IRC channel logs

2015-03-26.log

back to list of logs

<cjbarnes18>evening all o/
<cjbarnes18>am attempting my 1st try at installing the guix distribution onto a eeepc. but for some reason mu atheros nic is not active.
<cjbarnes18>I can see the module seems to have loaded OK.
<cjbarnes18>dmesg suggests it should be there.
<mark_weaver>cjbarnes18: so what goes wrong?
<cjbarnes18>mark_weaver: I get no eth0
<cjbarnes18>ifconfig has no output
<mark_weaver>they have different names on guixsd, because we use eudev
<mark_weaver>try ifconfig -a
<mark_weaver>well, it's dinner time here, gotta go for a while
<cjbarnes18>hmm, enp4s0 and wlp1s0
<cjbarnes18>thanks for your help
*Sleep_Walker waves
<Sleep_Walker>first succesful connection with connman :)
<Sleep_Walker>but there is still a lot of work left
<Kolt>is there a way to automatically resolve dependencies with guix?
<abnegate>when i do a "make install" of the guix from git, i get these errors: http://okturing.com/src/2755/body
<Kolt>is it 0.8.1?
<abnegate>it's guix from git
<abnegate>i cloned the git repo yesterday
<abnegate>however, i originally installed 0.8.1 and still have a lot of packages from that version installed
<Sleep_Walker>abnegate: and what if you use LIBPATH instead? :)
<Sleep_Walker>`guix environment guix' should provide you all the environment you need for building guix
<Sleep_Walker>(of course, if you have guix :)
***ljhms_ is now known as ljhms
<rekado_>Hi Guix, I'd appreciate it if someone could take a few minutes to review this patch: http://lists.gnu.org/archive/html/guix-devel/2015-03/msg00537.html
<Sleep_Walker>rekado_: I'd add space between semicolon and comment
<Sleep_Walker>and maybe some explanation why there are commented parameters would be nice
<Sleep_Walker>I can't find anything bad/wrong besides that :)
***Digit_ is now known as Digit
<rekado_>bleh, julia complains: "Target architecture mismatch. Please delete or regenerate sys.{so,dll,dylib}."
<rekado_>:-(
<rekado_>Sleep_Walker: thanks for taking a look.
<Sleep_Walker>yw
<civodul>mark_weaver: regarding the IceCat fixes, perhaps we should increase the version number in such cases?
<civodul>that would allow ^ in guix.el to pick it up, and would generally make the update more visible
<rekado_>I wonder why julia complains about a target architecture mismatch. Unless ARCH or MARCH is passed, a generic Julia will be built. I don't understand why this wouldn't work.
<rekado_>filed a bug report upstream. We may have to pass the MARCH make-flag to fix this.
<civodul>don't wait any longer or you'll have to make it APRIL instead
<rekado_>:-p
<Steap>civodul: you should take a break.
<civodul>:-)
<wingo>:)
***nully__ is now known as nully
<taylanub>I think I'm done implementing my pkg-config/propagated-inputs discrepancies finder, and lo and behold, it has a bunch of results for the first package I pass it for testing purposes already :D
<taylanub>we're probably missing a lot of propagated inputs, given Requires: in .pc files should be propagated. maybe it's also worrying, since propagated inputs aren't nice.
<Sleep_Walker>yeah
<Sleep_Walker>but it would be better to fix pkg-config so it won't be needed
<civodul>taylanub: yeah, this is a bit worrying
<civodul>nice work though!
<rekado_>still trying to update texlive-2014. "guix build --no-substitutes texlive-texmf-2014" simply times out with "starting phase `install'"
<rekado_>often it also just freezes at "waiting for locks or build slots..."
<rekado_>nothing else is being built..
<rekado_>I don't know why it would have to wait for build slots.
<rekado_>frustrating.
<rekado_>downloading substitutes for texlive-2014 won't work because the 1+GB files are almost always corrupt.
<Sleep_Walker>maybe we could split such big beasts into smaller outputs and create one meta package gathering them
<civodul>rekado_: i managed to update it with substitutes after the last core-updates merge
<civodul>what happens exactly when you say they are corrupt?
<civodul>i have to rush, but i'll be back later
<Sleep_Walker>damn
<Sleep_Walker>I missed the commit message change :E
<taylanub> https://notabug.org/taylanub/guix/src/master/missing-propagated-inputs.scm open to improvement
<civodul>taylanub: looks good!
<civodul>could you run it on a bunch of packages and post the results to the list?
*civodul still needs to migrate things from gitorious.org
<Sleep_Walker>taylanub: in "[[:space:]]*([^[:space:]]+)(.*)" - what do you expect to match in match2?
<Sleep_Walker>civodul: sorry, I noticed not fixed commit message after push (and I didn't want to push --force :( )
<taylanub>Sleep_Walker: the rest of the line
<taylanub>it iterates over the parts
<Sleep_Walker>you're splitting string on whitespace
<civodul>Sleep_Walker: bah, you'll have to be super extra careful next time
<civodul>but you were right not to use --force
<Sleep_Walker>I'm still quite new to cooperative mode of git with so high standard of commits :/
<Sleep_Walker>sorry
<civodul>np
<Sleep_Walker>good new is that was another necessary step to connman
<Sleep_Walker>I'm on that right now
<Sleep_Walker>it's superfast compared to wicd and supports VPNs
<Sleep_Walker>but I (or more probably others) still miss some decent full-featured client
<Sleep_Walker>bbl
<rekado><civodul> [18:04:54] what happens exactly when you say they are corrupt?
<rekado>^-- I don't have the exact error message here right now, but bzip2 complains about the archive being corrupt.
<civodul>hmmk
<mark_weaver>it sounds like maybe your connection is getting interrupted before completion
<civodul>yes
<rekado>yeah, that's what I'm thinking too.
<mark_weaver>some NAT routers do that
<rekado>it's quite frustrating. I wish I could just resume the download.
<rekado>maybe I should use wget and manually place the resulting file in the store.
<rekado>it only happens on the campus; never had the problem at home.
<mark_weaver>it's a drag that we feel the need to add complexity in the upper layers to work around broken crap in the lower layers (in this case, firewalls resetting connections for us)
<mark_weaver>but I suppose it would be good for us to implement the ability to resume downloads
<mark_weaver>i.e. detect that the connection was lost prematurely and then reconnect and ask for a partial file
<rekado>hmm, that's odd. With wget I downloaded the file in 3 minutes.
<rekado>Same host.
<rekado>in another terminal guix is still downloading the file at 300KiB/s
<civodul>sounds like a déjà vu
<taylanub>I had some bzip2-related errors when, I think, Hydra was down
<civodul>yeah
<civodul>rekado: it would be good if you could reproduce the transfer rate issue
*civodul -> zZz
<civodul>good night/day!
<taylanub>mark_weaver: master fails to build at the moment; could you review my patch that fixes it? Subject: "[PATCHES] Fix remaining references to substitute-binary."