IRC channel logs

2020-02-21.log

back to list of logs

<nckx>On the plus side, you can choose which terminal to add to /etc/config.scm. Then run sudo -E guix system reconfigure /etc/config.scm and reboot.
<PotentialUser-28>lfam: thanks! it works.
<PotentialUser-28>nckx: okay, ill try that.
<nckx>PotentialUser-28: Actually, if you use anything other than ‘st’ you'll probably have to edit your i3 configuration. Didn't think about that. But you're an i3 user, you'll figure it out.
<nckx>Don't forget to add dmenu and i3status as well if you want them.
<vagrantc>all this fiddling with guix on pinebooks, i finally saw an example of how to properly set ctrl:nocaps on guix system! :)
*vagrantc thanks janneke again
<PotentialUser-28>nckx: thanks a lot!!! everything works perfectly now.
<PotentialUser-28>lfam: thanks for your help too :)
<nckx>Welcome to Guix! \o/
<nckx>Now you have to fix all bugs.
<lfam>Cheers :)
<lfam>Yes, we are waiting ;)
<PotentialUser-28>haha, okay :D
<shader>how hard do you think it would be to integrate org-mode with haunt via guile's elisp support?
<hacktivista>I'm trying to use the guix environment --container as means for isolation of a wordpress application... it works mostly OK, but I have two problems:
<hacktivista>1- I have my own folder... say ~/guix/etc where I want to expose some files like php.ini php-fpm.conf my.cnf, etc... exposing files does not work and if I try to expose the entire ~/guix/etc=/etc it fails, with no errors... I can get around this creating soft links
<hacktivista>2- I think I reached a dead-end: `chown: invalid user : 'mysql'` if I see /etc/passwd indeed there's no mysql user, I can't use useradd inside a container... the unique option would be to place the mysql user by hand... that seems too hacky really
<hacktivista>I think this qualifies as a bug, as the container should consider the users created by the packages I'm using on it... where can I file a bug report or something related?
<hacktivista>My bad, I just saw the link above
<hacktivista>hmmm... now I'm thinking that probably the mysql user is never created... I could get around it with a configuration in my.cnf
<hacktivista>there are lots of applications which use custom users for its usage... how does guix handle this cases?
<roptat>hacktivista: each service can declare an extension to add users
<roptat>The mysql service defines a mysql user. I.m not sure, but maybe you need to reboot for the user to be created
<hacktivista>I'm not using guix as OS, but as package manage
<hacktivista>I really don't have
<hacktivista>'services'
<shader>are there any tools for generating golang packges, because dang there are a lot of dependencies on this project
<hacktivista>I want to use it as a cross platform dev partity means
<hacktivista>(cross platform as "it runs on every gnu/linux distro")
<lfam>shader: Not at the moment
<lfam>shader: I know that people want to make an importer for Go packages because, as you've found, the dependency graphs in Go are huge
<roptat>hacktivista, then I didn't get enough context :)
<nckx>NieDzejkob: Now I want to read ‘scheme-macro-phd.pdf’… Sorry for the rambling reply, I enjoyed this way too much.
<lfam>nckx: That bug 😂
<nckx>It was a wild ride.
<roptat>since I'm going to libreplanet and they're still looking for lightning talks, I was thinking I could present guix. What's a good topic for a 5 minute talk?
<roptat>I could go crazy and present the home manager ^^
<raghavgururajan>roptat TOPIC: How to unscrew a screwup in one step.
<raghavgururajan>roptat "--roll-back". ;-)
<genghis-sean>Hey #guix, I am new to guixSD and I'm having dns resolution issues. I think shepherd is running the networking service, but I don't know where to find logs that may be helpful.
<roptat>what does your config say? you're using the dhcp-client-service, or maybe the desktop-services?
<genghis-sean>I'm using desktop-services. Let me check for the dhcp-client-service
<roptat>I think the desktop-services include network-manager
<genghis-sean>yes, the manual seems to indicate that https://guix.gnu.org/manual/en/html_node/Desktop-Services.html#Desktop-Services
<roptat>so it should have created a /etc/resolv.conf, check that file
<roptat>also, what makes you think you have DNS issues?
<genghis-sean>ntpd is erroring with "error resolving pool"
<genghis-sean>the resolv.conf file has the router IP I'm connected to as the name server
<roptat>Should be fine…
<genghis-sean>The ethernet to the router was working when I first booted the system. I then ran guix pull followed by sudo guix system reconfigure /etc/config.scm and the new configuration has something broken
<roptat>oh, I get this too, it's not a DNS issue
<genghis-sean>I didn't check /etc/resolv.conf before when it was working, but I suppose I could roll-back and see what it was when it was working
<roptat>ntpd[421]: error resolving pool 0.guix.pool.ntp.org: Name or service not known (-2)
<genghis-sean>yea, that's it
<roptat>bubut ntpd is still running
<genghis-sean>would I check that with herd status?
<roptat>yes, herd status ntpd
<genghis-sean>It is running with value 340
<roptat>you could also check you have a running ntpd process, with "ps aux | grep ntpd" for instance (there should be two line: one for ntpd and one for grep)
<roptat>340 is its pid
<roptat>so the error is not fatal
<genghis-sean>oh ok. Yea it is running in ps
<roptat>I wonder why it can't resolve the name, since it exists: 0.guix.pool.ntp.org (204.2.134.164)
<roptat>maybe it tries to resolve too early?
<roptat>it seems to be a common thing even on other distros
<roptat>ok, not necessarily that IP, there seems to be many
<genghis-sean>I can't resolve other addresses with ping or curl so that's why I thought it was dns resolution
<roptat>Oh
<roptat>Can you try installing bind:utils?
<genghis-sean>The substitute can't resolve so it didn't work :(
<roptat>gah
<lfam>I would try restarting the nscd service
<genghis-sean>herd restart nscd?
<roptat>ah yes, good idea
<genghis-sean>it restarted, but still no dice
<lfam>So what's in /etc/resolv.conf
<roptat>is your computer connected? Can you ping an IP address?
<apteryx>process of the services in Shepherd must not detach themselves, right?
<apteryx>I'm experimenting with the nfs-service-type, and rpc.mountd is respawing too quickly and getting disabled by shepherd. When I launch it manually, it quickly detach. The process is shown in 'ps -eF'.
<genghis-sean>let me try pinging that IP you posted (204.2.134.164)
<genghis-sean>I can't ping IPs either
<genghis-sean>so I guess it's not DNS heh
<genghis-sean>I tried the router IP that it's connected to and still got "Network is unreachable"
<lfam>Is the link up?
<lfam>`ip link show`
<genghis-sean>it just has loopback so I don't think so
<apteryx>roptat: the ntpd error is just early in the boot, or ongoing?
<genghis-sean>ongoing
<lfam>Oh, I think you hit <https://bugs.gnu.org/39671>
<lfam>Did you reconfigure recently?
<genghis-sean>yes
<lfam>If so, reboot and pick an older generation from the GRUB menu
<genghis-sean>it was the first reconfigure after reboot
<roptat>only once for me
<genghis-sean>sorry after installation I mean
<lfam>Then after reboot, do `guix pull && guix system reconfigure ...`
<lfam>There was a bug for a few days on the master branch :/
<genghis-sean>oh ok
<genghis-sean>that's a straightforward fix
<lfam>It is now that the bug was fixed :)
<lfam>But yeah, at least you can go back to older systems from GRUB
<apteryx>roptat: yes I think it resolves too early (and last I check there didn't appear to be enough granularity in the dependency chain to fix that).
<roptat>ok, well it's not fatal so it's ok
<genghis-sean>I saw this "New
<genghis-sean>"New gdm session c1" followed by "Session c1 removed"
<genghis-sean>and thought that was normal
<apteryx>roptat: yeah. I'm just perfectionist and wish /var/log/messages had no extraneous error ;-)
<apteryx>there's also that pcspkr thing.
<lfam>genghis-sean: To be clear, do the `guix pull && guix system reconfigure ...` as root
<genghis-sean>I was wondering if guix pull needed to be run as root. The docs say do `guix pull && sudo system reconfigure ...`
<lfam>It gets tricky. I would just do it as root
<lfam>At least do `guix pull && sudo -E guix system reconfigure ...`
<lfam>I think there must be some bugs in `sudo guix system ...`
<genghis-sean>good to know
<nckx>lfam: Such as?
<genghis-sean>the docs should really be updated if that's the case
<lfam>I'm not sure but I feel like people are often getting messed by using sudo on its own here
<genghis-sean> https://guix.gnu.org/manual/en/guix.html#After-System-Installation
<raghavgururajan>genghis-sean I got same "New gdm session c1" followed by "Session c1 removed" error. GDM did not start.
<lfam>May just be when they do `sudo su`
<raghavgururajan>I reported it as a part of #39671
<drakonis>gdm is borked eh?
<nckx>genghis-sean, lfam: There aren't any known bugs with sudo guix on Guix system.
<genghis-sean>and this is gdm right? https://gitlab.gnome.org/GNOME/gdm
<drakonis>that exxplains why my little out of tree experiment had failed
<nckx>I just checked and ‘sudo guix’ does what you'd expect from the manual (use your regular user's guix).
<drakonis>incredible.
<lfam>Maybe it's just my setup but on Debian plain `sudo` breaks the environment for all the Guix stuff
<lfam>Leaves a very minimal environment
<nckx>The confusion may stem from other distributions, where sudo caan call either your Guix or root's Guix or prank-call your mother.
<genghis-sean>nckx: The manual points out that it runs your user's guix (because PATH is preserved) but doesn't explain when you would want to run root's guix vs your user's
<nckx>genghis-sean: There is no reason to run (or even ever pull) root's guix, unless you have some use case (and then you know what you're doing).
<nckx>So the manual doesn't cover that, just says what you should do.
<genghis-sean>that makes sense
<apteryx>shouldn't 'guix search' in M-x shell return all the results without nagging me with | less hint? This is a dumb terminal.
<nckx>lfam: IME on foreign distros you want to always explicitly use ‘sudo -E’ (use my guix, as root) or ‘sudo -i’ (use root's guix, as root) to be safe.
<nckx>Especially if you distro-hop.
<apteryx>lfam: sudo on Guix System is alright. On other system they mess with the PATH for security reasons. But not on Guix System.
<lfam>Alright
<apteryx>I mean, 'sudo guix system reconfigure ...' should work.
<apteryx>(agreed, it's confusing that sudo's behavior differs between distros)
<nckx>genghis-sean: On non-Guix Systems it can make sense to pull as root because the root profile serves as a poor person's system profile. Not the case on Guix System so you can plug your ears and just remember ‘sudo guix works’ for now.
<genghis-sean>I've been running guix on a non-Guix system for a while, but I didn't ever run it as root
<nckx>Good.
<nckx>Better that than the other way 'round I mean.
<genghis-sean>why sudo make me a sandwich when the sandwich gets better without all that shouting
<nckx>Hehe.
<apteryx>so... our nfs-service-type seems broken (rpc.mountd keeps respawning and Shepherd disables it)
<apteryx>I'll try to run what Shepherd does manually to get a hang of things
<apteryx>seems ntpd needs a switch to "not be daemonized" to prevent the same problem. I wonder if something's broken in Shepherd.
<guix-vits>nckx: thanks for operator |& ...
<jackhill>apteryx: (re M-x shell): have you seen the INSIDE_EMACS disucssion: https://lists.gnu.org/archive/html/guix-devel/2020-02/msg00029.html ? I believe it came up there.
<guix-vits>jackhill: link returns 404
<jackhill>guix-vits: hmmm, works for me ::shurg::. Anyways, looks for a thread with INSIDE_EMACS in the subject on the February archives of the guix-devel mailing list
<jackhill>it's about what the right behavior of guix tools should be when that variable is set. And also, that emacs only sets it sometimes (it seems to be related to modes that are derived from comint mode), but notably not in eshell.
<jackhill>which causes problems for the cool hyperlinking becasue eshell doesn't process or ignore the special characters
<guix-vits>jackhill: links is works, just i'd omitted l in .html. Thanks for summary.
<Kriocafr>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Kriocafr>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Kriocafr>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Cheeme>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Cheeme>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Cheeme>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Chetoe>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Chetoe>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<Chetoe>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<aphoi>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<aphoi>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<aphoi>hi, please kindly stay out of #freenode. the pool is closed. thanks.
<guix-vits>Woof!!!
<guix-vits>Woof!
<guix-vits>Woof-Woof-Woof!
<guix-vits>Woof?
***apteryx_ is now known as apteryx
<civodul>Hello Guix!
<gskru>Hello everyone!
<gskru>I just finished the installation of Guix (SD) on my pc and I'm receiving an error since the first boot which ends with "Error: You need to load the kernel first" Any ideas? My Guix installation is on an encrypted partition with separate /boot and /boot/efi partitions. Any ideas?
<civodul>hi gskru!
<guix-vits>gskru: btrfs?
<civodul>gskru: could you share a screenshot?
<gskru>No btrfs. Standard ext4. Where can I share the photo?
<guix-vits>gskru: ASCII :), please...
<gskru>Let me type it in...
<janneke>Hello Guix!
<guix-vits>sneek: later tell gskru: https://guix.gnu.org/contact/irc/ in kiwi, aren't there a button for sharing photos, etc...
<sneek>Got it.
<gskru>Error: No auch device: \n
<sneek>gskru, you have 1 message.
<sneek>gskru, guix-vits says: https://guix.gnu.org/contact/irc/ in kiwi, aren't there a button for sharing photos, etc...
<gskru>'gnu/store/{long-string}/linux-libre-5.1.2/bzImage' Not Found
<civodul>hey janneke!
<civodul>gskru: could you send all the details to bug-guix@gnu.org?
<gskru>there isnt a button to send images I'm on a phone
<civodul>namely: the config you chose (ext4, separate /boot, EFI, etc.) and a picture of the screen
<gskru>ok
<civodul>that way we can try to reproduce it and see
<civodul>TIA!
<davidl>When running guix gc Im getting an error which says "guix gc: error: statting `/gnu/store/.links/19851485104hfjdg9y132rlh': Invalid argument" - does anyone know what it means or what I should do about it?
<guix-vits>davidl: Stand-alone, or on-top?
<davidl>guix installed on a RedHat box
<davidl>guix-vits: so it first did everything I think it should, like deleting garbage, deleting /gnu/store/trash, deleting unused links... but then final line in CLI output was the error message I quoted.
<guix-vits>davidl: Maybe it's applicable: https://guix.gnu.org/manual/en/guix.html#Limitations-1
<guix-vits>"guix gc cannot access arbitrary links to profiles. ..."
<guix-vits>is ths `/gnu/store/.links/...' a simlink?
<guix-vits>*symlink
<davidl>not sure Ill check
<davidl>Ill also check that link
<davidl>I think I found the error in the code here: https://github.com/guix-mirror/guix/blob/7bd000d104f79f2b1c51c48c9d01f247189f9935/nix/libstore/gc.cc#L179
<davidl>as in where the error is thrown
<davidl>guix-vits: that file is not a symlink
<guix-vits>davidl: then my posts above is rather a flood...
<davidl>*link above should end with #L612
<davidl>guix-vits: not sure what you mean
<guix-vits>davidl: that means, my posts are useless, if the offending file isn't a symlink, which is the case here.
<davidl>ah ok
<guix-vits>davidl: re-running gc throw the same error?
<leoprikler>does anyone else encounter compilation errors since the introduction of copy-build-system?
<leoprikler>specifically its documentation seems borked due to missing strings
<nckx>leoprikler: Error?
<nckx>& good morning Guix.
<leoprikler>yep, guix.de.info fails to build
<leoprikler>there are some strange things going on with @end itemize not being on its own line and what not
<leoprikler>it might just be me, though
<leoprikler>I have a commit, that was never pushed, still on my local branch
<nckx>leoprikler: I couldn't reproduce this with ‘make’, then tried to ‘guix pull’ in an unrelated window and boom ☹ I've just removed the spurious indentation and that seems to fix it.
<guix-vits>nckx: indentation?
<jlicht>hey guix!
<nckx>‘Yes’. Because I don't know how to answer that.
<nckx>o/
<guix-vits>nckx: "removed the spurious indentation", what it's mean?
<nckx>guix-vits: Removed whitespace before Texinfo commands that was causing errors. See commit e90e64049ce160d28d1e8b3014badcc2b214627c.
<davidl>guix-vits: yes rerunning it gave the same error
<guix-vits>davidl: :(
***link2xt[nm] is now known as link2xt
<guix-vits>nckx: do you know anything about the davidl 's today first post issue?
<guix-vits>It's RedHat box
*nckx is summoned, reads backlog.
<nckx>davidl: The ‘statting’ error must be from one of the two other occurrences in that file. The first one only deals with lock files. Hence, it should be safe to just delete the offending file manually and try again.
<nckx>(In fact it should always be safe to delete files in /gnu/store/.links but don't do that 😛)
<davidl>nckx: yeah, I think it's the error on line 612 (edited above). Ok, I'll try and delete it, thanks. I would not have figured out whether it was safe myself.
***jluttine_ is now known as jluttine
<leoprikler>how do I increase the COLUMNS inside guix build?
<nckx>leoprikler: You could setenv it in a custom guix-daemon service. I just start a temporary guix-daemon from a shell with COLUMNS=9999999.
<nckx>Sorry to disappoint.
<leoprikler>and how do you talk with that daemon instead of the one that is always running?
<leoprikler>or do you disable it through shepherd and then re-enable it?
<nckx>leoprikler: You need to stop the main deamon. There can be only one.
<leoprikler>... and it still didn't work
<nckx>Foo.
<guix-vits>leoprikler, nckx: alias gbuild='COLUMNS=...' will not work?
<nckx>guix-vits: No, alias doesn't set variables like that.
<leoprikler>Okay, what am I doing wrong: `sudo sh -c 'COLUMNS=99999 guix-daemon --build-users-group=guixbuild' &`
<nckx>guix-vits: I must have typoed, now it does.
<nckx>leoprikler: Try exporting it.
<leoprikler>inside my current shell or inside sh -c?
<nckx>It shouldn't matter.
<nckx>If that doesn't work this will bug me all day because I've made it work before.
<leoprikler>What bugs me is that copy-build-system raises an error on copy-file
<leoprikler>Like wtf
<nckx>I'd ask ‘why’ but that would be trolling at this point.
<nckx>Nothing useful in the truncated version?
*guix-vits runned out of versions, sir! -- It's trunny, really very trunny!
<leoprikler>nope, just the path of the file that fails
<leoprikler>However, two things strike me as odd
<leoprikler>first, there's the sequence "//./" in the destination path, which kinda irks me
<nckx>I feel ya.
<leoprikler>and second, I have a path with subdirectories
<leoprikler>basically, I'm trying to port delft-icon-theme and get the following:
<leoprikler>`./Delft-Amber/places/16/gnome-fs-bookmark.svg' -> `/gnu/store/n37qwpab4i1mi0k20f371yzdfhnm0jzs-delft-icon-theme-1.10/share/icons//./Delft-Amber/places/16/gnome-fs-bookmark.svg'
<nckx>And c-b-s take's care of mkdir'ing all this junk?
<nckx>I haven't actually looked it it yet, being hours old.
<nckx>*takes, at, wtf.
<leoprikler>Okay, I found the bug
<leoprikler>It tries to install a broken symlink
<zzappie>Hi Guix!
<nckx>Hullo.
<zzappie>nckx: I am actually decided that I need to learn touchtyping
<guix-vits>zzappie: i'm recommend Ktouch app.
<guix-vits>but first is best to reduce difficulty, to practice with all keys.
<nckx>zzappie: Good! It's a valuable skill.
<zzappie>guix-vits: I find gtypist's texts very well made
<erkin>I learnt touchtyping on Dvorak with gtypist.
<erkin>Although I used the single space patch.
<erkin>Nobody uses double space between sentences any more.
*nckx also recommends Dvorak & gtypist.
<nckx>erkin: Guix does.
<zzappie>erkin: I didn't know that people did even
<erkin>Really?
<nckx>I do.
<erkin>zzappie: It used to be much more common.
<nckx>It's certainly more common in GNU/Emacs circles than outside of it.
<zzappie>nckx: do you use emacs?
*guix-vits found double space fascinating and easier to read. Also emacs deletes multiple sentences as one, if no double space after '.'
<nckx>Yes. But not as intensively as many here. Just mail & text editing, not the world.
<spk121>Double space between sentences was how I was taught in school, back in the olden times.
<nckx>guix-vits: Same.
<erkin>I technically use double space in documents I typeset with groff.
<nckx>Now it irks me when format=flowed mails have random single spaces in them.
<erkin>:-P
<nckx>‘Don't claim that I typed this sloppy trash, mail.’
<zzappie>nckx: then how do you handle the emacs keybindings on dvorak? Is it okay?
<erkin>Emacs keybindings are perfectly layout agnostic, unlike vi.
<erkin>vi is explicitly geared towards QWERTY.
<nckx>It's… OK. I don't rebind and my fingers haven't fallen off. Caps lock as control is far more important.
<nckx>I do use the ‘programmer Dvorak’ variant (stupid name) which places symbols in unshifted positions. But the letters are the same.
<erkin>I use the standard Dvorak with lots of personalisation piled upon it.
<ngz>I think Caps Lock as Ctrl is terrible, at least on a standard keyboard. It's not even symmetric.
<erkin>French, German, Turkish, Greek, Esperanto etc characters.
<erkin>I just press LCtrl with the corner of my palm, near the base of my left pinky.
<zzappie>ngz: I can't live without -option ctrl:swapcaps
<nckx>erkin: Ċỏṁpøŝẻ kê¥ 😉
<nckx>ngz: Luckily it's not mandatory.
<nckx>Yet.
<erkin>nckx: Hehe, I assigned compose to caps lock, but it gets tiring when you need to use it for common letters.
<nckx>erkin: Although I admit it's probably not ideal if it's your native language.
<nckx>Yeah.
<erkin>% setxkbmap -layout meow -option lv3:ralt_switch,compose:caps,terminate:ctrl_alt_bksp,compose:caps,terminate:ctrl_alt_bksp
<ngz>Of course. I'm just surprised that someone could find it even adequate. I'd argue there are better positions for Ctrl.
<erkin>oops, I mispasted
<erkin>The latter two are redundant.
<nckx>I switched to Dvorak exactly because it's not super-personalised. Used to be one of those annoying Colemak users. ‘Sorry, I can't use this computer.’ But everything supports the DSK.
<erkin>Oh that's definitely why I picked Dvorak over Colemak et al.
<zzappie>ngz: It actually was like that on old terminals. Sun was producing "Unix keyboards" that had control an capslock "swaped"
<erkin>You can find Dvorak everywhere.
<erkin>The personalisation isn't mandatory if I'm only typing in English, not coding APL, don't need mathematical symbols etc.
<zzappie>I was thinking recently about making my trakpad keys (under the space bar) send C-x and C-c
<ngz>zzappie: Interesting. I guess ergonomics wasn't a thing at this time.
<erkin>Haha
*dutchie wonders if it is worth switching to dvorak from colemak
<erkin>I mean, you could still say that. Our keyboard layouts are still staggered like typewriters.
<nckx>dutchie: You'll never know for sure. ;-) I'm very glad I did, and now I wish I'd never taken that detour. And the programmer variant is worth the mild unstandardness.
<dutchie>having to use the uk variant might reverse that
<dutchie>since I need to type £ somewhat frequently
<dutchie>still, I guess uk-dvorak is still more common than uk-colemak
<nckx>It's like emacs: I use a very stock configuration on purpose. There are people whose emacs is unusable by anyone but them. Both are fine, but both are trade-offs.
<dutchie>i like the dead-key accents in colemak too
<dutchie>and i am now very used to caps-as-backspac
<erkin>You can modify the layout with AltGr features.
<dutchie>e
<erkin>I've got λ on AltGr-l but you could put £ or ł there.
<erkin>(I use λ relatively often in Scheme.)
<nckx>It was never clear to me what exactly was standardised in Colemak and which was just bleeding through from the underlying (US?) intl support.
<dutchie>should probably not try switching keyboard layout until after i have finished writing my thesis
<zzappie>nckx: you also never need to open text editors when you have htings like tramp
<nckx>Mind you, I don't know that about DVP either but it feels more consistent across OSes.
<erkin>dutchie: Yeah, I recommend against that.
<nckx>dutchie: Ehm, yeah, there's a sharp edge between ‘learning pressure’ and ‘aversion therapy for life’. Don't associate your shiny new layout with suffering.
<erkin>Hahah
<erkin>It will get frustrating.
<nckx>dutchie: Backspace is the most frustrating key. I mean, I wouldn't trade Caps→control for it, but it is ridiculously unreachable.
<nckx>‘Teaches you not to make typoes’ but that only goes so far.
<zzappie>erkin: i think geiser has geiser-insert-lambda set to C-c \ by default
<nckx>Three cheers for C-h.
<erkin>One keybinding too many. :-)
<erkin>Err
<erkin>keystroke*
<nckx>If that's why people switch to emacs-for-everything, I completely understand.
<nckx>zzappie: WDYM by ‘you also never need to open text editors when you have htings like tramp’?
<nckx>I consider emacs a text editor, controversial as that may be 😉
<zzappie>nckx: I meant you dont ever need open unconfigured emacs or vi when you editing remotey and find your self in an unconfortoble zone when you use tramp.
<nckx>Understood.
<zzappie>Anyone knows can I use patch-sheebangs thing with guix cli?
<nckx>zzappie: ‘patch-shebangs’ itself is a phase which doesn't (really) lend itself to that, but you can use (guix build utils)'s patch-shebang procedure directly.
<alextee[m]>it's been a while since i noticed, but the source links are broken on the package browser website
<alextee[m]>for example: https://guix.gnu.org/packages/mpv-0.32.0/
<alextee[m]>clicking on package source takes you here https://git.savannah.gnu.org/cgit/guix.git/tree//gnu/store/zzbxvr0f2bkh65702jh91q4sv7gyy30l-guix-module-union/share/guile/site/3.0/gnu/packages/video.scm#n1359
<alextee[m]>which says file not found
<zzappie>nckx: Thanks ill try that
<nckx>alextee[m]: Thaw was reported but I don't remember if it was just here or as a real bug report.
<nckx>Having the latter would be nice.
<nckx>*that
<alextee[m]>idk how to search for bug reports should i send an email anyway?
<nckx>alextee[m]: Preferably not, let me search first. Though I don't have a clue what to search for.
<mfg>Hi, i updated Guix SD yesterday and it now doesn't find my WLAN Chip with the newest kernel ...
<mfg>Did something change?
<nckx>mfg: Sounds like this <https://debbugs.gnu.org/cgi/bugreport.cgi?bug=39671> which was fixed yesterday. Try pulling and reconfiguring again and let us know if the problem persists.
<Guest329>Hi, I'm new to guix and I asked my first question to the help-mailinglist of guix. But so far my mail never arrived in the actual list....
<mfg>nckx: nice, will do :)
<nckx>Guest329: ‘A first question about parallel environments’?
<nckx>I've approved that message. Nothing else is stuck in moderation (first-time posters need to be approved by a human, to wit: me).
<nckx>alextee[m]: http://logs.guix.gnu.org/guix/2020-02-04.log#175447
<Guest329>nckx: yes
<nckx>atw: Did you ever report that bug? ☝
<alextee[m]><civodul>atw: could you report it to bug-guix so we keep track of it?
<alextee[m]><atw>sure!
<nckx>alextee[m]: Hence my question.
<nckx>There's no link and my random keywords don't find anything.
***Guest329 is now known as benutto
<nckx>benutto: It's been delivered.
<benutto>thanks, and how long do you think until its visible?
<nckx>Visible? Oh, it should be in the archives alread, just waiting for the cache to clear.
<roptat>It arrived in my mailbox
<nckx>Mine too.
<roptat>Although your question is tricky
<nckx>Your subsequent mails will be delivered immediately.
<mfg>nckx: yep, it's fixed :)
<nckx>mfg: Great!
<nckx>Thank all the fine people in that bug report, it took some effort to coördinate & track down.
*nckx just got the message ‘How big can a Ubuntu PowerPoint file be if it needs to work in Africa?’ Now that's a tricky question…
<nckx>o/ Bye Guix.
<mfg>yes i skimmed through that thread :)
<mfg>Nice question though o.O
<benutto>@nckx: I can see it now, thanks very much
<zzappie>nckx: (patch-source-shebangs) worked. I think it could be useful. It is not exported in the module
<zzappie>anyone having issues upgraiding? Downlods from ci geting stucked
*zzappie got to go
<lfam>Sorry zzappie
<joshuaBPMan>hey guix!
<roptat>hey joshuaBPMan
<roptat>Sorry I didn't update the vpn service
<joshuaBPMan>roptat: that's really ok. I appreciate you giving me time to try to update it myself.
<joshuaBPMan>I think I'll update the bug report. We talked about it on irc, but it would be nice to have some further documentation.
<roptat>Sure, I'll probably have time this week-end
***emacsoma1 is now known as emacsomancer
<joshuaBPMan>roptat: I'm still racing you. :)
<joshuaBPMan>Also, it looks like some of the options in my worknig are not really helpful. one is only useful for windows for example.
<joshuaBPMan>so I'll write down the information about that in my updated bug report.
<roptat>Thank you
<roptat>I think tls-client is the difficult one, the rest should be easy to add, but you have to know whether they're for the client only or for both the client anl server
<lispmacs[work]>Is there an easy way to see what packages or store items are depending on a certain store item?
***PifyZ_ is now known as PifyZ
***PifyZ is now known as PifyZ_
<thomassgn>lispmacs[work]: have a look at guic's help, referers, references, requisites and derivers. I don't remember exactly how they all work.
<lfam>That's `guix gc`'s help
<thomassgn>haha, yes.
<civodul>is it me or supertuxkart 1.1 fails to start?
<lfam>civodul: It seems to work for me on Debian
<nckx>civodul: Segfaults fine here.
<bavier1>civodul: works for me
<nckx>‘Too many bones for skinning’.
<nckx>A haunting folk melody.
<lispmacs[work]>thomassgn: looks like guix gc has a --references option, I'll try that
<nckx> http://dpaste.com/30RB9F0
<lispmacs[work]>is shared-mime-info something regenerated every no package install? I always get the warning that it is conflicting in profile with xdg-mime-database
<lispmacs[work]>*every new
<lispmacs[work]>no, shared-mime-info has a package, but xdg-mime-database doesn't
<lispmacs[work]>same question, flipped around
<nckx>Is shared-mime-info supposed to be installed into profiles? (I don't know.)
<NieDzejkob>(I'd guess "no")
<nckx>It's not installed here.
<nckx>But then my .patch files open in LibreOffice when I accidentally xdg them, so don't mind me.
<thomassgn>anyone know of a VPS where installing guix is more comfortable than jumping through the unlimited hoops linode presents now a days? NixOS was really simple to install on linode a few years ago, but now I'm unable to make it work...
<mehlon>uhh I heard of that thing vultr?
<mehlon>dont know if its good but a lot of people seem to be using it
<mehlon>vultr.com
<ngz>Where would be the appropriate place to send an almost finished package (i.e., it builds, but raises a runtime error) that wouldn't mind to get some love? I'm a bit stuck on it.
<NieDzejkob>ngz: send a patch to guix-patches@ and mark it as [WIP PATCH] or [WIP]
<thomassgn>mehlon: thanks, looking at it now...
<ngz>NieDzejkob: done. I hope some fresh eyes will help finish it.
<nckx>thomassgn: What's changed @ Linode?
<NieDzejkob>nckx: Remind me to take another look at unoconv when substitutes for libreoffice become available again :D
<nckx>Muhaha.
<nckx>NieDzejkob: If you want (and your testing is appreciated), you can use --substitute-urls=https://guix.tobias.gr. Random stranger-binary-danger disclaimer applies.
<nckx>I could be anyone.
<nckx>But I'm not.
<mehlon>a dog?!
<mehlon>not that I would know.
<jackhill>nckx: (I'm reading through recent commits) I think emacs-rotate-text is another candidate for https homepage conversion :)
<nckx>jackhill: Then keep reading. 😉
<thomassgn>nckx: I don't know when, but they've got a new admin interface. I remember setting up nixos as a custom OS some years ago was more or less straightforward, now there's too much hidden. I gave up when I got errors about a partition 3031-309363438-... not being found.
<thomassgn>how can I stop the graphical install? :)
<thomassgn>choosing abort during the graphical install restarts the graphical install...
<nckx>thomassgn: As in ‘continue manually from a command line?’ Just ignore the graphical installer and switch to VT with C-M-F3.
<nckx>There's documentation on C-M-F2.
<nckx>thomassgn: Sounds like they changed their set-up to use network block devices or something like that. I don't know if we support that yet (and the installer on the Web site is a year old).
<thomassgn>ah, nckx I don't have access to modifier keys in the webterminal interface provided by vultr ... >_<
<nckx>Ai. I can't think of a trick right now.
<nckx>‘Aborting the installer‘ wouldn't just drop you to a terminal, though. It's not started from a shell. It's the only thing running on VT 1.
<nckx>It could support spawning a shell for uses like this, though.
<nckx>thomassgn: Does Vultr offer VNC? Is the ‘Web terminal’ a full-featured remote display viewer (like a HTML5 VNC client) or just a serial console?
<thomassgn>nckx: yea, I can't even see the full picture of the installer. Oh, idk. the url includes 'novnc'... ?
<thomassgn>I might just roll up a disk-image from a sort of minimal config and try from that
<nckx>thomassgn: Yeah, I'd just found a forum post calling it NoVNC, and that triggered a memory of a bug we had where the installer didn't like NoVNC :-/
<nckx> https://debbugs.gnu.org/cgi/bugreport.cgi?bug=36069
<thomassgn>haha, perfect.
<nckx>I'm not going to read the whole thing but there are some fixes in there you can try!
<nckx>Thanks for mentioning novnc though, that would've been extremely relevant information that most people wouldn't have provided.
<jackhill>nckx: ☺
<Ashjkaell>Hi everyone!, stupid question: To add small system services, the canonical solution is to define a service, a channel, and pull right?
<NieDzejkob>Ashjkaell: You can put them in your configuration. It's just Scheme code, after all.
<Ashjkaell>(There is shepherd's init.scm which says "add services here", but It doesn't seem to get read at startup, which is rather logical)
<Ashjkaell>NieDzejkob: Haha, that's the most reasonable solution I've ever head ;-)
<Ashjkaell>NieDzejkob: Many thanks, you saved me a lot of time; I was pulling from my own channel like an idiot for way longer than I'd like to admit...
<civodul>oops, ENOSPC on berlin
<NieDzejkob>unsolicited suggestion: set up an alert for (< free-space (* total-space reasonable-fraction))
<apteryx>civodul: ah, this prolly explains the ugly backtrace I'm seing on substitutes resolution
<civodul>NieDzejkob: we did think about it but, as crazy as it may seem, the blocker is getting an email setup there :-/
<civodul>if you or anyone has experience with it, it'd be great to have an opensmtpd config or something added to https://git.savannah.gnu.org/cgit/guix/maintenance.git/tree/hydra/berlin.scm
<civodul>also, this is largely due to a ridiculous mcron bug: https://issues.guix.gnu.org/issue/37237
<atw>idea: reproducible docker containers initiative called "save the whales"
<kirisime>I ran the guided guix system installer and while the last part was downloading substitutes the network stalled. And GUIX still doesn't recover from it, I killed the process and now the guided installer crashes after selecting a keyboard layout.
<atw>you say "still", I take in you've run into this before?
<kirisime>Can I restart the last part by running `guix system init /tmp /tmp/guix-inst/<path>-config.scm' or is there more magic to it?
<kirisime>atw: Practically every time I need to download lots of anything. My internet isn't high quality.
<atw>I don't see the harm in re-running the guix system init. I mean, the worst it can do is fail, right?
<kirisime>I suppose. It's a VM anyways.
<atw>if it failed while downloading substitutes, then I would expect the store to be partially populated with the stuff you need, and re-running guix system init will basically pick up where it left oyy
<kirisime>Well it doesn't seem like that's the right file either.
<atw>off
<NieDzejkob>check in /mnt/etc/config.scm?
<kirisime>The target partition is mounted at /tmp
<NieDzejkob>oh, ok. weird. /tmp/etc/config.scm, then
<kirisime>No dice, it only has guix-inst/, last-installer-error and .overlayfs-workdir/ on it. I guess the config is just gone by now.
<atw>ouch, that sucks. Is manual installation an option for you?
<kirisime>I rebooted and the guided installer crash is gone. If the network stalls again I'll write down the command line before killing the process.