Bugs with patches attached

Micah Cowan micahcowan at ubuntu.com
Wed Apr 9 19:41:56 BST 2008


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mackenzie Morgan wrote:
> On Wed, 2008-04-09 at 18:47 +0100, James Westby wrote:
>> The further concern is that when
>> they are seen we ask too much of the submitters and
>> so drive people away from contributing.
>
> One problem I have is although I could write patches for a couple bugs
> I've spotted in, for example, Universe packages, I don't know how to
> make a debdiff (or even what one is), but when you have a patch, that's
> what they ask for.

And, indeed, it's a complicated process.

Not only does one have to know enough to generate the original patch in
the first place, but with a debdiff, you now have to know how to
integrate that patch with whatever flavor of patching system is being
used in that particular source package (quilt? dpatch? dbs?) if there is
one.

You also have to get familiar with debuild to build the source package,
and then, quite likely, pbuilder to test it with (which takes some
amount of setup, first).

In short, it's a great investment for those who expect to be making many
such patches (say, package maintainers, and folks specifically
interested in becoming active in Debian (and derivatives) development),
and far less so for people who just want to get their freaking fix in. :)

> Making it easier to separate patch-writer from packager would, in my
> opinion, be helpful.

The trouble, particularly on the Ubuntu side of things, is that there
are so few developers in relation to the number of packages. Nobody
generally has the time to take on responsibility for properly packaging
someone _else's_ patch.

This would be helped perhaps by improving Ubuntu tendencies to forward
patches upstream. This is already fairly good, but could (to my memory)
be improved. There needs to be a widespread understanding that patches
should be checked against upstream first thing (current dev versions of
course), and if appropriate, left in their care (but tracked in
LaunchPad too, so it doesn't get lost if they don't rush to deal with
it); when that's not appropriate, forwarding to Debian probably is.

Upstream maintainers -> Debian -> Ubuntu, follows decreasing numbers of
resources, in general: best to try with the folks with the most, first.
The downside, of course, is that Ubuntu users must then wait, not only
for upstream to decide to apply the patch, but for it to trickle down to
them from Debian, as well. Still, for at least those packages that have
no real maintainers in Ubuntu, that seems like the best approach.
Critical patches may be an exception.

- --
Micah J. Cowan
Programmer, musician, typesetting enthusiast, gamer,
and GNU Wget Project Maintainer.
http://micah.cowan.name/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFH/Q3z7M8hyUobTrERAuuqAJ4/C1aeEKdbf19F6CDAvT4e1+a8dwCfZLHo
ZPAthSdTUulyevq2FgJeMR8=
=uKhh
-----END PGP SIGNATURE-----



More information about the Ubuntu-bugsquad mailing list