Re: %20Re: Debian Maintainers GR Proposal
* Benjamin BAYART <firstname.lastname@example.org> [070624 17:11]:
> Yes. So, the right solution if I want to help is:
> - first I spend a lot of time proving that I'm skilled enough to read
> crazy licenses in a language that is not mine
There is hardly any package thinkable, where one can be sure the maintainer
will never need that skill. Most stuff needs other software and
libraries. Or at least might some day include some snipped of code from
somewhere else. Do you know any other test that would ensure people do
at least know that they have to look for in such cases?
> - then I spend another lot of time proving I'm skilled enough to package
> complex stuff unrelated to my current skills (say python stuff, which
> I know nothing about, or trying to have a library not breaking
> everything in an upgrade)
Hardly any package will always stay the same and keep within the area
of its maintainer's deep knowledge. And most packages tend to develop
their ugly corner cases of things uninteresting for the maintainer but
important for the users. So seeing if a person is capable
to learn something new is a good test. I hope that if DMs are
introduced, noone would advocate someone for that without having checked
> - then I spend enother lot of time reading complex procedures about
> voting and so on in Debian
This is a valid point. But how complex is voting?
> - then I am granted the right to help fixing the bug I found a few
> months ago
There are few things where one needs special privileges to "help" fix
a bug. Perhaps access to some porting machine (though those most of the
time tend to not have installed the dependencies currently needed, so
asking elsewhere for an account or help is often faster, anyway).
The only thing left is uploading stuff, and even for this there are
> I'm not Debian-skilled enough to package KDE, or to package OpenOffice,
Noone expects big packages to be as perfect as the little ones. The
big problems are keeping up with the sheer amount of work and being able
> I cannot claim that I'm skilled enough to do NMU on anything.
I think NMU is much more easy than to maintain a package. Only
small specific changes are allowed and all one has to do is making sure
nothing breaks. When maintaining a package, one has to understand and
finally learn all its parts in the long run.
> Those last years, when I did find something broken in Debian, I just
> updated *my* computer to work properly, which is IMO stupid, since I can
> have contributed that.
Indeed. But if it is only about fixing some stuff, just sending
patches to the bts works quite well. (Well, sometimes it needs a bit,
and it does not hurt to check the package with the patch included if
it is properly applied). Perhaps with the big packages as exception,
but those you excluded yourself...
Bernhard R. Link