KCrash - test_crasher spins in a loop on Windows

Ben Cooksley bcooksley at kde.org
Tue Jul 23 19:02:15 BST 2019


On Wed, Jul 24, 2019 at 2:27 AM David Faure <faure at kde.org> wrote:
>
> On mercredi 17 avril 2019 08:52:23 CEST Ben Cooksley wrote:
> > On Sat, 13 Apr 2019, 07:35 Ben Cooksley, <bcooksley at kde.org> wrote:
> > > Hi all,
> > >
> > > Currently we have an issue where the test_crasher test will endlessly
> > > fork itself while trying to setup it's conditions to run.
> > >
> > > This leads to an issue where it will occupy a single CPU core of a
> > > builder indefinitely until it's killed (and when killed there are
> > > always two test_crasher.exe processes running, and if not killed
> > > simultaneously it won't move on to the next test)
> > >
> > > Any ideas on how to resolve this?
> > >
> > > (This has existed for some time, i've just not had the time to raise
> > > it as an issue)
> >
> > Gentle ping? Anyone with any thoughts on this?
>
> Is this still an issue?
>
> Does KCrash generally work on Windows?
>
> If nobody has time to debug this autotest on Windows, we could disable it...

To my knowledge the test is still broken on Windows yes.
I'm not sure to what extent KCrash is usable/works on Windows, I guess
that will depend on whether Dr Konqi has support for running a Windows
debugger?

For those who install Visual Studio, I did notice that it hooks into
the standard Windows crash reporting system allowing you to transfer
the application that crashed directly there for debugging, so I guess
the question is how much value Dr Konqi has there (given most of our
builds are done with MSVC)

>
> --
> David Faure, faure at kde.org, http://www.davidfaure.fr
> Working on KDE Frameworks 5
>
>
>

Cheers,
Ben


More information about the Kde-frameworks-devel mailing list