1 |
On Sat, 2020-03-07 at 18:49 +0100, Ulrich Mueller wrote: |
2 |
> > > > > > On Sat, 07 Mar 2020, Michał Górny wrote: |
3 |
> > Ebuilds. 183 of them. One is stuck on py2 but is included as only |
4 |
> > revdep. |
5 |
> |
6 |
> Just the ebuild being outdated doesn't sound like a sufficient reason |
7 |
> for removal of a package, at least not for those packages that install |
8 |
> applications for the end user. |
9 |
|
10 |
The list is almost exclusively about dev-python/, i.e. packages that do |
11 |
not install end-user applications but Python modules. |
12 |
|
13 |
> > But that's by no means all ebuilds like that, just a subset Python |
14 |
> > team doesn't see much of a point maintaining. |
15 |
> |
16 |
> Why had they been added then, in the first place? |
17 |
> |
18 |
|
19 |
You should ask the person who added them. Some of them probably used to |
20 |
have revdeps in the past but lost them (either because they switched to |
21 |
other deps or were removed). Some of them were added because somebody |
22 |
used it at some point. Some of them were added because someone thought |
23 |
it would be great idea to package a lot of Python modules because we |
24 |
can. |
25 |
|
26 |
Some of these someones have retired since. Some left the Python team. |
27 |
Some weren't ever part of it yet dumped packages on us. |
28 |
|
29 |
I could go on like this for much longer but what's the purpose? |
30 |
The point is, python@ has a lot of packages, we can't maintain them all. |
31 |
These packages weren't really maintained for at least a few months, so |
32 |
dropping them lets us focus on packages that do have dependencies or |
33 |
otherwise seem more useful. |
34 |
|
35 |
I mean, surely, we can try to test ~300 packages on py3.7 just to |
36 |
discover half of them were added without tests, large number have |
37 |
failing tests, some have silly mistakes... |
38 |
|
39 |
-- |
40 |
Best regards, |
41 |
Michał Górny |