RESOLVED FIXED 56511
Invalidate all NPObjects for a plug-in when that plug-in is destroyed 
https://bugs.webkit.org/show_bug.cgi?id=56511
Summary Invalidate all NPObjects for a plug-in when that plug-in is destroyed
Anders Carlsson
Reported 2011-03-16 17:56:02 PDT
Invalidate all NPObjects for a plug-in when that plug-in is destroyed
Attachments
Patch (12.10 KB, patch)
2011-03-16 18:01 PDT, Anders Carlsson 		no flags
DetailsFormatted DiffDiff
Patch (12.08 KB, patch)
2011-03-16 18:11 PDT, Anders Carlsson 		no flags
DetailsFormatted DiffDiff
Patch (14.75 KB, patch)
2011-03-16 18:35 PDT, Anders Carlsson 		no flags
DetailsFormatted DiffDiff
Patch (14.75 KB, patch)
2011-03-16 19:00 PDT, Anders Carlsson 		darin: review+
DetailsFormatted DiffDiff
Show Obsolete (3) View All Add attachment proposed patch, testcase, etc.
Anders Carlsson
Comment 1 2011-03-16 18:01:35 PDT
Created attachment 86013 [details] Patch
Anders Carlsson
Comment 2 2011-03-16 18:11:19 PDT
Created attachment 86015 [details] Patch
Anders Carlsson
Comment 3 2011-03-16 18:27:25 PDT
Comment on attachment 86015 [details] Patch This patch is incorrect.
Anders Carlsson
Comment 4 2011-03-16 18:35:47 PDT
Created attachment 86016 [details] Patch
Anders Carlsson
Comment 5 2011-03-16 19:00:24 PDT
Created attachment 86019 [details] Patch
Darin Adler
Comment 6 2011-03-16 19:49:20 PDT
Comment on attachment 86019 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=86019&action=review > Source/WebKit2/PluginProcess/PluginControllerProxy.cpp:86 > + // Get the plug-in so we can pass it to removePluginControllerProxy. The pointer is only > + // used as an identifier so it's OK to just get a weak reference. Aliasing can happen even when a pointer is just used as an identifier. In these cases I guess it’s OK because there’s no chance a new plug-in will be allocated before the remove call is made.
Anders Carlsson
Comment 7 2011-03-17 11:28:57 PDT
Committed r81370: <http://trac.webkit.org/changeset/81370>
WebKit Review Bot
Comment 8 2011-03-17 12:42:19 PDT
http://trac.webkit.org/changeset/81370 might have broken Leopard Intel Release (Tests)
Note You need to log in before you can comment on or make changes to this bug.