This should allow unskipping accessibility/image-map1.html and accessibility/image-map2.html
accessibility/image-map1.html is failing since r155022. http://trac.webkit.org/changeset/155022 The diff: --- /home/slave/webkitgtk/gtk-linux-64-release/build/layout-test-results/accessibility/image-map1-expected.txt +++ /home/slave/webkitgtk/gtk-linux-64-release/build/layout-test-results/accessibility/image-map1-actual.txt @@ -4,12 +4,18 @@ On success, you will see a series of "PASS" messages, followed by "TEST COMPLETE". -PASS body.childAtIndex(0).role is 'AXRole: AXLink' -FAIL body.childAtIndex(0).title should be AXTitle: Link1. Was AXTitle: . -PASS body.childAtIndex(1).role is 'AXRole: AXLink' -FAIL body.childAtIndex(1).title should be AXTitle: Link2. Was AXTitle: . -PASS body.childAtIndex(2).role is 'AXRole: AXLink' -FAIL body.childAtIndex(2).title should be AXTitle: Link3. Was AXTitle: . +Link1 role: AXRole: AXLink +Link1 title: AXTitle: +Link1 description: AXDescription: + +Link2 role: AXRole: AXLink +Link2 title: AXTitle: +Link2 description: AXDescription: + +Link3 role: AXRole: AXLink +Link3 title: AXTitle: +Link3 description: AXDescription: + PASS successfullyParsed is true TEST COMPLETE
Note that both tests are being run (i.e. are not skipped), and have GTK-specific baselines.
Created attachment 214972 [details] Patch proposal This patch mimics what the Mac does and fixes the issue. The situation has slighly changed since this bug was reported, but I checked results now and this patch seems to be the right one. As a plus, it allows us to share the expectations between GTK, EFL and Mac, which is good too.
Committed r158104: <http://trac.webkit.org/changeset/158104>