|
|
243a19 |
From d58a6811d830c2f2f850a03a18129c38cb732791 Mon Sep 17 00:00:00 2001
|
|
|
243a19 |
From: Yves Orton <demerphq@gmail.com>
|
|
|
243a19 |
Date: Tue, 13 Feb 2018 16:11:55 +1100
|
|
|
243a19 |
Subject: 5.26.1: fix TRIE_READ_CHAR and DECL_TRIE_TYPE to account for non-utf8
|
|
|
243a19 |
target
|
|
|
243a19 |
|
|
|
243a19 |
---
|
|
|
243a19 |
regexec.c | 14 ++++++++++----
|
|
|
243a19 |
t/re/re_tests | 1 +
|
|
|
243a19 |
2 files changed, 11 insertions(+), 4 deletions(-)
|
|
|
243a19 |
|
|
|
243a19 |
diff --git a/regexec.c b/regexec.c
|
|
|
243a19 |
index fa888823bd..cf81b07e30 100644
|
|
|
243a19 |
--- a/regexec.c
|
|
|
243a19 |
+++ b/regexec.c
|
|
|
243a19 |
@@ -1472,7 +1472,7 @@ Perl_re_intuit_start(pTHX_
|
|
|
243a19 |
#define DECL_TRIE_TYPE(scan) \
|
|
|
243a19 |
const enum { trie_plain, trie_utf8, trie_utf8_fold, trie_latin_utf8_fold, \
|
|
|
243a19 |
trie_utf8_exactfa_fold, trie_latin_utf8_exactfa_fold, \
|
|
|
243a19 |
- trie_utf8l, trie_flu8 } \
|
|
|
243a19 |
+ trie_utf8l, trie_flu8, trie_flu8_latin } \
|
|
|
243a19 |
trie_type = ((scan->flags == EXACT) \
|
|
|
243a19 |
? (utf8_target ? trie_utf8 : trie_plain) \
|
|
|
243a19 |
: (scan->flags == EXACTL) \
|
|
|
243a19 |
@@ -1482,10 +1482,12 @@ Perl_re_intuit_start(pTHX_
|
|
|
243a19 |
? trie_utf8_exactfa_fold \
|
|
|
243a19 |
: trie_latin_utf8_exactfa_fold) \
|
|
|
243a19 |
: (scan->flags == EXACTFLU8 \
|
|
|
243a19 |
- ? trie_flu8 \
|
|
|
243a19 |
+ ? (utf8_target \
|
|
|
243a19 |
+ ? trie_flu8 \
|
|
|
243a19 |
+ : trie_flu8_latin) \
|
|
|
243a19 |
: (utf8_target \
|
|
|
243a19 |
? trie_utf8_fold \
|
|
|
243a19 |
- : trie_latin_utf8_fold)))
|
|
|
243a19 |
+ : trie_latin_utf8_fold)))
|
|
|
243a19 |
|
|
|
243a19 |
/* 'uscan' is set to foldbuf, and incremented, so below the end of uscan is
|
|
|
243a19 |
* 'foldbuf+sizeof(foldbuf)' */
|
|
|
243a19 |
@@ -1496,7 +1498,7 @@ STMT_START {
|
|
|
243a19 |
switch (trie_type) { \
|
|
|
243a19 |
case trie_flu8: \
|
|
|
243a19 |
_CHECK_AND_WARN_PROBLEMATIC_LOCALE; \
|
|
|
243a19 |
- if (utf8_target && UTF8_IS_ABOVE_LATIN1(*uc)) { \
|
|
|
243a19 |
+ if (UTF8_IS_ABOVE_LATIN1(*uc)) { \
|
|
|
243a19 |
_CHECK_AND_OUTPUT_WIDE_LOCALE_UTF8_MSG(uc, uc_end - uc); \
|
|
|
243a19 |
} \
|
|
|
243a19 |
goto do_trie_utf8_fold; \
|
|
|
243a19 |
@@ -1519,10 +1521,14 @@ STMT_START {
|
|
|
243a19 |
uscan = foldbuf + skiplen; \
|
|
|
243a19 |
} \
|
|
|
243a19 |
break; \
|
|
|
243a19 |
+ case trie_flu8_latin: \
|
|
|
243a19 |
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE; \
|
|
|
243a19 |
+ goto do_trie_latin_utf8_fold; \
|
|
|
243a19 |
case trie_latin_utf8_exactfa_fold: \
|
|
|
243a19 |
flags |= FOLD_FLAGS_NOMIX_ASCII; \
|
|
|
243a19 |
/* FALLTHROUGH */ \
|
|
|
243a19 |
case trie_latin_utf8_fold: \
|
|
|
243a19 |
+ do_trie_latin_utf8_fold: \
|
|
|
243a19 |
if ( foldlen>0 ) { \
|
|
|
243a19 |
uvc = utf8n_to_uvchr( (const U8*) uscan, foldlen, &len, uniflags ); \
|
|
|
243a19 |
foldlen -= len; \
|
|
|
243a19 |
diff --git a/t/re/re_tests b/t/re/re_tests
|
|
|
243a19 |
index 410fceadac..78baed6ffc 100644
|
|
|
243a19 |
--- a/t/re/re_tests
|
|
|
243a19 |
+++ b/t/re/re_tests
|
|
|
243a19 |
@@ -1985,6 +1985,7 @@ AB\s+\x{100} AB \x{100}X y - -
|
|
|
243a19 |
/(?x)[a b]/xx \N{SPACE} yS $& # Note a space char here
|
|
|
243a19 |
/(?xx)[a b]/x \N{SPACE} n - -
|
|
|
243a19 |
/(?-x:[a b])/xx \N{SPACE} yS $& # Note a space char here
|
|
|
243a19 |
+(?il)\x{100}|\x{100}|\x{FF} \xFF y $& \xFF
|
|
|
243a19 |
|
|
|
243a19 |
# Keep these lines at the end of the file
|
|
|
243a19 |
# vim: softtabstop=0 noexpandtab
|
|
|
243a19 |
--
|
|
|
243a19 |
2.11.0
|
|
|
243a19 |
|