[Webkit-unassigned] [Bug 133512] [ATK] Missing 'selection-changed' signal when navigating a combo box with keyboard

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=133512


Mario Sanchez Prada <mario at webkit.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #232560|0                           |1
        is obsolete|                            |
 Attachment #232560|review?                     |
               Flag|                            |
 Attachment #233052|                            |review?
               Flag|                            |




--- Comment #12 from Mario Sanchez Prada <mario at webkit.org>  2014-06-13 08:25:16 PST ---
Created an attachment (id=233052)
 --> (https://bugs.webkit.org/attachment.cgi?id=233052&action=review)
Path proposal

(In reply to comment #11)
> I have been taking a look to this and I can't see why the combo-box test is timing
> out in the Mac, as I haven't found anything in the HTML/JavaScript code, nor in the
> two changes in RenderMenuList and AccessibilityMenuList, that tell me why it's not
> finishing.
>
> I'm currently trying to get a OS X box to try this out, but not sure how long that
> will take, so I would appreciate any help with this in the meanwhile, if possible.
> 

I finally got a Mac to try this out and indeed I managed to reproduce the timeout, which is good. However, I could also check that such a timeout has nothing to do with the changes in RenderMenuList.cpp and AccessibilityMenuList.cpp, but with the test itself, which is blocked after each call to eventSender.keyDown() for some reason.

Actually, executing the test manually I can see the popup showing up in the screen, and the test only finishes if I manually close it with Esc key. However, even if I do that I don't see any notification being fired, and a more in deep investigation reveled that the code inside the "if (!chidObjects.isEmpty() { ... }" is never been executed, and I'm not sure what ithat means. Perhaps that notification over the popup is not meant to be emitted in the Mac, but I'm not sure.

I'm the meantime, and in order to try to fix this issue at least in ATK based platforms, I'm attaching a new patch that adds this test to TestExpectations, together with other tests that might be in a similar situation.

Please review, thanks!

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.