Rebaseline the expected image for fast/forms/select-style.html for Mac. I had to include this in the following bug, but I forgot. Retry. Bug 25742: Fix RenderThemeChromiumWin::paintTextFieldInternal() not to hide background image. https://bugs.webkit.org/show_bug.cgi?id=25742
Created attachment 30994 [details] Proposed fix for 25742
Running the pixel test for fast/forms/select-style.html on Leopard 10.5.7 doesn't show any issues: $ ./WebKitTools/Scripts/run-webkit-tests --debug --pixel fast/forms/select-style.html [...] ** BUILD SUCCEEDED ** Running tests from /Volumes/Data/WebKit.git/LayoutTests Enabling pixel tests with a tolerance of 0.1% WARNING: Temporarily changing the main display color profile: The colors on your screen will change for the duration of the testing. This allows the pixel tests to have consistent color values across all machines. Testing 1 test cases. fast/forms . 0.57s total testing time all 1 test cases succeeded Are you sure this change is needed?
(In reply to comment #2) > Running the pixel test for fast/forms/select-style.html on Leopard 10.5.7 > doesn't show any issues: This is using ToT WebKit r44471.
Seems this is already fixed, per ddkilzer's comments.
Re-opening... This patch should be still applied. The reason why the run-webkit-tests passes is because select-style-expected.checksum has been already replaced by the patch in Bug 25742. run-webkit-tests doesn't use png file for comparison. It just compares generated image's checksum with the expected checksum. select-style-expected.png is out of date. We should replace or we'll show an incorrect image on test results.
Comment on attachment 30994 [details] Proposed fix for 25742 Forgot to set review flag
Comment on attachment 30994 [details] Proposed fix for 25742 It would be nice to mention (in the changelog) that this test needed to be rebaselined due to r43903, which added more to this test. At first, I was wondering why this patch was needed.
Created attachment 31535 [details] Proposed fix for 26207 (rev 2)
(In reply to comment #7) > (From update of attachment 30994 [details] [review]) > It would be nice to mention (in the changelog) that this test needed to be > rebaselined due to r43903, which added more to this test. > > At first, I was wondering why this patch was needed. > Added. Sorry for confusing.
Assign to levin for landing.
Committed as http://trac.webkit.org/changeset/45187.
(In reply to comment #11) > Committed as http://trac.webkit.org/changeset/45187. > Thank you.