[PATCH] regex with \p{IsAlpha} doesn't match properly #9720
Labels
Comments
From @ntyniAs reported by Srdjan <srdjan@catalyst.net.nz> in "abcd" =~ /x*\p{IsAlpha}/ while this supposedly equivalent match succeeds: "abcd" =~ /x*[[:alpha:]]/ Reported at 5.8.2, verified with 5.8.8, 5.10.0 and current blead. Output with -Dr: Compiling REx "x*\p{IsAlpha}" It seems that the problem has to do with losing the UNICODE_ALL flag Perl Info |
From @ntyni0001-Add-TODO-test-for-x-p-IsAlpha-match-failure.patchFrom bf52a1792a253b8bfd58b82d5b56581b2d6be397 Mon Sep 17 00:00:00 2001
From: Niko Tyni <ntyni@debian.org>
Date: Thu, 23 Apr 2009 00:00:27 +0300
Subject: [PATCH] Add TODO test for /x*\p{IsAlpha}/ match failure
As reported by Srdjan <srdjan@catalyst.net.nz> in
<http://bugs.debian.org/230144>, this match currently
fails:
"abcd" =~ /x*\p{IsAlpha}/
while this supposedly equivalent match succeeds:
"abcd" =~ /x*[[:alpha:]]/
---
t/uni/class.t | 9 ++++++++-
1 files changed, 8 insertions(+), 1 deletions(-)
diff --git a/t/uni/class.t b/t/uni/class.t
index 4620ca0..b755aae 100644
--- a/t/uni/class.t
+++ b/t/uni/class.t
@@ -4,7 +4,7 @@ BEGIN {
require "test.pl";
}
-plan tests => 5092;
+plan tests => 5094;
sub MyUniClass {
<<END;
@@ -71,6 +71,13 @@ is(($str =~ /(\p{Other::Class}+)/)[0], '@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_');
# make sure it finds class in other OTHER package
is(($str =~ /(\p{A::B::Intersection}+)/)[0], '@ABCDEFGHIJKLMNO');
+# see http://bugs.debian.org/230144
+{
+is(($str =~ /x*([[:alpha:]])/)[0], 'A');
+local $::TODO = "currently `contradicts stdclass'";
+is(($str =~ /x*(\p{IsAlpha})/)[0], 'A');
+}
+
# all of these should look in lib/unicore/bc/AL.pl
$str = "\x{070D}\x{070E}\x{070F}\x{0710}\x{0711}";
is(($str =~ /(\P{BidiClass: ArabicLetter}+)/)[0], "\x{070E}\x{070F}");
--
1.5.6.5
|
From @ntyni0002-Fix-x-p-IsAlpha-match-failure.patchFrom 33c4880b1d1e03d0b33adf341c6f93f9387f51ba Mon Sep 17 00:00:00 2001
From: Niko Tyni <ntyni@debian.org>
Date: Thu, 23 Apr 2009 00:03:52 +0300
Subject: [PATCH] Fix /x*\p{IsAlpha}/ match failure
regexec_flags() says 'Contradicts stclass...' because
the ANYOF_UNICODE flag is lost while cl_and()ing
the classes.
---
regcomp.c | 2 +-
t/uni/class.t | 3 ---
2 files changed, 1 insertions(+), 4 deletions(-)
diff --git a/regcomp.c b/regcomp.c
index e061528..150e7d0 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -3665,7 +3665,7 @@ S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp,
}
if (flags & SCF_DO_STCLASS_OR)
- cl_and(data->start_class, and_withp);
+ cl_or(pRExC_state, data->start_class, and_withp);
flags &= ~SCF_DO_STCLASS;
}
}
diff --git a/t/uni/class.t b/t/uni/class.t
index b755aae..3acaf76 100644
--- a/t/uni/class.t
+++ b/t/uni/class.t
@@ -72,11 +72,8 @@ is(($str =~ /(\p{Other::Class}+)/)[0], '@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_');
is(($str =~ /(\p{A::B::Intersection}+)/)[0], '@ABCDEFGHIJKLMNO');
# see http://bugs.debian.org/230144
-{
is(($str =~ /x*([[:alpha:]])/)[0], 'A');
-local $::TODO = "currently `contradicts stdclass'";
is(($str =~ /x*(\p{IsAlpha})/)[0], 'A');
-}
# all of these should look in lib/unicore/bc/AL.pl
$str = "\x{070D}\x{070E}\x{070F}\x{0710}\x{0711}";
--
1.5.6.5
|
|
jarich@perltraining.com.au - Status changed from 'new' to 'open' |
From @khwilliamsonThis was fixed in 5.13.11 |
|
@khwilliamson - Status changed from 'open' to 'resolved' |
From @khwilliamsonThis was fixed in 5.13.11 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Migrated from rt.perl.org#65054 (status was 'resolved')
Searchable as RT65054$
The text was updated successfully, but these errors were encountered: