Bug 42190 - Failing 2d.path.stroke.prune.curve philip canvas test
Summary: Failing 2d.path.stroke.prune.curve philip canvas test
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Layout and Rendering (show other bugs)
Version: 528+ (Nightly build)
Hardware: All OS X 10.6
: P2 Normal
Assignee: Matthew Delaney
URL:
Keywords:
Depends on: 42188
Blocks:
  Show dependency tree / graph
 
Reported: 2010-07-13 13:30 PDT by Matthew Delaney
Modified: 2010-07-20 00:07 PDT (History)
3 users (show)

See Also:

Attachments
Patch (4.77 KB, patch)
2010-07-16 22:00 PDT, Matthew Delaney 		no flags Details | Formatted Diff | Diff
Patch (5.73 KB, patch)
2010-07-17 19:13 PDT, Matthew Delaney 		kenneth: review+
commit-queue: commit-queue-
Details | Formatted Diff | Diff
Previous Patch (without extra whitespace deletion) (4.89 KB, patch)
2010-07-19 13:23 PDT, Matthew Delaney 		no flags Details | Formatted Diff | Diff
Show Obsolete (2) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Matthew Delaney 2010-07-13 13:30:46 PDT 
Safari 5 fails 2d.path.stroke.prune.curve
Comment 1 Matthew Delaney 2010-07-16 22:00:11 PDT 
Created attachment 61874 [details]
Patch
Comment 2 Andreas Kling 2010-07-17 18:41:01 PDT 
Comment on attachment 61874 [details]
Patch

>WebCore/html/canvas/CanvasRenderingContext2D.cpp:562
> +      
Stray whitespace.

>WebCore/html/canvas/CanvasRenderingContext2D.cpp:599
> +      
Ditto.

>WebCore/html/canvas/CanvasRenderingContext2D.cpp:613
> +      
Ditto.

>WebCore/html/canvas/CanvasRenderingContext2D.cpp:563
> +      FloatRect boundRect = m_path.boundingRect();
It's a bit sad that we have to do this, getting the boundingRect() can be rather expensive.
Qt will cache its results internally and reuse them when stroking/filling the path later on, but other back-ends may not be so lucky.
Unfortunately I don't have a better suggestion. ;(
Comment 3 Matthew Delaney 2010-07-17 19:13:55 PDT 
Created attachment 61896 [details]
Patch
Comment 4 Kenneth Rohde Christiansen 2010-07-19 06:25:32 PDT 
Comment on attachment 61896 [details]
Patch

Normally we keep style changes in separate patches. But the rest seems sensible to me, and Andreas Kling has had a look as well, so r=me.
Comment 5 WebKit Commit Bot 2010-07-19 08:25:40 PDT 
Comment on attachment 61896 [details]
Patch

Rejecting patch 61896 from commit-queue.

Failed to run "[u'/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/svn-apply', u'--reviewer', u'Kenneth Rohde Christiansen', u'--force']" exit_code: 1
Last 500 characters of output:
Core/ChangeLog
Hunk #1 succeeded at 1 with fuzz 3.
patching file WebCore/html/canvas/CanvasRenderingContext2D.cpp
Hunk #5 FAILED at 621.
Hunk #6 succeeded at 646 (offset 2 lines).
1 out of 6 hunks FAILED -- saving rejects to file WebCore/html/canvas/CanvasRenderingContext2D.cpp.rej
patching file WebCore/platform/graphics/cg/PathCG.cpp
patching file LayoutTests/ChangeLog
Hunk #1 succeeded at 1 with fuzz 3.
patching file LayoutTests/platform/mac/Skipped
Hunk #1 succeeded at 211 (offset -2 lines).

Full output: http://queues.webkit.org/results/3588225
Comment 6 Matthew Delaney 2010-07-19 13:23:46 PDT 
Created attachment 61986 [details]
Previous Patch (without extra whitespace deletion)

Ugh, it looks like deleting some whitespace in one of the otherwise untouched methods causes svn-apply to fail. Here's a new patch without those whitespace deletes that patches on my machine with no problem.
Comment 7 WebKit Commit Bot 2010-07-20 00:07:27 PDT 
Comment on attachment 61986 [details]
Previous Patch (without extra whitespace deletion)

Clearing flags on attachment: 61986

Committed r63727: <http://trac.webkit.org/changeset/63727>
Comment 8 WebKit Commit Bot 2010-07-20 00:07:33 PDT 
All reviewed patches have been landed.  Closing bug.