Bug 50652 – [chromium] tests failing on Chromium Win Release (Tests) bot

• Component: Tools / Tests
• Status: RESOLVED WONTFIX

[Tony Chang]

The following tests are failing on the Windows webkit.org test bot, but passing on the webkit canary after the DRT switch.

I wonder if the bot is configured differently?  Maybe something with cleartype?  Not sure about the filesystem failures.

Regressions: Unexpected text diff mismatch : (14)
  fast/filesystem/file-from-file-entry.html = TEXT
  fast/filesystem/filesystem-reference.html = TEXT
  fast/filesystem/flags-passing.html = TEXT
  fast/filesystem/op-copy.html = TEXT
  fast/filesystem/op-get-entry.html = TEXT
  fast/filesystem/op-get-parent.html = TEXT
  fast/filesystem/op-move.html = TEXT
  fast/filesystem/op-read-directory.html = TEXT
  fast/filesystem/op-remove.html = TEXT
  fast/filesystem/op-restricted-chars.html = TEXT
  fast/filesystem/op-restricted-names.html = TEXT
  fast/filesystem/simple-persistent.html = TEXT
  fast/filesystem/simple-temporary.html = TEXT
  http/tests/local/fileapi/send-sliced-dragged-file.html = TEXT

Regressions: Unexpected image mismatch : (19)
  css2.1/t1202-counter-09-b.html = IMAGE
  css2.1/t1202-counters-09-b.html = IMAGE
  fast/blockflow/japanese-lr-text.html = IMAGE
  fast/blockflow/japanese-rl-text.html = IMAGE
  fast/blockflow/japanese-ruby-horizontal-bt.html = IMAGE
  fast/blockflow/japanese-ruby-vertical-lr.html = IMAGE
  fast/blockflow/japanese-ruby-vertical-rl.html = IMAGE
  fast/blockflow/vertical-align-table-baseline.html = IMAGE
  fast/blockflow/vertical-baseline-alignment.html = IMAGE
  fast/css/line-height-determined-by-primary-font.html = IMAGE
  fast/forms/caret-rtl.html = IMAGE
  fast/repaint/japanese-rl-selection-clear.html = IMAGE
  fast/repaint/japanese-rl-selection-repaint.html = IMAGE
  fast/repaint/repaint-across-writing-mode-boundary.html = IMAGE
  fast/ruby/nested-ruby.html = IMAGE
  fast/text/atsui-negative-spacing-features.html = IMAGE
  fast/text/international/danda-space.html = IMAGE
  fast/text/international/hebrew-vowels.html = IMAGE
  svg/batik/text/textFeatures.svg = IMAGE

Regressions: Unexpected tests timed out : (3)
  fast/filesystem/async-operations.html = TIMEOUT
  fast/filesystem/file-writer-truncate-extend.html = TIMEOUT
  fast/filesystem/read-directory.html = TIMEOUT

Regressions: Unexpected image and text mismatch : (2)
  fast/forms/search-styled.html = IMAGE+TEXT
  fast/text/backslash-to-yen-sign-euc.html = IMAGE+TEXT

[James Robinson]

+cc Kinuko for the filesystem+DRT issues.

[Dimitri Glazkov (Google)]

The build.webkit.org bots are running Vista, btw.

[Tony Chang]

(In reply to comment #1)
> +cc Kinuko for the filesystem+DRT issues.

To be clear, these aren't DRT specific failures because they pass with DRT on the windows canary bot.

[Kinuko Yasuda]

(In reply to comment #3)
> (In reply to comment #1)
> > +cc Kinuko for the filesystem+DRT issues.
> 
> To be clear, these aren't DRT specific failures because they pass with DRT on the windows canary bot.

Not sure what's causing the diff.  Is there a way to get the test results from the webkit.org test bots?

[Tony Chang]

(In reply to comment #4)
> (In reply to comment #3)
> > (In reply to comment #1)
> > > +cc Kinuko for the filesystem+DRT issues.
> > 
> > To be clear, these aren't DRT specific failures because they pass with DRT on the windows canary bot.
> 
> Not sure what's causing the diff.  Is there a way to get the test results from the webkit.org test bots?

Yes! It's on the dashboard. E.g.,

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Ffilesystem%2Ffile-from-file-entry.html%20%20%20fast%2Ffilesystem%2Ffilesystem-reference.html%20%20%20fast%2Ffilesystem%2Fflags-passing.html&showExpectations=true&master=webkit.org

[Kinuko Yasuda]

(In reply to comment #5)
> (In reply to comment #4)
> > (In reply to comment #3)
> > > (In reply to comment #1)
> > > > +cc Kinuko for the filesystem+DRT issues.
> > > 
> > > To be clear, these aren't DRT specific failures because they pass with DRT on the windows canary bot.
> > 
> > Not sure what's causing the diff.  Is there a way to get the test results from the webkit.org test bots?
> 
> Yes! It's on the dashboard. E.g.,
> 
> http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Ffilesystem%2Ffile-from-file-entry.html%20%20%20fast%2Ffilesystem%2Ffilesystem-reference.html%20%20%20fast%2Ffilesystem%2Fflags-passing.html&showExpectations=true&master=webkit.org

Oh yeah right.  Great!

And every test seems to be failing with SECURITY_ERR when it requests a new filesystem.  It usually means it's failing to create a root directory in a unique temp directory...

[Tony Chang]

(In reply to comment #2)
> The build.webkit.org bots are running Vista, btw.

I can verify that the same tests in the 'fast' subdirectory fail for me on my (soon to be reclaimed) Vista box.  Maybe we should just re-image with whatever the canary bots are running?

[Tony Chang]

(In reply to comment #7)
> (In reply to comment #2)
> > The build.webkit.org bots are running Vista, btw.
> 
> I can verify that the same tests in the 'fast' subdirectory fail for me on my (soon to be reclaimed) Vista box.  Maybe we should just re-image with whatever the canary bots are running?

Alternately, I can check in win-vista results.

[Stephen Chenney]

LayoutTest failures for Chromium are being marked WontFix. The Bug is still accessible and referenced from TestExpectations.