IRC channel logs


back to list of logs

*mark_weaver extracted the security-critical patches from mozilla's repo to apply to icecat manually. doing a test build now.
<mark_weaver>jxself: I've opted for the approach above, until quidam releases a new version of icecat.
<jxself>OK. Did RMS saying anything interesting?
<mark_weaver>jxself: I decided that it wasn't worth taking his time. This solution is okay with me.
*mark_weaver just pushed security fixes to icecat
<zacts>lo. what's new in guixland?
<jxself>Yeah, this seems like a good solution. I can't believe I hadn't though of it myself.
<rekado_>I have a package definition ready, except for the test suite, which depends on quite a lot of other unpackaged perl modules.
<rekado_>Should I first package all of these perl modules before submitting a patch for this package, or can I submit a patch where the test suite is not yet enabled?
<Sleep_Walker>how can input be invalid?
<Sleep_Walker>guix package: error: gnu/packages/efl.scm:121:2: package `elementary-1.12.2' has an invalid input: (("efl" #<package efl-1.12.2 gnu/packages/efl.scm:47 343e3c0>))
<mark_weaver>gah, updating groff broke lots of builds on hydra, although it worked on my own system. hmm.
*mark_weaver investigates
<mark_weaver>ah, it's a parallel build problem.
<Sleep_Walker>still no idea?
<jxself>About sleep walking? Strap yourself in to stop it.
<Sleep_Walker>how can input be invalid?
<Sleep_Walker>guix package: error: gnu/packages/efl.scm:121:2: package `elementary-1.12.2' has an invalid input: (("efl" #<package efl-1.12.2 gnu/packages/efl.scm:47 343e3c0>))
<Sleep_Walker>jxself: the only movement I have? not a chance :b
<mark_weaver>Sleep_Walker: it looks to me like you've got a list like (("elf" ,elf)) where a package is expected.. if you paste your recipe, I can take a look.
<Sleep_Walker>mark_weaver: very good guess, thanks - too much parens
*mark_weaver works on applying the xorg security fixes
<mark_weaver>this is a massive pain :-(
<mark_weaver>I think maybe I should update our xorg first.
<civodul>Hello Guix!
<jxself>Ahoy there sir.
<jxself>Did you see RMS's suggestion?
<jxself>The "What To Call Guix" thread on gnu-system-discuss.
*mark_weaver works on upgrading mesa and xorg-server
*mark_weaver looks
<jxself>One eye on each task huh? :)
<mark_weaver>wow, that's a long name
<jxself>I don't think he was suggesting a name? But a description?
<davexunit>another long name suggestion? it's not "the gnu guix distribution of gnu/linux" again, is it?
<gry>"guix gnu/linux"?
<gry>Oh. Sorry. I should subscribe to that list. I am missing things...
<mark_weaver>gry: gmane has it, named "gmane.comp.gnu.system"
<jxself>Or the official site: ;)
<mark_weaver>I don't know why mesa's build system is so messed up. it keeps wanting to rebuild the build system.
<mark_weaver>I'm using "make -d" to figure out what's going on. I fixed the main build by creating a number of missing files in m4/* with their timestamps set to the epoch.
<mark_weaver>it got most of the way, but now it's trying to run msgfmt
<civodul>jxself: yes, i saw it
<civodul>watching the discussion, for now ;-)
<mark_weaver>maybe I'll just add gettext to the native inputs.
<Sleep_Walker>"avant-garde distribution" doesn't sound much appealing
<mark_weaver>I don't like it either
<jxself>Describing Guix as avant-garde is one thing.
<jxself>But I don't think his message addressed what people were really wanting to know/do.
<mark_weaver>the tmux-1.9a update didn't work. might be a bad source URI.
<mark_weaver>(not that I use tmux)
<civodul>Sleep_Walker: ↑ :-)
<civodul>in literature, being in the "avant-garde" usually means that your contemporary would consider you crazy
<civodul>only once dead would people think you were a genius
<civodul>i'm not sure that's a great goal for Guix ;-)
<mark_weaver>sweet, I finally got mesa-10.4.0 built with guix. now to update xorg-server and patch it.
<civodul>davexunit: did you find more clues on that nar issue?
<davexunit>civodul: I'm going to try your suggestions when I get a chance.
<davexunit>still at work, giving a presentation at the CiviCRM meetup here in the office later tonight, so I probably won't get a chance until tomorrow.
<davexunit>how I wish to return to happy scheme land.
<rekado>why does A.M.Szmidt keep saying that Guix is the name of the distribution and the package manager?
<civodul>rekado: feel free to ignore him here
<civodul>davexunit: heh, np
<civodul>good luck with that!
<davexunit>thanks. :)
<mark_weaver>hmm, I wonder if includes the fixes described in
<mark_weaver>I guess so, as they were released on the same day. that would be nice.
*mark_weaver wasn't looking forward to applying 37 patches
<mark_weaver>sweet, that already seems to have the patches applied.
<civodul>thanks for taking care of that!
<civodul>BTW, re the IceCat update: if the patches are available as individual files upstream, it's preferable to use 'origin' instead of adding them to the repo
<mark_weaver>civodul: they aren't. I had to cherry-pick them from the git repo (which is a mirror of the official mercurial repo)
<civodul>woow, ok
<mark_weaver>I had to search for them manually based on the bug numbers in the security advisory.
<civodul>well, one more reason to thank you!
<mark_weaver>glad to help :)
<civodul>it's always surprised me that it can be so hard to found the actual fix for a CVE
<civodul>i don't know why they don't include a patch on the or something
<mark_weaver>heh, that would be nice!