ResourceTiming entries for cached resources and XHR

• Component: New Bugs
• Status: RESOLVED FIXED

[Yoav Weiss]

ResourceTiming entries for cached resources and XHR

[Yoav Weiss]

Created attachment 278827 [details]
Patch

[Yoav Weiss]

Created attachment 278831 [details]
Patch

[Yoav Weiss]

Created attachment 278832 [details]
Patch

[Yoav Weiss]

This fixes a few issues with ResourceTiming that were preventing it from being useful:
* Resources that were being reused from MemoryCache were not getting their own entries. 
* XHR based resources were not being registered.

The patch also splits out the resource timing entry registration logic from CachedResourceLoader into its own class.

[Alex Christensen]

Comment on attachment 278832 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=278832&action=review

> Source/WebCore/loader/ResourceTimingInformation.cpp:14
> + *     * Neither the name of Google Inc. nor the names of its

I'm not sure this is the correct copyright.

> Source/WebCore/loader/ResourceTimingInformation.cpp:57
> +                initiatorDocument = document->parentDocument();

What if there are deeply nested iframes?  Should this be topDocument()?

> Source/WebCore/loader/ResourceTimingInformation.h:32
> +#ifndef ResourceTimingInformation_h
> +#define ResourceTimingInformation_h

#pragma once is the latest thing in WebKit.

[Yoav Weiss]

Created attachment 278864 [details]
Patch

[Yoav Weiss]

(In reply to comment #5)
> Comment on attachment 278832 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=278832&action=review
> 
> > Source/WebCore/loader/ResourceTimingInformation.cpp:14
> > + *     * Neither the name of Google Inc. nor the names of its
> 
> I'm not sure this is the correct copyright.

Indeed! Corrected

> 
> > Source/WebCore/loader/ResourceTimingInformation.cpp:57
> > +                initiatorDocument = document->parentDocument();
> 
> What if there are deeply nested iframes?  Should this be topDocument()?

I believe each iframe is reported as a resource in its parent document, so the above should by correct. FWIW, I didn't change the existing logic but just moved it from CachedResourceLoader to here.

> 
> > Source/WebCore/loader/ResourceTimingInformation.h:32
> > +#ifndef ResourceTimingInformation_h
> > +#define ResourceTimingInformation_h
> 
> #pragma once is the latest thing in WebKit.

Neat! Changed

[Yoav Weiss]

Comment on attachment 278864 [details]
Patch

Thanks for reviewing! :)

[WebKit Commit Bot]

Comment on attachment 278864 [details]
Patch

Clearing flags on attachment: 278864

Committed r200887: <http://trac.webkit.org/changeset/200887>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Joseph Pecoraro]

*** Bug 157259 has been marked as a duplicate of this bug. ***