[Debichem-devel] Bug#897191: Bug#897191: pymol: autopkgtest times out with new version while succeeded in the past

[Stuart Prescott]

Hi all,

> With the upload of version 2.1.0+dfsg-1 of pymol, the autopkgtest¹
> started to fail due to it timing out (> 11 hours) on 4 of the tests and
> additionally due to unexpected output to stderr (see the relevant part
> of the log copied below). Previous versions only required a couple of
> minutes. Therefore I assume a lot of tests got added (albeit no note in
> the changelog about that), otherwise it is really worth it to try and
> find out why it now takes so much longer.
> 
> A note about time outs. Time outs can often be worked around by
> splitting a test into multiple autopkgtest because each gets its own
> timeout. On the other hand, is 11 hours or more really worth it?
> Couldn't a limited set of the test be run?

This doesn't appear to actually be a case of a test that is simply slow but a 
test that has hung.

Adding 'set -x' to the test runner shows that it gets stuck (for the first 
time) on

pymol-2.1.0+dfsg/examples/cookbook/scenes2movie.pml

Running that script by hand shows why -- it's playing the movie in a loop 
forever. Does someone know pymol well enough to say that the "mplay" syntax 
has changed or how to limit the movie to just one run through? 

> Output to stderr normally causes a test to fail. One can overrule that
> by add allow-stderr to the test Restrictions if the output is expected
> and to be ignored.

The stderr output also indicates that the test runner think that the test run 
has failed. Once more with set -x in place, the runner is not finding 'PyMOL: 
normal program termination.' in the pymol output. 

Andreas, as the uploader of the current version, do you have a log from your 
local run of the autopkgtests so that we can compare it to the current one on 
ci.d.n and see if there's an obvious difference? (Also, was deleting the Vcs-* 
fields from d/control with a commit titled "Standards-Version: 4.1.4" 
intentional?)

cheers
Stuart

-- 
Stuart Prescott    http://www.nanonanonano.net/   stuart at nanonanonano.net
Debian Developer   http://www.debian.org/         stuart at debian.org
GPG fingerprint    90E2 D2C1 AD14 6A1B 7EBB 891D BBC1 7EBB 1396 
F2F7pymol-2.1.0+dfsg/examples/cookbook/scenes2movie.pml