A different stall problem using build http://nightly.webkit.org/start/trunk/26699

• Component: New Bugs
• Status: UNCONFIRMED ---

[Jim Oase]

This report was attached to an old problem report.  I was asked to start a new report.

Here is the build number
http://nightly.webkit.org/start/trunk/26699

Stall was on the Consumer Report site which stalled for over 30 seconds.

Jim

[Jim Oase]

Created attachment 16723 [details]
Terminal dump for stall

[Mark Rowe (bdash)]

There is nothing in the sample output that you attached that indicates that the application was unresponsive.  Did you generate this output while the application was beachballing?

[Jim Oase]

I am not sure if the site was still beach balling or not.  It took me a bit to find Termina, open it.  The first time I tried to execute       sample safari 10 10         Terminal complained so I retyped the command this time capitalizing Safari and it worked.

Jim

[Jim Oase]

Created attachment 16727 [details]
Stalled loading webkit.org

stalled for several seconds

Jim

[Mark Rowe (bdash)]

What do you mean by "stall"?  Does the spinning beachball appear, or does the page just take longer than you would expect to load?

                  8 -[NSHTTPURLProtocol determineErrorAndFail]
                    8 -[NSNetworkSettings connectedToInternet:]
                      8 SCNetworkCheckReachabilityByName
                        8 SCNetworkReachabilityGetFlags
                          8 __SCNetworkReachabilityGetFlags


That looks very much like CFNetwork determined was trying to determine whether you were connected to the Internet.  If one or more of your DNS servers was being slow it could take a second or two for this check to complete.

[Jim Oase]

What I see is the pin wheel on the tab.  However when i click again the page will immediately load what appears to be the previous page.  Then the click request appears to go to the internet, the page will go blank again,  the new page comes in and loads immediately.

Jim

[Mark Rowe (bdash)]

Ok, that sounds very much like what you're seeing is a problem at the network layer while loading the page.  That matches the CFNetwork information in the "sample" output.

[Robert Blaut]

(In reply to comment #7)
> Ok, that sounds very much like what you're seeing is a problem at the network
> layer while loading the page.  That matches the CFNetwork information in the
> "sample" output.
>
 
So I think the problem should be reported to http://bugreport.apple.com. CFNetwork isn't a part of WebKit. Is it correct?