IRC channel logs

2019-02-15.log

back to list of logs

<Elon_Satoshi>nckx: just like that? Can I have a look at the package?
<nckx>Elon_Satoshi: https://paste.debian.net/1067900/ — It's cheating, though, since we just inherit regular i3.
<nckx>I spent more time deleting some dead code from the original i3.
<nckx>* package.
<Elon_Satoshi>oh
<nckx>You sound disappoint.
<Elon_Satoshi>How do I install adobe source code pro in guix?
<Elon_Satoshi>oh here it is, font-adobe-source-code-pro
<colony>should installing glibc-utf8-locales take two hours?
<iyzsong>colony: nope, are your guix downloading it via ci.guix.info (substitutes) or building it?
<colony>I'm following the guide here: https://www.gnu.org/software/guix/manual/en/html_node/Binary-Installation.html
<colony>using guix in a foreign distro (Pop!_OS)
<colony>To be clear, this `guix pakage -i glibc-utf8-locales` is the first guix command I've run
<iyzsong>well, that's cool.. A common issue is forget to authorize the public key for the ci.guix.info substitutes server, so you end up with build it instead of download it.
<colony>(typos, sorry)
<colony>yeah that might be whats happening. Its downloading and building a whole bunch of stuff.
<iyzsong>if you see downloads from ci.guix.info, then that's ok. between step 7 (authorize the public key) and step 8 (Applicaiton Setup), you may want to run 'guix pull' to update guix to the latest version.
<colony>Yeah I did neglect to run that `guix --authorize` step. I just overlooked it. I'm killing the job and retrying now. Thanks for pointing that out
<colony>that did it.
<iyzsong>cool ^_^
<colony>:)
<ZombieChicken>Does guixsd usually include the kernel and the like in /boot?
<nckx>ZombieChicken: Never.
<ZombieChicken>Ah
<atw>ZombieChicken: my GuixSD installation includes a line in its /boot/grub/grub.cfg like "linux /gnu/store/5kyqaribmnis4fxmpaignnwcxi6nnln9-linux-libre-4.20.7/bzImage...", if that helps
<ZombieChicken>that is a bit of a problem then. Any way to get it to do so?
<nckx>ZombieChicken: Not ATM.
<nckx>or: write lots of Scheme code :-)
*nckx → zzz.
<atw>ZombieChicken: I'm curious about the constraints you're facing :)
<ZombieChicken>I'm trying to use a seperate boot partition with an encrypted root
<ZombieChicken>seperate boot is apparently okay, the docs say encrypted root is okay, so I'm not sure what's up
<atw>disk encryption is a bit beyond me, I'm afraid :)
<Elon_Satoshi>Has anyone tried running spacemacs on guix?
<ZombieChicken>I've done this kind of setup with gentoo before without any real problems
<Elon_Satoshi>I'm getting this error: Error (use-package): exec-path-from-shell/:init: Symbol's function definition is void: exec-path-from-shell-initialize
<ZombieChicken>just need the kernel and initramfs on /boot and things are easy
<Elon_Satoshi>ZombieChicken: I think GuixSD doesn't really work with a separate boot partition nowadays
<Elon_Satoshi>or at least a separate /boot partition
<atw>Elon_Satoshi: sounds like you have to get exec-path-from-shell installed, perhaps by doing M-x package-install-selected-packages? I think what you're encountering may be spacemacs-specific
<Elon_Satoshi>ZombieChicken: Why don't you try having the entire root system encrypted, and a BIOS boot partition that isn't mounted?
<ZombieChicken>BIOS boot partition?
<Elon_Satoshi>ZombieChicken: take a look at this https://wiki.archlinux.org/index.php/Dm-crypt/Encrypting_an_entire_system#Encrypted_boot_partition_(GRUB)
<Elon_Satoshi>atw: perhaps, but I only experience this problem installing spacemacs on GuixSD
<Elon_Satoshi>I also had to install it with emacs --insecure
<ZombieChicken>does guixsd support encrypted boot?
<atw>Elon_Satoshi: what do you mean by emacs --insecure?
<Elon_Satoshi>atw: that just makes emacs download packages with http instead of https
<atw>To my knowledge, emacs does not recognize such a command line flag
<atw>but in order to make SSL work on GuixSD, I believe you'll want nss-certs
<raghavgururajan>Yo All! What was that "homepage" thing you were talking about? What is the link? I was reading messages via log. Hard to see any links. LOL
<raghavgururajan>Elon_Satoshi Did you try wiping it?
<raghavgururajan>I wanted to ask this. What is life cycle of package management/processing in Guix System? For debian, it's like upstream-->package-->incoming-->unstable-->testing-->frozen-->stable. So how it is like for Guix????
<atw>raghavgururajan: depends on the package, see https://www.gnu.org/software/guix/manual/en/html_node/Submitting-Patches.html#Submitting-Patches number 7. For most packages, somebody writes a patch for an upstream release, sends a patch, and that gets merged into master.
<raghavgururajan>Ah! I see. Thanks!. So there no segregation of repositories? For example, Parabola has organised packages into different repositories like core, extra, testing etc.. So just one repo in guix??
<atw>there is one official repo, but there are also channels, which serve as extra sources for package definitions
*atw → zzz
<raghavgururajan>Those channels are not officially maintained by Guix right??
<raghavgururajan>Also, wikipedia page of all distros have a "update method" section in the box at the top right. But it's missing for GuixSD page. So whats the update method?
<dt3>isn't guix still alpha software, doesn't make much sense to have a pipeline from unstable -> stable, it's all being declared 'unstable'
<raghavgururajan>that's okay. I am asking about model of guix in general.
<tune>raghavgururajan: there are a lot of options for updating
<tune>often you start with a "guix pull" which is a bit like "apt update", pulling in fresh stuff, then you can do "guix package -u" to update installed packages
<tune>but there is also the option of declaring all your user's packages in a manifest file, and then you do "guix package -m ~/manifest.scm" for example
<tune>and then there are system updates where after doing a separate pull you'll do "guix system reconfigure /etc/config.scm"
<tune>you can have some packages in the system profile and some in a user profile, users can do unprivileged package management and update their stuff separate from system stuff
<tune>many people choose to keep their system package list small so that system updates are faster and you can get a new kernel and such quickly
<raghavgururajan>tune Ah Thanks. But I was referring to rolling vs LTS?
<raghavgururajan>Also, the branches are empty? https://git.savannah.gnu.org/git/guix.git/branches/
<raghavgururajan>Yo Guix! Can someone provide me a resource where I can learn general overview of guile scheme programming and extensions? I would like try packaging for first time in my life, so that I can contribute to the project. I am so excited.
<tune>oh, sorry. I think it's rolling
<tune>raghavgururajan: https://www.gnu.org/software/guix/manual/en/html_node/Defining-Packages.html
<tune>raghavgururajan: https://ambrevar.xyz/guix-packaging/
<tune>there are soem links I've bookmarked on packaging things, although I have yet to try it myself
<raghavgururajan>tune No issues. I see, but then why there are version numbers like 0.16?? Example, parabola being rolling model. don't have versioning. Where as, hyperbola being LTS model does.
<raghavgururajan>tune Thanks a lot!
<raghavgururajan>Let me know, once you tried. :)
<kmicu>raghavgururajan: could you share a link to an example with “GuixSD 0.16”?
<kmicu>Ah, I see it directely at https://www.gnu.org/software/guix/download/ “GNU Guix System 0.16.0”. That ‘0.16.0’ version is a thing inherited directly from Guix version.
<raghavgururajan>Hmm
<raghavgururajan>So it should be changed right??
<kmicu>It is a convention, it can be anything. On NixOS there is a rolling release channel with no versions and a stable channel with Ubuntu-like versioning (e.g. 18.09). At the same time Nix has a seperate versioning (e.g. 0.12.2).
<raghavgururajan>I am confused now. I thought rolling model means there is no versioning.
<kmicu>On Guix System and NixOS we can have both. They are not limited by traditional distros limitations. User can choose a stable releases like in Trisquel or rolling release like in Parabola (or both e.g. stable release for core sytem and rolling release for a user profile).
<raghavgururajan>Ah I am getting it. It's like hybrid. All packages are versioned, so one choose to use any (latest version or most stable version).
<raghavgururajan>That is so cooool!!!
<kmicu>It is and I miss it every time I need to use a machine w/o Guix/Nix.
<raghavgururajan>I feel ya! I am not willing to go back to hyperbola after trying guix system.
<decent-username>hello? O . o
<decent-username>I'm trying to install Guix on my Debian machine and I kind of did. But when trying to install a package it complains about not being able to install a locale. guile: warning: failed to install locale
<raghavgururajan>Check whether the locale of your current debian system is available in guix or check whther the naming convention for locales is different between debian and guix?
<iyzsong>decent-username: in short, install the glibc-utf8-locales packages as both root and user running guix, and make sure the user has the 'GUIX_LOCPATH' enviroment.
<decent-username>iyzsong: That's what I have to do. But how do I do that? It's not in the repositories apt uses.
<tune>I think you can just install it with guix itself
<decent-username>but guix isn't working. ;)
<tune>what did you do that didn't work? I recall having a problem like this in the past
<decent-username>I tried the binary install as explained here: https://www.gnu.org/software/guix/manual/en/html_node/Binary-Installation.html
<tune>make sure you've got the environment variable set and then do 'guix package -i glibc-utf8-locales'
<iyzsong>yes, the locale warning you got is just a warning, see "2.6 Application Setup" for detail..
<iyzsong>or glibc-locales, if your LANG isn't en_US.utf8, de_DE.utf8, etc..
<roptat>hi guix!
<decent-username>hi random person from the internet. o/
<roptat>hey, I don't feel random :-/
<decent-username>o . o I didn't mean to make you feel random. Mr. stranger
<decent-username>:D
<roptat>:D
<raghavgururajan>what are the available supplementary-groups for users?
<roptat>raghavgururajan, whatever is in %base-users, as well as additional groups you may define
<roptat>%base-user-accounts*
<roptat>actually %base-groups
<roptat>raghavgururajan, https://git.savannah.gnu.org/cgit/guix.git/tree/gnu/system/shadow.scm#n101
<raghavgururajan>How can I actually see the list?
<raghavgururajan>Great
<raghavgururajan>Thank you very much
<raghavgururajan>what is "lp" and "tape" ??
<roptat>lp is for bluetooth and printers I think
<roptat>I'm not sure about tape
<raghavgururajan>That is why I was not able to use bluetooth
<raghavgururajan>Thanks a lot
<raghavgururajan>Is "kmem" important?
<kmicu>raghavgururajan: lp is for printing and tape is archiving (tar = tape archiving)
<kmicu>(lp cuz https://en.wikipedia.org/wiki/Line_printer )
<raghavgururajan>ah!
<raghavgururajan>SO what is for bluetooth??
<raghavgururajan>In my computer, only root can use bluetooth, not other users.
<kmicu>raghavgururajan: Guix manual says “Users need to be in the @code{lp} group to access the D-Bus service.”. So roptat was right all along.
<raghavgururajan>Ah!
<raghavgururajan>Thank you both
*kmicu is dissapointed that there is no easily available info why bluez uses lp by default.
<raghavgururajan>Also, "plugdev" is not mentioned in guix git. Don't know why. I wouldn't be able to use removable devices right?
<kmicu>Plugdev is probably distro specific thing. We should ask what Guix System folks use for automounting removable sticks.
<kmicu>(It looks like go-to solution is a UDisks service.)
<tune>I just manually used the mount command earlier, but automounting would be nice. I did 'sudo blkid' to see what partition to mount and then just put it in /mnt/disk1.
<tune>I think when I had spacefm installed it was able to mount things easily somehow. It was broken so I uninstalled it a while ago, though.
<roptat>I use pcmanfm, and it's able to mount stuff
<roptat>I had to run my wm with dbus-launch for it to work first
<tune>hm okay. pcmanfm is what I use most of the time
<mikadozero>In my config.scm I have installed emacs-no-x when I try using ispell it says:
<mikadozero>Starting new Ispell process ispell with default dictionary...
<mikadozero>apply: Searching for program: No such file or directory, ispell
<mikadozero>Then I did guix package --search=ispell and I do not see ispell.
<mikadozero>So I added aspell and aspell-dict-en to my config.scm and did a reconfigure. It is still giving me the same message.
***rekado_ is now known as rekado
<bavier>argh, webkitgtk build fails if run with --cores=1
<mikadozero>Any suggestions on how I can try to fix this?
<rekado>mikadozero: ispell.el defines ispell-program-name.
<rekado>it tries to find aspell, then ispell, then hunspell, before settling on ispell.
<rekado>is aspell on your PATH?
<rekado>also check the value of exec-path in Emacs.
<rekado>that’s where executable-find looks for binaries.
<mikadozero>which aspell gives me /run/current-system/profile/bin/aspell
<mikadozero>rekado: I do not see a ispell.el in the .emacs.d or home directory.
<mikadozero>rekado: /run/current-system/profile/bin is in exec-path
<rekado>try M-x find-library RET ispell RET
<rekado>that’s how you can see ispell.el.
<mikadozero>rekado: Found ispell.el with C-h o ispell and followed the link to ispell.el
<mikadozero>rekado: /gnu/store/2ipzn6bjiribjyy16d08zk16j200mn42-emacs-no-x-26.1/share/emacs/26.1/lisp/textmodes/ispell.el.gz
<mikadozero>rekado: (defcustom ispell-program-name
<mikadozero> (or (executable-find "aspell")
<mikadozero> (executable-find "ispell")
<mikadozero> (executable-find "hunspell")
<mikadozero> "ispell") ...)
<kmicu>mikadozero: did you restart your emacs from the same shell you’ve intalled aspell?
<mikadozero>kmicu: No I have not restarted emacs from the same shell. I did eval-buffer ispell.el.gz
<rekado>(you don’t need to eval-buffer ispell.el)
<mikadozero>rekado: Thanks for pointing out M-x find-library I like it.
<mikadozero>I am in rcirc. I will try restarting emacs. I will be right back after I get rcirc open again.
<kmicu>@tell mikadozero You can start a second emacs instace for tests. No need to kill the irc session.
<mikadozero>I am back after restarting emacs.
<mikadozero>Looks like progress. Now I am getting a new error.
<mikadozero>ispell-init-process: Error: No word lists can be found for the language "en_US".
<kmicu>Time to install some dictionaries, not only aspell binary.
<mikadozero>kmicu: aspell-dict-en is in my config.scm but which aspell-dict-en says no aspell-dict-en
*mikadozero doing pull and reconfigure.
<nckx>mikadozero: aspell-dict-en is probably not a binary.
<nckx>So which won't work.
<mikadozero>nckx: Okay good point.
<kmicu>mikadozero: could you execute ‘aspell --help’ in your shell and check ‘Available Dictionaries:’ section?
<kmicu>(There is a comment in aspell package to ‘export ASPELL_CONF="dict-dir $HOME/.guix-profile/lib/aspell"’ or set ASPELL_DICT_DIR but I assume aspell should pick up installed dict-s in a fresh shell.)
<mikadozero>kmicu: Available Dictionaries:
<mikadozero> Dictionaries can be selected directly via the "-d" or "master"
<mikadozero> option. They can also be selected indirectly via the "lang",
<mikadozero> "variety", and "size" options.
<mikadozero>It does not list any specific dictionaries.
<kmicu>And nothing after that? Like en en-variant_0 en-variant_1 en-variant_2 en-w_accents ?
<kmicu>Is the same situation in a fresh shell?
<mikadozero>kmicu: Some blank lines then it goes to Available filters.
<kmicu>If still nothing then could you try ‘ASPELL_CONF="dict-dir $HOME/.guix-profile/lib/aspell" aspell --help’?
<mikadozero>kmicu: I just opened up bash in a new ansi-term in emacs and it the help output for Available dictionaries is the same.
<tune>emacs-stumpwm-mode fails to build! just saw it while searching for stumpwm and thought I'd grab it but I guess I'll avoid it for now
<mikadozero>kmicu: 'ASPELL_CONF="dict-dir $HOME/.guix-profile/lib/aspell" aspell --help' gives me the same output for available dictionaries.
<kmicu>mikadozero: did you install that dictionary with sudo to root profile or to your user profile?
<kmicu>Maybe you need to use /run/current-system/profile/lib/aspell intstead of $HOME/.guix-profile/lib/aspell
<mikadozero>kmicu: I installed it with a config.scm and guix system reconfigure
<mikadozero>kmicu: With this other path there is a long list of en dictionaries.
<kmicu>Ah, so it’s a system package. Do you see anything in ‘ls /run/current-system/profile/lib/aspell’?
<mikadozero>kmicu: Yes there is a lot in there.
<mikadozero>This looks like this is an issue with aspell and not emacs.
<mikadozero>When I run aspell -c file I get Error: No word lists can be found for the language "en_US".
<mikadozero>So emacs is just passing along the aspell error message.
<kmicu>mikadozero: and with ‘ASPELL_CONF="dict-dir /run/current-system/profile/lib/aspell" aspell -c file’ ?
<raghavgururajan>In this list (http://guix.info/packages/), how to differntiate between a package and a meta-package?
<mikadozero>kmicu: That work and aspell begins spell checking my config.scm
<colony>Hey all, i'm still getting the locale warning on guix package commands even though, as far as i can tell, I have a locale installed and the appropriate environment variables set. See https://pastebin.com/dmF86Xu6
<lfam>raghavgururajan: The list does not distinguish them
<lfam>colony: Okay, first of all, you can ignore those warnings safely. They are just *warnings*
<colony>I gotcha. they're just irritating
<lfam>colony: Next, this issue will go away soon: <https://bugs.gnu.org/34480>
<lfam>colony: Finally, you need to set GUIX_LOCPATH in two places. First, the environment where you run Guix and, second, the environment where the guix-daemon runs
<lfam>My advice is to ignore it
<colony>oh nice. Very cool.
<colony>thanks lfam, thats good enough for me.
<raghavgururajan>lfam It doesn't. I am looking for the one that does. Is there one?
<lfam>raghavgururajan: No. I recommend searching for packages with '(source #f)', such as the GNOME package
<raghavgururajan>Okay. What does "#f" mean?
<lfam>It means the package has no source
<raghavgururajan>Ah! Gotcha. Thanks@
<lfam>Technically the source would be the Guix package definition :)
<lfam>In Scheme #f is a false boolean value: https://www.gnu.org/software/guile/manual/html_node/Booleans.html
<lfam> https://www.cs.cmu.edu/Groups/AI/html/r4rs/r4rs_8.html
<raghavgururajan>Thanks a lot
<dustyweb>hello #guix friends!
<dustyweb>how goes Guix?
<mikadozero>kmicu: Exporting ASPELL_CONF in .bashrc makes it so I can do aspell -c file without the preamble. I am going to restart emacs to see if that gets it working in emacs as well.
<lfam>dustyweb: It goes well
<dustyweb>:)
<mikadozero>kmicu: Thank you for your help now ispell is working in emacs for me.
<mikadozero>The thing that got it all working was to add this export to .bashrc export ASPELL_CONF="dict-dir /run/current-system/profile/lib/aspell"
<ng0>oh, a online package list is back
<kmicu>Glad it works mikadozero ( ^_^)/
<phenoble>Hi everyone
<phenoble>I am following the guide on https://github.com/pjotrp/guix-notes/blob/master/GUIX-NO-ROOT.org for installing guix without root access, but I get errors that are not mentioned in there.
<lfam>phenoble: That third-party guide has not been updated since 2017. Probably it's gone a bit stale
<phenoble>lfam: Yes, I figured. Trying to get my head around what is going on there.
<lfam>Alright :)
<phenoble>I had intended to follow up with a question, but I couldn't even formulate one just yet. Let's see.
<lfam>Sometimes that's the hardest part!
<phenoble>indeed :)
<phenoble>Ok, so I'm trying to build/install the hello package, but that fails in the substitute step in fport_read with the Error "Is a directory". It looks as if the path that I am to supply to guix there was (now) a directory, but used to be a file in a previous version of guix.
<phenoble>Unfortunately I can't check what is the case, because I can't access that file hierarchy.
<phenoble>ah, ok, a typo - acl was supposed to be that file renamed
<phenoble>Now substitute vfails with "invalid access control list", looks like the public key is not valid
<Elon_Satoshi>Will Guix's 1.0.0 release be as amazing as Riot's?
<Elon_Satoshi>I know that's a slightly opinionated question but I like to look forward to things
<andswitch>'guix pull' gives me error: 'Git error: error inflating zlib stream'. How can I fix this?
<lfam>andswitch: What does `guix describe` say?
<lfam>phenoble: How did you authorize the key? And which key was it?
<phenoble>lfam: thanks, I got that resolved by overwriting the whole file once more with the one from the guide
<phenoble>lfam: now guix is refusing to to "guix perform-download: error: refusing to run with elevated privileges (UID 0)".
<phenoble>lfam: The guide states that proot should be run with the -0 flag for the daemon to fix that; though I started the daemon that way already. Puzzling.
<andswitch>lfam: pm'd you a screenshot of the situation
<lfam>andswitch: Does the error persist? I just did that step successfully?
<lfam>I mean, "I just did that step successfully."
<andswitch>yes, unfortunately
<lfam>Also FYI, there is always the `guix download` tool available as a simple downloader for when you don't have curl or wget
<andswitch>could it be https-related?
<lfam>Unfortunately the error message is not very helpful. I suppose?
<lfam>Does `guix download` successfully connect?
<andswitch>hm i think so
<tune>is anyone else running into icecat crashes? had it a few days ago and did a rollback, updated again today and it's still busted
<lfam>The message "error inflating zlib stream" comes from libgit2
<lfam>andswitch: How much RAM does your system have?
<andswitch>1G
<andswitch>660M free
<lfam>Is this a VM or bare metal?
<andswitch>VM
<lfam>Hm, I think it should be enough
<lfam>My next step would be tracing the guix-daemon with `strace -f` to try to get a little more info
<lfam>If it takes a little while to fail it may be running out of memory although I think 660M should be plenty
<lfam>You may also want to experiment by trying to clone the Git repo "by hand". That is, `git clone https://...`
<lfam>In any case this shouldn't happen so please send a bug report to <bug-guix@gnu.org> with as much detail as you can provide
<andswitch>ok, thanks
<andswitch>maybe not unrelated: last guix system reconfigure failed
<lfam>Hm, that shouldn't break anything but it also should not fail
<lfam>So that's another thing to look into
<apteryx>hello! I'm using the recently added docker service/package in Guix (for work), and I may have stumbled on some bug. When a dockerfile has a 'COPY' or 'ADD' directive, the relative path is taken not from the CWD where docker was run, but rather from some /var/lib/docker/tmp/docker-builderXXXXX/... path, so it fails finding anything.
<apteryx>if someone else is also using Docker, I'd be interested to know if this problem is reproducible on you side.
<lfam>Has anyone tried packaging magic-wormhole for Guix?
<apteryx>nevermind about the Docker issue, it's working the same on a non Guix system.
<g_bor>hello guix!
<g_bor>I have a problem.
<g_bor>I packaged prometheus.
<g_bor>It seems to work just fine, however it bundles a bunch of dependencies.
<g_bor>these reside in the vendor subfolder of the tarball.
<raghavgururajan>Query: Does anyone know where I can get a VPS with Guix System?
<ng0>mostly nowhere
<ng0>unless someone did convincing work
<g_bor>raghavgururajan: I believe there were some thread on the mailing list earlier...
<g_bor>I will try to dig it up
<ng0>some providers work with custom ISOs
<raghavgururajan>g_bor: That would be great. Can you DM me when you find it? Thanks!
<g_bor>I just had a quick look, but I can't find it right now...
<g_bor>I will try harder
<g_bor>oh, yes...
<g_bor>ng0: It was actually you who tried something with IN-Berlin in that direction...
***richi238 is now known as richi235
<ng0>did they tell you what they told me?
<g_bor>ng0: no, I don't think so. What did they tell you?
<g_bor>The mails start here by the way: https://lists.gnu.org/archive/html/guix-devel/2017-02/msg00420.html
<ng0>I had a GuixSD running on 1984 a long time ago. And given that guixsd becomes reliable etc they'd offer it in addition to debian. Hetzner is special. IN-Berlin says that you can ask them to mount an ISO, there's just too little people to support GuixSD
<ng0>this was never on-list, it was me as a customer of IN-Berlin
<ng0>i have config values somewhere for in-berlin.
<g_bor>ng0: So there is no pre-baked solution, but there are some opportunities, if you want to create your own. Is that about correct?
<jackhill>I've managed to dd the Guix install image to a disk on linode, and then sucessfully guix system init from that. It works, but more things need doing (like figuring out why the default kernel can't see the disks in paravirt mode).
<ng0>g_bor: more or less
<ng0>i mena, yes from my experience
<raghavgururajan>Thanks
<g_bor>so, regarding prometheus, should I try to unbundle the dependecies, or not
<g_bor>If I get it right this is common in go, but I'm not into go...
***slyfox_ is now known as slyfox
<ngz>I wonder if packaging Pharo is possible in Guix (with bootstrap considerations).
<civodul>ngz: Pharo relies (or used to rely?) on a VM image built literally decades ago
<civodul>they're working on fixing this, not sure what the status is
<ngz>civodul: You mean this: https://github.com/OpenSmalltalk/opensmalltalk-vm
<civodul>i didn't know this one, dunno if it's related
<civodul>actually we should ask DamienCassou :-)
<ngz>So, as long as Pharo relies on a non-bootstrappable VW, this is a big no-no for inclusion in Guix, right?
<ngz>VM*
<CompanionCube>why can't you package the VM alone
<CompanionCube>the image is technically just user data :)
<civodul>hmm welll :-)
<civodul>ngz: we do have non bootstrapped languages packages as you know
<civodul>but there are fewer of them now
<civodul>so... we could discuss it, but the big binary blob does seem bad from a bootstrapping viewpoint
*CompanionCube could see packaging the non-bootstrappable image as a no-go
<CompanionCube>but the VM itself is perfectly buildable without the image via the provided pre-generated source code
<civodul>"generated source" is an oxymoron :-)
<CompanionCube>no it's not
<civodul>what i mean is that, if it's generated, then it's not the "preferred form" for humans to edit
<civodul>ngz: but anyway, we should check the status of their bootstrapping effort
<CompanionCube>point
<g_bor>civodul: I have a question. I packaged prometheus, it seems to work, but the tarball includes a vendor directory, with a bunch of things, in source form. Should I give it a try to get those unbundled?
<civodul>g_bor: that'd be great
<ngz>hmm according to pharo.org site: "Pharo build has a fully new build process that supports its full bootstrap from sources. This will enable the production to specific (micro) images."
<CompanionCube>inb4 'port VMMaker to GNU Smalltalk and use that to bootstrap :p'
<civodul>g_bor: if it turns out to be tricky, you can at least submit what you have with explanations of what's causing troubles
<g_bor>ok, will have a look. Thanks.
<civodul>g_bor: at any rate, thanks for packaging prometheus!
<CompanionCube>apparently there is a bootstrap thing for Pharo but not for other dialects
<CompanionCube>'To bootstrap a new Pharo image you need the latest stable version of Pharo. For more information about bootstrapping, refer to guillep/PharoBootstrap.'
<CompanionCube>and then goes on to have a script
*CompanionCube has usually preferred other dialects but eh
<rain1>there is some incredibly promising work on ocaml bootstrap
<ArneBab>When I guix pull, I currently get "guix pull: error: Git error: unexpected HTTP status code: 502" — is that known? WHat can I do to fix that?
<civodul>ArneBab: looks like a transient error on the server side
<civodul>we seem to get that every now and then
<civodul>yeah git.sv.gnu.org seems to be inaccessible right now
<andswitch>civodul: how long does it usually last?
<civodul>andswitch: i don't know, it's usually short enough to not experience it but i sometimes notice it in the Cuirass logs
<ArneBab>damn — I need it right now because scribus does not build, and I need that for a project right now …
<civodul>bummer :-/
<ArneBab>yes — I should really have already installed it a few days ago
<vagrantc>hrm. guix git repository is giving me 504 errors
<vagrantc>from two different network perspectives
<ArneBab>I got 504 for a moment, now it’s back at 502
<ArneBab>does scribus build for you?
<lfam>Feel free to pull from this Github mirror which is only ~2 hours old: https://github.com/lfam/guix
<ArneBab>lfam: how do I tell guix to do that?
<ArneBab>(thank you!)
<lfam>ArneBab: `guix pull --url=https://github.com/lfam/guix`
<Sleep_Walker>was anybody successful with configuring VPN plugins for NetworkManager?
<ArneBab>lfam: thank you!
<lfam>The commits are signed so it should be safe™
<Sleep_Walker>and if so - do you have example system configuration?
<ArneBab>I run my vpn for work manually
<Sleep_Walker>well, that is nice for openvpn, but for L2TP/IPSEC it's much less comfortable
<ArneBab>I can imagine — but sadly I cannot really help with those
<ArneBab>I hope you get it working!