Make user-select:none not to affect editability

This patch makes CSS property |user-select:none| not to affect editability as IE and FireFox. 

However, selection behavior different among browsers:

- Chrome: selection works as |user-select:text|.
- FireFox: selection works as |user-select:text| for text input and textarea. In other elements, selection can be extended by cursor key. Mouse click always set caret to start of text.
- IE: Selection can not be extended neither cursor key nor mouse click.

* editing/selection/4866671.html: Replaced with editing/selection/user-select-none-in-editable.html

* editing/selection/5779984-1.html: Replaced with editing/selection/user-select-none-in-editable.html and user-select-js-property.html, for style.webkitUserSelect property.

* fast/events/standalone-image-drag-to-editable.html: Convert to text test instead of pixel test for updating reference images, since the final image is colored with selection color.

BUG=105876
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=152753

[yosin_UTC9, 2013-06-17 06:50:04]

Could you review this patch?
Thanks in advance.

[yosin_UTC9, 2013-06-19 03:31:41]

Hi, 
Could you review this patch?
Thanks in advance.

It seems Linux and Win bots are dead...

[tkent, 2013-06-19 03:50:40]

> editing/selection/4866671.html, editing/selection/5779984-1.html: Replaced
with editing/selection/user-select-non-ineditable.html 

non-ineditable -> none-in-editable

https://codereview.chromium.org/17225002/diff/12001/LayoutTests/editing/selec...
File LayoutTests/editing/selection/5779984-1.html (left):

https://codereview.chromium.org/17225002/diff/12001/LayoutTests/editing/selec...
LayoutTests/editing/selection/5779984-1.html:9: if (div.style.WebkitUserSelect
== undefined)
A test to check existence of style.WebKitUserSelect is missing in new tests of
this CL. Is it covered by other tests?

https://codereview.chromium.org/17225002/diff/12001/LayoutTests/editing/selec...
LayoutTests/editing/selection/5779984-1.html:11: div.style.WebkitUserSelect =
"none";
Ditto for updating style.WebKitUserSelect.

[yosin_UTC9, 2013-06-19 04:28:46]

On 2013/06/19 03:50:40, tkent wrote:
> > editing/selection/4866671.html, editing/selection/5779984-1.html: Replaced
> with editing/selection/user-select-non-ineditable.html 
> 
> non-ineditable -> none-in-editable
> 
Done.


>
https://codereview.chromium.org/17225002/diff/12001/LayoutTests/editing/selec...
> File LayoutTests/editing/selection/5779984-1.html (left):
> 
>
https://codereview.chromium.org/17225002/diff/12001/LayoutTests/editing/selec...
> LayoutTests/editing/selection/5779984-1.html:9: if (div.style.WebkitUserSelect
> == undefined)
> A test to check existence of style.WebKitUserSelect is missing in new tests of
> this CL. Is it covered by other tests?
> 
>
https://codereview.chromium.org/17225002/diff/12001/LayoutTests/editing/selec...
> LayoutTests/editing/selection/5779984-1.html:11: div.style.WebkitUserSelect =
> "none";
> Ditto for updating style.WebKitUserSelect.

Availability of "WebKitUserSelect" in JS is covered by
webexposed/css-properties-as-js-properties.html.
Setting CSS property from JS is covered by other tests by setting "WebKit*"
properties.

[tkent, 2013-06-19 04:32:20]

> Setting CSS property from JS is covered by other tests by setting "WebKit*"
> properties.

What's the test name?

[yosin_UTC9, 2013-06-19 04:39:19]

On 2013/06/19 04:32:20, tkent wrote:
> > Setting CSS property from JS is covered by other tests by setting "WebKit*"
> > properties.
> 
> What's the test name?
No test which use |style.WebKitUserSelect|.
I think tests for setting style.WebKitXXX are enough, e.g.

  flexbox.style.webkitJustifyContent = "flex-end";
  checkValues(100, 0, 200, 0);

https://code.google.com/p/chromium/codesearch#search/&q=style.WebKit%20file:W...

[tkent, 2013-06-19 05:05:27]

> I think tests for setting style.WebKitXXX are enough, e.g.

I don't think so.

[yosin_UTC9, 2013-06-19 06:41:37]

On 2013/06/19 05:05:27, tkent wrote:
> > I think tests for setting style.WebKitXXX are enough, e.g.
> 
> I don't think so.

Sure. Added user-select-js-property.html for testing WebKitUserSelect.
PTAL

P.S. Linux bot failure doesn't relate to this patch.

[tkent, 2013-06-19 08:03:47]

> Sure. Added user-select-js-property.html for testing WebKitUserSelect.

What we need is a test to verify dynamically-updated style.WebKitUseSelect
affects selection behavior.

[yosin_UTC9, 2013-06-19 09:11:17]

On 2013/06/19 08:03:47, tkent wrote:
> > Sure. Added user-select-js-property.html for testing WebKitUserSelect.
> 
> What we need is a test to verify dynamically-updated style.WebKitUseSelect
> affects selection behavior.
Updated.

PTAL

[tkent, 2013-06-19 21:10:10]

lgtm

[commit-bot: I haz the power, 2013-06-20 01:44:51]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/yosin@chromium.org/17225002/35001

[commit-bot: I haz the power, 2013-06-20 06:07:44]

Change committed as 152753