Bug 128022

Summary:

Add webView:didFailNavigation:withError: delegate method

Product:

WebKit

Reporter:

Anders Carlsson <andersca>

Component:

New Bugs

Assignee:

Anders Carlsson <andersca>

Status:

RESOLVED FIXED

Severity:

Normal

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	darin: review+

Description Anders Carlsson 2014-01-31 14:53:34 PST

Add webView:didFailNavigation:withError: delegate method

Comment 1 Anders Carlsson 2014-01-31 14:53:57 PST

Created attachment 222857 [details]
Patch

Comment 2 Darin Adler 2014-01-31 15:39:56 PST

Comment on attachment 222857 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=222857&action=review

> Source/WebKit2/UIProcess/API/C/WKPage.cpp:743
> +        virtual void didFailLoadWithErrorForFrame(WebPageProxy* page, WebFrameProxy* frame, uint64_t navigationID, const ResourceError& error, API::Object* userData) override

Won’t we get a warning because we are not using navigationID here?

> Source/WebKit2/WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:487
> +    WebDocumentLoader& documentLoader = static_cast<WebDocumentLoader&>(*m_frame->coreFrame()->loader().documentLoader());

Do we have a checked version of this cast? Should we?

Seems like coreFrame should return a reference, not a pointer.

Comment 3 Anders Carlsson 2014-02-01 09:24:21 PST

(In reply to comment #2)
> (From update of attachment 222857 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=222857&action=review
> 
> > Source/WebKit2/UIProcess/API/C/WKPage.cpp:743
> > +        virtual void didFailLoadWithErrorForFrame(WebPageProxy* page, WebFrameProxy* frame, uint64_t navigationID, const ResourceError& error, API::Object* userData) override
> 
> Won’t we get a warning because we are not using navigationID here?

I don't think we enable that warning.

> 
> > Source/WebKit2/WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:487
> > +    WebDocumentLoader& documentLoader = static_cast<WebDocumentLoader&>(*m_frame->coreFrame()->loader().documentLoader());
> 
> Do we have a checked version of this cast? Should we?

We're sure that all document loaders are created by WebKit so I don't think we do.

> 
> Seems like coreFrame should return a reference, not a pointer.

I think the member variable backing coreFrame can end up being nulled out.

Comment 4 Anders Carlsson 2014-02-01 09:27:04 PST

Committed r163243: <http://trac.webkit.org/changeset/163243>