[gentoo-dev] Re: Some ideas on how to reduce territoriality

[Steve Long]

Tiziano Müller wrote:

> Chris Gianelloni schrieb:
>> - arch-specific patches/dependencies - If someone is requesting KEYWORD
>> changes on a package and it requires a patch or additional dependencies
>> for your architecture, you are not only permitted, but really are
>> required to make the necessary changes to add support for your
>> architecture.
> And what is going to happen with the patch? Should go upstream, but
> who's responsible for that?
>
Er, the maintainer; if s/he's not bothered about the package compiling on
different archs, should s/he really be maintaining it? I doubt upstream
would appreciate that from a distro -- and it's not hard to file a quick
bug with a link to the gentoo one; a quick comment on the gentoo one and
any interested users can help upstream to triage it.

>> - metadata.xml changes
> With limitations.
>
Maintainer sounds like a definite no-no. Any others?

>> - Version bumps where the only requirement is to "cp" the ebuild
> Just "cp"'ing the ebuilds is the reason that so many ebuilds are still a
> nightmare and full of little nasty bugs.
>
> This is a complete no-go since there are so many things a careful
> maintainer has to consider (besides checking the packages changelog, the
> dependencies, the license, the docs, etc. he should also check the
> ebuild).
>
Yeah but if they can compile it and it works as an app (however that's
defined, this /is/ usr-land) what's the harm in bumping and allowing others
to test it? (This is unstable, I hope..) If they can't be bothered to do
that, how can they possibly claim to be testing it? (And why are they even
touching it if they're not interested? ;) [I dunno how make test fits into
this, either, as tests were broken for synfig.]

If the maintainer doesn't like it, well s/he's already got the new version
working on one arch/ machine (plus whichever user bugged the dev.) I can't
see anyone really bemoaning a new tester ;P


-- 
gentoo-dev@g.o mailing list