<crackedwheat>hi, i'm having some problems understanding guix installation. <crackedwheat>i took for comment claudette's husband for the heck of it but it turned out it was my username <crackedwheat>i have tried to correct it but cannot get rid of username claudette's husband. <crackedwheat>finally i removed my username and added another before doing guix system pull <Formbi>shouldn't Guix make the new user on its own? <DrimysWinteri>I have a bare guix install, nothing installed.. only console. I want to install Sway (and therefore wayland), How should I proceed? Is the a guide to do something like this? <DrimysWinteri>I have a bare guix install, nothing installed.. only console. I want to install Sway (and therefore wayland), How should I proceed? Is the a guide to do something like this? <roptat>so you're using the Guix System, right? not on a foreign distro? <roptat>the guix system uses a file (the installer should have created it as /etc/config.scm) that contains all of the system's configuration <roptat>that's where you want to declare sway as being installed. Then you reconfigure (you run guix system reconfigure /etc/config.scm) and boom! it's installed :) <roptat>there should be a line that says something like (package (append (list ...) %base-packages)) I think or simply (packages %base-packages) <DrimysWinteri>That's the problem.. I don't know where to declare it. I is enough to just write "sway" in the packages section? <roptat>I think it should look like (packages (append (list sway ...) %base-packages)) <roptat>or (packages (cons* sway ... %base-packages)), something like that <roptat>I don't remember which is generated by the installer <roptat>you can modify it just slightly to add sway, then reconfigure <roptat>but yes, just add sway to the packages list <roptat>ok, you can use this as your package declaration: (packages (cons* nss-certs sway %base-packages)) <roptat>don't use apostrophes for package names, as they create meaningless strings instead of using the package variable <roptat>also, make sure you use cons*, not cons (cons takes an element and a list, cons* takes multiple elements and a list) <roptat>(so you can use cons only to add one package to the list of %base-packages, whereas you can use cons* to add as many packages as you want to %base-packages) <roptat>try to add (use-modules (gnu packages wm)) at the top of your file <roptat>if you run "guix show sway" it will tell you "location: gnu/packages/wm.scm:1383:2" (that's filename:line number:column number) <roptat>from the filename, you can derive the name of the module: replace / with spaces and remove the extension. so gnu/packages/wm.scm gives you (gnu packages wm) <roptat>same with nss-certs btw, it tells you it's in gnu/packages/certs.scm, so (gnu packages certs) <roptat>since there are many modules under gnu/packages, we also have a more concise form to do it: use-package-modules, where you only specify the last part of the package, so (use-package-modules certs) is the same as (use-modules (gnu packages certs)) <roptat>and you can put as many module names you want in there, so (use-package-modules certs wm) imports both modules <roptat>use-package-modules is only available after you import the (gnu) module though, so make sure it's the first thing you import in your file ***caleb_ is now known as KE0VVT
<dissoc3>is the current nix package (2.3.7) failing to build for anyone else? <apteryx>I think I saw an issue created for that <apteryx>sneek: later tell civodul yes, I couldn't run guix deploy this morning. It was trying to transfer some 70 MiB ish of files and failing at 26% each time, with the traceback I pasted at that time. <dissoc3>thanks. yeah that's the same problem i was having <apteryx>perhaps just a GC library is missing <dissoc3>is there a way to set how many cores to use by default without passing it in the command? <apteryx>You can configure the guix-daemon on a Guix System to have those arguments <apteryx>(or hack the service definition for it if your are on a foreign distro). An alternative is the GUIX_BUILD_OPTIONS environment variable. <dissoc3>i'll have to look at the guix-daemon <str1ngs>dissoc3: GUIX_BUILD_OPTIONS="--cores=$(nproc) --max-jobs=1" <str1ngs>export GUIX_BUILD_OPTIONS="--cores=$(nproc) --max-jobs=1" is probably better <str1ngs>change cores to how many cores you want. and jobs to concurrent jobs <apteryx>see info '(guix)Common Build Options' or info '(guix)Using the Cnofiguration System'. The later has an example showing how to use the extra-options field of the guix-service-type. <peanutbutterandc>Quick question: does anybody know of any environment variables that I can use to point KDE/QT apps to use a theme in a non-standard location? <apteryx>dissoc3: reverting commit 62035656ca1 fixes it <apteryx>now let's see if we can fix the latest version of nix <apteryx>nothing sticks out from their release notes <apteryx>we need to apply that patch if we want the 2.3.7 version <dissoc3>so what would that include? adding a phase before build to replace those lines? <apteryx>strangely this look like it'd fix issues only if HAVE_BOEHMGC was false (not found). <apteryx>but we already have libgc as an input <apteryx>let's try passing the -enable-gc configure flag <dissoc3>sweeeet. so how did you determine that would let it build? <apteryx>the GC_free symbols come from libgc (The Boehm something GC library) <apteryx>and we had it as an input. The hint came from the patch I linked to above. <dissoc3>awesome. thanks for looking into it. <dissoc3>all because my attempt at writing a navit package failed. went to nix to look for it lol <apteryx>ah! I hope you'll give it a try again, when you feel like it :-) ***caleb_ is now known as KE0VVT
<apteryx>don't hesitate to ask help on help-guix@gnu.org. <dissoc3>for sure. i just wanted gps navigation software for tomorrow. i will get the package to work <bavier[m]1>I think I may have fixed a reproducibility bug in our "fpc" package. scrub all those build timestamps! <apteryx>weird, I can't display guix-patches ATM. <DrimysWinteri>hello, how can I see what is defined in %base-services from guix system? <apteryx>sneek: later tell DrimysWintery guix system search . | recsel -e 'location ~ "base.scm"', or look directly at the gnu/services/base.scm source file. ***caleb_ is now known as KE0VVT
<apteryx>efraim: in (gnu services linux), your zram service section is titled ";;; Kernel module loader." :-) <vits-test>Hallo Guixen. Is there any command that makes the System to boot _once_ to some generation, then to the previous one? I've my keyboard not working on U-boot prompt, so wonder. <vits-test>Probably i can make a service that runs `guix system --roll-back`? <vits-test>But that requires the system to successfully boot.. <str1ngs>vits-test: you can do guix system switch-generation 1 and then switch to the other after boot. if you have grub you can select the generation you want on boot <str1ngs>I'm assuming you are using uboot maybe <vits-test>It said "Das U-boot", but i think it not care about me much. <vits-test>But next time it will not, i'll need a serial, or if it's not so badly broken, maybe i'll learn how to ssh to it. <str1ngs>I ordered a serial cable from pine store 2 months ago. still have not got it. <str1ngs>I ended up ordering another one from amazon. <str1ngs>The following derivations would be built: <str1ngs> /gnu/store/fyf0a8cykpsavywf92awk3x1bk6pq4v1-webkitgtk-2.28.4.drv <str1ngs> /gnu/store/bwg7mkg7ndf8kmrw2mw69rh57kwwg17q-enchant-2.2.8.drv <str1ngs> /gnu/store/fhw4k0hhf4prr5w7dk64wj2ik745fsnp-geoclue-2.5.6.drv <str1ngs> /gnu/store/p20mspr75fss62f0w350cq56s6wz8rp9-libsoup-2.70.0.drv <str1ngs> /gnu/store/sr6v24a5q2lylqjws9f2dzwq5dwjwy32-php-7.4.9.drv <str1ngs>did you figure out if php is actually needed for libsoup? <vits-test>str1ngs: According to arch, for tests.. one sec <str1ngs>vits-test: btw LFS is a better resource then Arch <vits-test>Wow, LFS docs even explain what exact test it needed for. <str1ngs>there is also CLFS if you want to loss your sanity <str1ngs>vits-test: so it's only used for XMLRPC regression tests <str1ngs>tests are a good thing so can't avoid php <vits-test>str1ngs: didn't reached this in Manual yet. I just scroll it upside-down to refresh those things. Didn't changed much on laptop for long time, so almost forget how to set-up. <vits-test>There should be 'allow large adjustments', probably (2013 -> 2020) <str1ngs>vits-test: I think (service ntp-service-type) should just work. <vits-test>str1ngs: dmesg sais: "Linux version 5.8.5-gnu (nixbld@) ..." <str1ngs>uname -a are you expecting something else? <vits-test>str1ngs: i've only elogind enabled explicitly. (nixbld@) -- ? *vits-test "KASLR disabled due to lack of seed" -- noted for later wtf. <str1ngs>vits-test: does elogind provide dbus? <vits-test>str1ngs: I think "it may", but %desktop-services list dbus also. <str1ngs>yeah but if I use %desktop-services it will be many builds <str1ngs>say we have a bare minimun to get sway working that would be better <vits-test> (services (cons* (service elogind-service-type) %base-services))) <vits-test>str1ngs: Those didn't worked for me. "service restarting to fast" or similar err. <vits-test>str1ngs: Yes, but in lower than native resolution. <str1ngs>vits-test: thanks for that it saves me time. <vits-test>Some apps not working (taisei, kitty). Some working (blobwars, termite, emacs) <str1ngs>thats all that matters what other apps do you need? <str1ngs>nomad looks better on my phone the emacs :( <str1ngs>the widgets in emacs don't flow probably so half of emacs is off the screen. <vits-test>str1ngs: KDE isn't a option? Maybe it's default to X still. <str1ngs>oh no the phone uses wayland. I don't want to mess with it yet. *str1ngs php..... testing <vits-test>Shortness is sister of talent. Query "phosh linux" leads to Kali, "phosh" to actual Librem-phosh. <str1ngs>vits-test: wonder if we should check the build logs and see why webkitgtk has not been built for arm. <str1ngs>yes, but that won't say why it's not there <vits-test>str1ngs: guix build --log-file webkitgtk # tries to build webkitgtk <vits-test>Ah, commits can differ with the last attempt. <str1ngs>that's locally I mean the builds provided by ci.guix.gnu.org <vits-test>"Good enough". But how to find the commit to pull there? <str1ngs>vits-test: remind me C-g in <web-buffer> should stop loading the URI as well <str1ngs>I'm going to continue building locally I want to see what fails <str1ngs>it might be a resource issue knowing webkitgtk <str1ngs>so php has a test for each bug... and it has a lot of bugs. <vits-test>I pressed the link on webkitgtk page, and it sais that the evaluation by ci.guix happened on this commit. <vits-test>like: "good lord, make the next webkitgtk successful build happen AFTER the pull i am currently on" <str1ngs>it's time to version bump webkitgtk anyways the upstream release is 2.29.91 <str1ngs>well this is next not really stable. but I might play with it on x86_64 locally <str1ngs>vits-test: I'm going to run the guanlet on this webkitgtk build see how it pans out <str1ngs>if it builds there will be substitutes at least. <str1ngs>I'm building on a 8 core aarch64 and a 24 core x86_64 with qemu lets see who wins <vits-test>ZombieChicken: IDK for PBP, but rockpro works, more or less. <vits-test>ZombieChicken: There is u-boot-pinebook-pro-rk3399-bootloader <ZombieChicken>just talking to the people on the Pine64 IRC; apparently some blobs are needed for some firmware to work <str1ngs>ZombieChicken: hello, soon as my rockpro64 is running a little smother I'm going to start look at my PBP <ZombieChicken>str1ngs: Thanks. I may be getting a PBP in the next month or so <str1ngs>ZombieChicken: I would have looked at it by now, but I boot off of nvme support for that is not that great. <str1ngs>which you need to debug and fix uboot on PBP <vits-test>btw, this thing not needed for recent pulls (CONFIG_CGROUPS_FREEZER). <str1ngs>vits-test: that template wont work we need a template that you don't edit <str1ngs>it needs to be a modules that is used from a file <str1ngs>but it's a step in the right direction. <vits-test>str1ngs: btw, i'd forget if serial 1500000 or more. <str1ngs>Linux phone 5.7-pinephone #1 SMP PREEMPT Mon Aug 31 19:29:26 UTC 2020 aarch64 GNU/Linux <vits-test>str1ngs: guix pull.. what is webkit-is-there commit it was? <str1ngs>actually we should check failed logs :P <str1ngs>--------------------------------------------------------------------- <str1ngs>pcntl_unshare() with CLONE_NEWUSER [ext/pcntl/tests/pcntl_unshare_01.phpt] <vits-test>str1ngs: .phpt -- ohh.. it's sOo nice..! LET'SS KILLLLL IT WITTH FI-RE. <vits-test>btw, it's time to learn inferiors. Downgrade whole system ins't an option, i guess. <vits-test>str1ngs: guix package -n --with-commit=master=0f50b7e7b9191f81a5334d7314fbc656303d479e -i webkitgtk <vits-test>After realizing C-x C-x in M-x term, i wonder why i need terminal emulator? <rekado>vits-test: you need one to kill emacs when it gets stuck <rekado>vits-test: you can use âguix time-machineâ to build webkit with an older Guix <vits-test>rekado: I don't want to build WebKit on SBC :) <rekado>yes, if thereâs a substitute for an older version of Guix then you can get it with âguix time-machineâ <vits-test>rekado: `package --with-commit=master=0f50b7e7b9191f81a5334d7314fbc656303d479e -i` will not get the substitutes? <rekado>what matters is the resulting derivation <sneek>Welcome back civodul, you have 1 message! <sneek>civodul, apteryx says: yes, I couldn't run guix deploy this morning. It was trying to transfer some 70 MiB ish of files and failing at 26% each time, with the traceback I pasted at that time. <zzappie>peanutbutterandc: Hey, did you solve your modules issue? <peanutbutterandc>but yeah.... finally did that. And it uses (union-build too. Turns out, union build is supposed to be run on an empty path. Nonexistant, even. <peanutbutterandc>While someone (I forget who) did write a solution to the issue, I wasn't smart-enough to understand it. <peanutbutterandc>So yeah. Please do review those patches. There are a few things that I am concerned with in there: the name of the new package and stuff. Perhaps I could get some better suggestions? <msavoritias[m]1>Hi. I tried to follow the conversation above to install sway but I'm stuck <msavoritias[m]1><vits-test "msavoritias: ?"> I mean adding sway to my packages list roptat said above <zzappie>peanutbutterandc: Ah I see. Cool I'll check the logs out. <vits-test>msavoritias[m]1: Then (use-service-modules should work. <peanutbutterandc>zzappie, It might be a while before it is merged. I have a feeling there will be quite a bit of discussion in that thread. <vits-test>msavoritias[m]1: (use-modules (gnu) (gnu packages XYZ)) <zzappie>peanutbutterandc: Btw I'm not in a position to revew things here I'm stil learning :) <peanutbutterandc>zzappie, I see. Well, if you feel like it, do take a look at that patch and give your thoughts. (: <msavoritias[m]1><vits-test "sway needs elogind service."> I have already gnome so it should work out of the box right? <msavoritias[m]1>Mine though doesn't gave cons like the cookbooks in the packages list <vits-test>msavoritias[m]1: (append makes one list out of some lists <vits-test>(cons* take itemS and a list, and makes list <brendyyn>we have one commit with author date 1970 b3b33da07097af1c61d23f662902650dcdeb25af <msavoritias[m]1>Is it a problem that guix system reconfigure throws and error and I have to add /et c/config.scm <msavoritias[m]1>So I reconfigured the system and rebooted. But there is no option to start sway from gdm <vits-test>peanutbutterandc: info guile, 6.20.2 Using Guile Modules: see @@. But idk, just saw recently. <vits-test>msavoritias[m]1: IDK. try change to another TTY, and enter sway in console: is it works? *vits-test checks cookbook, sway. <peanutbutterandc>still... if we were to have a boilerplate package, we probably would not want to expose it to `guix search/show/and friends`.... <vits-test>peanutbutterandc: I was said that (module-ref (resolve-module '(gnu packages linux)) 'make-linux-libre*)) is de-facto the same (and it's in my config). <msavoritias[m]1>But I don't really care at this point. I will check if I can remove all of them and just start sway from command line <vits-test>msavoritias[m]1: There was some dir in ~/ gdm looking for optins (?). <civodul>it seems like there's a lot of good work there *civodul just went through a chunk of the guix-patches backlog <civodul>dear committers, join the Friday review party! :-) <civodul>dear contributors, now's a good time to reply to reviews! <rekado>PurpleSym: this looks pretty good. But are you sure that libuv isnât modified by the rstudio people? <emys>I tried to work on a guix package definition and I have trouble building guix <emys>I updated to the last origin/master checkout <emys>and when I reconfigure and make I get the error "No Guile development packages were found." <emys>gives "/gnu/store/wbpg666qkxhdwayssaix7jln0fs624gz-profile/bin/guile" <rekado>emys: I suggest running this all inside of âguix environment --pure guixâ <vits-test>`./bootstrap && ./configure --localstatedir=/var && make` <rekado>which version of Guix are you using for âguix environment --pure guixâ? <emys>ah, I think I forgot about localstatedir, let me try again <emys>guix (GNU Guix) 1.1.0rc2-1.9d0d27f <emys>with the localstatedir option in configure, I get a different error <emys>"configure: error: Guile-zlib is missing; please install it." <emys>I added guix environment --ad-hoc guile-zlib to the local environment, configure still thinks its missing <rekado>I suggest getting a more recent Guix. Yours is from April. <rekado>when configure doesnât find things itâs a good idea to check if youâre setting any obscure environment variables <rekado>I suggest 1) getting a new guix with âguix pullâ; 2) going through the output of âenvâ; 3) looking at the exact complaint in config.log <emys>I executed guix pull this morning <emys>on top of debian-testing <rekado>did you install the âguixâ package into your profile? Sounds like you did. <vits-test>emys: 2.7 of Manual sais, to upgrade the guix itself, one need to `sudo -i guix pull`, then `systemctl restart guix-daemon`. <vits-test>Or, if you'd pulled at morning, try re-login. <PurpleSym>The RStudio guys really like to vendor every single dependency for no good reason⌠<emys>vits-test, rekado thanks, with your help it works now <vits-test>emys: that was because guix in User's-profile? <emys>it seems so, all though I wonder why that would matter ***caleb_ is now known as KE0VVT
<emys>I have no knowledge of the guix internals of course <PurpleSym>rekado: Nope, ran a diff and the changes listed in the log above really are the only ones applied. Should be good. <allana>Hi guix! quick question, any dotnet core users here? fsharp? curious to know if there are any reasons not to include dotnet sdk and runtime as a package in guix. I'm not sure how easy it will be to package, but I do see it in the NixPkgs collection. <PurpleSym>Thank you very much civodul :) I can drop some local patches afterwards. <civodul>it's good to ping people when you see a patch is being forgotten! <civodul>(i said "people", not "civodul" ;-)) *vits-test :) `cat /sys/devices/system/cpu/cpu4/topology/package_cpus` -- i wonder, if that is possible to have a 1,5 cpu in package. <civodul>so you gave this LVM patch a try, right? <PurpleSym>Yes, been running the patches on my mergerfs-based storage since they were posted. *raghavgururajan peeps in <civodul>PurpleSym: isn't it a problem that 'target' is a list? <civodul>i vaguely remember that there was a limitation there, but i'm not sure if it's still holds <allana>raghavgururajan: ok, thanks for the correction. <vits-test>Is that bug, or feature: "cat: /sys/devices/system/cpu/cpu4/cpufreq/cpuinfo_cur_freq: Permission denied"? <alextee[m]>Hmm I think I will distro hop to nixos for a while, seems similar to guix with more things available <vits-test>"alextee did a bad joke. Deth to matrix users for that!"? <vits-test>alextee[m]: Do You know, do they have something like the LICENSE flag in Gentoo, to filter out the closed-source "happines"? *vits-test tryes to eat a .rar cactus <alextee[m]>I'm just having many problems with guix atm and can't be bothered to figure them out <alextee[m]>Old packages, systemd, and the package manager is a big mess <alextee[m]>I like archlinux and parabola but I find the guix/nixos style more organized now. Conventional distros seem messy <vits-test>alextee[m]: though NixOS system config a bit strange, like `pkg.system.wide { pkg.gnome; pkg.tetris; pkg.pkglinux.pkg; };` <vits-test>And when i'd specified a service, the building of config failed, as it needed to explicitly enable another service.. <vits-test>I think it will be a few weeks, not few months. *alextee[m] wonders if disk space will be a problem if he installs guix on top <rekado>PurpleSym: good, thanks for checking! In this case the only thing that probably needs changing is the commit message. <rekado>[arguments] and [inputs] should be on their own lines and should be followed by â:â <rekado>the patch line should be: * gnu/local.mk (dist_patch_DATA): Add it. <jonsger>vits-test: guix has this service dependency problem as well, try adding docker-service :p <PurpleSym>rekado: Heads up, Ludo also said he wanted to apply it above â (14:12) <civodul>rekado: yup i'm taking care of this one, will push shortly :-) <str1ngs>vits-test: I'm just building webkitgtk on 4546c0dd2f58fe9a1b1e78120a33b8c276970f62 <str1ngs>vits-test: if this builds you might need to switch to a channel derived from 4546c0dd2f58fe9a1b1e78120a33b8c276970f62 <str1ngs>vits-test: until we can get a fix into guix proper. then I think we can go back to HEAD <vits-test>str1ngs: Ah. Channels. I can do this, i think. <str1ngs>vits-test: since we pinned to 4546c0dd2f58fe9a1b1e78120a33b8c276970f62 it's not a big deal <str1ngs>the channel will be 4546c0dd2f58fe9a1b1e78120a33b8c276970f62+1 <str1ngs>vits-test: err 4546c0dd2f58fe9a1b1e78120a33b8c276970f62 plus 1 <zimoun>Does someone reported that emacs-ess seems broken with Emacs 27? The test "test-ess-roxy-literate" fails on my machine. Do I miss something? <zzappie>vits-test: so funny. Is word 'retrograd' actually used in english this way? <nly>hey, i don't know you. I'm not comfortable like this <joshuaBPMan>raghavgururajan: Guixoid RG? Really Great Reporting? <nly>could anyone recommend any company that offers bank card for crypto? <joshuaBPMan>Also my gnucode.me blog is getting pretty nice looking. I'm pretty happy about it. And it's powered by guix system! <joshuaBPMan>nly what do you mean? Like a company that helps you transfer $$ to bitcoin? <joshuaBPMan>raghavgururajan: Oh really? I didn't know your name was "Really Great". :) <nly>i'll be honest, i am looking for ways to spend crypto before i buy any <nly>i've heard there are bank cards that allow you to spend crypto on regular shop that take debit cards <joshuaBPMan>nly: hmmm. That would be kind of cool...but the one downside with crypto is that it takes sometimes days for transactions to clear. <joshuaBPMan>So you could use it perhaps on products that are getting shipped to you, but perhaps not at the store. Would be my guess. <zzappie>nly: I haven't used any. I came accross AdvCash once. But they all look a bit shady <wehlutyk>I have build failures for the `guix-jupyter` package, could someone tell me where's the best place to file a bug or find help to fix it? I'm new to the community (have been using GuixSD as main system for a month, and liking it a lot) <joshuaBPMan>wehlutyk: You can always submit a bug at via emailing bug-guix@gnu.org <zzappie>raghavgururajan: card issuing api whoa <joshuaBPMan>Or you can use emacs! take a look at your contributing section of your guix manual. It mentions how to use emacs to browse bugs. C-u M-x debbugs-gnu bug-guix RET n y I think <joshuaBPMan>wehlutyk: Yup. You can also email help-guix@gnu.org, if you don't think it's a bug, but you just want some help. <civodul>wehlutyk: i see, the issue is introduced by code that migrated from guix to guile-json <civodul>and now it's (temporarily) in both places, which causes problems <joshuaBPMan>I've been doing a lot of Haunt web development lately. <wehlutyk>civodul: how did you get there so fast? I couldn't find the full log in ci.guix.gnu.org (though what you say matches the build failures on my machine) <wehlutyk>joshuaBPMan: thanks! I used haunt for my last website and am thinking of migrating for my home page too <civodul>wehlutyk: i ran "guix build guix-jupyter" :-) <joshuaBPMan>wehlutyk: I'm thinking about getting guile-syntax-highlight set up. From the guix.gnu.org website, it looks pretty easy to do. <apteryx>chromium sometimes seems to have problems starting pulseaudio or something. The microphone will not be available and a 'hardware problem' message shown. A workaround seems to start pavucontrol before starting chromium. <joshuaBPMan>I honestly don't even remember what guix-jupyter does. :) <wehlutyk>basically what guix does to build systems and config, but for jupyter notebooks -- making them reproduceable <wehlutyk> * basically what guix does to build systems and config, but for jupyter notebooks -- making them reproducible <raingloom_>hoi! could someone look at my updated yggdrasil patches? ^u^ <raingloom_>it's not urgent, but it'd be nice to not wait another month. <zjgkkn>Hi! What's okay way for now to use tor + privoxy? tor-service-type and write own user shepherd service for privoxy? <zjgkkn>Switched to guix recently, and finally pretty happy with personal computing. Thanks you! <vits-test>zjgkkn: If privoxy can read it's settings from some arbitrary file, try to use it directly first. Just look at the files under gnu/services/, in the git-tree.. <zjgkkn>'guix edit' did not work for services? <vits-test>zjgkkn: try `guix system --help` for a short summary. <roptat>zjgkkn, guix edit works only with packages <telior>howdy guix! Not sure what the best practices are regarding the `guix system reconfigure` process: should I sudo edit the config.scm file on /etc or copy it to somewhere else, edit it and invoke `guix system reconfigure` using that copy? <leoprikler>you shouldn't have more than one admin editing files tho <telior>I see, I'll go with the sudo edit way then, ty :) <Noclip>'youtube-dl' shows an "CERTIFICATE_VERIFY_FAILED" error and is unable to download anything via https. Can someone reproduce this error? <Noclip>Eventually the issue is related to the python version. <telior>I get `/etc/config.scm:29:2: error: (service cups-service-type (cups-configuration (web-interface? #t) (extensions (list cups-filters escpr)))): invalid field specifier`, not sure what I'm doing wrong since I basically took those lines from the manual. I added a (gnu packages cups) argument to the use-modules bit, swapped a few specified packages <vits-test>(services (cons* (service cups...) %desktop-services)) <telior>I seem to be missing (gnu services cups) somewhere, I'm not sure if it's in the (use-modules (gnu) (gun packages cups) (gun services cups)), or (use-service-modules desktop networking cups) tho <vits-test>There is nice app, called recsel. `guix install recutils`. It can took some field, like: `guix system search cups| recsel -p name,location` <telior>thx, I've already installed cups with `guix package -i cups`, and already have the (use modules (gnu) (gnu packages cups)) in my config.scm, but when try to reconfigure I get `/etc/config.scm:22:4: error: cups-service-type: unbound variablehint: Did you forget a `use-modules' form?` <vits-test>telior: i just messed up around, but You'd seen the `guix system search` now. <vits-test>telior: You DONT need the cups package installed to use cups-service. <vits-test>The service should provide the needed things: for example, i do NOT need to install elogind, to use elogind-service: <telior>Oh, I thought the package provided the service <vits-test>So with tor, and the other services. Even fonts for console-font-service shouldn't be installed to (packages. <vits-test>But they, of course, can be installed. But why? <telior>I'm a bit confused because the manual says I need to install the package to configure, for example, support for epson printers: <telior>Suppose you want to enable the Web interface of CUPS and also add <telior>support for Epson printers via the escpr package and for HP <telior>printers via the hplip-minimal package. You can do that directly, <telior>like this (you need to use the (gnu packages cups) module): <telior>I'm guessing the package is needed for the added stuff that's not provided by the basic service? <telior>it's reconfiguring now, I'll report back in a bit, tyvm :) <sneek>Welcome back nckx, you have 1 message! <str1ngs>vits-test: see #nomad-browser for aarch64 webkitgtk fix <nckx>telior: It's not telling you to âinstallâ them (âguix installâ or â(packages escprâ). They need to be added to the (extensions ...) fields of the CUPS service, that's different. âInstallingâ them won't make them visible to the service. <vits-test>nckx: "Note: If you wish to use the Qt5 based GUI which comes with the hplip ___package__ then it is suggested that you install the âhplipâ __package__, either in your OS configuration file or as your user." <nckx>vits-test: I didn't see that quoted above. It was about printer support. What you quote is not an exception, it's the rule đ <nckx>vits-test: Thanks for the link. Glad it worked out. (What does âperseâ mean?) <vits-test>nckx: Stollman's answer to They in singular. <nckx>Oh, I'd seen a GNU mailing list derailment about that recently but didn't make the link. <nckx>I don't think it's awful but there's no need for it when âtheyâ is standard & has existed forever. <vits-test>nckx: in russian it's not polite to say "Đни", so i try avoid this. <nckx>I agree also think that this isn't a productive âdiscussionâ to rehash for the zillionth time. <nckx>vits-test: Interesting! I didn't know that. <vits-test>And it's a standard way to talk with strangers. <Noclip>str1ngs: Yes, why should my date not be right? <vits-test>nckx: It's like to reject the person individuality using this indirect form of "They". At least to my taste. <nckx>I can only reassure you that it's absolutely not impolite in English. Even those who oppose it wouldn't claim that (their - hah - point is that is not âcorrectâ, which is... wrong, and suspiciously postdates actual usage by a few centuries, just when gender is a hot topic again...). And now I'm guilty of furthering this discussion; time to tinker with my kernel some more! o/ <str1ngs>nckx: btw this is a wip patch to fix webkitgtk builds on aarch64 https://paste.debian.net/1162668 . I think *potentially* the version bump of php introduced some test failures. this is derived from 4546c0dd2f58fe9a1b1e78120a33b8c276970f62. which I'm kinda pinned to right now for working machines. <str1ngs>nckx: let me know if you think I should mail this or not. <nckx>Ah, hm. Is there an easy way to reproduce your tree & test it here? I have [access to] two [Guix] aarch64 machines that see too little useful use. <nckx>Would be nice to test whether it's reproducible and remove some asterisks from *potentially*. <str1ngs>nckx: you would need to guix pull --commit=4546c0dd2f58fe9a1b1e78120a33b8c276970f62 <str1ngs>the php tests are killer on arm machines I tested on three machines both failed. two native and one qemu <nckx>str1ngs: OK, so I don't need your pasted patch? <str1ngs>that patch just skips tests on aarch64. not ideal but it's interim atleast <nckx>jonsger reported âwhy you no test php update before push? it can't succeedâ after I had built it twice, so the test suite seems touchy at best. <str1ngs>I hear ya, with the tests failing with unshare that's kinda sketchy too. do we even want to test that in a chroot <str1ngs>does guix daemon chroot use unshare? <nckx>Yes, but I don't know if it falls back to something else etc. <str1ngs>I wish I had more powerful aarch64 to test this better. <str1ngs>maybe if I beowulf cluster all my teeny tiny arm machines together :P *nckx .oO powerful? aarch64? quid? <str1ngs>just to clarify my patch is based off of 46c0dd2f58fe9a1b1e78120a33b8c276970f62. but it could apply to tip as well. was just a caveat. I'm pretty sure there are no substitutes for webktigtk or php? <bdju>anyone else having horrendous icecat issues? I woke up to it having crashed or something, wasn't open. then trying to restore my 4 tabs sends my cpu to 99% and kicks up my fans and they never load after several minutes <nckx>Wow, no, but thanks for making me realise that I'm still on 68.10... <str1ngs>drakonis: pinephone with mobian. and a little guix sprinkled on top ***dddddd__ is now known as dddddd
<blackbeard[m]>Hello, after upgrading Emacs to 27.1 not Helm or Ivy work with M-x <nckx>guix copy --to=REMOTE /gnu/store/foo.drv doesn't do what I expect it to do (copy foo.drv to the remote store). All the dependencies seem to be, though. <joshuaBPMan>blackbeard[m] Ahhh. :) bummer. I had the same issue when I updated. haha. I moved everything over to ~/.config <raingloom>well, after emacs -q i moved my .emacs.d back to where it was and it's working. <raingloom>but mine crashed in a different way than others, so... this is still spooky. <raingloom>ah well. it's working now. if someone else reports the same bug, i'll look into it again. <ccao001>Hi Guix! Is anybody working on upgrading picard? <ccao001>I opened a pull request to add the mention that picard is packaged for guix. <ccao001>In case anybody would like to join in to clarify anything. I'm happy to help with getting picard updated in guix. <joshuaBPMan>blackbeard[m] I had some troubles too...Are use using emacs-seq in your config? <sneek>Welcome back jgart[m], you have 1 message! <raingloom>ccao001: have you already tried building newer versions? <raingloom>ok, then there is no duplication of effort if i just go for it :D <jgart[m]>raghavgururajan: bisq sounds like a nice addition. Let's talk more about it <ccao001>raingloom great! let me know if you run into any issues. I'd be happy to help in any way I can. Maybe mention to the picard devs that you are now working on upgrading picard to the current stable version <ccao001>raingloom can you share the build fail log with us? <ccao001>raingloom thank you! I'll be back in about 2 hours or so. I have to sign off for now <raingloom>no prob, see ya! i might be asleep by then though. <rekado>blackbeard[m]: what about .config/emacs ? <rekado>emacs 27 uses XDG directories when .emacs and .emacs.d donât exist, AFAIK <joshuaBPMan>blackbeard[m] You can still use .emacs.d Emacs will search in .config/emacs first. If that doesn't exist, then it will use .emacs.d <joshuaBPMan>blackbeard[m] run a guix hash -l emacs-seq. Check if any of those packages are on your system. <joshuaBPMan>also do a grep ~/.guix-profile/manifest 'emacs-seq'. <joshuaBPMan>That's how I found out that I was using emacs-seq. When I removed emacs-bug-hunter, my problems went away, <blackbeard[m]>I deleted everything in all the directories included .config <joshuaBPMan>blackbeard[m] Are you using emacs packages as packaged from guix? I am using a mix, which is probably not good.... <blackbeard[m]>I was sure I was using a new .emacs and .emacs.d but I rebooted to be safe <joshuaBPMan>I also think that guix's emacs should have only one autoloads file instead of several. That would make start up faster. <blackbeard[m]>joshuaBPMan: synonyms is not on guix or melpa. Only emacs wiki <joshuaBPMan>blackbeard[m] Awesome. I think Emacs 27.1 does seem a little faster and smoother <blackbeard[m]>joshuaBPMan: a thesaurus that works offline with a dictionary you have locally <pkill9>is there a CLI application for thesaurus? <joshuaBPMan>I know guix has a dictionary service, but I've never figured out how to use it...I have the service running.... <rekado>do I smell a cookbook entry being cooked upâŚ? <rekado>BTW: the reason for my signing problems with the new Emacs: my key expired⌠<joshuaBPMan>rekado Sure. I'm actually liking working on the cookbook. I've got one patch pending. <rekado>one on how to make the dictionary service work *rekado also doesnât know how to use it <joshuaBPMan>rekado Do you know of a decent guide for setting up a "master key" and several slave keys? That way you can sign with your slave key and revoke in in a year or two. Then your master key will let you create a new sub key? <rekado>I always set my keys to expire in a year, as a last resort in case I should lose them. <rekado>but I keep forgetting about this when they do expire :) <joshuaBPMan>rekado Ok. I suppose I can learn how to use the dictionary and then write up in the cookbook. <joshuaBPMan>Also, I may take a look at your free ebooks section...after I purchase an e-reader <rekado>âŚI have a free ebooks sectionâŚ? <rekado>no, itâs just a guide I found a long time ago <rekado>and itâs been my go-to âhow do I set up a GPG keyâ guide ever since <joshuaBPMan>That blog distributes ebooks....Also would guix developers be open to letting people "download" videos/ebooks/music ? I am thinking specifically of the video lectures of SICP <dustyweb>is there something to give more info when a service can't be started <dustyweb>as in, print out whatever errors it might have when trying to start it *rekado would also like to know <dustyweb>at the very least it would be nice to have a <dustyweb>"print out the command you would have run to start this service" command I think <joshuaBPMan>dustyweb I feel the pain. I just to deal with this w/ nginx. /var/log/nginx/errors.log...but it would have been nice if shepherd told me about those errors. <joshuaBPMan>dustyweb I know that the nginx service does that. "sudo herd restart nginx". When that fails, it prints out the command that it tried to run. <dustyweb>joshuaBPMan: seems like a nice feature... didn't work for tor tho <dustyweb>I figured out one way to do it manually :P <dustyweb>building /gnu/store/sz8y3b5zqr1h35n4d9gfw18xqbzsvd5l-upgrade-shepherd-services.scm.drv... <dustyweb>ok, cat that file, ok cat the file of the "out" in that scheme .drv, ok cat the tor entry <dustyweb>ok now reading the autogenerated scheme file I figured out the arguments it was called with :P <rekado>raingloom: yes, Iâm pretty sure they do <joshuaBPMan>I remember doing that at one point. That is a bit of a clunky way to do it, but it works. <dustyweb>sheperd is definitely an important but currently one of the weakest pieces in the guix system imo <dustyweb>(though I am completely booked until the end of the year personally!) <joshuaBPMan>I'm free! :) Though my development proless is somewhere between making a haunt site and contributing to the cookbook. <seepel>Hi Guix! I have a question about the light command that allows to change display brightness. I can't seem to run the command without sudo, it does come with some udev rules, is there something else I need to do to install the udev rules? <nckx>seepel: Yes. I use (service udev-service-type (udev-configuration (rules (list ⌠light âŚ)))). <nckx>You'll probably need modify-services if you use %base-services &c. <nckx>Let me test whether that actually works, since I only use it from scripts... <nckx>Oh, I see I also have it as a setuid-program. That's almost certainly redundant. <nckx>Yeah, the non-setuid one works fine too. <nckx>Oh boy, I *also* have a (simple-service 'backlight udev-service-type (list brightnessctl)) in there đł My configuration is such a mess... *nckx needs to do some autumn cleaning. <seepel>nckx: Do you know if there is someplace in the manual that would've told me this info had I been paying more attention? <seepel>Ahhh, actually I finally did find it in the base-services section <nckx>sneek: Happy to hear that. I didn't actually know; I just âknowâ these things... Don't hesitate to file bugs against the manual when something's missing. <nckx>Sigh, seepel, not sneek â <telior>vits-test: reporting back, the reconfigure worked fine, thx a lot :D <seepel>Hmmm, now my entire system seems to be re-downloading even though I reconfigured my system just the other day. I did run guix gc earlier, would that remove packages needed for my system configuration for any reason? <joshuaBPMan>seepel reconfigure redownloading the world is pretty common.