Bug 30107 – [Mac] dropEffect is incorrect when effectAllowed = "none"

• Component: WebKit Misc.
• Status: RESOLVED FIXED

[Daniel Bates]

The Mac build of WebKit does not comply with the HTML 5 specification for drag-and-drop with respect to the case when effectAllowed == "none" (*).

Currently, the resulting drop effect in the Mac build is determined by the followings mapping when effectAllowed == "none":


dropEffect:		Actual drop effect:	Expected drop effect:
"none"     -----------> "none"			"none"
"copy"     -----------> "copy"			"none"
"move"     -----------> "move"			"none"
"link"     -----------> "none"			"none"
"dummy"    -----------> "none"			"none"


As per section 7.9.2 of the HTML 5 specification <http://dev.w3.org/html5/spec/Overview.html#the-dragevent-and-datatransfer-interfaces>, regardless of the specified dropEffect, when effectAllowed == "none" the resulting dropEffect should be "none".

(*) It also does not comply with the spec. when effectAllowed == "uninitialized", but we should resolve this in a separate bug (after fixing bug #24731) so that we can bring HTML 5 compliance to both Mac and Windows builds in a platform-independent way.

[Daniel Bates]

Created attachment 40685 [details]
Patch

No test case is included in this patch because it uses the same test case as in the patch for bug #24731.

[Daniel Bates]

Created attachment 40686 [details]
Test case

Here is a test case based off of the test case in the patch for bug #24731. Compare the results under Safari and Firefox.

[Eric Seidel (no email)]

Comment on attachment 40685 [details]
Patch

You should still mention which test case this fixes.  If bug 24731 hasn't yet landed, you might consider landing the test with this change and skipping it on windows for now until bug 24731 can be fixed.

[Eric Seidel (no email)]

Comment on attachment 40685 [details]
Patch

cq- due to my comment above.

[Daniel Bates]

Before I'll land this patch, I'll add this test case to it and modify the file LayoutTest/platform/win/Skipped to skip this test.

(In reply to comment #3)
> (From update of attachment 40685 [details])
> You should still mention which test case this fixes.  If bug 24731 hasn't yet
> landed, you might consider landing the test with this change and skipping it on
> windows for now until bug 24731 can be fixed.

[Daniel Bates]

Created attachment 40735 [details]
Patch with test case

Patch with test case.

[Daniel Bates]

Committed r49203: <http://trac.webkit.org/changeset/49203>