This is a regression of https://bugs.webkit.org/show_bug.cgi?id=139405. The consensus in the last w3 csswg meeting was to follow the current specs for the non-intrisic SVG image case. Although the nine-piece algorithm for the non-intrisic SVG is not guaranteed to be meaningful in all cases, it is working as expected in general.
<rdar://problem/21995596>
Created attachment 262652 [details] Patch
Comment on attachment 262652 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=262652&action=review > Source/WebCore/ChangeLog:10 > + The nine-pieces algorithm should be applied to the border-image regardless > + whether the image has an intrinsic size or not. It is not guaranteed to have > + a meaningful border-image in all the cases of non-intrinsic size images. But > + it should work as expected in most of the cases. I am a little bit concerned about “should work as expected in most of the cases”. Is there consensus with other implementors on how this should work? Is the test coverage sufficient?
Comment on attachment 262652 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=262652&action=review >> Source/WebCore/ChangeLog:10 >> + it should work as expected in most of the cases. > > I am a little bit concerned about “should work as expected in most of the cases”. Is there consensus with other implementors on how this should work? > > Is the test coverage sufficient? This link https://lists.w3.org/Archives/Public/www-style/2015Aug/0170.html is part of the discussion about this issue on the www-style list:. The issue was also discussed in the last w3 css meeting https://wiki.csswg.org/planning/paris-2015. I also discussed it on webkit with David Vest https://bugs.webkit.org/show_bug.cgi?id=139405. And it seems the consensus is to follow the specs even with its unclarity with the case where the SVG itself has no intrinsic size but the elements inside it are fixed size. FireFox is still behaving the same as the shipped Safari; I am not sure about their plan. Chrome is behaving the same as WebKit with this patch. I attached a test case to this bug. I will add more tests for the case where the border-image show meaningful results.
Created attachment 262699 [details] test case
Actually you can try the test case which is attached to https://bugs.webkit.org/show_bug.cgi?id=139405
Created attachment 262898 [details] Patch
New test cases for the relative lengths cases were added. The intrinsic SVG with fixed lengths and the non-intrinsic SVG with relative lengths are the scenarios we care about. The non-intrinsic SVG with fixed lengths is not expected to show border-image correctly in all cases.
Comment on attachment 262898 [details] Patch Attachment 262898 [details] did not pass mac-ews (mac): Output: http://webkit-queues.webkit.org/results/272046 New failing tests: fast/borders/border-image-fill-no-intrinsic-size.html
Created attachment 262903 [details] Archive of layout-test-results from ews101 for mac-mavericks The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: ews101 Port: mac-mavericks Platform: Mac OS X 10.9.5
Comment on attachment 262898 [details] Patch Attachment 262898 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.webkit.org/results/272054 New failing tests: fast/borders/border-image-fill-no-intrinsic-size.html
Created attachment 262904 [details] Archive of layout-test-results from ews106 for mac-mavericks-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: ews106 Port: mac-mavericks-wk2 Platform: Mac OS X 10.9.5
Created attachment 262907 [details] Patch
Re-uploaded an new patch because a file was missing.
Comment on attachment 262907 [details] Patch Clearing flags on attachment: 262907 Committed r190883: <http://trac.webkit.org/changeset/190883>
All reviewed patches have been landed. Closing bug.