Use null for default value of event handler attributes in XMLHttpRequest....

Use null for default value of event handler attributes in XMLHttpRequest.

blogger expects typeof(new XMLHttpRequest().onreadystatuschange) is not 'undefined'.
In other browsers, it is 'object' (onreadystatechange is null).

cf.http://www.whatwg.org/specs/web-apps/current-work/#event-handler-attributes
 An event handler attribute, unless otherwise specified, can either have the value null or be set to a Function object.
 Initially, an event handler attribute must be set to null.

BUG=8536, 10552

[tony, 2009-04-16 17:33:59]

Does this fail in SFX (Safari's JS engine) as well?  If so, we should fix
upstream and add a layout test for it.
http://trac.webkit.org/wiki/Writing%20Layout%20Tests%20for%20DumpRenderTree

If it doesn't fail upstream, we should still write a layout test for it (unless
one already exists).

[Fumitoshi Ukai, 2009-04-16 18:28:46]

Thanks for reviewing.

On 2009/04/16 17:33:59, tony wrote:
> Does this fail in SFX (Safari's JS engine) as well?  If so, we should fix
> upstream and add a layout test for it.
> http://trac.webkit.org/wiki/Writing%20Layout%20Tests%20for%20DumpRenderTree
> 
> If it doesn't fail upstream, we should still write a layout test for it
(unless
> one already exists).

I believe this doesn't fail in Sarai's JS engine. Safari works fine on blogger.

So, should I put a layout test in webkit/data/layout_tests/chrome/fast/js/ ?

[Pam (message me for reviews), 2009-04-16 20:32:15]

This is not in my area of expertise, but it looks correct to me, with an
appropriate test.

- Pam

[Fumitoshi Ukai, 2009-04-16 20:59:01]

Thanks for reviewing.

I've added layout test for this change.
Could you take a look again, please?

Thanks.

On 2009/04/16 20:32:15, Pam wrote:
> This is not in my area of expertise, but it looks correct to me, with an
> appropriate test.
> 
> - Pam

[tony, 2009-04-16 21:25:51]

LGTM.  Can you try to send the test upstream?  We will then get it in a our
daily webkit merges and can remove the copy from layout_tests/chrome.

[Fumitoshi Ukai, 2009-04-16 21:41:54]

On 2009/04/16 21:25:51, tony wrote:
> LGTM.  Can you try to send the test upstream?  We will then get it in a our
> daily webkit merges and can remove the copy from layout_tests/chrome.

Thanks for reviewing.
I'm not committer yet, so could you submit this for me, please?

I'll try to send the test to webkit upstream.
Should I follow this procedure?
http://trac.webkit.org/wiki/Creating%20and%20Submitting%20Layout%20Tests%20an...

[tony, 2009-04-16 21:50:13]

On 2009/04/16 21:41:54, ukai wrote:
> On 2009/04/16 21:25:51, tony wrote:
> I'm not committer yet, so could you submit this for me, please?

Yes, I will commit it for you.

> I'll try to send the test to webkit upstream.
> Should I follow this procedure?
>
http://trac.webkit.org/wiki/Creating%20and%20Submitting%20Layout%20Tests%20an...

Yes, that's right.  Thanks!

[Fumitoshi Ukai, 2009-04-16 21:52:46]

On 2009/04/16 21:50:13, tony wrote:
> On 2009/04/16 21:41:54, ukai wrote:
> > On 2009/04/16 21:25:51, tony wrote:
> > I'm not committer yet, so could you submit this for me, please?
> 
> Yes, I will commit it for you.

Thanks!

> > I'll try to send the test to webkit upstream.
> > Should I follow this procedure?
> >
>
http://trac.webkit.org/wiki/Creating%20and%20Submitting%20Layout%20Tests%20an...
> 
> Yes, that's right.  Thanks!

Got it. Will do.

[Pam (message me for reviews), 2009-04-16 21:59:31]

Please don't add files to LayoutTests/chrome/. We're trying to unfork that
directory. Just send them upstream and wait for them to come back down.

http://codereview.chromium.org/67225/diff/6/10
File webkit/data/layout_tests/chrome/fast/js/resources/js-test-pre.js (right):

http://codereview.chromium.org/67225/diff/6/10#newcode102
Line 102: }
function shouldBeNull() already exists in the real version of this file (as
opposed to our forked copy).

[tony, 2009-04-16 22:02:31]

Ok, I'll only check in the .cpp change.  We should get the layout test in a
future merge.

[tony, 2009-04-17 00:07:18]

Actually, I can't land this for you.  This change is in third_party/WebKit/
which we pull from the upstream svn.webkit.org.  You need to make a bug in
bugs.webkit.org and attach a patch there.  You can include this test case with
that patch.

[Fumitoshi Ukai, 2009-04-17 00:21:51]

On 2009/04/17 00:07:18, tony wrote:
> Actually, I can't land this for you.  This change is in third_party/WebKit/
> which we pull from the upstream svn.webkit.org.  You need to make a bug in
> bugs.webkit.org and attach a patch there.  You can include this test case with
> that patch.

Ok.  I'll send this patch to upstream webkit.
Thanks!

[Fumitoshi Ukai, 2009-04-17 03:14:44]

On 2009/04/17 00:21:51, ukai wrote:
> On 2009/04/17 00:07:18, tony wrote:
> > Actually, I can't land this for you.  This change is in third_party/WebKit/
> > which we pull from the upstream svn.webkit.org.  You need to make a bug in
> > bugs.webkit.org and attach a patch there.  You can include this test case
with
> > that patch.
> 
> Ok.  I'll send this patch to upstream webkit.
> Thanks!

Submitted: https://bugs.webkit.org/show_bug.cgi?id=25257

[Fumitoshi Ukai, 2009-04-24 20:31:04]

On 2009/04/17 03:14:44, ukai wrote:
> On 2009/04/17 00:21:51, ukai wrote:
> > On 2009/04/17 00:07:18, tony wrote:
> > > Actually, I can't land this for you.  This change is in
third_party/WebKit/
> > > which we pull from the upstream svn.webkit.org.  You need to make a bug in
> > > bugs.webkit.org and attach a patch there.  You can include this test case
> with
> > > that patch.
> > 
> > Ok.  I'll send this patch to upstream webkit.
> > Thanks!
> 
> Submitted: https://bugs.webkit.org/show_bug.cgi?id=25257

Landed as: http://trac.webkit.org/changeset/42828