Bug 50849 - [Qt] StillImage::draw() should use ContextShadow
Summary: [Qt] StillImage::draw() should use ContextShadow
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Canvas (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC OS X 10.5
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on: 34479
Blocks:
  Show dependency treegraph
 
Reported: 2010-12-10 15:19 PST by Helder Correia
Modified: 2010-12-14 04:43 PST (History)
5 users (show)

See Also:


Attachments
Patch (10.52 KB, patch)
2010-12-10 15:24 PST, Helder Correia
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Helder Correia 2010-12-10 15:19:42 PST
The branch that checks for a shadow requirement in StillImage::draw() is executed when drawing a canvas onto a canvas with ctx1.drawImage(canvas2, x, y).  The current implementation supports solid shadows only.  And if transformations are present, the offset gets transformed as well.  Thus, ContextShadow must be used instead in order to support blur and correct offset transformations.
Comment 1 Helder Correia 2010-12-10 15:24:29 PST
Created attachment 76268 [details]
Patch
Comment 2 Eric Seidel (no email) 2010-12-13 00:17:43 PST
Comment on attachment 76268 [details]
Patch

LGTM.  Assuming your test case failed before this change.
Comment 3 WebKit Commit Bot 2010-12-13 00:56:43 PST
Comment on attachment 76268 [details]
Patch

Clearing flags on attachment: 76268

Committed r73889: <http://trac.webkit.org/changeset/73889>
Comment 4 WebKit Commit Bot 2010-12-13 00:56:49 PST
All reviewed patches have been landed.  Closing bug.
Comment 5 Eric Seidel (no email) 2010-12-13 11:15:43 PST
pfeldman seems to believe this caused failures on cr win/linux:
http://code.google.com/p/chromium/issues/detail?id=66683
That seems unlikely/impossible.
Comment 6 Pavel Feldman 2010-12-14 04:43:00 PST
(In reply to comment #5)
> pfeldman seems to believe this caused failures on cr win/linux:
> http://code.google.com/p/chromium/issues/detail?id=66683
> That seems unlikely/impossible.

Well, the new test introduced with this commit does not pass on chromium. Why unlikely/impossible?

Please follow the link to find the failure details: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#master=ChromiumWebkit&tests=canvas-draw-canvas-on-canvas-shadow.html

Click "Show results" link and also click the rightmost vertical red bar in the test timeline to see the commit that caused the regression. It this case it is http://trac.webkit.org/changeset/73889 that you reviewed.