[Webkit-unassigned] [Bug 35350] DumpRenderTree should allow tests with modal dialogs

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Jun 14 15:35:27 PDT 2010


https://bugs.webkit.org/show_bug.cgi?id=35350





--- Comment #9 from Prasad Tammana <prasadt at chromium.org>  2010-06-14 15:35:25 PST ---
(In reply to comment #8)
> (In reply to comment #7)
> > My broad conclusion is that NSApplication pretending to be in normal loop mode when its not, is interacting badly with the cleanup code from runModalForWindow.
> 
> That seems probably wrong to me.

Possible...

> > Its a whole another thing if you think we shouldn't be able to close showModalDialog from script
> 
> Does closing the dialog work in Internet Explorer? Since showModalDialog originates there, I’d like to match its behavior if possible.

I just tested this on IE - closing showModalDialog from script works on IE.  Interestingly closing showModalDialog from script also works on Safari.  Seems like Safari does not go through WebChromeClient::closeWindowSoon() when closing modal dialogs.  So its just applications that rely on WebChromeClient that are not able to close modal dialogs from script.

> > My plan B is to add abort method to LayoutTestController and invoke that from my test script.
> 
> Our decisions about the web platform should not be based on the needs of DumpRenderTree, so whether we can close a showModalDialog without using the test controller is something we should decide based on what we want on the web, not the needs of our test.

Agreed.  Its just that this seemed like a bug in WebKit which is why I tried to "fix" it.  If we decide that the current behavior (i.e preventing closing showModalDialog window from script), then I'll update my patch to add abortModal to LayoutTestController (which is test code) and leave the WebKit platform code untouched.

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


More information about the webkit-unassigned mailing list