CSS: Reduce style invalidation with sibling combinators

CSS: Reduce style invalidation with sibling combinators

Suppose we have a style rule
    .m + .n .o + .p { .. }

RuleFeatureSet::extractInvalidationSetFeatures is called more than once,
to find the rightmost compound selector of each sibling sequence:
  .p (the overall rightmost compound selector)
  .n (the rightmost compound selector of .m + .n)

If class n is added to, or removed from, an element, we don't want to
needlessly force that element's style to be recalculated.

For more context, see the Design doc https://goo.gl/z0Z9gn

BUG=378135
Committed: https://crrev.com/93e66016b67ce85ac8d7cbf3ad52b0e123ffa779
Cr-Commit-Position: refs/heads/master@{#356651}

[Eric Willigers, 2015-10-27 07:03:37]

ericwilligers@chromium.org changed reviewers:
+ esprehn@chromium.org, rune@opera.com, timloh@chromium.org

[rune, 2015-10-27 09:05:28]

https://codereview.chromium.org/1418133013/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/css/RuleFeature.cpp (right):

https://codereview.chromium.org/1418133013/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/css/RuleFeature.cpp:322: auto result =
extractInvalidationSetFeatures(ruleData.selector(), features, false, true);
Would increase readability to use enums instead of bool for the last two params.

Dunno if it would make sense to have a single parameter ExtractionFlags?

[Eric Willigers, 2015-10-28 00:13:07]

https://codereview.chromium.org/1418133013/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/css/RuleFeature.cpp (right):

https://codereview.chromium.org/1418133013/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/css/RuleFeature.cpp:322: auto result =
extractInvalidationSetFeatures(ruleData.selector(), features, false, true);
On 2015/10/27 09:05:28, rune wrote:
> Would increase readability to use enums instead of bool for the last two
params.
> 
> Dunno if it would make sense to have a single parameter ExtractionFlags?

I changed the code to use two enums.

[rune, 2015-10-28 14:32:20]

lgtm

[Eric Willigers, 2015-10-28 20:09:09]

The CQ bit was checked by ericwilligers@chromium.org

[commit-bot: I haz the power, 2015-10-28 20:09:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1418133013/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1418133013/20001

[commit-bot: I haz the power, 2015-10-28 21:20:35]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-10-28 21:21:34]

Patchset 2 (id:??) landed as
https://crrev.com/93e66016b67ce85ac8d7cbf3ad52b0e123ffa779
Cr-Commit-Position: refs/heads/master@{#356651}