RESOLVED FIXED 73627
[Mac] Reference count threading violation in FormDataStreamMac.mm 
https://bugs.webkit.org/show_bug.cgi?id=73627
Summary [Mac] Reference count threading violation in FormDataStreamMac.mm
Darin Adler
Reported 2011-12-01 23:01:23 PST
[Mac] Reference count threading violation in FormDataStreamMac.mm
Attachments
Patch (7.20 KB, patch)
2011-12-01 23:08 PST, Darin Adler 		no flags
DetailsFormatted DiffDiff
Patch (7.19 KB, patch)
2011-12-01 23:12 PST, Darin Adler 		sam: review+
DetailsFormatted DiffDiff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Darin Adler
Comment 1 2011-12-01 23:08:05 PST
Created attachment 117575 [details] Patch
mitz
Comment 2 2011-12-01 23:11:27 PST
Comment on attachment 117575 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=117575&action=review > Source/WebCore/platform/network/mac/FormDataStreamMac.mm:62 > + DEFINE_STATIC_LOCAL(NSMapTable *, streamFieldsMap, (NSCreateMapTable(NSNonRetainedObjectMapKeyCallBacks, NSNonOwnedPointerMapValueCallBacks, 1))); I don’t think DEFINE_STATIC_LOCAL makes sense for a pointer type. Those don’t have destructors.
Darin Adler
Comment 3 2011-12-01 23:11:50 PST
Comment on attachment 117575 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=117575&action=review >> Source/WebCore/platform/network/mac/FormDataStreamMac.mm:62 >> + DEFINE_STATIC_LOCAL(NSMapTable *, streamFieldsMap, (NSCreateMapTable(NSNonRetainedObjectMapKeyCallBacks, NSNonOwnedPointerMapValueCallBacks, 1))); > > I don’t think DEFINE_STATIC_LOCAL makes sense for a pointer type. Those don’t have destructors. You're right. I should just use a plain old "static" here.
Darin Adler
Comment 4 2011-12-01 23:12:44 PST
Created attachment 117576 [details] Patch
Darin Adler
Comment 5 2011-12-02 09:43:00 PST
Committed r101813: <http://trac.webkit.org/changeset/101813>
Note You need to log in before you can comment on or make changes to this bug.