Bug 71253

Added to the Skipped list: http://trac.webkit.org/changeset/98918

(In reply to comment #1) > Added to the Skipped list: http://trac.webkit.org/changeset/98918 Ooops, I broke it. Real skip landed in http://trac.webkit.org/changeset/98920.

Created attachment 113112 [details] failing images

I hope it works again for Qt since r99543. Can anyone confirm?

(In reply to comment #4) > I hope it works again for Qt since r99543. Can anyone confirm? Zoltán checked it and we got same results as the attached diff.

(In reply to comment #3) > Created an attachment (id=113112) [details] > failing images Could you reupload this in a way its viewable? as SVN user I'm not sure what to do with your patch.

(In reply to comment #6) > (In reply to comment #3) > > Created an attachment (id=113112) [details] [details] > > failing images > Could you reupload this in a way its viewable? as SVN user I'm not sure what to do with your patch. Just click here: https://bugs.webkit.org/attachment.cgi?id=113112&action=review red one is Mac png and green one in actual png on Qt.

(In reply to comment #7) > Just click here: https://bugs.webkit.org/attachment.cgi?id=113112&action=review. Ah great! So here's the analysis: > svg/zoom/page/zoom-img-preserveAspectRatio-support-1.html This flakiness is known, and tracked by another bug, will look it up. > svg/zoom/page/zoom-background-images.html > svg/zoom/page/zoom-svg-as-background-with-relative-size.html Difference between Qt/DRT and Mac/DRT. #define MinimumZoomMultiplier 0.5f #define MaximumZoomMultiplier 3.0f #define ZoomMultiplierRatio 1.2f We can only zoom in three times from DRT/Mac, zooming in 4 or 5 or 6 times, results in a scale factor < 0.5f, and is thus ignored! I'll fix the test, to avoid zooming in four times. > fast/backgrounds/size/contain-and-cover-zoomed.html Real regression, not sure what happens.

Created attachment 114058 [details] Patch Fix zoom problems.

Comment on attachment 114058 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=114058&action=review LGTM > LayoutTests/ChangeLog:8 > + Adress some problems seen on Qt/Chromium: Address > LayoutTests/ChangeLog:10 > + - Don't zoom out more than three times, otherwhise its ignored on Mac/DRT, where the minimum zoom factor is clamped to 0.5f. Otherwise

Comment on attachment 114058 [details] Patch Landed in r99566. Can anyone with a Qt build retry please? :-)

Comment on attachment 114058 [details] Patch Attachment 114058 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/10370086 New failing tests: svg/zoom/page/zoom-replaced-intrinsic-ratio-001.htm

Created attachment 114071 [details] result Still failing with r99569.

(In reply to comment #13) > Created an attachment (id=114071) [details] > result > > Still failing with r99569. Could you upload it again in a different format? $old-run-webkit-tests --platform mac -p --reset-results --add-platform-exceptions TEST_NAMES and then upload patch from "git diff --binary HEAD"

Created attachment 114078 [details] image-diff r99569 layout diff

Landed a rebaseline in r99572. I think there's one failure remaining only. I filed bug 71673 for the remaining issue with zooming SVGs in <object>.

(In reply to comment #16) > Landed a rebaseline in r99572. I think there's one failure remaining only. > I filed bug 71673 for the remaining issue with zooming SVGs in <object>. fast/backgrounds/size/contain-and-cover-zoomed.html still fails, because squirrelfish is still missin on the Qt png.

Created attachment 114219 [details] Fix two failing tests

Comment on attachment 114219 [details] Fix two failing tests Oops, this belongs to bug 71830, sorry for the confusion.

Can anyone update me, if there are still problems on Qt? The squirrelfish issue is still not solved, I guess?

No change with r99898. The results are the same as image-diff patch shows.

(In reply to comment #21) > No change with r99898. The results are the same as image-diff patch shows. zoom-img-preserveAspectRatio-support-1 problem should be fixed by r99937. I think the only remaining failure is contain-and-cover-zoomed for Qt. Hm, could you open it manually with a Browser, and then zoom in. I guess Squirrelfish appears, but then disappears after zooming, is that right?

Created attachment 114659 [details] image-diff2 I tried it with QtTestBrowser from r99937. There is a problem with the LayoutTests/fast/backgrounds/size/resources/SquirrelFish.svg: <?xml version="1." standalone="no"?> A zero is missing from the version. I uplodaed image-diff2 that shows the differencies with corrected svg file.

(In reply to comment #23) > Created an attachment (id=114659) [details] > image-diff2 > > I tried it with QtTestBrowser from r99937. > > There is a problem with the LayoutTests/fast/backgrounds/size/resources/SquirrelFish.svg: > <?xml version="1." standalone="no"?> > A zero is missing from the version. > > I uplodaed image-diff2 that shows the differencies with corrected svg file. Ah perfect! I forgot to fix contain-and-cover-zoomed.html - it still zomes out 4 times, instead of 3 times, which is ignored on Mac. Will upload a patch containing your fix as well to SquirrelFish.svg That's it we're finished :-)

(In reply to comment #23) > There is a problem with the LayoutTests/fast/backgrounds/size/resources/SquirrelFish.svg: I just wonder: shouldn't this affect fast/backgrounds/size/contain-and-cover.html as well? Is this skipped for Qt?

Created attachment 114663 [details] Patch

Comment on attachment 114663 [details] Patch r=me

Landed in r99941. This will require rebaelines for gtk/qt/chromium, I think.

(In reply to comment #25) > (In reply to comment #23) > > There is a problem with the LayoutTests/fast/backgrounds/size/resources/SquirrelFish.svg: > I just wonder: shouldn't this affect fast/backgrounds/size/contain-and-cover.html as well? Is this skipped for Qt? Its not on qt skip list. But it was also affected by this issue. run-webkit-tests -p --platform mac fast/backgrounds/size/contain-and-cover.html was failing with buggy svg file (diff 18%). With corrected svg there is only 0.01% diff.