Bug 158093

Example of existing use: http://trac.webkit.org/browser/trunk/LayoutTests/storage/indexeddb/resources/cursor-leak.js

Example of existing use: http://trac.webkit.org/browser/trunk/LayoutTests/storage/indexeddb/resources/cursor-leak.js

Created attachment 279842 [details] Patch

Comment on attachment 279842 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=279842&action=review r=me with a change suggestion > Source/WebCore/testing/GCObservation.h:50 > + class WeakOwner : public JSC::WeakHandleOwner { > + void finalize(JSC::Handle<JSC::Unknown>, void* context) final; > + }; > + WeakOwner m_weakOwner; I realized that you can get even simpler and delete this stuff. Weak<T> becomes falsy automatically after the value is GC'd, so you can delete the weak owner and wasCollected() can just return !m_observedValue. That's a lot nicer.

(In reply to comment #4) > Comment on attachment 279842 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=279842&action=review > > r=me with a change suggestion > > > Source/WebCore/testing/GCObservation.h:50 > > + class WeakOwner : public JSC::WeakHandleOwner { > > + void finalize(JSC::Handle<JSC::Unknown>, void* context) final; > > + }; > > + WeakOwner m_weakOwner; > > I realized that you can get even simpler and delete this stuff. Weak<T> > becomes falsy automatically after the value is GC'd, so you can delete the > weak owner and wasCollected() can just return !m_observedValue. That's a lot > nicer. Right - With the callback function approach we discussed in person, this approach was necessary. But for the bool check, it can be much simpler. As an added bonus, I'll make sure it builds! Thanks for your help on this.

Created attachment 279864 [details] Patch for EWS/landing

Same linker error the first patch was seeing... Weiiiiiiird.

Why did Mac Debug build fine (found the JSC symbol) but nobody else did (couldn't find it)? I'm definitely building debug locally... guess I'll kick off a release build.

Created attachment 279866 [details] Patch for EWS Attempt to fix the WebCoreTestSupport linker error with some JSC exports.

Okay, at this point I simply don't know why WebCoreTestSupport can't find these symbols while linking against JavaScriptCore, even though WebCore obviously does. I also don't know why the debug build works fine, but release does not. I'll need the help of somebody else more knowledgable about the build system, exports, libs, etc.

I was missing JSCJSValueInlines.h *slaps forehead*

Created attachment 279891 [details] Patch for EWS + landing

Comment on attachment 279891 [details] Patch for EWS + landing Attachment 279891 [details] did not pass mac-ews (mac): Output: http://webkit-queues.webkit.org/results/1386641 Number of test failures exceeded the failure limit.

Created attachment 279893 [details] Archive of layout-test-results from ews103 for mac-yosemite The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: ews103 Port: mac-yosemite Platform: Mac OS X 10.10.5

(In reply to comment #14) > Created attachment 279893 [details] > Archive of layout-test-results from ews103 for mac-yosemite > > The attached test failures were seen while running run-webkit-tests on the > mac-ews. > Bot: ews103 Port: mac-yosemite Platform: Mac OS X 10.10.5 These failures cannot be related to this patch.

(In reply to comment #15) > (In reply to comment #14) > > Created attachment 279893 [details] > > Archive of layout-test-results from ews103 for mac-yosemite > > > > The attached test failures were seen while running run-webkit-tests on the > > mac-ews. > > Bot: ews103 Port: mac-yosemite Platform: Mac OS X 10.10.5 > > These failures cannot be related to this patch. Seems much more likely that the W3C web platform test server encountered a problem during the test run.

Comment on attachment 279891 [details] Patch for EWS + landing Rejecting attachment 279891 [details] from commit-queue. Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.webkit.org', '--bot-id=webkit-cq-02', 'validate-changelog', '--check-oops', '--non-interactive', 279891, '--port=mac']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit Geoff Garen found in /Volumes/Data/EWS/WebKit/LayoutTests/ChangeLog does not appear to be a valid reviewer according to contributors.json. /Volumes/Data/EWS/WebKit/LayoutTests/ChangeLog neither lists a valid reviewer nor contains the string "Unreviewed" or "Rubber stamp" (case insensitive). Full output: http://webkit-queues.webkit.org/results/1386722

Comment on attachment 279891 [details] Patch for EWS + landing Rejecting attachment 279891 [details] from commit-queue. Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.webkit.org', '--bot-id=webkit-cq-03', 'validate-changelog', '--check-oops', '--non-interactive', 279891, '--port=mac']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit Geoff Garen found in /Volumes/Data/EWS/WebKit/LayoutTests/ChangeLog does not appear to be a valid reviewer according to contributors.json. /Volumes/Data/EWS/WebKit/LayoutTests/ChangeLog neither lists a valid reviewer nor contains the string "Unreviewed" or "Rubber stamp" (case insensitive). Full output: http://webkit-queues.webkit.org/results/1386760

(In reply to comment #18) > Geoff Garen found in /Volumes/Data/EWS/WebKit/LayoutTests/ChangeLog does not > appear to be a valid reviewer according to contributors.json. This kind of nonsense is why I hate the commit queue. :P

Landed manually in http://trac.webkit.org/changeset/201422

Usually the right fix for a linker bug like this is to #include JSCInlines.h. Debug builds don't see the failure because they don't do the inlining thing.