Created attachment 71566 [details] fix patch

Created attachment 71567 [details] same patch but less intrusive. I talked to yael two weeks ago about it, and even set her a patch for review. it is essentially the same as yours. could we go this way, instead?

Comment on attachment 71566 [details] fix patch see https://bugs.webkit.org/show_bug.cgi?id=48134#c2

(In reply to comment #2) > Created an attachment (id=71567) [details] > same patch but less intrusive. > > I talked to yael two weeks ago about it, and even set her a patch for review. it is essentially the same as yours. could we go this way, instead? No problem. :) I can see the effort of making non-DRT mode working, but if it's not 100% accurate, it would cause confusion. I'd rather not to make it work for non-DRT.

> I can see the effort of making non-DRT mode working, but if it's not 100% accurate, it would cause confusion. I'd rather not to make it work for non-DRT. Why? :) The other method was not being accurate for me too, so I did not put the patch up for review. <select> related tests were failing on both qt and mac for me. What do you see?

Synthesizing events that are exactly like the ones *** This bug has been marked as a duplicate of bug 16735 ***

Sorry, hit "Commit" by accident. But see bug 16735 for related info.

(In reply to comment #7) > Sorry, hit "Commit" by accident. But see bug 16735 for related info. It seems the problem has been noticed and has been worked on for a long time. But I don't see an immediate close of the bug. Before WebCore is fixed, I guess we still have to use eventSender.

Created attachment 71580 [details] fix patch 2 In this patch, I use Antonio's suggestion to convert event names in spatial-navigation-utils.js. It has less impact on the current code and also, in the future, when bug 16735 is fixed, we can simply roll back the change to support both DRT and non-DRT mode. The failed multiple-select case will be addressed in a separate bug.

Comment on attachment 71580 [details] fix patch 2 Rejecting patch 71580 from commit-queue. Failed to run "['./WebKitTools/Scripts/webkit-patch', '--status-host=queues.webkit.org', 'build-and-test', '--no-clean', '--no-update', '--test', '--quiet', '--non-interactive']" exit_code: 2 Last 500 characters of output: ing 21625 test cases. fast/events/spatial-navigation/snav-single-select.html -> timed out Sampling process 67159 for 10 seconds with 10 milliseconds of run time between samples Sampling completed, processing symbols... Sample analysis of process 67159 written to file /Users/eseidel/Library/Logs/DumpRenderTree/HangReport.txt Exiting early after 1 failures. 7696 tests run. 208.48s total testing time 7695 test cases (99%) succeeded 1 test case (<1%) timed out 2 test cases (<1%) had stderr output Full output: http://queues.webkit.org/results/4721022

Created attachment 71592 [details] fix patch 3

Comment on attachment 71580 [details] fix patch 2 Cleared Antonio Gomes's review+ from obsolete attachment 71580 [details] so that this bug does not appear in http://webkit.org/pending-commit.

Comment on attachment 71592 [details] fix patch 3 Clearing flags on attachment: 71592 Committed r70481: <http://trac.webkit.org/changeset/70481>

All reviewed patches have been landed. Closing bug.

You probably know this already, but changing events generated using event.initKeyboardEvent() to be like real ones shouldn't be approached lightly. See also: bug 9933 and bug 16735.

(In reply to comment #15) > You probably know this already, but changing events generated using event.initKeyboardEvent() to be like real ones shouldn't be approached lightly. See also: bug 9933 and bug 16735. Thanks, Alexey. I created the bug with this title hoping to fix the root cause but instead I just changed the test to use eventSender. Hopefully, someone will fix the above two bugs so we can run the test from browser.

Revision r70481 cherry-picked into qtwebkit-2.1.x with commit 8a14b20 <http://gitorious.org/webkit/qtwebkit/commit/8a14b20>