User wishes (was: gtk-gnutella)
Hi,
On Sun, Dec 04, 2005 at 10:19:58PM +0100, Andreas Barth wrote:
> * Mike Dornberger (Mike.Dornberger@gmx.de) [051204 22:09]:
> > The last sentence seems a bit contradictorily to me. Some messages back,
> > the answer to a mail with "I'd like to see package foo in volatile" was
> > something like "Only the maintainer of foo might ask for inclusion",
> > IIRC there was even something like "Don't file wishlist bugs against bar
> > asking for inclusion into volatile". I got the impression, the users had
> > to "secretly convince" the maintainer to upload to volatile.
>
> Well, the reason why we need the maintainer's upload is:
> - bug reports reported via the debian BTS go to the maintainer,
> - this makes sure (well, at least it makes our chances pretty ok :) that
> upgrading to the next stable release works well
I fully understand this an I think it is good that way. But that's not what
I meant. I got that "secretly convince"-impression from the messages [1] -
[6].
To me, these messages say: It must be the package foo maintainer's innermost
wish to include it into volatile and that wish must not come from there,
that someone asked (in a public way) to do so. (See esp. [4].) (Maybe I
exaggerate a bit here.)
Granted, my impression may also come from the fact, that at the time these
replies have been written, there were no (written) guidelines for inclusion
available the poster could have been pointed at and that often the replies
were quite terse.
[1] http://lists.debian.org/debian-volatile/2005/06/msg00027.html
[2] http://lists.debian.org/debian-volatile/2005/06/msg00028.html
[3] http://lists.debian.org/debian-volatile/2005/09/msg00024.html
especially
[4] http://lists.debian.org/debian-volatile/2005/06/msg00032.html
[5] http://lists.debian.org/debian-volatile/2005/07/msg00016.html
[6] http://lists.debian.org/debian-volatile/2005/09/msg00031.html
There were also some messages [7] - [9] that maybe had some good points
about including one or the other package to volatile, that got no
discussion, even not one answer (at least at the d-v mailing list).
[7] http://lists.debian.org/debian-volatile/2005/06/msg00036.html
[8] http://lists.debian.org/debian-volatile/2005/07/msg00008.html
[9] http://lists.debian.org/debian-volatile/2005/07/msg00011.html
> > Maybe it is a good idea to put some sentences about user wishes to the
> > volatile website. Is a non-maintainer-initiated discussion desired here,
> > a non-DD-initiated one also?
>
> We don't mind about formal status in reviews, but rather about good
> reasons.
What have reviews to do with user wishes? Hm, I'm going to get the
impression, we don't mean the same thing when we talk about "users" and
"user wishes" here.
A quote from www.d.o/dev/d-v:
"Fellow developers are encouraged to join this discussions, so that the
debian-volatile team would know what changes the users like, and what not."
That is the only place that seems to talk remotely about "user wishes". But
it is in the "...for developers" section and not that easy to spot. So here
"users" could mean "fellow developers (that use the volatile repository)".
Or is this to be read as "fellow developers can tell the volatile team what
the users of their packages want/expect"?
In either way it seems to say (together with reading the archive; see my
statements above), it is not wanted that an end user like me, who has some
computers running Debian/sarge and helps a fried or two to get their
machines running Debian stable, asks for a package to be included into
volatile. Or better: ... that an end user [...] starts a discussion what
others think if package foo could be a good candidate for volatile - and if
the conclusion is more "yes" than "no", asks the maintainer of foo, if he is
willing to do the extra work to maintain yet another version of foo for
volatile.
But maybe I'm getting this all wrong. Maybe because of some unfortunate
wording of the website? So what do users of volatile (e. g. all those people
who have volatile in their sources.list) expect or wish?
One good point can be read in [10]. Arno van Amersfoort speaks there of a
"dependency hell" one has often to go through when using backported
packages. So maybe this could be a gap that sloppy could fill.
Hm, I thought I have read something about 'packages in volatile should only
close important (and above) bugs' but I cannot find some such wording on
either of the both volatile websites any more. (But maybe I mean the phrase
"should only contain changes to stable programs that are necessary to keep
them functional" on the v.d.n page.) Maybe that could be dropped for sloppy
and replaced with something like "should not depend on software not in
stable/main or may have one dependency to a (source-)package inside volatile
(sloppy) that does not depend on anything outside stable/main".
That could allow packages with additional features, like for example Azureus
>= 2.3.0.6 (a Java GUI BitTorrent client). This one supports the distributed
tracker and database, what comes in handy if the original tracker cannot be
reached for some reason. The version in stable (2.2.2.2) doesn't have this
feature. (I backported that one myself with nearly no problems. Replacing
the dependencies from gtk 3.1 to gtk 3.0 packages was nearly anything that
needed to be done. It has one dependency on a package not in sarge that
could be built without problems inside sarge, so that's why I added that
"... or may have one dependency ..." above. ;^) But that one could maybe
also be dropped. I have to talk to the maintainer of Azureus anyway. :-) )
[10] http://lists.debian.org/debian-volatile/2005/06/msg00031.html
Greetings,
Mike Dornberger
PS: It took me some days to write this mail. I hope that all the websites I
refer to haven't changed too much since I started.
Reply to: