Discussion:
Bug#938157: python-scipy: Python2 removal in sid/bullseye
(too old to reply)
Sandro Tosi
2020-01-13 02:00:02 UTC
Permalink
Package: src:python-scipy
Version: 1.2.2-7
Severity: normal
Tags: sid bullseye
Usertags: py2removal
we finally reach the point were src:python-scipy produces only leaf
binary packages (excluding packages that are not in testing because RC
already), so i think it's time to file for its removal, and continue
with src:scipy.

I've copied maintainers and Drew (whos been doing all the work
recently); if i dont receive a NACK within a week, i'll file for the
removal of src:python-scipy.

Regards,
Sandro
Drew Parsons
2020-01-13 04:40:01 UTC
Permalink
Post by Sandro Tosi
we finally reach the point were src:python-scipy produces only leaf
binary packages (excluding packages that are not in testing because RC
already), so i think it's time to file for its removal, and continue
with src:scipy.
I've copied maintainers and Drew (whos been doing all the work
recently); if i dont receive a NACK within a week, i'll file for the
removal of src:python-scipy.
Thanks Sandro. There is RC bug#946624 affecting python-scipy, with a
counterpart in #946625 for python3-scipy. Something subtle has changed
in the syntax for skipping tests, I haven't figured out what yet. Help
needed.

Since that problem needs to be fixed anyway for python3-scipy, I propose
dealing with that first before decommissioning python-scipy. I expect
that whatever the fix is that's needed for python3-scipy, the same patch
will fix the tests for python-scipy. That way we can put python-scipy
to sleep on a clean bed.

Drew
Scott Kitterman
2020-01-13 05:10:02 UTC
Permalink
Post by Drew Parsons
Post by Sandro Tosi
we finally reach the point were src:python-scipy produces only leaf
binary packages (excluding packages that are not in testing because RC
already), so i think it's time to file for its removal, and continue
with src:scipy.
I've copied maintainers and Drew (whos been doing all the work
recently); if i dont receive a NACK within a week, i'll file for the
removal of src:python-scipy.
Thanks Sandro. There is RC bug#946624 affecting python-scipy, with a
counterpart in #946625 for python3-scipy. Something subtle has changed
in the syntax for skipping tests, I haven't figured out what yet. Help
needed.
Since that problem needs to be fixed anyway for python3-scipy, I propose
dealing with that first before decommissioning python-scipy. I expect
that whatever the fix is that's needed for python3-scipy, the same patch
will fix the tests for python-scipy. That way we can put python-scipy
to sleep on a clean bed.
I don't follow this at all. It sounds to me like you are suggesting not
removing a package that is ready to be removed due to python2-rm until it can
be fixed. Any bugs on the python2 package are gone as soon as it is. Please
don't delay removing packages in order to fix them first.

Scott K
Scott Kitterman
2020-01-13 05:30:02 UTC
Permalink
Post by Drew Parsons
Post by Drew Parsons
Thanks Sandro. There is RC bug#946624 affecting python-scipy, with
a
Post by Drew Parsons
counterpart in #946625 for python3-scipy. Something subtle has changed
in the syntax for skipping tests, I haven't figured out what yet.
Help
Post by Drew Parsons
needed.
Since that problem needs to be fixed anyway for python3-scipy, I propose
dealing with that first before decommissioning python-scipy. I
expect
Post by Drew Parsons
that whatever the fix is that's needed for python3-scipy, the same patch
will fix the tests for python-scipy. That way we can put
python-scipy
Post by Drew Parsons
to sleep on a clean bed.
what is the benefit of fixing python-scipy, when we can remove it
right now? i feel like we could just focus on fixing python3-scipy
and
use the time/effort for python-scipy for something else :)
It's just that there is no extra effort required in this case.
python3-scipy does indeed need the effort taken, but once that effort
is
done, it's no more effort to apply the same fix to python-scipy. So
yes, please do focus on fixing python3-scipy! :)
Yes, maybe so, but delaying the python2-rm blocks other work and thus delays and complicates getting that major effort completed. Let's not.

Scott K
Drew Parsons
2020-01-14 02:20:01 UTC
Permalink
Post by Scott Kitterman
Post by Drew Parsons
It's just that there is no extra effort required in this case.
python3-scipy does indeed need the effort taken, but once that effort
is
done, it's no more effort to apply the same fix to python-scipy. So
yes, please do focus on fixing python3-scipy! :)
Yes, maybe so, but delaying the python2-rm blocks other work and thus
delays and complicates getting that major effort completed. Let's
not.
exactly, i really dont understand what's the value of fixing something
that's gonna be removed anyway.
I think i'm just gonna wait a couple more days and then file a RM
Fair enough then. I'm out-voted :)

Loading...