<pkill9>jlicht: are you julien lepiller? <pkill9>roptat: can android run graphical linux software from guix when you install guix on your phone? (since you wrote a blog post on installing guix on your phone i am asking you) <roptat>pkill9: never tried, but I don't think android is running an x server ***lurch is now known as optima
<efraim>mbakke: I'm looking at the commits in core-updates, you forgot to change swig's input to guile-2.2 <snape>ScaredySquirrel: because of a ddos attack <snape>git clone <url>, or guix pull --url=<url> <ScaredySquirrel>what the heck it says when I try to ping google.com that hte network is unreachavle <snape>hmm, that's another issue ;-) <lurch>what is the default guix url again? <snape>leoprikler: talking on IRC is a good sign that the cable is alright ;) <reepca>could be that the network isn't allowing ping requests to go through. I've had that happen on a university campus before, where I could make http requests fine but couldn't ping anything. <leoprikler>There are also some people running one machine to report the errors they encounter on another. <emyles>Anyone know why I get "guile[xxxx]: Libgcrypt warning: missing initialization - please fix the application" for guix-daemon? ***sneek_ is now known as sneek
<alextee[m]>but im not sure how to add it to the list in gdm <leoprikler>is this a full-blown desktop like GNOME or is it more like openbox? <alextee[m]>but i haven't tried it. it just looks really good in the screenshots, and i think it's mostly for freebsd distros <leoprikler>it might suffice if you just package it and add it to the packages field of your operating-system <reepca>I get a bunch of "QSslSocket: cannot resolve <some-symbol>" when running murmurd (from "mumble" package) <leoprikler>reepca: Has this not yet been reported to bug-guix? <reepca>leoprikler: great question, I am not sure. My preliminary web searches don't indicate so, but that might not be the best tool for the job. <reepca>I guess searching through email history is more effective than duckduckgo when you're subscribed to the relevant list <reepca>although that seems to refer to specifically the mumble binary, not murmurd, and it looks like the same issue occurs there. <snape>I find downloading the .mbox files and searching inside extremely effective <reepca>snape: out of curiosity, what do you use to search them? Just grep? <Franciman>I tried creating an environment containing cabal-install and ghc-8.4 packages *raghav-gururajan was just out of hectic week. Glad to be back on Guix. <raghav-gururajan>Folks! How do I separately upgrade system packages/components? For example, if just want to upgrade only one service and only one package in the system config. <zig>clearly not the good day to do a release (because of savannah issues). <zig>raghav-gururajan: it was for anybody. I was just stating the bad situation I am in. <zig>raghav-gururajan: about your question, I don't know :) <zig>I need to change it in channels.scm? <zig>PurpleSym: It does work to update guix channel, but I want to update both channels I have. And doing a guix pull --url=... only updates guix channel <zig>nevermind, it worked actually. <nomr>does anyone know what factors influence the hash in a store path? I'm trying to figure out when a substitute becomes invalid <reepca>nomr: depends on the "type" of the store path - whether it's from a derivation, plain text directly added to the store with explicit references, or some arbitrary file with no references - as well as the hash of the file, the name it's given, the references it has, the name of the store, the hash type, and maybe some other things I've forgotten. <nomr>Thanks reepca. I think this is a derivation file. I don't suppose you know what scheme func calculates that to aid debugging? <reepca>I believe there's a nice comment on exactly how it's done on line 77 of nix/libstore/store-api.cc *reepca goes looking for the scheme function *nomr goes looking for the nix source <reepca>nomr: that's actually in the guix source <reepca>by "is a derivation file", do you mean a file produced by a derivation, or an actual .drv file? <nomr>A .drv file; I think they are used in comparing substitutes with the build server? <raghav-gururajan>Folks! Under TO-DO list, I saw " As KDE Plasma (5.8 or 5.12) LTS follows the same goals than our packaging guidelines in stability and security concerns, we have decided to remove KDE Plasma and use KDE Plasma (5.8 or 5.12) LTS as our default provider of Desktop environment.". <zimoun>hi! Is savannah.gnu.org still down? <bandali>there may be intermittent issues, but it seems fairly okay to me <bandali>zimoun, hmm, did you retry a number of times? <zimoun>yes, initially I got The requested URL returned error: 502. Then Failed to connect to git.savannah.gnu.org port 443:ion refused <bandali>these things are why we can't have nice things <wingo>are mailing lists also not delivering? <bandali>wingo, yeah i haven't received any mail in my gnu account <bandali>zimoun, right. that page explains it, no? <bandali>in short, either post a comment in the bug tracker of some project <bandali>or do the other two things listed there <bandali>if you can't do those, email savannah hackers and ask them to make an exemption for your account <zimoun>bandali: I already post items to Guix trackers. Comment a bug. Send patch. And I have been removed then. Ok, I will try again... <bandali>zimoun, to be specific, the trackers have to savannah trackers, not debbugs <zimoun>bandali: ah... subtile difference. :-) So where do I find this tracker? <nckx>where group= the project name. <snape>reepca: either emacs-ag + emacs-rmail (because it's fast), or emacs-gnus <nckx>they obviously don't know what's going on. <bandali>and kind of a subtle difference indeed; it'd be nice to have that wiki article clarified <nckx>Guess I'll restore the /topic warning⌠<nckx>Wait, mailing lists too⽠<snape>reepca: with emacs-gnus (way slower) you can import the files with 'G f', or browse the lists interactively with 'B' <snape>with emacs-rmail, you can just open the .mbox files and M-x rmail-mode <bandali>i haven't received in my @gnu.org for a couple of hours ***ChanServ sets mode: +o nckx
***nckx changes topic to 'GNU Guix | â ď¸ Savannah (âguix pullâ, mailing lists) is under attack: https://frama.link/ddos | 1.0.1 is out! get it at https://guix.gnu.org | videos: https://guix.gnu.org/blog/tags/talks/ | bugs & patches: https://issues.guix.gnu.org | paste: https://paste.debian.net | Guix in high-performance computing: https://hpc.guix.info | This channel's logged: http://logs.guix.gnu.org'
***ChanServ sets mode: -o nckx
<zimoun>bandali: does the password need time to be propagated? Because I am able to login to the webpage but then cloning as member tells me "Permission denied, please try again" <bandali>also note that there are a number of ongoing issues affecting gnu infra <bandali>i'd suggest waiting a bit if possible <snape>zimoun: you need to register your ssh key on the Savannah page <snape>does 'guix pull' handle SSH auth? <snape>well, answer is: no ÂŻ\_(ă)_/ÂŻ <bandali>so i can't put an ssh repo in my channels config file? <snape>I believe you can't. I just tried 'guix pull --url=my-member-account' <zimoun>bandali, snape: thank you. Hum? do not know if I fail somewhere or if something is currently broken. Wait and see. :-) <snape>it's guile-git which doesn't support it <str1ngs>well maybe currently guix pull does not support SSH. but libgit2 does so probably that could be added somehow <snape>(I need that feature for Cuirass) <snape>str1ngs: yes, that's what the link I just pasted is about <snape>zimoun: I don't think a "Permission denied" has anything to do with the ddos attack <snape>waiting usually doesn't fix that kind of error ;) <zimoun>snape: I agree. :-) I was preferring the option: password needs time to propagate. :-) ***lurch is now known as optima
<jlicht>pkill9: jlicht --> Jelle Licht ;-) <jlicht>bah, the ddos thing is kind of annoying :/ <anadon>Who even wants Savannah down? I don't understand why. <jlicht>time to move to all of the infrastructure to gnunet ;-) <bandali>keeping in mind gnunet's main server's failure a few months ago ;) <bandali>whereby they lost a bunch of stuff, iirc <anadon>I think there is some discussion on a CAS, but it is still being worked on. <anadon>What is the practical difference between radicle and IPFS, since the former is built ontop of the other. <bavier>bandali: but that would not affect guix's use of gnunet to distribute artifacts <str1ngs>radicle is like a decentralized gitlabs or github could also replace ML patches <bandali>bavier, sure, i was mostly kidding :) <str1ngs>anadon: basically it supports p2p issues, patches/proposals <bandali>i'd love for gnunet to be usable and for guix and others to adopt it <bavier>bandali: ok, just thought I'd clarify ;) <bandali>it could be our saving grace from the current broken internet <jackhill>issues.guix.gnu.org is giving me 502 bad gateway <anadon>How should I submit my magic_enum package for review and acceptance into the main channel? <jackhill>anadon: oh, thanks. Somehow I missed that it was more than just Savannah. It pays to double check, thanks! <bandali>yeah at this point it's a whack-a-mole <bandali>e.g. debbugs goes down, fsf sysadmins bring it up (restarting the web server), and again, and again <anadon>Newbie problem, why is the following failing with a message saying /bin/ and /usr/bin are missing? `guix environment --ad-hoc --bootstrap --pure zsh` <nojr>Hi, I'm getting guix pull: error: Git error: failed to connect to git.savannah.gnu.org: Network is unreachable <nojr>Hi, I'm getting error: Git error: failed to connect to git.savannah.gnu.org: Network is unreachable <nojr>Hi, I'm getting error: Git error: failed to connect to git.savannah.gnu.org: Network is unreachable <bavier>anadon: probably something in your shell init scripts that reference /bin and /usr/bin <bavier>anadon: those dirs don't exist on the Guix System by default <anadon>bavier Now what package do I need to install for basics like `ls`? <apteryx>Should we always squash related changes in a single commit? It seems more concise to have them seperated, but they *do* are related and reverted just one wouldn't work. <sneek>Welcome back apteryx, you have 1 message. <sneek>apteryx, lfam says: Cool, I hope it's working out :) <nomr>You know, BSV can hold git repos, never loses data; probably other blockchains can do this too, costs some one-time money per megabyte. Gateways exist for those without software. An idea to include among the others. <bavier>ScaredySquirrel: I've not seen that before <bandali>i had this weird bug where launching icecat would cause my X process to sue 100% cpu <bandali>this was earlier this year on guix system <alextee[m]>ScaredySquirrel: xorg? happens in inefficient gtk apps <bandali>haven't tried reproducing it again tho <bandali>i don't recall seeing it in other distros i used <alextee[m]>i had a problem with 100% xorg usage with my software too but it was fixed with optimization <ScaredySquirrel>ok so I install icecat and then next it does not appear in the Xfce apps menu <zimoun>why with trivial-build-system ((guix build utils)) is called 2 times? For example, see the package mail.scm:mb2md. <lispmacs>hi, I'm trying to use emacs-guix inside the emacs from Guix, inside a foriegn (distro) on my work computer. But I keep getting Guile errors when I try to use any emacs-guix commands. I don't have any trouble with this on my Guix System at home. <lispmacs>I see variable guix-guile-program is set to a bin from store, so that shouldn't be a problem <nckx>zimoun: It's not called, it's imported. #:modules tells *Guix* to make the module(s) *available* in the build environment, (use-modules âŚ) [as in any *Guile* code] actually *loads* it. <lispmacs>maybe incorrect guile libraries being loaded or something? <nckx>zimoun: You'll see this in packages with all build systems, not just trivial-. For example, packages that mix the gnu-build-system with another. It's just more common (almost mandatory) to import modules yourself in trivial-. <civodul>zimoun: in mb2md, #:modules says which modules must be _imported_ in the build env (in the chroot), whereas use-modules just puts them in scope <zimoun>nckx: yes, sorry for the wrong word. But why it appears 2 times: in argumentsL #:modules ((guix build utils)) and then in #:builder with (use-modules ((guix build utils))) <nckx>zimoun: What's not clear in my best-effort explanation? <zimoun>nckx, civodul: thank you. I get it. So I will look for package where there is #:modules and no use-modules. :-) <lispmacs>hmm, in Guix repl am seeing error "no code for module (gcrypt hash)" <zimoun>nckx: your explanation is clear. :-) It just appears to me redundant. And does exit cases where modules are required in the build environment but not load? <zimoun>I was expecting only #:modules. But ok, I can live with that. :-) <zimoun>nckx, civodul: I answer to my question. Yes there are. :-) And what is the difference between #:modules and #:imported-modules? <civodul>zimoun: i think #:modules in trivial-build-system is like #:imported-modules in other build systems <civodul>lispmacs: normally, (gcrypt hash) should be visible in "guix repl" <zimoun>civodul: hum? not sure to really understand all the differences. But ok. <lispmacs>civodul: i start emacs-26.3 from the bash shell, then M-x guix RET p n (or some other guix command). <lispmacs>civodul: if I run ,use (gcrypt hash) I get same error message, no code for module <lispmacs>%load-path is giving me ("/gnu/store/yb24670f1by3kmfqpqgh1h2zczn6ml0a-emacs-guix-0.5.1.1/share/guile/site/2.2" "/home/christopher/.config/guix/current/share/guile/site/2.2" "/gnu/store/x5zg395pmrlrr43n154g5k1j2gw0w4qn-guix-1.0.1-6.0ed97e6/share/guile/site/2.2" "/gnu/store/q62w3wc35nyhjvr4bcqy563wkqx0c425-emacs-geiser-0.10/share/geiser/guile/" "/gnu/store/9alic3caqhay3h8mx4iihpmyj6ymqpcx-guile-2.2.4/share/guile/2.2" <lispmacs>"/gnu/store/9alic3caqhay3h8mx4iihpmyj6ymqpcx-guile-2.2.4/share/guile/site/2.2" "/gnu/store/9alic3caqhay3h8mx4iihpmyj6ymqpcx-guile-2.2.4/share/guile/site" "/gnu/store/9alic3caqhay3h8mx4iihpmyj6ymqpcx-guile-2.2.4/share/guile") <civodul>lispmacs: ah, then i think you probably need to have guile-gcrypt installed in your profile, because Emacs-Guix uses Geiser to run "guile" directly, not "guix repl" <lispmacs>civodul: does it, interesting. I did install guile-gcrypt from guix, but I see that the %load-path doesn't point at its location in ~/.guix-profile, or anything in ~/.guix-profile <lispmacs>civodul: so, do I need to update geiser-guile-load-path variable then? <lispmacs>or, where does *Guix REPL* pull it's load path from, anyway? <lispmacs>(emacs-guix) info page indicates the is supposed to be a guix-load-path variable, but it doesn't seem to exist <lispmacs>adding the correct profile path to geiser-guile-load-path seems to have worked <lispmacs>curious, you would think that would have been mentioned in the manual <lispmacs>hmm, I can run the commands now, but if I try to search packages I get a million (gnu pacakge *) failed to load errors <lispmacs>where are the (gnu packages) definitions stored? <zig>lisp: guix edit package-of-your-choice <lispmacs>zig: that seems to take me to the built package, but where is the actual package definition. Like, where it guix git repo stored? <zig>guix edit foobar will open the file with the definition of foobar. <zig>That way you can find the root of (gnu packages) for your user <lispmacs>zig: ok, okay. Though, I guess I really need to figure out what symlink points there, eh? <lispmacs>or I'll have to adjust it each time I run a guix pull...? <lispmacs>zig: okay, it looks like that is in /var/guix/profiles <lispmacs>yeah, that should work, found the gnu/packages dir <lispmacs>hmm, it is showing up in the load path, but still getting those errors <lispmacs>lispmacs: think I have to give up for today <civodul>hey! Savannah's down again? (i just saw the message as i logged in to Freenode) <rekado>and maybe automate the mirroring <jackhill>is inherit a guile/scheme concept or a guix concept? <nckx>jackhill: It's an explicit feature of some Guix records IIRC. I forget the name. It's not universal. <nckx>âGood newsâ is that the latest wave of 502s was actually due to a plain old bug which went unnoticed because of the DoS. <nckx>That misconfiguration's fixed now. <nckx>Is everyone here successfully guix pulling then? <nckx>jackhill: Hah, literally (guix records) đ <jackhill>nckx, rekado thanks. I'm interesting in undertanding it better to create locally-weird packages. I guess they best way to learn is to poke around in the source âş <nckx>The debbugs downage seems to have been fixed as well đ ***ChanServ sets mode: +o nckx
***nckx changes topic to 'GNU Guix | 1.0.1 is out! get it at https://guix.gnu.org | videos: https://guix.gnu.org/blog/tags/talks/ | bugs & patches: https://issues.guix.gnu.org | paste: https://paste.debian.net | Guix in high-performance computing: https://hpc.guix.info | This channel's logged: http://logs.guix.gnu.org'
***ChanServ sets mode: -o nckx
<civodul>i feel more of an incentive to fix checkout authentication now... <nckx>Not really a mirror, just a static copy of rekado's checkout. <nckx>If we script it properly (well, using the expensivest brand of duct tape) we should use --mirror or I fear refs will go wonky. <nckx>I'd have gladly taken care of it but didn't get much response so that was meh. *rekado doesnât know what --mirror does <rekado>and what âgitâ command does this switch apply to? <nckx>rekado: git clone --mirror, git push --mirror (don't use one without the other). Makes sure all refs & tags & stuff are actually in sync. Tedious & error-prone to script âby handâ. <nckx>It might not matter âin practiceâ for us, but it's one of those things that doesn't matter until it does (âŚand upstream is down :-). <rekado>I created a bare repo. Itâs not a checkout. <rekado>I suppose we can simply fetch from savannah periodically and then do âgit reset --hard upstreamâ where âupstreamâ is the savannah repo. <nckx>git reset --hard takes a branch, not origin, as argument. <nckx>But apart from that probable typo that's fine, I still don't think it's a 1:1 mirror but OK. *rekado just read about --mirror <rekado>okay, let me replace this with a proper mirror <jlicht>how do Guix System using folks do anything related to bluetooth? <jlicht>nckx: lol, thanks! This is what I have now as well (except for the <nckx>That's enough for my BT mouse. Haven't tried fancy futuristic things like OBEX transfers to my 2007 mobile. <nckx>I used to use \t (hence the â-eâ remnant) before I added⌠a second device đŽ <jlicht>is there anything that falls inbetween "use gnome" and "use bluetoothctl"? *nckx would also like to know. <pkill9>jlicht: i've been using bluetoothctl <jlicht>if only there were a convenient emacs interface for pairing/trusting/connecting BT devices :-) <rekado>uhm, if we can only pull via ssh then doing this on ci.guix.gnu.org means that someoneâs credentials would have to lie on that server. <rekado>maybe we can just create a read-only account on savannah for that purpose? <nckx>rekado: Sounds like the safest solution apart from just using HTTPS (which would freeze the mirror if this exact same thing happened again, but at least people could pull).