RESOLVED FIXED 72631
[Chromium] fast/filesystem/file-writer-abort-depth.html occasionally timeouts
https://bugs.webkit.org/show_bug.cgi?id=72631
Summary [Chromium] fast/filesystem/file-writer-abort-depth.html occasionally timeouts
Attachments
Steve Block
Comment 1 2011-11-17 10:30:08 PST
Steve Block
Comment 2 2011-11-18 08:02:15 PST
Also seeing a crash in fast/filesystem/file-writer-abort.html on all Chromium Linux bots, eg http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Linux%20(dbg)/builds/440 [31095:31372:7140143115020:FATAL:file_writer_delegate.cc(225)] Check failed: total_bytes_written_ <= allowed_bytes_to_write_ || allowed_bytes_to_write_ < 0. Backtrace: base::debug::StackTrace::StackTrace() [0x712bf2] logging::LogMessage::~LogMessage() [0x6a114a] fileapi::FileWriterDelegate::Write() [0x24bfd65] fileapi::FileWriterDelegate::OnDataWritten() [0x24c0078] base::internal::Invoker1<>::DoInvoke() [0x24c1417] base::Callback<>::Run() [0xc4454e] net::FileStream::AsyncContext::RunAsynchronousCallback() [0x1fb25d6] base::internal::Invoker1<>::DoInvoke() [0x1fb4452] base::Callback<>::Run() [0x68b197] net::CancelableCallback::Run() [0x1fb3922] base::internal::Invoker1<>::DoInvoke() [0x1fb45b2] base::Callback<>::Run() [0x68b197] MessageLoop::RunTask() [0x6a64a3] MessageLoop::DeferOrRunPendingTask() [0x6a65bd] MessageLoop::DoWork() [0x6a6ddf] base::MessagePumpLibevent::Run() [0x70790d] MessageLoop::RunInternal() [0x6a6125] MessageLoop::RunHandler() [0x6a5fd8] MessageLoop::Run() [0x6a58eb] base::Thread::Run() [0x1cfa3e6] base::Thread::ThreadMain() [0x1cfa566] base::(anonymous namespace)::ThreadFunc() [0x6ee8a3] start_thread [0x7f3555b439ca] 0x7f355379170d
Adam Klein
Comment 3 2011-11-22 11:01:05 PST
Widening expectation to Mac DEBUG.
Vsevolod Vlasov
Comment 4 2011-12-05 04:26:58 PST
fast/filesystem/file-writer-abort-depth.html also crashes on win release, win7 release. Updated expectations: Committed r101989: <http://trac.webkit.org/changeset/101989>
John Knottenbelt
Comment 5 2012-01-09 03:39:57 PST
Adam Klein
Comment 6 2012-03-02 15:51:53 PST
More widening, file-writer-abort-depth.html also crashes on Mac debug. Kinuko, have you had a chance to dig into this at all?
Adam Klein
Comment 7 2012-03-02 15:53:37 PST
Here's one of the asserts from a Snow Leopard debug run: 2012-02-24 17:27:08,272 1062 single_test_runner.py:202 DEBUG worker/6 Stacktrace for fast/filesystem/file-writer-abort-depth.html: [1074:-1340764160:883054682632:ERROR:file_stream_metrics.cc(37)] RecordFileError() error = 9 source = 1 record = 0 Assertion failed: (ptr_ != NULL), function operator->, file ../base/memory/ref_counted.h, line 249. No stack, unfortunately.
Kinuko Yasuda
Comment 8 2012-03-02 16:59:49 PST
(In reply to comment #7) > Here's one of the asserts from a Snow Leopard debug run: > > 2012-02-24 17:27:08,272 1062 single_test_runner.py:202 DEBUG worker/6 Stacktrace for fast/filesystem/file-writer-abort-depth.html: > [1074:-1340764160:883054682632:ERROR:file_stream_metrics.cc(37)] RecordFileError() error = 9 source = 1 record = 0 > Assertion failed: (ptr_ != NULL), function operator->, file ../base/memory/ref_counted.h, line 249. > > No stack, unfortunately. Sorry these days I haven't taken a closer look at WebKit bugs-- just filed the corresponding chromium bug since the crash is likely coming from the chrome code. http://code.google.com/p/chromium/issues/detail?id=116639 I will take a look at the quota related DCHECK error. I hope Eric has a better idea FileWriter.abort related errors-- or we might have introduced a new error in the recent callback refactoring. Anyway I'll take a look.
Kinuko Yasuda
Comment 10 2012-05-17 22:39:16 PDT
(In reply to comment #9) > After landing http://crrev.com/136513 it looks fine now. > > http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Ffilesystem%2Ffile-writer-abort Now it looks pretty stable in terms of CRASH, but it seems to be occasionally seeing timeouts. I removed CRASH expectations from test_expectations.txt. As for timeouts I'll keep an eye on it.
Kinuko Yasuda
Comment 11 2012-06-07 06:00:56 PDT
Now all tests look green. Closing!
Note You need to log in before you can comment on or make changes to this bug.