***xgqtd is now known as xgqt
<kraai>The package description of ledger-agent seems to be wrapped incorrectly. Is there a way to fix it? ***namtsui_ is now known as namtsui
<apteryx>GNUtoo: hello! you need to unquote the parameter values <apteryx>from the REPL: `(#:tests? (let ((system ,(or (%current-target-system) (%current-system)))) (string-prefix? "x86_64-linux" system))) -> $2 = (#:tests? (let ((system "x86_64-linux")) (string-prefix? "x86_64-linux" system))) <iskarian>sneek: later tell zacchae[m]: I think efraim is also working on go cross-compilation :) <efraim>sneek: later tell vagrant come join us in #guix-risc-v :) <sneek>Welcome back efraim, you have 1 message! <sneek>efraim, lfam says: Okay, let me know how I can help <efraim>yeah, i put the go cross compilation on hold while I was working on getting ... something to import correctly (I think from guix/utils) <iskarian>efraim, in cross compilation does anything in the actual build process change other than setting GOOS/GOARCH/CC_FOR_TARGET/CXX_FOR_TARGET/PKGCONFIG_FOR_TARGET? <EdLin>I seem to be having problems booting the guix OS, I've blacklisted pcspkr, removed my wifi card so it won't stop when there isn't blobs, but it now stops at udev. <EdLin>I tried it within ventoy first, but now I've dedicated another USB key to it and it still won't finish booting past an early stage from the USB key <efraim>looking at it I'm not sure what part I didn't like <efraim>oh, that's what it was, I'm not passing (guix utils) correctly somewhere <leoprikler>bost: nope, only reliable shebang still is #!/bin/sh <leoprikler>if you want to do something else, you need to package your script <bost>leoprikler: Ok thanks. (Well then I need to learn how to package stuff) <efraim>sneek: later tell lfam the go cross-build error I was seeing is this: go install: cannot install cross-compiled binaries when GOBIN is set <pkill9>does anyone know of any work done on packaging python-adblock? <pkill9>this could be wrapped into the qutebrowser package and then would have adblock-style adblocking <pkill9>where it blocks elements instead of basic hosts <pkill9>probably is it uses a mixture of rust and python <pkill9>well it's a wrapper over a rust library <zacchae[m]>So, I tried making a minimal arm64 image (without syncthing), and it crashed saying that the most likely problem is that the disk ran out of space when it was building linux-libre. The disk in question is 32 GB so I doubt that. However, when I check it with fdisk, it says the two partitions are each <5GB. Is fdisk not reliable for a failed image write, or do I need to specify something in order for the entire disk to be used? <sneek>Welcome back zacchae[m], you have 1 message! <sneek>zacchae[m], iskarian says: I think efraim is also working on go cross-compilation :) <zacchae[m]>iskarian: oh neat, though if I understand correctly, I should be able to install syncthing on the arm device so long as I make the image raw-efi and volitile <GNUtoo>apteryx: thanks a lot, I'll try as soon as the current compilation finishes *GNUtoo tried the unquote at another place before, like ,(%current-target-system) if I recall well <vldn>what is the best way to submit version bumps with little adjustments to the package? <GNUtoo>The manual has a section on sending patches <GNUtoo>The neat thing is that the patch tracking system also works by mail <GNUtoo>So it looks like a mailing list and the usual tools for sending patches to mailing lists work *GNUtoo has a hard time to find it each time, so I bookmarked it <GNUtoo>vldn: and for the commit message I typically copy the ones that already do something similar <slyfox>Oh, i'm surprised './pre-inst-env guix refresh -u sway' does not work as is. <vldn>needs new meson version > 55 and wlroots etc. <vldn>setting meson-55 to meson-next could be enough <vldn>1.6.1 works so far fine for me :) no bugs and old bugs from 1.5.1 are gone <vldn>(really ugly package, too much modules etc, but works :D) <bost>It looks like the 'disk-image' gone missing: `guix help system | grep -c disk-image` return 0. <bost>Does anyone have an explanation for that? <GNUtoo>bost: maybe it's because it has been replaced by guix system image <GNUtoo>The thing is that you now have to select an image type, 'guix system --list-image-types' list them <GNUtoo>for instance if you want a standalone image that works in a vm, you can use qcow2 for instance <bost>GNUtoo: I though the same but I get an error when I try `guix system image`. <bost>GNUtoo: and also for `guix system image --image-type=iso9660` and `guix system image --image-type=iso9660 my-os.scm` <GNUtoo>AFAIK the later one is supposed to work <GNUtoo>So the error will tell what ig going wrong <GNUtoo>(like some package failing, no more space, etc) <muradm>can one use "make -j8" on guix source tree? does it speed up things? <slyfox>i think it allows running guile in parallel to build .go files out of .scm (bulk of the time) *GNUtoo didn't check if that speeds up things but it's supposed to work once you run ./bootstrap.sh and configure --localstatedir=/var like explained in the manual <bost>GNUtoo: I have a guix in a VM and I'd like to create a bootable USB and install it on my machine along the other OSes I have already. <bost>GNUtoo: but the 'disk-image' seems to be gone so I'm stuck. Ugh. <bost>GNUtoo: `guix system image --image-type=iso9660 my-os.scm` leads to 'guix system: error: failed to load 'my-os.scm': No such file or directory' <GNUtoo>ah that's because there is no file named my-os.scm in the current directory <GNUtoo>You're supposed to create that file <GNUtoo>There are up to date examples in the manual or in guix source code <GNUtoo>but it's not for building the installer image <GNUtoo>it's for building complete systems <GNUtoo>for the installer I think there is a shortcut <muradm>slyfox: thet is what expected, just wander if guile can handle that and does things in parallel <GNUtoo>You probably need to get that gnu/system/install.scm <slyfox>muradm: i'm using 'make -j4' for a few years and had no problems with it <muradm>if one ever did guix pull, gnu/system/install.scm can be found in ~/.cache/guix/checkouts/....../gnu/system/install.scm <muradm>slyfox: ok, will try that on next iteration.. thanks <bost>GNUtoo: muradm: Thanks for telling me where to get the install.scm from <muradm>./pre-inst-env guix build <package> is there any way to force rebuild of package? <GNUtoo>muradm: I'm not sure, maybe with --rounds=2 <bost>You know, the sentence 'Have a look at gnu/system/install.scm in the source tree' doesn't really help. Source tree of what? And how to get it? - that's not really clear. <muradm>i want to run ./pre-inst-env guix build --rounds=2 my-package <muradm>bost: source tree of guix, as it was mentioned, if you did ever run "guix pull" in your vm, it is literally pulling the source tree of guix also <muradm>so ~/.cache/guix/checkouts/....../ where dots is a long directory name, mine is ~/.cache/guix/checkouts/pjmkglp4t7znuugeurpurzikxq3tnlaywmisyr27shj7apsnalwq/ <bost>muradm: I meant the explanation in the manual is too vague. <slyfox>does './pre-inst-env guix build --check my-package' do what you need? <muradm>that is where you can find source tree of guix which used by "guix ...." commands by default <bost>muradm: It'd be great to add there: "Run `find ~/.cache/guix/checkouts -name install.scm` to find the install.scm" <muradm>slyfox: yeah, --check started building, thanks, let's see how it finishes <GNUtoo>Nice, maybe the manual needs to be updated for adding --check <muradm>bost: that file you need to find for you reference, normally your *.scm files should be prepared and managed by you i think, at least this is my understanding *GNUtoo probably didn't pay enough attention when sending patches as he just checked if the 2 "builds" were the same, not what it did before <muradm>once you get used to "look at guix source" that questions will be gone :) <muradm>bost: i.e. just look in install.scm but have your own my-install.scm based on install.scm for instance, because if you will want to change/customize something, you won't be able to do it on ~/.cache/guix/checkouts/... <muradm>also after "guix pull" whatever is under checkout may/will change without your control <muradm>so once you used "gnu/system/install.scm" next time it might be different thing <muradm>just keep in mind that every system you seed with guix, you need a scheme file for it under your control, since it is your system <muradm>be it installation media, your laptop, cloud server or whatever <muradm>that is how guix works, you declare your system in my-system1.scm and then ask guix to make it <muradm>in some or another way depending on what is inside my-system1.scm <vldn>mhh is there a possibility to add reallife examples to the docpages, or some additional info that is not bound to a function programmatically? <apfel>hi there, ci.guix.gnu.org seems to be down, i know there is some mastadon status overview, but i cant find it right now. I just want to know what is going on and when it will be up again. <muradm>also having issue with accessing ci.guix.gnu.org <bost>muradm: Oh this is crazy. It doesn't work against a simple copy of the install.scm. `guix system image -t iso9660 install.scm` returns 'guix system: error: failed to load 'install.scm': No such file or directory'. It seems like the install.scm must be inside a git repository. E.g. `cd ~/.cache/guix/checkouts/<...>; guix system image -t iso9660 gnu/system/install.scm` <bost> -t iso9660 gnu/system/install.scm` <vldn>ci.guix.gnu.org is online for me <vldn>and substitutes are pulled correctly <vldn>okay no pulling for me anymore, can someone confirm? <bricewge>CI seems down here also, as well as data.guix.gnu.org <htsr>even guix.gnu.org seems to be down <vldn>website down too for me, but ci.guix.gnu.org answers to ping <vldn>lets wait, maybe patch day :D <donc>is it just me or is guix.gnu.org down? <muradm>donc: not only guix.gnu.org, not you only <roptat>looks like berlin is down? bordeaux is still online though <vldn>possible to set bordeaux via channel-with-substitutes-available? <Noisytoot>guix substitute: warning: try `--no-substitutes' if the problem persists <Noisytoot>guix substitute: error: connect*: Connection timed out <Noisytoot>substitution of /gnu/store/gm9h9j42gwgd395hpdk43rxaqypjx7lk-guix-core failed <Noisytoot>guix pull: error: corrupt input while restoring archive from #<closed: file 7f967720ba80> <Noisytoot>How can I use bordeaux.guix.gnu.org for substitutes? <vldn>sudo guix-daemon --build-users=guixbuild --substitute-urls="http <vldn>s://bordeaux.guix.gnu.org" <vldn>and right before sudo herd stop guix-daemon <roptat>if you've updated recently, you're already using bordeaux along with berlin (ci) <roptat>well, unless you have a custom value in your os declaration <roptat>or you're on a foreign distro, the systemd unit is not updated automatically <Noisytoot>roptat, I was one commit behind on Guix System (without a custom value), but it wasn't falling back to bordeaux <Zambyte>Is the guix website offline? I can't seem to reach it <lfam>Indeed it seems to be unresponsive <sneek>lfam, efraim says: the go cross-build error I was seeing is this: go install: cannot install cross-compiled binaries when GOBIN is set <lfam>It looks like all the infrastructure is down Zambyte, including the substitute server <pkill9>I'm having networking issues getting substitutes too <moshy>Guess it's time to make coffee while everything builds. It's like loading from tapes all over again <vldn>doesn't bordeaux work @pkill? <pkill9>hmm it seems not to have many substitutes <vldn>:( haven't used it either but someone said it is working <pkill9>hmm actually doesn't seem to be working for me <pkill9>i think i just stopped it trying to use normal guix substitute server so it stopped timing out <pkill9>yea it's buidlign the world for me <pkill9>it tries to build that, also same with tmux and linux <pkill9>do I need to authorize it like the normal guix servers? <vldn> (q #7D602902D3A2DBB83F8A0FB98602A754C5493B0B778C8D1DD4E0F41DE14DE34F#) <vldn>look on the website for instructions <qyliss>I'm on hotel wi-fi so entirely possible it's just me <podiki[m]>main guix site, issues, substitute server all seem down since this morning (for the last 5 hours or so?) <qyliss>oh yes, I see people talking about it half an hour ago <moshy>Anyone happen to have trouble with xfce4-whiskermenu-plugin? Got it installed but it doesn't appear as a selectable panel item <podiki[m]>and bootstrappable.org (coincidence or is it hosted by guix?) <pkill9>i dont get why guix keeps compiling new stuff on reconfigure as root when i haven't run guix pull as root <pkill9>and i've only made minor changes to the system configuration file *dzho has grabbed the iso now <podiki[m]>I've never run guix pull as root, should I? i just use sudo guix system reconfigure as needed <podiki[m]>and it picks up things I've pulled, like when adding a package I needed in system <iskarian>efraim, do you think it makes sense to keep the Go "tests" output? It includes toolchain tests, but it's not actually usable unless you create a directory union with "out" and "tests" and then set GOROOT to that union directory. <Zambyte>Hi, I'm trying to set up cgit, but I'm not quite sure how I can "overload" values from %cgit-configuration-nginx, I'm pretty new to lisp in general :D This is what I have so far if anyone can help: https://dpaste.com/DE9AATBMJ <iskarian>Zambyte, you can do (nginx-server-configuration (inherit %cgit-configuration-nginx) ...) where ... are the values you want to override <Zambyte>Hm, this is what I get when I try to do that: guix system: error: Invalid value for field nginx: #<<nginx-server-configuration> listen: ("80" "443 ssl") server-name: ("git.mysite.tld") root: #<package cgit@1.2.3 gnu/packages/version-control.scm:914 7f92f5de38c0> locations: (#<<nginx-location-configuration> uri: "@cgit" body: ("fastcgi_param SCRIPT_FILENAME $document_root/lib/cgit/cgit.cgi;" "fastcgi_param <Zambyte>PATH_INFO $uri;" "fastcgi_param QUERY_STRING $args;" "fastcgi_param HTTP_HOST $server_name;" "fastcgi_pass 127.0.0.1:9000;")>) index: ("index.html") try-files: ("$uri" "@cgit") ssl-certificate: "/etc/letsencrypt/live/mysite.tld/fullchain.pem" ssl-certificate-key: "/etc/letsencrypt/live/mysite.tld/privkey.pem" server-tokens?: #f raw-content: ()> <iskarian>Zambyte, did you replace `%cgit-configuration-nginx` with `(nginx-server-configuration ...)`? <iskarian>It looks like nginx takes a `nginx-server-configuration-list` <iskarian>perhaps it should be `(nginx (nginx-server-configuration-list (nginx-server-configuration (inherit %cgit-configuration-nginx) (listen '("80" "443 ssl")) ...))` <iskarian>rather, `(nginx (nginx-server-configuration-list (list (nginx-server-configuration (inherit %cgit-configuration-nginx) (listen '("80" "443 ssl")) ...)))` <Zambyte>weird, I can't figure out where nginx-server-configuration-list is defined. It's saying it's an unbound variable <Zambyte>It's not define-public'd in cgit.scm :\ <iskarian>oh, it's the syntax for define-configuration <iskarian>so `(nginx (list (nginx-server-configuration (inherit %cgit-configuration-nginx) (listen '("80" "443 ssl")) ...))` <iskarian>I did say I'm not familiar with services :p <Zambyte>That did not give me any errors and is building. Thank you! :) <doncatnip>maybe you can make use of it .. its a config i use for a vm of one of my projects <iskarian>doncatnip, thanks, examples are always helpful :) <Noisytoot><pkill9> i dont get why guix keeps compiling new stuff on reconfigure as root when i haven't run guix pull as root <Noisytoot>(ci.guix.gnu.org, not bordeaux.guix.gnu.org) <pkill9>that doesn't explain why it's trying to build/download new derivations <pkill9>it should already have all derivations required to build the system, since I haven't updated the revision of guix and I've already got everything in order to build current system <dstolfa>pkill9: that's really odd, do you always reconfigure as root, or did you perhaps reconfigure as another user last time? <muradm>#49969 updated, minor fix for pam-mount configuration verbosity, also default greetd package is now greetd-next