Bug 43230

Summary: REGRESSION: Memory leak within JSParser::JSParser
Product: WebKit Reporter: Mark Rowe (bdash) <mrowe>
Component: JavaScriptCoreAssignee: Alexey Proskuryakov <ap>
Status: RESOLVED FIXED    
Severity: Normal CC: ap, commit-queue, ggaren, oliver
Priority: P1 Keywords: InRadar, Regression
Version: 528+ (Nightly build)   
Hardware: PC   
OS: OS X 10.6   
URL: http://build.webkit.org/results/SnowLeopard%20Intel%20Leaks/r64324%20(9455)/DumpRenderTree7-leaks.txt
Attachments:
Description Flags
proposed fix none

Description Mark Rowe (bdash) 2010-07-29 18:50:44 PDT
The build bot shows memory leaks from within JSParser::JSParser at <http://build.webkit.org/results/SnowLeopard%20Intel%20Leaks/r64324%20(9455)/DumpRenderTree7-leaks.txt>.  I believe these showed up shortly after the new JavaScript parser was turned on.  Oliver said he was going to look at them then but I don’t think a bug was ever filed.
Comment 1 Mark Rowe (bdash) 2010-07-29 18:51:14 PDT
<rdar://problem/8254215>
Comment 2 Geoffrey Garen 2010-08-03 11:16:41 PDT
Is this a bug in ThreadSpecific?
Comment 3 Alexey Proskuryakov 2010-09-01 15:16:02 PDT
I think that this is a misuse of ThreadSpecific. It can't possibly destroy data for all threads in its own destructor.

template<typename T>
inline ThreadSpecific<T>::~ThreadSpecific()
{
    pthread_key_delete(m_key); // Does not invoke destructor functions.
}
Comment 4 Alexey Proskuryakov 2010-09-01 17:13:23 PDT
Created attachment 66301 [details]
proposed fix
Comment 5 WebKit Commit Bot 2010-09-02 05:58:24 PDT
Comment on attachment 66301 [details]
proposed fix

Clearing flags on attachment: 66301

Committed r66665: <http://trac.webkit.org/changeset/66665>
Comment 6 WebKit Commit Bot 2010-09-02 05:58:29 PDT
All reviewed patches have been landed.  Closing bug.