IRC channel logs

<goo>Hello. I would like to package a soft from a debian 8 to debian 13 install in guix. Can I package a static build to support it please?

<ieure>goo, Your question is unclear. What do you want to package?

<goo>Un want to package torch7

<goo> https://codeberg.org/libertIA/nn and https://github.com/torch/torch7

<goo>ieure

<ieure>goo, What do you mean by a static build?

<ieure>goo, You want to package binaries provided by upstream?

<goo>I mean a build done in the OS debian 8 and portable to debian 13

<goo>yes

<ieure>Where does Guix fit in here?

<goo>because I want to package for debian 13

<goo>I am a noob

<apteryx>goo: with guix you dot need a static build; you can use 'guix pack', and it'll build the package the usual way (linking ot guix-provided shared libraries), and it'll work on any systems having a sufficient kernel

<goo>oh my god great!

<apteryx>the dependencies of a package packed/installed with guix are completed provided

<apteryx>you could think of 'guix pack something' as a complete distribution of something

<goo>nice

<goo>then if I pack with good software, the package will be ok

<apteryx>it's like running it in a chroot in some ways. it's a bit heavy on disk if it's just for one application, but it works everywhere.

<goo>but if I do not want to worry about having good dependencies in newest os I can set static buuild?

<apteryx>all the dependencies will be taken from guix, so it's pretty isolated from your operating system in terms of libraries and such

<goo>sounds already better than docker lol

<goo>nice

<goo>perfect

<goo>I am going to use a debian 8 vm. I will use guix in 8 and then deploy it to debian 13

<ieure>goo, It's not a static build, but Guix will bundle the binary and all the libraries it depends on.

<ieure>Why are you using Debian 8? That's very, very old.

<goo>perfgect

<goo>ieure: I am confused

<apteryx>I'm biased but I like the transparency offered by Guix better. With Docker there tends to me some magical part used, mostly in the form of FROM dependencies. In most cases with Guix FROM is just a commit has of the Guix commit hat was used to build/pack a software (there could be some extra channels involved, if you use that).

<goo>Does guix pack debian 8 to debian 13 install?

<goo>apteryx absolutely particulary the Dockerfile

<ieure>I don't believe Debian 8 has a package for Guix.

<goo>no pb I will install it

<apteryx>if you mean is the result of 'guix pack' compatible across debian 8 to 13, it should be, yes.

<goo>apteryx I agree with you the Dockerfile litteraly:

<goo>-replace nftable / iptable

<goo>- does build AND run commands

<ieure>goo, Why are you using such an old version of Debian? Debian 8 "jessie" is nearly 10 years old, and has been out of support for 5.

<goo>- custom cd

<goo>etc...

<goo>ieure comaptibility

<goo>please I just want to use it

<ieure>With what?

<goo>apteryx : nice! great ! I am very happy ! it is clear

<goo>I will do it

<apteryx>OK! if you have any issues, it'll be interesting to hear them!

<ieure>goo, Do you know Scheme?

<goo>no

<ieure>goo, Okay. Just curious.

<apteryx>is there a way to express first friday of the month in a Shepherd timer's calendar event?

<meaty>hm, it indeed appears that although there's an importer for it, nobody has ever packaged anything for php.

<apteryx>let's package gnu social ^^

<meaty>apteryx: lol we can hack it out on a giant pr once the help-wanted label is put in

<apteryx>:-)

<apteryx>we don't have such a label yet?

<meaty>apteryx: see https://codeberg.org/guix/guix/issues/67 and continued on guix-devel

<mange>apteryx: For the first Monday thing, you could try this wild hack? https://blog.healthchecks.io/2022/09/schedule-cron-job-the-funky-way/

<sneek>mange, you have 2 messages!

<sneek>mange, janneke says: that issue (https://codeberg.org/Codeberg/Community/issues/1562) seems to be exactly what i'm talking about. sending email with a noreply@ just seems ... fascist to me?

<sneek>mange, janneke says: s/fascist/obnoxious/ ?

<mange>Sorry, you said first Friday, not first Monday. Either way, same idea might work.

<janneke>rather: codebergs needs time, and they are already working on it, as the issue shows

<mange>Looking at the shepherd source, I don't think the hack will work. :( That said, looking at fit-day I think it might intersect #:days-of-month and #:days-of-week, so it might just work the "obvious" way. #:days-of-month (1 2 3 4 5 6 7) #:days-of-week '(friday)

<apteryx>mange: that'd be cool! (calendar-event #:days-of-month '(1) #:days-of-week '(sunday) #:hours '(0) #:minutes '(0)) -> #<<calendar-event> seconds: (0) minutes: (0) hours: (0) days-of-month: (1) months: (1 2 3 4 5 6 7 8 9 10 11 12) days-of-week: (0)>

<apteryx>I guess it works!

<apteryx>thank you

<apteryx>hm, I should drop #:days-of-month, that's extraneous

<mange>I think #:days-of-month '(1) will only run if sunday is the first day of the month.

<mange>There's a comment in the Shepherd which makes me think it intersects them: "DAYS* is empty because the DAYS/WEEKDAYS intersection is empty--e.g., ;; no Friday 13th this month. Try the next month."

<apteryx>ah, that's my issue. I can have it run every Sunday, but I don't see how to have it run on the *first* Sunday

<mange>So you say #:days-of-month '(1 2 3 4 5 6 7) to restrict it to the first week, then #:days-of-week '(sunday) to restrict it to Sunday.

<apteryx>that's why you used 1 2 3 4 5 6 7, I see, to ensure the Sunday matched is somewhere in the first week

<mange>The combination of the two is "the first Sunday of the Month".

<mange>... Not sure why I capitalised "month" there.

<apteryx>makes sense

<apteryx>thanks!

<mange>Once you configure a service with it "herd status $service" should tell you if it's worked by showing you the next times the service will run.

<apteryx>civodul: moin!

<apteryx>mange: it works!

<mange>Huzzah!

<civodul>hi there!

<civodul>wazup?

<apteryx>hi! I was looking at the file system performance on Berlin; it was pretty slow with the ongoing btrfs balance.

<apteryx>I've prepped this: https://codeberg.org/guix/maintenance/pulls/1/files, based on feedback from the fine folks in #btrfs

<apteryx>I don't think it'll do miracles, but if the file system is seriously fragmented, the balance with -dusage=70 shoud help.

<apteryx>fun, the forgejo repo now has a 'manifest.scm', and the doc some instructions for building/testing from source using Guix! https://codeberg.org/forgejo/forgejo/pulls/8038/files

<civodul>apteryx: i think btrfs-balance ran for almost 24h, no?

<civodul>ACTION never had to worry about file system fragmentation in the past 20+ years

<apteryx>civodul: I think the file system and/or the array performance in general is bad at the moment. I'm oping that's all due to fragmentation, but there is added IO from rsync since recently

<apteryx>you can see it in 'sudo atop -d 1'

<apteryx>hoping*

<apteryx>I sent a wall of text email to guix-sysadmin ;-)

<apteryx>fragmentation is inherent to CoW file systems

<apteryx>rsync is currently using 100% of IOs

<noe>oof

<apteryx>civodul: by the way, having rsync fetch the nars and narinfo regularly means their atime will always be refreshed, and nothing will ever get cleared according to the retention policy, right? that's also bad.

<civodul>apteryx: uh, good point

<civodul>that’s rsync of /var/cache/guix/publish or another one?

<civodul>the one that’s using 100% of IOs

<civodul>we should ionice these things

<apteryx>I'm not sure how to tell

<apteryx>the rsyncd.log is not that useful for long running connections

<apteryx>another thing; rsyncd.log doesn't appear rotated

<civodul>right, it’s not in ‘herd files log-rotation’

<civodul>but apparently it’s written directly by rsync, so shepherd doesn’t know about it

<civodul>either we let shepherd do the logging, or we add it as an “external file” for ‘log-rotation’

<civodul>(the former is nicer)

<apteryx>OK, sounds good

<apteryx>civodul: haven't found a nice way to check what rsync is doing via logs, but csysdig is proving useful here

<apteryx>try e.g. sudo csysdig --modern-bpf, filter with F4 for rsync, RET to 'dig' into it, then F2 and select Files view -> shows current nars being downloading

<jakef>can guix graph see inside computed-file inputs?

<sughosha>I want to configure my system with working plymouth. Has anyone done it with their system?

<g_bor>civodul: quick question, my key has the user id changed, as my e-mail changed, but it is otherwise the same that was recorded in the keyring in the repo. Is any action needed to avoid problems, or it should be good as is?

<mra>howdy guix o/

<Guest4374>Trying to build dmenu inside "$guix shell --development dmenu" fails with "/gnu/store/i7nv87h4a3vdnmqq9dim825ji6cr1r2w-profile/include/X11/Xft/Xft.h:40:10: fatal error: ft2build.h: No such file or directory"

<Guest4374>The error message is pretty descriptive but i am too inexpierenced to know what exactly i am looking for. The --development flag made things very easy so far for me. Would appreciate it a lot if anyone could point me directions

<reyman>Hi

<reyman>I search an exemple of lightdm install, my gdm is broken, impossible to login with latest pull

<reyman>The error in debug log message is not clear

<reyman>I have a gdm session exited with status 1

<reyman>Perhaps there a re some local file to delete ?

<reyman>To reset gdm state ?

<JodiJodington>reyman: not familiar with GDM, but maybe there's a GDM entry in `~/.cache` or in `~/.local/cache` or `~/.local/state` etc. maybe doing a `find ~/ -type d -iname "*gdm` would help you find some possible cached state

<JodiJodington>just to be safe maybe do `*gdm*` instead of `*gdm`

<reyman>Thanks JodiJodington, i will try that

<apteryx>reyman: GDM state is on a tmpfs, because we've had issues in the past where it'd crash with stale files or wrong permissions. So if you reboot in theory it should be good.

<apteryx>it's under /var/lib/gdm IIRC or similar

<jakef>oops, sorry if you just got a notification apteryx

<jakef>i'm trying to figure out how to put a label on my PR

<jakef>accidentally added you as a reviewer

<apteryx>don't worry, I'm alredy swimming in notifications ^^'

<apteryx>jakef: I left a LGTM :-)

<jakef>wew thanks :)

<mra>apteryx: sorry, i think that some of that pool of notifications may be my fault :P

<apteryx>I'm exaggerating; it's just a bit annoying to have to process both the emails and the web notifications separately.

<apteryx>(I've unwatched the guix/guix repo so that I only receive notifications for the teams I'm part of, that helps)

<mra>ah, i can imagine. my inbox is a largely neglected mess

<apteryx>I'll probably disable all email notifications and live in the browser (since we can't reply from the emails anyway)

<mra>that seems wise, yeah. i need to check out fj.el to see how it is

<apteryx>or maybe fj.el? haven't tried that yet

<apteryx>I suspect it'd need a lot of love to be able to do what the web ui can do

<mra>yeah, it seems interesting, but it isn't as good as the web ui, i imagine

<mra>apparently forgejo is next in line for support by magit, but my understanding is that adding new forges to magit is very low priority

<xelxebar>Anyone here using /etc/hosts to blacklist domains?

<xelxebar>I've been using https://github.com/StevenBlack/hosts

<xelxebar>with hosts-file.

<civodul>sneek: later tell g_bor this is fine: the OpenPGP user ID doesn’t matter for authentication purposes

<sneek>Will do.

<xelxebar>But need to figure out a good way to update to a hosts-service-type.

<civodul>apteryx: same for me! having both email and Forgejo notification is terrible

<civodul>*notifications

<civodul>(fj.el is sweet BTW)

<civodul>hmm i think the new pre-push hook isn’t doing anything

<apteryx>hehe, nice to know! I'll try it in time.

<apteryx>I'm still on the case of git-repo-go

<civodul>it’s not authenticating commits for me

<apteryx>ah?

<civodul>well, i think; i just pushed and i don’t see the usual auth and check-channel-news messages in the Magit buffer

<mra>oh, fj.el isn't on melpa. huh

<civodul>but it’s in Guix :-)

<apteryx>pre-push reads ok to me

<civodul>apteryx: yeah i just wonder about matching on $2

<civodul>but apparently $2 is supposed to be the URL, so it should work

<apteryx>the case "$2" in has been there for many years

<apteryx>you can put some false in the perform_checks() and see if it aborts

<apteryx>or maybe set -x to print which lines get executed

<mra>civodul: i don't actually use guix, unfortunately

<mra>i'd like to switch, but i use storage in a raid configuration, which makes zfs support quite important to me

<mra>maybe fj.el is in nixpkgs...

<apteryx>you don't need to switch your OS to start installing emacs packages with guix :-)

<mra>i do have guix installed on nixos, actually, but it sort of just dies if i ever try to use it. haven't had the energy to debug it

<apteryx>oh! interesting. maybe nix has a killswitch ^^'

<apteryx>I seem to recally some NIX variable that guix-daemon used that could conflict, not sure if this is still an issue

<apteryx>recall*

<mra>i'm not sure. i might look into it at some point if i find the energy

<mra>i'd rather just switch to using guix as an operating system. i just gotta do the work to make that possible for me :)

<xelxebar>mra: Did you just manually install, or is the nixpkgs derivation for Guix broken?

<reyman>@apteryx i reboot multiple time without success

<xelxebar> https://github.com/NixOS/nixpkgs/blob/master/pkgs/by-name/gu/guix/package.nix

<reyman>The problem persist, so is i jndersrand that could be another thing than cache

<apteryx>are you using the xorg session with gnome-shell?

<apteryx>this stopped working recently, there's an issue reported

<reyman>I also try to create another user

<reyman>Oh

<apteryx>it works fine with wayland

<apteryx>it shows just a grey background and a cursor for the mouse IIRC. it's reproducible in QEMU

<reyman>I set wayland true in gdm configuration

<apteryx>hm. then I don't know :-/

<apteryx>nothing of use under /var/log/gdm ?

<reyman>I have warning

<apteryx>is there a way to configure the pull branches to be fetched but *without* crowding the branches view in magit?

<apteryx>I have fetch = +refs/pull/*/head:refs/pullreqs/* in my .git/config but this has that problem

<mra>xelxebar: the nixpkgs derivation seems to be busted, although I may be using it incorrectly. I just added services.guix.enable = true to my flake

<reyman>I have multiple errors, libmutter failed to make thread kms thread high priority

<reyman>Apteryx is it easy to switch to lightdm ? I cannot work actually if i cannot login...

<xelxebar>mra: That's unfortunate. It'd be worth pinging one or more of the maintainers listed in the package metadata.

<ieure>mra, Can't speak for others, but I dislike putting stuff on MELPA.

<apteryx>reyman: shouldn't be too difficult. you need to remove gdm from %desktop-services

<ieure>They always want to review my code. I don't care to have them review it.

<ruther>ieure: do they decline it if they don't like it or do they just comment?

<mra>ieure: ah, I didn't know that! I've never actually submitted anything to melpa

<ruther>ieure: beccause if they just comment, it makes sense to me - they should go over the code to prevent malicious code additions, so why not send their reviews when they already went through it

<ieure>ruther, That's a good question, I haven't resisted enough to find out, I just took my ball and went home.

<mra>xelxebar: I might do so at some point. I'm currently putting the little energy that I have to spare into guix

<mra>what can I say, I like parentheses

<meaty>Should we use `` '' or " " in descriptions? I see both in the codebase

<meaty>or the devilish option “ ” >:)

<meaty>...which actually does occur several times also :o

<apteryx>meaty: `` '', or @emp{}

<apteryx>@emph{}, if appropriate

<gargaml>Hi! Still exploring the basics of Guix :) Still having issues with

<gargaml> the sound configuration. I'm on the path to try pipewire but first

<gargaml> I'd like to understand something about pulseaudio. When I type sudo

<gargaml> herd status, pulseaudio doesn't appear to be a service managed by

<gargaml> herd. But when I look at running processes (top, proced, ...) I can

<gargaml> actually see that pulseaudio is running. In my config I rely on

<gargaml> %desktop-services in which we can find the pulseaudio service, but is it only this that make pulseaudio

<gargaml> running?

<gargaml>(sorry for the formatting issue bad buffer management...)

<ruther>gargaml: pulseaudio server is started by the first client that accesses pulseaudio

<ruther>gargaml:so there is no need to start pulseaudio by shepherd

<ruther>gargaml:pulseaudio-service-type is for configuring pulseaudio, it is not to start it

<gargaml>ruther: thanks!

<gargaml>it seems that pulseaudio on my system doesn't manage my sound hardware well, on other distros using pipewire, the problem seems solved. I read that installing pipewire requires to rely on guix home. Are there other ways to switch to pipewire without going through guix home?

<ruther>gargaml: you can definitely run it manually, it' not so hard. How did you find out that pulseaudio is the one having problems with your hardware?

<ruther>gargaml: it is very unusual pulseaudio would have problems with hardware as kernel is the one managing hardware... pulseaudio just uses what the drivers give it to userland

<gargaml>I compared firmware versions, drivers and the only difference I can spot seems related to pulseaudio

<ieure>Not sure why that would make a difference, shouldn't they both be using ALSA under the hood?

<gargaml>I thought pipewire would have an alsa replacement but you're right this is not the case

<gargaml>so basically it's either a firmware or a driver issue and nothing else. I'm gonna keep digging, thanks

<ieure>Yeah, ALSA works fine, no reason to replace the plumbing here.

<dlowe>you can use aplay to test out alsa directly

<meaty>is there a way to delete PRs? I messed up an agit push and accidentally sent the whole rust-team branch to itself alongside my changes :,(

<ruther>meaty: no, because that would possibly make inconsistent state. So instead you can only close PRS

<ieure>meaty, You can close it and open a new one, but you can also force-push a PR branch to get it back into whatever state you want.

<meaty>I understand

<graywolf>Hi :) Any committer that could take a look at 78482 ? I think it would be nice addition.

<gargaml>dlowe: I'm trying to do this and right now things seem to push towards snd_hda_intel arguments to prevent auto detection.

<graywolf>Is the bot syncing to savannah down?

<ekaitz>graywolf: re 78428: does it have a possible non-intended consequence?

<bavier>graywolf: also re 78482, what are the intended consequences? (I'm not super familiar with LANG)

<ekaitz>i mean, i could just push it... it feels harmless but who knows

<graywolf>Guile in shellcontainer will be running with UTF-8 support. If

<ekaitz>oh great! dyslexia strikes again!

<graywolf>If user already has LANG set, there will be no change, the user requested one will be used

<graywolf>But in absence, we are currently running effectively in C, so C.UTF-8 should be fairly minimal change

<graywolf>For anything that is not Guile

<ekaitz>it feels harmless and looking good

<ekaitz>bavier: do you plan to push? or should I?

<bavier>ekaitz: go ahead!

<ekaitz>hehe so if everything fails it will be my fault

<ekaitz>I'm ready for the responsibility! :)

<bavier>feel free to add a Reviewed-by for me.

<graywolf>Thank you both :)

<ekaitz>graywolf: 5c5c5a6e586923a2a6d9e2991cdc281dd9d0a7ef

<ekaitz>your email user is a tilde, you are great

<graywolf>Thanks for the push; re email, I have many problems with it. But it *is* a valid address, so I just ask people to fix their validations (e.g. now you can use ~ email on savannah :) )

<ekaitz>i have problems with mine just because the tld is not recognized so i'm not surprised

<graywolf>It is surprising how many things are technically legal on the left side of @ :)

<graywolf>Surprising amount of systems are storing emails case-insensitive, that is one fight I gave up on

<ekaitz>and how many emails are validated with a very simple regex

<ekaitz>that's also a lot of fun

<graywolf>Well I am always temped to use https://pdw.ex-parrot.com/Mail-RFC822-Address.html for the validations

<peanuts>"RFC 822 - STANDARD FOR THE FORMAT OF ARPA INTERNET TEXT MESSAGES" https://datatracker.ietf.org/doc/html/rfc822

<graywolf>Could be fun explaining that in a code review

<ekaitz>haha classic! i love it

<dlowe>if [^()<>@,;:\\".\[\] \000-\031]+ were collapsed down the whole thing would be a lot shorter

<nico_>I'm trying to install guix system with legacy boot. In partitioning, do I need a boot partition, or can I remove the fat32 EFI partition and leave the 2TB ext4 (or btrfs) cryptroot partition?

<nico_>Or maybe I can leave the default partition schema for one partition disk encryption and not install bootloader on efi partition.

<bavier>anyone else unable to build guix with commit ad8cb7af8fc572bb3d11cd0344bd2bed3b9d653f ?

<bavier>huh, this commit only exists on savannah...

<ieure>Can't even get to Savannah to look at the commit.

<bavier>huh, oops, no, exists on codeberg: https://codeberg.org/guix/guix/commit/ad8cb7af8fc572bb3d11cd0344bd2bed3b9d653f

<bavier>oh, was this commit force-reverted? it doesn't exist on master anymore.

<ieure>Hmm, must have been? Anything on guix-devel? I think civodul is the only one with the power to do that.

<bavier>I couldn't find anything, no mention of issues on the initial patch thread.

<ruther>how do force pushes to master work on savannah anyway? I thought branches have to be removed so that you can push different HEAD, is it the same with master?

<ruther>why was it force reverted anyway? Doesn't seem like the signature would be wrong... so what else would lead someone to force push master?

<bavier>idk why a force revert instead of a more selective revert. The commit leads to build failures: https://issues.guix.gnu.org/issue/78668#11

<graywolf>ruther: Is it possible dannym pushed it to savannah instead of the codeberg?

<graywolf>Last shared commit on both masters is 325058efde2509536604c1e7cd705d48aeb5ce74

<graywolf>That would possibly also explained why changes from codeberg are no longer mirrored to savannah (since the histories diverged)

<ruther>graywolf: yeah, it is possible as Danny is still member of the group - member of the group

<ruther>- https://savannah.gnu.org/project/memberlist.php?group=guix

<ruther>graywolf: even force push to codeberg would explain it, it can be seen with mesa-updates - the bot doesn't seem capable of resolving that

<graywolf>Yeah, but for codeberg it would need to be done by Ludo, right?

<graywolf>I would expect an email or something, so push-by-mistake to savannah seems more likely

<ruther>graywolf: I am not sure what people have admin rights in the codeberg organization

<graywolf>hm, right

<graywolf>Luckily the savannah version does not pull. That should prevent guix from pull, so after it is force-pushed to resync with codeberg, users should be able to pull without any warning. Maybe?

<graywolf>does not build*

<graywolf>Actually, I am not sure. Does anyone know what is the behavior here?

<ruther>graywolf: guix doesn't save anything until it is built... so it won't even know what commit it tried to build yesterday etc... so why would there be any warning?

<graywolf>I was not sure what information is stored during the checkout. So this is a good news.

<ruther>graywolf: the checkout is basically stateless. It is just fetched and resetted to the correct commit

<graywolf>Now I am curious what happens when I do two pulls of different commits at the same time.

<graywolf>ACTION is off to experiment

<RamonaZero>hi! :0 is there any documentation on the disk space requirements for the guix system? I have a 64GB SSD laptop (that I can't upgrade...) but i plan to just use text-based work nothing with the desktop environment

<ekaitz>RamonaZero: that should work if you carefully choose the packages you install. Guix allows for a lot of costumization. But also is a little bit wasteful with space because of the functional package management.

<ekaitz>RamonaZero: I have currently 69G in use in my / partition (doesn't include /home) and I have 3 web browsers installed, and many packages that I could clean with `guix gc`

<RamonaZero>that's very nice! :0

<ekaitz>if you just want a minimal setup you'll probably be able to go pretty low in disk usage, but remember that if you want to be able to roll-back you'll need to have copies of things and so on.

<pastor>Hello, please, could someone check this issue? Its ready to merge, if there is a reason not to, I would appreciate some feedback:

<pastor> https://issues.guix.gnu.org/78580

<ekaitz>pastor: me pillas un poco mal, pero lo intento mirar estos días

<bavier>RamonaZero: I ran Guix on a small netbook with 40GB root partition for a long time. IIRC I was running the lightweight ratpoison DE and mostly text-based stuff. I was even building a lot of packages locally and never had disk-space problems.

<RamonaZero>oh wow, that's actually nice! :0

<RamonaZero>im trying to find a compatible wifi card xP im surrounded by intel cards Dx

<ieure>RamonaZero, ath9k is the most recent WiFi card that doesn't need blobs. It's getting pretty aged at this point.

<pastor>Gracias ekaitz! PD: me gustó leer tus blogs sobre tu trabajo con el bootstrap de RISC-V

<RamonaZero>yeah i heard ath9k is pretty versatile!

<dlowe>I had a lot of luck running guix on a pinebook pro

<ekaitz>pastor: that's great, I'm not reviewing your post because I'm specifically writing one about GNU Mes' performance

<ekaitz>your patch*

<ieure>RamonaZero, As long as your versatility doesn't include "supports WiFi technology newer than 2009."

<pastor>Don't worry, I understand. I'm actually very greatful that we have you in this comunity. I really want to see more Guix + RISC-V!

<ieure>'cos.... it doesn't

<RamonaZero>802.11b is where its at! :0

<RamonaZero>actually i do have a ethernet-USB adapter, so it might be working

<RamonaZero>its not a main laptop, just a projects one (a fun one), so constant wifi isn't really needed

<ekaitz>pastor: aaaaaww shut up! <3

<mra>i'm not actually sure if my laptop's wifi card is supported. I picked it out of a bin of parts for ¥500, so I don't know the model

<ieure>RamonaZero, Nice. I ran Guix System that way for a good while. Finally made the jump and reinstalled my daily driver.

<pastor>BTW, ekaitz, since you have been involved in NLNet grants, are the finish dates of grants stablished somewhere? For example, I would like to know the reoadmap for this one: https://nlnet.nl/project/Pre-Scheme

<ekaitz>pastor: that's a great question. You normally have 1 year to finish, but that doesn't mean you finish in time.

<ekaitz>they reserve the money for you for 1 year, let's say. If you don't finish in time you lose it. (they are actually more flexible than that, but the summary is as I said)

<pastor>And is the status public? Could I know if Pre-Scheme or the Guile daemon are going well or if they are droped?

<ekaitz>they are not public by default, it's on the hands of the grantees to decide if they share info

<pastor>I see, thanks for the insight. Seems like we will never now until they are completed :'(

<ekaitz>the guile daemon for guix is done by cbaines and he told me not that long ago that the project was a little bit stuck, as he had a lot of work to do with other things

<ekaitz>you could find the explanation for that in the irc archive

<ekaitz>about the pre-scheme... i think you could find the author in #guile-steel

<pastor>Oh, right, I will check it out. One day I hope we can have full time developer in Guix

<ekaitz>we have a few in practice

<pastor>Do we?

<ekaitz> https://guix.gnu.org/en/blog/2025/guix-user-and-contributor-survey-2024-the-results-part-3/

<goo>pastor : remunerated? If yes please mememe

<goo>also

<ekaitz>well, full time we have one

<goo>I am not a guix expert

<ekaitz>but I think he's quite spread in maaaany things

<pastor>goo: yes, I meant remunerated! Of course

<goo>oh nice!

<goo>I was going to use guix actively

<goo>I am interested in

<goo>could we do a visio please?

<pastor>Just to clarify, it was a with...

<pastor>wish*

<goo>Also I am not as string as gnutoo lol

<RamonaZero>i have never heard of emacs-exwm, but i love emacs, and this is amazing o.o

<goo>ah okl

<ieure>I would love to get paid to hack on Guix full-time. Seems unlikely, though, and even more unlikely to approach what I make now.

<ieure>RamonaZero, EXWM is awesome. I've been daily-driving it for a long time now. Seven years? Something like that.

<RamonaZero>daaaamn, ill give it a try

<RamonaZero>i was going to use emacs anyway o.o my favorite editor

<ieure>RamonaZero, It's kind of shitty. But shitty in ways I mostly like.

<ekaitz>ieure: I could get paid more doing other things, but i'm happier hacking around

<ieure>RamonaZero, I wrote a bunch of stuff to make it work more how I like. They're all in Guix. exwm-ss, exwm-firefox, exwm-mff, blight, discomfort, lemon are the main ones.

<ieure>RamonaZero, Source repos all at https://codeberg.org/emacs-weirdware

<RamonaZero>oh nice! :0 ill take a look!

<ieure>ekaitz, Yeah.... gotta pay the mortgage, though.

<ekaitz>ieure: sure... yeah. Our context oftentimes decides for us. It's sad. Probably the humanity would benefit more if you and many others could decide what to do.

<pastor>ieure: we need to make it ubiquitous so it brings money in

<ieure>pastor, That'd be awful nice.

<ekaitz>i think there are many business opportunities that can start from guix

<ieure>Though if I could choose to do anything, it'd probably be repairing vintage electronics. Have a backlog of that stuff you wouldn't believe.

<ekaitz>that could justify working on guix almost full time

<ieure>People just keep giving me antique TVs for free!

<ieure>And I keep taking them.

<pastor>ieure: ubiquitous and buggy so they have to pay for the fixes 😈

<ieure>pastor, Well, we're halfway there.

<pastor>:')

<ieure>Most recent TV... Not sure what year, 1950s most probably. https://retro.social/system/media_attachments/files/114/237/230/538/267/045/original/9aa80cacac151593.jpg

<ieure>And you can see two more in the background, a real nice Zenith, and an Admiral.

<pastor>ieure: that looks quite retro indeed

<ekaitz>ieure: I understand the mortage now... you gotta keep all those. :)

<ieure>ekaitz, I really have to cut down on the hoard, honestly.

<ieure>But it's so massive, it has its own gravity, people keep giving me stuff to add to it.

<ieure>This is the latest... https://retro.social/system/media_attachments/files/114/564/761/703/218/180/original/27f0159a774d9723.jpg

<ieure>That wasn't free.... just cheap.

<pastor>ieure: wtf. Are you living in a castle?

<ieure>pastor, I wish. I rent space for the overflow.

<pastor>That thing look masive xD

<ieure>It's regular pinball machine size.

<g_bor>Hello, what is the general stance on this: https://codeberg.org/guix/guix/issues/406

<sneek>Welcome back g_bor, you have 1 message!

<sneek>g_bor, civodul says: this is fine: the OpenPGP user ID doesn’t matter for authentication purposes

<pastor>There is what looks like a completely useless machine not fit for the modern era in the background o_0

<RamonaZero>o.o so i know this is for guix, but is there any repo or project on reverse engineering the intel wireless firmware? im pretty familiar with RE'ing and magic, i'd like to look into writing an open source firmware driver :0

<RamonaZero>(at least the possibly of it XD)

<ekaitz>g_bor: idk, probably requires a GCD? but I like.

<ieure>pastor, Machine in the background of the second photo is a Computer Space, the first mass-produced coin-op video game. https://en.wikipedia.org/wiki/Computer_Space

<g_bor>civodul: thanks

<reyman>After hours, i finally found that gdm works better if the gnome service is also installed...

<reyman>(service gnome-desktop-service-type)

<ieure>pastor, I got mine autographed by Nolan Bushnell last year. He designed it, gameplay, board layout, cabinet shape, everything. He also founded Atari and Chuck E. Cheese.

<pastor>ieure: I see you have an impresive collection

<reyman>So i found my problem apteryx , when i finally read in the long debug list that gdm search fail.desktop ... because gnome.desktop is not found.

<ieure>pastor, I've been at it a while.

<reyman>So stupid.

<pastor>BTW, ieure, I wanted to ask you. Will it take much to package Zen Browser in Guix, given your work on Librewolf?

<pastor> https://zen-browser.app/

<pastor>It's built on top of Firefox. I'm guessing the packaging will be similar to Librewolf

<ieure>pastor, Not sure. Didn't that just get discontinued?

<pastor>ieure: really?

<ieure>Or am I thinking of a different one?

<pastor>This one seems pretty active, last commit is from 4h ago

<ieure>Okay, must be thinking of a different one.

<pastor>ieure: is the Librewolf package made in a way that it can be used as a starting point for other Firefox forks?

<g_bor>ekaitz: sounds good to me, I would also like to see some feedback from translation team members, I think they would be in the best position both to propose and to evaluate the change

<ekaitz>g_bor: of course! in my case I didn't take part in the translation specially because I didn't want to deal with another platform. If we move to codeberg I may try to take part.

<ieure>pastor, All the Firefoxen (meaning: Firefox (in nonguix), IceCat, Mullvad, Tor Browser, LibreWolf) packaging is similar, but none of them have clean extension mechanisms. They're all different enough that I don't know whether that'd be a good idea.

<ieure>pastor, IceCat, Mullvad, and Tor Browser track the ESRs (though IceCat is one ESR behind). Firefox has both ESR and Rapid Release packages. LibreWolf only tracks Rapid Release.

<ieure>pastor, Best bet is to pick one that shares the ESR or RR property with Zen. I am personally not motivated to work on Zen Browser packages, but happy to offer feedback.

<vasi>I'm building guix from git on another distro, following the instructions in the manual. After I `make install`, I tried running `guix git authenticate ...`, but it errors: `Authenticating commits 9edb3f6 to 5c5c5a6 (100,660 new commits)...

<vasi>▕ ▏guix git: error: signature verification failed for commit e29c57ab81517424b03579147910553d92246212`. I can't seem to find any logging of _why_ it failed. Any ideas what I should do?

<pastor>ieure: cool! Thanks, I don't think I will find time soon, but I will remember to address you if I get to it.

<ieure>pastor, Sounds good!

<RamonaZero>the guix install system is sleek AF :0

<ieure>pastor, Looking at the Zen Browser source, I think LW is going to be the closest to what it's doing, packaging-wise -- they're both based on the Rapid Release channel with additional patches and such. Zen Browser build system looks a lot more complicated, they seem to be using NPM and Python.

<ieure>I appreciate that they have instructions to build from source: https://docs.zen-browser.app/guides/building

<ekaitz>vasi: I think i know the reason for that

<ekaitz>vasi: pull the `keyring` branch

<vasi>oh i did pull the keyring branch already

<ekaitz>vasi: when?

<vasi>uh a couple minutes after `git clone`?

<ekaitz>vasi: did that happen today? or before the 30th of may?

<RamonaZero>not sure how long its supposed to take, but its sure taking its time XD or it the install didn't like it

<RamonaZero>(to be fair this is a pretty low laptop... like celeron)

<ieure>RamonaZero, It takes a bit, but the low-end CPU won't help.

<ieure>RamonaZero, Are you in the USA?

<RamonaZero>yeah :0

<ieure>RamonaZero, The default Guix substitute servers are in Europe, which makes the downloading kind of slow for the US.

<RamonaZero>oh, no like its installed fine, just actually booting it up o.o

<ieure>RamonaZero, Oh, I see. Boot shouldn't take all that long.

<RamonaZero>like its a black screen, but it definitely is doing *something* since power management seems to be working like the screen dimming

<RamonaZero>maybe it doesn't like the emacs-exwm D:

<ieure>RamonaZero, Nah, that doesn't do anything until you log in.

<ekaitz>RamonaZero: did you try to log open the virtual terminals (how are they called?) ctrl+alt+F2...?

<ekaitz>maybe you are just stuck in the login?

<RamonaZero>nah, the virtual terminals (?) Don't seem to work :0

<RamonaZero>wait

<RamonaZero>XD

<RamonaZero>i think i know the issue

<RamonaZero>i had to switch from UEFI to BIOS mode xP

<ieure>RamonaZero, Huh, strange. My Guix System machines boot fine with UEFI.

<ieure>I generally set stuff to UEFI-only these days.

<vasi>ekaitz: today

<ekaitz>vasi: hm! i told you that because when I became a committer and uploaded my key, some people told me that my commits could not be verified and they didn't pull the keyring branch

<ekaitz>it's a common problem

<ekaitz>vasi: guix git: successfully authenticated commit 5c5c5a6e586923a2a6d9e2991cdc281dd9d0a7ef

<ekaitz>hmmmm

<ekaitz>weird

IRC channel logs

2025-06-03.log