IRC channel logs


back to list of logs

<gnutec>Install guix with gnome (XFCE had only 10MB) is a good idea. Comes with Browser, PDF reader, Image view e many other basic app. I just need install emacs, transmission, gnumeric, abiword and VLC (to open MP4, MKV and others).
<leoprikler>instead of vlc, you can also add some gst-plugins and totem will eat those formats as well :)
<peanutbutterandc>Hey there, I have been getting these build failures for a custom package and I would really appreciate some help in reviewing them. I am new to guix and packaging in general
<peanutbutterandc>Here is my package definition as of right now:
<peanutbutterandc>and here is the build failure log:
<peanutbutterandc>I am rather surprised by "Can't locate IO/ in @INC (you may need to install the IO::String module)". It is defined as a (propagated input) in the chordpro package definition
<peanutbutterandc>Also, there's this line about requiring Test::CPAN::Changes, and it is this one: and yet, guix import cpan Test::CPAN::Changes gives me "guix import: error: failed to download meta-data for package 'Test::CPAN::Changes'"
<nckx>peanutbutterandc: On that CPAN page you can see the actual package name is CPAN::Changes, already in Guix as perl-cpan-changes.
<peanutbutterandc>nckx, Thank you; You've become my unofficial supervisor in my journey to guix mastery. Haha
<nckx>As for IO::String, all I can say is add it as an input to perl-font-ttf itself, since lib/Font/TTF/ actually tries to ‘use IO::String’ (i.e. it's not a transient dependency).
<peanutbutterandc>Now, another question is that, since this package is only needed to run tests, where should this input be declared? In the same block as the other inputs? (propagated-input?)
<nckx>Test inputs (like all build-time-only inputs) are usually native-inputs.
<nckx>Usually = always unless good reason 🙂
<peanutbutterandc>nckx, I see. Please forgive me but what is a transient dependency?
<nckx>Sorry, my typo. I meant to say transitive.
<peanutbutterandc>nckx, The wikipedia page seems a bit too complex for me. Could you please explain it to me in simple terms?
<nckx>I only mentioned it since you talked about propagation above. It *would* have been possible that Font::TTF didn't contain a single use of IO::String, but that it was use'ing another module that needed it. *Had* that been the case, the error message would probably have been the same, but the fix different: instead of adding IO::String to Font::TTF's inputs, you'd add it to <that module>'s.
<nckx>I hope that makes sense, I'm tired and about to go to bed.
<peanutbutterandc>nckx, I will try to re-read and make sense of it. Thank you very much. Good Night (from GMT+5:45)
<nckx>Actually, why do you think that ‘[IO::String] is defined as a (propagated input) in the chordpro package definition’ would solve ‘Can't locate IO/ in @INC’ for the perl-font-ttf package?
<nckx>perl-font-ttf never gets chordpro as input.
<nckx>Were you perhaps confused?
<nckx>One of us is 🙂
<peanutbutterandc>nckx, I don't really know. I'm using the simpleton assumption that all interpreted libraries go under propagated input. And perhaps I chose the wrong package to declare the propagated inputs in. This is my first attempt at this, really. I'm still a huge n00b.
<nckx>(UTC+2 here)
<nckx>peanutbutterandc: Oh, no, propagation doesn't mean ‘interpreted’ at all (although there's a certain correlation).
<peanutbutterandc>nckx, Yes, it seems I have a lot to learn... I will try to figure things out from here based on the information you have given me. I'd hate to keep you awake for now. Perhaps if I haven't figured this out later, I could ask you again.
*nckx → 😴
<thomassgn>trying to run sway, got error: "[backend/drm/atomic.c:57] LVDS-1: Atomic commit failed (modeset): permission denied" Anyone familiar with this?
<ison[m]>thomassgn: I just know one way sway differs from i3 is it includes display and input properties in the config file. So maybe you just need to configure them. Does running "swaymsg -t get_outputs" list LVDS-1?
<peanutbutterandc>If anyone here has the time, I would be very grateful if they could review my (first ever) package definition: I would, in the very near future, like to be able to be competent enough to add this to the gnu distribution itself. Thank you (:
<vagrantc>have you run "guix lint" on it?
<peanutbutterandc>vagrantc, No, not yet. I don't think I have read up on `guix lint`. I will look it up right away. Thank you very much
<vagrantc>it checks for common issues in packages
<oriansj>is there any particular reason why guix pull is resulting in a full bootstrap build of guile? Even though bootstrapped gcc already exists on the system?
<leoprikler>probably so that user guix can function without depending on the underlying system
<leoprikler>might be helpful on foreign systems
<oriansj>leoprikler: that would be fine but bootstrapped gcc is part of /gnu/store/'s contents
<oriansj>aka, you already did all the work to bootstrap guix on that system; so why are you doing it all over again for guix pull?
<leoprikler>oriansj: perhaps one of the packages changed in some way and needs rebuilding?
<leoprikler>I'm pretty sure it won't build packages again that it already has
<leoprikler>that said, `guix gc` might inadvertently delete some
<leoprikler>especially with grafts
<ng0>if I asked this before, I'm sorry. Does linux-libre have "net/ieee80211_radiotap.h" ?
<ng0>I'm trying to gather information for
<leoprikler>doesn't seem like it
<leoprikler>(did `guix build linux-libre --source` and then opened the tar.xz with file-roller)
<leoprikler>ended in the wrong dir, include/net does have ieee80211_radiotap.h
<ng0>i supposed this is 4.x or whatever Linux is on now? 5.x?
<leoprikler>i did it with 5.x
<ng0>ah. can you publish the file somewhere so that I can diff it to linux and to the original netbsd source? or does linux-libre have a full source now?
<leoprikler>just do `guix build linux-libre[@VERSION] --source`
*ng0 writes this from NetBSD 9
<ng0>I'll search the source
<leoprikler>do you have a Guix somewhere?
<ng0>not at the moment
<ng0>oh too early
<ng0>i could just grab the release tarball
<ng0>thanks :)
<reepca>sanity check: how exactly can an umounted filesystem that hdparm reports as having "readonly = 0 (off)" be considered "read-only" by fsck?