New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 717389 link

Starred by 0 users

Issue metadata

Status: Fixed
Owner:
Last visit > 30 days ago
Closed: Sep 2017
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

virtual/gpu/fast/canvas clip failures on debug

Project Member Reported by suzyh@chromium.org, May 2 2017

Issue description

The following two tests

virtual/gpu/fast/canvas/canvas-clip-rule.html
virtual/gpu/fast/canvas/canvas-path-context-clip.html

are failing consistently on

WebKit Linux Trusty (dbg) : crash
WebKit Win7 (dbg) : timeout

Flakiness dashboard link:
https://test-results.appspot.com/dashboards/flakiness_dashboard.html#testType=webkit_tests&tests=virtual%2Fgpu%2Ffast%2Fcanvas%2Fcanvas-clip-rule.html%2Cvirtual%2Fgpu%2Ffast%2Fcanvas%2Fcanvas-path-context-clip.html

First failures:
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20%28dbg%29/builds/1484
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Win7%20%28dbg%29/builds/9671

The crash log in Linux shows
"""
crash log for renderer (pid <unknown>):
STDOUT: <empty>
STDERR: ../../build/linux/debian_jessie_amd64-sysroot/usr/lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/debug/safe_iterator.264:
STDERR:     error: attempt to dereference a past-the-end iterator.
STDERR: 
STDERR: Objects involved in the operation:
STDERR: iterator "this" @ 0x0x7fff9690c038 {
STDERR:   state = past-the-end;
STDERR:   references sequence @ 0x0x7fff9690c038
STDERR: }
"""

The range of patches on those first failures are
https://chromium.googlesource.com/chromium/src/+log/75011c3d5ea9d22e2317d6faa7d9901eb2d2a2ca%5E..20c0b3059ef47f07a3e09e40f1dcc167da4817d8?pretty=fuller

I haven't had any luck identifying a culprit.
 

Comment 1 by suzyh@chromium.org, May 2 2017

Cc: junov@chromium.org
Adding failure test expectations in https://codereview.chromium.org/2856653006

Comment 2 by suzyh@chromium.org, May 2 2017

Potentially related:  issue 717019 

Comment 3 by vabr@chromium.org, May 2 2017

I am able to reproduce locally, attempting a bisect.

Comment 4 by vabr@chromium.org, May 2 2017

The bisect showed https://codereview.chromium.org/2847763004 as the culprit. I will revert that speculatively.
Project Member

Comment 5 by bugdroid1@chromium.org, May 2 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/5c9d124ac2131f4765f86902ff4cf711ed96c37e

commit 5c9d124ac2131f4765f86902ff4cf711ed96c37e
Author: vabr <vabr@chromium.org>
Date: Tue May 02 09:53:45 2017

Revert of Make InsertTextCommand not to apply style for empty selection (patchset #1 id:1 of https://codereview.chromium.org/2847763004/ )

Reason for revert:
Speculative revert, this seems to have broken some virtual/gpu/fast/canvas/ tests. More info on the associated bug.

BUG= 717389 

Original issue's description:
> Make InsertTextCommand not to apply typing style for empty selection
>
> This patch makes |InsertTextCommand::DoApply()| not to call |ApplyStyle()| for
> applying typing style when selection after inserting text is empty since
> |ApplyStyle()| doesn't work with empty selection.
>
> The  issue 714311  and the attached test case insert text into OPTION element to
> get empty selection after insertion, since we can't place selection inside
> OPTION element.
>
> BUG= 714311 
> TEST=run_webkit_unit_tests --gtest_filter=InsertTextCommandTest.WithTypingStyle
>
> Review-Url: https://codereview.chromium.org/2847763004
> Cr-Commit-Position: refs/heads/master@{#468080}
> Committed: https://chromium.googlesource.com/chromium/src/+/89209614959b6d3b2d4c4e8b015232663b4fcd87

TBR=xiaochengh@chromium.org,yoichio@chromium.org,yosin@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG= 714311 

Review-Url: https://codereview.chromium.org/2853213002
Cr-Commit-Position: refs/heads/master@{#468590}

[modify] https://crrev.com/5c9d124ac2131f4765f86902ff4cf711ed96c37e/third_party/WebKit/Source/core/editing/BUILD.gn
[modify] https://crrev.com/5c9d124ac2131f4765f86902ff4cf711ed96c37e/third_party/WebKit/Source/core/editing/commands/InsertTextCommand.cpp
[delete] https://crrev.com/0cd36b0f1fa2b9778ec33ec9db4ad9bcba094c7e/third_party/WebKit/Source/core/editing/commands/InsertTextCommandTest.cpp

The tests are still failing on the bots and I'm able to reproduce the crash locally after the revert in #5. I think I'll commit suzy@'s patch to add failure expectations.

Comment 7 by vabr@chromium.org, May 2 2017

SGTM, I am relanding my revert.
Project Member

Comment 8 by bugdroid1@chromium.org, May 2 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/6192661b83f875baac41bd4ff9c1d238a702d89c

commit 6192661b83f875baac41bd4ff9c1d238a702d89c
Author: vabr <vabr@chromium.org>
Date: Tue May 02 15:32:46 2017

Reland of Make InsertTextCommand not to apply style for empty selection (patchset #1 id:1 of https://codereview.chromium.org/2853213002/ )

Reason for revert:
Hmm, interestingly enough, the tests were still broken in the build which had the revert: https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20%28dbg%29/builds/1512

So I am relanding it.

Original issue's description:
> Revert of Make InsertTextCommand not to apply style for empty selection (patchset #1 id:1 of https://codereview.chromium.org/2847763004/ )
>
> Reason for revert:
> Speculative revert, this seems to have broken some virtual/gpu/fast/canvas/ tests. More info on the associated bug.
>
> BUG= 717389 
>
> Original issue's description:
> > Make InsertTextCommand not to apply typing style for empty selection
> >
> > This patch makes |InsertTextCommand::DoApply()| not to call |ApplyStyle()| for
> > applying typing style when selection after inserting text is empty since
> > |ApplyStyle()| doesn't work with empty selection.
> >
> > The  issue 714311  and the attached test case insert text into OPTION element to
> > get empty selection after insertion, since we can't place selection inside
> > OPTION element.
> >
> > BUG= 714311 
> > TEST=run_webkit_unit_tests --gtest_filter=InsertTextCommandTest.WithTypingStyle
> >
> > Review-Url: https://codereview.chromium.org/2847763004
> > Cr-Commit-Position: refs/heads/master@{#468080}
> > Committed: https://chromium.googlesource.com/chromium/src/+/89209614959b6d3b2d4c4e8b015232663b4fcd87
>
> TBR=xiaochengh@chromium.org,yoichio@chromium.org,yosin@chromium.org
> # Not skipping CQ checks because original CL landed more than 1 days ago.
> BUG= 714311 
>
> Review-Url: https://codereview.chromium.org/2853213002
> Cr-Commit-Position: refs/heads/master@{#468590}
> Committed: https://chromium.googlesource.com/chromium/src/+/5c9d124ac2131f4765f86902ff4cf711ed96c37e

TBR=xiaochengh@chromium.org,yoichio@chromium.org,yosin@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 717389 

Review-Url: https://codereview.chromium.org/2854123002
Cr-Commit-Position: refs/heads/master@{#468642}

[modify] https://crrev.com/6192661b83f875baac41bd4ff9c1d238a702d89c/third_party/WebKit/Source/core/editing/BUILD.gn
[modify] https://crrev.com/6192661b83f875baac41bd4ff9c1d238a702d89c/third_party/WebKit/Source/core/editing/commands/InsertTextCommand.cpp
[add] https://crrev.com/6192661b83f875baac41bd4ff9c1d238a702d89c/third_party/WebKit/Source/core/editing/commands/InsertTextCommandTest.cpp

Project Member

Comment 9 by bugdroid1@chromium.org, May 2 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/79091e1a33ad072fa80e69cb7a9b031a0c982009

commit 79091e1a33ad072fa80e69cb7a9b031a0c982009
Author: suzyh <suzyh@chromium.org>
Date: Tue May 02 16:13:45 2017

Add failure expectations for canvas clip tests

virtual/gpu/fast/canvas/canvas-clip-rule.html and
virtual/gpu/fast/canvas/canvas-path-context-clip.html
are crashing on WebKit Linux Trusty (dbg)
and timing out on WebKit Win7 (dbg).
Added entries to TestExpectations.

TBR=junov@chromium.org
BUG= 717389 

Review-Url: https://codereview.chromium.org/2856653006
Cr-Commit-Position: refs/heads/master@{#468662}

[modify] https://crrev.com/79091e1a33ad072fa80e69cb7a9b031a0c982009/third_party/WebKit/LayoutTests/TestExpectations

Cc: -junov@chromium.org
Labels: -Sheriff-Chromium
Owner: junov@chromium.org
Status: Assigned (was: Untriaged)
suzy@'s patch in #1 has been committed and the test failures are now ignored. 

junov@: Could you please take a look the failing layout tests, or reassign to someone else who can take a look? 
Labels: -Pri-1 BugSource-Chromium Test-Layout PaintTeamTriaged-20170502 Pri-2
LayoutTest failures are not a P1.
Project Member

Comment 12 by bugdroid1@chromium.org, May 10 2017

Labels: merge-merged-3071
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/bd2a15d23a22163e93b7510711f22322c5a4844c

commit bd2a15d23a22163e93b7510711f22322c5a4844c
Author: Yoshifumi Inoue <yosin@chromium.org>
Date: Wed May 10 09:02:11 2017

Reland of Make InsertTextCommand not to apply style for empty selection (patchset #1 id:1 of https://codereview.chromium.org/2853213002/ )

Reason for revert:
Hmm, interestingly enough, the tests were still broken in the build which had the revert: https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty%20%28dbg%29/builds/1512

So I am relanding it.

Original issue's description:
> Revert of Make InsertTextCommand not to apply style for empty selection (patchset #1 id:1 of https://codereview.chromium.org/2847763004/ )
>
> Reason for revert:
> Speculative revert, this seems to have broken some virtual/gpu/fast/canvas/ tests. More info on the associated bug.
>
> BUG= 717389 
>
> Original issue's description:
> > Make InsertTextCommand not to apply typing style for empty selection
> >
> > This patch makes |InsertTextCommand::DoApply()| not to call |ApplyStyle()| for
> > applying typing style when selection after inserting text is empty since
> > |ApplyStyle()| doesn't work with empty selection.
> >
> > The  issue 714311  and the attached test case insert text into OPTION element to
> > get empty selection after insertion, since we can't place selection inside
> > OPTION element.
> >
> > BUG= 714311 
> > TEST=run_webkit_unit_tests --gtest_filter=InsertTextCommandTest.WithTypingStyle
> >
> > Review-Url: https://codereview.chromium.org/2847763004
> > Cr-Commit-Position: refs/heads/master@{#468080}
> > Committed: https://chromium.googlesource.com/chromium/src/+/89209614959b6d3b2d4c4e8b015232663b4fcd87
>
> TBR=xiaochengh@chromium.org,yoichio@chromium.org,yosin@chromium.org
> # Not skipping CQ checks because original CL landed more than 1 days ago.
> BUG= 714311 
>
> Review-Url: https://codereview.chromium.org/2853213002
> Cr-Commit-Position: refs/heads/master@{#468590}
> Committed: https://chromium.googlesource.com/chromium/src/+/5c9d124ac2131f4765f86902ff4cf711ed96c37e

TBR=xiaochengh@chromium.org,yoichio@chromium.org,yosin@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 717389 

Review-Url: https://codereview.chromium.org/2854123002
Cr-Commit-Position: refs/heads/master@{#468642}
(cherry picked from commit 6192661b83f875baac41bd4ff9c1d238a702d89c)

Review-Url: https://codereview.chromium.org/2872243002 .
Cr-Commit-Position: refs/branch-heads/3071@{#499}
Cr-Branched-From: a106f0abbf69dad349d4aaf4bcc4f5d376dd2377-refs/heads/master@{#464641}

[modify] https://crrev.com/bd2a15d23a22163e93b7510711f22322c5a4844c/third_party/WebKit/Source/core/editing/BUILD.gn
[modify] https://crrev.com/bd2a15d23a22163e93b7510711f22322c5a4844c/third_party/WebKit/Source/core/editing/commands/InsertTextCommand.cpp
[add] https://crrev.com/bd2a15d23a22163e93b7510711f22322c5a4844c/third_party/WebKit/Source/core/editing/commands/InsertTextCommandTest.cpp

Status: Fixed (was: Assigned)

Sign in to add a comment