diff options
Diffstat (limited to 'src/nvim/regexp.c')
| -rw-r--r-- | src/nvim/regexp.c | 26 |
1 files changed, 18 insertions, 8 deletions
diff --git a/src/nvim/regexp.c b/src/nvim/regexp.c index fd1b6116f2..193c68860d 100644 --- a/src/nvim/regexp.c +++ b/src/nvim/regexp.c @@ -258,6 +258,7 @@ #define RE_MARK 207 /* mark cmp Match mark position */ #define RE_VISUAL 208 /* Match Visual area */ +#define RE_COMPOSING 209 // any composing characters /* * Magic characters have a special meaning, they don't match literally. @@ -1256,12 +1257,6 @@ static regprog_T *bt_regcomp(char_u *expr, int re_flags) if (reg(REG_NOPAREN, &flags) == NULL) return NULL; - /* Small enough for pointer-storage convention? */ -#ifdef SMALL_MALLOC /* 16 bit storage allocation */ - if (regsize >= 65536L - 256L) - EMSG_RET_NULL(_("E339: Pattern too long")); -#endif - /* Allocate space. */ bt_regprog_T *r = xmalloc(sizeof(bt_regprog_T) + regsize); @@ -2030,6 +2025,10 @@ static char_u *regatom(int *flagp) ret = regnode(RE_VISUAL); break; + case 'C': + ret = regnode(RE_COMPOSING); + break; + /* \%[abc]: Emit as a list of branches, all ending at the last * branch which matches nothing. */ case '[': @@ -4105,10 +4104,12 @@ regmatch ( status = RA_NOMATCH; } } - // Check for following composing character. + // Check for following composing character, unless %C + // follows (skips over all composing chars). if (status != RA_NOMATCH && enc_utf8 && UTF_COMPOSINGLIKE(reginput, reginput + len) - && !ireg_icombine) { + && !ireg_icombine + && OP(next) != RE_COMPOSING) { // raaron: This code makes a composing character get // ignored, which is the correct behavior (sometimes) // for voweled Hebrew texts. @@ -4173,6 +4174,15 @@ regmatch ( status = RA_NOMATCH; break; + case RE_COMPOSING: + if (enc_utf8) { + // Skip composing characters. + while (utf_iscomposing(utf_ptr2char(reginput))) { + mb_cptr_adv(reginput); + } + } + break; + case NOTHING: break; |