Re-attach plugin renderers in the standard way.

Re-attach plugin renderers in the standard way.

When embedded content changes its source or type, it potential needs to
change its renderer. That was done by reattaching the renderer using a
combination of SubtreeStyleChange plus an explicit reattach in a custom
willRecalcStyle callback. This CL uses a standard lazy reattach instead.

Three svg tests changed their expected repaint rectangles due to the
following reason:

The reattach in willRecalcStyle did not re-attach any whitespace siblings
like the standard reattach does (certainly wrong, but I don't know if it
actually is observable as a display property change is necessary to affect
the whitespace renderers afaik). Moving to standard reattach makes
reattachment of a whitespace sibling do paint invalidation for the closest
block ancestor.

It's possible that we can improve whitespace re-attachment by:

1. Not invalidate block containers for whitespace renderers (will they ever
   render anything?)

2. Detect if we need to reattach whitespace siblings before we do the actual
   reattachment.

R=esprehn@chromim.org,leviw@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=182868

[rune, 2014-09-23 12:09:33]

PTAL

[rune, 2014-09-29 17:25:05]

pling

[leviw_travelin_and_unemployed, 2014-09-29 18:55:26]

Great. LGTM.

[rune, 2014-09-29 19:35:51]

The CQ bit was checked by rune@opera.com

[commit-bot: I haz the power, 2014-09-29 19:36:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/596773002/1

[commit-bot: I haz the power, 2014-09-29 19:37:03]

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

[commit-bot: I haz the power, 2014-09-29 19:37:03]

Failed to apply patch for Source/core/html/HTMLObjectElement.cpp:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file Source/core/html/HTMLObjectElement.cpp
  Hunk #1 FAILED at 256.
  Hunk #2 FAILED at 331.
  2 out of 2 hunks FAILED -- saving rejects to file
Source/core/html/HTMLObjectElement.cpp.rej

Patch:       Source/core/html/HTMLObjectElement.cpp
Index: Source/core/html/HTMLObjectElement.cpp
diff --git a/Source/core/html/HTMLObjectElement.cpp
b/Source/core/html/HTMLObjectElement.cpp
index
3d588cdd406fedfea8d5204e2d01b0c788e2cd72..721e3f1cc308ebecb8c3f1b8d1c93f504ad2fad6
100644
--- a/Source/core/html/HTMLObjectElement.cpp
+++ b/Source/core/html/HTMLObjectElement.cpp
@@ -256,7 +256,7 @@ void HTMLObjectElement::reloadPluginOnAttributeChange(const
QualifiedName& name)
     }
     setNeedsWidgetUpdate(true);
     if (needsInvalidation)
-        setNeedsStyleRecalc(SubtreeStyleChange);
+        lazyReattachIfNeeded();
 }
 
 // FIXME: This should be unified with HTMLEmbedElement::updateWidget and
@@ -331,7 +331,7 @@ void HTMLObjectElement::childrenChanged(const
ChildrenChange& change)
 {
     if (inDocument() && !useFallbackContent()) {
         setNeedsWidgetUpdate(true);
-        setNeedsStyleRecalc(SubtreeStyleChange);
+        lazyReattachIfNeeded();
     }
     HTMLPlugInElement::childrenChanged(change);
 }

[rune, 2014-09-29 19:47:08]

The CQ bit was checked by rune@opera.com

[commit-bot: I haz the power, 2014-09-29 19:48:03]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/596773002/20001

[esprehn, 2014-09-29 20:48:46]

esprehn@chromium.org changed reviewers:
+ esprehn@chromium.org

[esprehn, 2014-09-29 20:48:46]

lgtm

This is how I feel about this patch: http://i.imgur.com/EWNQmzd.gif

[commit-bot: I haz the power, 2014-09-29 22:07:14]

Committed patchset #2 (id:20001) as 182868