OpenBSD Journal

BSD ports collection

Contributed by Dengue on from the geez,-there-goes-the-neighborhood dept.

Both proof and Alex wrote to let me know of Chris Coleman's new crusade , a unified ports tree for the various *BSD's. I think once the novelty of the idea wears off, both a shortage of volunteers, and platform specific differences will probably sound it's death knell. Based on the comment's at Dæmonnews , it's generating more sound than fury. Don't get me wrong, it's a good idea, but the level of work required might better be spent creating new ports.

(Comments are closed)

  1. By Will Barton () on

    It sounds like a good idea on the surface -- a unified set of software for all the BSDs, but when you look deeper, there is a reason why each project maintains its own ports collection.

    FreeBSD tries to catter to the masses a lot more than OpenBSD, thus FreeBSD needs a few more applications easily availible for it than OpenBSD. Why do you need mozilla on an OS thats primarily used for Routers/Firewalls/Servers? (yes, I use OpenBSD on my laptop as well, but Im slightly odd to begin with :)

    I really don't think this project will get very far, yes you can transplant the ports collection from FreeBSD (which is really what their talking about), but it just seems wrong.

  2. By niekze () on

    i just want XMMS :) and the new version of Xevil.

  3. By Anonymous Coward () on

    Well, I don't think anyone is arguing that we should have fewer ports for OpenBSD; and while a unified ports system sounds neat - what would OpenBSD really gain from it?

    Here are some things that -I- would like to see out of ports:

    1. More ports which can be rigourously and continuously audited for security for inclusion in the base install (nothing in particular stands out; but the majority of the ports are -not- subject to the kinds of scrutiny that the base system is; so installing from them [which often happens] can potentially defeat the 'secure by default' install). 'Course if they're auditted that much, then they probably will get moved out of the ports section. ;)

    2. A CD with the source code from the ports tree - not just the tar.gz of a skeletal tree as currently included. Alternately, a complete set of pkg's on CD matching up to the ports tree @ the time of the release. This way I don't need to depend on my flakey and slow net connection to grab useful ports - I can just mount the CD. I believe that the project was attempting something similar to this, but the page has remained out of date for a while. Note, that this desire is unlikely to be tied directly to an 'official' OpenBSD project due to the fact that many ports have restrictions or licensing issues not quite suited towards distribution with the system.

    3. More ports related to other locales being expanded upon. This is mostly a personal issue as I also speak Japanese; and feel that internationalization is a very very very good goal when it comes to computers. Better yet, let's get internationalized support into the default install so that we can use OpenBSD no just in romance and Germanic languages, but non-Indo-European ones. Sanskrit, Chinese, etc. That's a huge project, and thankfully work is already going on in the ports tree now - continuing those efforts will ultimately be beneficial for the worldwide appeal of OpenBSD.

    In counter to these desires, I think there are misconceptions which come up time and time again which people should really think about with relation to ports, and the BSD's and in particular OpenBSD (seeing how that's what we're focusing on @

    Misconception i.
    OpenBSD ports are audited for security purposes.

    This is generally not true. In this case the argument that 'a unified ports system might compromise OpenBSD's security' is erroneous, or at least no more true than OpenBSD's own ports system compromising OpenBSD's security. ;) I've seen arguments about OpenBSD's ports and security before, that's why I thought it worth mentioning.

    Misconception ii.
    OpenBSD doesn't need ports due to it's security/firewall/whatever status.

    I know a lot of people just use it as a firewall - but the people doing only that with OpenBSD I personally don't really consider dedicated OpenBSD -users- so it's important to remember that while the users may admire security, they also enjoy getting lots of other things done with OpenBSD. Thus, ports become useful in some situations, and are something worth pursuing.

    Misconception iiiA.
    FreeBSD's domination in the field of ports is something to be pursued/emulated/etc.

    Guess which free OS has even -more- programs available? Does that mean it's better? No. So what if FreeBSD has 3500+ ports, do you (or even your fleet of servers and hundreds of thousands of users as the case may be) really think they're all worth running? OpenBSD has never played the game of penis waving that I've noticed, so I don't think it should start now. Already there are more ports than I know what to do with, and more added all the time. I personally haven't really felt a lack in my usage of OpenBSD, maybe I'm outnumbered. However, I've also witnessed others who -did- feel such a lack, exercise some great productivity & take a FreeBSD/whatever port and make it work under OpenBSD only to later be incorporated into the ports tree itself. This misconception sorta relates to the next (I just thought it better to flesh it into two parts)

    Misconception iiiB.
    The strengths of FreeBSD (smp, speed, large ports) correspond to weaknesses in OpenBSD.

    Like I said, this is related - but people need to sit back and remember that FreeBSD is really only targetting two (and fundamentally one with i386) architectures. That gives them an advantage in that they only have to focus their efforts in a couple of areas to get things working. While we may see smp/speed/ports as an advantage for FreeBSD. FreeBSD sees lack of architectures as a disadvantage. However, both projects are and is working to change these perceived disadvantages. Ultimately, in the case of FreeBSD - being more cross-architecture may ultimately lose them the newest features and speediest performance, so who knows if that's a win for them. Meanwhile, OpenBSD is continuing on its own path - and is actually cutting -out- architectures to an extent. I see this as a good thing, I mean we don't need to keep following NetBSD's goals; and security is doing great things for OpenBSD, and other OS's (which are taking notice) as a whole. Meanwhile, adding smp, speed, and ports helps us along too - but I for one am relieved that OpenBSD isn't in a race to be the 'fastest' or most cutting edge. It's not that it doesn't add functionality, but it does so at its own pace - when it's proven to be secure and stable, and often minimal. I like these things, I think most OpenBSD users do too. The strengths of one OS don't entail that there's a weakness in another.

    And finally

    Misconception iv.
    A unified ports system will cut down on the effort required by each BSD.

    I'm not going to set about to disprove the above, but in order for it -not- to be false; a lot of brilliance will have to take place. Having a central location for 3 separate development teams to work on ports doesn't lessen the work load for any one of them. If a program is already written to be cross-platform/compilable, then the effort required by each team currently is minimal->none - if it's not designed that way, it takes tons of effort. Somehow I doubt that having a unified ports system will change that; unless it creates (as its goal) to be a repository for highly cross-platform/architecture/etc. source code. That doesn't sound like the current goal. Moreover, if that becomes the goal [which sounds like a good one given the intended purpose] it's going to consist of a very -small- repository due to the few programs which adhere to such an agenda. Given that, it no longer becomes a place to have -all- the ports for the BSD's in one place. I think that should this take place, it's actually probably a better idea to get it off on the right foot, and start off with a small resource of highly cross platform well written ports; and then slowly expand on those so that the entire repository is consistent. That'd be my design goal at least; unfortunately, I don't think it's the goal in mind given the article.

  4. By jdube () on

    I think it's a great idea. Someone mentioned how *BSD is most used for gateways / firewalls / etc, but wouldn't it be nice to have more people running it as a desktop? Personally, I've only ever run OBSD as a desktop. I have always envied the FreeBSD ports, however. XMMS, mozilla, etc etc etc would be great additions and get a lot of support for the BSDs I am sure.

    Perhaps if we got more user friendly, BSD would catch up with Linux. This would be a Good Thing in my opinion. The fact that Loki is planning to make BSD games is amazing, and a huge step in the right direction. I know quite a few people who would switch to BSD if it had a better useability factor (by this I mean, for example, the ports tree to be bigger to nab the Debian apt-get users and perhaps even a GUI kernel config tool to get a lot of other users).

    The (from what I have seen) negative response to this must not be much different than when Linux started getting more attention. If you don't like a unified ports package (which would also make it easier for developers - make one port instead of many) then perhaps there should be a choice. If I understand correctly it is a different person working on this than the people working on the Free and Open BSD (not sure if Net has one, sorry) ports tree, so continuing development probably wouldn't be that hard.

  5. By Fisherman () on

    Idea: let the ports maintainers could chat a bit among one another, and see what really could/could not be done. I'm sure they could figure this one out wayyyy better than a bunch of posters on (this one included!).


Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]