IRC channel logs


back to list of logs

<ng0>whoever asked on the mailinglist earlier today: gnunet latest head builds, runs, but -gtk misses some features with latest head, you can see that the refactoring is still going on
<ng0>maybe works. don't know. haven't really checked, logs read like either me wrong config or bad build
<ng0>sneek: later tell iyzsong: it seems your mailbox is or was full/quota maxed out. one of the messages I CC'd you at came back
<ng0>so i think i missed some messages... a very long thread about the coming soon 0.11 release.
<ng0>new tor version was released. I updated it, for anyone looking to review and push.
<ng0>good night
<lfam>Is Manolis here?
<lfam>I'm trying to merge master into core-updates-next and I'm not sure how to resolve a conflict in (gnu packages commencement):
<lfam>I *think* that kernel-headers-boot0 should win, because the commit message (e41e39451) introduced it as a new variable in late June, but I'm not sure
<demotri>I tried to install Guix under QEMU, but it failed:
<demotri>make[1]: *** [temacs] Error 1
<demotri>make[1]: Leaving directory '/tmp/guix-build-emacs-no-x-24.5.drv-0/emacs-24.5/src'
<lfam>Similar conflict in (gnu packages cross-base), pasted in an annotation to my previous paste
<lfam>demotri: Is there anything else?
<demotri>lfam: ld returnd 1 exit status
<demotri>Makefile:664:recipe for target 'temacs' failed
<lfam>demotri: Can you paste the whole thing on
<demotri>lfam: I don't know how to get it out of qemu.
<demotri>And there are more errors.
<demotri>Is there any way to restart? I have the feeling that repeatable 'guix system init ...' will build/install more/other packages.
<lfam>There is probably some message that says why the target failed.
<lfam>You can do `guix system init` again. It sounds like it failed while building packages and that should be fine to restart.
<lfam>I assume you are using one of the GuixSD 0.10.0 USB installer images?
<demotri>lfam: yes. 0.10.0 USB
<lfam>phant0mas: I remembered your IRC handle :) Can you look at the paste of the merge conflict that I linked above? I'm merging master in to core-updates-next.
<demotri>lfam: I started from scratch, currently it looks better. Also, I'm logging to file.
<lfam>Okay, that should make it easier to debug if there are problems
<lfam>There are a lot of conflicts in gnu/tests/base.scm and gnu/tests/install.scm when merging master into core-updates-next.
<lfam>Okay, for (gnu tests base), it seems that master should win all the conflicts
<lfam>And likewise for (gnu tests install)
<lfam>Huh. Merging master into core-updates-next caused the mcron documentation to be duplicated in doc/guix.texi. Git didn't even report a conflict there.
<lfam>I wonder if I can amend the merge...
<ng0>hi. the moving of shells broke some packages
<ng0>is someone working on that already?
<ng0>ACTION gives it a try
<ng0>we broke the po files too.. i think i'm lamost finished with the patch
<ng0>done. i can not build all the packages in the files where i changes tcsh to shells, can hydra just do that?
<ng0>it's proof enough to work for me when make and make check runs fine
<ng0>someone review and apply (there's a typo in one of the commit messages)
<ng0>fixes broken master build
<ng0>why did we apply a patch series which breaks this and does not take occurences of removed modules in the first place.
<ng0>*does not replace
<ng0>ACTION afk
<ng0>civodul: ^
<civodul>Hello Guix! :-)
<ng0>hi. oh, probably doesn't make much sense becasue you just connected
<ng0>master broke because of some deleted modules, i sent a patch which fixes it
<ng0>should've been checked more carefully.
<fall__>is it normal for the initial guix system init command to stop on the line of "base-initrd.drv"?
<ng0>fall__: idon't think so
<ng0>can you provide more info?
<fall__>im just wondering, im very new to guix(sd) and linux distros in general
<ng0>i don't know what the output should be when it finishes successfully, too long since i ran it.
<fall__>i follow the manual exactly and im using the usb 0.10.0 files
<ng0>not 0.10.1?
<fall__>i am pretty sure it was 0.10.0, but im not on my workplace right now, so i cant tell you for sure right now
<ng0>i'm multitasking much at the moment, so maybe someone else can help before i provide false info :)
<ng0>there will be a release of 0.11 soon btw
<ng0>ah, definitely 0.10.0
<ng0>working on two 0.10x software systems at the same time xD
<fall__>also, the only thing i dont do that is in the manual is i do not create the partitions as there is already a partition created and mounted by default
<fall__>im also using virtualbox for it
<ng0>there's a guide for qemu in the manual
<ng0>i meant to say, we do not officially support virtualbox, but some people have guixsd working on virtualbox. there might be threads on the guix-devel or help-guix mailinglists where people ran into problems
<fall__>i know, ive seen that, why would the system init stop a certain point if i just use the usb file and convert it to a vdi file? not like it doesnt boot or anything, it just does nothing or is extremely slow after that one line
<ng0>honestly i don't know. stick around a while, maybe someone can help
<civodul>fall__: it's hard to tell what's happening
<civodul>maybe you should try 0.11 when it's out (real soon)
<jlicht>hello guix
<civodul>hey jlicht!
<GNUtoo-irssi>hi, guix system vm-image --image-size=4G example.scm now works
<GNUtoo-irssi>I've the VM that starts and all
<GNUtoo-irssi>however I end up with a guile shell instead of bash
<GNUtoo-irssi>I used the .scm from the manual here: 7.2.1 Using the Configuration System
<GNUtoo-irssi>I removed the user account creation and changed the locales
<GNUtoo-irssi>also, changing " (packages (cons tcpdump %base-packages))" with busybox instead of tcpdump didn't work:
<GNUtoo-irssi>In procedure module-lookup: Unbound variable: busybox
<iyzsong>GNUtoo-irssi: 'busybox' is defined in 'gnu/packages/busybox.scm', to make it available, you can add '(use-packages-module busybox)' or '(use-modules (gnu packages busybox))' at the begin of your system config file.
<sneek>Welcome back iyzsong, you have 1 message.
<sneek>iyzsong, ng0 says: it seems your mailbox is or was full/quota maxed out. one of the messages I CC'd you at came back
<iyzsong>ng0: ah, thanks :-
<ng0>do you think I could ask for push access? I know my own view on my capabilities is bad because of own insecurities.
<iyzsong>yes, I don't see why you can't.
<ng0>okay, i'll see to request on savannah later today.
<iyzsong>welcome in advance :-)
<GNUtoo-irssi>ok, thanks
<ng0>grothoff told me i should maybe ask for push to gnunet, not dealing with trivial fixes..i though the patch i sent this morning to guix also to be trivial (but a typo introduced in the commit) to fix the broken master. bothering people about it.. well okay, QA.
<jlicht>hmm, I a nice command yesterday getting the exact environment of the guix daemon in case of test failure
<jlicht>can't seem to find it now
<jlicht>it was some env $(which bash) command
<jlicht>I saw *
<ng0>i'm not sure if i sent the request for inclusion at savannah already, i thought hitting Enter would make a linebreak
<ng0>in any case, you either get one very short or two requests now
<ng0>am i the only one experiencing a broken master? i pull my checkouts from a local untouched guix master which is synced.
<ng0>Sleep_Walker: neomutt, do you mean you do not experience the problems i had?
<civodul>jlicht: maybe "guix build something -K" and then "cd /tmp/guix-build* ; source environment-variables"
<Sleep_Walker>ng0: I have no crash
<ng0>then i'd say commit what i sent most recently
<Sleep_Walker>yeah, that's what I meant :)
<jlicht>civodul: found it!
<rekado>gtk-doc is broken; the tests fail.
<jlicht>especially the noprofile and norc things were nice
<phant0mas>is lfam here?
<phant0mas>sneek: later tell lfam use kernel-headers
<sneek>Will do.
<phant0mas>sneek: botsnack
<ng0>i don't know what went wrong this morning.. now i can build again
<jlicht>node started failing a test, but the node source hasn't changed
<jlicht>Does that mean that either the test result is non-deterministic, or that something with the inputs might have changed?
<davexunit>could be either or both
<davexunit>core-updates was just merged
<davexunit>so lots has changed
<brendyn>Anyone running guix on their personal computers yet?
<brendyn>GuixSD i mean
<civodul>ACTION does
<davexunit>yeah, a bunch of us.
<civodul>jlicht: the failure seems to be deterministic
<brendyn>Keen to move over one day
<brendyn>Busy with study atm though
<civodul>jlicht: it was mentioned in the core-updates thread on guix-devel
<Sleep_Walker>ACTION does ;b
<jlicht>civodul: I just saw it, yes. I suspect it's related to a change in openssl, checking right now...
<davexunit>ACTION notices that 0.11 has been tagged :)
<davexunit>jlicht: I'm also building node right now
<davexunit>re-downloading the world first, though. :)
<jlicht>davexunit: I'd be much obliged if you can share your findings, if any :-)
<davexunit>jlicht: no guarantees since I'm at work, but when I have an idle moment I will poke at it.
<davexunit>jlicht: this is a tricky one
<davexunit>6.3.1 is out. gonna try upgrading and see where that gets us.
<davexunit>jlicht: good news: that test does not fail in 6.3.1. bad news: 3 tests fail.
<myglc2>hello guix
<myglc2>Hi alezost, sorry about the run around yesterday.
<alezost>myglc2: hi! no problem :-)
<ng0>now it builds
<ng0>wtf was wrong this morning
<ng0>no.. stil lfails on other machine
<GNUtoo-irssi>thanks, it built with busybox, I also figured out that having the guile interpreter meant here that I was in the initramfs...
<GNUtoo-irssi>the manual says (with the example) ;; Assuming /dev/sdX is the target hard disk, and "my-root" is the label of the target root file system.
<GNUtoo-irssi>(file-systems (cons (file-system (device "my-root") (title 'label) (mount-point "/") (type "ext4")) %base-file-systems))
<GNUtoo-irssi>I had that error in the past already, I should look at how I fixed it, I guess that I must change title 'label to 'path and put a patch instead of an uuid
<jlicht>davexunit: do you by any chance have log output for it?
<davexunit>jlicht: here is the relevant portion
<GNUtoo-irssi>that should do: (device "/dev/sda3")
<ng0>gtk-doc fails.
<ng0>did someone see and fix this? i did not follow all the channel
<jlicht>davexunit: hmm, I'll have a poke at it later today. The localhost one seems similar to something I've already seen in guix, the dgram might actually be something weird in the source
<jlicht>and the third one, who knows :-)
<ng0>can someone look at this thread? I'm super confused if it's just locally me or also other people affected.. guix pull pulls a broken master on all my guix installations.
<rekado>ng0: I already pushed one of these patches. Thanks.
***civodul changes topic to 'GNU Guix | | 0.11.0 is out! | videos: | bugs: | patches: | paste: | log:'
<civodul>0.11.0 \\o/
<civodul>ACTION goes get some fresh air :-)
<ng0>rekado: yep. now it compiles again. thanks for pushing :)
<paroneayea>hey I should annoucne that
<rekado>civodul: yay!
<ng0>too much parallel work.. i forgot that libunique does not fail, it is gtk-doc
<ng0>good thing i did only sent the patch for libunique then
<sirgazil>0.11.0 ♫ Thank you everyone for the hard work :)
<ng0>i like the transparency and clear content of guix commit messages. gentoo just dropped guile-2 and i don't know why.
<ajgrf>sorry, didn't mean to type that
<ajgrf>actually, does anyone know where to find eelco's papers on nix?
<ajgrf>i found a reference to it on the guix web site but i can't actually find it anywhere
<itorres>sorry, wrong link, it is
<itorres>(got there from the bibliography in Guix article)
<lfam>How can I add a wifi-service to the "bare-bones" OS config template? I tried adding wicd-service, and it demanded dbus, so I added the dbus-service, and it then told me that I had provided the networking service more than once
<sneek>Welcome back lfam, you have 1 message.
<sneek>lfam, phant0mas says: use kernel-headers
<lfam>phant0mas: Thanks!
<lfam>Well, I don't really need to do that, but I would like to figure out how. I do need to strip my system down because I want to rebuild the system based on some changes to core pcakages
<jlicht>I keep running into a ` fatal error: linux/limits.h: No such file or directory` error
<jlicht>when building a piece of software I'm packaging
<lfam>jlicht: Have you tried setting -I in #:make-flags or #:configure-flags?
<Sleep_Walker>civodul: thanks for bringing up bug #20067, I wasn't able to find it by myself, I'll try that approach
<jlicht>lfam: that... might be a really smart idea.
<lfam>Heh. It's not typical for limits.h to be missing.
<jlicht>lfam: I guess I'll figure it out, but seeing as it's a configure script that wraps a waf script, I might need some hacks to get this working
<jlicht>seems to work! Thanks a bunch
<lfam>I'm glad it was easy enough to already be working!
<jlicht>well, I just got past this hurdle. More than enough to go ;)
<sirgazil>Howdy! The release announcement uses "" to download the public keys, but it never works for me. I get this message:
<sirgazil>But the server indicated in the manual works fine.
<sirgazil>civodul: about "" ↑
<civodul>hi sirgazil!
<civodul`>ah right, is kinda broken
<civodul>i forgot to update that
<Sleep_Walker>damn, it seems I diverged again, everything is building locally
<lfam>Does anyone know how to add the wicd-service to the bare-bones template?
<ng0>is there a smarter way than like the patches download in bash.scm to handle the download and applying of the patches for netcat-openbsd?
<ng0>the patches.. here or here. see the discussion(s) on our mailinglist
<Orioa>im getting ready to install
<lfam>ng0: I'm not sure if there's a better way than that or not.
<lfam>ng0: But, I didn't really like the response from the Debian developer. It seems like nobody is responsible for those thousands of lines of patches to OpenBSD's netcat code.
<ng0>and also no further responses..
<lfam>Does anybody actually understand and maintain all those patches?
<ng0>openbsd upstream is
<ng0>openbsd upstream seems like it is done, upstreaming according to debian did not work, last commits to linux port 2012 same as the patches
<lfam>ng0: OpenBSD is maintaining the patches in Debian? The Debian git repo shows no work since 2012
<ng0>hit enter too soon
<Sleep_Walker>and what about leting it rot?
<ng0>for me personally it's okay.. i use it and have found no way to work around what I need for some parts in ssh i use
<Sleep_Walker>there is GNU netcat, busybox netcat, socat, ...
<ng0>i need it for.. one moment
<lfam>OpenBSD's native netcat code is maintained, but the Debian package is allowed to rot. I think that Debian should drop the package entirely.
<ng0>Host *.onion ProxyCommand nc-xlocalhost:9050 -X5 %h %p etc
<lfam>nmap also has a netcat implementation
<Orioa>hmm can't seem to get it to download
<lfam>Orioa: What link?
<ng0>started using this in gentoo, did not find another netcat implementation, but i also tried not many
<ng0>orginal netcat did not work out
<Orioa>the download page keeps timing out
<ng0>Orioa: for guix?
<lfam>Orioa: Can you paste the link that is failing?
<Orioa>got it to go to the http download list
<Orioa>not sure what one to download
<lfam>So, it's not timing out anymore?
<lfam>As for which one to download, what are you trying to do? Install the operating system, or the standalone package manager?
<Orioa>i went to the link at bottom of page
<Sleep_Walker>ng0: socat can do socks4 proxy, I'm not sure about socks5 and its difference
<Orioa>operating system
<lfam>Orioa: On what architecture? x86_64 (64 bit) or i686 (32 bit)
<Orioa>x86_64 (64bit)
<lfam>Then you want guixsd-usb-install-0.11.0.x86_64-linux.xz and the associated .sig file
<ng0>Sleep_Walker: i think i'll make this one (netcat-openbsd) a candidate for out-of-tree maintenance on my side.
<ng0>or maybe just don't
<Sleep_Walker>sure, I just want to help you with maintained alternative
<ng0>i think i tried socat.. i don't know. i have to try again.
<ng0> do they not know about onion subdomains, like www.123425asdfg.onion git.12345asdfg.onion etc? hrm
<pmikkelsen>Hello guix!
<sapientech>hi everyone, created a package and installed it with guix package -f package_name.scmw
<sapientech>when i list installed, it shows up, however i am not able to use it in guix environment commands, for example guix environment --ad-hoc package_name does not work
<rekado>sapientech: use GUIX_PACKAGE_PATH
<sankey>or --load-path=
<rekado>if the module is “(my packages foo)” in a file at “${GUIX_PACKAGE_PATH}/my/packages/foo.scm” then setting the GUIX_PACKAGE_PATH environment variable will tell Guix to pick it up.
<davexunit>the package must be in a proper module to use the load paths
<davexunit>-f works with files that evaluate to a package object
<davexunit>which is a different thing
<sapientech>okay thanks guys, now i imagine it is best to put it in a version controlled guix branch and do ./pre-inst-env? id like to do that, but have had trouble with it thus far
<sapientech>that way i could simply put the package in gnu/packages/
<pmikkelsen>I don't think you have to put it in a branch of the guix source-tree.
<civodul`>grrr is it just me having connectivity issues?
<Sleep_Walker>probably interest in new release :)
<Sleep_Walker>but yes, I had that as well
<civodul`>heh, ok :-)
<pmikkelsen>I did not :)
<davexunit>judging by the comments here
<davexunit>it seems that we still have a bit of an "identity crisis"
<davexunit>people don't understand why guix exists when there is nix
<davexunit>or they think it is simply nix with a different language
<civodul>i guess we need to explain more :-)
<civodul>OTOH, it's probably expected rivalrish behavior
<civodul>same situation we see between competing compilers, languages, distros, etc.
<pmikkelsen>I think a large amount of the phoronix readers are not very concerned about freedom tbh, so maybe that explains it somehow
<civodul>yeah, possibly
<pmikkelsen>does anyone know when savannah will be running again?
<civodul>pmikkelsen: it works for me; what problem do you experience?
<Gamayun>Hydra is up but savannah is down?
<pmikkelsen>civodul: I get an error 500 when i try to do a git pull or guix pull :/
<Gamayun>Same here.
<ng0>shouldn't the offloaded build keep the failed dir with -K passed to build? because this doesn't happen
<sapientech>davexunit: pmikkelsen got it working, thanks
<pmikkelsen>sapientech: Nice
<civodul>ng0: the failed dir is not transferred back to the calling machine
<civodul>this would be a nice improvement
<ng0>I assumed it was like this.. yes, that would be nice
<Sleep_Walker>asciidoc fixed, sway works \\o/
<paroneayea> yeep
<paroneayea>looks like we aren't using any fosshub hosted stuff
<pmikkelsen>Are there any haskell users in here
<Sleep_Walker>as I can't see wanderlust package, I assume you're using something else in emacs for mail - what it is?
<Sleep_Walker>people on this channel
<Sleep_Walker>ideally emacs fanatics :b
<ng0>sime use Gnus, some mu/mu4e, some notmuch, etc
<Sleep_Walker>ng0: thanks
<ng0>one could always package wanderlust
<Gamayun>ACTION uses Gnus
<Gamayun>Sleep_Walker: You could also get it from MELPA? ;)
<civodul`>paroneayea: fun; i didn't know about fosshub, but i guess i didn't miss much ;-)
<paroneayea>civodul`: heh :)