Memory leak with named regexp captures #10702
Comments
From r.khamidullin@transcredit.ruPlease, I have a memory leak. Sample with memory leak Sample without memory leak -- Characteristics of this binary (from libperl): -- Р���ам Хамид�ллин |
From @iabynOn Wed, Aug 18, 2010 at 04:36:10AM -0700, r.khamidullin wrote:
Can you please supply a short, complete program that demonstrates the -- |
|
The RT System itself - Status changed from 'new' to 'open' |
From @SBECK-githubCreated by @SBECK-githubThere is a memory leak in both perl 5.10 and 5.12 that occurs when named The following script illustrates the bug: ============================= $a1 = '(?<a>\d)'; $a2 = $a1; $b = '\d'; #$string = "12"; while (1) {
|
|
@iabyn - Status changed from 'open' to 'stalled' |
From @SBECK-githubJust FYI, I just checked this against perl 5.14.1 and 5.15.0 and the |
|
@SBECK-github - Status changed from 'new' to 'open' |
From @pluI think I'm experiencing the same bug. I've narrowed it down to a simple |
From [Unknown Contact. See original ticket]I think I'm experiencing the same bug. I've narrowed it down to a simple |
From @pluChanging status from stalled to open, hope that's ok. If not, just |
|
@cpansprout - Status changed from 'stalled' to 'open' |
From data@cpan.orgYou may want to merge #77304 into this one, since they are the same. Or On Thu Oct 07 04:40:37 2010, sbeck wrote:
PATH=/home/sulbeck/local/bin:/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin:/usr/X11R6/bin:/bin:/sbin:/usr/games:/usr/lib/mit/bin:/usr/lib/mit/sbin:/usr/lib/qt3/bin:/etc:.
multi:/home/sulbeck/local/lib/perl5/site_perl/5.12.1:/home/sulbeck/local/lib/perl5/site_perl:/home/sulbeck/local/lib/perl5/5.12.1/x86_64-
multi:/home/sulbeck/local/lib/perl5/5.12.1:/usr/lib/perl5/site_perl/5.12.1/x86_64-
multi:/usr/lib/perl5/site_perl/5.12.1:/usr/lib/perl5/vendor_perl/5.12.1/x86_64-
|
From @pluAttached a fix. Please review that patch. I have to admit to I have no experience at all |
From @plu0001-Fix-78266-Memory-leak-with-named-regexp-captures.patchFrom f753a93ee39594777931d507a8acbcab61183436 Mon Sep 17 00:00:00 2001
From: Johannes Plunien <plu@pqpq.de>
Date: Tue, 13 Dec 2011 11:39:52 +0400
Subject: [PATCH] Fix #78266: Memory leak with named regexp captures.
---
regcomp.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/regcomp.c b/regcomp.c
index b1b8e6f..9e9fac4 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -5418,6 +5418,10 @@ Perl_reg_named_buff_fetch(pTHX_ REGEXP * const r, SV * const namesv,
return newRV_noinc(MUTABLE_SV(retarray));
}
}
+
+ if (ret)
+ SvREFCNT_dec(ret);
+
return NULL;
}
--
1.7.7
|
From @pluAdded a testcase. |
From @plu0002-Add-testcase-for-78266.patchFrom 7d77e85aabc4cf50f733820cb2b793ffc240840c Mon Sep 17 00:00:00 2001
From: Johannes Plunien <plu@pqpq.de>
Date: Tue, 13 Dec 2011 12:29:47 +0400
Subject: [PATCH 2/2] Add testcase for #78266.
---
t/op/svleak.t | 19 ++++++++++++++++++-
1 files changed, 18 insertions(+), 1 deletions(-)
diff --git a/t/op/svleak.t b/t/op/svleak.t
index 5eb090c..4eee3c8 100644
--- a/t/op/svleak.t
+++ b/t/op/svleak.t
@@ -13,7 +13,7 @@ BEGIN {
or skip_all("XS::APItest not available");
}
-plan tests => 19;
+plan tests => 20;
# run some code N times. If the number of SVs at the end of loop N is
# greater than (N-1)*delta at the end of loop 1, we've got a leak
@@ -141,3 +141,20 @@ leak(2, 0,
},
"rcatline leak"
);
+
+{
+ my $RE = qr/
+ (?:
+ <(?<tag>
+ \s*
+ [^>\s]+
+ )>
+ )??
+ /xis;
+
+ "<html><body></body></html>" =~ m/$RE/gcs;
+
+ leak(5, 0, sub {
+ my $tag = $+{tag};
+ }, "named regexp captures");
+}
--
1.7.7
|
From @cpansproutOn Tue Dec 13 00:09:08 2011, plu wrote:
This patch looks correct to me, so I have applied it as b28f4af. -- Father Chrysostomos |
|
@cpansprout - Status changed from 'open' to 'resolved' |
From @cpansproutOn Tue Dec 13 00:31:24 2011, plu wrote:
And this one as a17b855. -- Father Chrysostomos |
From @avarOn Tue, Dec 13, 2011 at 18:32, Father Chrysostomos via RT
I had already pushed smoke-me/avar-plu-RT-78266 earlier today as In this case we're leaking memory because we're allocating something, I've pushed a better fix to blead in |
From @demerphqOn 13 December 2011 19:11, �var Arnfjörð Bjarmason <avarab@gmail.com> wrote:
Concur that this is a better fix. This one was my fault. Sorry people. Yves -- |
From r.khamidullin@transcredit.ruJohannes Plunien, Thank You very much. I can not check that fix at the moment. 13.12.2011 12:09, Johannes Plunien via RT пи�е�:
-- Р���ам Хамид�ллин |
From r.khamidullin@transcredit.ruJohannes Plunien, full job. :) Regards! 13.12.2011 12:31, Johannes Plunien via RT пи�е�:
-- Р���ам Хамид�ллин |
Migrated from rt.perl.org#78266 (status was 'resolved')
Searchable as RT78266$
The text was updated successfully, but these errors were encountered: