Re: Taking over package into Debian Python Team maintenance and fixing bug (Was: mdp: FTBFS with numpy 1.21 (in experimental): dh_auto_test: error: pybuild --test --test-pytest -i python{version} -p 3.9 returned exit code 13)
- To: Yaroslav Halchenko <debian@oneukrainian.com>, Debian Python <debian-python@lists.debian.org>
- Cc: 999526@bugs.debian.org, Tiziano Zito <opossumnano@gmail.com>
- Subject: Re: Taking over package into Debian Python Team maintenance and fixing bug (Was: mdp: FTBFS with numpy 1.21 (in experimental): dh_auto_test: error: pybuild --test --test-pytest -i python{version} -p 3.9 returned exit code 13)
- From: Andreas Tille <andreas@an3as.eu>
- Date: Tue, 18 Apr 2023 15:22:05 +0200
- Message-id: <ZD6Zfc7GX/bYgDCt@an3as.eu>
- In-reply-to: <ZD3BO3s/FpeiGR6S@bilena>
- References: <ZD0Ojl8zfcbUPyVl@an3as.eu> <ZD3BO3s/FpeiGR6S@bilena>
Hi Yaroslav,
Am Mon, Apr 17, 2023 at 05:59:23PM -0400 schrieb Yaroslav Halchenko:
> Hi Andreas,
>
> Thank you very much for offering help. I think Tiziano would not mind,
> so please feel very welcome to a) for the sake of b) or any other
> goodness you would like to bring ;)
https://tracker.debian.org/news/1429393/accepted-mdp-36-2-source-into-unstable/
https://salsa.debian.org/python-team/packages/mdp
> Note though that MDP is pretty much inactive project since a few years
> back. It seems it is still used by some and somewhat maintained
> upstream, so might indeed be worthwhile keeping afloat in Debian but I
> would not cry if it got RMed.
We have some rdepends:
$ apt rdepends python3-mdp
python3-mdp
Reverse Depends:
Recommends: pan-machine-learning
Recommends: science-machine-learning
Enhances: python3-sklearn
Recommends: python3-pymca5
Recommends: science-machine-learning
> After/if packaging moves to a new repo on salsa, we can submit a
> PR to add an empty out debian/ and add stub debian/README to that
> upstream repo to signal that packaging moved to salsa.
See above. A finding that came unexpected after adding autopkgtest
is that 32bit architectures can't cope with 'float128', 'complex256'
data types. Since I wanted to avoid further trouble in hard freeze
I decided to restrict architectures to 64bit for the moment.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: