another pytest hang

[Rebecca N. Palmer]

Today's snakemake 5.18.1 build also hung, but at the end of the tests 
rather than the start: when stopped by the timeout, it reported that the 
tests had passed but taken ~16x longer than normal.  The same package 
builds normally in a local Ubuntu chroot, and in Debian.

Both snakemake and statsmodels use the pytest test runner.  This has 
upstream reports of both the at-start and at-end hangs, but they mostly 
seem to be closed as not enough information / obsolete pytest version 
(i.e. what we would call Expired or Invalid) without actual progress:
https://github.com/pytest-dev/pytest/issues?q=is%3Aissue+hang

Should I keep requesting retries of these?  File an Ubuntu bug against 
pytest?  Both?  Or something else?

On 11/05/2020 10:32, Rebecca N. Palmer wrote:
> The statsmodels 0.11.1-2 autopkgtest timed out on Ubuntu arm64 and 
> ppc64el, but succeeded in roughly the usual amount of time on Ubuntu 
> amd64 and Debian arm64.
> 
> In both cases the last output was pytest's "collecting ... ", implying a 
> hang at the test collection (not test running) stage, but not 
> identifying which file it happened in.
> 
> Does it make sense to retry them?
(this was done, successfully)