QA » openQA Project

https://github.com/os-autoinst/os-autoinst/pull/1546 , already merged, and https://github.com/mudler/Mojo-IOLoop-ReadWriteProcess/pull/13

if you don't do further work but wait for https://github.com/mudler/Mojo-IOLoop-ReadWriteProcess/pull/13 please set to "Feedback".

the PR was accepted, the release including the change is https://github.com/mudler/mojo-ioloop-readwriteprocess/releases/tag/0.28 but also https://build.opensuse.org/package/show/devel:languages:perl/perl-Mojo-IOLoop-ReadWriteProcess is still at 0.27 . Probably because the perl autoupdate still does not work but tinita is looking into that. However https://build.opensuse.org/package/show/devel:languages:perl:autoupdate/perl-Mojo-IOLoop-ReadWriteProcess has the new version already. So created a submission with osc sr -m "Update to 0.28" devel:languages:perl:autoupdate perl-Mojo-IOLoop-ReadWriteProcess devel:languages:perl creating https://build.opensuse.org/request/show/839403

SR accepted, now forwarded for openSUSE:Factory with https://build.opensuse.org/request/show/840304

0.28 is in now by the looks of it. Does the timeout still need adjusting?

As mentioned in the PR:

This saves measurably time, e.g. the os-autoinst test t/18-qemu-options.t
needs 28 seconds to execute on my system without this change and 11 seconds
with this change.

I'm not sure how much there's left to optimize. And yes, I still need to adjust the timeout. However, if we're actually enforcing that within the testsuite we should also specify in the dependency files that we require at least 0.28.

hm, actually we would only strongly depend on 0.28 for "tests" - good that we can encode that separately :)

mkittler wrote:

PR for adding the time limit test helper to os-autoinst-common to use it for t/18-qemu-options.t to cover AC2: https://github.com/os-autoinst/os-autoinst-common/pull/3

ok, not a bad idea. But why not just rely on what is already there in https://github.com/os-autoinst/os-autoinst/blob/master/t/18-qemu-options.t#L90 ?

PR for changing the specific time check: https://github.com/os-autoinst/os-autoinst/pull/1552

PR for overall timeout: https://github.com/os-autoinst/os-autoinst/pull/1553

PR is merged, https://build.opensuse.org/package/live_build_log/devel:openQA/os-autoinst/openSUSE_Factory_PowerPC/ppc64le failed now with

[  248s] 3: ./17-basetest.t .......................... ok
[  248s] 3: Bailout called.  Further testing stopped:  test exceeds runtime limit of '5' seconds
[  248s] 3: FAILED--Further testing stopped: test exceeds runtime limit of '5' seconds
[  248s] 3/3 Test #3: test-perl-testsuite ..............***Failed   86.27 sec

looks a bit like the test already passed but maybe a dirty subprocess is stuck or needs longer to finish

Likely 5 seconds are too low in OBS anyways. I'll add timeout scaling like for the CI: https://github.com/os-autoinst/os-autoinst/pull/1556

merged. Please wait for results in OBS.

That should be enough waiting :) Thank you so far. I wanted to check what travis CI records as runtime for the individual test so that we can check on the current value. But I realized we don't record the timer output. Anyway, unless you plan for more work feel welcome to set to "Resolved" now