IRC channel logs

2019-03-05.log

back to list of logs

***rvgn is now known as Guest37358
***rvgn-net is now known as rvgn
<mikadoZero>adfeno: From 4.2 Invoking ‘guix package’ in the manual. "After having rolled back, installing, removing, or upgrading packages overwrites previous future generations. Thus, the history of the generations in a profile is always linear."
<mikadoZero>adfeno: I do not know a direct command. But there may be one.
<adfeno>Well, the good news is that I found a way to do that
<nckx>adfeno: I think the only option for now is deleting gc roots directly from /var/guix/profiles/per-user/foo.
*nckx is not up to date tho'.
<adfeno>As the user who owns that `guix pull' generation (remember that `guix pull' as well as `guix package' work per-user), do: rm /var/guix/profiles/per-user/adfeno/current-guix-2-link
<adfeno>the "2" in "current-guix-2-link" being the number you want to remove. Then do `guix gc'.
<nckx>That's more or less it.
<ngz>adfeno: I think you can do guix package -p ~/.config/guix/current --delete-generations=1m for example
<adfeno>Also, "per-user/adfeno" relates to my username/LOGNAME, so yours can be different mikadoZero.
<nckx>Nice. --list-generations works too.
<adfeno>ngz: But doesn't `guix package -p ~/.config/guix/current --delete-generations=1m' deal only with guix-profile generations?
<adfeno>What I was asking at first was how to delete those appearingi in `guix pull --list-generations'.
<ngz>That's what I'm talking about
<nckx>adfeno: See the ‘-p ~/.config/guix/current’.
<nckx>‘deal only with guix-profile generations’ → hm? That's what you're trying to delete? A guix profile? No.
<nckx>*?
<adfeno>Actually I'm trying to delete a `guix pull' generation, not a `guix package' one.
<nckx>The profile containing guix, not all of your packages.
<nckx>adfeno: Yes. We understand.
<adfeno>Hm... In that case you might be right on using the -p example, although I didn't test it yet.
<adfeno>I did see this example being given in `info guix pull' but from what I understood from that, it would be deleting the `guix package' generations, so it seems to be a misunderstanding on my part.
<adfeno>Sorry for that
<nckx>adfeno: It literally says ‘This ‘~/.config/guix/current’ profile works like any other profile created by ‘guix package’ (*note Invoking guix package::). That is, you
<nckx>can list generations, roll back to the previous generation—i.e., the previous Guix—and so on’
<adfeno>Hm... Oh! Awesome! :D
<adfeno>Now I'll go on and do another round of guix-pull and uprade :D
<nckx>Guix is awesome. ;-)
<ngz>I find documentation about profiles pretty terse.
<ngz>About "-p PROFILE": "Use PROFILE instead of the user’s default profile."
<adfeno>First time since 2010 that I find myself managing a limited storage space. Last year got the HDD internal storage replaced by an SSD one, because the old one died and I do need a trusty storage that I can move around with my laptop, but now have the caveat of less space.
<nckx>Hm. Terse is the right word.
<nckx>OTOH we shouldn't bury an explanation of what a profile is under the documentation for an option. Is there an ‘Introduction to profiles’ that could refer to?
<ngz>I don't see anything like that.
*nckx hasn't really read the manual in years :-/
<adfeno>The / partition isn't a problem for me today, what is worrying me is /home since I like to seed some at-least-legally-shareable torrents. I can't even imagine how I'll deal with this when I start to make and seed my own video publications/recordings.
<nckx>adfeno: How big is your store?
<adfeno>3,6 GiB
<nckx>Ooh. Tiny! :-D
<adfeno>Indeed it is tiny
<adfeno>My last message about /home didn't involve /gnu/store, thankfully, as the / partition isn't affected by my storage space limitations that much.
<adfeno>Taking note that the / partition has a total of 38 GiB, with 37% usage.
<adfeno>And /home's total is 66 GiB, 93% being used.
*nckx has never partitioned a drive and never will.
*nckx obviously has EFI and swap partitions but you know what I mean.
<adfeno>I used not to partition storages so much too. But I found it easier when doing distro updating.
<adfeno>s/dating/grading/
<nckx>du -hs /gnu/store → 113G /gnu/store
<nckx>I should probably throw out some garbage…
<adfeno>I must go now, my family wants to watch a film now.
<adfeno>Bye! :D
<nckx>o/
<nckx>lsl88: Sorry for not saying hullo in the outreachy channel! I wasn't actually on-line then. I just saw it today.
<lsl88>hi guix!!! goodbye as an intern
<lsl88>hello as a contributor :0
<lsl88>:)
<lsl88>nckx: hhahaha ok, no worries
<nckx>lsl88: Goodbye and welcome!
<apteryx>Has anyone had luck with the Thunar file manager? It lacks icons for me, despite having installed gnome-icon-theme, arc-icon-theme, and hicolor-icon-theme.
<vagrantc_>adwaita-icon-theme ?
*vagrantc_ just guessing wildly
<tune>neovim still segfaulting
<tune>did anyone look into this yet?
<tune>I don't think I reported it but I thought someone else mentioned it happening to them too
<mikadoZero>apteryx: As alternative that will not have icon issues (because they have no icons)You could try a file manager like dired in emacs, ranger or nnn.
<rvgn-net>@tune you mean rust.drv?
***rvgn-net is now known as rvgn
<mikadoZero>tune: I saw something about vim/neovim segfaulting on the bug mailing list.
<rvgn>First I did "guix pull", it went well. Then I did "guix package -u", the process stuck at "building /gnu/store/qsx56ga4fmyn0wzjrfg6fwj3bxrdr5zj-rust-1.24.1.drv...
<rvgn>- 'build' phase" for a very long time.
<rvgn>Like a really long time. It's been 5 hrs.
<tune>that's normal
<tune>took me more than a full night's sleep to build rust
<tune>if you wait for the build server to finish building it, you can just download a binary
<tune>you can check how much stuff is built with 'guix weather'
<rvgn>@tune Ah I see. Okay.
<brettgilio>Hey all. I installed libedit, which usually includes the editline/history.h header, but it seems it only comes with editline/readline.h on guix. Any thoughts on this?
<ebrasca>Hello I get this error with gpg in terminal http://ix.io/1CAu
<ebrasca>In emacs I get this error http://ix.io/1CAv
<vagrantc_>guix package -A gnupg ... guix package -i gnupg
<vagrantc_>i think that's the package name
<ebrasca>I have gnupg
<ebrasca>installed
<vagrantc_>which gpg2
<vagrantc_>maybe it's not in your $PATH
<vagrantc_>or is there gnupg@1 vs. gnupg@2 ?
<ebrasca>gpg --version -> "gpg (GnuPG) 2.2.13 ..."
<vagrantc_>gpg2 --version ?
<ebrasca>zsh: command not found: gpg2
<vagrantc_>i think there was another package that provided gpg2 ... gnupg-as-something
<ebrasca>it is not gnupg2
<vagrantc_>gnupg-2.0 looks like it should keep a symlink
<ebrasca>It is gnupg and I have version 2.2.12
<vagrantc_>yes, there's another package called "gnupg-2.0" that provides gpg2
<ebrasca>In web page I see gnupg 2.2.12
<vagrantc_>yes, there is also that version ... you can install them both
<ebrasca>vagrantc_: guix package: error: gnupg-2.0: unknown package
<vagrantc_>the package "gnupg" is 2.2.12, the package "gnupg-2.0" is 2.0.x and there's also "gnupg1" that's 1.x
<vagrantc_>what version of guix are you running?
<ebrasca>vagrantc_: last one
<vagrantc_>ebrasca: that does not help
<vagrantc_>ebrasca: every commit is a different version :)
<vagrantc_>guix --version
<ebrasca>guix (GNU Guix) 5ae458befc803fe357210fef343cf322d9f4bade
<ebrasca>vagrantc_: ^
<vagrantc_>gnupg-2.0 is marked as superceded, so maybe it doesn't allow you to install it
<vagrantc_>probably just need to configure git to use "gpg"
<ebrasca>how?
<vagrantc_>i've never had to do that explicitly, so not sure :/
<vagrantc_>ebrasca: are you running guix on top of another distro?
<ebrasca>no
<ebrasca>GuixSD
<ebrasca>I am in GuixSD
<vagrantc_>got it
<apteryx>mikadoZero: thanks for the reply. I already use dired for myself, but for my wife it's a bit much :-)
<apteryx>Nautilus would be a nice easy option, but I'm trying to minimize the bloat.
<mikadoZero>ebrasca: `git config user.signingkey your-key-id` to configure git to know about your signing key.
***vagrantc_ is now known as vagrantc
<ebrasca>mikadoZero: I have configured my key.
<ebrasca>mikadoZero: I have cp my .gitconfig from gentoo.
<vagrantc>yeah, it looks like it can't find the "gpg2" binary ... why it's looking for that specifically seems like a mystery
<vagrantc>there anything in your .gitconfig that might indicate which gpg binary to use?
<ebrasca>Yea
<ebrasca>[gpg] program = gpg2
<ebrasca>...
<vagrantc>just comment that out, and try again
<ebrasca>New error
<ebrasca>error: gpg falló al firmar la data
<vagrantc>if you could also set your locale to C while debugging, the copy-pasting would be easier for those who don't speak ... portugese?
<vagrantc>export LC_ALL=C
<ebrasca>vagrantc: It is spanish from spain...
<ebrasca>here : error: gpg failed to sign the data
<ebrasca>vagrantc: I searched for someting like "LC_ALL=C" , thanks!
<ebrasca>Probably I need to start pinentry daemon, I don't know how to make it in GuixSD.
<vagrantc>there are several pinentry packages ... do you have any installed?
<vagrantc>pinentry should start when needed in recent gpg versions
<ebrasca>I have pinentry and pinentry-emacs
<ebrasca>Maybe not pinentry-emacs.
<ebrasca>Nop I have pinentry-emacs
<ebrasca>vagrantc: Do you have some idea why it is not auto starting?
<vagrantc>i had a similar problem once, but it just went away eventually
<tune>how can I restart ssh on guix? I tried to guess but "sudo herd restart ssh/sshd/openssh/opensshd" all failed
<tune>I couldn't accesss my computer from my phone and I usually can
<lfam>tune: as root `herd restart ssh-daemon`
<lfam>We should make it accept more names
<tune>ah thank you
<ebrasca>vagrantc: What do you mean with "eventually"?
<ebrasca>vagrantc: Do you mean rebot?
*ebrasca is going to bed.
<feisho>Hi! I have problems setting up my personal channel. I try to import a python module from pypi. I ran the corresponding guix command and added the code to my channel. when i run 'guix package -A <x>' the package is found. However, when I run 'guix package -i <x>' I get an unbound variable error (for python-setuptools). I guess I'm doing something stupid but I cant figure out what :)
<feisho>I'm importing (gnu package python) in the module btw
***vagrantc_ is now known as vagrantc
<keerthikota>I'm intrested to participate in Gsoc and want to work on Guile based build tool and i don't have any information regarding the mentors of that project,can anyone help me in contacting with mentor of this project
<g_bor>hello guix!
<g_bor>hello keerthikota!
<g_bor>I am partially resposnible to coordinate the GSoC effort here, I will have a look at the project page to see who the mentor supposed to be.
<g_bor>ok, I see that Pjtor Prins is marked here as a primary contact.
<g_bor>If you can pm me you email address, the I will connect you.
<roptat>hi guix!
***heaven- is now known as heaven|somewhere
***heaven|somewhere is now known as heaven-
<g_bor>hello roptat!
<rekado_>oof, I wanted to build the Java bindings for OpenMPI, but that increases the closure size from 370.8 MiB to 873.2 MiB.
<rekado_>tune: see https://issues.guix.info/issue/34616; you’re welcome to participate in the discussion.
<g_bor>hello!
<g_bor>regarding the neovim segfault, it seems to be fixed upstream.
<g_bor>Can we try it?
<g_bor> https://github.com/neovim/neovim/issues/9630
<roptat>yes, please :)
<g_bor>ok, I will look around. It seems that we should update libtermkey.
<roptat>yes, I think that's it
<jlicht>hey guix!
<g_bor>roptat: it seems that we have the updated libtermkey.
<g_bor>Can you check your version?
<roptat>sure
<roptat>I'm at libtermkey 0.20
<roptat>I'll run guix pull
<g_bor>hello guix!
<g_bor>I need a bit of help with someone more experienced in licensing than me.
<g_bor>Currently the outreachy video repository is not licensed correctly.
<g_bor>What steps would be needed to get this straight?
<roptat>g_bor[m], after guix pull, I'm still at libtermkey 0.20
<roptat>ah, nevermind, I was in a weird environment
<roptat>it's 0.21.1
<g_bor>hello roptat!
<g_bor>did it fix the problem?
<roptat>g_bor, it didn't :/
<roptat>I get the same backtrace from gdb
<g_bor>roptat: ok, noted
<nly>How can i fetch multiple git repositories for a single package? I am probably doing something terrible in `chromium-uBlock-origin' (source ...) field: http://nly.info.tm:9001/guix/pkgs.scm
<rekado_>nly: you would use native-inputs for that.
<nly>thanks
<htgoebel>Hi
<htgoebel>I packages zbar, the bar-code scanner software suite. Into which .scm file should it go?
<nckx>htgoebel: aidc.scm.
<nckx>(The name is a bit obscure, to be sure.)
<htgoebel>nckx: okay. Will do so.
<htgoebel>Is there some "glob-move" easily availalbe in guix? Like "mv $out/lib/libzabrqt.* $otherour/lib/"
<nckx>htgoebel: Not that I know of (and personally I'm happy for it); use RENAME or INSTALL-FILE, and FIND-FILES.
<roptat>(for-each (lambda (file) (rename-file file (string-append "otherout/lib/" (basename file)))) (find-files "out/lib/" "libzabrqt.*"))?
<roptat>or something like this :)
<nckx>Nitpick: "^libzabrqt\\."
<nckx>Iz regex.
<roptat>ah, you're right
*nckx was writing almost exactly that code snippet when the question came. Weird.
<nckx>Is there a cmake analogue to ‘make -C subdir’? I want to run ‘cmake foo/bar.cmake’ manually since the main CMakeLists.txt makes the rule dependent on finding systemd.
<nckx>I don't know if that's even possible.
*nckx iz cmake noob.
*nckx tries ‘cmake -P’
*nckx dislikes cmake.
<nckx>A story in 3 acts.
<bavier>so, I have to use Slack (for reasons), and it doesn't "support" icecat. I can't find a user-agent addon that I can install, because addons.mozilla.org insists I need to install Firefox first :/
<nckx>bavier: about:config?
<bavier>nckx: thanks, I was trying to avoid that, but once I looked there I realized there was an outdated override already present. it all works now :)
<jlicht>bavier: I did not have any problems using slack on icecat
<jlicht>besides having to use slack, that is
<nckx>bavier: I'm curious: did it not work because of the outdated override, or because they blacklist Icecat?
<jlicht>I did have to disable the provided add-ons that come with icecat in that profile though
<bavier>nckx: outdated override. I deleted that and it works now
<nckx>👍
<bavier>so, the warning on about:config is relevant :P
<htgoebel>roptat: Thanks (Sorry for the late answer, my internet connection broke down)
<htgoebel>Now I fail to look over several (out1 "/lib" "libzbar.*") (out2 "/include/zbar" ".*.\\h") elements
<htgoebel>See https://paste.debian.net/1071668/ Error is procedure for-each: List of wrong length
<ebrasca>Hi i get
<ebrasca>error: gpg failed to sign the data fatal: failed to write commit object
<ebrasca>I have gnupg and pinentry.
<nckx>htgoebel: Indentation of ‘args’ in that snippet looks off.
<nckx>Like it's the last line of your inner lambda.
<nckx>OK, no, it's just the formatting.
<htgoebel>Just the indention is wrong.
<ebrasca>How I can fix my problem?
<nly>How can I delete multiple phases at once? (modify-phases %standard-phases (delete 'configure)) http://nly.info.tm:9001/guix/pkgs.scm (in package ublock-origin)
<pkill9>i guess with a for-each
<bavier>nly: 'modify-phases' is a shortcut for (alist-delete ...), which you could use in a fold expression: (fold alist-delete %standard-phases '(configure build check install strup))
<nly>thanks
<mbakke>Apparently Debians "93% reproducible" is somewhat exaggerated: https://lists.debian.org/debian-devel/2019/03/msg00017.html
<nly>Why is this erroring out? `(arguments #:modules ((srfi srfi-1)) ...) in package ublock-origin http://nly.info.tm:9001/guix/pkgs.scm
<nly>I need srfi-1 for `fold'
<nly>Sry, I meant this: (arguments `(#:modules ((srfi srfi-1)) ..))
<nckx>ebrasca: I have ‘pinentry-program /home/nckx/.guix-profile/bin/pinentry-tty’ in my .gnupg/gpg-agent.conf because once upon a time GPG didn't find it by default; maybe that's still the case?
<kutsurak>Hi #guix! Quick question: are there no substitutes for rust after version 1.23.0?
<roptat>kutsurak, rust takes a very long time to build
<roptat>so there might be missing substitutes :/
<roptat>the build farms will eventually build the newer versions though
<ebrasca>Rust take like 1 hour in my PC.
<ebrasca>kutsurak: 1 hour is not much
<kutsurak>True, 1 hour is acceptable. Thanks rotat and ebrasca!
<rekado_>bavier: there’s a (barely working) slack client for Emacs. But it also works in Epiphany.
*rekado_ found a way to use bluetooth audio without pulseaudio
<bavier>rekado_: thanks, didn't know!
<nly>>found a way to use bluetooth audio without pulseaudio how?
<rekado_>bluealsa
<rekado_>part of the bluez-alsa package.
<hulten>Hello.
<nly>thanks, i didn't quite get it though
<hulten>My 'guix package' commands keep hanging on rust ('check' phase).
<nly>How can i make available a directory from another package? There is a shell script in package ublock-origin which looks for the other package's files in ../uAssets... (http://nly.info.tm:9001/guix/pkgs.scm)
<ebrasca>nckx: Do I need to restart for it to take efect?
<bavier>anyone here have experience with c++ std::tuple initializers?
<nly>I don't
<nly>can copy-file handle folders as well?
<rekado_>nly: copy-recursively does
<rekado_>nly: regarding your earlier question: you may be able to just link the directory
<nly>nice, Thanks
<nly>looks nicer now: http://nly.info.tm:9001/guix/pkgs.scm
<nly>`ublock-origin'
<nly>I can't find anything in (guix build utils) to link to the other directory, maybe something like this '(derivation->output-path (package-derivation (open-connection) ublock-origin-assets))
<bavier>nly: you can use (assoc-ref inputs "ublock-origin-asserts") to get the store directory of that input, then just use the 'symlink' procedure
<nly>nice, thanks
<nly>uhh, how can i symlink to ../ ?
<bendersteed>hi there guys
<nly>is there any difference in this: (assoc-ref %build-inputs "ublock-origin-assets")?
<bavier>nly: no
<bavier>bendersteed: hi. s/guys/folks/
<bendersteed>bavier: 👍
<nly>Thanks bavier and all, this does it i think. http://nly.info.tm:9001/guix/pkgs.scm
<nly>What are your thoughts on having web-browser extensions in guix?
<bavier>I think it's generally a good idea, even better if the browsers know how to pick them up automatically from the user's profile
<nly>hmm, any clues how to do the latter?
<mbakke>nly: Maybe you'll find some clues here: https://lists.gnu.org/archive/html/guix-devel/2017-01/msg01037.html
<rvgn>Hello Guix!
<tune>rekado_: thanks for the link. I have replied to the issue with some info of my own now
<tune>(in reference to 34616, the neovim issue)
<nly>thanks
<g_bor>hello guix!
<g_bor>I had a look around the neovim segfault.
<g_bor>I am making a slow, but steady progress on this.
<g_bor>I could narrow it to a single call, so I now know whick strequal call segfaults.
***vagrantc_ is now known as vagrantc
<mikadoZero>I do not see a emacs-mu4e package. Is mu4e included in the mu package?
<nckx>mikadoZero: Yes.
<nckx>mu4e is part of mu.
<mikadoZero>nckx: Thanks