Don't force recalc if there are no shadow boundary crossing rules

Don't force recalc if there are no shadow boundary crossing rules

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=164382

[esprehn, 2013-12-20 19:11:42]

lgtm

https://codereview.chromium.org/119693002/diff/20001/Source/core/dom/Element.cpp
File Source/core/dom/Element.cpp (right):

https://codereview.chromium.org/119693002/diff/20001/Source/core/dom/Element....
Source/core/dom/Element.cpp:1624: return
document().ensureStyleResolver().treeBoundaryCrossingRules().isEmpty() ? Inherit
: Force;
I'm not sure if this will work with adding <style> elements to ShadowRoots since
sorvell said that a <style> in one ShadowRoot should apply to all ShadowRoots.
Does that work yet? We might need to do setNeedsStyleRecalc on each ShadowRoot
in that case.

It might be nice to add a method in StyleResolver.h like
hasTreeBoundaryCrossingRules() const { return ... } to hide the isEmpty() check
since we might not expose that API. We can do that in the future if you don't
want to mess with it in this patch.

[esprehn, 2013-12-20 19:34:32]

Actually I think this breaks dynamically inserting style sheets in
ShadowTreeStyleSheetCollection::updateActiveStyleSheets where we do
host()->setNeedsStyleRecalc(). You should try doing
shadowRoot.appendChild(style) in script after doing offsetTop and such, I think
you broke some things.

[ykyyip, 2013-12-30 00:49:53]

On 2013/12/20 19:34:32, esprehn wrote:
> Actually I think this breaks dynamically inserting style sheets in
> ShadowTreeStyleSheetCollection::updateActiveStyleSheets where we do
> host()->setNeedsStyleRecalc(). You should try doing
> shadowRoot.appendChild(style) in script after doing offsetTop and such, I
think
> you broke some things.

It seems like when you dynamically add a style sheet, it'll set
childNeedsStyleRecalc on the shadow root and not hit the early return in
ShadowRoot::recalcStyle. Is there another case I'm missing?

[commit-bot: I haz the power, 2013-12-30 17:48:11]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ykyyip@chromium.org/119693002/20001

[esprehn, 2013-12-30 17:48:19]

On 2013/12/30 00:49:53, ykyyip wrote:
> On 2013/12/20 19:34:32, esprehn wrote:
> > Actually I think this breaks dynamically inserting style sheets in
> > ShadowTreeStyleSheetCollection::updateActiveStyleSheets where we do
> > host()->setNeedsStyleRecalc(). You should try doing
> > shadowRoot.appendChild(style) in script after doing offsetTop and such, I
> think
> > you broke some things.
> 
> It seems like when you dynamically add a style sheet, it'll set
> childNeedsStyleRecalc on the shadow root and not hit the early return in
> ShadowRoot::recalcStyle. Is there another case I'm missing?

Where is that code? I'm pretty sure this is going to break something, but lets
land it and see what happens.

[commit-bot: I haz the power, 2013-12-30 19:02:55]

Change committed as 164382