Recalc style or reattach when adding or removing the audio controls attribute

Recalc style or reattach when adding or removing the audio controls attribute

In label-contains-other-interactive-content.html there is an audio
element which is clicked twice, first without the controls attribute and
then again after adding it.

When there is no controls attribute the element doesn't have a layout
object, so that amounted to clicking at (0,0). Due to the bug addressed
here, adding the controls attribute also didn't cause the layout object
to be created when it should, so the second click was also at (0,0).

These two clicks add up to a double click, triggering a selection in the
page. This is the cause of the test failure in a seemingly unrelated
change: https://codereview.chromium.org/838003005/

Change label-contains-other-interactive-content.html to not (fail to)
click on the audio element with no controls attribute, as it's far from
obvious that clickElement(element) will click (0,0) instead.

BUG=497120

[philipj_slow, 2015-06-05 12:44:46]

The CQ bit was checked by philipj@opera.com to run a CQ dry run

[commit-bot: I haz the power, 2015-06-05 12:44:53]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1170583002/1

[philipj_slow, 2015-06-05 12:48:00]

philipj@opera.com changed reviewers:
+ mstensho@opera.com, rune@opera.com

[philipj_slow, 2015-06-05 12:48:01]

PTAL. I'm actually not at all happy with how HTMLAudioElement::parseAttribute()
looks, is there no better way to say "delete your layout object and create a new
one if needed"?

[commit-bot: I haz the power, 2015-06-05 14:39:25]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-06-05 14:39:25]

Dry run: This issue passed the CQ dry run.

[rune, 2015-06-08 07:45:40]

The HTML spec says:

"The audio element, when it is exposing a user interface, is expected to be
treated as a replaced element about one line high, as wide as is necessary to
expose the user agent's user interface features. When an audio element is not
exposing a user interface, the user agent is expected to force its 'display'
property to compute to 'none', irrespective of CSS rules."

Couldn't we just add a UA rule saying:

audio:not([controls]) { display: none !important }

[philipj_slow, 2015-06-08 11:33:59]

On 2015/06/08 07:45:40, rune wrote:
> The HTML spec says:
> 
> "The audio element, when it is exposing a user interface, is expected to be
> treated as a replaced element about one line high, as wide as is necessary to
> expose the user agent's user interface features. When an audio element is not
> exposing a user interface, the user agent is expected to force its 'display'
> property to compute to 'none', irrespective of CSS rules."
> 
> Couldn't we just add a UA rule saying:
> 
> audio:not([controls]) { display: none !important }

That is excellent, I wanted to do exactly that but didn't think there was a way
to express a negative attribute selector. Live and learn:
https://codereview.chromium.org/1166743004/