IRC channel logs


back to list of logs

<mark_weaver>or: guix build --log-file <package-name>
<sapientech>unfortunately, the one it showed, was the one that i ran right after and then cancelled
<sapientech>imma just rebuild
<mark_weaver>yeah, it only keeps that most recent log of any given derivation.
<mark_weaver>*the most recent
<mark_weaver>sapientech: also, it's good to pass --keep-failed to "guix build", so that it will keep the failed build directory in case of failure.
<mark_weaver>to whomever is asking for another core-updates evaluation: please wait until we have the results of 'wip-gd-fix' branch, so I can merge that into master and core-updates before asking for another evaluation.
<sapientech>good point, thanks
<mark_weaver>sapientech: yeah, looks like something on your host system is interfering with the build.
<mark_weaver>I have to go afk for a while...
<sapientech>mark_weaver: after some research, confirming this is an emacs, not guix issue. trying a workaround....
<mark_weaver>sapientech: it's possible that other builds will fail for the same reason
<sapientech>workaround worked, will see if other builds fail too
<mark_weaver>okay, 'wip-gd-fix' was successful. cherry-picked to master and merged to core-updates
<lfam>Huh. I'm trying to update btrfs-progs to the latest (4.6.1), but the test suite fails because it can't find `e2fsck`. Of course, the directory containing `e2fsck` is on PATH, and the code that checks for it is just doing `which e2fsck`. So... somehow the build process is ignoring PATH.
<lfam>The error message was misleading — `which` itself was missing.
<lfam>I've seen that one before!
<piyo>hard to believe 'which' is optional
<mark_weaver>piyo: 'which' is in %base-packages, so unless you specifically remove it, it's present on every GuixSD system.
<mark_weaver>but the set of packages provided implicitly by gnu-build-system is a smaller set. usually 'which' is not needed by builds.
***tschwing_ is now known as tschwinge
<amz3>hello, I'm trying to install nodejs but it always try to compile it
<amz3>how can I troubleshot that?
<amz3>I am on guixsd
<amz3>I did guix pull recently
<iyzsong>amz3: (latest build of node) says 'the build is no longer available', it's due to the hydra disk broken rectently..
<amz3>is there builds for 32bit architecture ? because i am on 32bits
<iyzsong>same, it's not available too.
<amz3>I need to find a trick to make it compile :)
<iyzsong>good luck :-
<amz3>I tried to lower the cpufreq but i still fails to compile
***jjmarin1 is now known as jjmarin
<ng0>I found a patch from february this year of libunique which was not merged. there was no reply in the second thread after ludo' reviewed it.
<ng0>are there problems with newer libunique? else i would package it, i think it's something optional i need
<sneek>Welcome back civodul, you have 2 messages.
<sneek>civodul, jlicht says: something SPARQL-like would be amazing, althoug for my use-case, I could make it work with simple filters and maps
<sneek>civodul, lfam says: Do you think that this error message when pushing a new branch to Savannah is related to the new signed-commit hook? "remote: fatal: bad object 0000000000000000000000000000000000000000". Mark got that message a few minutes ago, after some brief downtime
<civodul>bah, yeah
<civodul>mark_weaver: the hook i gave at doesn't properly deal with this case
<ng0>libreoffice update is blocking me from test building... I should look into this offloading build
<ng0>libosinfo by Rene is also not merged, february. i will mark those i find for myself as guix::patch:wip to see if i can get them past the finish line
<civodul>mark_weaver: i think we must let core-updates build now, and stop further master builds
<civodul>ACTION runs GuixSD from core-updates
<ng0>civodul: have you tested the gnunet-download-manager.scm in gnunet recently, is it still functional with guile-2.0.11? I just discovered it
<civodul>i think it's not :-
<civodul>ACTION has to go
<ng0>okay, thanks. I'll consider to debug as I learn more
<amz3>does anyone knows how to patch a binary to make it work under guixsd?
<ng0>depends on what needs to be patched
<ng0>and the type of binary
<ng0>gnunet-svn, gnunet-gtk-svn update: expect both to be ready for review either today or tomorrow on the list
<ng0>building too much at the same time slows down everything here atm
<ng0>i'll add building optional features to gnunet-svn as time passes, the base is done.
<ng0>there are optionals i added in gentoo which i need to question upstream for what the dependencies do etc.
<amz3>at last I compiled node, I don't need patchelf
<ng0>laptop smells really warm.. compiling libreoffice, texlive, and some third application is not recommended
<ng0>at the same time i mean
<ng0>well it is finished. i have used this for a while here myself.. but it will only be fully functional with the gnunet service i need to write… unless i can debug it, but form my experience it will be easier with the service (how it's intended to be run)
<ng0>because of errors like: ng0@shadowwalker ~$ Jul 30 14:58:25-224568 vpn-8973 ERROR `gnunet-helper-vpn' is not SUID, refusing to run.
<ng0>the standard case is create gnunet group+user + gnunet-dns, setup gnunet with gnunet user, add regular user to gnunet group
<ng0>i declare it functional enough to sent the patches, but it suffers from this bug as the other gnunet versions. therefore i'll write the service asap.
<ng0>alright, sent. thanks for reviewing. I'll be off :)
<myglc2>Anyone care to share tips re best way to interact with guix mail lists using emacs?
<myglc2>I've been using gmane, which seems to be in jeopardy
<myglc2>Hmm... guess I should just use directly, eh?
<alezost>myglc2: I just use gnus and gmane; what jeopardy do you mean?
<amz3>what console emulator do you use?
<amz3>terminal emulator*
<myglc2>alezost: gmane search glitches + lost post + above linke suggest gmane might go away :-(
<mark_weaver>sneek: later tell civodul: I needed to focus hydra on 'master' in order to deploy the 'gd' security update.
<sneek>Will do.
<alezost>myglc2: ouch, thanks for the link!!
<amz3>I need to use a recent version of firefox to do some web stuff
<amz3>what do you recommend me to do? outside using another distro
<mark_weaver>amz3: out of curiosity, what's your use case? Firefox 45 is ESR now, and GNU IceCat should be based on that, but there's only one IceCat developer, stretched thin at the moment.
<mark_weaver>amz3: would epiphany, based on the most recent stable webkitgtk, be sufficient?
<mark_weaver>unfortunately, upstream Firefox steers users to non-free software, so we can't include it directly.
<amz3>mark_weaver: I require firefox 48 for an extension that extends the devtools
<amz3>it won't work with epiphany
<mark_weaver>has firefox 48 even been released yet?
<mark_weaver>when I go to the firefox download page, it offers me 47.0.1
<mark_weaver>what extension is it?
<mark_weaver>(just curious)
<amz3>it's called 'redux dev tools' it's for developing frontend apps using redux framework, a framework that takes insipiration from functional programming
<amz3>for instance the have a single global immutable state
<ng0>amz3: did you ask for firefox on the list a while ago?
<amz3>ng0: no, I know it's not going to be included in guix
<ng0>replies can lag behind i am learning new elayout
<ng0>i ask because i passed a wip of ff to someone
<ng0>abandoned it because of icecat and because through discussion I found out how to package torbrowser, when someone will get to it or for my personal notes on how to apply gentoo packaging of it to guix. i'm not sure if it was 100% done discussing it, but at least pretty complete
<ng0> well, this is one more minus for patchwork. It does not display/recognize more than 1 appended patch per message
<wgreenhouse>ng0: your notes on torbrowser packaging for guix would definitely be interesting to me
<wgreenhouse>even if in an incomplete form
<ng0>I can bump the thread with some added details if it needs some, else I'll just point you to it if i remember irc nicks.
<ng0>it replicates in some parts what we do in gentoo inofficial build
<wgreenhouse>ng0: thanks! even just putting it up on a paste site would be great, since gmane is down. :(
<ng0> guix-devel archive, last 3 months. should've had either firefox or torbrowser as part of subject/content.
<ng0>i'll search for the direct link now
<ng0>gmane... down? oh
<wgreenhouse>ng0: yes, although the nntp gateway might come back. cf.
<wgreenhouse>ng0: looks like the right thread, yes?
<ng0>interesting. i expected running an archive to be much trouble, but i thought a group was behind gmane
<ng0>correct. that's the right thread, you just have to find the infomration i can not piece together right now. for eventual debugging i might or might not be of use.
<wgreenhouse>adam sjøgren and some others helped out but mostly it was just larsi
<wgreenhouse>well, thank you for mentioning it so that I could start looking
<ng0>the problems email has.. you have to figure out the fixes it does not intend to have yourself
<ng0>i'm happy with the work around notmuch gives me
<wgreenhouse>yes, provided you can easily download a large archive from somewhere, a system like that is very nice
<ng0>icecat: 45.x iirc. torbrowser, moving one release candidate range above, 45.2.0 at the moment
<ng0>oh.. 6.0.2 was released i think, i should version bump
<ng0>the funny part will be the language files etc..
<wgreenhouse>yeah, torbrowser 6.x = firefox 45.x
<ng0>they do only release them like this:
<ng0>and also for extra fun 45.2 vs 45.0 has some files moved in place etc
<civodul>this gmane story is terrible
<sneek>Welcome back civodul, you have 1 message.
<sneek>civodul, mark_weaver says: I needed to focus hydra on 'master' in order to deploy the 'gd' security update.
<ng0>torbrowser 6.x is 45.2 not 45.x
<ng0>well that sentence was stupid. but there is a difference between icecat and torbrowser
<ng0>at least when i looked at them, last time before the icecat update
<mark_weaver>civodul: core-updates is not yet usable on i686. I need security updates in the meantime, and I guess I'm not alone.
<civodul>hi mark_weaver
<civodul>mark_weaver: sure; any idea what's missing on i686?
<civodul>i can help sort it out
<mark_weaver>I just restarted a bunch of i686 jobs, as a result of gnupg-2.1.13 failing a test the first time.
<mark_weaver>but with any rebuild, there are many transient failures like this that come up