There are several roles like SplitterRole, ParagraphRole, etc. that are only currently used on some platforms, like GTK. Let's get rid of the platform-specific logic in WebCore/accessibility, and just map them to the proper native roles on each platform in the platform-specific subdirectories. This will help avoid subtle bugs where a test behaves differently on one platform because other logic in WebCore uses the role, and also make it easier for platforms to take advantage of more roles as the native roles get expanded.
Created attachment 160457 [details] Patch
Comment on attachment 160457 [details] Patch looks good
Please wait for approval from abarth@webkit.org, dglazkov@chromium.org, fishd@chromium.org, jamesr@chromium.org or tkent@chromium.org before submitting, as this patch contains changes to the Chromium public API. See also https://trac.webkit.org/wiki/ChromiumWebKitAPI.
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass chromium-ews (chromium-xvfb): Output: http://queues.webkit.org/results/13600214
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/13601172
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass efl-ews (efl): Output: http://queues.webkit.org/results/13597217
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass qt-ews (qt): Output: http://queues.webkit.org/results/13596205
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass qt-wk2-ews (qt): Output: http://queues.webkit.org/results/13599202
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass win-ews (win): Output: http://queues.webkit.org/results/13591338
Comment on attachment 160457 [details] Patch Attachment 160457 [details] did not pass cr-android-ews (chromium-android): Output: http://queues.webkit.org/results/13605140
Sorry, this won't compile under after 94302 is re-landed.
Created attachment 162705 [details] Rebase
Created attachment 162709 [details] Rebase
Created attachment 162719 [details] Rebase
Created attachment 162786 [details] Patch for landing
Comment on attachment 162786 [details] Patch for landing Rejecting attachment 162786 [details] from commit-queue. New failing tests: platform/chromium/accessibility/chromium-only-roles.html accessibility/canvas-fallback-content.html accessibility/div-within-anchors-causes-crash.html accessibility/adjacent-continuations-cause-assertion-failure.html accessibility/inline-continuations.html Full output: http://queues.webkit.org/results/13772837
Created attachment 162882 [details] Patch for landing
Comment on attachment 162882 [details] Patch for landing Clearing flags on attachment: 162882 Committed r127936: <http://trac.webkit.org/changeset/127936>
All reviewed patches have been landed. Closing bug.
(In reply to comment #18) > (From update of attachment 162882 [details]) > Clearing flags on attachment: 162882 > > Committed r127936: <http://trac.webkit.org/changeset/127936> This change has caused multiple accessibility tests to fail. See <http://build.webkit.org/results/Apple%20MountainLion%20Debug%20WK1%20(Tests)/r127946%20(650)/results.html>.
(In reply to comment #20) > (In reply to comment #18) > > (From update of attachment 162882 [details] [details]) > > Clearing flags on attachment: 162882 > > > > Committed r127936: <http://trac.webkit.org/changeset/127936> > > This change has caused multiple accessibility tests to fail. See <http://build.webkit.org/results/Apple%20MountainLion%20Debug%20WK1%20(Tests)/r127946%20(650)/results.html>. Filed bug 96173.
It also broke three windows tests: accessibility/adjacent-continuations-cause-assertion-failure.html accessibility/deleting-iframe-destroys-axcache.html accessibility/div-within-anchors-causes-crash.html But all the diff's are along the lines of : client editable text - grouping + client editable text editable text - grouping + client editable text - grouping + client editable text - grouping + client editable text editable text editable text I suspect that due to Windows accessibility limitations that these 3 tests just need Windows specific results.
(In reply to comment #22) > It also broke three windows tests: > > accessibility/adjacent-continuations-cause-assertion-failure.html > accessibility/deleting-iframe-destroys-axcache.html > accessibility/div-within-anchors-causes-crash.html Filed this bug and uploaded a patch for review: https://bugs.webkit.org/show_bug.cgi?id=96219