[fab] future of rpm [in fedora]

Fri Nov 3 10:31:18 UTC 2006

Hi,

for a long time I have been thinking how it would be great to get so called
'soft dependencies' into rpm and actually make use of them for fedora extras.
From what I heard at the time even apt was already able to make use of them
(and thought for myself that adding it to yum would follow in time). When I
asked why we did not upgrade devel to rpm >= 4.4.3 I was told that it was to
late in the devel cycle to do that. So I stopped bothering about it for a while.
Lately I started my promised work on better repo to web interface and started
to look at the rpmlib api. That was the time I began remembering why there was
no new version of rpm in fc6 and thought that now would be a good time to get a
the newer version into devel so all issues with it can be worked out in time
for the release. So I began by looking at bugzilla and found #174307 which
really got me angry. Since then I have been thinking. Reading (rpms changelogs,
f-a-b, lwn.net and so on) and gathering information about the topic. So here
are my toughts about this evenso nobody will care I strongly feel that now is
the time for me to make a stand:

First on thing that is bothering me about the bugzilla ticket and the
resulting discussions in the various places. Paul Nasrat who this bug was
assigned to as component owner did not (up to now) respond to the bug. If I
would do so for an extras bug people would consider it rude (and they sure are
right to do so) but that might still happen because I am a volunteer. The bug
was filed on 2005-11-27 so one would think that there would have been some time
in the past ~2 years to respond in _any_ way to this. So now before the flame
starts: I don't want to accuse Paul or anybody. This is just a fact. Now from
the information I gathered I see that rh was/is having some problems with the
way jbj is handling rpm development and on how to proceed with the issue. That
is perfectly ok but why not state this clearly in the bugreport, give infos
about the state of this and so on? Fedora is supposed to be a community
project. I know that for the Core stuff other considerations (in respect to
RHEL) have to be considered and that this is a big reason why the community is
not as much involved and asked about core decisions but not being involved does
not mean that the community is not informed and kept up to date about decisions
or that nobody is responding not even with a superficial answer (or that only a
few selected non rh people know). For me this is one of the biggest reasons to
get upset about working on fedora stuff: the lack of communication that pops up
now and then about 'sensitive' topics. I really hope that we can draw a line
one of these days and that this stops.

The second thing that is bothering me about the whole upgrading/forking/... rpm
issue is this: jbj from a technical pov is really someone one would like to
have on board for working on rpm and enhancing/fixing it. I think nobody will
argue about this: jbj knows the code inside out. I don't know what issues rh
and jbj had that he got fired and that lead to the tense situation that is
there now. To tell you the truth I don't care and I think it does not really
matter in respect to the future of rpm in fedora or even to the future of rpm
as a whole. Why did the idea arise to fork rpm in the first place (as a side
note: I am using fork merely as 'fork' because I want to skip the whole 'who is
upstream rpm...' discussion)? From what I read it is because people (mostly rh
folks) are against some of the changes/enhances that newer rpm versions have
undergone. Well that is perfectly fine but why can't these people speak out
what they don't like and even more importantly speak out why they don't like
it (and saying because I like my coffee black does not count)? In addition we
still have a packaging board that could say not to use an extension ect. even
if it is ther. Just think about this for a while before you go on to read the
next block.

Ok that what do I want: I want to point out my strongest disbelieve that
forking rpm away from jbj is the right thing to do. We are, and that no matter
was is a fault of not taking action in the past 2 years, in a position where we
need to make progress and take a decision but please let it be the right one.
We are all grown up (at least more or less) and we cannot let some conflicts
that don't really have to do with rpm let as believe that a fork is that
solution we seek. There have been enough examples in the open source world where
this lead to so many problems for developers as well as for people using the
products and lastly blaming it on 'linux' and I think we all agree that we want
to draw people to linux and not push them away from it. That happens often
enough already. So yes it feels good to say that. I hope that you are still with
me and reading on and not tearing my mail apart, I would really like to be
heard and given the chance to try this:

So I am just a seemingly inactive member of FESCo and a packager with a bunch
of unimportant packages and somebody who usually does not even write mails to
lists (so think about what it means that I wrote this mail) I am offering
myself to do this:

rpm as a open source software needs to be just that: open sources software.
Nothing rh or any company has its finger on (is this a valid english analogy?).
As an os project rpm should have a _team_ of people working on it. So my idea
would be to not conflict with each other but to work together and work things
out.
How? I strongly believe that jbj would be willing to do some changes on
rpm and work on a joint effort without reservation if rh people (and fedora
community people) would do so without reservation. I don't expect anybody to
become buddies as a result of this but for the sake of rpm and all the projects
using it: lets get down on a table together and talk about this and I mean a
technical discussion. Forget or lay aside any differences that don't have to do
with rpm and start making decisions about rpm and its future. I am willing to
talk to jbj and trying to get him to sit down at this table but the question is
are you? There are so many ways the issues with new versions of rpm can be
worked out that forking seems like an easy way out for the time being but could
really be avoided.
What I would expect to come out of this? That those people gather who would
like to see rpm getting active development and form a team. A team in which
decisions are made as a team and in which people can live with (attention big
word coming up) democratic decisions. This way we can prevent another disaster
in the open source world and even show that situations like this can be
overcome if people are willing.

Ok before I go on with writing even more here (and be sure I could) I hope you
think about this before answering to my mail. I don't want this to turn into
another huge thread with no results, flames or backbiting. Stay technical and
political in an rpm project sense and work this out. I did not loose my
humor yet but before I do one last word: Please.

Andreas
-- 
Andreas Bierfert               | http://awbsworld.de      | GPG: C58CF1CB
andreas.bierfert at lowlatency.de | http://lowlatency.de     | signed/encrypted
phone: +49 2402 102373         | cell: +49 173 5803043    | mail preferred
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/fedora-advisory-board/attachments/20061103/eeab243c/attachment.sig>