<s34n>Is it normal for every. single. operation. to be extremely slow? Or do I have some unique problem? <s34n>Maybe it's just everything that has to hit a gnu.org server... <s34n>It seems to hang sometimes (often) in then middle of `guix pull` or guix package install while downloading packages <rekado_>can you show a concrete reproducible example? <rekado_>do you mean that downloading from ci.guix.gnu.org is slow? <s34n>It has been displaying this for the last 10 minutes: <s34n> python-sphinxcontrib-devhelp-1.0.2 10KiB 307KiB/s 00:00 [##################] 100.0% <s34n>I don't understand that question <rekado_>FWIW the server is in Germany and has a 10G uplink <s34n>it's not actually downloading (according to the status prints) <s34n>The last package is 100% downloaded <s34n>It just hasn't started the next one <s34n>I mean, I can't explain the 307Kib/s speed. That's extremely slow. But stalled is different than slow <rekado_>you shouldn’t wait for 10mins on a stalled download <cbaines>The file is very small, so the speed measurement isn't really meaningful <s34n>cbaines: that was my thought <s34n>rekado_: if by "abort" you mean CTRL+C, that doesn't work <s34n>this is the 3rd time of this <cbaines>have you checked if there are builds going on in the background? <s34n>guix package install <packagename> <s34n>(I think. I can't look at history right now and I'm brand-new to guix) <rekado_>is your system completely unresponsive? <cbaines>s34n, you can try with --verbosity=2 that might produce more output ***terpri__ is now known as terpri
*rain peeks out from behind a corner <rain>could someone look at my Yggdrasil patches? ^u^ <dftxbs3e>nckx, ohh... again! I didnt even receive the warning mail nope.. <sneek>dftxbs3e, you have 1 message! <dissoc>how do i use uconv in a package definition? i get "failed with status 127" when trying to (invoke "uconv" ...). there's not package def for uconv <dissoc>NieDzejkob: i tried that with native-inputs but doesnt work <dissoc>does the 127 error always mean command not found? <mroh>dissoc: uconv is in package icu4c <apteryx>sanity check: UEFI *is* supported with our installer image, correct? <apteryx>as in, flashing a USB key with the iso should be able to boot on a UEFI machine <anadon>OK, what's the trick to getting meson to work? <apteryx>anadon: I'm afraid that's too vague to be answered. <anadon>It might involve some frustration <anadon>`The 'meson==0.53/2' diistribution was not found and is required by the application` when I run `meson`. <apteryx>oh! It seems it wants an exact version? What's ours? <anadon>There is a whole python stacktrace, but that's less interesting. <apteryx>so your string was 'meson==0.53.2' right? <apteryx>are you using the meson-build-system? <anadon>No, I'm using it for a coding project I'm setting up. <anadon>Did I do something dumb, or is the problem really non-obvious? <apteryx>what is that something complaining about meson? is it meson itself? <anadon>I think so. Just a bare `guix install meson && meson` causes the error locally. This is a side install along side Ubuntu 20.04. <apteryx>perhaps something's not right in your meson build file <anadon>When I removed it and used Ubuntu's, meson worked without issue. <apteryx>anadon: even when in a empty directory? <apteryx>ah, you may be having some interference with the PYTHONPATH of your system. Try using 'guix environment --pure --ad-hoc meson' <apteryx>I don't have that issue (Guix System) <anadon> Curious, it works when fully isolated. But in the polluted host envirnment I can see from the python stack trace that it is using guix's version of python. <apteryx>and picking up Meson libs from your host PYTHONPATH <anadon> I removed the Ubuntu version of meson before trying all that. <apteryx>perhaps you have pip installed versions laying around? <anadon>If so, I didn't type that. Let me double check. <apteryx>or perhaps some ubuntu package propagates meson libraries in your host - <apteryx>it'd be good to wrap our meson script though <apteryx>to lessen the likelyhood of this happening <anadon> only pip and setuptools are installed over the system and user profile. <apteryx>you could strace it and check which files it loads <apteryx>it's possible even without PYTHONPATH it just looks user /usr/lib/python/site-packages or such <anadon>It is doing a ton of lstat, fstat, and opens. <anadon>All the paths looks to be correct guix-ified paths. <anadon>Looking at the launcher, it does looks like it is related to how meson uses python modules. <anadon>It does mean that I can use a bigger invocation to make it work for now. <anadon>Ack. GCC 10.1 doesn't support C++20 directly yet. Darn you godbolt.org for spoiling me! <apteryx>what do you mean by "bigger invocation" <apteryx>do you mean the Guix specific launcher of meson, or otherwise? <anadon>Just running meson from a full guix envirnment each time. <bdju>does anyone know a way to view an xml file with its styling in a terminal? <bdju>I want to curl an xml file from the web and pipe it to something to display it nicely <apteryx>anadon: you may want to look up the meson package definition, and see if it's properly wrapped <apteryx>in a script that sets useful variables such as PYTHONPATH <apteryx>but in your case that seem not be the issue... <anadon>In a bit. It is already tomorrow for me... <apteryx>janneke: thanks for making the doc crisp clear w.r.t. hurd-vm and childhurd :-) <bdju>efraim: would it? does it support xml? I don't know anything about it really but just piping into it didn't seem to work at least ***wxie1 is now known as wxie
***bionicbeaver is now known as aligngator
<efraim>bdju: sorry, I was AFK. I thought it might, unfortunately it doesn't look like it <efraim>IIRC python-pygments can do some nice display coloring with xml but I've never actually used it <janneke>apteryx: yw, thanks for the suggestion :-) ***apteryx is now known as Guest37148
***apteryx_ is now known as apteryx
<DVulgaris>Hello, I'm debating between installing Parabola or Guix on my T400 (after I Libreboot it). <DVulgaris>One of the obvious benefits is not using systemd (in Parabola, OpenRC is supported, but it is flakey), another one is being source-based (though, how much can a C2D with 4 GiB of RAM at most handle?) especially if I'm going for a more minimal system (probably Wayland with Sway - the heaviest thing will probably be FireFox). <DVulgaris>I hope someone can help me with this or point me to a comparison page. <DVulgaris>Because I just read the headline that the bootstrap seed was reduced to 25%, which renewed my interest in Guix. <efraim>The desktop I currently use has a dualcore chip about 10 years old and I have 4 GB of RAM. SSD makes all the difference in the world. There's a couple of packages I refuse to build on my machine but it's not really a problem <EatThem>right channel for guix system? does somebody have a simple working wpa_supplicant config with a declared wifi profile in config.scm? *civodul uses NetworkManager <efraim>I'd like to air a general complaint that the meson itself doesn't like having different outputs and it makes it near impossible to split those programs into multiple outputs <bonz060>Hi. Does anyone have any experience packaging things in racket? Like if you want to package something that requires a `raco pkg install` somewhere; how'd you go about dealing with that? <aligngator>rekado: aah, this is the option for "no tabs". Cool, he-he, thanks. <bonz060>aligngator: Thanks! Let me have a look at that \m/\m/ <bonz060>aligngator: There's place where they mention packaging racket packages(similar to how pypi packages can be packaged). There was also talk about a racket build system but I reckon that never really materialised :( <aligngator>bonz060: try ask those people directly about the current state: something possible was discussed in mailing lists. <civodul>rekado: from a colleague of mine actually (a ReScience person) <civodul>i find it nice but i thought somebody must have done it before, no? ***Kimapr__ is now known as Kimapr
<EatThem>someone has a Working wpa_supplicant config? *rekado uses NetworkManager <rekado>civodul: spacemacs is too fancy-looking for me <rekado>I’m using the elegant-emacs settings now. I think it’s nice <wdkrnls>Great article! Reminded me of an old quotation I had read somewhere long ago for which I can now add two sentences: "Writing is nature's way of letting you know how sloppy your thinking is. Math is nature's way of letting you know how sloppy your writting is. Code is nature's way of letting you know how sloppy your math is. Guix is nature's way of letting you know how sloppy your code is." <civodul>rekado: do we need a package for it? <mbakke> /etc/wpa_supplicant.conf is managed outside of guix (by hand or wpa_gui) <efraim>does the cmake-build-system honor the #:substitutable? flag? <civodul>mbakke: commit 9edb3f66fd807b096b48283debdcddccfea34bad is problematic <apteryx>is there some place in Guix where basic colors are set for Emacs? <civodul>the stricter rules in "guix pull" report it as unauthorized <civodul>one of its parents is e5f95fd897, which lacks .guix-authorizations <civodul>i thought we rebased 'staging' on top of the .guix-authorizations commit, no? <efraim>civodul: looks like adding substitutable? doesn't cause a full rebuild, i'll push the patch in a bit <civodul>anyway, we'll have to live with that <mbakke>civodul: I thought so too, at least 'make authenticate' was happy :-/ <civodul>i'll change "make authenticate" to do the exact same thing as "guix pull" <civodul>"guix pull" will ignore historical authorizations <mbakke>so we should have rebased all merges on staging (as it was ancient), instead of just the one that added Brice's key <civodul>perhaps i wasn't clear, sorry about that! <mbakke>and I should have been more thorough/careful, I just assumed .guix-authorizations was already merged since 'make authenticate' was happy :-) <civodul>mbakke: is it a coincidence that the SHA1 ends in "bad"? :-) <civodul>there's a tradition of making elaborate jokes in this project <civodul>which reminds me that the systemd package hasn't landed <mbakke>heheh, if I could control the git commit SHA1 identifiers there would be a whole lot more subtle jokes :-) <mbakke>I still have the systemd package in my local tree and have to fix it every so often :P <NieDzejkob>reminds me that the sha256 of the ntl package starts with 1lisp :P <NieDzejkob>oh, even better, it references the latest version of guile, it starts with 1lisp30 <civodul>mbakke: that reminds me of "vanity" Tor onion names <civodul>it's easier to achieve because you can generate as many key pairs as you want until you obtain the desired effect <civodul>though we could have a "RANDOM" file in the tree to allow people to have vanity commit IDs <mbakke>I guess I can 'git commit --amend' enough times to find a suitable SHA for the commit <mbakke>the fun stops when I need to 'git pull --rebase' though <mbakke>expect lots more 'merge branch origin/master into master' from now on :P <NieDzejkob>I suspect a good first use would be the commit that adds it as a package <mbakke>civodul: so 'guix pull' need to special case the range b6d18fbdf6..e5f95fd897 (99 commits), right? <mbakke>would having it start at 9edb3f66fd807b096b48283debdcddccfea34bad be an option (muahaha)? <mbakke>that is truly awesome, thanks for this civodul ! :-) <mbakke>once the dust settles, let's start preparing a 1.2.0 release :-) <civodul>mothacehe has made great progress fixing tests but could use help i guess ;-) <mbakke>ah yes, strong authentication and vastly improved 'guix pull' performance, this will be the best release ever :-) <bonz060>Hi. I'm trying to build guix. But after running `guix environment -C guix --ad-hoc bash vim guile-json guile-gcrypt`, I get `No Guile development packages were found. <bonz060>I'm on the lates tip of guix on master: b6329d026b6b5ac58cfcfeb26e786aacf70bd636 ***calher is now known as KE0VVT
<bonz060>aligngator: I've tried that too, but that doesn't work :( <aligngator>bonz060: `guix environment --pure guix && ./bootsrap && ./configure --localstatedir=/var && make check` too? <aligngator>"configure: error: C++ compiler 'g++' does not support the C++11 sta"? Scarry thing. <civodul>i hope we won't have to reopen it ;-) <EatThem>is there a.command to show package options for usage in config.csm like config-file="" *civodul tried "guix time-machine --commit=619972f7b5 -- time-machine -- describe" and it actually works! <leoprikler>apteryx: afaik there are no Guix-specific basic emacs themes <EatThem>okay the right name are service definitions :D <aligngator>EatThem: `guix system search SERVICE_NAME| recsel -p name,location` <aligngator>EatThem: You can without it, but it is a really cool tool. <EatThem>so i get the config Location, but a oneliner that shows all available options for that package would be cool <apteryx>I don't know where I got these or if I had come up with these myself, but I think these look good for the most part: '(ansi-color-names-vector ["#242424" "#e5786d" "#95e454" "#cae682" "#8ac6f2" "#333366" "#ccaa8f" "#f6f3e8"]) <aligngator>EatThem: Probably something with `awk` acting on the manual.html :) <NieDzejkob>mbakke: I'm considering pushing #40469 to c-u, any objections? <EatThem>the man is Inside guix too like info guix <EatThem>so grep would be fine but no built in command? :( <apteryx>You can search inside info manuals with C-s, if this is what you are asking about. <rekado>NieDzejkob: I don’t understand the rationale behind this patch <rekado>why introduce “python-output” instead of using the “outputs” field? <NieDzejkob>rekado: Many of the useful phases of the python build system assume "out" <mbakke>NieDzejkob: perhaps the phases can check for a "python" output and fall back to "out"? <NieDzejkob>mbakke: Good idea, I think that would handle the same usecases <lispmacs[work]>In Gnu Shepherd User Services tutorial, XDG_CONFIG_DIR is not defined in my environment. Should I add some HOME directory to XDG_CONFIG_DIRS, and should I do this from .bash_profile, or...? <rekado>NieDzejkob: I looked at it, but I still think that using “outputs” (a conventional “python” output for example) would be better *rekado sees that jupyter is still broken :-/ <rekado>unless someone already has a fix ready I’m going to unbreak it now <mbakke>unfortunately it's mixed with lots of indentation changes, I haven't had time to study it yet <rekado>can we just revert the problematic commits first? <rekado>it’s not a good look to have a popular package like that broken for more than a week. <rekado>the potential fix also looks pretty ugly to me <rekado>wrapping all those package definitions in prompt-toolkit-2-instead-of-prompt-toolkit <rekado>I don’t see why we can’t just upgrade python-prompt-toolkit and be done with it <rekado>oh, I see it’s the other way around <rekado>then python-prompt-toolkit should be python-prompt-toolkit-3 and python-prompt-toolkit-2 should be python-prompt-toolkit <mbakke>in the following message he also identified a missing propagation somewhere, it would be good to untangle this mess under less pressure :-) <bonz060>aligngator: yeah man. I'm using gcc 6.3.0 which probably threw the error. I fixed it now by using the gcc-toolchain packageid within guix. Much thanks for your help! <mbakke>I wonder how conda or pip install jupyter avoids these conflicting dependencies in the graph <rekado>I have a simple patch to fix this. Can build jupyter, python-ipython, ptython, etc. <rekado>python2-ipython doesn’t build but that’s also broken in current master <rekado>I’m building all packages that have ipython among their inputs <civodul>rekado: a simple fix or a revert would be nice <rekado>we should probably go through the ever-growing list of broken packages one day and fix what we can <rekado>python-ipython-documentation is also broken <rekado>something related to texlive-union <rekado>same problem with python-matplotlib-documentation <rekado>the problem seems to be with permissions on /gnu/store/522gvdar50vmskfd8jaam7av6r3nxish-texlive-generic-pdftex-51265/share/texmf-dist/fonts/map/pdftex/updmap/pdftex_ndl14.map <rekado>maybe I missed something obvious that prompted the complexity in Edouard’s patches. <Bryophyllum>Hello o/ I'm having an issue with upgrading a fresh Guix installation. I have only root account configured, and despite running `guix pull` prior to running `guix system recofingure`, I'm receiving warnings that I should consider running 'guix pull'. <mbakke>Bryophyllum: perhaps your PATH does not contain ~/.config/guix/current/bin ? It should be the first entry, or at least before ~/.guix-profile/bin <mbakke>rekado: thanks a lot for taking care of the ipython issue, LGTM <mbakke>I wonder how hard it would be to implement the systemd Restart=on-failure/on-success semantics in the Shepherd <mbakke>AFAICT Shepherd currently does not know whether a process has exited at all without polling <Bryophyllum>mbakke: Power cycling the VM did the trick, though, this shouldn't be needed to begin with, but thanks for your suggestion ^^ <apteryx>mbakke: shepherd uses some signal fd to know IIRC <apteryx>ah, no, that's for communication with the daemon... hmm <apteryx>mbakke: if shepherd is PID 1, it doesn't do polling, according to shepherd.scm <apteryx>and I was right that when it's PID1, it just seems to listen to signals from its children to know when they die. <apteryx>It was implemented in a recent commit of Ludovic, 'shepherd: Use 'signalfd' when possible.' <mbakke>apteryx: oh, thanks. So it should be possible to store the exit status in some slot during 'handle-SIGCHLD' or thereabouts. <apteryx>I guess! My knowledge pretty much ends here -> . <mbakke>uff, the ssi_status field of the siginfo struct can contain exit status or signal (ie SIGCHLD) <dufresnep>Hi! Had installed guix on Manjaro yesterday... I did guix pull and to my surprise, lot of packages were building... util-linux failed on ipcs/mk-rm-sem and mk-rm-shm ... should I open an issue? <apteryx>dufresnep: if util-linux is not yet know to have problems, definitely! <dufresnep>Is there a web interface to report bugs... or is it by sending on the bugs mailing-list? <dufresnep>found my answer: "or email bug-guix@gnu.org to submit a bug report." <nckx>dufresnep: You can respond to existing bugs using issues.guix.gnu.org, but not create new ones (yet). <mbakke>dufresnep: did you forget to enable binary substitutes? ***janneke_ is now known as janneke
<nckx>apteryx: It's known, but only for an unlucky few, luckily. <dufresnep>Well... I kind of know it is possible to enable binary substitutes... but would have to reread docs to know how <apteryx>dufresnep: how is it related? You can upgrade your kernel without substitutes (given some time & heat). <dufresnep>my idea of guix pull is/was that it would git pull the new .scm files etc... like I said I was surprised it began building stuff <apteryx>guix pull needs to build guix itself (when not using substitutes), which requires a bootstrap, hence downloading and building stuff <mbakke>dufresnep: to enable substitutes you can run: sudo guix archive --authorize < /var/guix/profiles/per-user/root/current-guix/share/guix/ci.guix.gnu.org.pub <mbakke>it will create /etc/guix/acl with the signing key of the CI server <dufresnep>looks like my kernel 5.6.11 is 'affected by the issue'... would enabling substitutes be ok to do without changing kernel? *nckx PSA: {dmitri,sergei}.tobias.gr currently down for planned maintenance and totally not ‘oh shite they still can't build util-linux either’. <nhfcz>Has the bug where guix would complain about locale not found been fixed yet? <bonz060>In guix, when running a binary from a given profile, is there a way to make the command refer to binaries from that profile. I'd like somescript.rkt in `~/opt/myprofile/bin/racket somescript.rkt` to refer to binaries inside `~/opt/myprofile`. How do you do that? <mbakke>nhfcz: guix complains if it can not find locales for your current version of glibc, if that is the case you probably need to upgrade guix for the root user (on a foreign distro), or reconfigure && reboot (on Guix System). <nhfcz>What if it complains about a fresh install? <mbakke>nhfcz: if you ran 'guix pull' that is expected, as the glibc has been updated since 1.1.0 was released <nhfcz>What is the first thing I should do after a fresh install? <mbakke>I think updating is a popular post-install activity <nhfcz>Last time I used GNU guix, I substituted right after installation. Install guide never mentioned updating. <mbakke>hacking on the shepherd was a delightful experience, now to try writing some tests for this thing <apteryx>mbakke: yay :-) the code base is refreshingly light, right? <mbakke>apteryx: indeed, about 2.5k SLoC :-) <jsoo>sneek: later tell NieDzejkob: how do you use racer currently? <nhfcz>Just so I understand what i'm doing. After a fresh install I pull and then do -u to update everything. Am I right? <jsoo>nhfcz: let me think. i may be wrong, but due to guix' declarative nature, your system should be up to date already <mbakke>nhfcz: are you on Guix System or a foreign distro? <dufresnep>just to let you know that guix pull completed (with substitutes enabled) <mbakke>nhfcz: on a foreign distro, I think the root user has a single package installed: glibc-locales <mbakke>"guix pull && guix package -u" will update that <mbakke>Bryophyllum: looks like the server is currently down :/ <bdju>there is an actual wishlist on the libreplanet wiki dufresnep *jsoo changing rust-1.19 in order to have a different rust-1.44 <sneek>Welcome back NieDzejkob, you have 1 message! <sneek>NieDzejkob, jsoo says: how do you use racer currently? <roptat>it fails on almost every commit, but I don't see any error in the log <mbakke>nhfcz: can we trace back a bit, did you install guix using the shell installer? are you getting the locale warning for the root user or your current user (or both)? <nhfcz>Here's what happened the last time I installed guix <nhfcz>I installed the manual step by step way. Did the application step. All as root. And it gave me a constant warning about locales. <mbakke>nhfcz: right, there are two possible sources for that warning, one is that $GUIX_LOCPATH is not set (or the wrong value) for the current user; the other is that the guix-daemon is using the wrong locales <jsoo>NieDzejkob: ah ok, thanks anyways. the rust 1.4X patches are excellent <nhfcz>So next time I install guix, what should I do to avoid this? <Kozo>Hey, what is Guix's alternative to putty? <nhfcz>Putty? Isn't that the windows ssh utility? <Kozo>Sorry, more specifically, I need to console into an Aruba switch to configure it. I need the serial portion of putty <mbakke>nhfcz: just make sure GUIX_LOCPATH is set (and glibc-locales is installed) for each user. When there is an update to glibc-locales, you may have to restart guix-daemon.service as well. <Kozo>I can serial using emacs... what have I been missing out on <dufresnep>On Manjaro... I 'guix package -i hello' ... but I can't execute it directly ... it's not in /home/paul/.config/guix/current/bin:/home/paul/.local/bin (where only guix and guix-daemon are) <NieDzejkob>guix should tell you something along the lines of "consider setting $GUIX_PROFILE and then sourcing this file" <NieDzejkob>but I don't remember what that file is called... <dufresnep>I see the file ... did . ~/.guix-profile/etc/profile ... but don't seems to change my PATH <NieDzejkob>dufresnep: look at the file with a text editor and see what variables it's setting <dufresnep>echo PATH="${GUIX_PROFILE:-/gnu/store/3kpv7imn0g3fvfzzibxh7y85k7dwkgsa-profile}/bin${PATH:+:}$PATH" <dufresnep>PATH=/home/paul/.config/guix/current/bin:/home/paul/.config/guix/current/bin:/home/paul/.config/guix/current/bin:/home/paul/.config/guix/current/bin:/home/paul/.local/bin:/usr/local/bin:/usr/bin:/bin:/usr/local/sbin:/usr/lib/jvm/default/bin:/usr/bin/site_perl:/usr/bin/vendor_perl:/usr/bin/core_perl:/var/lib/snapd/snap/bin <dufresnep>echo echo ${GUIX_PROFILE:-/gnu/store/3kpv7imn0g3fvfzzibxh7y85k7dwkgsa-profile} return /home/paul/.config/guix/current and it is already 3 times at the start of my PATH <mbakke>dufresnep: are you sure you sourced ~/.guix-profile/etc/profile? looks like you have only sourced ~/.config/guix/current/etc/profile (multiple times) <dufresnep>but the two seems to do the same to my $PATH <dufresnep>cat ~/.guix-profile/etc/profile have export PATH="${GUIX_PROFILE:-/gnu/store/vlldq86kkjhdigzpcgb6dnlliyhvnk7s-profile}/bin${PATH:+:}$PATH" (and the path have hello file) <apteryx>if so, unset it then source again ~/.guix-profile/etc/profile <dufresnep>echo $GUIX_PROFILE give /home/paul/.config/guix/current <dufresnep>I am afraid not understanding what is supposed to be my value for GUIX_PROFILE <dufresnep>Ah... it is supposed to be empty... hello works now! <dufresnep>frankly I am still confused... because when I install a package... guix hint me to GUIX_PROFILE="/home/paul/.guix-profile" <rekado>dufresnep: it should not be empty. <rekado>dufresnep: it should have the value that the hint tells you <rekado>note that /home/paul/.guix-profile and /home/paul/.config/guix/current are different profiles <dufresnep>the first for having guix ... the other for installed packages ?? <rekado>the former is just for “guix pull” <rekado>the other is the default location for installed packages <jsoo>dufresnep: it might be worthwhile to explore the contents of those directories a little to get a feel for what they are <dufresnep>yeah... that's what I am doing... they are the opposite of what I said <jsoo>dufresnep: there is some documentation on profiles, also: see `info guix profile' <jsoo>(plus more). they (profiles) are quite central to operating the guix package manager <dufresnep>they are the kind of things you would want to work without trying to understand them ;-) <dufresnep>also $VAR I understand ... ${ some VAR some } ... confuse me <apteryx>dufresnep: info '(bash) Shell Parameter Expansion' <apteryx>${PARAMETER:-WORD} -> If PARAMETER is unset or null, the expansion of WORD is substituted. Otherwise, the value of PARAMETER is substituted. <dufresnep>thanks! ... I will reboot and see how much trouble it is to have things works again <dufresnep>Ok! I added export PATH=/home/paul/.guix-profile/bin:$PATH to my .profile and things works pretty much like I expect... thanks everyone! <dufresnep>I had already export GUIX_PROFILE="/home/paul/.config/guix/current" and . "$GUIX_PROFILE/etc/profile" and xport GUIX_LOCPATH="$HOME/.guix-profile/lib/locale" in my .profile file <NieDzejkob>dufresnep: you shouldn't need to put the directory in $PATH yourself <NieDzejkob>there are many other variables that will need to be handled like this <NieDzejkob>$GUIX_PROFILE is not a variable that should have one specific value, it should be set to the path of each profile before you load it <NieDzejkob>this should all be handled automatically by /etc/profile <rekado>I installed another build node and it’s been idle the whole day <NieDzejkob>rekado: let's hope cbaines' build coordinator improves things ***dingenskirchen1 is now known as dingenskirchen
<civodul>rekado: i think we still haven't deployed the fixes reepca made <civodul>rekado: BTW, i just noticed that logs. is down