As best I can tell from looking at the test results, it looks like shift+click is getting treated as a regular click. Skipping the test for now.
Created attachment 55257 [details]
Comment on attachment 55257 [details]
Ok, but please follow up with some of the Qt/Gtk folks by email to make sure this is how they'd like this handled.
Committed r58896: <http://trac.webkit.org/changeset/58896>
My intuition says this is related to the fact that GTK+ uses Mac editing behavior by default, but this test is checking the user-agent for Mac-specific strings.
(In reply to comment #4)
> My intuition says this is related to the fact that GTK+ uses Mac editing
> behavior by default, but this test is checking the user-agent for Mac-specific
As best I can tell, that's not the case here. The test results don't match windows or mac editing behavior. They match the behavior of doing a regular click.
There are some other bugs with GTK+ EventSender around double-clicks. I have one patch for it and will try to get it ready for review tomorrow. I wouldn't be surprised if it's failing both because Qt and GTK+ use the Mac editing style and because of bugs in DRT. I'll try to follow up when I have a moment.
Coming from the email contact. I believe Martin is the GTK+ authority of choice for such issues, so I'll let him handle the case since he's already on it =D
test is still failing on Gtk and Qt. Reopening.
This patch is failing on GTK+ for two reasons:
1. There is no support in the EventSender for mouseUp and mouseDown modifiers.
2. The test uses the user-agent to figure out the editing style.
I'll post patches soon.
From a quick look, it appears that number 1 is also an issue with the Qt DRT.
Still reproducible at r79447
This doesn't seem to be an issue any longer.