Came in on a TM -> JM merge, needs to go out before we can merge TM -> m-c.
12926 INFO TEST-START | /tests/dom/tests/mochitest/dom-level2-core/test_setAttributeNS01.html
--DOCSHELL 184F0128 == 7
--DOMWINDOW == 842 (1827BCD0) [serial = 5862] [outer = 00000000] [url = http://mochi.test:8888/tests/dom/tests/mochitest/dom-level2-core/files/staffNS.xml]
--DOMWINDOW == 841 (185B5A80) [serial = 5859] [outer = 00000000] [url = http://mochi.test:8888/tests/dom/tests/mochitest/dom-level2-core/files/staffNS.xml]
--DOMWINDOW == 840 (0F2B0900) [serial = 5856] [outer = 00000000] [url = http://mochi.test:8888/tests/dom/tests/mochitest/dom-level2-core/files/staffNS.xml]
WARNING: NS_ENSURE_TRUE(mState == STATE_TRANSFERRING) failed: file e:/builds/moz2_slave/tm-w32-dbg/build/netwerk/base/src/nsSocketTransport2.cpp, line 1889
before 856064, after 856064, break 00000000
++DOMWINDOW == 841 (17BCD8C8) [serial = 5867] [outer = 07FFC5A0]
pldhash: for the table at address 17C68850, the given entrySize of 48 probably favors chaining over double hashing.
++DOCSHELL 17C687E8 == 8
++DOMWINDOW == 842 (1827BCD0) [serial = 5868] [outer = 00000000]
WARNING: Subdocument container has no frame: file e:/builds/moz2_slave/tm-w32-dbg/build/layout/base/nsDocumentViewer.cpp, line 2438
++DOMWINDOW == 843 (0ED2F808) [serial = 5869] [outer = 1827BC80]
--DOCSHELL 186C2010 == 7
command timed out: 5400 seconds elapsed, killing pid 3012
WARNING: NS_ENSURE_SUCCESS(rv, 1) failed with result 0x80520012: file e:/builds/moz2_slave/tm-w32-dbg/build/parser/htmlparser/src/nsExpatDriver.cpp, line 711
WARNING: NS_ENSURE_SUCCESS(rv, rv) failed with result 0x80530005: file e:/builds/moz2_slave/tm-w32-dbg/build/content/base/src/nsContentUtils.cpp, line 2100
WARNING: NS_ENSURE_SUCCESS(rv, rv) failed with result 0x80530005: file e:/builds/moz2_slave/tm-w32-dbg/build/content/base/src/nsGenericElement.cpp, line 2574
12927 INFO TEST-PASS | throw_INVALID_CHARACTER_ERR
program finished with exit code 1
This only appeared in two of the last four pushes to TM, and seems to stop on a different test pretty much every time. I'm guessing this is not getting stuck, just hitting an absolute timeout on how long the mochitest can take.
Seems like a reasonable bet to me: "Windows debug" is our synonym for "slow" so if, as I finally heard, you've got a perf impact, timing out on Windows debug is not-too-surprising fallout from that.
This hasn't happened on any platform in the JM branch since the patch for bug 670185 went in (which was, in large part, cutting out work done during GC). How should this get marked as resolved?
Fixed by bug 670185