Closed baldurmen closed 2 months ago
Hello.
I remember having issues with that test suite, but tests were just plainly failing, not running indefinitely.
I'm not sure the _sleep()
function is to blame, it could also be that the watcher fails to stop at the end of the test.
You say it fails on ARM architectures? I will have to dust off my old Raspberry Pi then :P
Ok I think I found it. It was a thread that failed to get the notification to stop.
I tested it on a Raspberry Pi 1 B+ (armv6l
architecture then) running Raspbian Bullseye (Python 3.9). I did get a never ending test but oddly enough it wasn't test_move_out
but test_change
. It doesn't seem to stall now.
I'm also getting a varying number of failing tests but haven't tackled them. I guess the sluggish Pi doesn't like those time dependent tests.
Could you please test again if that's ok now? Do you need me to publish a new version to run the tests?
Sorry for not replying earlier! That patch looks good to me :) I've upgraded the Debian package to the latest release. If that doesn't fix it, I'll come back :)
Thanks again for maintaining supysonic!
Hi!
We're running the supysonic testsuite fairly often in Debian (via the DebCI project) and it seems the
tests.base.test_watcher.AudioWatcherTestCase.test_move_out
test is somewhat flaky and causes timeouts in our infra (we kill tests after ~2h40).We've seen this on multiple different architectures (
arm64
,armel
,armhf
,i386
,ppc64el
ands390x
), but not onamd64
(which you are probably developing on :P). Here are a few examples:I'm not exactly sure why this happens, but I guess it's related to the
_sleep()
function called in this test?I'd be happy to try to debug things some more if you want me to test specific things :P For now, I've disabled this test in our package.
Cheers,