[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: task & skills

On 00-12-03 Martin Michlmayr wrote:
> * Adrian Bunk <bunk@fs.tum.de> [20001203 00:01]:
> > When I see that a package with missing build dependencies (except
> > debhelper) that has in debian/copyright a "Copyright: not yet known"
> > and 2 other bugs is enough to become a Debian maintainer
> > (http://lists.debian.org/debian-newmaint-discuss-0011/msg00103.html)
> > I can't see any quality standards that are needed.

> The NM you're talking about knew about the missing Copyright and also
> knew that he would never upload such a package to woody.  He had

Well, why didn't he then fixed this first before applyin to new
maintainer? There is in my opinion no reason to create a debian package
of a software if it's missing a copyright.

> The Build-Depends: I didn't count but I guess about 80 - 90% get it

Why? There are lists where you can look for you package and check what
build-dependencies it has. Also I wonder how you can guess that the
number is so high? I think it would be only about 10-20& judging from
looking at the bug reports.

> wrong.  I point them to the manual, tell them why it's necessary, tell
> them about setting up a chroot environment (which, BTW, is not
> documented well at all).  And I have them fix it.  There are packages

And that should be the part that the sponsor should do. We have the
concept of sponsorship for new maintainers, so that they have a debian
developer as guide who checks there packages and explains all the stuff
to them, so that they know how to produce a high quality debian package.
In my opinion it's not the job of the application manager to fix
packages. If a new maintainer has no sponsor, he should first get one
assigned, who will guide him and check the packages. After he passed
this test, he should be processed by the NMs and not earlier.

> in Woody from experienced developers with wrong build dependencies.

What do you want to state with this sentence? You forgot that it's
easier to send a mail to an experined developer telling him his package
has wrong build-dependency then sending such a mail to a new maintainer.
The experienced developer knows what's wrong and how we has to fix his
package, while you have to explain all the stuff to the new maintainer.

> When a build depdencency is broken, they will sooner or later get a
> bug report, and then fix it (I remember filing a bug against one of

If they are not lazy and just ignoring it. 

> your packages, Adrian -- removing a unneeded build dependency is not a
> nice thing either, especially when you build depend on the package you
> want to build).

Sorry, but this is lame. Did you ever take a look how much stuff Adrian
is doing for debian and qa in debian? If not, I would suggest that you
first examine this before blaming him for such small bug.

> I know it's all in the manuals, but it's a lot to learn for NMs.  Why
> don't you write better documentation?  e.g. a "check list before

Because it's not needed. I know some people in the NM queue personally
and I can tell you, that we don't need more documentation, but a bit
more sponsoring and willing new maintainers. The people I know in the NM
queue read the documention and if somehthing is unclear to them, I
explain it and then undersand it and fix their packages. And this just
happens, because they are willing to learn and apply for new maintainer,
to contribute somehting back to the distribution that they are using and
not because they just want to have a nice new email-address.

> uploading my package" which I proposed earlier in this thread (and to

This is not needed since there's lintian and a lot of sponsors who can
check the rest of the package for bugs, that lintian won't catch.

          Debian Developer and Quality Assurance Team Member
    1024/26CC7853 31E6 A8CA 68FC 284F 7D16  63EC A9E6 67FF 26CC 7853

Attachment: pgpm6O8O7Mqw8.pgp
Description: PGP signature

Reply to: