IRC channel logs


back to list of logs

<civodul>bah, i had to stop cuirass on berlin due to the send-files mess i created
<civodul>i'll re-reconfigure and restart it tomorrow morning
*civodul -> zZz
***catonano_ is now known as catonano
<drakonis>hmm, how come there's kde plasma in the repos but no service for it?
<brendyyn>drakonis: simply because the work hasn't been done I think. I heard there was someone working on kde
<drakonis>i'm aware
<drakonis>hartmut goebbel
<brendyyn>confucius say one who know need not ask.
<drakonis>was looking into some more awareness on the matter
<Aurora_v_kosmose>The certificate for's program's git ( appears to be expired.
<roptat>Aurora_v_kosmose, that's on rekado_'s server
<bavier[m]1>I see that our "boinc" package seems to be fixed now (haven't checked in a while, maybe the update back in April?)
<bavier[m]1>anyhow, I'm concerned that there might be a FSDG issue with the client recommending non-free software to run on the users computer. Has this topic been brought up before?
<bavier[m]1>I know at least a few of the recommended BOINC projects run software that is not freely licensed.
<vits-test>Holy duct tape. That's so good when one can change email address in some account without login to previous email service.
<Gooberpatrol66>was there at some point a "guix profile" command? i'm seeing references to it.
<vits-test>Gooberpatrol66: where?
<brendyyn>how can i include guix environment in a shell script such that the code after it is run inside the environments shell
<Gooberpatrol66>vits-test: someone mentioned it in an old irc log
<vits-test>brendyyn: example: `guix environment --ad-hoc tor -- sh -c 'tor'`
***apteryx is now known as Guest33093
***apteryx_ is now known as apteryx
<efraim>sneek: later tell civodul I might have a blog post follow-up to with a tensorflow example and the package that uses tensorflow-native
<efraim>sneek: botsnack
<brendyyn>Doom Emacs' Scheme module conflicts with emacs-guix. Disabling fixed all the weird bugs I was getting.
<civodul>Hello Guix!
<sneek>civodul, you have 1 message!
<sneek>civodul, efraim says: I might have a blog post follow-up to with a tensorflow example and the package that uses tensorflow-native
<efraim>I saw the blog post used --input=foo=fancy-custom-foo. I can dig up the email where switching to the native tensorflow sped up the code by a factor of 10 (IIRC)
<civodul>hey efraim, interesting!
<civodul>that's probably one of these cases where they didn't bother implementing run-time selection
<civodul>it'd be nice to run the Clear Linux technique
<civodul>too bad that link is dead
<efraim>that's where took me too
<janneke>hey civodul, i noticed that ci's %core-packages (still?) has gcc-5 instead of gcc-7 and guile-2.2 instead of 3.0 -- is that what we want?
<civodul>janneke: oh, probably not!
<civodul>well, gcc-final gets built anyway, transitively
<civodul>but still it'd be more useful to have gcc > 7 there
<civodul>hmm "Timestamp out of range; substituting 2514-05-30 01:53:03.999999999"
<brendyyn>"occurrences of the generic fftw package are replaced by fftw."
<brendyyn>last fftw should be fftw-avx i think
<civodul>brendyyn: oops, thanks
<civodul>i'll fix it
<efraim>might as well update the clearlinux link at the same time
<civodul>yup, done
<efraim>can I mount /data from my host into a guix system container by adding it into the filesystem config of the os file?
<jgart[m]>what is the main benefit of having bags as an intermediary format between packages and derivations? How are bags useful for processing? Any help understanding this would be much appreciated
<rekado_>since the Emacs upgrade I can no longer sign commits; no password prompt appears
<rekado_>any ideas?
<vits-test>rekado_: M-x pinentry-start?
<vits-test>Maybe it's just not started automatically.
<efraim>rekado_: shot in the dark, is it not just the emacs upgrade but also a change to kernel 5.8?
<rekado_>vits-test: I don’t seem to have that.
<rekado_>efraim: I’m still on 5.4
<efraim>I have a wierd bug where I can't sign anything on 5.8, thought it might be related
<rekado_>FWIW I have .gpg-agent.conf with “pinentry-program …” and the agent is running
<janneke>rekado_: updated last tuseday and works for me...
*janneke notes that doesn't really help other than "it could probably work"
<civodul>rekado_, nckx: for the record i'm still in the process of reconfiguring berlin
<civodul>bumps on the road: i cannot offload due to the send-files mess, but i cannot build locally due to
<civodul>so i'm building remotely with GUIX_DAEMON_SOCKET
<civodul>hmm i could also run an older daemon actually
<vits-test>rekado_: Try with `guix show emacs-pinentry`?
<rekado_>this just shows the package…?
<vits-test>And the instructions how to set it up.
<rekado_>civodul: I’m out of the loop. Don’t know what the send-files mess refers to.
<rekado_>vits-test: I don’t actually want to use emacs-pinentry
<rekado_>I didn’t have to in the past and I don’t want Emacs to handle passphrases directly
<vits-test>OK. I though as an temporary workaround (in case i'll work).
<efraim>is the guile for boot-program in (gnu system vm) supposed to be guile-2.2?
<civodul>rekado_: ah np, i'll post a summary on the list :-)
<civodul>i just wanted to let you know in case you were fiddling with the machine or wondering what was going on
<rekado_>I probably won’t have time to fiddle with the machine in the next three months
<civodul>sure, np
<raghavgururajan>Hello Guix!
<raghavgururajan>What are the good ways to test services?
<civodul>hey raghavgururajan!
<vits-test>raghavgururajan: make them units, and run them on systemd, then check `systemctl status SERVICE`.
<civodul>there are "system tests" for that
<civodul>easiest way is to just spawn a VM with your service (with 'guix system vm')
<raghavgururajan>vits-test, systemd? o.O
<janneke>grrr gnus' C-f (forward-mail) does some weird inlining instead of forwarding the mail
*janneke though all 27.1 upgrade issuse were handled
<vits-test>janneke: btw, is there a way to defeat the thing, when gnus changes newlines in attachments?
<raghavgururajan>civodul, o/ Right now, I have been via VM. That is, after every change in a service, I will be redoing `guix system vm`. Is this what you meant or?
<janneke>vits-test: ??
<vits-test>janneke: i sent a patch with gnus, but `git apply` didn't worked.
<vits-test>nly said, i'd weird newlines, or so.
<janneke>vits-test: i never had that
<vits-test>janneke: OK, thanks. That good, btw.
<civodul>janneke: oh i use S o m to forward messages
<civodul>raghavgururajan: yes!
<civodul>and for automated testing, see and (gnu tests ...)
<janneke>civodul: hmm, that has the same message-mangling effect here
<janneke>-------------------- Start of forwarded message --------------------
<janneke>Subject: ....
<janneke>instead of properly forwarding the full e-mail
<vits-test>O:D gnus \:<
<vits-test>work or not to work?
<raghavgururajan>civodul: Thanks! The problem I face with that is, the relevant errors are not shown. For example, I modified gdm and it did not start witout any errors in the screen.
<janneke>re gnus: a default changed -- terrible choice, but fixable for me :-(
<civodul>ah, i still haven't upgraded
<civodul>anyway good to know
<civodul>hope i can learn from y'all when i upgrade :-)
<mroh>I think, one of the most important parts missing for the emacs27 upgrade is removing the emacs-seq pkg, as it kills emacs27 completely and in lot's of different places. see and
<janneke>re gnus: message-forward-included-headers needs to be set to nil to get the old behaviour back ("are you really, really sure you don't want to mangle the mail before forwarding?" ;-)
*raghavgururajan just bought Emacs 26.1 manual from FSF store. 😤
*vits-test has it pictures?
<c4droid>Emm, I cannot building my dwm configuration in guix, but it can be build at my arch linux. the default dwm use what version of gcc?
<rekado_>what error do you get?
<c4droid>initializer element is not constant
<rekado_>have you tried a different version of the GCC toolchain?
<c4droid>tried at gcc 10, it can be compiled
<c4droid>Maybe I want to pack the gcc 10 only for compile dwm..
<rekado_>we have a package for gcc-toolchain@10
<rekado_>if GCC 10 is known to work you can install that.
<raghavgururajan>Folks! How do I pass cores and mem arguments to VM script?
<leoprikler>/path/to/ [ARG...]
<nckx>Good morning, unpaid Guix employees.
<c4droid>nckx: Good evening
<nckx>civodul: <reconfigure> Thanks for the heads-up, it's good to announce such state here. Did it work?
<c4droid>Maybe I should waiting the offical update for gcc, that way can solve my dwm configuration problem. :)
<c4droid>I mean.. the gcc depend of dwm
<civodul>howdy, nckx!
<civodul>it did!
<civodul>i'll email guix-sysadmin
<efraim>how long has guix system container supported --expose and --share?
<vits-test>efraim: It was, when i joined there (~January).
<raingloom>hey, i'm finally finishing up my yggdrasil service code and accompanying docs, and i'm wondering if i should export the accessor functions of the yggdrasil-configuration struct?
<raingloom>eg.: gnu/services/ssh.scm does not seem to export anything beyond the basics, but gnu/services/networking.scm exports several accessors.
<civodul>efraim: quite some time!
<brendyyn>raingloom: yggdrasil sounds awesome to have in guix to play with
<raingloom>brendyyn: indeed :) i have some ideas for combining it with offloading to make some ad-hoc distributed computing stuff.
<raingloom>(but that will need lots of fixes for the offloading code in guix)
<civodul>raingloom: take a look at the Guix Build Coordinator that cbaines has been working on
<civodul>it's much more flexible, and probably better suited to some P2P-style of offloading
<raingloom>civodul: OwO thanks!
<brendyyn>raingloom: im not an expert but probably it would be best to export everything. if someone wants one of the functions they will not be able to get it without @@. using them one can take an existing configuration and use it to create a modified version perhaps?
<raingloom>brendyyn: yes, that was my intuition as well, but i assumed there was a reason for not exporting everything in ssh.scm
<civodul>yeah in general, if you'd like to use something that's not exported, it's best to ask
<civodul>usually there's a good reason, sometimes not :-)
<str1ngs>civodul: Hello, how do I turn guix output off ANSI escapes for links? I assume that's a environment variable? eshell really doesn't like them lol
<civodul>str1ngs: ah, you can set INSIDE_EMACS=
<civodul>er, INSIDE_EMACS=t
<civodul>even eshell in Emacs 27 doesn't support links?
<str1ngs>eshell technically is not a terminal just a shell.
<str1ngs>civodul: that works thanks. wonder if there is a way to do termcap for this. I'm more curious than anything else.
<str1ngs>I have the opposite problem in Nomad I need to enable that feature in the terminal. and have nomad handle them.
<civodul>no idea, but if there are better ways to handle this, we should try it
<civodul>you run a terminal emulator inside Nomad/Emacsy?
<str1ngs>yes, M-x terminal. it uses VTE
<civodul>oh neat, didn't know that
*civodul tries
<str1ngs>it's not like comint-mode or anything. maybe some day.
<str1ngs>I have a new nomad, g-golf and emacsy declaration here I need to mail in.
<civodul>in the Nomad currently found in Guix i don't get a prompt
<civodul>maybe i should try from master
<civodul>anyway, interesting!
<vits-test>civodul: devel
<str1ngs>civodul: I have a patched that updates nomad for guix. do M: (%default-shell "/bin/sh")
<str1ngs>that might not work actually
<civodul>i see
<civodul>yeah i get an emacsy/klecl exception
<civodul>i'll try again later :-)
<civodul>Emacsy was (and still is) a great idea
<str1ngs>nomad is a bit of a moving target, will mail the patch today. 0.2.0-alpha-100-g6a565d3 100 commits difference
<str1ngs>i've been rolling some subtitle fixes back into Emacsy which is nice
<vits-test>str1ngs: spare a WebKit for lessers, kind sir?
<str1ngs>vits-test: oh I think I can build that now. I got busy with the pinephone
<civodul>BTW (not related), readline has been flickering somewhat; is it just me?
<str1ngs>the rockpro64 has a recent guix pull now
<str1ngs>civodul: in bash? or guile?
<str1ngs>what terminal do you use? and Xwayland or X11?
<str1ngs>or Wayland I mean.
<str1ngs>civodul: there was a change from a variable binding to parameter for %default-shell .M-: (set! %default-shell "/bin/sh") should work for guix systems.
<str1ngs>or you can add it to ~/.nomad
<str1ngs>vits-test: that's strange php is a libsoup dependency. do you know what it is used for?
<str1ngs>vits-test: I'm assuming libsoup is used by webkitgtk
<vits-test>str1ngs: No, idk what libsoup is. Yes, it's used by webkitgtk.
<str1ngs>vits-test: libsoup is a glib http library. think of it like curl but for gtk applications
<vits-test>.. with PHP!!!
*vits-test gigles.
<str1ngs>yeah that's the odd part, unless it's being used for testing or something.
<str1ngs>is it a native input?
<vits-test>native-input Yes.
<vits-test>to libsoup
<str1ngs>well you can use php as a shebang
<str1ngs>I'd look into it more though. I'd assume it's there for a good reason though.
<vits-test>Why they moved to github?
<str1ngs>who Archlinux?
<vits-test>Yes. I press the "Sources" button.. and found myself on GitHub. How's that?
<vits-test>Cool, is steel there.
<nckx>It's temporary.
<nckx>They are moving to GitLab and for some reason involved a detour through GitBuh.
<nckx>That was a typo but I like it.
<nckx>Found sauce:
<vits-test>Hell. w3m + github = Ж.
<str1ngs>technically Arch uses SVN so all this stuff is probably mirrors
<str1ngs>also there whole SVN has no copyright.
*str1ngs gringes
<nckx>Hence the move to git.
<nckx>(Not because copyright, just because SVN.)
<janneke>svn is great, you don't need a browser to constantly check your internet connection
<str1ngs>last time I checked there PKGBUILD's don't have copyright which is a problem unto it's self.
<str1ngs>also you can checkout only what you want with SVN partly why they have used it.
<str1ngs>Arch is bottom up, not top down like guix.
<nckx>janneke: 😉
<janneke>"a better cvs" -- just when we discovered distributed vc's -- whut?
<vits-test>str1ngs: still, unlike black magic from some other disros, the PKGBUILDs are readable.
<vits-test>ebuilds, srcrpms, .debian..
<str1ngs>civodul: I mailed a patch that updates Guix's Nomad and dependencies. Should make terminal usable now. apologies.
<civodul>str1ngs: readline flickering happens in xterm
<civodul>thanks for the patch!
<str1ngs>vits-test: alpine is more like PKGBUILDS. but with posix shell only hahah
<janneke>str1ngs: a shell/terminal in nomad -- sweet!
<str1ngs>civodul: I've been meaning to send that patch for a couple of days. the duck will never be in a row I think :)
<str1ngs>I need to add support for shell arguments as well.
<str1ngs>janneke: aye, it's not like comint-mode though. but its way faster then ansi-mode atleast lol
<str1ngs>janneke: I call it termianl mode, since there is no underlying <text-buffer> like <text-buffer> has.
<dustyweb>Guix (b/c of its collaboration with Software Heritag) gets a few mentions in the latest episode of FOSS & Crafts
<str1ngs>janneke: also ibuffer-mode \o/
<civodul>hey dustyweb, nice! thanks for the heads-up
*civodul adds to listen-to list :-)
<dustyweb>civodul: :)
<janneke>str1ngs: it's a great start! what is ibuffer-mode?
<str1ngs>janneke: it's like Emacs M-x ibuffer. a buffer to manage buffers
<str1ngs>janneke: I find it translates better to web buffers. switch-to-buffer does not give enough context for web buffer. namely the URI
<janneke>oh my, that looks like it should have replaced C-x C-b?
*janneke feels like crawling out from under a stone
<str1ngs>ibuffer will probably be C-x C-b for Nomad
<str1ngs>alt east till switch-to-buffer completion is better. I don't know how it can improve tough for URI buffers.
<janneke>"why on earth doesn't list-buffers allow regexp marking"
<janneke>apparently, they hid that function somewhere else ;-)
<str1ngs>janneke: BTW did you now switch-to-buffer takes a class?
<str1ngs>switch-to-buffer in emacsy
<janneke>that's weird? you want to switch to an instance?
<str1ngs>(set! buffer-classes (cons* <terminal> buffer-classes)) and it will complete the class. so you can dynamically create buffers with switch-to-buffer interactively
<janneke>ah, it also takes a class -- nice
<str1ngs>say you do M-x: switch-to-buffer RET Name? terminal RET Class? <terminal>
<janneke>so, how's text editing comping along?
<str1ngs>will create a terminal buffer named terminal
<str1ngs>janneke: terrible, I provide only enough to handle minibuffer *scratch* and *Messages*
<str1ngs>janneke: mainly the reasoning is I want to focus on browsing
<str1ngs>no recreate Emacs... not yet anyways.
<str1ngs>and the redraw of <text-buffer>'s is terrible implemented needs to be redone.
<janneke>str1ngs: kind of why i asked -- i tried something like that in my guimax (guile-gi + emacsy), but got stuck there
<str1ngs>janneke: I just sent a patch to update Guix nomad to the latest. you can try *scratch* it does work. but there are limitations.
<janneke>i realised that emacs has some great (or at least "working") abstraction for text buffers, being able to render on different backends
<str1ngs>yes I think it uses custom pixbuf stuff. nomad can do it nicely but it needs much work.
<str1ngs>like a TextView per window
<civodul>janneke: yeah what's up with guimax, it sounded pretty cool? :-)
<str1ngs>janneke is busy with Hurd! :P
<zzappie>janneke: what is guimax?
<civodul>yup :-)
*zzappie my connection is laggish today
<janneke>str1ngs: and some bootstrapping here and there ;-)
<janneke>zzappie: "<janneke> str1ngs: kind of why i asked -- i tried something like that in my guimax (guile-gi + emacsy)"
<str1ngs>janneke: still thinking on text buffers. the emacsy abstraction part is very good IMO. it's redrawing to a toolkit that needs work. My current method is rather aggressive. if the underlying text buffer changes redraw the while thing lol
<peanutbutterandc>Hey guix.... I'm trying to use a guix module inside a custom phase. but I get an error 'no code for module (guix whatever i imported)'
<janneke>civodul: didn't work on it for a year, the pressing need was paused -- meanwhile guile-gi has moved on ...
<peanutbutterandc>I even added a #:modules in (arguments before the actual phase.... but it still doesn't quite work.... how do I fix that?
<janneke>civodul: and also, i got stuck with the text-buffer abstraction -- it needs rethinking
<zzappie>janneke: I just thought it was something I could have been aware of
<janneke>zzappie: ah, not really it's an abandoned toy atm
<zzappie>peanutbutterandc: I am not 100% sure but if you trying to load non-guix modules it might not work. They might not be in build environment.
<peanutbutterandc>zzappie, I'm trying to load a guix module. And I have seen some of the thingies using (ice-9 match) or sth, so I guess non-guix modules are not off limits. The question is: there is probably a proper way for guix modules and another way for non-guix ones. I just don't know where to find the answer....
<zzappie>peanutbutterandc: ah ice-9 is always aroud :)
<vits-test>peanutbutterandc: Maybe paste the definition?
<zzappie>If you are loading guix module then I don't know...
<peanutbutterandc>vits-test, Here: please don't laugh at me for doing that thing with (guix build union)
<peanutbutterandc>I'm just trying to see if I can make icon-themes available without using (propagated-inputs)
<zzappie>peanutbutterandc: mabe try also add (use-module (guix build union))
<peanutbutterandc>zzappie, It was there before, but I removed it to see if that'd fix it
<peanutbutterandc>but I'll try again
<vits-test>cd; mkdir test; cd test; guix -L . imglapse
<vits-test>It's builds now.
<vits-test>Maybe There some modules under the directory You're in?
<vits-test>I didn't changed a definition at all.
<peanutbutterandc>vits-test, wait.... what !? I am doing the same thing essentially.... `guix buile -L . --verbosity=2 imglapse`
<vits-test>peanutbutterandc: try plase the imglapse.scm in some stand-aside dir. Like $HOME/test.
<vits-test>Without any subdirs.
<peanutbutterandc>vits-test, on it, cap'n!
<peanutbutterandc>vits-test, No sir... it does not work for me... not even like that
<peanutbutterandc>this is strange
<zzappie>this is strange
<peanutbutterandc>guix show works
<peanutbutterandc>guix build does not build
<vits-test>The actual build process successfully built icons. Keeps building.
<vits-test>peanutbutterandc: Is the file named imglapse.scm?
<peanutbutterandc>Yes sir
<vits-test>Aha. No syntax error, then.
<vits-test>The build failed.
<vits-test>Sorry, i misunderstanded.
<peanutbutterandc>vits-test, It's all right. For a moment there I thought I might have to spin up a time-machine
*str1ngs .... lets do the time warp.... again.
*str1ngs 🎵🎵
<vits-test>peanutbutterandc: try gnu build union
<vits-test>find|grep build/union ---> ./guix/build/union.scm
<vits-test>instead of (guix build union)
<str1ngs>vits-test: git grep my comrade
<vits-test>str1ngs: Ich bin meow. Ich bin kitty aber nicht :)
<str1ngs>also use grep -R . instead of find| grep :P
<c4droid>Emm, how to change the spec packages gcc version, I'm stuck in here.
<peanutbutterandc>Hmmm...? I'm confused
<vits-test>str1ngs: Only a few super-russians can in grep -R. Ama justa 9-classes-only whelp ;P
<vits-test>peanutbutterandc: 43 line
<str1ngs>vits-test: git grep is better it ignores most of what you don't want to grep
<vits-test>peanutbutterandc: in (arguments
<apteryx>weird: guix deploy -> (system-error "open-file" "~A: ~S" ("No such file or directory" "/gnu/store/id1il9jzq72pcrm3l7zri0d4vxl1zhig-remote-exp.scm")
<peanutbutterandc>vits-test, why gnu build union instead of guix build union?
<peanutbutterandc>vits-test, Because the module is in guix/build/union.scm
<vits-test>Ah yes.
<vits-test>I was badly mistaken again.
<apteryx>full backtrace:
<str1ngs>apteryx: does the remote have the same guix describe?
<apteryx>it's on a guix a week behind
<zzappie>vits-test: peanutbutterandc: won't build for me too. Have no idea what's wrong
<apteryx>it shouldn't matter, no?
<peanutbutterandc>We need an expert on #:modules to see this
<apteryx>perhaps more guile-ssh issues? Seems to occur non deterministically during the copying of store items.
<apteryx>mothacehe: I see where that bash loop guix deploy came from ;-)
<apteryx>bah. I'll reconfigure it locally.
<mothacehe>apteryx: haha, good news is that recent civodul seem to greatly improve the situation :)
<apteryx>perhaps deploy needs a working guile-ssh both ways (on the local *and* remote machine?)
<apteryx>yeah, I'd have those fixes on the local machine, but not the remote I think
<vits-test>peanutbutterandc: for all packages in gnu/packages, there is #:builder stright after #:modules. And it's always a 'trivial-build-system'.
<mothacehe>Cuirass on berlin is catching up, pleasant to watch it build things!
<peanutbutterandc>vits-test, Yes, I realize that... I am just experimenting with union and phases
<zzappie>peanutbutterandc: I have disconneted for a while
<zzappie>peanutbutterandc: try this one out
<ArneBab_>is anyone working on getting gccemacs into guix?
<joshuaBPMan>Hey guix, has anyone had any issues with using org-babel-load-file with emacs 27?
<zzappie>doesn't work but at least imports modules :)
<joshuaBPMan>I'm getting a lisp error, exceeds max-depth
<peanutbutterandc>zzappie, that does kinda' get a little further
<vits-test>How's that: (guix build union) exports union-build. One can't just use the union-build, still. What's gives?
<str1ngs>vits-test: arguments happen in the guix daemon. so you need #:modules which passes it to the daemon
<str1ngs>from my limited understanding
<mroh>joshuaBPMan: try to remove the emacs-seq pkg.
<peanutbutterandc>str1ngs, Even that does not do it
<joshuaBPMan>mroh thanks!
<joshuaBPMan>I'll give that a try.
<peanutbutterandc>str1ngs, could you please take a look at this: ?
<joshuaBPMan>mroh: I don't think I have seq installed.
<joshuaBPMan>it's not installed in package-list-package. Nor have I installed it via guix
<vits-test>str1ngs: BTW, it's a madness.
<peanutbutterandc>vits-test, lol
<joshuaBPMan>mroh maybe I'm wrong....(require 'seq) I guess I have it somewhere.
<peanutbutterandc>I'm trying to have an icon-theme (union-build) inside of out rather than be a (propagated-input)
<str1ngs>joshuaBPMan: you need coreutils
<str1ngs>vits-test: welcome to hell, will you be staying long?
<joshuaBPMan>str1ngs may I ask why?
<str1ngs>joshuaBPMan: coreutils provides seq
<mroh>joshuaBPMan: it's included in emacs, so require works. Do you have any of those `guix refresh -l emacs-seq` installed?
<raghavgururajan>I am facing issue with testing gdm service in a VM. The gdm doesn't start without errors on the screen. How can I debug it?
<joshuaBPMan>mroh I seem to have them installed.
<joshuaBPMan>Building the following 5 packages would ensure 9 dependent packages are rebuilt: jmacs@27.1 emacs-psc-ide@0.1.0-1.7fc2b84 emacs-org-roam@1.2.1 emacs-erc-status-sidebar@0.1-1.ea4189a emacs-flycheck-haskell@0.8-2.32ddff8
<joshuaBPMan>str1ngs I'm also a little surprized that coreutils is not installed in the profile by default
<vits-test>str1ngs, peanutbutterandc: Maybe then make a package object (not-public), and insert it in (inputs?
<str1ngs>joshuaBPMan: you mean the system profile?
<vits-test>Just how it's done with sdl-union.
<joshuaBPMan>str1ngs I guess I meant the user profile. I assume the system profile has it installed by default
<pkill9>coreutils is in %base-packages
<str1ngs>joshuaBPMan: Nothing is installed in the user profile by default that a good thing. Personally I just add it to my manifest
<nckx>seq != emacs-seq
<str1ngs>context is king I guess :)
<joshuaBPMan>str1ngs gotcha. I would like to use manifests soon.
<raghavgururajan>nckx: o/
<raghavgururajan>Billion dollar man!
<nckx>I wish.
<joshuaBPMan>nckx: what do you mean?
<raghavgururajan>I am struggling with testing gdm service in a VM.
<raghavgururajan>The gdm doesn't start without errors on the screen.
<nckx>joshuaBPMan: A wild coreutils appeared and did not seem relevant to the discussion at hand.
<raghavgururajan>I looking for a way to see errors and warnings.
<pkill9>i dunno if what i said was relevant to the discussion either
<pkill9>i should go do something
<joshuaBPMan>nckx: Are you saying that I don't need to install coreutils?? And instead I should focus on removing emacs-seq?
<pkill9>i just saw something and felt the need to say something
<raghavgururajan>Werid thing is that, when I logged in as root and did `herd restart gdm`, I got no gdm found. o.O
<nckx>o7 thank you for your vigilance.
<joshuaBPMan>pkill9 You're welcome to chime in anytime. :)
<nckx>joshuaBPMan: Coreutils are very useful (although I disagree that they should be magically added to empty profiles :-) and things get weird when they aren't there, but they won't help you with the emacs ‘seq’ package which has nothing to do with the seq(1) command.
<nckx>That's all I know.
<str1ngs>joshuaBPMan chimes are okay, gongs are kinda pushing it. :P
<joshuaBPMan>nckx: gotcha. Thanks for clarifying
<joshuaBPMan>str1ngs :)
<joshuaBPMan>unfortunately installing core-utils did not help. So now I'll look at removing emacs-seq
<nckx>raghavgururajan: I don't think I'm your man. I've literally never used GDM. I know it from reputation as an undebuggable mess. Isn't there something matching ‘find /var/log -name "*gdm*" ’?
<raghavgururajan>nckx: It is with any service. When I was tesing gnome-shell using SLiM (as gdm didn't work), I was given "Oh no something went wrong" screen. I was not about to get out of that, so that I do something in terminal.
<civodul>apteryx: guile-ssh is not used on the remote site
<civodul>you're having troubles with 'guix deploy'?
<raghavgururajan>ALT+F1 is changing my host, instead of guest.
<civodul>& mothacehe too? ↑
<nckx>raghavgururajan: I'd check tty12... once you can switch VTs 😉
<nckx>raghavgururajan: You can use ‘sendkey’ or something in the Qemu debug console (I think that's what it's called; this is all from memory). You can get to it from one of the menus.
<raghavgururajan>nckx: I am looking to change terminals inside VM, so that I can debug.
*nckx being super helpful \o/
<raghavgururajan>Ah I see.
<nckx>Let me start a VM here.
<peanutbutterandc>nckx, May I ask you how I'm supposed to use #:modules to use guix modules inside #:phases here: please? I can't seem to make heads or tails of it
<peanutbutterandc>I know (guix build union) is probably not meant to be used like that.... but I'm trying to experiment with making the icon theme available to the package without adding it (the icon theme) as a (propagated input)
<raghavgururajan>Even on full screen cannot change VTs inside VM.
<nckx>raghavgururajan: I don't expect you to. Try C-M-2, sendkey ctrl-alt-f12 <RET>, C-M-1.
<joshuaBPMan>mroh I guess I'm not certain how to follow your advice. package-list-packages shows no "seq." guix package -I emacs-seq...shows nothing. The only thing that I see is guix refresh -l seq....and it lists 5 packages...only 1 of which I have installed "jmacs", which is a custom emacs package...How do I go about "uninstalling" seq ?
<nckx>OK, so my Qemu can't even open the ‘View’ menu.
<raghavgururajan>nckx: Works! Thank you!
<joshuaBPMan>well cd .guix-profile; find . -name *seq* found seq...guix has it installed in the site-lisp directory.
<joshuaBPMan>also cd .guix-profile; grep -r "seq" shows something is pulling it in.
<joshuaBPMan>ok... apparently bug-hunter is pulling in seq.
<nckx>peanutbutterandc: To fix your module import error:
<nckx>Then I get ‘In procedure delete-duplicates: Wrong type argument in position 1 (expecting list): "/gnu/store/7wd4xn6c2mlqsflx1gy9m5znnvfk83i0-breeze-icons-5.70.0"’ which is probably just a thinko or usage error. Didn't look into it.
<peanutbutterandc>nckx, May I please ask why one has to do that? Or could you please guide me to the relevant section in the docs/code/etc. please? (I'd hate to waste your time over something that I myself am not sure will work)
<peanutbutterandc>the #:modules gets the modules ready...?
<peanutbutterandc>what does #:modules do? and #:imported-modules ? o.O
*nckx hopes someone els respons because they are AFK
<vits-test>peanutbutterandc: ?
<vits-test>but it's build -f, as it isn't module.
<vits-test>If it works at all?
*raghavgururajan found one and only error line for gnome-shell: "Exited with status code 1" 😏
<str1ngs>vits-test: does your rockpro64 graphics work out of the box? I get now display at all?
<peanutbutterandc>vits-test, whoa!
<peanutbutterandc>vits-test, You fixed it!
<vits-test>peanutbutterandc: No, it's not module any more.
<peanutbutterandc>vits-test, I am already going the route with and have crossed a few bars and have an issue now.... and I'm debugging it....
<vits-test>str1ngs: I did had troubles with graphics: but only at the same time when system didn't booted at all.
<str1ngs>vits-test: I have a hiDPI monitor I probably need to set a resolution or something.
<peanutbutterandc>vits-test, perhaps I might be able to retain the module too? Thank you for this
<str1ngs>vits-test: I'm working on the nomad-git aarch64 depends regards to substitutes
<vits-test>str1ngs: highest on my is 1440x900 (native is 1680x1050). But some Armbians claim 4K HD, blah-blah.
<str1ngs>vits-test: do you have a fbset command for that?
<vits-test>1680x1050 isn't working now.
<vits-test>fbset -xres 1680 -yres 1050 -vxres 1680 -vyres 1050
<vits-test>But that is console.
<vits-test>sway cant in this res.
<vits-test>peanutbutterandc: I there no way to point the app to icons? The union for data seems like overhead.
<peanutbutterandc>vits-test, I have no idea... I'm just trying to learn stuff as I go along...
<str1ngs>vits-test: awww yeah have graphics
<str1ngs>pro software tip, make sure the HDMI cable is connected!
<vits-test>str1ngs: ah You little...
<vits-test>str1ngs: setfont LatGrkCyr-12x22.psfu.gz
<vits-test>or latarcyrheb-sun32.psfu.gz
<str1ngs>vits-test: can we go back to normal guix kernel?
<vits-test>str1ngs: IDK, but last time i'd tried, it's not boot.
<vits-test>It's worked for nckx on some arm-machine, though.
<vits-test>vagrantc may know more.
<vits-test>* a lot more.
<str1ngs>vits-test: okay I'll stay on 4546c0dd2f58fe9a1b1e78120a33b8c276970f62 for now. at-least for substitutes regards to that kernel
<vits-test>BTW, normal kernel builds looooonger.
<str1ngs>I'll build it on my Ryzen 9 3900X. but the source patching phase takes a good portion too
<str1ngs>err forget the name of the phase where it deblobs
<vits-test>str1ngs: tar cvf, or so. one thread.
<str1ngs>usb-c does not work with this kernel btw. think that's a 5.8 issue though
<vits-test>str1ngs: 5.4?
<str1ngs>Linux puck 5.8.5-gnu #1 SMP PREEMPT 1 aarch64 GNU/Linux
<vits-test>We also have this option on Guix. Though without any ARM-targeted patches, it seems less probable that older kernels will help.
<str1ngs>pine need to focus on mainline not distro specific patches.
*zzappie going afk
<str1ngs>vits-test: pro tip ./profiles/aarch64/bin/guix describe
<str1ngs>vits-test: if I use that guix it always deals with aarch64 on x86_64
<dissoc3>when i do a guix pull i constantly get "guix substitute: error: host name lookup error: Name or service not know. fallback seems to work but im not sure why the substitution host name error is occurring. did it change?
<vits-test>str1ngs: How did You built that profile?
<vits-test>dissoc3: foreign distro?
<vits-test>nscd is running?
<dissoc3>vits-test: im not sure if i understand but im running the complete guix system
<str1ngs> vits-test: guix pull --system=aarch64-linux -p ~/profiles/aarch64 . make sure ~/profiles exists first. you need qemu service though. also not faster if you have a slow machine
<str1ngs>vits-test: use --commit= though so you get the right commit
<vits-test>dissoc3: ping
<efraim>str1ngs: how do you like that CPU? I'm thinking of picking one up at the end of the year
<vits-test>str1ngs: Thanks.
<vits-test>efraim: Ryzen do You mean?
<dissoc3>vits-test: i cant ping it but dig returns A record
<dissoc3>vits-test: sorry. i can indeed ping it
<vits-test>dissoc3: pull again?
<str1ngs>efraim: it's really useful, I mainly got it when I created the qtwebengin package. also now when building webkitgtk. I was going to get 3950X but I think 3900X is cost wise.
<vits-test>dissoc3: That's strange that "name or service" not known for
<str1ngs>efraim: I also need more ram 16G is kinda low for 3900x
<str1ngs>dissoc3: do you have and not ?
<efraim>I wasn't sure between 32G and 64G
<str1ngs>I'd use minimum 32G if you are using all 24 core/threads
<efraim>I figured the zram service could help me out while I decide
<vits-test>str1ngs: dissoc3 are on System, so it should be right by default.
<str1ngs>vits-test: true, missed that thanks.
<efraim>The mobo available locally has 4 slots so it looks like I'll be ordering 16G modules
<vits-test>буржуи :)
<dissoc3>str1ngs: vits-test: for some reason it worked this time. i dont have any clue why it didnt. it was like this for a few days
<vits-test>dissoc3: WiFi?
<str1ngs>efraim: I have a mini atx so only two slots
<dissoc3>vits-test: nope. good network connection. no changes to dns or anything. complete mystery to me
<vits-test>Indeed, then.
<nckx>Critical security fix in gnupg. OK to push to master with 341 dependents?
*str1ngs is a nobody :(
<str1ngs>on the bright side you can blame. but nobody will know who I am!
<str1ngs>blame me*
<str1ngs>vits-test: what are you using for network NM does not work right? right now I'm using static
<vits-test>nckx: wanna help peanutbutterandc? He stil having fun with (union-build. (sorry)
<vits-test>str1ngs: `sudo dhclient`
<str1ngs>what the heck
<vits-test>it's in %base-packages.
<str1ngs>use dhcp service now?
<str1ngs>I can't be bothered making a config
<nckx>Sorry, I'm working. ‘[Announce] [security fix] GnuPG 2.2.23 released’ is just a good interrupt (subject) line 🙂
*str1ngs is lazy
<nckx>peanutbutterandc, vits-test ☝
<vits-test>str1ngs: just run it from command-line.
<vits-test>It's already there.
<peanutbutterandc>vits-test, The hero I didn't deserve. But the one I needed. Very badly. :D lol Thank you
<str1ngs>vits-test: ummm you know kinda hard to do that on a headless system :P
<peanutbutterandc>I have been looking at that for quite some time now
<peanutbutterandc>and I just have had a small new idea to deal with it
<vits-test>str1ngs: What, HDMI cable didn't helped!?
<str1ngs>lets do this the Guix way!
<str1ngs>vits-test: even with HDMI I run it headless
<str1ngs>I would have to switch HDMI inputs. login and run dhcpclient
<str1ngs>no beuno
<str1ngs>prays (service dhcp-client-service-type) is enough!
<dissoc3>im writing a package that uses the cmake build system. it fails and in the log file all I get is a simple error message. is there a build log or something that would give me more info? i ran guix package -K so i can see the build files. something in there to look at?
<peanutbutterandc>vits-test, Okay... so my idea does not seem to work. On to your code then! (:
<vits-test>peanutbutterandc: I didn't managed to make this copied snipped from sdl-union (gnu/packages/sdl.scm) to work as a module, though.
<vits-test>dissoc3: There is a command.
<vits-test>dissoc3: guix build --log-file
<vits-test>Also when You get the error, the guix prints whrere the log is.
<dissoc3>vits-test: is that the same thing as the /var/log/guix/drvs/xx/something.drv.bz2 file? if so that's the file with the basic error command
<dissoc3>error message*
<vits-test>dissoc3: what message it is?
<dissoc3>vits-test: make: *** [Makefile:155: all] Error 2
<vits-test>And nothing before?
<vits-test>dissoc3: btw, paste the definition to So many people will have an opportunity to see.
<dissoc3>i think i figured out how to debug it. i was just confused on how to actually go about it. the Makefile mentioned calls another Makefile in a different directory with a log file with the info i think i need.
<joshuaBPMan>mroh: hey your suggestion seemed to work. Should I file a bug report that emacs-seq is breaking org-bable-load-file somewhere?
<leoprikler>emacs-seq breaking everything is a known bug IIRC
<leoprikler>you can certainly add to that discussion
<joshuaBPMan>leoprikler : Where should I add that discussion? On the official emacs mailing list or on ?
<str1ngs>vits-test: (service dhcp-client-service-type) should work atleast network came up for me and dhclient is running.
<vits-test>str1ngs: BTW, is ntp-service working?
<str1ngs>vits-test: have not tested that yet
<str1ngs>I'm still enjoying the 1970s
<vits-test>str1ngs: aren't it 2013-th?
<str1ngs>I might have dated myself there
<mroh>joshuaBPMan: it's a known issue, we have at least 2 of them. It's also known on debian and nix at least, so no need to report it, imho.
<joshuaBPMan>mroh thanks.
<joshuaBPMan>well my emacs works well, but magit is being slightly annoying...something about transient not being defined.
<joshuaBPMan>I'll get there eventually.
<rekado_>make sure you aren’t mixing any packages from (m)elpa and from Guix
<rekado_>check that there’s no unexpected code in ~/.emacs.d/