Saturday, 15 July 2023

+1 maintenance report

+1 report, week of July-10-2023

### ocaml-sexplib0 cluster: ppx-import vs frama-c ###

ocaml-sexplib0 was waiting on ppx-import, which was waiting on adt of frama-c
with a ppx-import trigger. frama-c/ppx-import retests run, passed on retry.

A red herring while analyzing this: frama-c, when autopkgtested locally, trips
shell-fail due to a test which wants package alt-ergo. alt-ergo was removed for
license reasons. That might normally be a problem but the test is marked
skip-not-installable so adt doesn't actually fail and it's just a distraction.

### ocaml-sexplib0 cluster: ocaml-sedlex ###

I don't recognize the failure and can't reproduce the test failure locally, and
retries hadn't been done yet. I ran the retests, which passed.

### ocaml-sexplib0 cluster: utop vs ocaml-dune (LP: #2027333) ###

ocaml-sexplib0 -> lwt -> utop reported being uninstallable.
libutop-ocaml-dev : Depends: libdune-ocaml-dev-v9190 but it is not installable

The debian ocaml policy, as documented in the dh-ocaml package, wants
ocaml:Provides and ocaml:Depends in the control file, and the ocaml:Provides is
not present in ocaml-dune. Bug opened in Debian with debdiff adding this, and
uploaded.

### ocaml-sexplib0 cluster: ocaml-qcheck vs ocaml-dune (LP: #2027522) ###

Separately from the previous item, ocaml-qcheck fails autopkgtest. qcheck uses
dune-ocaml as part of building examples, which needs an explicit dune project
file. Also wants a package name tweak s/dune/ocaml-dune/ in the test
dependency list. Bug opened in Debian with debdiff addressing these, and
uploaded. ocaml-qcheck tests rerun with ocaml-dune trigger.

### nbs: libarcus3 vs cura-engine (LP: #2026769, LP: #2026778) ###

libarcus3 needed cura-engine to move along. cura-engine needed a fix backported
from cura-engine upstream. Uploaded, forwarded to Debian.

However, I neglected to notice that armhf was also not building for an
unrelated issue. Fixed in Debian but not yet uploaded, so I incorporated
https://salsa.debian.org/3dprinting-team/cura-engine/-/commit/7acf586b03a18e21bda8c2195a333f6af0f7c469
and reuploaded.

### fenics-dolfinx ###

fenics-dolfinx calls doxygen, which aborts during build, only on amd64 and on
mantic (but not sid). The abort due to a warning:
error: Failed to extract bounding box from generated diagram file
/src/pkg/cpp/doc/latex/d2/d97/classdolfinx_1_1io_1_1FidesWriter__inherit__graph.pdf
(warning treated as error, aborting now)

Newer cairo is changing the output of dot in a way that doxygen doesn't care
for.

Unknowingly, ahasenack and I investigated this in parallel. See LP: #2026834
for his writeup.

### selinux-python cluster: refpolicy (LP: #2027733) ###

refpolicy autopkgtest failed as follows:
Setting up selinux-policy-default (2:2.20221101-9) ...
Updating selinux default policy (this step might take a moment)...Problems
processing filecon rules
Failed post db handling
Post process failed
semodule: Failed!
failed.

semodule can take a verbose argument, which provided useful debugging
information. Following that trail lead to http://bugs.debian.org/1038968 .

/var/lib/selinux/mls/tmp/modules/100/spamassassin/cil

Uploaded with duplicate rule removed, as suggested in debbug.

### selinux-python cluster: libsemanage ###

Internal compiler error on riscv64, pass on rebuild.

### quick look ###

* nbs: sight vs libinsighttoolkit5.2 (LP: #2026781) - FTBFS with gcc-13

-Dan

--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel