Bug 29742 – [Chromium] Change window.toString() callback to match JSC

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Nate Chapin]

The V8 bindings currently access the global object wrapper instead of the DOMWindow wrapper in the toString() callback.  We should be accessing the DOMWindow wrapper, since we currently output [object global] instead of the expected [object DOMWindow].

[Nate Chapin]

Created attachment 40124 [details]
patch

[Eric Seidel (no email)]

You should list what test this affects.  Assuming that it affects a test?  If it doesn't then we need to add a test.

Adam and Dimitri are probably your two best WebKit reviewers here.

[Adam Barth]

Comment on attachment 40124 [details]
patch

This looks great, but please list which LayoutTests this fixes.

[Nate Chapin]

I initially started looking at it for LayoutTests/fast/events/event-trace.html, and from a run of all layout tests, it appears it will fix 2 and allow us to use default
expectations on about a dozen others.  There are some others that we're also
failing for other reasons that will still fail, but will be closer to
expectations.

I'll see if I can assemble a full list.

I can try and compile a full list if that would be helpful.(In reply to comment #3)
> (From update of attachment 40124 [details])
> This looks great, but please list which LayoutTests this fixes.

[Nate Chapin]

The complete list, to the best of my knowledge, is: LayoutTests/fast/dom/Window/atob-btoa.html
LayoutTests/fast/dom/Window/window-lookup-precedence.html
LayoutTests/fast/dom/Window/window-postmessage-args.html
LayoutTests/fast/dom/Window/window-properties.html
LayoutTests/fast/events/event-trace.html
LayoutTests/fast/events/event-view-toString.html
LayoutTests/fast/js/eval-cross-window.html
LayoutTests/fast/js/eval-keyword-vs-function.html
LayoutTests/fast/js/toString-and-valueOf-override.html
LayoutTests/http/tests/security/cross-frame-access-custom.html
LayoutTests/http/tests/security/cross-frame-access-put.html

[Adam Barth]

Comment on attachment 40124 [details]
patch

Awesome!

[Nate Chapin]

Comment on attachment 40124 [details]
patch

cq-

The Chromium WebKit canary is compile-broken at the moment, and I would like to verify precisely how the downstream expectations need to be changed before this gets integrated, so I want to wait on committing.

[Nate Chapin]

Committed as r48841.