Mercurial > vim
changeset 4997:8a7d3a73adab v7.3.1243
updated for version 7.3.1243
Problem: New regexp engine: back references in look-behind match don't
work. (Lech Lorens)
Solution: Copy the submatches before a recursive match.
author | Bram Moolenaar <bram@vim.org> |
---|---|
date | Wed, 26 Jun 2013 12:42:44 +0200 |
parents | 5759342cde79 |
children | de6925a3a1b5 |
files | src/regexp_nfa.c src/testdir/test64.in src/testdir/test64.ok src/version.c |
diffstat | 4 files changed, 24 insertions(+), 4 deletions(-) [+] |
line wrap: on
line diff
--- a/src/regexp_nfa.c +++ b/src/regexp_nfa.c @@ -290,10 +290,11 @@ static void nfa_dump __ARGS((nfa_regprog #endif static int *re2post __ARGS((void)); static nfa_state_T *alloc_state __ARGS((int c, nfa_state_T *out, nfa_state_T *out1)); +static void st_error __ARGS((int *postfix, int *end, int *p)); +static int nfa_max_width __ARGS((nfa_state_T *startstate, int depth)); static nfa_state_T *post2nfa __ARGS((int *postfix, int *end, int nfa_calc_size)); static void nfa_postprocess __ARGS((nfa_regprog_T *prog)); static int check_char_class __ARGS((int class, int c)); -static void st_error __ARGS((int *postfix, int *end, int *p)); static void nfa_save_listids __ARGS((nfa_regprog_T *prog, int *list)); static void nfa_restore_listids __ARGS((nfa_regprog_T *prog, int *list)); static int nfa_re_num_cmp __ARGS((long_u val, int op, long_u pos)); @@ -3469,6 +3470,7 @@ typedef struct #ifdef ENABLE_LOG static void log_subsexpr __ARGS((regsubs_T *subs)); static void log_subexpr __ARGS((regsub_T *sub)); +static char *pim_info __ARGS((nfa_pim_T *pim)); static void log_subsexpr(subs) @@ -3508,7 +3510,8 @@ log_subexpr(sub) } static char * -pim_info(nfa_pim_T *pim) +pim_info(pim) + nfa_pim_T *pim; { static char buf[30]; @@ -3532,6 +3535,7 @@ static void clear_sub __ARGS((regsub_T * static void copy_sub __ARGS((regsub_T *to, regsub_T *from)); static void copy_sub_off __ARGS((regsub_T *to, regsub_T *from)); static int sub_equal __ARGS((regsub_T *sub1, regsub_T *sub2)); +static int match_backref __ARGS((regsub_T *sub, int subidx, int *bytelen)); static int has_state_with_pos __ARGS((nfa_list_T *l, nfa_state_T *state, regsubs_T *subs)); static int state_in_list __ARGS((nfa_list_T *l, nfa_state_T *state, regsubs_T *subs)); static void addstate __ARGS((nfa_list_T *l, nfa_state_T *state, regsubs_T *subs, nfa_pim_T *pim, int off)); @@ -4319,8 +4323,6 @@ check_char_class(class, c) return FAIL; } -static int match_backref __ARGS((regsub_T *sub, int subidx, int *bytelen)); - /* * Check for a match with subexpression "subidx". * Return TRUE if it matches. @@ -5195,6 +5197,10 @@ nfa_regmatch(prog, start, submatch, m) || t->state->c == NFA_START_INVISIBLE_BEFORE_FIRST || t->state->c == NFA_START_INVISIBLE_BEFORE_NEG_FIRST) { + /* Copy submatch info for the recursive call, so that + * \1 can be matched. */ + copy_sub_off(&m->norm, &t->subs.norm); + /* * First try matching the invisible match, then what * follows.
--- a/src/testdir/test64.in +++ b/src/testdir/test64.in @@ -380,6 +380,9 @@ STARTTEST :call add(tl, [2, '\(a\)\(b\)\(c\)\(dd\)\(e\)\(f\)\(g\)\(h\)\(i\)\1\2\3\4\5\6\7\8\9', 'xabcddefghiabcddefghix', 'abcddefghiabcddefghi', 'a', 'b', 'c', 'dd', 'e', 'f', 'g', 'h', 'i']) :call add(tl, [2, '\(\d*\)a \1b', ' a b ', 'a b', '']) :call add(tl, [2, '^.\(.\).\_..\1.', "aaa\naaa\nb", "aaa\naaa", 'a']) +:call add(tl, [2, '^.*\.\(.*\)/.\+\(\1\)\@<!$', 'foo.bat/foo.com', 'foo.bat/foo.com', 'bat']) +:call add(tl, [2, '^.*\.\(.*\)/.\+\(\1\)\@<!$', 'foo.bat/foo.bat']) +:call add(tl, [2, '^.*\.\(.*\)/.\+\(\1\)\@<=$', 'foo.bat/foo.bat', 'foo.bat/foo.bat', 'bat', 'bat']) :" :"""" Look-behind with limit :call add(tl, [2, '<\@<=span.', 'xxspanxx<spanyyy', 'spany'])
--- a/src/testdir/test64.ok +++ b/src/testdir/test64.ok @@ -866,6 +866,15 @@ OK 2 - \(\d*\)a \1b OK 0 - ^.\(.\).\_..\1. OK 1 - ^.\(.\).\_..\1. OK 2 - ^.\(.\).\_..\1. +OK 0 - ^.*\.\(.*\)/.\+\(\1\)\@<!$ +OK 1 - ^.*\.\(.*\)/.\+\(\1\)\@<!$ +OK 2 - ^.*\.\(.*\)/.\+\(\1\)\@<!$ +OK 0 - ^.*\.\(.*\)/.\+\(\1\)\@<!$ +OK 1 - ^.*\.\(.*\)/.\+\(\1\)\@<!$ +OK 2 - ^.*\.\(.*\)/.\+\(\1\)\@<!$ +OK 0 - ^.*\.\(.*\)/.\+\(\1\)\@<=$ +OK 1 - ^.*\.\(.*\)/.\+\(\1\)\@<=$ +OK 2 - ^.*\.\(.*\)/.\+\(\1\)\@<=$ OK 0 - <\@<=span. OK 1 - <\@<=span. OK 2 - <\@<=span.