Relax HTMLBodyElement's scrollTop/Left property set and get operations.

Relax HTMLBodyElement's scrollTop/Left property set and get operations.

After bug 303131 (r159035) and bug 157855 (r158864) are fixed, Blink started to
match other Web engines with regards to documentElement/body::scrollTop/Left
behaviors, according to the DocType mode. It turns out that these changes broken
some Web content relying on Blink quirks behavior.

That said, this CL takes a more conservative approach, where it was chosen to relax
the HTMLBodyElement bits of the CLs above, restoring their original behavior, but adding
an UseCount/DeprecationWarnings instead.

We are still more compliant with other Web engine with regards to documentElement's
scrollTop/Left operations, though.

R=Julien Chaffraix

BUG=304768, 304816, 304773, 305099
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=159230

[Julien - ping for review, 2013-10-08 18:07:14]

lgtm

https://codereview.chromium.org/26489005/diff/1/Source/core/html/HTMLBodyElem...
File Source/core/html/HTMLBodyElement.cpp (right):

https://codereview.chromium.org/26489005/diff/1/Source/core/html/HTMLBodyElem...
Source/core/html/HTMLBodyElement.cpp:258: // FIXME: The specification is not
clear about what is the expected behavior here:
The specification is clear now no?

Maybe we should keep a new FIXME about matching other browsers.

[mike.sherov, 2013-10-08 23:37:14]

On 2013/10/08 18:07:14, Julien Chaffraix wrote:
> lgtm
> 
>
https://codereview.chromium.org/26489005/diff/1/Source/core/html/HTMLBodyElem...
> File Source/core/html/HTMLBodyElement.cpp (right):
> 
>
https://codereview.chromium.org/26489005/diff/1/Source/core/html/HTMLBodyElem...
> Source/core/html/HTMLBodyElement.cpp:258: // FIXME: The specification is not
> clear about what is the expected behavior here:
> The specification is clear now no?
> 
> Maybe we should keep a new FIXME about matching other browsers.

Yeah, it seems as if the spec is slightly ambiguous: if you read the setter
steps carefully: http://dev.w3.org/csswg/cssom-view/#dom-element-scrolltop, it
doesn't forbid the body element from setting a value, per se. Although that
might be an oversight...
http://dev.w3.org/csswg/cssom-view/#dom-element-scrolltop

What does Firefox and IE actually do here? Also, this might be worth posting to
www-style about for clarification.

[commit-bot: I haz the power, 2013-10-09 11:48:39]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/a1.gomes@sisa.samsung.com/26489005/9001

[mike.sherov, 2013-10-09 12:09:44]

On 2013/10/08 23:37:14, mike.sherov wrote:
> On 2013/10/08 18:07:14, Julien Chaffraix wrote:
> > lgtm
> > 
> >
>
https://codereview.chromium.org/26489005/diff/1/Source/core/html/HTMLBodyElem...
> > File Source/core/html/HTMLBodyElement.cpp (right):
> > 
> >
>
https://codereview.chromium.org/26489005/diff/1/Source/core/html/HTMLBodyElem...
> > Source/core/html/HTMLBodyElement.cpp:258: // FIXME: The specification is not
> > clear about what is the expected behavior here:
> > The specification is clear now no?
> > 
> > Maybe we should keep a new FIXME about matching other browsers.
> 
> Yeah, it seems as if the spec is slightly ambiguous: if you read the setter
> steps carefully: http://dev.w3.org/csswg/cssom-view/#dom-element-scrolltop, it
> doesn't forbid the body element from setting a value, per se. Although that
> might be an oversight...
> http://dev.w3.org/csswg/cssom-view/#dom-element-scrolltop
> 
> What does Firefox and IE actually do here? Also, this might be worth posting
to
> www-style about for clarification.

I brought this up on the www-style ML, and I think this clears it up: 
http://lists.w3.org/Archives/Public/www-style/2013Oct/0287.html

In standards mode, document.body.scrollTop can return >0, but only the element's
scroll value itself. This is different from returning the viewport's (window)
scrollY. I think this patch should be amended.

[commit-bot: I haz the power, 2013-10-09 13:26:58]

Change committed as 159230

[chrisgao (Use stgao instead), 2013-10-11 18:23:06]

Hi,

The change with regards to documentElement/body::scrollTop/Left
behaviors will have side effect on ChromeDriver.

Next time when there is another change, it is appreciated if you could post a
comment on this bug https://code.google.com/p/chromedriver/issues/detail?id=569.
It will save us some time on figuring out what's broken.

Thanks,
Shuotao Gao

[tonikitoo_, 2013-10-11 18:28:49]

On 2013/10/11 18:23:06, chrisgao wrote:
> Hi,
> 
> The change with regards to documentElement/body::scrollTop/Left
> behaviors will have side effect on ChromeDriver.
> 
> Next time when there is another change, it is appreciated if you could post a
> comment on this bug
https://code.google.com/p/chromedriver/issues/detail?id=569.
> It will save us some time on figuring out what's broken.
> 
> Thanks,
> Shuotao Gao

Sure thing. Behavior of body.scrolltop/left should stay as it was, in the
meanwhile, so any breakage should have been gone by now.

I will also notify blink-dev about the plan here, and can I also notify the bug
you mentioned. Doing it shortly.

[Julien - ping for review, 2013-10-11 19:51:24]

@chrisgao: Why don't you just CC the people involved on this bug? Without
knowing about the breakage, it's hard for anyone to keep you updated. Also such
bugs are good data points when we assess the impact of a change.

[chrisgao (Use stgao instead), 2013-10-11 19:54:32]

Thanks for the reply and suggestion!
Will do the cc immediately.