WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED WONTFIX
101381
On ML, Chromium DRTs are killed very frequently while running tests in fast/forms
https://bugs.webkit.org/show_bug.cgi?id=101381
Summary
On ML, Chromium DRTs are killed very frequently while running tests in fast/f...
Ryosuke Niwa
Reported
2012-11-06 11:38:54 PST
I’m seeing lots of DRT instances being killed on Mountain Lion: [170/942] fast/forms/week/week-appearance-pseudo-classes.html failed unexpectedly (image diff) stopping DumpRenderTree(pid 85580) timed out, killing it [187/942] fast/forms/week/week-appearance-pseudo-elements.html failed unexpectedly (image diff) [202/942] fast/forms/week/week-appearance-style.html failed unexpectedly (image diff) stopping DumpRenderTree(pid 85612) timed out, killing it stopping DumpRenderTree(pid 85578) timed out, killing it [220/942] platform/chromium/fast/forms/suggestion-picker/date-suggestion-picker-appearance.html failed unexpectedly (image diff) stopping DumpRenderTree(pid 85592) timed out, killing it stopping DumpRenderTree(pid 85588) timed out, killing it [294/942] fast/forms/range/range-thumb-height-percentage.html passed unexpectedly stopping DumpRenderTree(pid 85586) timed out, killing it [356/942] fast/forms/file/input-file-re-render.html failed unexpectedly (image diff) stopping DumpRenderTree(pid 85615) timed out, killing it stopping DumpRenderTree(pid 85606) timed out, killing it stopping DumpRenderTree(pid 85601) timed out, killing it [380/942] platform/chromium/fast/forms/suggestion-picker/time-suggestion-picker-appearance-rtl.html failed unexpectedly (image diff) [384/942] platform/chromium/fast/forms/suggestion-picker/time-suggestion-picker-appearance-with-scroll-bar.html failed unexpectedly (image diff) [385/942] platform/chromium/fast/forms/suggestion-picker/time-suggestion-picker-appearance.html failed unexpectedly (image diff) [407/942] fast/forms/basic-buttons.html failed unexpectedly (image diff) [413/942] fast/forms/blankbuttons.html failed unexpectedly (image diff) stopping DumpRenderTree(pid 85568) timed out, killing it
Attachments
Add attachment
proposed patch, testcase, etc.
Ryosuke Niwa
Comment 1
2012-11-06 11:39:45 PST
I’ll note that this is ninja build since build-webkit --chromium no longer works.
Dirk Pranke
Comment 2
2012-11-06 13:58:52 PST
I haven't been able to reproduce this using a tip-of-tree/canary chromium build on my mac pro (12 cpu, 24gb RAM) running either just fast/forms or fast/dom. Each folder has produced a couple of "killing it" messages, but nowhere near what rniwa's log showed. I will try a webkit checkout next.
Dirk Pranke
Comment 3
2012-11-06 15:52:41 PST
So, synced to the same pair of revisions in chromium and webkit (webkit
r133678
, chromium
r166297
), I continue to see more timeouts in the webkit checkout than the chromium checkout. This makes me suspect that there's something different about how we're building DRT in the webkit checkout. Adding JamesR and Thakis to see if they have any ideas ...
Dirk Pranke
Comment 4
2012-11-06 15:55:40 PST
Note that I don't have an ~/.gyp/include.gypi, so I'm getting the default builds for both checkouts, which I believe are static builds, if that matters.
James Robinson
Comment 5
2012-11-06 16:03:43 PST
(In reply to
comment #4
)
> Note that I don't have an ~/.gyp/include.gypi, so I'm getting the default builds for both checkouts, which I believe are static builds, if that matters.
diff the binaries? Is it DumpRenderTree that's different or the wrapper scripts driving it (maybe try copying the binary from one output dir to the other to test)?
Dirk Pranke
Comment 6
2012-11-06 16:17:59 PST
The binaries are different. I would assume the wrapper scripts are the same (but haven't actually checked this). Running the webkit binary in the chromium checkout produces the same timeouts; running the chromium binary in the webkit checkout does not timeout. This makes me think that it's something about how things are being built.
Ryosuke Niwa
Comment 7
2012-11-07 11:45:01 PST
This is seriously making it impossible to test V8 bindings. Tests run insanely slowly :( Just running 685 tests is taking well over 5 minutes on my MacPro...
Eric Seidel (no email)
Comment 8
2012-11-14 01:16:59 PST
Is this your command? new-run-webkit-tests --release --chromium fast/forms
Ryosuke Niwa
Comment 9
2012-11-14 02:45:53 PST
(In reply to
comment #8
)
> Is this your command? > new-run-webkit-tests --release --chromium fast/forms
--debug.
Dirk Pranke
Comment 10
2012-11-14 11:19:43 PST
This was originally --release, right? At any rate, I have no reason to think this has something to do w/ NRWT, so clearing that part from the bug subject.
Ryosuke Niwa
Comment 11
2012-11-14 11:31:52 PST
(In reply to
comment #10
)
> This was originally --release, right? > > At any rate, I have no reason to think this has something to do w/ NRWT, so clearing that part from the bug subject.
No. I never use release builds.
Eric Seidel (no email)
Comment 12
2012-11-14 11:37:55 PST
(In reply to
comment #11
)
> No. I never use release builds.
Where is memegen when you need it.
Dirk Pranke
Comment 13
2012-11-14 11:57:50 PST
(In reply to
comment #11
)
> (In reply to
comment #10
) > > This was originally --release, right? > > > > At any rate, I have no reason to think this has something to do w/ NRWT, so clearing that part from the bug subject. > > No. I never use release builds.
Oh, that's right, I remember now, that you were using debug but I reproduced it in both release and debug.
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug