Hi,

During a test rebuild of HDF5's reverse dependcies against HDF5 2.1.0
currently in experimental, med-fichier FTBFS with 2 failing tests.

Patching src/ci/MEDfileCompatibility.c to relax the HDF5 version check
works, but then another test fails more badly in tests/unittests/c:

FAIL: Test_MEDnStructElement_Rd.sh

The output from this failing test is:

../../../../tests/unittests/c/Test_MEDnStructElement.c [48] : _nstructelement = 2
HDF5: infinite loop closing library
      L,T_top,F,P,P,FD,VL,VL,E,SL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL
Aborted

Best,
_g.

- -- System Information:
Debian Release: 12.0
Architecture: amd64 (x86_64)

Kernel: Linux 6.1.0-44-amd64 (SMP w/12 CPU threads; PREEMPT)
Kernel taint flags: TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCgAdFiEEoJObzArDE05WtIyR7+hsbH/+z4MFAmo71XYACgkQ7+hsbH/+
z4OaTAgArv4QEy7D2MxZOAZnZ5hhYCFv3OdeHCHURnY9yKcqQdb5ZlYcnVwqdDgO
bsgfRetSrgV2xxdLP4yyicJZcJRuAQ9/L/FlKfAO1IDt2h6IjrVoB/AhlWhLYyj6
zYqO1gq4lUvTh6YSzs20gfnl8FR7ggiiuXXXjNPzlPeSLeFRu5rQbZ8slanTAIqm
WAal3dNZI08ih/Gyi/+Z7gJakcfHDyRhy99wBVeYr3ZoSb8ogITz+Rn0tAinRqmc
hb+pEkOLFBPrZl2tHb2DzTliQAjhCLFetBTmyWSj7mqBPC7sylQhF6ztcvaPfutY
3WS23Qd+rfq7OFkChXCWupseh6ih+Q==
=SRrH
-----END PGP SIGNATURE-----

Hi Gilles,

I was looking into this yesterday/today
I've tracked down the failures and prepared three
patches:

  * fix the Python bindings for current SWIG/Python versions to avoid an
    early FTBFS
  * relax the HDF5 version compatibility check, this will fix some
  * explicitly close the MED file in Test_MEDnStructElement, avoiding the
    HDF5 library shutdown error with HDF5 2.1.0 (manifesting as the
    infinite loop you've described below. (a quick $searchengine check
    seems that this might be a bug in HDF5.)

With these patches applied, the test suite completes successfully.

I've uploaded the packages for a testbuild to
https://debusine.debian.net/debian/developers/work-request/1090706
(ignore the FreeCAD autopkgtest failure, that's a bad newly added test)

I've also locally updated the package to upstream version 4.1.1, which is
available and can be provided as part of the update.

As the package is currently under RFA, how would you like me to proceed?
I can certainly add my name to Uploaders to share the load, but I have
no experience with HDF5 ecosystem...

I'm attaching the (three) patches as a diff to the current pacakge in
the archives.
check
_nstructelement = 2
L,T_top,F,P,P,FD,VL,VL,E,SL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,F
L,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,
FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL
,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,FL,F
L,FL,FL,FL,FL,FL,FL,FL
LANGUAGE not set