Bug 115372

Summary: [Qt][WK2] r149292 made to fail hundreds of pixel tests, just rebaselining needed
Product: WebKit Reporter: Zoltan Arvai <zarvai>
Component: Tools / TestsAssignee: Zoltan Arvai <zarvai>
Status: RESOLVED INVALID    
Severity: Normal CC: jturcotte, kadam, noam
Priority: P2 Keywords: Qt
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
Bug Depends on:    
Bug Blocks: 79666, 110355    

Description Zoltan Arvai 2013-04-29 13:26:07 PDT
Most of the failing image results got grey cover after r149292. This could affect about 800 tests.
http://build.webkit.sed.hu/results/x86-64%20Linux%20Qt%20Release%20WebKit2%20%28Pixel%20Tests%29/r149293%20%283785%29/results.html

There were around 100 failure before the patch, that is "normal".
http://build.webkit.sed.hu/results/x86-64%20Linux%20Qt%20Release%20WebKit2%20%28Pixel%20Tests%29/r149291%20%283784%29/results.html

Also, I see about 100 text result that expanded with (drawsContent 1), but I assume it is normal and just a rebase needed that I can handle.

Can you take a look on this issue, please?
Comment 1 Zoltan Arvai 2013-04-30 06:04:05 PDT
After talking with Noam turned out the gray ones are not really regression, but normal behavior of WRT. I'll rebase those and we will see if any real regression happened.
Comment 2 Zoltan Arvai 2013-05-02 07:09:30 PDT
After rebaselinig the text expected results based on x86-32 Linux Qt Release WebKit2 bot, it seems repaint results differ after the patch:

eg:
fast/box-shadow/shadow-repaint.html

--- /ramdisk/qt-linux-32-release-webkit2/build/layout-test-results/fast/css-grid-layout/grid-element-change-columns-repaint-expected.txt
+++ /ramdisk/qt-linux-32-release-webkit2/build/layout-test-results/fast/css-grid-layout/grid-element-change-columns-repaint-actual.txt
@@ -1,7 +1,2 @@
 This test checks that changing the grid-columns lead to a repaint. The final grid element should be 250px * 50px, the grid item should be 50px * 50px. No trace of the elements before the grid-columns change should be seen.
-(repaint rects
-  (rect 8 38 100 50)
-  (rect 8 38 50 50)
-  (rect 8 38 250 50)
-)
 

There are other similar tests:
http://build.webkit.sed.hu/results/x86-32%20Linux%20Qt%20Release%20WebKit2/r149485%20%2835207%29/results.html

fast/repaint/layer-full-repaint.html
fast/repaint/overflow-flipped-writing-mode-table.html
fast/repaint/selection-gap-absolute-child.html
fast/repaint/selection-gap-fixed-child.html
fast/repaint/selection-gap-flipped-absolute-child.html
fast/repaint/selection-gap-flipped-fixed-child.html
fast/repaint/selection-gap-transformed-absolute-child.html
fast/repaint/selection-gap-transformed-fixed-child.html
svg/repaint/repaint-webkit-svg-shadow-container.html

Is this result correct or it's a regression?
Comment 3 Zoltan Arvai 2013-06-20 01:47:59 PDT
Skippend on QT WK2 in http://trac.webkit.org/changeset/151773.
Please, unskip it with the proper fix.
Comment 4 Zoltan Arvai 2013-06-20 01:49:18 PDT
(In reply to comment #3)
> Skippend on QT WK2 in http://trac.webkit.org/changeset/151773.
> Please, unskip it with the proper fix.

I'm sorry. This was the wrong window. Sorry for the noise.