Arch Hurd?

I have always been interested in the GNU Hurd. This probably stems from the endless discussions on Slashdot about how (in my interpretation) microkernels should be full of awesome but none have really managed to obtain the greatness that they deserve. I always thought the status of Hurd was so far off being useful that there was no point in looking into it further. However, I recently read the Hurd status page and there was a picture of a GUI, doing useful spreadsheet type stuff.

My interest was piqued… Combining that with the joys of building a cross-compiler for an operating system or architecture you do not actually have access too (yes, I am a sad, sad person) and you get a Hurd cross compiler. I built a few packages and even managed to get (a slightly patched) pacman built. Then, having wasted much time, I moved on.

Several months pass and there is a post on the Arch forums, with someone trying to compile a GNU operating system for themselves. I mentioned my previous endeavours and somewhat surprisingly others seem interested in the possibility of making a Hurd distro. Well, Arch users are a weird bunch…

And so, Arch Hurd was born. There is a website, so there is no stopping now! The current status is a bunch of scripts that create a quite up-to-date cross-compiling toolchain (glibc-2.10.1, binutils-2.19.1 and gcc-4.4.2), which can be used to build the GNU Mach kernel, the Hurd, coreutils and bash (the latter two being more updated than the versions in Arch!). That is not far from a minimally bootable (but completely useless) system. Then we can all bask in the microkernally goodness.

12 thoughts on “Arch Hurd?

  1. Pingback: uberVU - social comments

  2. Wow, i’ve never used Hurd but great job, seems to be very interesting and shows how Arch can be adapted to other kernel than linux.

  3. Pingback: Arch Hurd « Seamos realistas…

  4. I was always interested in microkernels too, but I don’t really like the Hurd because, well, it’s GNU (meaning: the opposite of KISS principles for many things). Also, it destroyed my Linux partitions last time I tried to install it, but it was the first release of Debian GNU/Hurd so it has probably improved since. I saw a demonstration at Libre Software Meeting 2008 and I was quite impressed with the “everything is a filesystem” approach, especially with XMLFS.

    I still think that the short term future for microkernels is Minix 3 rather that Hurd, because it’s simpler and I would say more functional. I tried to port Pacman to it earlier but I have to give up because of time issues, maybe I will give it a try again later (and your port to Hurd could prove quite useful – actually the most difficult part is not really Pacman itself but libalpm).

  5. I’m glad to see your effort :-)

    Except by the fact hurd doesn’t have a lot of drivers (compared with linux) and there is no sound subsystem like ALSA, Hurd can be used well (with some performance issues, off course).

    Did you send a e-mail to hurd-devel list? they will be very happy to see your initiative and, in my opinion, you will have a lot of help from there.

    I’ll try to help you too.

    BR,
    Carlos.

  6. @Carlos Costa: I was looking around the Hurd mailing lists today and, from what I can see, there is work on updating the driver glue code they’re using so it’ll work with Linux 2.6 drivers, as well as stuff from the BSDs.
    Hopefully we’ll have access to more or less all of the hardware with a Linux driver soonish :)

  7. Pingback: 2010 sera nerdz : Arch Hurd — Hippopota.me/blog : le blog de deK

  8. Pingback: Spork Bomb :: Arch Hurd

  9. Good luck with the project!

    Some time ago someone started with gentoo-hurd but sadly the project is dead. I hope you can ‘finish’ the project!

  10. Here are some good news from the Hurd :)
    Thanks for this project, i hope installing it soon and provide graphisms if I can and if I have time.
    Good continuation !