diff options
| author | Kenichi Handa <handa@m17n.org> | 2007-03-23 05:44:14 +0000 |
|---|---|---|
| committer | Kenichi Handa <handa@m17n.org> | 2007-03-23 05:44:14 +0000 |
| commit | 6482db2eb982c10a91d875719ed5de4d1d994065 (patch) | |
| tree | ac0cd27c72e355cd0f28c2ebf18e69ba6ff6afae /src/regex.c | |
| parent | d85d38392e338f66053a6a6f1017720660239338 (diff) | |
| download | emacs-6482db2eb982c10a91d875719ed5de4d1d994065.tar.gz emacs-6482db2eb982c10a91d875719ed5de4d1d994065.tar.bz2 emacs-6482db2eb982c10a91d875719ed5de4d1d994065.zip | |
(analyse_first): Fix for multibyte characters in "case
charset:" and "case categoryspec:".
Diffstat (limited to 'src/regex.c')
| -rw-r--r-- | src/regex.c | 37 |
1 files changed, 20 insertions, 17 deletions
diff --git a/src/regex.c b/src/regex.c index 4fbd80232c6..81c37730a9a 100644 --- a/src/regex.c +++ b/src/regex.c @@ -4103,17 +4103,21 @@ analyse_first (p, pend, fastmap, multibyte) if (!!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))) ^ not) fastmap[j] = 1; - if ((not && multibyte) - /* Any leading code can possibly start a character +#ifdef emacs + if (/* Any leading code can possibly start a character which doesn't match the specified set of characters. */ - || (CHARSET_RANGE_TABLE_EXISTS_P (&p[-2]) - && CHARSET_RANGE_TABLE_BITS (&p[-2]) != 0)) - /* If we can match a character class, we can match - any multibyte characters. */ + not + || + /* If we can match a character class, we can match any + multibyte characters. */ + (CHARSET_RANGE_TABLE_EXISTS_P (&p[-2]) + && CHARSET_RANGE_TABLE_BITS (&p[-2]) != 0)) + { if (match_any_multibyte_characters == false) { - for (j = 0x80; j < (1 << BYTEWIDTH); j++) + for (j = MIN_MULTIBYTE_LEADING_CODE; + j <= MAX_MULTIBYTE_LEADING_CODE; j++) fastmap[j] = 1; match_any_multibyte_characters = true; } @@ -4145,6 +4149,7 @@ analyse_first (p, pend, fastmap, multibyte) fastmap[j] = 1; } } +#endif break; case syntaxspec: @@ -4167,20 +4172,18 @@ analyse_first (p, pend, fastmap, multibyte) if (!fastmap) break; not = (re_opcode_t)p[-1] == notcategoryspec; k = *p++; - for (j = (multibyte ? 127 : (1 << BYTEWIDTH)); j >= 0; j--) + for (j = (1 << BYTEWIDTH); j >= 0; j--) if ((CHAR_HAS_CATEGORY (j, k)) ^ not) fastmap[j] = 1; - if (multibyte) + /* Any leading code can possibly start a character which + has or doesn't has the specified category. */ + if (match_any_multibyte_characters == false) { - /* Any character set can possibly contain a character - whose category is K (or not). */ - if (match_any_multibyte_characters == false) - { - for (j = 0x80; j < (1 << BYTEWIDTH); j++) - fastmap[j] = 1; - match_any_multibyte_characters = true; - } + for (j = MIN_MULTIBYTE_LEADING_CODE; + j <= MAX_MULTIBYTE_LEADING_CODE; j++) + fastmap[j] = 1; + match_any_multibyte_characters = true; } break; |