Chromium bug: crbug.com/77172 Description: The attached testcase contains a plugin and a textarea embedded in an iframe. Open the testcase in Chromium, click on the plugin, then click back in the textarea and try to type some text. Expected result: You should be able to type into the textarea. Actual result: You can't type into the textarea, even though the focus ring is draw around it.
Created attachment 89354 [details] testcase
Created attachment 89397 [details] fix Uses FocusController to handle the focus change, just as would happen for a non-plugin element (see the debugging in the Chromium bug for more details on why).
Ping. Can one of you take a look at this? It's a small patch, and the issue it fixes is pretty nasty.
Comment on attachment 89397 [details] fix How do we test this?
If this is only reproducible in the full Chromium browser, we could/should still land a manual-test.
Comment on attachment 89397 [details] fix View in context: https://bugs.webkit.org/attachment.cgi?id=89397&action=review Can't we use eventSender to test this? > Source/WebKit/chromium/ChangeLog:4 > + You forgot bug title & bug url.
(In reply to comment #4) > How do we test this? (In reply to comment #6) > Can't we use eventSender to test this? I have very little experience with WebKit development; I'll try to figure this out myself I guess (the docs seem pretty sparse), but pointers would be a lot more helpful than rhetorical questions I don't know the answers to. (In reply to comment #6) > You forgot bug title & bug url. The bug link is there. I wasn't aware the title was necessary; I'll add it. (Perhaps the prepare-ChangeLog script should update the template if people are expected to include it, since I don't see any mention of it in either the template or the "how to make a patch" documentation.)
(In reply to comment #7) > I have very little experience with WebKit development; I'll try to figure this out myself I guess (the docs seem pretty sparse), but pointers would be a lot more helpful than rhetorical questions I don't know the answers to. EventSender allows you to send arbitrary mouse events / keyboard events in layout tests. I don't think there's any documentation about eventSender but you can see a couple of examples if you grep on LayoutTests. > (In reply to comment #6) > > You forgot bug title & bug url. > > The bug link is there. I wasn't aware the title was necessary; I'll add it. (Perhaps the prepare-ChangeLog script should update the template if people are expected to include it, since I don't see any mention of it in either the template or the "how to make a patch" documentation.) You're supposed to put the bug title & bug url in the consecutive lines and then have a blank line before describing what you're fixing.
Created attachment 90114 [details] fix v2 Test case added, based on looking at some of the surrounding test--thanks for the pointers. It runs as both a layout test and an interactive test (although the two modes work differently since I can't simulate clicks in normal mode).
Comment on attachment 90114 [details] fix v2 Adding cq+ since stuart isn't a committer.
The commit-queue encountered the following flaky tests while processing attachment 90114 [details]: http/tests/xmlhttprequest/cross-origin-no-credential-prompt.html bug 52249 (author: ap@webkit.org) The commit-queue is continuing to process your patch.
Comment on attachment 90114 [details] fix v2 Clearing flags on attachment: 90114 Committed r84270: <http://trac.webkit.org/changeset/84270>
All reviewed patches have been landed. Closing bug.
The commit-queue encountered the following flaky tests while processing attachment 90114 [details]: http/tests/misc/favicon-loads-with-icon-loading-override.html bug 58412 (author: alice.liu@apple.com) The commit-queue is continuing to process your patch.
http://trac.webkit.org/changeset/84270 might have broken SnowLeopard Intel Release (Tests) The following tests are not passing: plugins/mouse-click-iframe-to-plugin.html
Am I missing something in the test structure? It's all run in one pass, no waiting for callbacks, so I didn't think I needed notifyDone based on looking at other cases. It ran fine for me locally on Snow Leopard.
I skipped the test in http://trac.webkit.org/changeset/84279 for WebKit2 but it doesn't seem to be the right fix then. Please re-enable the test on WebKit2 when you fixed the test.
Faling on Windows 7: http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r84283%20(11851)/results.html
Updated Skipped files in http://trac.webkit.org/changeset/84287. The bug for tracking the mac/win failure is https://bugs.webkit.org/show_bug.cgi?id=58924