how to debug test build issues form Debian servers

[olivier sallou <olivier.sallou@gmail.com>]

Hi,

I have a package that fails some tests on a few architectures.

Looking at logs from those servers, it seems that it is quite random (not always same arch, or same tests), could be due to some timeouts in tests etc...

I am not upstream author, impacted tests are not easy to debug. And log show no detail except log failure count (and which test).

Should I create a new package version (upstreamversion-X) and upload it to sid then wait for the build of each arch to check if my patch work ? and create new version again if patch is not the expected one ?

This could lead to many patched package upload just for test....

Indeed, test failure seem related to the build/test servers (lower RAM, higher cpu load.... generating timeout?).

Thanks for any suggestion.

Olivier