Use deferred style recalc for non-blocking stylesheets.

Use deferred style recalc for non-blocking stylesheets.

Non-blocking stylesheets like alternate stylesheets cause a synchronous style
recalc when finished loading. This is quite the opposite of what one would
expect as these stylesheets should not contribute to the style at all. We use
deferred style recalc to batch up the style recalcs instead.

It might be possible to further improve this by not doing a style recalc at
all, but that requires that we detect when a non-blocking stylesheet becomes
relevant by another link element setting the preferred stylesheet set to the
same name as that of an alternate stylesheet already added as non-blocking.

R=dglazkov@chromium.org
BUG=368659
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=173012

[ojan, 2014-04-30 18:27:11]

lgtm

I think nearly all our uses of RecalcStyleImmediately are wrong and are a
leftover from old code that we haven't gotten around to fixing yet. It'd be
awesome if we could get rid of RecalcStyleImmediately entirely, but I'm not sure
if that's really possible.

[commit-bot: I haz the power, 2014-04-30 18:28:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rune@opera.com/260713005/1

[commit-bot: I haz the power, 2014-04-30 18:36:21]

Change committed as 173012

[rune, 2014-04-30 19:40:29]

On 2014/04/30 18:27:11, ojan wrote:
> lgtm
> 
> I think nearly all our uses of RecalcStyleImmediately are wrong and are a
> leftover from old code that we haven't gotten around to fixing yet. It'd be
> awesome if we could get rid of RecalcStyleImmediately entirely, but I'm not
sure
> if that's really possible.

Yes, I think so too. I thought it would be good to understand/remove on a
case-by-case basis, and then finally get rid of the enum.

[ojan, 2014-04-30 19:41:25]

> Yes, I think so too. I thought it would be good to understand/remove on a
> case-by-case basis, and then finally get rid of the enum.

Totally.