Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 722830 - dev-python/spyder-4.2.3: stable request
Summary: dev-python/spyder-4.2.3: stable request
Status: RESOLVED WONTFIX
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: Stabilization (show other bugs)
Hardware: All Linux
: Normal normal (vote)
Assignee: Python Gentoo Team
URL:
Whiteboard:
Keywords: ALLARCHES, STABLEREQ
Depends on: 779523
Blocks:
  Show dependency tree
 
Reported: 2020-05-13 09:27 UTC by Andrew Ammerlaan
Modified: 2021-04-18 07:40 UTC (History)
7 users (show)

See Also:
Package list:
dev-python/spyder-4.2.3 dev-python/spyder-kernels-1.10.2
Runtime testing required: No
nattka: sanity-check+


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Andrew Ammerlaan gentoo-dev 2020-05-13 09:27:59 UTC
Package seems to be stable and fully functional, I use it a lot.

Tests on this package and the dependencies that are maintained by me are passing for me.
Comment 1 NATTkA bot gentoo-dev 2020-05-13 09:44:32 UTC Comment hidden (obsolete)
Comment 2 NATTkA bot gentoo-dev 2020-05-13 09:48:38 UTC Comment hidden (obsolete)
Comment 3 NATTkA bot gentoo-dev 2020-05-13 09:56:34 UTC Comment hidden (obsolete)
Comment 4 NATTkA bot gentoo-dev 2020-07-07 20:16:54 UTC Comment hidden (obsolete)
Comment 5 NATTkA bot gentoo-dev 2020-07-07 21:50:55 UTC Comment hidden (obsolete)
Comment 6 NATTkA bot gentoo-dev 2020-07-12 17:01:14 UTC Comment hidden (obsolete)
Comment 7 NATTkA bot gentoo-dev 2020-07-13 08:01:02 UTC Comment hidden (obsolete)
Comment 8 NATTkA bot gentoo-dev 2020-07-13 08:05:42 UTC Comment hidden (obsolete)
Comment 9 NATTkA bot gentoo-dev 2020-07-13 08:45:03 UTC Comment hidden (obsolete)
Comment 10 NATTkA bot gentoo-dev 2020-07-13 08:48:57 UTC Comment hidden (obsolete)
Comment 11 NATTkA bot gentoo-dev 2020-07-13 08:52:56 UTC Comment hidden (obsolete)
Comment 12 NATTkA bot gentoo-dev 2020-08-22 07:17:13 UTC Comment hidden (obsolete)
Comment 13 NATTkA bot gentoo-dev 2020-08-22 07:34:19 UTC Comment hidden (obsolete)
Comment 14 NATTkA bot gentoo-dev 2020-08-22 07:41:51 UTC Comment hidden (obsolete)
Comment 15 NATTkA bot gentoo-dev 2020-08-22 07:53:00 UTC Comment hidden (obsolete)
Comment 16 NATTkA bot gentoo-dev 2020-08-22 08:37:23 UTC Comment hidden (obsolete)
Comment 17 NATTkA bot gentoo-dev 2020-08-22 08:54:28 UTC Comment hidden (obsolete)
Comment 18 NATTkA bot gentoo-dev 2020-08-22 09:13:05 UTC Comment hidden (obsolete)
Comment 19 NATTkA bot gentoo-dev 2020-09-18 15:57:06 UTC Comment hidden (obsolete)
Comment 20 NATTkA bot gentoo-dev 2020-09-18 16:01:08 UTC Comment hidden (obsolete)
Comment 21 NATTkA bot gentoo-dev 2020-09-18 16:09:15 UTC Comment hidden (obsolete)
Comment 22 NATTkA bot gentoo-dev 2020-09-24 07:02:45 UTC Comment hidden (obsolete)
Comment 23 Joonas Niilola gentoo-dev 2020-09-24 07:05:34 UTC
Do we really need to stabilize spyder? It'll cause a lot of work for everyone. I feel like the nature of this package doesn't really live well in stable systems anyway.
Comment 24 NATTkA bot gentoo-dev 2020-09-24 07:06:31 UTC Comment hidden (obsolete)
Comment 25 Andrew Ammerlaan gentoo-dev 2020-09-24 08:24:08 UTC
> I feel like the nature of this package doesn't really live well in stable systems anyway.

Why not?

Almost all of the dependencies are stable now, the 4.1.4 -> 4.1.5 bump introduced some new dependencies though and that is why this is now failing the sanity-check. In fact this is the reason why I did not remove the 4.1.4 version in my PR, as it was very close to becoming stable. But it seems someone else removed that version from the tree and now I have to wait 30 days before asking for stabilization of the new dependencies introduced in the 4.1.5 bump.
Comment 26 Joonas Niilola gentoo-dev 2020-09-24 08:42:56 UTC
(In reply to Andrew Ammerlaan from comment #25)
> 
> Almost all of the dependencies are stable now, the 4.1.4 -> 4.1.5 bump
> introduced some new dependencies though and that is why this is now failing
> the sanity-check. In fact this is the reason why I did not remove the 4.1.4
> version in my PR, as it was very close to becoming stable. But it seems
> someone else removed that version from the tree and now I have to wait 30
> days before asking for stabilization of the new dependencies introduced in
> the 4.1.5 bump.

Exactly. If it were stable, it couldn't have been removed and it'd slow down the cleaning process of python project. And there seem to be humongous amount of deps not stabilized, and some of the deps seem to have truly ancient versions stabilized. Overall I feel like the stabilization process of this will cause huge amount of work to everyone, not just to you and p-m project. There are other projects involved. Just weigh the advantages and disadvantages. Even this stabilization bug has been going on for 4 months now...

You should wait for actual users to ask for stabilization. Not all of my packages are stable either. And looking at deps here, it'd be easier to run ~ instead of editing package.accept_keywords so I doubt anyone running this is on stable system.

And overall I just feel like it's too error-prone with all the deps.
Comment 27 Joonas Niilola gentoo-dev 2020-09-24 09:01:00 UTC
I just want to make it clear that in the end you decide. If you are yourself on a stable system, then that's a use case.

I'm just looking into areas where we are constantly behind, and spending 4 months on a stabilization bug where there are possibly 0 stable users doesn't help anyone in Gentoo. Then having multiple new packages with no prior stable being forced to go through another stabilization round before being able to clean older versions. And then how long will it take when 5 comes out... Just think about the long-term effects this has, overall, not just on spyder. Is it worth it? Although now it looks like you're closer to the target.

Have you yourself ran tests/spyder on a stable system so that the "downgraded" deps actually work?
Comment 28 Andrew Ammerlaan gentoo-dev 2020-09-24 09:24:20 UTC
> And there seem to be humongous amount of deps not stabilized

Well yes, there is after the 4.1.5 bump because I finally managed to sort of fix the test phase. As it is now, if you install this with USE="-test -doc" the only package.accept_keywords entries needed are spyder, spyder-kernels, qdarkstyle and qtawesome. The last two of those have a very infrequent release cycle so I think it's worth stabilizing those two at least. But I guess I'll close this as there is no point in keeping it open anyway until those new pytest deps are stable.

> And overall I just feel like it's too error-prone with all the deps.

And this is precisely why I copy the dependencies exactly as specified by upstream into the ebuild. It is the only way to make sure everything works without having to test compatibility with every single version of everything. And it is why I'm getting a bit annoyed with bugs such as these: https://bugs.gentoo.org/743466 https://bugs.gentoo.org/715290 . I completely understand that it is not preferable to have deps on exact versions or on older versions, but if upstream specifies it that way in their setup.py, then presumably they have a reason for that. And I do not have the time to test the whole thing to see if it maybe is compatible with a newer version of foo/bar, despite upstream suggesting that it is not. And if I do test it against the newer version of foo/bar I usually find that it is indeed not compatible.

> Even this stabilization bug has been going on for 4 months now...

That is my fault, I had put the whole thing in one request. And only later I realized that that is not the best nor most efficient way to do it. I should have started with requesting stabilization on some of the smaller deps with infrequent release cycles and slowly moved my way upward, instead of doing the whole thing in one go.

> Have you yourself ran tests/spyder on a stable system so that the "downgraded" deps actually work?

Well no, I haven't done that (yet) as I was waiting on the dependencies to be stabilized.

I'll close this for now, and maybe come back to it after the 5.0.0 version bump, and after the new pytest and sphinx deps are stable.
Comment 29 Andrew Ammerlaan gentoo-dev 2020-09-24 09:49:17 UTC
> If you are yourself on a stable system, then that's a use case.

Just to add, back in the day when I first started using Gentoo, I was one of those users who wanted to use spyder on a stable amd64 system. To accomplish this I had a *very* large package.accept_keywords file, and at some point I just gave up and switched to the unstable ~amd64 arch. And this is something I think we should avoid, users shouldn't be impelled to move to the unstable arch just because they want to use some package foo/bar. And this is exactly why I opened this stabilization request in the first place, and to some extent it has been successful as the amount of lines required in package.accept_keywords has already been reduced to 4.
Comment 30 Joonas Niilola gentoo-dev 2020-12-15 12:31:07 UTC
I guess it's go-time again?
Comment 31 Andrew Ammerlaan gentoo-dev 2020-12-15 12:34:54 UTC
> I guess it's go-time again?

Not yet, we need the new python-language-server and plugins to be stable first (and the 30 day waiting period hasn't passed yet for those packages)
Comment 32 NATTkA bot gentoo-dev 2020-12-22 16:49:43 UTC Comment hidden (obsolete)
Comment 33 Sam James archtester Gentoo Infrastructure gentoo-dev Security 2021-03-02 08:18:13 UTC
If we're doing this - and I'm happy to - let's ALLARCHES the heck out of everything please where applicable?
Comment 34 Andrew Ammerlaan gentoo-dev 2021-03-02 08:37:49 UTC
(In reply to Sam James from comment #33)
> If we're doing this - and I'm happy to - let's ALLARCHES the heck out of
> everything please where applicable?

Sure, will do
Comment 35 NATTkA bot gentoo-dev 2021-03-20 08:29:06 UTC Comment hidden (obsolete)
Comment 36 NATTkA bot gentoo-dev 2021-03-20 10:49:06 UTC Comment hidden (obsolete)
Comment 37 NATTkA bot gentoo-dev 2021-03-20 11:01:14 UTC Comment hidden (obsolete)
Comment 38 Thomas Deutschmann (RETIRED) gentoo-dev 2021-03-30 16:13:37 UTC
dev-python/spyder appears untestable:

> !!! Multiple package instances within a single package slot have been pulled
> !!! into the dependency graph, resulting in a slot conflict:
> 
> dev-python/pytest:0
> 
>   (dev-python/pytest-4.6.11-r1:0/0::gentoo, ebuild scheduled for merge) USE="-test" PYTHON_TARGETS="python3_8 (-pypy3) -python3_7 -python3_9" pulled in by
>     <dev-python/pytest-5.0[python_targets_python3_7(-)?,python_targets_python3_8(-)?,python_targets_python3_9(-)?,-python_single_target_python3_7(-),-python_single_target_python3_8(-),-python_single_target_python3_9(-)] required by (dev-python/pytest-faulthandler-1.6.0:0/0::gentoo, ebuild scheduled for merge) USE="-test" PYTHON_TARGETS="python3_8 -python3_7-python3_9"
>     ^                  ^^^
> 
>     (and 1 more with the same problem)
> 
>   (dev-python/pytest-6.2.2-1:0/0::gentoo, installed) USE="-test" PYTHON_TARGETS="python3_8 (-pypy3) -python3_7 -python3_9" pulled in by
>     >=dev-python/pytest-5[python_targets_pypy3(-)?,python_targets_python3_7(-)?,python_targets_python3_8(-)?,python_targets_python3_9(-)?,-python_single_target_pypy3(-),-python_single_target_python3_7(-),-python_single_target_python3_8(-),-python_single_target_python3_9(-)] required by (dev-python/pytest-mock-3.5.1:0/0::gentoo, ebuild scheduled for merge) USE="-test" PYTHON_TARGETS="python3_8 (-pypy3) -python3_7 -python3_9"
>     ^^                  ^
>
Comment 39 Andrew Ammerlaan gentoo-dev 2021-03-31 07:53:56 UTC
(In reply to Thomas Deutschmann from comment #38)
> dev-python/spyder appears untestable:
> 
> > !!! Multiple package instances within a single package slot have been pulled
> > !!! into the dependency graph, resulting in a slot conflict:
> > 
> > dev-python/pytest:0
> > 
> >   (dev-python/pytest-4.6.11-r1:0/0::gentoo, ebuild scheduled for merge) USE="-test" PYTHON_TARGETS="python3_8 (-pypy3) -python3_7 -python3_9" pulled in by
> >     <dev-python/pytest-5.0[python_targets_python3_7(-)?,python_targets_python3_8(-)?,python_targets_python3_9(-)?,-python_single_target_python3_7(-),-python_single_target_python3_8(-),-python_single_target_python3_9(-)] required by (dev-python/pytest-faulthandler-1.6.0:0/0::gentoo, ebuild scheduled for merge) USE="-test" PYTHON_TARGETS="python3_8 -python3_7-python3_9"
> >     ^                  ^^^
> > 
> >     (and 1 more with the same problem)
> > 
> >   (dev-python/pytest-6.2.2-1:0/0::gentoo, installed) USE="-test" PYTHON_TARGETS="python3_8 (-pypy3) -python3_7 -python3_9" pulled in by
> >     >=dev-python/pytest-5[python_targets_pypy3(-)?,python_targets_python3_7(-)?,python_targets_python3_8(-)?,python_targets_python3_9(-)?,-python_single_target_pypy3(-),-python_single_target_python3_7(-),-python_single_target_python3_8(-),-python_single_target_python3_9(-)] required by (dev-python/pytest-mock-3.5.1:0/0::gentoo, ebuild scheduled for merge) USE="-test" PYTHON_TARGETS="python3_8 (-pypy3) -python3_7 -python3_9"
> >     ^^                  ^
> >

Last time I did this I think I used --nodeps to emerge dev-python/pytest-mock, which is far from a perfect solution.

I should just update this stable request to a newer version of spyder. The newer ebuilds have RESTRICT="test" because the test suite is too flaky (see also the in-ebuild comments)
Comment 40 NATTkA bot gentoo-dev 2021-04-01 19:15:43 UTC Comment hidden (obsolete)
Comment 41 NATTkA bot gentoo-dev 2021-04-01 20:13:11 UTC Comment hidden (obsolete)
Comment 42 NATTkA bot gentoo-dev 2021-04-01 22:09:25 UTC
All sanity-check issues have been resolved