Use style invalidation for more pseudo classes.

Use style invalidation for more pseudo classes.

Support descendant invalidation sets for :checked, :disabled, :enabled,
:indeterminate, :target, :link, and :visited. Pseudo state changes for those
classes were recalculated with SubtreeStyleChange in didAffectSelector.

Instead, schedule invalidation sets when these pseudo states change.

This change does not fully fix 388965 as there are still several
SubtreeStyleChange instances left.

The pseudo flags could be removed from the SelectRuleFeatureSet since they
will now be present through invalidation sets in the m_cssRuleFeatureSet.

R=esprehn@chromium.org
BUG=388965
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=182507

[rune, 2014-09-10 08:13:18]

Rebased, ptal.

[rune, 2014-09-16 18:52:53]

ping

[esprehn, 2014-09-19 04:55:18]

lgtm, this is going to make us call shadowWhereNodeCanBeDistributed more, but we
shouldn't optimize that until we see it on a profile. The old system was built
to avoid that, but I'm not sure it was really needed.

https://codereview.chromium.org/540533004/diff/20001/LayoutTests/fast/css/inv...
File LayoutTests/fast/css/invalidation/link-pseudo.html (right):

https://codereview.chromium.org/540533004/diff/20001/LayoutTests/fast/css/inv...
LayoutTests/fast/css/invalidation/link-pseudo.html:35: link.offsetTop; // Force
recalc.
I often use getComputedStyle(link).color; to avoid also causing a layout so your
test is more targeted.

https://codereview.chromium.org/540533004/diff/20001/Source/core/css/RuleFeat...
File Source/core/css/RuleFeature.h (right):

https://codereview.chromium.org/540533004/diff/20001/Source/core/css/RuleFeat...
Source/core/css/RuleFeature.h:87: inline bool
hasSelectorForPseudoType(CSSSelector::PseudoType pseudo) const
remove inline, this in the class def so it's inline already. I might also make
this out of line since there's only one caller.

[rune, 2014-09-23 12:23:16]

https://codereview.chromium.org/540533004/diff/20001/LayoutTests/fast/css/inv...
File LayoutTests/fast/css/invalidation/link-pseudo.html (right):

https://codereview.chromium.org/540533004/diff/20001/LayoutTests/fast/css/inv...
LayoutTests/fast/css/invalidation/link-pseudo.html:35: link.offsetTop; // Force
recalc.
On 2014/09/19 04:55:17, esprehn wrote:
> I often use getComputedStyle(link).color; to avoid also causing a layout so
your
> test is more targeted.

getComputedStyle doesn't recalc/layout enough to make the updateStyleAnd...
below be unaffected by the shouldBe() output above.

https://codereview.chromium.org/540533004/diff/20001/Source/core/css/RuleFeat...
File Source/core/css/RuleFeature.h (right):

https://codereview.chromium.org/540533004/diff/20001/Source/core/css/RuleFeat...
Source/core/css/RuleFeature.h:87: inline bool
hasSelectorForPseudoType(CSSSelector::PseudoType pseudo) const
On 2014/09/19 04:55:17, esprehn wrote:
> remove inline, this in the class def so it's inline already. I might also make
> this out of line since there's only one caller.

Done.

[rune, 2014-09-23 12:23:28]

The CQ bit was checked by rune@opera.com

[commit-bot: I haz the power, 2014-09-23 12:24:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/540533004/40001

[commit-bot: I haz the power, 2014-09-23 12:24:16]

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

[commit-bot: I haz the power, 2014-09-23 12:24:17]

Failed to apply patch for Source/core/css/RuleFeature.cpp:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file Source/core/css/RuleFeature.cpp
  Hunk #1 FAILED at 212.
  1 out of 1 hunk FAILED -- saving rejects to file
Source/core/css/RuleFeature.cpp.rej

Patch:       Source/core/css/RuleFeature.cpp
Index: Source/core/css/RuleFeature.cpp
diff --git a/Source/core/css/RuleFeature.cpp b/Source/core/css/RuleFeature.cpp
index
23e821930aa27db742e350dcc2bac6983686346a..67f68d5844550d6af7d4f28ed67c1c2f08cdbe1a
100644
--- a/Source/core/css/RuleFeature.cpp
+++ b/Source/core/css/RuleFeature.cpp
@@ -212,9 +212,21 @@ DescendantInvalidationSet*
RuleFeatureSet::invalidationSetForSelector(const CSSS
     if (selector.match() == CSSSelector::Id)
         return &ensureIdInvalidationSet(selector.value());
     if (selector.match() == CSSSelector::PseudoClass) {
-        CSSSelector::PseudoType pseudo = selector.pseudoType();
-        if (pseudo == CSSSelector::PseudoHover || pseudo ==
CSSSelector::PseudoActive || pseudo == CSSSelector::PseudoFocus)
-            return &ensurePseudoInvalidationSet(pseudo);
+        switch (selector.pseudoType()) {
+        case CSSSelector::PseudoHover:
+        case CSSSelector::PseudoActive:
+        case CSSSelector::PseudoFocus:
+        case CSSSelector::PseudoChecked:
+        case CSSSelector::PseudoEnabled:
+        case CSSSelector::PseudoDisabled:
+        case CSSSelector::PseudoIndeterminate:
+        case CSSSelector::PseudoLink:
+        case CSSSelector::PseudoTarget:
+        case CSSSelector::PseudoVisited:
+            return &ensurePseudoInvalidationSet(selector.pseudoType());
+        default:
+            break;
+        }
     }
     return 0;
 }

[rune, 2014-09-23 12:41:36]

The CQ bit was checked by rune@opera.com

[commit-bot: I haz the power, 2014-09-23 12:42:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/540533004/60001

[commit-bot: I haz the power, 2014-09-23 14:49:21]

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

[commit-bot: I haz the power, 2014-09-23 14:49:23]

Try jobs failed on following builders:
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/28371)

[rune, 2014-09-23 16:03:08]

The CQ bit was checked by rune@opera.com

[commit-bot: I haz the power, 2014-09-23 16:04:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/540533004/60001

[commit-bot: I haz the power, 2014-09-23 16:48:42]

Committed patchset #4 (id:60001) as 182507