After seven months of comprehensive development, the popular Unix software packaging tool RPM Package Manager (RPM) was released as stable version 5.0.0. This version builds portably on a wide variety of Unix platforms and includes initial support for XAR as a package format, among other features.
Post a Comment
Member since:
2005-07-06
Member since:2005-10-11
Member since:2005-07-01
RE: Another question, what is exactly the XAR format?
Why xar is interesting
http://code.google.com/p/xar/wiki/whyxar
Member since:2005-07-24
Member since:2006-08-01
Member since:2006-03-14
Redhat could always change their minds. They have shown that they are not averse to doing so in the past (see inclusion of mono).
If it turns out to be much better, whilst maintaining compatibility with older rpms, then it is not inconceivable that they could adopt it.
Member since:2005-07-24
"""
Redhat could always change their minds. They have shown that they are not averse to doing so in the past (see inclusion of mono).
"""
There is no mono in RedHat. Perhaps you are thinking of Fedora.
At any rate, it would be a mistake, and a step backwards, to allow such a critical component to be controlled by such an uncooperative developer. There was a reason that RedHat terminated Mr. Johnson. And there was a reason that Fedora spearheaded forking the project.
The XFree86->Xorg fork, painful as it was, has proven highly beneficial. In the same way, I believe that this fork will also result in long term benefits to the community.
Member since:2006-08-01
> If I remember correct it was somekind of fork?
Jeff Johnson RPM : http://rpm5.org/ .
Red Hat/Fedora/SuSE/Mandriva RPM : http://www.rpm.org/ .
One of the reasons of Jeff Johnson fork :
https://bugzilla.redhat.com/show_bug.cgi?id=119185
Member since:2005-07-06
The full story:
http://lwn.net/Articles/196523/
- Gilboa
Member since:2005-08-20
Different side of the story: http://www.kuro5hin.org/story/2006/6/5/101431/9311
While I seriously doubt that Jeff wrote that, it's still a funny read.
Marcus
Member since:2005-07-06
Member since:2005-08-18
Member since:2005-11-02
This is very funny. I just tried this with a Debian box that has read only /usr and... aptitude update worked, install failed, remove failed. No package db inconsistencies were introduced. Installing and removing after /usr was remount,rw worked normally.
Gotta love RPM!
Member since:2005-11-12
wow, reading through that bug was kind of painful. It is unfortunate that people respond in a work environment with nastiness. There is no need for that. Johnson makes a reasonable argument that typing 'yum update' should not leave your system in a fragmented state if only part of a package is installed. I would think transactional accounting would be a requirement for updating a system. It makes me relieved that I got out of the software industry. It is annoying when good ideas get blocked based on egos and bad middle management.
Edited 2008-01-06 16:05
Member since:2005-07-24
.
Member since:2005-07-06
I don't think you understood what was Johnson said (and why he got canned.)
The reporter claimed that the RPM state must always be consistent. RPM should have failed gracefully when trying to install a package on RO storage. (instead of being corrupted - read: stating that a certain application is installed when it isn't)
Johnston claimed that if the user is stupid enough to try and install RPMs on a read only storage, the RPM is not required to keep a consistent state when things get busted.
Software 101 - no matter how stupid your user get, your application must not crash and/or stay in an inconsistent state.
- Gilboa
Member since:2005-07-06
Member since:2005-07-06
EDIT: Checked who you are.
As a general rule, before you start threatening people with law-suites (I assume that this is the next logical step - once you finish sending your cease and desist messages) you should -really- check where they live and learn something (or two) about their local law.
Somehow your response here doesn't surprise me a bit given your past behavior.
- Gilboa
Edited 2008-01-07 12:53
Member since:2005-07-06
Just so Mr. "Cease and desist"'s lawyer won't chase me into oblivion, please replace:
"(and why he got canned.)" with
"(and why, he was allegedly [according to certain sources] got canned)"
- Gilboa "hiding under a rock; switching off the Internet; changing my forwarding address to Iraq" Davara.
Member since:
2006-03-20
Member since:2006-11-24
Member since:2006-08-01
Member since:2007-07-31
Member since:2007-07-23
Since Linus forked it 
I think the tongue in cheek part was missed by 1 or 2 people here. Plus 1, back to 0.
On topic: The Red Hat/Fedora/SuSE/Mandriva kind of RPM has never failed us. I am not sure if I fully understand the benefits of this RPM5, but if it adds something, why not?
Cheers.
Member since:2008-01-07
For details of what rpm-5.0 adds, try reading the rationale
for adoption of rpm-5.0 by OpenPKG at
http://trainofthoughts.org/blog/2008/01/06/rpm5-vs-rpm/
Member since:2006-01-03
Member since:2006-10-08
"Well, rpm is available out-of-the box on AIX..."
And AIX is a UNIX. Another UNIX is FreeBSD where RPM is available, too.
But I may ask a question. The article's description is "RPM Package Manager (RPM)". The FreeBSD ports collection lists RPM as "The Red Hat Package Manager". So, is RPM the correct abbreviation for this title or is it an recursive acronym (RPM = RPM Package Manager)?
Regarding UNIX vs. Linux, see linear algebra: the "is a", "is part of" and "is like a" identity, subset and similarity relations. :-)
Member since:2007-12-28
Member since:
2005-07-06
What is the official version of RPM?
Is it...
http://www.rpm.org/
http://rpm5.org/
http://wraptastic.org/blog/
Member since:2006-01-16
the official rpm is at: http://www.rpm.org/
rpm5 is a fork.
Member since:
2006-01-17
A fork, which might catch up if a disto embraces it. (e.g. if someone at Red Hat/Suse/Mandriva notices). A problem might be that it was removed RPMv3 package support which is a LSB requirement (but that can be handled by a separate RPM version even in the same system).
Now I believe deb/dpkg has advantage because it doesn't have N forks and support dist-upgrades (to upgrade to the new upstream version), while rpm's big advantage IMHO is multi-arch package support. So both have their set of advantages and drawbacks.
Maybe big distros should overcome the NIH syndrome and concentrate on creating a new common package standard (what ever happened to deb2 proposal?) as a base for their systems, because now both are very similar. I think big point should be to add to it a system for signed 3rd party and per-user (installed for a local user(s)) packages, a la ZeroInstall.
Member since:2005-07-06
Member since:2007-03-28
Member since:2005-07-06
Member since:2005-07-24
Member since:2006-12-28
Dude, seriously, lay off!
As far as I and many others are concerned, AdamW is a well respected contributer to this forum. Your sad attempts to sound like you have any legal muscle when it comes to people points you disagree with is pathetic.
Your not making any friends pretending you have some weight to throw around. This is an online forum, if you don't like whats being said, go somewhere else.
Member since:2005-08-18
Member since:2005-07-06
Member since:2008-01-07
There's a bit more to rpm-5.0 than links to kuroshin and
bugzilla #119185. The issues are several years old if
you bothered to look at times and dates.
Why do you think I care what rpm version Fedora, SuSE, Mandriva
or any other vendor distribution uses? rpm5.org is now a full fledged,
vendor neutral, OSS project. That was never the case when I worked
@redhat.com, and (if you look at the membership and roadmaps
at the"official" rpm.org site) is very much not the case now.
For those interested in the differences between rpm.org and rpm5.org,
I again point to the rationale behind OpenPKG's decision to assist
setting up infrastructure for rpm5.org:
http://trainofthoughts.org/blog/2008/01/06/rpm5-vs-rpm/
Member since:2005-08-18
Member since:2005-07-24