Bug 89494 - [Chromium] fast/canvas/webgl/gl-teximage.html needs rebaseline
Summary: [Chromium] fast/canvas/webgl/gl-teximage.html needs rebaseline
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebGL (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Jeff Timanus
Depends on: 86275
  Show dependency treegraph
Reported: 2012-06-19 11:45 PDT by Jeff Timanus
Modified: 2012-08-08 15:13 PDT (History)
7 users (show)

See Also:

Patch (1.24 KB, patch)
2012-08-07 12:21 PDT, Jeff Timanus
no flags Details | Formatted Diff | Diff
Patch (10.55 KB, patch)
2012-08-07 13:02 PDT, Jeff Timanus
no flags Details | Formatted Diff | Diff
Patch (7.04 KB, patch)
2012-08-08 13:21 PDT, Jeff Timanus
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Jeff Timanus 2012-06-19 11:45:18 PDT
wkbug.com/86275 causes platform/chromium/virtual/gpu/fast/canvas/webgl/gl-teximage.html to fail with text differences. 

A new rebaseline is required.
Comment 1 Jeff Timanus 2012-08-07 12:21:31 PDT
Created attachment 156971 [details]

EWS trial to capture current state of failing test output.
Comment 2 Jeff Timanus 2012-08-07 13:02:28 PDT
Created attachment 156982 [details]
Comment 3 Jeff Timanus 2012-08-07 13:06:39 PDT
Comment on attachment 156982 [details]


White-space modifications were a result of the webkit-patch tool.
Comment 4 Kenneth Russell 2012-08-08 02:07:07 PDT
gman, zmo, if one of you could please review this perhaps enne could marshal it through.
Comment 5 Adrienne Walker 2012-08-08 10:33:11 PDT
Comment on attachment 156982 [details]

In general, you don't need a review to land a rebaseline.  You can just land with a comment about it being unreviewed.

If you're asking for a review, I'd probably say that the test should be modified so that it behaves the same on all platforms rather than checking in a different expectation.  The fact that the virtual/gpu directory prints 50% for every value but the checked in expectations says 49% or 51% (but still passes) seems like a wart in the test case that should be removed.  Maybe it shouldn't print any percent if it's a passing value.
Comment 6 Jeff Timanus 2012-08-08 13:21:18 PDT
Created attachment 157278 [details]
Comment 7 Jeff Timanus 2012-08-08 13:23:13 PDT
Comment on attachment 157278 [details]

This patch removed the output of the percentage of passing pixels in DRT runs, but not when running in Chrome.
Comment 8 Adrienne Walker 2012-08-08 14:16:38 PDT
Comment on attachment 157278 [details]

R=me.  Thanks for that change.
Comment 9 WebKit Review Bot 2012-08-08 15:13:39 PDT
Comment on attachment 157278 [details]

Clearing flags on attachment: 157278

Committed r125099: <http://trac.webkit.org/changeset/125099>
Comment 10 WebKit Review Bot 2012-08-08 15:13:44 PDT
All reviewed patches have been landed.  Closing bug.