[webkit-reviews] review granted: [Bug 45790] ASSERTION FAILED: m_loadEventDelayCount : [Attachment 67823] Patch

[bugzilla-daemon at webkit.org]

Adam Barth <abarth at webkit.org> has granted Tony Gentilcore
<tonyg at chromium.org>'s request for review:
Bug 45790: ASSERTION FAILED: m_loadEventDelayCount
https://bugs.webkit.org/show_bug.cgi?id=45790

Attachment 67823: Patch
https://bugs.webkit.org/attachment.cgi?id=67823&action=review

------- Additional Comments from Adam Barth <abarth at webkit.org>
View in context:
https://bugs.webkit.org/attachment.cgi?id=67823&action=prettypatch

> WebCore/WebCore.xcodeproj/project.pbxproj:2388
> -		893C47DF123EF4A9002B3D86 /* JSDirectoryEntryCustom.cpp in
Sources */ = {isa = PBXBuildFile; fileRef = 893C47DE123EF4A9002B3D86 /*
JSDirectoryEntryCustom.cpp */; };
>		893C47CC123EEBA2002B3D86 /* JSEntryCustom.cpp in Sources */ =
{isa = PBXBuildFile; fileRef = 893C47CA123EEBA2002B3D86 /* JSEntryCustom.cpp
*/; };
> +		893C47DF123EF4A9002B3D86 /* JSDirectoryEntryCustom.cpp in
Sources */ = {isa = PBXBuildFile; fileRef = 893C47DE123EF4A9002B3D86 /*
JSDirectoryEntryCustom.cpp */; };

These changes seems spurious.  Please land without them.  If the project file
is wrong, we can fix it in a separate commit.

> WebCore/dom/AsyncScriptRunner.cpp:81
> +    RefPtr<Document> document(m_document);

Usually we'd call this variable protector and then just m_document through the
function.

> WebCore/dom/AsyncScriptRunner.cpp:87
>	   scripts[i].first->execute(scripts[i].second.get());

Does this execute script when the document is detached from the frame?	It's an
invariant that we never execute scripts on behalf of detached documents.  (Not
really an issue for this patch, just something to think about.)