In perl.git, the branch blead has been updated
<http://perl5.git.perl.org/perl.git/commitdiff/d981ef2485de59a9de4d1fe1ddfbd76889abe0b9?hp=b57e41186b2ceb48bef4f0588dcd19e105cc8a38>
- Log -----------------------------------------------------------------
commit d981ef2485de59a9de4d1fe1ddfbd76889abe0b9
Author: Karl Williamson <pub...@khwilliamson.com>
Date: Tue Jan 18 16:08:08 2011 -0700
regexec.c: Remove break statements from macros
This is so future coders won't be tempted to rely on them.
M regexec.c
commit 6895a8aa39da30e5d27a918f8082897ce5f21779
Author: Karl Williamson <pub...@khwilliamson.com>
Date: Tue Jan 18 16:06:47 2011 -0700
regexec.c: Don't rely on break stmts in macros
It is safer and clearer to have the break statement in each case statement
at
the source level
M regexec.c
commit 8e9da4d422fd1fb3711b88105d4e58f6b8f88877
Author: Karl Williamson <pub...@khwilliamson.com>
Date: Tue Jan 18 16:01:11 2011 -0700
regexec.c: Fix /a complements
This showed up only on some systems in the current test suite, but
processing
eg, \D has to care about the target string being utf8.
M regexec.c
-----------------------------------------------------------------------
Summary of changes:
regexec.c | 57 ++++++++++++++++++++++++++++++++++++++++++++++++---------
1 files changed, 48 insertions(+), 9 deletions(-)
diff --git a/regexec.c b/regexec.c
index 748e047..b59b8bf 100644
--- a/regexec.c
+++ b/regexec.c
@@ -1333,8 +1333,7 @@ if ((!reginfo || regtry(reginfo, &s))) \
} \
else { \
REXEC_FBC_CLASS_SCAN(CoNd); \
- } \
- break
+ }
#define REXEC_FBC_CSCAN_PRELOAD(UtFpReLoAd,CoNdUtF8,CoNd) \
if (utf8_target) { \
@@ -1343,8 +1342,7 @@ if ((!reginfo || regtry(reginfo, &s))) \
} \
else { \
REXEC_FBC_CLASS_SCAN(CoNd); \
- } \
- break
+ }
#define REXEC_FBC_CSCAN_TAINT(CoNdUtF8,CoNd) \
PL_reg_flags |= RF_tainted; \
@@ -1353,8 +1351,7 @@ if ((!reginfo || regtry(reginfo, &s))) \
} \
else { \
REXEC_FBC_CLASS_SCAN(CoNd); \
- } \
- break
+ }
#define DUMP_EXEC_POS(li,s,doutf8) \
dump_exec_pos(li,s,(PL_regeol),(PL_bostr),(PL_reg_starttry),doutf8)
@@ -1676,128 +1673,170 @@ S_find_byclass(pTHX_ regexp * prog, const regnode *c,
char *s,
isALNUM_LC_utf8((U8*)s),
isALNUM_LC(*s)
);
+ break;
case ALNUMU:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_WORD(),
swash_fetch(RE_utf8_perl_word,(U8*)s, utf8_target),
isWORDCHAR_L1((U8) *s)
);
+ break;
case ALNUM:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_WORD(),
swash_fetch(RE_utf8_perl_word,(U8*)s, utf8_target),
isWORDCHAR((U8) *s)
);
+ break;
case ALNUMA:
+ /* Don't need to worry about utf8, as it can match only a single
+ * byte invariant character */
REXEC_FBC_CLASS_SCAN( isWORDCHAR_A(*s));
+ break;
case NALNUMU:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_WORD(),
swash_fetch(RE_utf8_perl_word,(U8*)s, utf8_target),
! isWORDCHAR_L1((U8) *s)
);
+ break;
case NALNUM:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_WORD(),
!swash_fetch(RE_utf8_perl_word, (U8*)s, utf8_target),
! isALNUM(*s)
);
+ break;
case NALNUMA:
- REXEC_FBC_UTF8_CLASS_SCAN( !isWORDCHAR_A(*s));
+ REXEC_FBC_CSCAN(
+ !isWORDCHAR_A(*s),
+ !isWORDCHAR_A(*s)
+ );
+ break;
case NALNUML:
REXEC_FBC_CSCAN_TAINT(
!isALNUM_LC_utf8((U8*)s),
!isALNUM_LC(*s)
);
+ break;
case SPACEU:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_SPACE(),
*s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s,
utf8_target),
isSPACE_L1((U8) *s)
);
+ break;
case SPACE:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_SPACE(),
*s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s,
utf8_target),
isSPACE((U8) *s)
);
+ break;
case SPACEA:
+ /* Don't need to worry about utf8, as it can match only a single
+ * byte invariant character */
REXEC_FBC_CLASS_SCAN( isSPACE_A(*s));
+ break;
case SPACEL:
REXEC_FBC_CSCAN_TAINT(
isSPACE_LC_utf8((U8*)s),
isSPACE_LC(*s)
);
+ break;
case NSPACEU:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_SPACE(),
!( *s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s,
utf8_target)),
! isSPACE_L1((U8) *s)
);
+ break;
case NSPACE:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_SPACE(),
!(*s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s,
utf8_target)),
! isSPACE((U8) *s)
);
+ break;
case NSPACEA:
- REXEC_FBC_UTF8_CLASS_SCAN( !isSPACE_A(*s));
+ REXEC_FBC_CSCAN(
+ !isSPACE_A(*s),
+ !isSPACE_A(*s)
+ );
+ break;
case NSPACEL:
REXEC_FBC_CSCAN_TAINT(
!isSPACE_LC_utf8((U8*)s),
!isSPACE_LC(*s)
);
+ break;
case DIGIT:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_POSIX_DIGIT(),
swash_fetch(RE_utf8_posix_digit,(U8*)s, utf8_target),
isDIGIT(*s)
);
+ break;
case DIGITA:
+ /* Don't need to worry about utf8, as it can match only a single
+ * byte invariant character */
REXEC_FBC_CLASS_SCAN( isDIGIT_A(*s));
+ break;
case DIGITL:
REXEC_FBC_CSCAN_TAINT(
isDIGIT_LC_utf8((U8*)s),
isDIGIT_LC(*s)
);
+ break;
case NDIGIT:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_POSIX_DIGIT(),
!swash_fetch(RE_utf8_posix_digit,(U8*)s, utf8_target),
!isDIGIT(*s)
);
+ break;
case NDIGITA:
- REXEC_FBC_UTF8_CLASS_SCAN( !isDIGIT_A(*s));
+ REXEC_FBC_CSCAN(
+ !isDIGIT_A(*s),
+ !isDIGIT_A(*s)
+ );
+ break;
case NDIGITL:
REXEC_FBC_CSCAN_TAINT(
!isDIGIT_LC_utf8((U8*)s),
!isDIGIT_LC(*s)
);
+ break;
case LNBREAK:
REXEC_FBC_CSCAN(
is_LNBREAK_utf8(s),
is_LNBREAK_latin1(s)
);
+ break;
case VERTWS:
REXEC_FBC_CSCAN(
is_VERTWS_utf8(s),
is_VERTWS_latin1(s)
);
+ break;
case NVERTWS:
REXEC_FBC_CSCAN(
!is_VERTWS_utf8(s),
!is_VERTWS_latin1(s)
);
+ break;
case HORIZWS:
REXEC_FBC_CSCAN(
is_HORIZWS_utf8(s),
is_HORIZWS_latin1(s)
);
+ break;
case NHORIZWS:
REXEC_FBC_CSCAN(
!is_HORIZWS_utf8(s),
!is_HORIZWS_latin1(s)
);
+ break;
case AHOCORASICKC:
case AHOCORASICK:
{
--
Perl5 Master Repository
