See http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=accessibility%2Floading-iframe-updates-axtree.html . The blame lists here are http://trac.webkit.org/log/?verbose=on&rev=124373&stop_rev=124371 and http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=/trunk/src&mode=html&range=149513:149497 . The former doesn't have anything obvious in it. The latter has http://src.chromium.org/viewvc/chrome?view=rev&revision=149510 , but I tried reverting that on the Chromium side for a few hours and the bot didn't stop crashing, so unless it's not rebuilding correctly, that's not the cause either. Dominic, I'm hoping that since you understand the accessibility stuff a bit you might be able to investigate this.
I don't see anything in the blamelists either. It made sense to try reverting the accessibility-related change I just landed, but I think it's unlikely to be related; in particular none of that code should be running in DumpRenderTree. I'll try debugging it directly.
(1) Crashing on Win dbg (1) as well (but not Win, dunno why). (2) Flaky -- sometimes this passes.
Can repro locally on Win 7, but it's not easy - it's flaky and usually passes for me. When it does fail, here's the stack trace I get: Backtrace: RtlEnterCriticalSection [0x777D22B2+18] WTF::Mutex::lock [0x004D2A4B+27] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\threadingwin.cpp:307) WTF::Locker<WTF::Mutex>::Locker<WTF::Mutex> [0x0042B480+32] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\locker.h:38) WTF::addIterator<void *,WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *>,WTF::KeyValuePairKeyExtractor<WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *> >,WTF::PtrHash<void *>,WTF::HashMapValueTraits<WTF::HashTraits<void *>,WTF::HashT [0x0048CB37+87] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\hashtable.h:1289) WTF::HashTableConstIterator<void *,WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *>,WTF::KeyValuePairKeyExtractor<WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *> >,WTF::PtrHash<void *>,WTF::HashMapValueTraits<WTF::HashTraits<void *> [0x0048DA8C+44] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\hashtable.h:132) WTF::HashTable<void *,WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *>,WTF::KeyValuePairKeyExtractor<WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *> >,WTF::PtrHash<void *>,WTF::HashMapValueTraits<WTF::HashTraits<void *>,WTF::HashTra [0x0048D5CF+63] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\hashtable.h:470) WTF::HashTable<void *,WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *>,WTF::KeyValuePairKeyExtractor<WTF::KeyValuePair<void *,CppBoundClass::PropertyCallback *> >,WTF::PtrHash<void *>,WTF::HashMapValueTraits<WTF::HashTraits<void *>,WTF::HashTra [0x0048CE69+57] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\hashtable.h:387) WTF::HashMap<void *,CppBoundClass::PropertyCallback *,WTF::PtrHash<void *>,WTF::HashTraits<void *>,WTF::HashTraits<CppBoundClass::PropertyCallback *> >::end [0x0048B95B+91] (d:\src\gitchrome3\src\third_party\webkit\source\wtf\wtf\hashmap.h:285) CppBoundClass::hasProperty [0x00489359+73] (d:\src\gitchrome3\src\third_party\webkit\tools\dumprendertree\chromium\testrunner\cppboundclass.cpp:223) CppNPObject::hasProperty [0x00488D40+32] (d:\src\gitchrome3\src\third_party\webkit\tools\dumprendertree\chromium\testrunner\cppboundclass.cpp:181)
The test is crashing on Linux as well: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=accessibility%2Floading-iframe-updates-axtree.html
Updated test expectations: http://trac.webkit.org/changeset/129200
Flakiness dashboard didn't record crash recently, but the result was changed to TIMEOUT by http://trac.webkit.org/changeset/142385