On 12/12/2019 16.33, Graham Inggs wrote: > On 2019/12/12 16:25, Andreas Beckmann wrote: >> I can reproduce the error in the silx tests in sid. This could be >> related to pocl being rebuilt for the hwloc 1->2 transition. > > Silx's autopkgtests started failing in unstable on 2019-11-15 [1]. > Pocl was binNMU'd for hwloc15 on 2019-11-27 [2].
> [1] https://ci.debian.net/packages/s/silx/unstable/amd64/ The first failures were due to the "python3.8 supported" transition (and something numpy not being ready, yet). But anyway you are right, the test on 2019-11-26 already failed the same way still using the old hwloc. Actually on 2019-11-23 this pocl-specific failure happened for the first time, but before that, the test did not get far enough... I can reproduce it in current sid with both python3.7 and python3.8 ... Attached is the debci artefacts diff from the last good one to the first with this failure. There is the python3.8 stuff going on, but also the transition to OpenMPI 4. I would really need a smaller testcase to reproduce this bug for further debugging. Andreas
silx_3491834_artifacts.diff.gz
Description: application/gzip
