Bug#1115118: deal.ii: autopkgtest regression on riscv64

Graham Inggs ginggs at debian.org
Sun Sep 14 13:09:33 BST 2025


Hi Sebastian

> the autopkgtest of deal.ii fails on riscv64 (maybe due to the trilinos
> transition):
...
> See https://ci.debian.net/packages/d/deal.ii/testing/riscv64/64460939/
> for details.

The log to which you refer, shows that deal.ii from testing was
installed for the autopkgtest:

130s Get:532 http://mirrors.tuna.tsinghua.edu.cn/debian testing/main
riscv64 libdeal.ii-dev riscv64 9.6.2-4 [2,258 kB]

Given the trilinos transition, this was probably expected to fail.

However, the equivalent migration tests scheduled for amd64 [1],
ppc64el [2] and s390x [3], show the following, respectively:

 55s Get:541 http://deb.debian.org/debian unstable/main amd64
libdeal.ii-dev amd64 9.6.2-4+b3 [2,259 kB]

886s Get:534 http://deb.debian.org/debian unstable/main ppc64el
libdeal.ii-dev ppc64el 9.6.2-4+b3 [2,259 kB]

 48s Get:523 http://deb.debian.org/debian unstable/main s390x
libdeal.ii-dev s390x 9.6.2-4+b3 [2,258 kB]

deal.ii from unstable was installed, and the autopkgtests passed.
Unfortunately, due to long build times, the tests on riscv64 were run
more than a week after the tests on these architectures, and a lot has
changed in unstable and testing since then, including a new version of
openmpi.

I retried the riscv64 autopkgtest with the same trigger as requested
by britney, but with an additional pin on deal.ii from unstable and it
passed [4].

350s Get:504 http://mirrors.tuna.tsinghua.edu.cn/debian unstable/main
riscv64 libdeal.ii-dev riscv64 9.6.2-4+b3 [2,259 kB]

Regards
Graham


[1] https://ci.debian.net/packages/d/deal.ii/testing/amd64/64145024/
[2] https://ci.debian.net/packages/d/deal.ii/testing/ppc64el/64144872/
[3] https://ci.debian.net/packages/d/deal.ii/testing/s390x/64186307/
[4] https://ci.debian.net/packages/d/deal.ii/testing/riscv64/64538437/



More information about the debian-science-maintainers mailing list