IRC channel logs


back to list of logs

<jxself>I suppose it's time to compile more kernels while I sleep and see how things go.
***tschwing_ is now known as tschwinge
***Svetlana|2 is now known as gry
<civodul>Hello Guix!
<svetlana>Hello. Could you remind me where the guix www is located? I'd like to translate the front page (again). I don't have notes from my past take at that.
<civodul>let us know if you have any comments/suggestinos
<svetlana>Oh, but there was a git repo for that somewhere.
<svetlana>Suggestion was to start using gnun. But uh. I ought to finish 1 page -- the front page -- before you put effort into changing the infrastructure.
<civodul>if you want you can convert the page to GNUN
<civodul>i'm not sure exactly what it takes, so your help would definitely be welcome
<svetlana>What git repo is the page in though?
<civodul>it's actually a (cough...) CVS repo
<svetlana>Oh, that is ok.
<svetlana>In my head, I use "git" as a meta term for "git, svn, cvs, ..." set of things. I'm in the wrong there, quite likely.
<civodul>you can get the details by going to and then "source code" -> "web pages"
<svetlana>Thank you!
<svetlana>Do you know what fsf uses to host its websites?
<svetlana>(Guix website? A special data center? VPS?)
<rekado_>civodul: do you happen to know if there's a public repository for IXIN?
<davexunit>good morning, my avant-garde friends.
<rekado_>: )
<mark_weaver>heh :)
<mark_weaver>so, updating xorg-server required updating a bunch of the *proto packages, which in turn forced a rebuild of the X libraries and everything on top of it.
<mark_weaver>so, I figured I might as well update everything in xorg.scm. and almost everything in there had at least a minor version bump.
<mark_weaver>I also updated dbus, glib, pixman, cairo and some others to the latest.
<mark_weaver>currently rebuilding my system to make sure everything still works and shake out any remaining problems.
<nkar>mark_weaver: re multiple postgres versions on hydra: some may be redundant. maybe I forgot to clean up after upgrading, but that was a long time ago, and I can't recall anything.
<mark_weaver>if OpenGL ES 1.1 / 2.0 support is important to anyone, they can enable it by packaging libxml2-python
<mark_weaver>nkar: ah, okay, thanks. I didn't realize you were on that list :)
<mark_weaver>relatively new nick, and I thought you were relatively new to guix
*davexunit needs to figure out issues with opengl on guix
<mark_weaver>well, we've been running a very old mesa, maybe that's a problem?
<mark_weaver>the newest mesa will be in my upcoming updates.
<davexunit>maybe. I think I had issues with the mesa libs not being on the load path.
<davexunit>so no webgl in icecat, minetest ran horribly.
<mark_weaver>so much to do :)
<davexunit>if only guix could be my full time job.
<mark_weaver>working on guix seems to be somewhat addictive :)
<davexunit>it's good fun. there's a good range of things to hack on.
<mark_weaver>davexunit: while packaging the latest Mesa, I noticed that we don't have Mark Kilgard's GLUT library, which is linked prominently from
<mark_weaver>instead we have something called "freeglut".
<mark_weaver>well, I wrote "instead", but I don't know if they're interchangeable or not.
<davexunit>there's a nonfree GLUT implementation
<mark_weaver>I'm rather ignorant of GL. what's your take on this?
<davexunit>could it be that one?
<davexunit>I'm not a glut user, so I just don't know all the details.
<davexunit>glut and glu are pretty old school.
<davexunit>but plenty of stuff uses it.
<mark_weaver>nevermind then. if someone wants it we can revisit then.
<davexunit>yeah, we'll see.
<mark_weaver>Mark Kilgard's GLUT library is hosted at so I would hope it's free.
<mark_weaver>sorry, wrong link, this is the one:
<davexunit>the licensing is weird...
<davexunit>/* This program is freely distributable without licensing fees and is
<davexunit> provided without guarantee or warrantee expressed or implied. This
<davexunit> program is -not- in the public domain. */
<mark_weaver>that's rather vague
<davexunit>don't know what to make of it. I can't find a single file in the release that gives a more clear license.
<mark_weaver>hydra is really clogged up waiting for these texlive transfers:
<davexunit>we got knuth'd.
<mark_weaver>looks like debian uses freeglut, and doesn't have this other one.
<mark_weaver>heh :)
<davexunit>I guess this is the nonfree one because of the extremely vague licensing statement.
<mark_weaver>yeah, I'd rather not touch that.
<mark_weaver>the fact that debian (apparently) doesn't have it is enough for me.
<mark_weaver>texlive is such a monolithic monster.
<mark_weaver>whenever I look at it, I think it would be so much better to package the individual pieces, but on the other hand, I'm not volunteering for that job either :)
<davexunit>there be dragons.
<mark_weaver> is signed with a mystery key that I can't find anywhere.
<mark_weaver>so I held off on updating it, although maybe that's silly since most of the packages aren't signed at all.
<mark_weaver>civodul: I've updated a very large number of packages, and would ideally like to order the commits in topological order. is there anything handy to help me do that?
<mark_weaver>anything to top-sort a set of packages?
<mark_weaver>maybe it doesn't matter.
<mark_weaver>if I were rewriting hydra, I'd probably use such a thing to determine the build order, instead of the "build steps" thing that's currently done.
<mark_weaver>actually, if I just had the full dependency graph, I'd compute its SCC.
<mark_weaver>well, nevermind, I suppose it doesn't matter.
*mark_weaver goes afk
<civodul>mark_weaver: i don't think there's such a tool, though maybe what bavier wrote for 'guix refresh -l' could help
<civodul>otherwise there's a topological-sort for derivations, in (guix store) IIRC
<civodul>rekado_: re IXIN, good question, dunno
<rekado_>I mailed a patch to ttn directly that fixes problems with the guile2-based tool
<civodul>oh good
<civodul>we should tweak him into getting involved in that discussion
<rekado_>but I must say that I don't quite understand how IXIN fits into this whole info-must-die discussion.
<civodul>it's meant to be an Info replacement, AIUI
<rekado_>I mean, it doesn't attempt to "solve" any of the "problems" that esr sees with info.
<civodul>actually, esr's main concern is Texinfo, not Info
<civodul>that is, he wants the whole shebang to die, not just Info ;-)
<civodul>(that's a choice of word that is very esr-esque)
<rekado_>IXIN makes it easier for renderes to tweak the output of the format, but it actually operates on the XML dump of texinfo.
<civodul>that would be under the assumption that Texinfo is still used
<rekado_>it's not quite clear what an alternative input format (other than Texinfo or its XML dump) to be used with IXIN would look like. And that's where the bikeshedding can be resumed. "Org-mode as an input to IXIN! --- No, DocBook! --- How about Asciidoc to XML to IXIN to SXML...?"
<rekado_>I hope ttn will shed some light on how IXIN fits in.
<civodul>thing is, i have no problems with Texinfo
<civodul>and i did look at the other options
<rekado_>well, neither have I.
<davexunit>texinfo is pretty good, I haven't seen a better documentation format.
<civodul>the whole "hipsters find it archaic" argument seems pointless to me
<civodul>and i'm astonished to see rms buy it without further ado
<civodul>anyway, i'm irritated, this is no good ;-)
<davexunit>and it only *looks* archaic because no has bothered to do some web design work to make the html output better
<civodul>well, rekado did :-)
<civodul>which reminds me i need to take the next step on bug-gnulib
<bavier>civodul: I might have to scratch an itch and do up 'guix import cpan'
<civodul>bavier: yes, i've noticed it would be very useful, go ahead :-)
<davexunit>very useful
<davexunit>in this topsy turvy world that is held together by perl scripts
<rekado_>I started working on 'guix import cpan' but noticed that the metadata provided by isn't all that rich.
<rekado_>then I dropped it...
<bavier>rekado_: I noticed the same
<bavier>things could get messy
<rekado_>there is, which might be better, but it looks like we'd have to scrape most interesting stuff from the huge value assigned to the "pod" key (that's just text, not structured data).
<civodul>i think it'd be helpful even if some of the fields are still left to be filled out
<davexunit>that's how the pypi importer is
<davexunit>it fills out what it can, but leaves the dependencies and such for you to fill out.
<civodul>right, same for "gnu"
<bavier>should be doable for cpan as well
<Fullbreed>Anybody home
<jxself>No. Wait for the beep, and then leave a message.
*jxself beeps
<Tsutsukakushi>hai, remember the problem i had with the usb being mounted as sda?
<Tsutsukakushi>another person that i know had the same problem today
<Tsutsukakushi>it might be problem in the usb image
<Tsutsukakushi>both machines are x86_64
<jxself>I have the same issue too but with a totally different distro.
<jxself>So I don't think it's Guix-specific, whatever makes that happen.
<jxself>My X60s with libreboot does it too.
<civodul>Tsutsukakushi: actually what's the problem?
<Tsutsukakushi>the usb is the /dev/sda
<Tsutsukakushi>and the disk sdb
<jxself>But when I boot from the HDD then it becomes /dev/sda...
<Tsutsukakushi>if you install grub to sda it goes to the usb, if you install it to sdb it looks for the os from the sdb
<Tsutsukakushi>something like that
<jxself>When I installed Trisquel I told it to install to /dev/sdb.
<jxself>But it boots just fine.
<Tsutsukakushi>i thought it was just my computer but happened to other person too
<Tsutsukakushi>both are bios if that matters
<civodul>i would just install to whatever the disk is called, no?
<jxself>"...the order in which their corresponding device nodes are added is arbitrary"
<civodul>would have been nice to have a Guix image at
<jxself>Send them a link to the current on
<civodul>there's just the installation image here
<civodul>but we could send one based on demo-os.scm