[Webkit-unassigned] [Bug 165670] Add a warning for unused link preloads.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=165670

Yoav Weiss <yoav at yoav.ws> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |youennf at gmail.com

--- Comment #28 from Yoav Weiss <yoav at yoav.ws> ---
(In reply to comment #24)
> (In reply to comment #23)
> > Comment on attachment 304495 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=304495&action=review
> > 
> > > Source/WebCore/loader/cache/CachedResourceClient.h:48
> > > +    virtual bool markAsReferenced() const { return true; }
> > 
> > I'm not sure this is named correctly.  shouldMarkAsReferenced?
> 
> Yeah, I'll rename
> 
> > Isn't there already a way to tell if a load is a preload?
> 
> There is already a way to tell if a certain request is a preload request or
> a preloaded resource.
> However, a preloaded resource can be referenced (e.g. `<link rel=preload
> href=foo as=script><script src=foo>`).
> Current tracking if a resource is referenced is based on the addition of
> CachedResourceClients to it. So I figured it'd be cleanest to add that "is
> this referencing" info to the resource client.
>  
> > 
> > > LayoutTests/TestExpectations:1260
> > > +http/tests/preload/single_download_preload.html [ DumpJSConsoleLogInStdErr ]
> > 
> > Does this make these tests not accidentally print out a warning message?
> 
> Yeah
> 
> > Is there no way to do this from within the test?
> 
> Don't know. If there is, I'd be happy to switch to that method

Hmm, the debug bot failure seems to indicate that it's not taking this directive into account. Is that a known issue? Is there a way to get around that?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170320/6ad0e7c8/attachment.html>