Hooked up the test data!

Gilles Caulier caulier.gilles at gmail.com
Mon Oct 3 19:31:38 BST 2022


Hi Steven,

Please take a look to this MR in QtAV :

https://github.com/wang-bin/QtAV/pull/1109/files/705ad03cc01f9f710963d0923de80c0b1623232d..d63024fffb4e7029f3bff97e29ede78725487a80

I'm sure that code will not be merged, as maintainers declare the
project as end of life. But at least, perhaps there are few ideas to
retain here, as for ex the way to init opengl context.

If it's not easy to backport, just ignore.

Gilles

Le lun. 3 oct. 2022 à 16:36, Gilles Caulier <caulier.gilles at gmail.com> a écrit :
>
> Le lun. 3 oct. 2022 à 16:18, Steven Robbins <steve at sumost.ca> a écrit :
> >
> > Hello,
> >
> > I have populated and pushed (a fork of) digika-test-data [1] and hooked it up
> > to the main digikam repo -- again, in my fork [2].  After Ben pointed out that
> > perl was not available on all platforms, I just put in a "git clone" line into
> > cmake and it works fine on the three testers.
> >
> > So now my problem is the actual tests.  They run fine on my system -- I have
> > tested both ffmpeg 4 and 5 -- but fail on all three testers.  The failure is
> > right in the beginning -- at the load() call.  I had been supposing it was
> > because the codec wasn't supported, but I changed codecslist.cpp to always
> > print and the codec used in the file is listed.  So I'm boggled.
> >
> > Is there any way to get a shell on one of these systems to do test build/
> > debug?
>
> Hi Steven,
>
> I think it's not possible, for security reasons.
> The CI VM access are limited from users.
> The best way is to ask Ben directly if he can capture the information
> that you need directly.
> Another solution is to use cmake to print details of system
> configuration at configuration time.
>
>
> Best
>
> Gilles


More information about the Digikam-devel mailing list