_used = 2 + slen;
+ char *t = malloc(slen);
+ if (!t) {
+ return NULL;
+ }
+ strlcpy(t, ct, slen);
+ if (__collate_load_error) {
+ return t;
+ }
+ char *r = __collate_substitute(t);
+ free(t);
+ return r;
+}
+
+
static token
lex (void)
{
@@ -991,7 +1023,7 @@
if (MB_CUR_MAX > 1)
{
/* In multibyte environment a bracket expression may contain
- multibyte characters, which must be treated as characters
+ multibyte characters, which must be treated as characters
(not bytes). So we parse it by parse_bracket_exp_mb(). */
parse_bracket_exp_mb();
return lasttok = MBCSET;
@@ -1014,20 +1046,39 @@
characters. We can do this because we assume
regex has checked for syntax errors before
dfa is ever called. */
- if (c == '[' && (syntax_bits & RE_CHAR_CLASSES))
- for (c1 = 0; prednames[c1].name; ++c1)
- if (looking_at(prednames[c1].name))
- {
- int (*pred) PARAMS ((int)) = prednames[c1].pred;
-
- for (c2 = 0; c2 < NOTCHAR; ++c2)
- if ((*pred)(c2))
- setbit_case_fold (c2, ccl);
- lexptr += strlen(prednames[c1].name);
- lexleft -= strlen(prednames[c1].name);
- FETCH(c1, _("Unbalanced ["));
- goto skip;
- }
+ if (c == '[' && (syntax_bits & RE_CHAR_CLASSES)) {
+ if (lexleft >= 1 && (*lexptr == '.' || *lexptr == '=')) {
+ int used = 0;
+ char *match = try_collating_thing(lexptr, lexleft, &used);
+ if (!match) {
+ dfaerror(_("invalid collating element or class"));
+ } else {
+ char *cp = match;
+ for(; *cp; ++cp) {
+ setbit(*cp, ccl);
+ }
+ free(match);
+ }
+ lexptr += used;
+ lexleft -= used;
+ FETCH(c1, _("Unbalanced ["));
+ goto skip;
+ } else {
+ for (c1 = 0; prednames[c1].name; ++c1)
+ if (looking_at(prednames[c1].name))
+ {
+ int (*pred) PARAMS ((int)) = prednames[c1].pred;
+
+ for (c2 = 0; c2 < NOTCHAR; ++c2)
+ if ((*pred)(c2))
+ setbit_case_fold (c2, ccl);
+ lexptr += strlen(prednames[c1].name);
+ lexleft -= strlen(prednames[c1].name);
+ FETCH(c1, _("Unbalanced ["));
+ goto skip;
+ }
+ }
+ }
if (c == '\\' && (syntax_bits & RE_BACKSLASH_ESCAPE_IN_LISTS))
FETCH(c, _("Unbalanced ["));
FETCH(c1, _("Unbalanced ["));