[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [linux-distros-dev] A bit of a reboot
|
Andrew Overholt wrote:
Reasons for existence / Goals
=============================
We started the project after a few different distribution maintainers
found that they were complaining about the same things (builds being
difficult, lack of autotools support in CDT, package maintenance
difficulties, etc.). The goals at the time were two-fold: 1) be a
place for distributions to collaborate and 2) provide tools specific to
Linux developers and packagers.
I think the reasons are more valid, and more reasonable as Eclipse
grows. Though perhaps 1) and 2) don't converge as much as they should.
I'm guilty of completely ignoring 1) in favour of 2). This is partly
because people already do tasks associated with 1) regarding ChangeLog
and I don't have too. I prefer to ignore it; guilty as charged ;) To be
honest, I prefer updates sites to RPMs as it more the "Eclipse" way
(imo). I wonder if these is a way to mesh them together?
Current status
==============
We have some Linux-specific tools that are doing well: the ChangeLog
plugin and the RPM specfile editor.
I've had a lot of fun hacking on ChangeLog and this project has provided
a good solid safe-haven from the beginning. In that sense, and in my
own eyes, this project has been a complete success.
Unfortunately, little cross-distro collaboration has actually taken
place. I'd like to think this is because we all have lots to deal with
in real life and in our respective distros. What can do to improve
here? Since we all fight the same battles with builds and stuff, it
would be good to work together and not duplicate effort.
I've been periodically involved with Eclipse over several years, and
packaging has evolved constantly throughout that time, but not in a
consistent manner. That is not a criticism - it is to be expected as it
mirrors how Eclipse has grown. Eclipse has grown constantly but not
consistently.
To be sure, distribution packaging requires a lot of investment of time
to do a "good enough" job, and a considerable investment of time do an
excellent job. I think it is not something you can pick up for a few
hours a week and do. However, the work that has been going on in the
last two years has moved what was a delicate, complicated and tricky job
to a more consistent approach - at least in Linux. It is easy to be
disillusioned, the improvements to process and tools are sometimes only
apparent to those that work - or have worked on them. But from a point
of view as a code contributor, the folks working on 1) are doing a super
job. It takes a knowledge of history to know just how much better
things are.
But you are right, now this work has been done, it's good to look at the
cross-distribution collaboration issues.
I think we need to do a few things:
1. Make an update site for the tools we have.
Agreed. I still have my old one hosted where ChangeLog was originally
developed:
http://people.redhat.com/pmuldoon/
If I remember correctly, it was all done via a wizard, and it wasn't too
hard. (on another note, I should take that site down ;)
2. Have an actual release! Then we won't be the bad Eclipse project who
never really followed the processes :)
A Linux tools based release separate from the distribution releases
would be neat. Then we could follow a feature path instead of a Fedora 9
or Ubuntu release path. The big drawback with distribution release paths
is the functionality of the different Eclipse Linux Tools is not
consistent across distributions.
We need to have a plan in place soon so I'll be working on something
more formal than this email. I really welcome all feedback so please
let me know whatever you're thinking -- even if it's "you're stupid" ;) .
Thanks for asking, and starting the conversation!
Regards
Phil Muldoon