aniani/vim
0
0
Fork 1
mirror of https://github.com/vim/vim.git synced 2025-09-24 03:44:06 -04:00
Code Activity

patch 8.2.0938: NFA regexp uses tolower ()to compare ignore-case

Browse Source 
Problem:    NFA regexp uses tolower() to compare ignore-case. (Thayne McCombs)
Solution:   Use utf_fold() when possible. (ref. neovim #12456)
This commit is contained in:
Bram Moolenaar
2020-06-09 19:34:54 +02:00
parent dcf59c37d0
commit 59de417b90
5 changed files with 30 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/diff.c
View File

@@ -747,7 +747,7 @@ diff_write_buffer(buf_T *buf, diffin_T *din)
// xdiff doesn't support ignoring case, fold-case the text. // xdiff doesn't support ignoring case, fold-case the text.
c = PTR2CHAR(s); c = PTR2CHAR(s);
c = enc_utf8 ? utf_fold(c) : MB_TOLOWER(c); c = MB_CASEFOLD(c);
orig_len = mb_ptr2len(s); orig_len = mb_ptr2len(s);
if (mb_char2bytes(c, cbuf) != orig_len) if (mb_char2bytes(c, cbuf) != orig_len)
// TODO: handle byte length difference // TODO: handle byte length difference

1
src/macros.h
View File

@@ -93,6 +93,7 @@
#define MB_ISUPPER(c) vim_isupper(c) #define MB_ISUPPER(c) vim_isupper(c)
#define MB_TOLOWER(c) vim_tolower(c) #define MB_TOLOWER(c) vim_tolower(c)
#define MB_TOUPPER(c) vim_toupper(c) #define MB_TOUPPER(c) vim_toupper(c)
#define MB_CASEFOLD(c) (enc_utf8 ? utf_fold(c) : MB_TOLOWER(c))
// Use our own isdigit() replacement, because on MS-Windows isdigit() returns // Use our own isdigit() replacement, because on MS-Windows isdigit() returns
// non-zero for superscript 1. Also avoids that isdigit() crashes for numbers // non-zero for superscript 1. Also avoids that isdigit() crashes for numbers

14
src/regexp_nfa.c
View File

@@ -5459,7 +5459,7 @@ find_match_text(colnr_T startcol, int regstart, char_u *match_text)
{ {
c1 = PTR2CHAR(match_text + len1); c1 = PTR2CHAR(match_text + len1);
c2 = PTR2CHAR(rex.line + col + len2); c2 = PTR2CHAR(rex.line + col + len2);
if (c1 != c2 && (!rex.reg_ic || MB_TOLOWER(c1) != MB_TOLOWER(c2))) if (c1 != c2 && (!rex.reg_ic || MB_CASEFOLD(c1) != MB_CASEFOLD(c2)))
{ {
match = FALSE; match = FALSE;
break; break;
@@ -6271,11 +6271,11 @@ nfa_regmatch(
} }
if (rex.reg_ic) if (rex.reg_ic)
{ {
int curc_low = MB_TOLOWER(curc); int curc_low = MB_CASEFOLD(curc);
int done = FALSE; int done = FALSE;
for ( ; c1 <= c2; ++c1) for ( ; c1 <= c2; ++c1)
if (MB_TOLOWER(c1) == curc_low) if (MB_CASEFOLD(c1) == curc_low)
{ {
result = result_if_matched; result = result_if_matched;
done = TRUE; done = TRUE;
@@ -6287,8 +6287,8 @@ nfa_regmatch(
} }
else if (state->c < 0 ? check_char_class(state->c, curc) else if (state->c < 0 ? check_char_class(state->c, curc)
: (curc == state->c : (curc == state->c
|| (rex.reg_ic && MB_TOLOWER(curc) || (rex.reg_ic && MB_CASEFOLD(curc)
== MB_TOLOWER(state->c)))) == MB_CASEFOLD(state->c))))
{ {
result = result_if_matched; result = result_if_matched;
break; break;
@@ -6713,7 +6713,7 @@ nfa_regmatch(
result = (c == curc); result = (c == curc);
if (!result && rex.reg_ic) if (!result && rex.reg_ic)
result = MB_TOLOWER(c) == MB_TOLOWER(curc); result = MB_CASEFOLD(c) == MB_CASEFOLD(curc);
// If rex.reg_icombine is not set only skip over the character // If rex.reg_icombine is not set only skip over the character
// itself. When it is set skip over composing characters. // itself. When it is set skip over composing characters.
if (result && enc_utf8 && !rex.reg_icombine) if (result && enc_utf8 && !rex.reg_icombine)
@@ -6882,7 +6882,7 @@ nfa_regmatch(
// cheaper than adding a state that won't match. // cheaper than adding a state that won't match.
c = PTR2CHAR(rex.input + clen); c = PTR2CHAR(rex.input + clen);
if (c != prog->regstart && (!rex.reg_ic if (c != prog->regstart && (!rex.reg_ic
|| MB_TOLOWER(c) != MB_TOLOWER(prog->regstart))) || MB_CASEFOLD(c) != MB_CASEFOLD(prog->regstart)))
{ {
#ifdef ENABLE_LOG #ifdef ENABLE_LOG
fprintf(log_fd, " Skipping start state, regstart does not match\n"); fprintf(log_fd, " Skipping start state, regstart does not match\n");

19
src/testdir/test_regexp_utf8.vim
View File

@@ -355,4 +355,23 @@ func Test_ambiwidth()
set regexpengine& ambiwidth& set regexpengine& ambiwidth&
endfunc endfunc
func Run_regexp_ignore_case()
call assert_equal('iIİ', substitute('iIİ', '\([iIİ]\)', '\1', 'g'))
call assert_equal('iIx', substitute('iIİ', '\c\([İ]\)', 'x', 'g'))
call assert_equal('xxİ', substitute('iIİ', '\(i\c\)', 'x', 'g'))
call assert_equal('iIx', substitute('iIİ', '\(İ\c\)', 'x', 'g'))
call assert_equal('iIx', substitute('iIİ', '\c\(\%u0130\)', 'x', 'g'))
call assert_equal('iIx', substitute('iIİ', '\c\([\u0130]\)', 'x', 'g'))
call assert_equal('iIx', substitute('iIİ', '\c\([\u012f-\u0131]\)', 'x', 'g'))
endfunc
func Test_regexp_ignore_case()
set regexpengine=1
call Run_regexp_ignore_case()
set regexpengine=2
call Run_regexp_ignore_case()
set regexpengine&
endfunc
" vim: shiftwidth=2 sts=2 expandtab " vim: shiftwidth=2 sts=2 expandtab

2
src/version.c
View File

@@ -754,6 +754,8 @@ static char *(features[]) =
static int included_patches[] = static int included_patches[] =
{ /* Add new patch number below this line */ { /* Add new patch number below this line */
/**/
938,
/**/ /**/
937, 937,
/**/ /**/