Mailing List Archive

Dale posted on Sun, 02 Oct 2011 06:49:04 -0500 as excerpted:

> I know the KDE3 folks subscribe here and wanted to let you know about
> this. I got this after I done my sync:
>
> Building database (/var/cache/eix) ..

> [1] "kde-sunset" /var/lib/layman/kde-sunset (cache:
> parse|ebuild*#metadata-flat#assign)
> Reading category 17|155 ( 10%): app-misc .. *
> *
> "/var/lib/layman/kde-sunset/app-misc/releaseforge/
releaseforge-1.3.ebuild":
> * Deprecation Warning: Usage of distutils.eclass in packages not
> supporting installation
> * for multiple Python ABIs in EAPI <=2 is deprecated and will be
> banned on 2011-06-01.

The short version of the story is that there have been some changes and
that message can be ignored, for now.

The longer version... well, let's say that a certain former gentoo/python
project dev is just that, former, and that the disruptively high rate of
python eclass code churn and function deprecation and removal has slowed
WAAAYYY down. It'll probably pick back up slightly as others pick up the
pieces and development gets back to normal, but it should remain far more
modest and less disruptive than it was, as that wasn't normal (or, it
turns out, acceptable for the rest of gentoo) either.

Similarly, it's rather unlikely that the methods used to stabilize
python-3 will be repeated any time soon. (Heh, at least they didn't move
python-2 out of the tree to a python-sunset overlay... but the
circumstances there are quite different as unlike kde/qt, upstream python
continues to support python-2, and unlike python, not so much of the rest
of the tree is so widely affected by kde. The kde thing was more
upstream policy, the python thing more a gentoo-only thing.)

That's why the named date has passed without any action, but the warning
remains. Everything's more or less frozen in place as the folks picking
up the pieces assess the situation and decide where to go from here.
From what I've read, many of the changes have already been done in the
python overlay, and they have to figure out whether to move them to the
tree, but slower, or to revert them in the python overlay, or if a
different solution needs to be worked out and tested in the python
overlay, before moving to the tree.

(FWIW, I believe said developer remains active in the python overlay,
helping to straighten things out or at least explain where he was going
with the changes so what to do with them can be better worked out, and he
does still post to gentoo-dev occasionally, of course now without a gentoo
address, so he's still active in the community and I guess helping to
pick up the pieces. What happens from here on out, however, remains to
be seen. Maybe he'll return as a dev, having learned his lesson, perhaps
he's not even interested, but at least he's still around, so AFAIK
there's still some bridges remaining unburned.)

So, um, yeah, the deprecation message can be ignored for now. It'll
probably be removed at some point. If they do decide to kill that
functionality, etc, expect some further warning and likely discussion on
the dev list (which I know both you and I follow) first.

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman

Duncan wrote:
> Dale posted on Sun, 02 Oct 2011 06:49:04 -0500 as excerpted:
>
>> I know the KDE3 folks subscribe here and wanted to let you know about
>> this. I got this after I done my sync:
>>
>> Building database (/var/cache/eix) ..
>> [1] "kde-sunset" /var/lib/layman/kde-sunset (cache:
>> parse|ebuild*#metadata-flat#assign)
>> Reading category 17|155 ( 10%): app-misc .. *
>> *
>> "/var/lib/layman/kde-sunset/app-misc/releaseforge/
> releaseforge-1.3.ebuild":
>> * Deprecation Warning: Usage of distutils.eclass in packages not
>> supporting installation
>> * for multiple Python ABIs in EAPI<=2 is deprecated and will be
>> banned on 2011-06-01.
> The short version of the story is that there have been some changes and
> that message can be ignored, for now.
>
> The longer version... well, let's say that a certain former gentoo/python
> project dev is just that, former, and that the disruptively high rate of
> python eclass code churn and function deprecation and removal has slowed
> WAAAYYY down. It'll probably pick back up slightly as others pick up the
> pieces and development gets back to normal, but it should remain far more
> modest and less disruptive than it was, as that wasn't normal (or, it
> turns out, acceptable for the rest of gentoo) either.
>
> Similarly, it's rather unlikely that the methods used to stabilize
> python-3 will be repeated any time soon. (Heh, at least they didn't move
> python-2 out of the tree to a python-sunset overlay... but the
> circumstances there are quite different as unlike kde/qt, upstream python
> continues to support python-2, and unlike python, not so much of the rest
> of the tree is so widely affected by kde. The kde thing was more
> upstream policy, the python thing more a gentoo-only thing.)
>
> That's why the named date has passed without any action, but the warning
> remains. Everything's more or less frozen in place as the folks picking
> up the pieces assess the situation and decide where to go from here.
> > From what I've read, many of the changes have already been done in the
> python overlay, and they have to figure out whether to move them to the
> tree, but slower, or to revert them in the python overlay, or if a
> different solution needs to be worked out and tested in the python
> overlay, before moving to the tree.
>
> (FWIW, I believe said developer remains active in the python overlay,
> helping to straighten things out or at least explain where he was going
> with the changes so what to do with them can be better worked out, and he
> does still post to gentoo-dev occasionally, of course now without a gentoo
> address, so he's still active in the community and I guess helping to
> pick up the pieces. What happens from here on out, however, remains to
> be seen. Maybe he'll return as a dev, having learned his lesson, perhaps
> he's not even interested, but at least he's still around, so AFAIK
> there's still some bridges remaining unburned.)
>
> So, um, yeah, the deprecation message can be ignored for now. It'll
> probably be removed at some point. If they do decide to kill that
> functionality, etc, expect some further warning and likely discussion on
> the dev list (which I know both you and I follow) first.
>


This is good. I just didn't want to not post and it cause someone
problems with say a new install. I posted issues I had with a new
install of KDE3 a while back too. I just wanted to post in case someone
was unaware of the issues I ran into. They could have been caused by me
but they could have been something else too. I posted to let some dev
that knows more than me can sort it out.

At least I know all is well. Sounds like python 3 is taking a bit of a
nap.

Dale

:-) :-)