IRC channel logs
2023-01-13.log
back to list of logs
<akirakyle>But I'm lazy and didn't want to spend more time debugging this <lfam>What kernel version are you using akirakyle? <lfam>Well, I'm sure your git-blame points to the first-order culprit. I do wonder if other kernel versions might work <akirakyle>There haven't been binary substitutes available for a few months due to timeouts in deblobing, which I think nckx just recently bumped up the timeout value, so there's now 6.0.18 availabe <lfam>Oh, I'm so glad to hear that. I ran out of steam on that issue. Glad to hear a critical mass of complaints emerged and spurred a fix :) <akirakyle>But then I ran into this issue with qemu when trying to reconfigure <lfam>You are an aarch64 Guix pioneer <lfam>You should get a shirt or something <lfam>Is your machine fast enough to comfortably keep debugging? <akirakyle>ha :) ngl it's been tough, but it's workable since I only try to update once in a blue moon, then usally give up when it fails <akirakyle>It's an m1 so it's plenty fast enough, but since I'm running guix system inside qemu, it's pretty memory constrained since my machine only has 8GB available, so I try not to build locally <lfam>I'm curious which model it is. I'm considering picking one up <akirakyle>If I had money and time, I'd buy some mac minis, get guix system running on bare metal on them (via asahai) and see how well they could keep up with substitutes <lfam>Those new ones are nice! I use them at work (video production) <akirakyle>I'm on a m1 macbook air, best hardware I've ever owned (but ofc macos is meh) <lfam>The 14" MBP is also super nice <lfam>And the air can compile reasonably, given that it's fanless? <akirakyle>I've never really stress tested it compiling on all cores (I only give qemu four) but single threaded perf is fantastic, and I've never had tex documents compile so fast <lfam>If you're compiling qemu and not dying of boredom, it can't be that bad <lfam>Well, hang in there re: Guix support for it. I think the situation will keep improving for a while <rekado>akirakyle: I’m trying “guix time-machine --commit=53a1fce25afdaf81c964c8867f3b9cce61b846ba -- system reconfigure config.scm” <Apo11o[m]>I’m trying to install guix in gnome boxes but it’s stuck on ssh-keygen: generating new host keys: RSA <lfam>For how long has it been stuck? <lfam>That's not normal. I can guess at what's wrong <Apo11o[m]> waits for the response while watching ssh-keygen xD <lfam>It's likely that the virtual machine can not create enough entropy to initialize the kernel's entropy pool. This is a problem in VMs. I would search the web for info on how gnome boxes deals with this issue. Other virtual machine hosting programs have mechanisms to provide random data to VMs, to avoid this problem <lfam>Search terms may be "entropy" "random" "rng" <lfam>You *could* try mashing on the keyboard while making sure that the VM is "in focus". That is, making sure it would be the program to receive your key presses <lfam>I don't know if this works in VMs or not <lfam>Just random key presses for like 30 seconds <lfam>That's what I would try first <Apo11o[m]>Random key presses just writes whatever I type <lfam>The kernel collects entropy from any source it can imagine <lfam>It won't mess anything up in the VM <Apo11o[m]>Started searching about kernel entropy and ended up looking into scihub unrelated articles lol. I switched to using the QEMU image and everything works fine, should have done that from the start <Apo11o[m]>My bad, but thanks lfam for the quick response <lfam>Well I'm glad it's working for you now <lfam>It's a fascinating subject :) <lfam>The QEMU image has some special sauce to avoid this issue <Apo11o[m]>Thermodynamics are truly fascinating, more so as related to biology. I guess I won’t have any issues if I test my config in the QEMU image? <lfam>I think that, in a VM, you can't meaningfully test things like bootloaders, disk partitioning, etc. By default, QEMU does not support ping / ICMP. But most things will work <johnabs[m]>Hi all, does anyone here use emacs-company-coq per chance? I'm seeing some strange behavior in the package that may warrant a bug report, but I usually like to get a second opinion since I'm pretty new to this stuff. <lechner>there is no shame in filing a bogus bug report. you would not be the first <mirai>lechner: still using connman? <lfam>What's a simple-to-use no fuss tool for making a slideshow presentation that's packaged in Guix? <lfam>Just text and maybe a couple pictures <lfam>Hm, without Emacs! I don't use Emacs and I'm looking for no-fuss <lfam>I'll use the 'sent' program! Super simple <Apo11o[m]>I thought every guix user uses emacs and lives in EXWM xD have you checked viewnior? Haven’t used it but heard it’s cool <lfam>Viewnior sounds nice, but I meant like a slideshow that's mostly text, for giving a presentation to a group <oriansj>DigitalKiwi: only if someone has to deal with a Windows environment. Emacs has a couple options for slideshows <oriansj>DigitalKiwi: I guess their kink is pain and they wish to inflict it on all others... <DigitalKiwi>they said they didn't want to use emacs not that they wanted to use vim ;D <oriansj>DigitalKiwi: you know evil mode exists on emacs right? <oriansj>also vim isn't pain if you can think modal <oriansj>why quit when the daemon can just rule your computer with an iron fist? <lechner>thanks to everyone for Guix, by the way! this thing is so stable (famous last words) <bumble[m]>also, for me, Guix seems to boot faster than the arch linux install it replaced, nice <oriansj>lechner: rolling releases can't by definition be stable <patched[m]>When updating, org-mode no longer seems to work. In emacs -Q, doing M-x org-mode yields Invalid function: org-assert-version. I have no issues if I revert back to the previous generation. <KarlJoad>What is the state of pipewire on Guix system right now? <zoglesby>I am using it right now to listen to music, works fine. <KarlJoad>zoglesby: Could I see a config? Are you running it through the system-level or user-level? <trevdev[m]>If I wanted to package something written in golang that wasn't some lib/extension of the language, what package file should I put it in? golang-xyz? golang-apps? Maybe I should make a contextual file such as notetaking or cli-apps? The context is a cli driven zettlekasten app <KarlJoad>zoglesby: Awesome sauce! Did you have to remove anything PulseAudio-related from the system level? <zoglesby>No, and piprewire-pulse works to bridge the gap for things that want pulse <KarlJoad>Noice. Thanks again! This will be super useful! <KarlJoad>Is auto-mounting NFS shares in the system config possible? I cannot seem to make it work with CIFS. <fiesh>KarlJoad: like (file-system (mount-point "/yolo") (type "nfs") (create-mount-point? #t) (options "lazytime")) ? <fiesh>forgot something like (device "192.168.123.123:/yolo") there, sorry <KarlJoad>fiesh: Something like that, yes. I have (file-system (mount-point "/mnt/store") (type "cifs") (device "//hostname.blah/store") (mount? #f) (create-mount-point? #t) (mount-may-fail? #t)) in mine. <fiesh>oh you mean "auto-mounting" as in "mount when needed and it works", not just "at boot time"? guess you need something like autofs then, not sure <patched[m]><patched[m]> "When updating, org-mode no..." <- Turns out I needed to remove eln-cache, now it works properly. <KarlJoad>fiesh: Either or. This mount is my personal NAS and I am the only one using this machine, so auto-mounting and mounting at boot time are roughly equivalent here. <fiesh>KarlJoad: but then just setting `mount?` to true should do the trick? or set options "_netdev,user,noauto" to allow a user to do it when they need it <lechner>fiesh / does that nfs mount actually work? mirai suggests a different solution <KarlJoad>fiesh: Ok. I was thinking it was the mount? variable. <fiesh>lechner: ah hmm I only do it with user and noauto and assumed it would work without that -- if it doesn't that's really a gotcha <fiesh>KarlJoad: sorry, maybe I'm all wrong <KarlJoad>fiesh: No worries. I will have to tinker over the weekend. I have to get work done tomorrow anyways. <cobra>Hey, I'm having trouble with Xorg on my Guix System installation. When I try to run Xorg for any purpose, my video output freezes until i SSH into the machine and kill Xorg. My Xorg log says that it couldn't load driver "intel", and I have an i965 GPU <lechner>fiesh / on the system level i believe we have an issue with network not being up in time 60300 <fiesh>lechner: oh I see -- but it doesn't seem like a smart setup to use NetworkManager for a static connection with a static NFS mount to me to be honest though <lechner>the other ones don't work either, for me <lechner>mirai_ engineered something with nm-oline <the_tubular>Hey, lechner just reporting that killing gdm seems to have solved my issue <bumble[m]>I've added TERM=xterm to the environment variables in the home configuration <bumble[m]>but when I start a new shell, and echo $TERM the value is always 'foot' <bumble[m]>holy cow, I added font-google-noto to my home configuration package list and am running reconfigure <bumble[m]>hate to wonder how huge the CJK variant will be :( <patched[m]><bumble[m]> "does anyone know of a way to..." <- alias emacs to TERM=xterm emacs? <cobra>Probably not the place to ask, but if i915 GPUs need proprietary firmware, how does Trisquel work with it? <unmatched-paren>cobra: i don't know. maybe if you were using trisquel it defaulted to software rendering and you didn't notice? <cobra>Any way I can force software rendering then? <cobra>Wait, you're sure that it needs proprietary firmware? I found someone on reddit who has done the exact same thing I'm trying to do. <unmatched-paren>and I missed "librebooted"; now i'm 190% confident it doesn't have an i915 gpu :) <cobra>What made me think it was i915 <cobra>Anyways, i915 or not, Xorg isn't working <unmatched-paren>and if it does, you'll need to use nonfree firmware to get it to work, i think <bumble[m]>`(initrd-modules (cons "i915" %base-initrd-modules)` <bumble[m]>I use a machine with i915 and that's how it is loaded in the global guix system config <unmatched-paren>it *might* only be applicable for gen 12 intel, though? this laptop is gen12 and has an i915 which needed proprietary firmware <unmatched-paren>cobra: i'd recommend you check your laptop's specs to see the type of gpu and generation <cobra>00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03) <cobra>h-node says it uses the i915 driver <cobra>So i'd have to use the evil channel? Sucks. <rekado>the evil is not in using proprietary software; the evil is in being made to use it. <bumble[m]>speaking of "evil" the guix logo looks like a taurus symbol --is this auspicious? Or does it have an evil esoteric meaning that is hidden from the profane? <rekado>bumble[m]: it’s easier than that: it’s the horns of a gnu <rekado>unmatched-paren: I’m aware of that, but the phrase has crept into this channel’s vernacular in a way that may create “facts” or misunderstandings. <unmatched-paren>bumble[m]: rekado is part of the conspiracy! You see, Taurus is a STAR sign, the Guix website has PENTAGRAM STARS on it, and the PENTAGRAM is a SATANIC mark!!!! Therefore... [continues for the length of a mid-sized novella] <rekado>bumble[m]: and a gnu’s head is the logo of the GNU project. Guix aims to fulfill the vision of the GNU project in several ways: a fully free operating system, and an overarching means to give people more power over their software <rekado>unmatched-paren: gotta tie it into the time cube <unmatched-paren>rekado: Of course. See Book 5, Part 9, Chapter 4 of my upcoming book that will enlighten us all, available at all good dodgy conspiracy websites! <bumble[m]>I'm trying to use unmatched-paren's swaybar config and there are ... seems like there are font characters that aren't found on the system, or something <bumble[m]>unmatched-parens' waybar config looks like this <bumble[m]>but on my machine, the icons look like broken font glyph <Kabouik>Thanks rekado! So I have redone a guix pull and guix package -u, but I still get the same issue in R. Isn't that how I am supposed to update r-guix-install before retrying the installation of the R package that requires jq.h? <rekado>Kabouik: do you have a definition of r-jqr in ~/.Rguix? <rekado>that’s something that guix.install doesn’t handle gracefully <rekado>it keeps adding imported package definitions to that personal package module, which will override whatever packages may have been added to Guix later <rekado>I guess I’ll need to change it so that it prefers packages from Guix proper <futurile>I'm trying to create/build a package (virtualenvwrapper). It has a dependency virtualenv-clone, which I've already packaged. At the moment I'm trying to debug the build of virtualenvwrapper: so I do --keep-build and now want to go into that environment. I normally do `guix shell python <package dependencies> --container` and then mess with the build in /tmp/. But, I don't know how to tell guix to <futurile>install the local dependency (virtualenv-clone). I can't figure out how to tell `guix shell` about this local package (it is built and in my /gnu/store). Anyone know how I should be doing this? <rekado>futurile: I don’t know what you mean by “local dependency”, but two things: “guix shell -D yourpackage” gives you an environment for hacking on “yourpackage”, and /tmp/guix-build…/ has an environment file with all variables it set during the build. <jpoiret>if your local package is defined in a file (and it is the value returned by the last statement) you can do `guix shell -f yourfile.scm` <futurile>rekado: the package I'm working on (virtualenvwrapper) depends on another package (virtualenv-clone). The dependency package (virtualenv-clone) is not in Guix yet, I've got an scm file for it - so I can build it and install it on my machine. That's what I mean that it's "local" - the package is not known by guix yet. So when I try `guix shell python virtualenv-clone` the output is `error: <futurile>python-virtualenv-clone unknown package` <Parnikkapore_m>ah, just reread the qn, is there a reason guix shell --container python -f yourPackageDefinition.scm wouldn't work? <futurile>jpoiret: ah thanks - yes that's an option I didn't know about - that works. <futurile>Parnikkapore_m: that's working - just got an environment with the 'local' package and my other dependencies <mekeor[m]>does anybody have guix system installed on a pinephone? :D <jonsger>it don't think so. but phodina[m]1 is working towards that goal, can't link the patches here as they are on an external repo... <Kabouik>Yes rekado (sorry I'm out of the office often this morning) <rekado> Kabouik: I’ll try to think of something a little more sophisticated than exploiting GUIX_PACKAGE_PATH then :) <rekado>after all, we can evaluate arbitrary Scheme expressions and build whatever the result is <rekado>I guess we could use that to effectively implement package fallbacks <jonsger>ACTION counts only 20 days left until Guix Days :) <PotentialUser-22>Hello! Hope everyone here is having a great Friday. May I ask if someone here can help me to package an application and add it to Guix library? The software is written in golang and I'm not sure how to get it built correctly. <PotentialUser-22>is there possibly a conversion tool to take an RPM SPEC file and tool it for use by Guix? <munksgaard>PotentialUser-22: Unless you have some strange dependencies, it might be pretty straightforward to package using the go-build-system? Have you looked at how other go packages are defined in guix( <apteryx>I'm trying to debug it but hints welcome! <elmoonfxm[m]>I will be teaching anyone on how to earn $20k from the crypto market,with a minimum of $200. Send a message to know more on Telegram. <civodul>apteryx: %current-system is not a triplet <civodul>to check whether you're doing a native build (as opposed to cross-compilation), you can write: (not (%current-target-system)) <civodul>oh but make-u-boot-package seems to be doing complicated things <civodul>not sure what's giong on there, forget what i wrote :-) <apteryx>civodul: is it (doing complicated things) ? :-) the idea is to use #:target and let the magic happen instead of doing a "native" cross-compilation build <apteryx>(which doesn't currently work because of python dependencies of u-boot, which in this mode are cross-compiled too, but I'm working toward fixing that) <apteryx>but that's unrelated to the issue at hand <civodul>dunno this 'same-arch?' things looks fishy to me, but "git blame" says i'm the last one who touched it, d'oh! <civodul>comes from 3bfee8ff0255f43018aedbde0b8710cd8905745f <apteryx>could probably be made a simple variable, unless I'm missing how things get delayed <civodul>it's delayed on purpose, or the fluids would be evaluated at the wrong time, as in the initial commit above <apteryx>even if this is a procedure rather than a package? <civodul>it's trying to second-guess what you're trying to do <civodul>like, you're cross-compiling, and the thing says "nope, you don't actually need to cross-compile, let me do that differently" <apteryx>I think I was surprised by something there; if #:target is not #f, it trigger the cross-compilation builder even if #:target is the same as #:system <civodul>hmm it's actually more complicated because the user is not asking for cross-compilation, as with --target <apteryx>I'm not 100% sure what I wrote is accurate, but I seem to recall trying to remove it and had to keep it <apteryx>it seems to me that guix should be smart and realize that when #:target is the same as #:system, that's a native build? is it the case? <civodul>no, it shouldn't be "smart" in the general case <civodul>i'm not on the bootloaders team though, so i'll let others figure it out :-) <apteryx>civodul: by the way, %current-system *is* a triplet <apteryx>I added a comment in my patch series above it to remember <apteryx>returns "x86_64-linux" on my machine <civodul>%current-target-system is a triplet though (and yes, that's a source of confusion) <apteryx>ah, so that's what my comment was originally aimed at, eh <apteryx>seems the struct-vtable crash occurs when calling expand-cache on u-boot-malta <apteryx>the other u-boot packages do not have such problem <klavul>anyone is packaging kandria, a new actionRPG in common lisp? <apteryx>fun, the reproducer ends up being: (supported-package? u-boot-malta) -> In procedure struct-vtable: Wrong type argument in position 1 (expecting struct): #f <lechner>sneek / later tell the_tubular / good to hear! please file a bug if possible. <mirai>neither is "longer" but I'm not sure which one is clearer <peterpolidoro>what is the scheme command to find the parent directory of (dirname (current-filename)) <bjc>the first is better, imho. match syntax is harder to read, so should be reserved for when you need more compex matching <mirai>peterpolidoro: yeah, try it in a repl <mirai>you can use arbitrary paths too <Guest39> is there an easy way to continue a guix build (while developing a package)? <mirai>bjc: match can get tricky yeah, the first one looks more "obvious" though I was under the impression the second one was a little more "compact" <civodul>kinda weird, it seems ci.guix is not processing x86 builds <bjc>at this point in my life, i strongly prefer longer and more obvious to shorter and more clever =) <apteryx>I guess the bug I'm after is because u-boot-malta is MIPS, which is no longer a supported architecture by guix <apteryx>this trips the supported-systems-procedure procedure somewhere, with an unexpected #f <peterpolidoro>when launching a guix shell container is it possible to expose the parent directory using some relative path command? <acrow>peterpolidoro: I think you can just add either --expose or --share for any other directories. <acrow>peterpolidoro: Maybe for a parent directory you might first want to create a link? <nckx>sneek: later tell peterpolidoro: Yes, relative --expose file names are supported. <acrow>nckx: a more persistent as well as authoritative response. :) <nckx>ACTION read that as authoritarian for a sec 😳 <lechner>one is trust in self, the other trust in another <nckx>I'm not sure what Peter's trying to do TBH, so maybe your symlink idea would make more sense 🤷 <apteryx>eh: (package->bag linux-libre-headers-cross-mips64el-linux-gnuabi64 (%current-system) #f) -> In procedure struct-vtable: Wrong type argument in position 1 (expecting struct): #f <dthompson>my morning: laptop had 2 hard restarts. suspected it was an overheating thing. realized thermald wasn't part of the default desktop service set so I added it to my system but it didn't work. realized that our version was tool old for intel alder lake cpus. updated to 2.5.1. thermald works! hope that helps someone else out there. <civodul>dthompson: interesting, worth sharing on the list as others are likely to stumble on this issue! <apteryx>I don't know about thermald, but perhaps it should be included in the %desktop-services ? <lechner>dthompson / civodul / thanks for the update! thermald also did not work locally, but my equipment did not restart due to otherwise sufficient cooling <dthompson>civodul: sure, I'll share. the package upgrade was very simple so I pushed an update to master. <dthompson>I don't know for sure that I was having thermal issues, but it seems like the most likely scenario. hopefully that won't happen again now that thermald is running. <apteryx>dthompson: help-guix would be fine yes <apteryx>or if you think it should be part of %desktop-services, you could start a discussion on guix-devel <elevenkb>I'm having trouble using `simple-scan` and `sane-airscan` - even though `airscan-discover` shows my scanner it doesn't show up in Simple Scan? <lechner>elevenkb / which model scanner, please? <elevenkb>would def. like to avoid that... you would have to use a USB-A to USB-A cable methinks and i don't have any of those lying around. <elevenkb>also maybe would be a good idea to configure `avahi-service-type`. <lechner>yes, definitely, but airscan is already working? <dthompson>apteryx: I don't know if it *should* be part of desktop services, but maybe there should be a laptop services list or something. <dthompson>for now I'll just mention that thermald will work for alder lake cpus now :) <lfam>USB A to A? Don't see those too often <lfam>Way more likely to be A to B <lechner>yeah, i don't know (and the other party is gone). in my experience, consumer equipment either comes with cables or uses cables that are commonly accessible <lembrun[m]>not a guix question but what do people do to have syntax highlighting for their custom macros ? <lembrun[m]>Should I put the face related code in .dir-locals.el ? <mirai>What's the proper way to use (generate-documentation) ? <mirai>or configuration->documentation <podiki[m]1>for any committers here using the pre-push hook for make authenticate: would that be better as a guix shell --pure -D guix rather than just "make"? <podiki[m]1>I can install autoconf and automake@1.16.3 of course, but seems like a good spot for guix shell <mirai>I don't think calling the repl with (with-output-to-file "/tmp/doc.txt" (lambda () (configuration->documentation 'my-configuration))) is what everyone's doing <apteryx>mirai: ,m (gnu packages your-module) <apteryx>then (configuration-documentation 'config-record-symbol) <apteryx>actually, refer to (gnu services mcron), at the bottom of the module <mirai>apteryx: it's not too easy to copy the repl output <apteryx>yeah, it's ad-hoc. we should devise a way so that make generates it and includes it itself, at build time <mirai>it ends up inserting blank spaces, etc. <mirai>as right now the documentations are prone to get desync'd <apteryx>agreed. still an improvement over the full manual sync of not using define-configuration, though <mirai>some people might edit guix.texi when they should have changed .scm instead <apteryx>comments are there to help with this, but yeah, there's a risk <apteryx>which is kind of a deal breaker for my VNC use case (the session switching menu is blank) <mirai>patch 2/2 syncs documentation between .scm and .texi <apteryx>I think patch 1/2 fixes #57958, right? <mirai>apteryx: I don't think it does? <mirai>item @code{assets} (default: @code{(#<package adwaita-icon-theme@@42.0 gnu/packages/gnome.scm:2870 7f018c9fd630> #<package gnome-themes-extra@@3.28 gnu/packages/gnome.scm:12233 7f018ca2f0b0> #<package hicolor-icon-theme@@0.17 gnu/packages/gnome.scm:3042 7f018c9fd2c0>)}) (type: list-of-file-likes) <mirai>@item @code{assets} (default: @code{(adwaita-icon-theme gnome-themes-extra hicolor-icon-theme)}) (type: list-of-file-likes) <mirai>I've thought those % lines of #57958 were comments <mirai>apteryx: that lightdm issue, could it be related to gdm missing icons? <apteryx>there should be entries (at least text) for GNOME, Xfce, ratpoison in my case <apteryx>the menu contains an icon and title per .desktop found, IIRC <mirai>btw, there's a missing gpg key for vnstat <mirai>guix refresh -u vnstat fails <apteryx>that seems more a vnstat problem than a guix one <apteryx>usually it proposes for you to retrieve the key, does it no? <mirai>it asks but it's not expecting for user input <lechner>sorry about the bot. long GPG fingerprints seems to be 40 hex long, as well <apteryx>gnupg prompts could be improved, certainly <mirai>lechner: do you get my messages? <gnucode>I started working on re-submitting the endlessh service patches. I am getting it to use (define-configuration to learn how to use that. <mirai>gnucode: I've got some patches that use define-configuration <mirai>from simple to not-so-simple <lechner>Hi, can #:cargo-inputs also take the new variable style inputs already? (for greetd) <mirai>apteryx: I think I wrote the wrong answer <mirai>I remember there being some funky issues with it not rendering empty lists <mirai>but I remembered why that condition looks like that <mirai>an empty list of strings/anything-except-packages would also satisfy that (list-of package?) predicate <apteryx>that's surprising? why does it behave that way? <apteryx>package? matches only package objects <mirai>hmm... I think this had to do with config->documentation <gnucode>mirai: I have some not-working patches. :) <gnucode>But it's fun to do it via the "right way" instead of using (define-record-type* <gnucode>I just wish the generated documentation looked better. <mirai>@item @code{assets} (default: @code{(#f #f #f)}) (type: list-of-file-likes) <mirai>the glitch gremlin has paid me a visit <mirai>apteryx: patch 2/2 should be OK to cherry-pick <mirai>gnucode: the generated docs (with configuration->documentation) tend to look alright <randommmmmm>Do you know how to delete old installed package that are now removed in Guix ?? <mirai>ah, you want automated removal <randommmmmm>mirai: but when I run "guix remove" it does not remove everything <mirai>I'm not aware if guix has a `not-in-repo' / `deprecated' package remover <mirai>gnucode: #60788 and #59866 should have some examples <mirai>randommmmmm: no, don't touch the store folder <mirai>monkeying with the /gnu/store directory invites wrath from Olympus <podiki[m]1>well /gnu/store is easily regenerated, as long as you didn't delete something really essential? <podiki[m]1>at least one doesn't have to backup /gnu/store, you get it from installing your system configuration and profiles, that's what you want backed up <mirai>ah, a rebuild scared the glitch gremlin away <mirai>@item @code{assets} (default: @code{(adwaita-icon-theme gnome-themes-extra hicolor-icon-theme)}) (type: list-of-file-likes) <mirai>apteryx: you actually spotted a consistency issue <mirai>the predicate satisfies any empty list and returns () <mirai>but empty lists should be displayed as '() <gabber>does cuirass not build for a foreign riscv64 system? i configured a spec to build `guix` for x86_64, aarch64 and riscv64 but only the former two are being built. the riscv64 doesn't show up in the evaluation and when i click on "Add a specification" in cuirass-web i can not select riscv64 as a system..? <mirai>huh, alists are rendered as '() but empty lists are () <apteryx>is there a way to inspect the content of fields without the record accessors? <mirai>you mean without using foo-configuration-<fieldnamehere> <mirai>I'm *guessing* I used to have a different format string <gnucode>apteryx: yes...I think so...just a sec... <mirai>look at serialize-configuration under gnu/services/configuration.scm <f3n1x>haó guixers of the world!... say i've already 'guix install(ed) emacs' package <f3n1x>hey gabber ! ... question: while in an old laptop as the one i'm typing this from now, am i wrong saying that if remove the package and compile it, may i perceive a performance improvement? <f3n1x>In case is worth compiling emacs locally... how to ? thanks, thanks, thanks ! <gabber>if you somehow configure the package (or the compiler) to optimize better for your machine, yes <gabber>but imho it's not usually worth the effort for small-ish things like text editors <gabber>you can force guix to build by invoking `guix build emacs --no-substitutes` <f3n1x>ah, nice to know, weeeell,not so slow...it is just that i'm probably trying to accomplish so many different things-tabs-buffers at the same time ? xD <gabber>are you running out of memory? maybe it's worth trying to debug emacs itself. there's quite a bunch of tools for emacs introspection <dthompson>ACTION sends a joycond service patch to the list <podiki[m]1>what does joycond do? gamepad/joystick configuration or some such? <podiki[m]1>f3n1x: just running guix build should give you exactly the same thing as the binary substitute unless you make a change to the package definition; you could look into "tunable" packages though for some that let you set options <elevenkb>Is there a list of instructions for using `sane-airscan` ? 
