<htsr>leoprikler: what's the differences between normal and devel? how can I know which one is installed? <leoprikler>it seems devel also includes documentation for experimental features <leoprikler>oh my, the online manual (at least devel) highlights parentheses <rekado>htsr: “devel” is a moving target. It’s always going to be the latest unreleased state. The normal one is for the last official release. <leoprikler>only those that come with an ISO and everything? <rekado>releases are announced, get a proper version number, and have signed tarballs, etc <htsr>what's the default value of --cores for guix pull? I want more speed ;) <reepca>when build-cores is not specified, the value returned by (current-processor-count) is used. <reepca>er, when --cores is not specified (build-cores is what the variable is named internally) <htsr>I'm trying to create a new service cron using the scron package but I have this error : <sneek>Welcome back htsr, you have 1 message. <sneek>htsr, nckx says: Note that --cores only affects build phases that respect something like ‘make -j’. Some don't. You can use --max-jobs build more *packages* in parallel. However, this means you'll have anything from 1 to (* cores jobs) jobs running at once (-c6 -M6 = possible load average of 36). You'll need to find a balance. <htsr>nckx: thanks for the info! <htsr>guix system: error: exception caught while executing 'start' on service 'cron': <htsr>Throw to key `match-error' with args `("match" "no matching pattern" "/gnu/store/lban3cimaz4c9vz54lvzk3zi3rikx685-scron-0.4/bin/crond")'. <htsr>well no, how do you debug shepherd services? <htsr>I've finally fixed my service! <bsima>does guix interop with nix at all? I have a bit of nix code and I'm wondering how to port it ***oldosfan_ is now known as oldosfan
<bsima>ah, i found in the README that the derivations can be re-used. <bsima>oh i also found then (gnu services nix) module, looks like i can use this to run regular nix stuff ***ng0_ is now known as ng0
***ng0_ is now known as ng0
<bdju>yes, I'm using sddm because it can start a wayland session and gdm couldn't <efraim>bdju: which DE are you using? can you share your os config? I'm trying to get enlightenment-wayland to work <efraim>thanks. looks like my problem is with enlightenment-wayland itself then <leoprikler>by the way, what stops Guix GDM from using Wayland? On other systems, GNOME Wayland is already standard <bdju>I was thinking the same. maybe our version is out of date. I never really looked into it that much <leoprikler>3.30 should not be "out of date" w.r.t. other distros <bdju>seems to be one of the inputs if I do "guix edit gdm" and then search for "wayland" <bdju>oh I think it jumped me to gnome-settings actually <bdju>wayland is not an input for gdm that I see <leoprikler>It appears, that GDM needs to be compiled with --enable-wayland-support to support Wayland <leoprikler>The build still succeeds despite them -- what should I do? <mbakke>leoprikler: those errors are "normal" and can be ignored <pinoaffe>leoprikler: okay, I'll guess I'll let that be for now, thanks for the info! <verbasidor>I was really a fan of GuixSD and Guix package manager , but after what these kids shown from disrespectful and back stabbing RMS nah its not a welcomed community to begin with <verbasidor>disgusting behavior , so all who participated/supported that you deserve a fuck you and tfffoooo on your face <sneek>Welcome back Formbi, you have 1 message. <sneek>Formbi, nckx says: Someone was able to reproduce your bug: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=37688 . Seems likely that ‘--fallback’ was a red herring; the command simply happened to work the second time because it's somewhat random. Not good, but at least you're not alone. ***apteryx_ is now known as apteryx
<roptat>is there anything we could turn into a bug report to improve things a bit for users? <dongcarl>I want to reply to him as he seemed like he wants to contribute back: "What I haven’t gotten to yet, is using Guix to study, change and maybe even contribute back to projects. This includes Guix itself. The [guix enviornment] command is critical in this field." <roptat>actually, it's the first time I hear about it :) <dongcarl>It's like Hacker News but more technical I would say (and sometimes more noise) <civodul>dongcarl: i think we should turn this post into a list of actionable bug reports <dongcarl>That makes sense, I'm asking for an invite to Lobsters and seeing if I can engage with the author <dongcarl>bandali: It would seem that you have a lobsters account, would you be able to extend an invite? <bandali>dongcarl, sure, /msg me your email ? <roptat>it took more than a year, but the patch is commited :) <dongcarl>roptat: that's great! all ant builds will benefit from this? <cbaines>I've just ended up in the Guix interactive installer (almost by accident), it's pretty exciting! <roptat>any ant build that uses the tstamp task, yes <htsr> how can I start a script at each reboot? By modifying the reboot command of shepherd? <jgart>Any help or guidance is greatly appreciated <htsr>jgart: darkhttp is a package so you need to use something like this (file-append darkhttpd "/bin/darkhttp")) <mbakke>htsr: do you mean running a script at shutdown? <htsr>mbakke: no, after a reboot <mbakke>roptat: nice! should we patch our ant in the mean time? <civodul>cbaines: the interactive installere for Guix System you mean? <cbaines>yep, seems to start automatically with the ISO image <jgart>htsr: my darkhttpd lives in /gnu/store/hash-number/bin/darkhttpd <jgart>htsr: does the file-append procedure add the hash automatically? <mbakke>htsr: tricky :-) I don't think the shepherd has any hooks for that, nor mcron. <mbakke>though i suppose you could hijack the 'reboot' command and add a bit of state for the startup scripts <htsr>jgart: yes the darkhttpd correspond to /gnu/store/hash-darkhttpd-version/ <yo99>is there any way to get KDE on GuixSD? <jgart>htsr: thank you, I'm taking a look now <roptat>jgart, basically, file-append takes a package object and a path, and appends the path to the store path of the package <roptat>so yes, the hash is part of it :) <htsr>jgart: you should rename your service to darkhttpd-service. And does darkhttpd create a pid file? In /var/run/ for example <roptat>and you don't have to guess it yourself : <htsr>the scron package is kinda useless without a service, when I will have more time in a few weeks I might send my service definition for review <htsr>I need to improve it before <jgart>htsr: darkhttpd does not seem to create a pid file. I checked in /var/run and there is nothing from darkhttpd there <jgart>Should I be looking elsewhere? <htsr>jgart: so you don't need #:pid-file <htsr>what's the command to show the path of an installed package? <htsr>leoprikler: thanks! weird name btw <leoprikler>for the command? it actually builds it if it isn't already built <roptat>well, it's the command to build the package, which ends by giving you the store path for the package <roptat>if the package is already built, it does nothing, but gives you the store path <jgart>leoprikler: dpaste.com/3M97JDY <jgart>that's what I get from `ps aux | grep darkhttpd` <leoprikler>okay, so we've confirmed that it is indeed running <jgart>yes, darkhttpd is currently serving the content <jgart>it's hosting a few reveal.js slides for testing purposes here: jgart.xyz <htsr>not very user friendly ;) "build" is kinda scary if you just want the path <leoprikler>"get-path" is less friendly if you don't expect a build ;) <leoprikler>alternatively, you can dive into your profile and follow symlinks <htsr>leoprikler: what about putting this information in guix describe? <leoprikler>guix describe describes Guix itself, not packages <htsr>and what's the command to have the sources of a package? not the definition <htsr>guix download the link from guix edit? <jgart>leoprikler: dpaste.com/2YM4N23 <jgart>I tried starting that with `sudo shepherd -c darkhttpd-service.scm` <jgart>it did not return an error but `herd status` does not see it after I ran that <leoprikler>That is actually not a well-formed shepherd config. <leoprikler>You're mixing Guix procedures with Shepherd ones. <bluekeys>I seem to have a login manager that I can't remember installing, but I don't see it in the list of installed packages. Any ideas how I can get rid? <leoprikler>jgart: try adding the define to your /etc/config.scm and then add a (service darkhttpd-service-type) to services <leoprikler>bluekeys: your login manager should be part of /etc/config.scm (or whichever file you use in its stead) <bluekeys>It does look like gdm to me too leoprikler, could I remove it and just put an entry in xinitrc to load my window manager? <htsr>don't forget to add (use-package-modules web) in your /etc/config.scm and (service darkhttpd) <leoprikler>bluekeys: if you just want to autologin, you should be able to configure that as part of gdm-service-type <leoprikler>there are also some other desktop managers that you can try instead <leoprikler>I'm not entirely sure why you'd want to go completely without one -- what distro are you coming from? <bluekeys>I'm using stumpwm, I've used various distros, wouldn't say I came from one in particular. <bluekeys>I'd like to just switch the laptop on, and be logged in as my standard user, security be damned. <leoprikler>Set auto-login? to #t and default-user to your user in gdm-configuration <bluekeys>Nice, except I don't have gdm-configuration section... Can I just add it? <leoprikler>you will have to do something along the lines of (modify-services %desktop-services (gdm-service-type config => ...)) <htsr>mbakke: I've looked at the reboot function of shepherd, it use the libc reboot function "as declared in <sys/reboot.h>" :/ might be too tricky for me <htsr>can't find the reboot.h file in include/sys/ ... libc is in glibc package right? <bluekeys>Ahh, it's a little complicated for me... I really don't understand scheme at all yet. leoprikler, i'm trying to read what you've referenced in my emacs, but lots of characters are not displaying properly, if I could fix that, then I could read the manual. The characters are showing as squares with 4 characters inside <leoprikler>jgart: you need the darkhttpd-service type that you've defined earlier in the wrong file <htsr>jgart: you need the darkhttpd-service type in your config.scm <leoprikler>bluekeys: afterwards it's "m System Configuration RET m Using the Configuration System RET" <bluekeys>I can't see the guix manual properly in emacs. It doesn't show many of the characters properly. I'll try the config you just explained... brb <jgart>hmmm do you know of a template config.scm that I can look at that shows the define command how you are describing? <leoprikler>copy the define of your other file and paste it into your config.scm before the (operating-system ...) part <jgart>leoprikler: ok let me try that <jgart>leoprikler: do I run `guix system reconfigure config.scm`? <jgart>leoprikler: ok thanks, let me try <htsr>you might kill darkhttpd so you're sure it's started from shepherd <htsr>if I create a new package for the debian cron, should I be able to start scripts after reboot with @reboot or is it incompatible with guix? <leoprikler>htsr: depends on what debian-specific "inputs" that cron needs <jgart>after trying to reconfigure my system I am getting this error: <jgart>guix system: error: python-pep8-stdlib-tokenize-compat.patch: patch not found <jgart>I was trying to guix system reconfigure config.scm in order to see if my new darkhttpd service definition im my config.scm would not fail <leoprikler>do you yourself reference that patch anywhere? If not, your Guix is probably broken <jgart>leoprikler: hmm no I've never referenced it anywhere <jgart>I haven't guix system reconfigured root user in a while <jgart>leoprikler: could that have broken my system? <leoprikler>With "your guix", I meant ~/.config/guix/current <jgart>that was run without sudo priviledges <leoprikler>can you do a guix pull first and then a reconfigure? <leoprikler>you did not by accident run "guix install guix", did you? <jgart>leoprikler: do i run that as a sudo user? <jgart>leoprikler: I haven't run guix install guix <leoprikler>guix pull as normal user and don't install guix in any profile <jgart>leoprikler: ok will remember that. Thanks! running `guix pull && guix system reconfigure /etc/config.scm` now <jgart>after running the above two commands I get the same error <jgart>guix system: error: python-pep8-stdlib-tokenize-compat.patch: patch not found <jgart>guix describe: error: failed to determine origin <jgart>hint: Perhaps this `guix' command was not obtained with `guix pull'? Its version string is 1.0.1-9.f63e493. <jgart>/home/jgart/.guix-profile/bin/guix <jgart>hmm I don't remember running `guix install guix` ever <jgart>leoprikler: I've run "guix pull" as user on several occasions <jgart>just finished running guix package -r guix <leoprikler>your guix should point to ~/.config/guix/current/bin/guix <leoprikler>run `which guix` again (perhaps in a new terminal) <jgart>I get /home/jgart/.config/guix/current/bin/guix <leoprikler>now `hash guix` for safety and then `sudo guix system reconfigure ...` <jgart>ok I ran both of those. Just waiting now <jgart>I got the failed to install locale this time but it continued through the process <jgart>leoprikler: thank you by the way for helping me with this <jgart>leoprikler: do you build in an alternate directory? my / folder on my vps tends to get filled up easily when installing or updating software <leoprikler>builds themselves are run in /tmp, but packages are stored in /gnu <jgart>leoprikler: I've been using guix gc but given that I am running guix in small vps partition it fills up to 100% before it can finish it's task sometimes <jgart>my root partition is only 8GB <leoprikler>guix deploy builds packages on your local machine and then sends them to remote <leoprikler>that said, 8GB is still quite limiting, especially when Emacs is 1GB in size <jgart>does the local machine also have to have guix package manager installed? <jgart>My local machine is running arch linux. I do have guix installed. I installed it from the binary that's available <jgart>leoprikler: so I run guix deploy from my arch linux machine that has guix installed on it? <jgart>guix deploy will take care of sending the binaries over to the server? <jgart>ok now I have some guix deploy homework to do :) <leoprikler>It doesn't seem able to do user profiles, though, so you have to install all packages system-wide <jgart>leoprikler: hmm good to know. Thanks again for all the help. see you later! <bluekeys>Hi guix, my emacs info windows dont show right, I think it's because of my locale or font choice. I'm using fira-code and guix is set to locale of en_GB.utf8. Did I do something wrong? <divansan`>that while trying to guix system reconfigure .scm <roptat>divansan`, I wonder what /gnu/store/bmvvwq2lx7ap3sfy* is? <roptat>maybe also share your config and the output of "guix describe" <roptat>looks like your agetty service doesn't load the module <roptat>looks like 62c2217570fb3a3437e8316d3679b7fe13db0884 is related, too <roptat>how did you run "guix system reconfigure" exactly? with sudo? as root? <roptat>can you try "sudo guix describe" (if you used sudo) or guix describe as root, if you ran reconfigure as root? does it give you the same commit? <roptat>if so, then I think you have found a bug, could you send a report to bug-guix@gnu.org? <kirisime>I'm trying to increase my file descriptor limit, but I can't seem to get it to work in GNOME. <kirisime>I've added a pam-limits-service, and the limits are set correctly when logging in in a tty and with su <kirisime>But logging in to GNOME with gdm doesn't respect the configuration generated by the pam-limits-service and I'm not sure how to work around it