RESOLVED FIXED 21525
55 StructureID leaks on Wikitravel's main page 
https://bugs.webkit.org/show_bug.cgi?id=21525
Summary 55 StructureID leaks on Wikitravel's main page
Cameron Zwarich (cpst)
Reported 2008-10-10 11:15:24 PDT
After loading and closing http://wikitravel.org/en/Main_Page I see 60 StructureIDs leaked, 5 of which are the ones allocated by JSGlobalData. I will post allocation backtraces for them.
Attachments
Stack traces of leaking allocations (389.75 KB, text/plain)
2008-10-10 11:18 PDT, Cameron Zwarich (cpst) 		no flags
Details
Somewhat evil fix (3.82 KB, patch)
2008-10-11 01:18 PDT, Cameron Zwarich (cpst) 		no flags
DetailsFormatted DiffDiff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Cameron Zwarich (cpst)
Comment 1 2008-10-10 11:18:10 PDT
Created attachment 24268 [details] Stack traces of leaking allocations Lots of the things leaking are things that are likely to be extended by a JS framework, like StructureIDs created in ObjectPrototype and static function slots in ArrayPrototype. There are also a lot of generic puts on arrays that create leaking StructureIDs.
Cameron Zwarich (cpst)
Comment 2 2008-10-10 15:36:48 PDT
One particular leak has been spun off as bug 21533.
Cameron Zwarich (cpst)
Comment 3 2008-10-11 01:18:32 PDT
Created attachment 24291 [details] Somewhat evil fix This fixes the bug, as well as all other StructureID leaks I could find. It is a bit evil, because it is a default argument on a virtual method, and subclasses won't pass it back down to the base class implementation in JSObject. I will probably add an extra default bool argument to make sure people deal with it.
Cameron Zwarich (cpst)
Comment 4 2008-10-11 13:11:40 PDT
Comment on attachment 24291 [details] Somewhat evil fix Oops, that fix was meant for bug 21533, but I have a new one to post there.
Cameron Zwarich (cpst)
Comment 5 2008-10-11 13:59:44 PDT
This was fixed by r37508, the fix for bug 21533.
Note You need to log in before you can comment on or make changes to this bug.