Bug 62907

Summary:

[chromium] force rebaseline-chromium-webkit-tests to use the chromium port

Product:

WebKit

Reporter:

Tony Chang <tony>

Component:

New Bugs

Assignee:

Tony Chang <tony>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

dpranke, webkit.review.bot

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	none

Description Tony Chang 2011-06-17 13:54:35 PDT

[chromium] force rebaseline-chromium-webkit-tests to use the chromium port

Comment 1 Tony Chang 2011-06-17 13:56:27 PDT

Created attachment 97649 [details]
Patch

Comment 2 Dirk Pranke 2011-06-17 13:58:48 PDT

Comment on attachment 97649 [details]
Patch

This is fine. See the email I just sent you, though, I don't think this explains rniwa's problem (it might fix it, but there's probably another bug in here somewhere). 

Ideally we'd find whatever ImageDiff was built, but that can certainly be a separate patch.

Comment 3 Tony Chang 2011-06-17 14:02:28 PDT

See the ChangeLog description.  Even with a tolerance of 0, it doesn't do an exact pixel match.  Here's the code in question:
http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/cg/ImageDiffCG.cpp#L118

It calculates a distance and if the distance is small enough, it considers the pixels the same.

Comment 4 Tony Chang 2011-06-17 14:02:59 PDT

*** Bug 61833 has been marked as a duplicate of this bug. ***

Comment 5 Dirk Pranke 2011-06-17 14:06:54 PDT

Wow, that just seems like a bug. It seems like --tolerance=0 should mean "exact match".

Comment 6 WebKit Review Bot 2011-06-17 14:48:25 PDT

Comment on attachment 97649 [details]
Patch

Clearing flags on attachment: 97649

Committed r89174: <http://trac.webkit.org/changeset/89174>

Comment 7 WebKit Review Bot 2011-06-17 14:48:31 PDT

All reviewed patches have been landed.  Closing bug.