-----BEGIN PGP SIGNED MESSAGE-----
On 08/06/2011 01:24 PM, Jorge Manuel B. S. Vicetto wrote:
> On 06-08-2011 11:59, Markos Chandras wrote:
>> On 08/06/2011 12:49 PM, Fabian Groffen wrote:
>>> On 06-08-2011 11:00:16 +0100, Markos Chandras wrote:
>>>> Oh come on Jorge. You know what I mean by slacking arches. I am
>>>> not talking about punishing them. Maybe drop stable keywords
>>>> or drop keyword from X package and shrink their tree so they
>>>> can keep up with the load.
>>>>> This should be the arch more developers use daily and is
>>>>> likely the one with more members (herd count). Also, one
>>>>> should remember the time it takes to compile, test or debug
>>>>> an issue in a recent amd64 system or an old / slow box with
>>>>> an "exotic arch" varies substantially. Not to mention that
>>>>> the amount of testing done on "exotic arches" varies
>>>>> substantially between projects.
>>>> I am aware of the problems and this is way I want a solution.
>>> And what solution do you have in mind (in your Council role)?
>> Drop stable keywords for certain arches and/or remove their
>> keywords from X packages. The idea is to keep only a single and
>> smaller portage tree which would be much more easier to manage.
>> Pretty much the same situation as MIPS. It worked pretty well on
>> MIPS, so it will work on these arches too
> If you talk to Mike, Raúl and Matt, I'm sure they'll tell you that
> trying to get an arch out of testing status to supported is a
Why would you wanna do that? The situation is highly unlikely to change
in the future. It is better to have a fully working testing tree than
pretending to have a usable stable tree. I would accept a solution were
only @system is in stable and everything else is in ~testing.
> AFAIK that was already done once for arm and is happening now for
> About the arches with >200 stable bugs, should we have dropped KDE
> from the tree when for years we had >300 open bugs? Should we stop
> doing releases if we have >100 open bugs? What about mysql with >50
> open bugs, etc?
You can't really compare an architecture with a package can you?
> Furthermore, there have been many complaints from arches with a high
> number of stable bugs that by the time they were working on a stable
> bug, a maintainer either dropped the version they were testing or
> somewhere else in the tree someone decided to drop their keywords
> and got their tree broken.
We can't keep old ebuilds around forever just because an arch decides to
act on these bugs after 10 months. I'd say it again, I don't blame them,
they have real lives too.
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.18 (GNU/Linux)
-----END PGP SIGNATURE-----