Hello Paul, I think that I found the clue: Python3.10 interprets the following lines -------------8<------- file src/einsteinpy.egg-info/requires.txt ---- [:implementation_name == "cpython"] numba!=0.49.0,>=0.46 -------------8<------------------------------------------------------
no exactly as Python3.9 did, and the result of ${python3:Depends} includes
python3-numba as a consequence.
So I shall test another modification: wiring the Python3 dependencies,
and declaring python3-numba as a recommendation.
Best regards, Georges.
Paul Gevers a écrit :
> Hi,
>
> [Why not in the bug report? Feel free to quote me.]
>
> On 07-05-2022 15:44, Georges Khaznadar wrote:
> > Paul Gevers a écrit :
> > > Where are you seeing this? I only see failures:
> > > https://ci.debian.net/packages/e/einsteinpy/
> >
> > The tests are still failing for Salsa/CI; however the bug 1010215 was
> > different: it was genuinely due to the wrong usage of the symbol np
> > (defined after "import numpy as np"), instead of the string "numpy".
> > This parameter finishes by reaching sympy.lamdify, and the current
> > version of sympy does no longer tolerate this wrong usage.
> >
> > Please take a look at the begin of Bug#1010215's report for more
> > information.
> >
> > I took the freedom to patch einsteinpy's test syntax because sympy was
> > marked for AUTORM if nothing was done for a few weeks.
> >
> > Besides, you are true: now, the failed tests report a broken dependency
> > on python-numba3, which is not yet compiled with/for Python3.10.
>
> Your NMU einsteinpy grew a *new* dependency on python3-numba, which has RC
> bug 1000336 for a while already, so don't expect it to migrate. Hence, your
> fix can't migrate to testing as-is.
>
> Paul
--
Georges KHAZNADAR et Jocelyne FOURNIER
22 rue des mouettes, 59240 Dunkerque France.
Téléphone +33 (0)3 28 29 17 70
signature.asc
Description: PGP signature
