Fedora Linux Support Community & Resources Center

Sections ›› Home | Forums | Guidelines | Forum Help | Fedora FAQ | Fedora News 

Go Back   FedoraForum.org > The Community Lounge > News

News Community information and the latest Fedora related news

Closed Thread
 
Thread Tools Search this Thread Display Modes
  #1  
Old 2006-12-14, 11:09 AM CST
bob's Avatar
bob Online
Administrator
 
Join Date: Jul 2004
Location: Colton, NY; Junction of Heaven & Earth.
Age: 64
Posts: 16,880
The Future of RPM - by Max Spevack

There has been a lot of discussion in the past few months about RPM -- its
present state, its future plans, and its leadership team. In particular,
the Fedora Project has received numerous requests asking us, "what are you
guys doing about RPM?"

Here is our answer, in a few words. Then if you want more, you can read
the rest of this note:

The Fedora Project is leading the creation of a new community around RPM.
One in which the leaders can come from Fedora, from Red Hat, from Novell,
from Mandriva, or from anywhere. Job #1 is to take the current RPM
codebase and clean it up, and in doing so work with all the other people
and groups who rely on RPM to build a first-rate upstream project.

==========

The Fedora Board has spoken with Fedora stakeholders both inside and
outside Red Hat, developers/maintainers in Novell, and other parties who
rely on RPM as the foundation for their distributions. We wanted to make
sure those parties agreed that this was the right thing to do for their
respective communities. We touched base with some of these people at the
recent LSB conference, and the overwhelming community opinion there was in
favor of what we are outlining here.

At the most fundamental level, we begin with two points:

(1) RPM is an important piece of technology, not just for Fedora or for
Red Hat, but for many other distributions and users. Its stability and
maintenance are critical.

(2) Red Hat realizes the need to build a strong community of contributors
around RPM, that the upstream of RPM needs to be handled in a manner which
allows contributors and developers to have maximum freedom in their
modifications, and that those modifications can be easily shared across
distributions.

Expanding on that:

(3) RPM, as a file format, is good at what it does and capable of being
the core of a Linux distribution. From the Fedora perspective, we are not
particularly interested in making any grand deviations from it at this
time.

(4) RPM, as an application, has a fairly mature feature set that we are
very interested in stabilizing and bug fixing. Furthermore, we want to
make sure that RPM is a stable and simplified base for the building of
other technologies on top of it. Down the road, we might be interested in
exploring a variety of new features, but we don't believe that should be
the initial focus of our efforts.

Ultimately, the Fedora Project and Red Hat are committed to seeing RPM be
as healthy and vibrant as many other large open source projects -- GNOME,
Xorg, etc -- consumed and contributed to by many companies, users,
distributions, and developers. Our overall goal for RPM is to ensure that
is has consistency, reliability, and stability.

We switch now to a handy Q&A format:

Q -- So what, specifically, are you doing with RPM? And where is the work
going to happen?

We have set up a new repository, wiki, and webspace -- external to any
distribution or company -- for RPM, to which anyone can contribute. A
reboot of the upstream, if you will. We don't expect that everyone will
be running the same version of RPM, or run with the same patches, but we'd
like for there to be a single place that everyone can refer to as
upstream, and be able to contribute patches.

There is already a contributor base that exists around RPM -- engineers
within Red Hat, Novell, Mandriva, and other organizations. We don't want
to leave those people behind -- we want to do a better job of
collaborating and accepting their work.

Everything will live at rpm.org, with a relaunched wiki, code repository,
and mailing lists. As for rpm.org itself, its hosting and maintainership
is outside of Red Hat, and is being generously provided by Duke
University.

Q -- How is that different from what currently exists?

What we're doing here is collecting together everyone who has a stake in
the future of RPM and building a healthy community around it. This
involves major bug fixing, development work, performance work and making
regular, predictable releases. As it stands today, we don't have these
things. This is a good first step. Could you call it a fork? Maybe. But
we're doing it because we think it's the right thing to do, for
distributions all the way down to the individual users of RPM.

Q -- Where is all this stuff going to happen? What's the public mailing
list and wiki? What *EXACTLY* is Fedora or Red Hat going to do?

Short answer -- http://rpm.org

Over the past few years, engineers from Red Hat and other companies, as
well as a community of independent contributors, have been working on and
maintaining their own versions of RPM -- sometimes sharing patches,
sometimes not. It is important that these contributions move through an
upstream process like many other projects do, in order to maintain a
healthy community and proper checks and balances.

To that end, Red Hat is adding an additional engineer that works full time
on upstream issues including patch reviews, community development, etc.
Additionally other Red Hat engineers will contribute to RPM like any other
open source project -- working on the release-engineering parts of RPM
such as rpmbuild, and doing maintenance work.

Additionally, here are some of our initial goals:

* Give RPM a full technical review, based off of RPM 4.4.2. This is the
common base for Novell and Red Hat. Look what vendors have on top of
4.4.2 and work towards a shared base. Figure out which pieces or code
paths are unnecessary, poorly implemented, or receive little to no use,
and either clean them up or clear them out. Make RPM simpler.

There's a lot of folks out there who are using RPM, including the various
Red Hat/Fedora based distros, Suse, and Mandriva, just to name a few.
Simplificaion and focus on the parts of RPM that are core to these
stakeholders is a good way to start.

* In turn, this gives us a chance to do a better job with bug fixes.
Squashing bugs that already exist, or closing out bugs that are related to
parts of RPM that are superfluous.

* Give RPM the stability that it needs to continue to be the cornerstone
of many distributions.

* Enhance the rpm-python bindings, which includes understanding and
gathering together the work that already exists in this area.

Most importantly, this work will be done in the community, fully
transparent with the help of the community and RPM stakeholders outside of
Red Hat or Fedora. This is all about incremental steps, not blowing
everything away and starting from scratch.

Q -- When is all of this happening?

Starting now. Planning and review happening over the next 3-6 months, at
rpm.org. Implementation happening appropriately alongside that planning,
as in most any free software project. Initially, Paul Nasrat is the
primary developer/maintainer dedicated to RPM from Red Hat. At the same
time, we want to make sure that leadership has a chance to develop and
emerge, rather than be mandated.

Q -- How did we end up here?

This is the part of the email in which Red Hat takes some accountability
for the current situation:

* Several years ago, the maintainer of RPM worked for Red Hat. When he
left, he continued his own work on RPM, which he acknowledges is a fork.
And that's fine -- we support anyone's right to fork, since forking is one
of the paths to innovation in open source software.

* Red Hat didn't commit the necessary resources to RPM following that
departure.

* RPM, without a strong upstream, has languished as a result.

* The community has (rightfully) been demanding that the situation be
fixed, and this is the first step in that effort.

--
Max Spevack
__________________
Linux & Beer - That TOTALLY Computes!
Registered Linux User #362651

Don't forget to comment when your problem is solved - others will be searching for solutions too!
  #2  
Old 2006-12-14, 12:26 PM CST
duanecu's Avatar
duanecu Offline
Registered User
 
Join Date: Mar 2006
Location: South Bend, IN - USA
Age: 38
Posts: 34
Will this fix the incredibly slow yum usage?
  #3  
Old 2006-12-14, 01:18 PM CST
leigh123linux's Avatar
leigh123linux Offline
Administrator
 
Join Date: Oct 2006
Location: Hampshire UK
Posts: 16,644
Quote:
Originally Posted by duanecu
Will this fix the incredibly slow yum usage?
No_____________________________
  #4  
Old 2006-12-14, 02:10 PM CST
tomcat's Avatar
tomcat Offline
Registered User
 
Join Date: Aug 2005
Location: EU, Germany
Age: 36
Posts: 1,185
Nice to see that they will work harder on rpm now.
__________________
Powered by Fedora & CentOS | Windows-free since 2002
  #5  
Old 2006-12-14, 02:52 PM CST
Firewing1's Avatar
Firewing1 Online
Administrator
 
Join Date: Dec 2004
Location: Canada
Age: 18
Posts: 9,129
I wonder if there could be a way to analyze the %post scriptlets for ldconfig. I've noticed that when you're updating many packages with so's in /usr/lib, each one runs ldconfig. Correct me if I'm wrong, but usually only one ldconfig would be needed at the end unless you're updating major packages like glibc... Which would speed up rpm a lot if we could manage to get this type of analysis working.
Firewing1
__________________
[+] My open source software and blog
[+] Some of my howtos: (for full list, click here)
  #6  
Old 2006-12-15, 12:58 AM CST
fedski's Avatar
fedski Offline
Registered User
 
Join Date: Dec 2006
Location: North Germany
Age: 38
Posts: 10
Quote:
Originally Posted by Firewing1
I wonder if there could be a way to analyze the %post scriptlets for ldconfig. I've noticed that when you're updating many packages with so's in /usr/lib, each one runs ldconfig. Correct me if I'm wrong, but usually only one ldconfig would be needed at the end unless you're updating major packages like glibc... Which would speed up rpm a lot if we could manage to get this type of analysis working.
Firewing1
Right. I can only say thats one important point.
  #7  
Old 2006-12-15, 05:36 AM CST
johannlo's Avatar
johannlo Offline
Registered User
 
Join Date: Jul 2005
Location: Melbourne, Australia
Age: 30
Posts: 744
http://linux.slashdot.org/linux/06/12/15/040258.shtml

it begins again - don't you love it when geeks duke it out over their favourite 'nix distro, m$ must be falling over themselves laughing
__________________
the phases of 'nix troubleshooting

For
| google >
Next
  #8  
Old 2006-12-15, 05:44 AM CST
CD-RW's Avatar
CD-RW Online
Registered User
 
Join Date: Nov 2006
Posts: 459
Quote:
Originally Posted by bob
Additionally, here are some of our initial goals:

* Give RPM a full technical review, based off of RPM 4.4.2. This is the
common base for Novell and Red Hat. Look what vendors have on top of
4.4.2 and work towards a shared base. Figure out which pieces or code
paths are unnecessary, poorly implemented, or receive little to no use,
and either clean them up or clear them out. Make RPM simpler.

There's a lot of folks out there who are using RPM, including the various
Red Hat/Fedora based distros, Suse, and Mandriva, just to name a few.
Simplificaion and focus on the parts of RPM that are core to these
stakeholders is a good way to start.

* In turn, this gives us a chance to do a better job with bug fixes.
Squashing bugs that already exist, or closing out bugs that are related to
parts of RPM that are superfluous.

* Give RPM the stability that it needs to continue to be the cornerstone
of many distributions.

* Enhance the rpm-python bindings, which includes understanding and
gathering together the work that already exists in this area.

Most importantly, this work will be done in the community, fully
transparent with the help of the community and RPM stakeholders outside of
Red Hat or Fedora. This is all about incremental steps, not blowing
everything away and starting from scratch.

Q -- When is all of this happening?

Starting now. Planning and review happening over the next 3-6 months, at
rpm.org. Implementation happening appropriately alongside that planning,
as in most any free software project. Initially, Paul Nasrat is the
primary developer/maintainer dedicated to RPM from Red Hat. At the same
time, we want to make sure that leadership has a chance to develop and
emerge, rather than be mandated.

Max Spevack
I agree totally with this decision. I was under the impression that an important piece of software like RPM was already maintained like this. I have bookmarked the new website, and will be watching development closely. As I'm new to Fedora (from openSuSE) I really need to learn RPM properly. Gettind rid of the un-used features would make the learning curve alot simpler. Possible a GUI like smart PM would be a help, as well as the CLI version.
  #9  
Old 2006-12-15, 09:26 AM CST
sailor's Avatar
sailor Offline
Administrator
 
Join Date: Mar 2004
Location: San Antonio, Texas
Age: 51
Posts: 3,992
This is good news...I wish that the 3rd party repos could do a little work and end the problem with mixed repositories...
__________________
sailor
Fedora 11, RHEL 5, PCLinuxOS - registered linux user #362635
Check out Linux Google,
Fedora Faq and Fedora News
Please read the Fedora Forum Guidelines
  #10  
Old 2006-12-15, 04:30 PM CST
sej7278 Offline
Registered User
 
Join Date: Sep 2004
Posts: 1,818
i don't really think there's anything wrong with rpm. yum on the other hand needs to be killed off in a big way!

everyone knows it's always been dog slow, and even the fc6 version with bits written in c is not much faster, plus you don't seem to be able to abort it with ctrl-c if it decides to just hang.

don't get me started on mixed repositories, it seems these days that you have a choice of livna on its own or dries+freshrpms, and if you mix atrpms with anything (including fedora-extras!) it seems to break.

my rpm database has been corrupted unrecoverably on two installs of fc6 as well, although i've got a feeling that is to do with the bloody applet/daemon; although it happened to me on fc4 once too iirc, never happened back in the redhat days before yum (6.1 to 9.0)

and why does updatedb and friends always decide to kick off when you're doing something cpu/io intensive - surely it should check if you're idling first before it decides to slow your machine to a crawl?

lately i've been using debian again (sarge, not that ubuntu crud) and don't like it at all, but i've got to admit that apt-get is infinitely faster, and being able to fetch from the install cd is great too.

i've been using freebsd and solaris again too actually, and at least yum seems to be better than ports and pkgadd.....

i really don't care for a gui, i never use yumex, pup or synaptic.
  #11  
Old 2006-12-15, 09:06 PM CST
b00st Offline
Registered User
 
Join Date: Apr 2005
Posts: 258
This is good news. I'm also not too happy with yum, I hope it gets replaced in the future.
  #12  
Old 2006-12-15, 11:56 PM CST
Finalzone's Avatar
Finalzone Offline
Community Manager
 
Join Date: Mar 2004
Location: Vancouver, Canada
Posts: 1,979
Yum is here to stay. The issue is related to the mirror servers and the use of python.
__________________
CPU: AMD Athlon(tm) 64 X2 Processor 4200+ Manchester overlocked to 2.61GHz
Motherboard: Gigabyte GA-K8NSC-939 nForce3
Videocard: BFG NVidia Geforce 7800GS OC
Memory: 4 x 1GB DDR-RAM OCZ EL Platinium
Storage: 1.2 TB Total HDD
Media: built-in soundcard, 16X DVD burner, 16X DVD-ROM drive
OS: Fedora 12 Constantine x86-64, Rawhide and Windows XP Services Pack 3
  #13  
Old 2006-12-16, 08:21 AM CST
marco75 Offline
Registered User
 
Join Date: Mar 2005
Posts: 35
I have used both yum and apt-rpm to interface with the RPM system on my Fedora Core 3 installation.
Apt is clearly superior; There's already a very stable GUI for it (synaptic), unlike yumex, which constantly stalls and crashes.

Rather than "enhancing rpm-python bindings", I think python is not suitable for the task. It is far too slow, resource hungry and buggy. Dump yumex and standardise to apt-rpm.

I'd like Red Hat to be much more standardized, become the "default Linux". It was the first commercially supported distro, has certificates and training courses. The other major Linux fork of interest is debian, its offspring such as Xandros, Ubuntu and Knoppix certainly leave RedHat/Fedora behind when it comes to software management.

Rather than the plethora of sys-config-* tools, I think webmin should become the standard configuration interface. It just needs a facelift (pretty RedHat icons) and some adjusting to the Fedora environment (there's no binary for FC5, only the noarch package, I think it does need a few tweaks to run 100%)
The advantage of webmin is that you can run it over a network, even from a windows workstation, and once its extensive interface is mastered, can be used to control many *nix-like operating systems.

Of course, that's just my opinion.
  #14  
Old 2006-12-16, 08:23 AM CST
marco75 Offline
Registered User
 
Join Date: Mar 2005
Posts: 35
Probably not. They want to "enhance rpm-python bindings", which means a continuation of the donkey-humping disaster that is yum/yumex.
  #15  
Old 2006-12-16, 10:41 AM CST
sej7278 Offline
Registered User
 
Join Date: Sep 2004
Posts: 1,818
webmin is horrible, jees it breaks half of the config files when it rewrites them, so they will no longer work with system-config-*

i don't think python (or mirrors) is really the problem with yum, the new fc6 version has c modules for the speedy bits, i just think maybe the logic is a bit screwed up and it can't rely on rpm for dependency resolution so has to implement that itself. it just needs rewriting from scratch or replacing.

and i couldn't care less for a gui for yum, redhat needs to concentrate the effort on getting yum working/replaced rather than some stupid gui or "enhancing rpm".

it is odd that although mandriva, suse etc all use rpm, it's only redhat that ever seems to contribute to it - hehe now that novell is effectively sponsered by microsoft, they could throw in some resources to fix yum
Closed Thread

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Fedora 7 - by Max Spevack bob News 0 2007-05-08 05:51 PM CDT
Interview with Max Spevack from the Fedora project hackmeister Fedora Focus 0 2006-06-15 08:06 AM CDT
What do you see in the future for Technology? I SEE A BRIGHT and PENGUINY FUTURE!!! zivalican_elite Wibble 4 2005-10-25 08:38 AM CDT
What do you see in the future for Technology? I SEE A BRIGHT and PENGUINY FUTURE!!! zivalican_elite Fedora Focus 1 2005-10-20 03:51 AM CDT
What do you see in the future for Technology? I SEE A BRIGHT and PENGUINY FUTURE!!! zivalican_elite Programming 1 2005-10-20 03:50 AM CDT

Automatic Translations (Powered by Powered by Google):
Afrikaans Albanian Arabic Belarusian Bulgarian Catalan Chinese Croatian Czech Danish Dutch English Estonian Filipino Finnish French Galician German Greek Hebrew Hindi Hungarian Icelandic Indonesian Italian Japanese Korean Latvian Lithuanian Macedonian Malay Maltese Norwegian Persian Polish Portuguese Romanian Russian Serbian Slovak Slovenian Spanish Swahili Swedish Taiwanese Thai Turkish Ukrainian Vietnamese Yiddish

All times are GMT -7. The time now is 10:58 AM CST.

TopSubscribe to XML RSS for all Threads in all ForumsFedoraForumDotOrg Archive
logo



All trademarks, and forum posts in this site are property of their respective owner(s).

FedoraForum.org is privately owned and is not directly sponsored by the Fedora Project or Red Hat, Inc.

Privacy Policy | Term of Use | Posting Guidelines | Archive | Contact | Founding Members
Designed By Ewdison Then | Powered by vBulletin ©2000-2009, Jelsoft Enterprises Ltd.
FedoraForum is Powered by Open Source Projects and Products
Translated to other languages thanks to NLP-er 2.3.8