Another Summer of Code project


From a commenter on a previous post: Gentoo has a Google Summer of Code project porting portage to DragonFly, by student Naohiro Aota.  I had no idea this was happening – this is interesting!

Posted by     Categories: DragonFly, Google Summer of Code     12 Comments
12 Comments on Another Summer of Code project

Respond | Trackback

  1. Anonymous says:

    Portage is a great thing, it’s much better in handling upgrades and stuff like that than pkgsrc. But i would better like Nix (http://nixos.org/nix/) for dragonfly.

  2. [...] Zobacz reszt? artyku?u: Another Summer of Code project · DragonFly BSD Digest [...]

  3. Chris Turner says:

    Please no!

    This is a total waste of time!

    It might work, perhaps – and then we’ll get package fixes fragmented all over the place, and it will likely be abandoned if it is ever used.

    I’d much rather see someone fix broken packages on dragonfly.
    or improving pkgsrc, if they are interested.

    Do these people even contemplate *mentioning* it to DF lists..
    to see what the DF COMMUNITY is most interested in?

    ARRGH!

    ok. off me soapy-box. best of luck to them.

  4. Petr says:

    Personally I think I would probably use portage instead of pkgsrc. There’s issues with pkgsrc that have been there for ages like updating packages that simply do not get enough developer attention (for whatever reason). I upgrade packages couple of times a year and its always pain with pkgsrc. I think DF should go the Solaris way: stick to one desktop (preferrably KDE, not Gnome) and make it work real well, but still allow for people to install something else (and thus risk breakage) via pkgsrc or portage, or even a home-brewed package manager.

  5. DEVELOPMENT ELEPHANT says:

    I also think this is a wast of time, it would be much more productive to create a pkgsrc tool that works in a similar fashion to the PC-BSD PBI builder, and give an option of installing packages with isolated dependencies.

  6. Anonymous says:

    2Petr
    That’s way to go if you dream of “conquering the desktop market” or some shit like this, but DF is not doing it.

    2DEVELOPMENT ELEPHANT
    PBI is totally retarded idea.

    Of course it *is* waste of time, it’s gentoo we are talking about. They already have portage/FreeBSD and portage/NetBSD, and portage/MacOSX and what not, and nobody is using it. So if they want to waste some time on DF port, it just doesn’t hurt.

  7. Petr, I’d point out that packages not getting attention is the same in portage as in pkgsrc – I’d have to see some hard numbers comparing update rates before I thought pkgsrc was any worse off. Anecdotally, the material I’m using out of pkgsrc tends to follow newer versions even faster than I can upgrade to them, which is fine with me.

    Chris – Summer of Code projects don’t have to be useful; the goal is learning. Even if it isn’t used, the student will know more about packaging software. There would be a negative effect if we started thrashing between packaging systems as a project, or something like that, but that’s not happening.

    I’m mildly optimistic that the DESTDIR changes in pkgsrc are going to get rid of the remaining problems with pkgsrc upgrading; past that, there are aspects of pkgsrc that are really appealing – cross-platform, with regular stable releases, plus there’s mechanisms to create complete binary builds. I don’t think you find that in ports or portage.

  8. blinkkin says:

    It’s interesting to see how pkgsrc and portage makes to other platforms. Portage is ported mostly by Gentoo developers, while pkgsrc is picked up by users of particular opereting system. Good example of this is GSoC 2010. MINIX has pkgsrc project this year and Gentoo Foundation chose porting Portage to DragonFly BSD.

    Has pkgsrc more cons than Portage? Probably documentation isn’t so good and infrastructure is more dirty. Which of these two has more success in adaptations? I would dispute, especially when you can see a presentation like this (pkgsrc and OpenSolaris).

    Personally I see pkgsrc more as framework than package manager. Most of the existing solutions tries to integrate with system as much as possible. Pksrc is quite opposition – just look on Haiku with mostly non-POSIX environment. Pkgsrc is working even there – quote from pkgsrc 2010Q1 announcement: “Haiku port is mostly done”.

    Sorry for poor english quality, it isn’t even my second language.

  9. Lazarus says:

    Is the world not a better place for having more than one type of tree? Apples are fine, pears are nice, but you orange loving mofos can go to hell?

    I’d like to point out that it is no more a waste of time than people working on six or more different BSD based OSes or the countless Linux distributions or the countless other Unix-like or non-Unix like OSes in existence.

    This person is pursuing an interest and who are we to question what he does with his time and energy.

  10. DEVELOPMENT ELEPHANT says:

    I take back my earlier statement, live and let live, every one should be encouraged to participate in open source projects even if the goals are in competition with your own.
    Thank you Lazarus for reminding me of this.

  11. Petr says:

    Justin,
    I was referring to the process of updating packages installed on your system. Hell, pkgsrc can’t even update itself, let alone 3rd party apps like DEs and media players etc. Every time there is an update to core pkgsrc packages (bmake, pkg_install etc) I have to remove them manually and then re-bootstrap. It’s nice that pkgsrc fetches dependancies and applies patches and that it runs on every possible platform and architecture but realistically pkgsrc is only usable only on small servers with handful of packages installed, but desktop and beyond? its simply unworkable. You can’t have your workstation with 300+ packages installed down for a week or more because you only wanted to update one application, but pkgsrc decided to remove every package, and a couple of the new packages fail to compile resulting in TOTAL ABSOLUTE MESS.

    Before moving to BSD years ago I was Gentoo user for a while. Some/lot Portage packages might be broken initially, but Portage’s usability is simply skyhigh above pkgsrc/fbsd ports.

  12. Hi there, I run Gentoo’s GSoC program. I just tripped over this page when searching for gentoo summer of code.

    We’re happy to let students pursue the ideas they propose, as we’re all about what gets them excited. If a student really wants to work on Gentoo and chooses to experiment with a new BSD-based port, we think Gentoo will benefit from increasing the generality of our existing code as well as gaining a new contributor. Our primary goal at the end of the summer is gaining new developers. My assumption is that if DF were a student’s primary interest, that’s where he would apply. Please let me know if you have any questions; I’m pretty easy to find on the Internet.

Respond

Comments

Comments