|
|
2c161b |
From 1a076cf2730d3a1fbb174af6f56c554691f4cdc3 Mon Sep 17 00:00:00 2001
|
|
|
2c161b |
From: Yves Orton <demerphq@gmail.com>
|
|
|
2c161b |
Date: Mon, 19 Feb 2018 13:49:46 +1100
|
|
|
2c161b |
Subject: [PATCH] v5.24.3: fix TRIE_READ_CHAR and DECL_TRIE_TYPE to account for
|
|
|
2c161b |
non-utf8 target
|
|
|
2c161b |
MIME-Version: 1.0
|
|
|
2c161b |
Content-Type: text/plain; charset=UTF-8
|
|
|
2c161b |
Content-Transfer-Encoding: 8bit
|
|
|
2c161b |
|
|
|
2c161b |
Petr Pisar: Ported to 5.24.0.
|
|
|
2c161b |
|
|
|
2c161b |
Signed-off-by: Petr Písař <ppisar@redhat.com>
|
|
|
2c161b |
---
|
|
|
2c161b |
regexec.c | 14 ++++++++++----
|
|
|
2c161b |
t/re/re_tests | 1 +
|
|
|
2c161b |
2 files changed, 11 insertions(+), 4 deletions(-)
|
|
|
2c161b |
|
|
|
2c161b |
diff --git a/regexec.c b/regexec.c
|
|
|
2c161b |
index 3fd12c4..05e148f 100644
|
|
|
2c161b |
--- a/regexec.c
|
|
|
2c161b |
+++ b/regexec.c
|
|
|
2c161b |
@@ -1451,7 +1451,7 @@ Perl_re_intuit_start(pTHX_
|
|
|
2c161b |
#define DECL_TRIE_TYPE(scan) \
|
|
|
2c161b |
const enum { trie_plain, trie_utf8, trie_utf8_fold, trie_latin_utf8_fold, \
|
|
|
2c161b |
trie_utf8_exactfa_fold, trie_latin_utf8_exactfa_fold, \
|
|
|
2c161b |
- trie_utf8l, trie_flu8 } \
|
|
|
2c161b |
+ trie_utf8l, trie_flu8, trie_flu8_latin } \
|
|
|
2c161b |
trie_type = ((scan->flags == EXACT) \
|
|
|
2c161b |
? (utf8_target ? trie_utf8 : trie_plain) \
|
|
|
2c161b |
: (scan->flags == EXACTL) \
|
|
|
2c161b |
@@ -1461,10 +1461,12 @@ Perl_re_intuit_start(pTHX_
|
|
|
2c161b |
? trie_utf8_exactfa_fold \
|
|
|
2c161b |
: trie_latin_utf8_exactfa_fold) \
|
|
|
2c161b |
: (scan->flags == EXACTFLU8 \
|
|
|
2c161b |
- ? trie_flu8 \
|
|
|
2c161b |
+ ? (utf8_target \
|
|
|
2c161b |
+ ? trie_flu8 \
|
|
|
2c161b |
+ : trie_flu8_latin) \
|
|
|
2c161b |
: (utf8_target \
|
|
|
2c161b |
? trie_utf8_fold \
|
|
|
2c161b |
- : trie_latin_utf8_fold)))
|
|
|
2c161b |
+ : trie_latin_utf8_fold)))
|
|
|
2c161b |
|
|
|
2c161b |
/* 'uscan' is set to foldbuf, and incremented, so below the end of uscan is
|
|
|
2c161b |
* 'foldbuf+sizeof(foldbuf)' */
|
|
|
2c161b |
@@ -1475,7 +1477,7 @@ STMT_START {
|
|
|
2c161b |
switch (trie_type) { \
|
|
|
2c161b |
case trie_flu8: \
|
|
|
2c161b |
_CHECK_AND_WARN_PROBLEMATIC_LOCALE; \
|
|
|
2c161b |
- if (utf8_target && UTF8_IS_ABOVE_LATIN1(*uc)) { \
|
|
|
2c161b |
+ if (UTF8_IS_ABOVE_LATIN1(*uc)) { \
|
|
|
2c161b |
_CHECK_AND_OUTPUT_WIDE_LOCALE_UTF8_MSG(uc, uc_end - uc); \
|
|
|
2c161b |
} \
|
|
|
2c161b |
goto do_trie_utf8_fold; \
|
|
|
2c161b |
@@ -1497,10 +1499,14 @@ STMT_START {
|
|
|
2c161b |
uscan = foldbuf + skiplen; \
|
|
|
2c161b |
} \
|
|
|
2c161b |
break; \
|
|
|
2c161b |
+ case trie_flu8_latin: \
|
|
|
2c161b |
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE; \
|
|
|
2c161b |
+ goto do_trie_latin_utf8_fold; \
|
|
|
2c161b |
case trie_latin_utf8_exactfa_fold: \
|
|
|
2c161b |
flags |= FOLD_FLAGS_NOMIX_ASCII; \
|
|
|
2c161b |
/* FALLTHROUGH */ \
|
|
|
2c161b |
case trie_latin_utf8_fold: \
|
|
|
2c161b |
+ do_trie_latin_utf8_fold: \
|
|
|
2c161b |
if ( foldlen>0 ) { \
|
|
|
2c161b |
uvc = utf8n_to_uvchr( (const U8*) uscan, foldlen, &len, uniflags ); \
|
|
|
2c161b |
foldlen -= len; \
|
|
|
2c161b |
diff --git a/t/re/re_tests b/t/re/re_tests
|
|
|
2c161b |
index 7e8522d..ab7ddbb 100644
|
|
|
2c161b |
--- a/t/re/re_tests
|
|
|
2c161b |
+++ b/t/re/re_tests
|
|
|
2c161b |
@@ -1968,6 +1968,7 @@ ab(?#Comment){2}c abbc y $& abbc
|
|
|
2c161b |
(?:.||)(?|)000000000@ 000000000@ y $& 000000000@ # [perl #126405]
|
|
|
2c161b |
aa$|a(?R)a|a aaa y $& aaa # [perl 128420] recursive matches
|
|
|
2c161b |
(?:\1|a)([bcd])\1(?:(?R)|e)\1 abbaccaddedcb y $& abbaccaddedcb # [perl 128420] recursive match with backreferences
|
|
|
2c161b |
+(?il)\x{100}|\x{100}|\x{FF} \xFF y $& \xFF
|
|
|
2c161b |
|
|
|
2c161b |
# Keep these lines at the end of the file
|
|
|
2c161b |
# vim: softtabstop=0 noexpandtab
|
|
|
2c161b |
--
|
|
|
2c161b |
2.14.3
|
|
|
2c161b |
|