The OpenPKG project released version 2.0 of their unique RPM-based cross-platform Unix software packaging facility. OpenPKG 2.0 consists of 473 selected (from a pool of 700) packages which include latest versions of popular Unix software like Apache, Bash, BIND, GCC, INN, MySQL, OpenSSH, Perl, Postfix, PostgreSQL, Samba, Squid, teTeX and Vim — all carefully packaged for easy deployment on 16 different Unix platforms, including FreeBSD 4.9/5.2, Debian GNU/Linux 2.2/3.0/3.1, Red Hat Linux 9, Red Hat Enterprise Linux 3, Fedora Core 1, SuSE Linux 8.2/9.0, and Sun Solaris 2.6/8/9/10. Highlights in OpenPKG 2.0 are the upgrade to RPM 4.2.1, custom RPM extensions for more complete packaging, new platform product and technology identification, usage of UUIDs, etc.
Cross-platform packaging facility OpenPKG 2.0 released
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I notice UnixWare is listed as depreciated.
yea, this is what we need. a 68th package manager. the freedom of choice is what brought me to *nix, but i think the community goes to far when it comes to choice. how about a defacto of no more then 30 different choices… sound good? i give this project a year.
FreeBSD has ports. Why does it need something like this?
And I suppose the 600 different text editors, 100 different antivirus software, 200 different persoal firewalls, 400 web browsers and 150 email clients for MS Windows also bother you? What about the 300 different ice flavors and the 500 different kinds of chocolates in the supermarket? Or the 300 different kinds of cars?
i think he means crucial things like… installers. are there 300 different installer packeages on windows?
anyway, this thing *might* be interesting. but what is needed is not a cross platform installer, we need a cross distribution installer.
subject says it all…
They are using the mainstream package manager RPM
and their project is already successfully since over
3 years, so you perhaps should rethink your statements…
“are there 300 different installer packeages on windows?”
InstallShield (Pro, Express, insert more versions here), Inno Setup, WISE, NSIS, MSI, MS Cab Extractor, WinZip Self-Installer, Paquet Builder, pure zip files, custom self-written installers, etc. etc. etc.
Zaphod:
i think he means crucial things like… installers.
and ice cream isn’t crucial? don’t get me started on chocolate chips…
but seiriously i think that OpenPkg is interesting in that it could save a lot of effort wasted repackaging software for different distros.
Funny thing is, ALL those installers have exactly the same end user effect. Next next finish, in some variation, installs software *sucessfully*. Then in 90% of the cases that software is in the installed software list and can easily be removed in a similar fashion.
The problem in Linux, isn’t so much the installers, but the distributions. Even considering one package format, like rpm, you likely cannot sucessfully install it and have it work right unless the rpm is designed for your distribution. Also rpm does not have dependency resolution, which is needed to avoid dependency hell. Sure other tools like apt4rpm do, but even then you still have the distro specific package problem. Each distribution needs it’s own curated package set.
This is what end users want to go away. A standard way of packaging programs so that any distribution can install it properly. Maybe autopackage will help. Remains to be seen.
why do people keep asking for packages to be compatible with all distros? its like asking for these packages to be installable too, in OSX, BSD or UNIX! people should realize that every linux distro should be considered as a different OS.
we got portable code so why not aim for portable packages?
Although I think choice is a very good thing in the OSS world, I can see you’re underlying meaning.
OSS tends to re-invent the wheel when ever a developer has a slightly different idea than the previous developer. The problem with that concept is there are no true standouts among the different projects that would promote a defacto standard. There are a few that come close, but even they have their drawbacks that prohibit them from pushing ahead of the pack.
So naturally the anwser for distros providers and OSS sofware vendors is to promote as many of them as possible in hopes that at least a few will become breakout projects. Unfortunately, many never do and they all remain part of a large group of, what appears to be, uncolaborating OSS projects who’s goals are to provide a better way of doing the same thing that the others are trying to do, thus creating the misconception of choice overload and lack of organization.
Eventually though, a project will break out from the rest and become a world class project. That’s why competition is good. We as end users have to be patient and focus on what works best for us and provide feedback to those projects we think will be the most benificial to us.
If you ever have to manage an heterogeneous network with several flavors of UNIX, from Solaris to various Linux distros and flavors of BSD, you can trust me when I say that you will greatly appreciate the value of a multi-platform packaging system.
I don’t know OpenPKG much, but I’m a developer of pkgsrc (http://www.pkgsrc.org/) which started as the packaging system of NetBSD.
Using the same tools and the same source tree for all your packages, whatever the underlying OS is, means two things:
1. You don’t have to learn about specificities of a given distro or OS reguarging configuration and such
2. A given package will have the same behaviour on every OS. PHP, PERL and such will be compiled with the same options on all the supported OSes.
Is the value of portable packaging systems such as OpenPKG and pkgsrc clearer now?
Those are some good reasons but there are more.
I think the main thing all of us are annoyed by at some stage is only having the choice of:
-binary packaged for another distro
-source
Neither are convenient and contribute to the complexity and sometimes difficulty of installing things.
I think to say each distro should be considered a different OS is silly.
Choice is good but only… _ONLY_ if it is standardised and the choices are different implementations of that standard.
Currently a more popular distro may be chosen for the package support, not because it is better.
But no distro has all packages of everything.
Choice is good, but distro cant/should not be “standardized’. That was the point I was raising. People expect all the distros to be just the same.If distros are all the same then where is the choice? If it is “standardized” would the only difference left with distros is the GUI appearance (wallpaper/icons)? Because one cant tinker on the inner workings of the OS because if they do they risk breaking the “standard”. Another example, if an NT app does not work in XP or vise versa do you cry foul? No because you treat NT and XP as different OSes. Hence my point, if everyone treats distros as different OSes then one will accept why this and that is not available for his/her distro.
**when a say “standard” I am referreing to insanity’s difinition of it.**