Bug#1069745: magics-python: wrong arch: any packaging builds potentially uninstallable packages
Hi Alastair, Am Thu, Apr 25, 2024 at 08:33:58AM +0100 schrieb Alastair McKinstry: > Yes, thanks for this. You are perfectly welcome. Its relieving to work in a cooperative team (compared to my experiences in DPT). ;-) > I'm moving all packages I have in science-team to Science Maintainers. That's fine and even helps making your packages covered by some Blends tools. In case I might stumble upon a package of yours that is not yet maintained by Science Maintainers I take this as a sign I can safely do so when I feel some other work on this package is needed. Kind regards Andreas. > regards > > Alastair > > On 24/04/2024 07:23, Andreas Tille wrote: > > Control: tags -1 pending > > thanks > > > > Hi Alastair, > > > > as far as I learned in past example cases its rather by chance that > > you did not set > > > > Maintainer: Debian Science Maintainers > > > > > > in packages that are residing in science-team space on Salsa. I took > > the freedom to fix this in Git as well as applying the patch for this > > bug + routine-update (doing several tiny polishing changes). > > > > I hope you don't mind if I'll upload those changes this evening. > > > > Kind regards and thanks for maintaining this package in Debian Science > > team > > Andreas. > > > -- > Alastair McKinstry, > GPG: 82383CE9165B347C787081A2CBE6BB4E5D9AD3A5 > e: mckins...@debian.org, im: @alastair:mckinstry.ie > > -- https://fam-tille.de
Bug#1069745: magics-python: wrong arch: any packaging builds potentially uninstallable packages
Hi Andreas Yes, thanks for this. I'm moving all packages I have in science-team to Science Maintainers. regards Alastair On 24/04/2024 07:23, Andreas Tille wrote: Control: tags -1 pending thanks Hi Alastair, as far as I learned in past example cases its rather by chance that you did not set Maintainer: Debian Science Maintainers in packages that are residing in science-team space on Salsa. I took the freedom to fix this in Git as well as applying the patch for this bug + routine-update (doing several tiny polishing changes). I hope you don't mind if I'll upload those changes this evening. Kind regards and thanks for maintaining this package in Debian Science team Andreas. -- Alastair McKinstry, GPG: 82383CE9165B347C787081A2CBE6BB4E5D9AD3A5 e: mckins...@debian.org, im: @alastair:mckinstry.ie
Bug#1069745: magics-python: wrong arch: any packaging builds potentially uninstallable packages
Control: tags -1 pending thanks Hi Alastair, as far as I learned in past example cases its rather by chance that you did not set Maintainer: Debian Science Maintainers in packages that are residing in science-team space on Salsa. I took the freedom to fix this in Git as well as applying the patch for this bug + routine-update (doing several tiny polishing changes). I hope you don't mind if I'll upload those changes this evening. Kind regards and thanks for maintaining this package in Debian Science team Andreas. -- https://fam-tille.de
Bug#1069745: magics-python: wrong arch: any packaging builds potentially uninstallable packages
Package: magics-python
Version: 2:1.5.8-1
Severity: important
Tags: patch
User: ubuntu-de...@lists.ubuntu.com
Usertags: origin-ubuntu noble ubuntu-patch
Hi Alastair,
Working to resolve per-arch uninstallability of python3-magics++ in Ubuntu
for the upcoming release, I found significant issues in the packaging that
should be resolved.
- The package build-depends on libeccodes-dev, but does not use it
- It depends on libmagplus3v5, but nothing ensures that this package is
only built for architectures on which libmagplus3v5 is available (both
libmagplus3v5 and libeccodes have *similar* portability issues, but not
identical; in Ubuntu we're now in the situation that we have architectures
where libeccodes-dev is available but libmagplus is not)
- It doesn't actually *use* libmagplus3v5, so this dependency is wrong;
what it does use is libMagPlus.so, from libmagics++-dev, which is resolved
via python3-ecmwflibs
- But the package is missing an actual dependency on python3-ecmwflibs, so
installing python3-magics++ by itself results in a python module that
fails to import
- But also, this package is a pure-python module containing no
architecture-dependent code, so it should be Architecture: all anyway
instead of Architecture: any (which also makes the per-arch
installability issues go away)
- And because it's architecture-independent it should build-depend on
python3 - not python3-dev, which is for binary modules.
Please find attached a patch addressing these issues, which has been
uploaded to Ubuntu.
Thanks,
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer https://www.debian.org/
slanga...@ubuntu.com vor...@debian.org
diff -Nru magics-python-1.5.8/debian/control magics-python-1.5.8/debian/control
--- magics-python-1.5.8/debian/control 2022-04-17 00:12:44.0 -0700
+++ magics-python-1.5.8/debian/control 2024-04-23 17:10:54.0 -0700
@@ -4,8 +4,7 @@
Maintainer: Alastair McKinstry
Build-Depends: debhelper-compat (= 13),
dh-sequence-python3,
- libeccodes-dev,
- python3-dev,
+ python3,
python3-setuptools,
python3-pytest-runner
Standards-Version: 4.6.0
@@ -14,8 +13,8 @@
Vcs-Git: https://salsa.debian.org:/science-team/magics-python.git -b
debian/latest
Package: python3-magics++
-Architecture: any
-Depends: libmagplus3v5 , ${misc:Depends},
+Architecture: all
+Depends: python3-ecmwflibs, ${misc:Depends},
${python3:Depends},
python3-simplejson,
python3-jinja2
