Patch 8.2.2278
Problem: Falling back to old regexp engine can some patterns.
Solution: Do not fall back once [[:lower:]] or [[:upper:]] is used.
(Christian Brabandt, closes #7572)
Files: src/regexp.c, src/regexp_nfa.c, src/testdir/test_regexp_utf8.vim
*** ../vim-8.2.2277/src/regexp.c 2020-09-07 18:53:18.383974577 +0200
--- src/regexp.c 2021-01-02 17:34:22.627683939 +0100
***************
*** 294,299 ****
--- 294,300 ----
static char_u *regparse; // Input-scan pointer.
static int regnpar; // () count.
+ static int wants_nfa; // regex should use NFA engine
#ifdef FEAT_SYN_HL
static int regnzpar; // \z() count.
static int re_has_z; // \z item detected
***************
*** 381,386 ****
--- 382,390 ----
static char_u *cstrchr(char_u *, int);
static int re_mult_next(char *what);
static int reg_iswordc(int);
+ #ifdef FEAT_EVAL
+ static void report_re_switch(char_u *pat);
+ #endif
static regengine_T bt_regengine;
static regengine_T nfa_regengine;
***************
*** 2662,2668 ****
if (prog == NULL)
{
#ifdef BT_REGEXP_DEBUG_LOG
! if (regexp_engine != BACKTRACKING_ENGINE) // debugging log for NFA
{
FILE *f;
f = fopen(BT_REGEXP_DEBUG_LOG_NAME, "a");
--- 2666,2672 ----
if (prog == NULL)
{
#ifdef BT_REGEXP_DEBUG_LOG
! if (regexp_engine == BACKTRACKING_ENGINE) // debugging log for BT
engine
{
FILE *f;
f = fopen(BT_REGEXP_DEBUG_LOG_NAME, "a");
***************
*** 2686,2691 ****
--- 2690,2698 ----
&& called_emsg == called_emsg_before)
{
regexp_engine = BACKTRACKING_ENGINE;
+ #ifdef FEAT_EVAL
+ report_re_switch(expr);
+ #endif
prog = bt_regengine.regcomp(expr, re_flags);
}
}
*** ../vim-8.2.2277/src/regexp_nfa.c 2020-12-21 18:23:56.687287184 +0100
--- src/regexp_nfa.c 2021-01-02 17:40:10.658093030 +0100
***************
*** 253,258 ****
--- 253,264 ----
static int *post_start; // holds the postfix form of r.e.
static int *post_end;
static int *post_ptr;
+
+ // Set when the pattern should use the NFA engine.
+ // E.g. [[:upper:]] only allows 8bit characters for BT engine,
+ // while NFA engine handles multibyte characters correctly.
+ static int wants_nfa;
+
static int nstate; // Number of states in the NFA.
static int istate; // Index in the state vector, used in alloc_state()
***************
*** 306,311 ****
--- 312,318 ----
return FAIL;
post_ptr = post_start;
post_end = post_start + nstate_max;
+ wants_nfa = FALSE;
rex.nfa_has_zend = FALSE;
rex.nfa_has_backref = FALSE;
***************
*** 1707,1712 ****
--- 1714,1720 ----
EMIT(NFA_CLASS_GRAPH);
break;
case CLASS_LOWER:
+ wants_nfa = TRUE;
EMIT(NFA_CLASS_LOWER);
break;
case CLASS_PRINT:
***************
*** 1719,1724 ****
--- 1727,1733 ----
EMIT(NFA_CLASS_SPACE);
break;
case CLASS_UPPER:
+ wants_nfa = TRUE;
EMIT(NFA_CLASS_UPPER);
break;
case CLASS_XDIGIT:
***************
*** 2137,2145 ****
// The engine is very inefficient (uses too many states) when the
// maximum is much larger than the minimum and when the maximum is
! // large. Bail out if we can use the other engine.
if ((nfa_re_flags & RE_AUTO)
! && (maxval > 500 || maxval > minval + 200))
return FAIL;
// Ignore previous call to nfa_regatom()
--- 2146,2160 ----
// The engine is very inefficient (uses too many states) when the
// maximum is much larger than the minimum and when the maximum is
! // large. However, when maxval is MAX_LIMIT, it is okay, as this
! // will emit NFA_STAR.
! // Bail out if we can use the other engine, but only, when the
! // pattern does not need the NFA engine like (e.g. [[:upper:]]\{2,\}
! // does not work with with characters > 8 bit with the BT engine)
if ((nfa_re_flags & RE_AUTO)
! && (maxval > 500 || maxval > minval + 200)
! && (maxval != MAX_LIMIT && minval < 200)
! && !wants_nfa)
return FAIL;
// Ignore previous call to nfa_regatom()
*** ../vim-8.2.2277/src/testdir/test_regexp_utf8.vim 2020-12-21
14:54:28.844116987 +0100
--- src/testdir/test_regexp_utf8.vim 2021-01-02 17:34:22.627683939 +0100
***************
*** 510,515 ****
--- 510,561 ----
bwipe!
endfunc
+ " Check that [[:upper:]] matches for automatic engine
+ func Test_match_char_class_upper()
+ new
+ let _engine=®expengine
+ " Test 1: [[:upper:]]\{2,\}
+ set regexpengine=0
+ call setline(1, ['05. ПЕСНЯ О ГЕРОЯХ муз. А. Давиденко, М. Коваля и Б.
Шехтера ...', '05. PJESNJA O GJEROJAKH mus. A. Davidjenko, M. Kovalja i B.
Shjekhtjera ...'])
+ call cursor(1,1)
+ let search_cmd='norm /\<[[:upper:]]\{2,\}\>' .. "\<CR>"
+ exe search_cmd
+ call assert_equal(4, searchcount().total, 'TEST 1')
+ set regexpengine=1
+ exe search_cmd
+ call assert_equal(2, searchcount().total, 'TEST 1')
+ set regexpengine=2
+ exe search_cmd
+ call assert_equal(4, searchcount().total, 'TEST 1')
+
+ " Test 2: [[:upper:]].\+
+ let search_cmd='norm /\<[[:upper:]].\+\>' .. "\<CR>"
+ set regexpengine=0
+ exe search_cmd
+ call assert_equal(2, searchcount().total, 'TEST 2')
+ set regexpengine=1
+ exe search_cmd
+ call assert_equal(1, searchcount().total, 'TEST 2')
+ set regexpengine=2
+ exe search_cmd
+ call assert_equal(2, searchcount().total, 'TEST 2')
+
+ " Test 3: [[:lower:]]\+
+ let search_cmd='norm /\<[[:lower:]]\+\>' .. "\<CR>"
+ set regexpengine=0
+ exe search_cmd
+ call assert_equal(4, searchcount().total, 'TEST 3 lower')
+ set regexpengine=1
+ exe search_cmd
+ call assert_equal(2, searchcount().total, 'TEST 3 lower')
+ set regexpengine=2
+ exe search_cmd
+ call assert_equal(4, searchcount().total, 'TEST 3 lower')
+
+ " clean up
+ let ®expengine=_engine
+ bwipe!
+ endfunc
" vim: shiftwidth=2 sts=2 expandtab
*** ../vim-8.2.2277/src/version.c 2021-01-02 17:06:12.538765972 +0100
--- src/version.c 2021-01-02 17:35:54.359249438 +0100
***************
*** 752,753 ****
--- 752,755 ----
{ /* Add new patch number below this line */
+ /**/
+ 2278,
/**/
--
How many light bulbs does it take to change a person?
/// Bram Moolenaar -- [email protected] -- http://www.Moolenaar.net \\\
/// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\ an exciting new programming language -- http://www.Zimbu.org ///
\\\ help me help AIDS victims -- http://ICCF-Holland.org ///
--
--
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php
---
You received this message because you are subscribed to the Google Groups
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/vim_dev/202101021644.102GiI1Y1893653%40masaka.moolenaar.net.
