Make Window.prototype.toString return [object Window] at all times.

Make Window.prototype.toString return [object Window] at all times.

The old code was incorrectly falling back to
Object.prototype.toString.call which does not work as expected without
full support for @@toSringTag

BUG=239915
R=adamk, jsbell

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=195193

[arv (Not doing code reviews), 2015-05-08 21:36:53]

PTAL

[arv (Not doing code reviews), 2015-05-08 21:42:24]

https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
File LayoutTests/http/tests/security/cross-frame-window-to-tostring-expected.txt
(right):

https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
LayoutTests/http/tests/security/cross-frame-window-to-tostring-expected.txt:7:
FAIL Object.prototype.toString.call(window) should be [object Window]. Was
[object global].
This has a bunch of FAIL in it. I think I got the expectations and they should
pass when we add support for @@toStringTag

https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
File LayoutTests/http/tests/security/cross-frame-window-to-tostring.html
(right):

https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
LayoutTests/http/tests/security/cross-frame-window-to-tostring.html:15:
shouldBe("Object.prototype.toString.call(window)", "'[object Window]'");
WebIDL does not mention `[object WindowPrototype]` any more so I'm sticking to
this for now.

[adamk, 2015-05-08 22:03:04]

https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
File LayoutTests/http/tests/security/cross-frame-window-to-tostring.html
(right):

https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
LayoutTests/http/tests/security/cross-frame-window-to-tostring.html:15:
shouldBe("Object.prototype.toString.call(window)", "'[object Window]'");
On 2015/05/08 21:42:24, arv wrote:
> WebIDL does not mention `[object WindowPrototype]` any more so I'm sticking to
> this for now.

Oh huh, we currently return [object Object] for this...

Why is this different in WebIDL from everything else? I see the following text
at the end of http://heycam.github.io/webidl/#interface-prototype-object :

The class string of an interface prototype object is the concatenation of the
interface’s identifier and the string “Prototype”.

[arv (Not doing code reviews), 2015-05-08 22:08:42]

On 2015/05/08 22:03:04, adamk wrote:
>
https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
> File LayoutTests/http/tests/security/cross-frame-window-to-tostring.html
> (right):
> 
>
https://codereview.chromium.org/1138583002/diff/1/LayoutTests/http/tests/secu...
> LayoutTests/http/tests/security/cross-frame-window-to-tostring.html:15:
> shouldBe("Object.prototype.toString.call(window)", "'[object Window]'");
> On 2015/05/08 21:42:24, arv wrote:
> > WebIDL does not mention `[object WindowPrototype]` any more so I'm sticking
to
> > this for now.
> 
> Oh huh, we currently return [object Object] for this...
> 
> Why is this different in WebIDL from everything else? I see the following text
> at the end of http://heycam.github.io/webidl/#interface-prototype-object :
> 
> The class string of an interface prototype object is the concatenation of the
> interface’s identifier and the string “Prototype”.

You are right. I'm not sure why I couldn't find that... maybe I was looking at a
different version.

[adamk, 2015-05-09 00:59:22]

Looks like some real failures in event-view-toString?

[arv (Not doing code reviews), 2015-05-09 01:17:00]

On 2015/05/09 00:59:22, adamk wrote:
> Looks like some real failures in event-view-toString?

I was counting on the try bot to find the test to update ;-)

[arv (Not doing code reviews), 2015-05-11 14:45:14]

Use WindowPrototype in tests and fix a test failure

[arv (Not doing code reviews), 2015-05-11 14:46:58]

Fix one test line

[arv (Not doing code reviews), 2015-05-11 15:11:33]

PTAL

[adamk, 2015-05-11 15:46:40]

lgtm

[jsbell, 2015-05-11 16:18:09]

lgtm

https://codereview.chromium.org/1138583002/diff/40001/LayoutTests/http/tests/...
File LayoutTests/http/tests/security/cross-frame-window-to-tostring.html
(right):

https://codereview.chromium.org/1138583002/diff/40001/LayoutTests/http/tests/...
LayoutTests/http/tests/security/cross-frame-window-to-tostring.html:13:
shouldBe("window + ''", "'[object Window]'");
These could use shouldBeEqualToString to avoid double-quoting the 2nd arg.

[arv (Not doing code reviews), 2015-05-11 16:42:55]

On 2015/05/11 16:18:09, jsbell wrote:
> lgtm
> 
>
https://codereview.chromium.org/1138583002/diff/40001/LayoutTests/http/tests/...
> File LayoutTests/http/tests/security/cross-frame-window-to-tostring.html
> (right):
> 
>
https://codereview.chromium.org/1138583002/diff/40001/LayoutTests/http/tests/...
> LayoutTests/http/tests/security/cross-frame-window-to-tostring.html:13:
> shouldBe("window + ''", "'[object Window]'");
> These could use shouldBeEqualToString to avoid double-quoting the 2nd arg.

Yup. I initially used shouldBe from a different file (I wonder how many copies
there are in LayoutTests) and that file had no shouldBeString. Turned out that
that file overrode window.toString so I switched back to the standard one.

Updatin the test.

[arv (Not doing code reviews), 2015-05-11 16:45:38]

shouldBeEqualToString

[arv (Not doing code reviews), 2015-05-11 16:46:22]

Joshua, Thanks for the review.

CQ'ing...

[arv (Not doing code reviews), 2015-05-11 16:46:27]

The CQ bit was checked by arv@chromium.org

[arv (Not doing code reviews), 2015-05-11 16:46:27]

The patchset sent to the CQ was uploaded after l-g-t-m from jsbell@chromium.org,
adamk@chromium.org
Link to the patchset: https://codereview.chromium.org/1138583002/#ps60001
(title: "shouldBeEqualToString")

[commit-bot: I haz the power, 2015-05-11 16:47:34]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1138583002/60001

[commit-bot: I haz the power, 2015-05-11 18:26:51]

Committed patchset #4 (id:60001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=195193

[arv (Not doing code reviews), 2015-06-02 21:54:38]

A revert of this CL (patchset #4 id:60001) has been created in
https://codereview.chromium.org/1163943002/ by arv@chromium.org.

The reason for reverting is: Going to disable --harmony-tostring until our
bindings have been updated..