Bug 31003 – Using dataTransfer format = "text" not HTML 5 compliant in Safari, but is in Chrome

• Status: NEW
• Component: WebKit Misc.

[Daniel Bates]

Section 7.9.2 of the HTML 5 spec <http://dev.w3.org/html5/spec/Overview.html#datatransfer> states: "If format is the value 'Text', then it must be treated as 'text/plain'".

For reasons that are unclear at the moment, the Chrome build correctly implements this, but the Safari build does not.

Safari allows format 'text', but not 'Text'.

Chome allows format 'Text', but not 'text'.

By the spec, only 'Text' is valid.

[Daniel Bates]

Created attachment 42304 [details]
Layout test

Modified the layout test for bug #30266 to show this bug.

[Daniel Bates]

Created attachment 42305 [details]
Self-contained test

For convenience, here is a self-contained version of the layout test.

[Eric Seidel (no email)]

What does IE allow?  "Text" came from IE.  If IE allows "text" as well, then we should allow both.  If IE does not allow "text" then we should match the spec and IE I guess.

[Oliver Hunt]

ClipboardWin.cpp is looking for text rather than Text, mac safari does a compare to Text so this appears to be a win only bug

[Daniel Bates]

Oliver, have you tried running the self-contained test <https://bugs.webkit.org/attachment.cgi?id=42305>?

I just ran it using r50423 and it fails. I haven't tried running the test with the ToT (r50425), but the latest commits don't appear to address this issue.

(In reply to comment #4)
> ClipboardWin.cpp is looking for text rather than Text, mac safari does a
> compare to Text so this appears to be a win only bug

[Daniel Bates]

I mean to say I ran the test with r50423 on my Mac. So, this is also a Mac issue. I haven't yet looked at the code with respect to this bug.

[Ahmad Saleem]

Using 'self-contained' test, all browsers (STP174, Chrome Canary 117 & Firefox Nightly 117) show:

FAIL event.dataTransfer.types should contain format "Text", but it does not.