Bug 108284 – Snapshotted plug-in should use shadow root

• Component: Plug-ins
• Status: RESOLVED FIXED

[Dean Jackson]

A snapshotted plug-in should use an internal shadow root, so that it can display a message to the user indicating what is happening. Currently, this is implemented by retrieving an image that is burnt into the Renderer.

As part of this, we'll provide a default styling that could be customized by ports using the extraPlugInsStyleSheet() function. For example, Mac will grab this content from its injected bundle to provide a platform-like UI.

[Radar WebKit Bug Importer]

<rdar://problem/13112495>

[Dean Jackson]

Created attachment 185545 [details]
Patch

[Jon Lee]

Comment on attachment 185545 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185545&action=review

Unofficial r=me with adjustments to style calc calls

> Source/WebCore/html/HTMLPlugInImageElement.cpp:146
> +    if (displayState() == DisplayingSnapshot && document()->page() && document()->page()->settings()->plugInSnapshottingEnabled()) {

Is the setting check required? The displayState should only ever not be Playing if the setting is enabled.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:355
> +        recalcStyle(Force);

Is this second recalcStyle necessary?

[Dean Jackson]

(In reply to comment #3)
> (From update of attachment 185545 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=185545&action=review
> 
> Unofficial r=me with adjustments to style calc calls
> 
> > Source/WebCore/html/HTMLPlugInImageElement.cpp:146
> > +    if (displayState() == DisplayingSnapshot && document()->page() && document()->page()->settings()->plugInSnapshottingEnabled()) {
> 
> Is the setting check required? The displayState should only ever not be Playing if the setting is enabled.
> 
> > Source/WebCore/html/HTMLPlugInImageElement.cpp:355
> > +        recalcStyle(Force);
> 
> Is this second recalcStyle necessary?

Fixed both of these.

[Dean Jackson]

Created attachment 185659 [details]
Patch

[Dean Jackson]

Updated patch after r141066, which changed the way you attach shadow roots and their subtrees.

[Early Warning System Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/16257054

[Jon Lee]

Comment on attachment 185659 [details]
Patch

Need to appropriately remove of startLabelImage() and references to LabelSize from other ports.

[Jon Lee]

(In reply to comment #8)
> (From update of attachment 185659 [details])
> Need to appropriately remove of startLabelImage() and references to LabelSize from other ports.

... which means removing ChromeClient::plugInStartLabelImage(). Or alternatively, don't remove LabelSize.

[WebKit Review Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16218894

[Build Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://queues.webkit.org/results/16250104

[Peter Beverloo (cr-android ews)]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass cr-android-ews (chromium-android):
Output: http://queues.webkit.org/results/16265001

[EFL EWS Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass efl-ews (efl):
Output: http://queues.webkit.org/results/16267005

[Early Warning System Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass qt-wk2-ews (qt):
Output: http://queues.webkit.org/results/16269023

[Build Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/16268051

[Build Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/16278112

[WebKit Review Bot]

Comment on attachment 185659 [details]
Patch

Attachment 185659 [details] did not pass cr-linux-debug-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16269119

[Dimitri Glazkov (Google)]

Comment on attachment 185659 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185659&action=review

> Source/WebCore/html/HTMLPlugInImageElement.cpp:291
> +void HTMLPlugInImageElement::didAddUserAgentShadowRoot(ShadowRoot* root)

Elliott's been refactoring around this. Need to cc him so that he don't stomp over ya here.

[Elliott Sprehn]

Comment on attachment 185659 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185659&action=review

> Source/WebCore/html/HTMLPlugInImageElement.cpp:337
> +    recalcStyle(Force);

Why do you need to manually cause the style recalc or set all these needs properties? This seems really wrong. Just like doing appendChild() in JS, we handle all of that automatically here.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:351
> +        attach();

Adding a shadow root will already cause a reattach. You don't need to do it manually. See: ElementShadow::addShadowRoot

> Source/WebCore/html/HTMLPlugInImageElement.cpp:362
> +    attach();

reattach(). What's the reason for wanting to destroy and recreate the render tree here?

> Source/WebCore/rendering/RenderSnapshottedPlugIn.cpp:171
> +    firstChild()->style()->setWidth(Length(newSize.width(), Fixed));

I think you should be able to do this with just CSS? It doesn't seem you should need to layout the child twice here.

[Dean Jackson]

Created attachment 185810 [details]
Patch

[Early Warning System Bot]

Comment on attachment 185810 [details]
Patch

Attachment 185810 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/16270281

[Elliott Sprehn]

Comment on attachment 185810 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185810&action=review

Looks great, two small things.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:295
> +    RefPtr<Element> shadowContainer = doc->createElement(divTag, ASSERT_NO_EXCEPTION);

It would be nicer to use HTMLDivElement::create(doc) instead here since you wouldn't need that ASSERT_NO_EXCEPTION.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:357
> +    attach();

Lets use reattach()

[WebKit Review Bot]

Comment on attachment 185810 [details]
Patch

Attachment 185810 [details] did not pass cr-linux-debug-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16292035

[WebKit Review Bot]

Comment on attachment 185810 [details]
Patch

Attachment 185810 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16266339

[Dean Jackson]

(In reply to comment #18)
> (From update of attachment 185659 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=185659&action=review
> 
> > Source/WebCore/html/HTMLPlugInImageElement.cpp:291
> > +void HTMLPlugInImageElement::didAddUserAgentShadowRoot(ShadowRoot* root)
> 
> Elliott's been refactoring around this. Need to cc him so that he don't stomp over ya here.

Oh, I thought that was the result of the refactoring :)

[Build Bot]

Comment on attachment 185810 [details]
Patch

Attachment 185810 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://queues.webkit.org/results/16264360

[Dean Jackson]

(In reply to comment #9)
> (In reply to comment #8)
> > (From update of attachment 185659 [details] [details])
> > Need to appropriately remove of startLabelImage() and references to LabelSize from other ports.
> 
> ... which means removing ChromeClient::plugInStartLabelImage(). Or alternatively, don't remove LabelSize.

This is just a side effect of this patch depending on https://bugs.webkit.org/show_bug.cgi?id=108273

I'll run against EWS before landing, once that other one has committed.

[Dean Jackson]

Created attachment 185828 [details]
Patch

[Dean Jackson]

New patch with fixed for comments from Dimitri and Elliot. Sorry, this is going to fail the bots once more and cause bugzilla email :( Like I said, it's expecting https://bugs.webkit.org/show_bug.cgi?id=108273

[WebKit Review Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass cr-linux-debug-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16270314

[Build Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://queues.webkit.org/results/16251495

[Early Warning System Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/16273347

[WebKit Review Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16269363

[Peter Beverloo (cr-android ews)]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass cr-android-ews (chromium-android):
Output: http://queues.webkit.org/results/16270324

[EFL EWS Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass efl-ews (efl):
Output: http://queues.webkit.org/results/16266378

[Build Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/16303061

[Jon Lee]

Comment on attachment 185828 [details]
Patch

Unofficial r=me.

[Elliott Sprehn]

Comment on attachment 185828 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185828&action=review

Unofficial r=me on this. Two nits.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:337
> +    return (childContext.isOnUpperEncapsulationBoundary() && HTMLElement::childShouldCreateRenderer(childContext));

nit: extra parens are not needed.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:343
> +    if (displayState() == DisplayingSnapshot && !userAgentShadowRoot()) {

Why do you both assert about it and check here? Early return is usually considered better as well.

[Early Warning System Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass qt-wk2-ews (qt):
Output: http://queues.webkit.org/results/16297473

[Build Bot]

Comment on attachment 185828 [details]
Patch

Attachment 185828 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/16301707

[Dean Jackson]

(In reply to comment #38)
> (From update of attachment 185828 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=185828&action=review
> 
> Unofficial r=me on this. Two nits.
> 
> > Source/WebCore/html/HTMLPlugInImageElement.cpp:337
> > +    return (childContext.isOnUpperEncapsulationBoundary() && HTMLElement::childShouldCreateRenderer(childContext));
> 
> nit: extra parens are not needed.

Removed.

> 
> > Source/WebCore/html/HTMLPlugInImageElement.cpp:343
> > +    if (displayState() == DisplayingSnapshot && !userAgentShadowRoot()) {
> 
> Why do you both assert about it and check here? Early return is usually considered better as well.

Good point. I've removed the test in the conditional, and swapped it around to an early return.

I won't upload a new patch to avoid the EWS build fail emails :)

[Simon Fraser (smfr)]

Comment on attachment 185828 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185828&action=review

> Source/WebCore/css/plugIns.css:42
> +/*
> + * This is the UA StyleSheet for <object> and <embed> elements.
> + *
> + * Such elements, when snapshotted (paused), will contain a ShadowRoot
> + * with the following structure:
> + *
> + * <object>
> + *   #ShadowRoot
> + *     <div class="snapshot-container">
> + *       <div class="snapshot-overlay"></div> <!-- e.g. for dimming content -->
> + *         <div class="snapshot-label">
> + *           <span class="snapshot-title">Snapshotted Plug-In</span>
> + *           <span class="snapshot-subtitle">Click to restart</span>
> + *         </div>
> + *       </div>
> + *     </div>
> + *
> + */

It's a shame to bog down UA stylesheet parsing with big comments. Or is this stripped out before we convert to const char[] ?

> Source/WebCore/css/plugIns.css:55
> +object::-webkit-snapshotted-plugin-content * {
> +    cursor: hand;
> +    pointer-events: none;
> +    -webkit-user-select: none;
> +}

Is this necessary on all child nodes or just some container?

> Source/WebCore/css/plugIns.css:91
> +    color: #444;

You have used every one of rgba(), keyword and hex values for colors!

> Source/WebCore/html/HTMLPlugInImageElement.cpp:74
> +    , m_snapshotImage(0)

No need to init a RefPtr.

> Source/WebCore/html/HTMLPlugInImageElement.cpp:311
> +        String clientTitleText = document()->page()->chrome()->client()->plugInStartLabelTitle();

Do you need to null-check chrome()->client()?

[Dean Jackson]

Comment on attachment 185828 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=185828&action=review

>> Source/WebCore/css/plugIns.css:42
>> + */
> 
> It's a shame to bog down UA stylesheet parsing with big comments. Or is this stripped out before we convert to const char[] ?

Yes, it's stripped out.

>> Source/WebCore/css/plugIns.css:55
>> +}
> 
> Is this necessary on all child nodes or just some container?

Removed

>> Source/WebCore/css/plugIns.css:91
>> +    color: #444;
> 
> You have used every one of rgba(), keyword and hex values for colors!

Thank you!! :)

All rgb or rgba now.

>> Source/WebCore/html/HTMLPlugInImageElement.cpp:74
>> +    , m_snapshotImage(0)
> 
> No need to init a RefPtr.

Fixed.

>> Source/WebCore/html/HTMLPlugInImageElement.cpp:311
>> +        String clientTitleText = document()->page()->chrome()->client()->plugInStartLabelTitle();
> 
> Do you need to null-check chrome()->client()?

No. No other use of this does.

[Dean Jackson]

Created attachment 187360 [details]
Landing EWS test

Landing Patch for EWS Test

[Early Warning System Bot]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/16428824

[WebKit Review Bot]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass cr-linux-debug-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16445893

[WebKit Review Bot]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/16429840

[Peter Beverloo (cr-android ews)]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass cr-android-ews (chromium-android):
Output: http://queues.webkit.org/results/16432857

[Build Bot]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://queues.webkit.org/results/16435830

[Build Bot]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/16445954

[EFL EWS Bot]

Comment on attachment 187360 [details]
Landing EWS test

Attachment 187360 [details] did not pass efl-ews (efl):
Output: http://queues.webkit.org/results/16432937

[Dean Jackson]

Created attachment 187381 [details]
Another build test

[Dean Jackson]

chromium, qt and gtk are failing to link looking for snapshottedPlugInLabelTitle() and snapshottedPlugInLabelSubtitle() both of which landed in http://trac.webkit.org/changeset/142311

I'm not sure why it can't find those methods.

mac & mac-wk2 build error seems unrelated.

[Early Warning System Bot]

Comment on attachment 187381 [details]
Another build test

Attachment 187381 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/16465014

[Dean Jackson]

Created attachment 187384 [details]
Patch

[Dean Jackson]

Duh! I forgot to update platform localized strings.

[Dean Jackson]

The gtk failure is the same unrelated one that mac was hitting in InlineBox.

[Dean Jackson]

Committed r142333: <http://trac.webkit.org/changeset/142333>

[Dean Jackson]

Rolled out and landed again in http://trac.webkit.org/changeset/142343

[Kent Tamura]

(In reply to comment #59)
> Rolled out and landed again in http://trac.webkit.org/changeset/142343

Let me roll it out.  This change is very suspicious about a test crash.

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#group=%40ToT%20-%20webkit.org&showExpectations=true&tests=inspector%2Fprofiler%2Fselector-profiler-url.html

If the roll out won't resolve the crash, I'll re-land it.

[Kent Tamura]

Reverted r142343 for reason:

It might make inspector/profiler/selector-profiler-url.html crashy.

Committed r142400: <http://trac.webkit.org/changeset/142400>

[Dean Jackson]

(In reply to comment #61)
> Reverted r142343 for reason:
> 
> It might make inspector/profiler/selector-profiler-url.html crashy.
> 
> Committed r142400: <http://trac.webkit.org/changeset/142400>

Has it made a difference? I can never understand those flakiness pages.

[Kent Tamura]

(In reply to comment #62)
> > Reverted r142343 for reason:
> > 
> > It might make inspector/profiler/selector-profiler-url.html crashy.
> > 
> > Committed r142400: <http://trac.webkit.org/changeset/142400>
> 
> Has it made a difference? I can never understand those flakiness pages.

Yes, the revert fixed the crash.



Stack trace on Apple MountainLion Debug WK1 follows.  Other ports had similar stack.


Process:         DumpRenderTree [11741]
Path:            /Volumes/VOLUME/*/DumpRenderTree
Identifier:      DumpRenderTree
Version:         0
Code Type:       X86-64 (Native)
Parent Process:  Python [11327]
User ID:         501

Date/Time:       2013-02-10 07:48:29.816 -0800
OS Version:      Mac OS X 10.8.2 (12C54)
Report Version:  10

Crashed Thread:  0  Dispatch queue: com.apple.main-thread

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x000000000000000f

VM Regions Near 0xf:
--> 
    __TEXT                 000000010ed93000-000000010ee2f000 [  624K] r-x/rwx SM=COW  /Volumes/VOLUME/*

Application Specific Information:
CRASHING TEST: inspector/profiler/selector-profiler-url.html

Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0   com.apple.WebCore             	0x0000000110bc1cdc WebCore::CSSRule::hasCachedSelectorText() const + 12 (CSSRule.h:107)
1   com.apple.WebCore             	0x0000000110bc16fd WebCore::CSSStyleRule::selectorText() const + 45 (CSSStyleRule.cpp:82)
2   com.apple.WebCore             	0x00000001112ebd39 WebCore::SelectorProfile::startSelector(WebCore::CSSStyleRule const*) + 57 (InspectorCSSAgent.cpp:172)
3   com.apple.WebCore             	0x00000001112e7c12 WebCore::InspectorCSSAgent::willMatchRule(WebCore::StyleRule*, WebCore::StyleResolver*) + 146 (InspectorCSSAgent.cpp:964)
4   com.apple.WebCore             	0x000000011135f27e WebCore::InspectorInstrumentation::willMatchRuleImpl(WebCore::InstrumentingAgents*, WebCore::StyleRule*, WebCore::StyleResolver*) + 78 (InspectorInstrumentation.cpp:607)
5   com.apple.WebCore             	0x00000001120c501a WebCore::InspectorInstrumentation::willMatchRule(WebCore::Document*, WebCore::StyleRule*, WebCore::StyleResolver*) + 106 (InspectorInstrumentation.h:1205)
6   com.apple.WebCore             	0x00000001120af70d WebCore::StyleResolver::collectMatchingRulesForList(WTF::Vector<WebCore::RuleData, 0ul> const*, WebCore::StyleResolver::MatchRequest const&, WebCore::StyleResolver::RuleRange&) + 253 (StyleResolver.cpp:696)
7   com.apple.WebCore             	0x00000001120af5fc WebCore::StyleResolver::collectMatchingRules(WebCore::StyleResolver::MatchRequest const&, WebCore::StyleResolver::RuleRange&) + 1052 (StyleResolver.cpp:518)
8   com.apple.WebCore             	0x00000001120b01b2 WebCore::StyleResolver::matchUARules(WebCore::StyleResolver::MatchResult&, WebCore::RuleSet*) + 162 (StyleResolver.cpp:675)
9   com.apple.WebCore             	0x00000001120b05eb WebCore::StyleResolver::matchUARules(WebCore::StyleResolver::MatchResult&) + 171 (StyleResolver.cpp:1193)
10  com.apple.WebCore             	0x00000001120b0289 WebCore::StyleResolver::matchAllRules(WebCore::StyleResolver::MatchResult&, bool) + 57 (StyleResolver.cpp:750)
11  com.apple.WebCore             	0x00000001120ae46b WebCore::StyleResolver::styleForElement(WebCore::Element*, WebCore::RenderStyle*, WebCore::StyleSharingBehavior, WebCore::RuleMatchingBehavior, WebCore::RenderRegion*) + 1179 (StyleResolver.cpp:1445)
12  com.apple.WebCore             	0x0000000110e8591b WebCore::Element::styleForRenderer() + 283 (Element.cpp:1330)
13  com.apple.WebCore             	0x0000000110e85abd WebCore::Element::recalcStyle(WebCore::Node::StyleChange) + 397 (Element.cpp:1350)
14  com.apple.WebCore             	0x0000000110e85fdd WebCore::Element::recalcStyle(WebCore::Node::StyleChange) + 1709 (Element.cpp:1419)
15  com.apple.WebCore             	0x0000000110e85fdd WebCore::Element::recalcStyle(WebCore::Node::StyleChange) + 1709 (Element.cpp:1419)
16  com.apple.WebCore             	0x0000000110c71cde WebCore::Document::recalcStyle(WebCore::Node::StyleChange) + 1166 (Document.cpp:1804)
17  com.apple.WebCore             	0x0000000110c6dedc WebCore::Document::updateStyleIfNeeded() + 396 (Document.cpp:1852)
18  com.apple.WebCore             	0x0000000111e934ed WebCore::RenderView::repaintSelection() const + 45 (RenderView.cpp:563)
19  com.apple.WebCore             	0x0000000110fc053e WebCore::FrameSelection::focusedOrActiveStateChanged() + 78 (FrameSelection.cpp:1691)
20  com.apple.WebCore             	0x0000000110fc0a6a WebCore::FrameSelection::setFocused(bool) + 106 (FrameSelection.cpp:1739)
21  com.apple.WebCore             	0x0000000110f40c0d WebCore::FocusController::setFocused(bool) + 269 (FocusController.cpp:227)
22  com.apple.WebKit              	0x000000011021d8f1 -[WebHTMLView becomeFirstResponder] + 609 (WebHTMLView.mm:3844)
23  com.apple.AppKit              	0x00007fff956758e8 -[NSWindow makeFirstResponder:] + 702
24  com.apple.WebKit              	0x00000001101aa98e -[WebDefaultUIDelegate webView:makeFirstResponder:] + 62 (WebDefaultUIDelegate.m:99)
25  com.apple.CoreFoundation      	0x00007fff8f6e863c __invoking___ + 140
26  com.apple.CoreFoundation      	0x00007fff8f6e84d7 -[NSInvocation invoke] + 263
27  com.apple.CoreFoundation      	0x00007fff8f6e86a9 -[NSInvocation invokeWithTarget:] + 57
28  com.apple.WebKit              	0x00000001102b7e09 -[_WebSafeForwarder forwardInvocation:] + 361 (WebView.mm:3128)
29  com.apple.CoreFoundation      	0x00007fff8f6e3737 ___forwarding___ + 775
30  com.apple.CoreFoundation      	0x00007fff8f6e33b8 _CF_forwarding_prep_0 + 232
31  com.apple.WebKit              	0x000000011019c3d1 WebChromeClient::makeFirstResponder(NSResponder*) + 129 (WebChromeClient.mm:778)
32  com.apple.WebCore             	0x00000001109dfa81 WebCore::Chrome::focusNSView(NSView*) + 369 (ChromeMac.mm:46)
33  com.apple.WebCore             	0x000000011237f209 WebCore::Widget::setFocus(bool) + 201 (WidgetMac.mm:113)
34  com.apple.WebCore             	0x0000000110c77d37 WebCore::Document::setFocusedNode(WTF::PassRefPtr<WebCore::Node>, WebCore::FocusDirection) + 2071 (Document.cpp:3414)
35  com.apple.WebCore             	0x0000000110f42c1d WebCore::FocusController::setFocusedNode(WebCore::Node*, WTF::PassRefPtr<WebCore::Frame>, WebCore::FocusDirection) + 973 (FocusController.cpp:620)
36  com.apple.WebCore             	0x0000000110e880f2 WebCore::Element::focus(bool, WebCore::FocusDirection) + 306 (Element.cpp:1894)
37  com.apple.WebCore             	0x000000011116ef9f WebCore::HTMLInputElement::focus(bool, WebCore::FocusDirection) + 111 (HTMLInputElement.cpp:377)
38  com.apple.WebCore             	0x0000000111556bac WebCore::jsElementPrototypeFunctionFocus(JSC::ExecState*) + 444 (JSElement.cpp:1991)
39  ???                           	0x00002ae2a6801045 0 + 47152944386117
40  com.apple.JavaScriptCore      	0x000000010f197fe4 JSC::JITCode::execute(JSC::JSStack*, JSC::ExecState*, JSC::JSGlobalData*) + 84 (JITCode.h:135)
41  com.apple.JavaScriptCore      	0x000000010f190719 JSC::Interpreter::execute(JSC::EvalExecutable*, JSC::ExecState*, JSC::JSValue, JSC::JSScope*) + 2169 (Interpreter.cpp:1321)
42  com.apple.JavaScriptCore      	0x000000010f230565 JSC::globalFuncEval(JSC::ExecState*) + 997 (JSGlobalObjectFunctions.cpp:523)
43  com.apple.JavaScriptCore      	0x000000010f195289 JSC::Interpreter::executeCall(JSC::ExecState*, JSC::JSObject*, JSC::CallType, JSC::CallData const&, JSC::JSValue, JSC::ArgList const&) + 1561 (Interpreter.cpp:1058)
44  com.apple.JavaScriptCore      	0x000000010efa4a52 JSC::call(JSC::ExecState*, JSC::JSValue, JSC::CallType, JSC::CallData const&, JSC::JSValue, JSC::ArgList const&) + 306 (CallData.cpp:40)
45  com.apple.WebCore             	0x000000011168ee52 WebCore::JSInjectedScriptHost::evaluate(JSC::ExecState*) + 482 (JSInjectedScriptHostCustom.cpp:304)
46  com.apple.WebCore             	0x000000011168b82a WebCore::jsInjectedScriptHostPrototypeFunctionEvaluate(JSC::ExecState*) + 410 (JSInjectedScriptHost.cpp:310)
47  ???                           	0x00002ae2a6801045 0 + 47152944386117
48  com.apple.JavaScriptCore      	0x000000010f197fe4 JSC::JITCode::execute(JSC::JSStack*, JSC::ExecState*, JSC::JSGlobalData*) + 84 (JITCode.h:135)
49  com.apple.JavaScriptCore      	0x000000010f19525f JSC::Interpreter::executeCall(JSC::ExecState*, JSC::JSObject*, JSC::CallType, JSC::CallData const&, JSC::JSValue, JSC::ArgList const&) + 1519 (Interpreter.cpp:1055)
50  com.apple.JavaScriptCore      	0x000000010efa4a52 JSC::call(JSC::ExecState*, JSC::JSValue, JSC::CallType, JSC::CallData const&, JSC::JSValue, JSC::ArgList const&) + 306 (CallData.cpp:40)
51  com.apple.WebCore             	0x0000000111425572 WebCore::JSMainThreadExecState::call(JSC::ExecState*, JSC::JSValue, JSC::CallType, JSC::CallData const&, JSC::JSValue, JSC::ArgList const&) + 146 (JSMainThreadExecState.h:56)
52  com.apple.WebCore             	0x0000000111f3b329 WebCore::ScriptFunctionCall::call(bool&, bool) + 537 (ScriptFunctionCall.cpp:140)
53  com.apple.WebCore             	0x000000011125bae5 WebCore::InjectedScriptBase::callFunctionWithEvalEnabled(WebCore::ScriptFunctionCall&, bool&) const + 213 (InjectedScriptBase.cpp:91)
54  com.apple.WebCore             	0x000000011125bbc9 WebCore::InjectedScriptBase::makeCall(WebCore::ScriptFunctionCall&, WTF::RefPtr<WebCore::InspectorValue>*) + 137 (InjectedScriptBase.cpp:108)
55  com.apple.WebCore             	0x000000011125bd7a WebCore::InjectedScriptBase::makeEvalCall(WTF::String*, WebCore::ScriptFunctionCall&, WTF::RefPtr<WebCore::TypeBuilder::Runtime::RemoteObject>*, WebCore::TypeBuilder::OptOutput<bool>*) + 74 (InjectedScriptBase.cpp:120)
56  com.apple.WebCore             	0x0000000111256e6c WebCore::InjectedScript::evaluate(WTF::String*, WTF::String const&, WTF::String const&, bool, bool, bool, WTF::RefPtr<WebCore::TypeBuilder::Runtime::RemoteObject>*, WebCore::TypeBuilder::OptOutput<bool>*) + 316 (InjectedScript.cpp:72)
57  com.apple.WebCore             	0x00000001113aadf8 WebCore::InspectorRuntimeAgent::evaluate(WTF::String*, WTF::String const&, WTF::String const*, bool const*, bool const*, int const*, bool const*, bool const*, WTF::RefPtr<WebCore::TypeBuilder::Runtime::RemoteObject>&, WebCore::TypeBuilder::OptOutput<bool>*) + 600 (InspectorRuntimeAgent.cpp:92)
58  com.apple.WebCore             	0x00000001113ab043 non-virtual thunk to WebCore::InspectorRuntimeAgent::evaluate(WTF::String*, WTF::String const&, WTF::String const*, bool const*, bool const*, int const*, bool const*, bool const*, WTF::RefPtr<WebCore::TypeBuilder::Runtime::RemoteObject>&, WebCore::TypeBuilder::OptOutput<bool>*) + 179
59  com.apple.WebCore             	0x00000001112a6bd1 WebCore::InspectorBackendDispatcherImpl::Runtime_evaluate(long, WebCore::InspectorObject*) + 1553 (InspectorBackendDispatcher.cpp:1662)
60  com.apple.WebCore             	0x00000001112c4789 WebCore::InspectorBackendDispatcherImpl::dispatch(WTF::String const&) + 1833 (InspectorBackendDispatcher.cpp:6335)
61  com.apple.WebCore             	0x00000001112d8390 WebCore::InspectorController::dispatchMessageFromFrontend(WTF::String const&) + 96 (InspectorController.cpp:357)
62  com.apple.WebCore             	0x000000011135862e WebCore::InspectorBackendDispatchTask::onTimer(WebCore::Timer<WebCore::InspectorBackendDispatchTask>*) + 110 (InspectorFrontendClientLocal.cpp:94)
63  com.apple.WebCore             	0x0000000111358a43 WebCore::Timer<WebCore::InspectorBackendDispatchTask>::fired() + 115 (Timer.h:106)
64  com.apple.WebCore             	0x000000011229b356 WebCore::ThreadTimers::sharedTimerFiredInternal() + 294 (ThreadTimers.cpp:119)
65  com.apple.WebCore             	0x000000011229b0e9 WebCore::ThreadTimers::sharedTimerFired() + 25 (ThreadTimers.cpp:94)
66  com.apple.WebCore             	0x0000000111fcb513 WebCore::timerFired(__CFRunLoopTimer*, void*) + 67 (SharedTimerMac.mm:167)
67  com.apple.CoreFoundation      	0x00007fff8f6b1da4 __CFRUNLOOP_IS_CALLING_OUT_TO_A_TIMER_CALLBACK_FUNCTION__ + 20
68  com.apple.CoreFoundation      	0x00007fff8f6b18bd __CFRunLoopDoTimer + 557
69  com.apple.CoreFoundation      	0x00007fff8f697099 __CFRunLoopRun + 1513
70  com.apple.CoreFoundation      	0x00007fff8f6966b2 CFRunLoopRunSpecific + 290
71  com.apple.Foundation          	0x00007fff8c68489e -[NSRunLoop(NSRunLoop) runMode:beforeDate:] + 268
72  DumpRenderTree                	0x000000010edac009 runTest(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 5017 (DumpRenderTree.mm:1375)
73  DumpRenderTree                	0x000000010edaabfa runTestingServerLoop() + 282 (DumpRenderTree.mm:832)
74  DumpRenderTree                	0x000000010edaa5f5 dumpRenderTree(int, char const**) + 405 (DumpRenderTree.mm:887)
75  DumpRenderTree                	0x000000010edac7f9 main + 105 (DumpRenderTree.mm:925)
76  libdyld.dylib                 	0x00007fff8f51f7e1 start + 1

[Dean Jackson]

Removed the offending style rule that was causing the crash. It's not really related to this commit, so I'll file another bug.

Committed r142507: <http://trac.webkit.org/changeset/142507>