Update: There is still one problem that prevents us from deploying the new network test server image.
The problem only shows within the Digia network; I cannot reproduce it locally. This makes it somewhat tedious to debug, if somebody from Digia has more insights or wants to help debugging, I would be happy to hear. So far I have sent patches to Tony, who applied them for me and pasted me the results, which makes turnaround time quite long. What I think the problem is: We get 2 Socks messages in 1 TCP packet, so the read notifier only fires once, then we never parse the 2nd Socks message. Peter On 02/06/2014 01:54 PM, Sarajärvi Tony wrote: > Hi > > We've discovered (just lately) that in some cases incoming TCP packages are > bundled together. Somewhere along the line in our network code the first TCP > packet is drawn from buffers, but the remainders are left there. It's a > random thing, which we can reproduce in our test rig where we develop the new > network test server (that's supposed to replace this old one as soon as we > get it 100% working). Peter is working on the problem it and hopefully we'll > get a patch for it soon. > > Reagards, > -Tony > >> -----Original Message----- >> From: development-bounces+tony.sarajarvi=digia....@qt-project.org >> [mailto:development-bounces+tony.sarajarvi=digia....@qt-project.org] On >> Behalf Of BogDan >> Sent: 6. helmikuuta 2014 14:26 >> To: Mandeep Sandhu; Thiago Macieira >> Cc: development@qt-project.org >> Subject: Re: [Development] CI broken again? >> >> Hi, >> >>> >>> In your case, the QDnsLookup failure is caused by your code. It happened >> 100% >>>> of the time that the QDnsLookup test was executed and it did not happen >> when >>>> your change wasn't present. >>>> >>> >>> >>> It can't be related to my change as the same tests pass 100% of the time on >> my local setup (Ubuntu 13.10) and moreover my changes do not kick-in if the >> name server is not specified (which is the case with all the failing test >> cases). >>> >>> >>> The latest test failures seemed to be due to timeout rather than anything >> else. The 'waitForDone()' function is exiting with a failure due to DNS >> lookup >> timeout. The current timeout is set to 10 secs. This is quite large, but >> still, >> could there be a network issue with the CI setup? >>> >>> >> >> Ok, maybe his patch has some problems, but I'm pretty sure that my patch >> was staged without his patch a few times and it still failed :( . A little >> help >> from CI folks will be very appreciated! >> >> Is a little bit frustrating to fix a problem in a few minutes and wait a >> few days >> to integrate it :). >> >> Cheers, >> BogDan. >> >> _______________________________________________ >> Development mailing list >> Development@qt-project.org >> http://lists.qt-project.org/mailman/listinfo/development _______________________________________________ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
