<reepca-laptop>So I turned off the wifi on my laptop and plugged in the desktop's connection and sure enough, download got stuck at 100%. So I tried a different cord and a different port on the router, but it's still getting stuck. Bizarre. Anyone know a good free software router to recommend? <rk4>[though i'm unclear on the wifi blob situation :\ - they have atheros SoCs which is promising] <reepca>So I'm trying to work around the weird router issue by using wget to download stuff and then using guix download file:///... but some sources have snippets associated with them, and the builder for the source-fetching derivation wants to do the downloading itself. Any ideas how to work around it now? <leungbk>i've been stuck on this all day; can someone help me out? <rk4>reepca: if i was trying to fight a weird router, i'd try a vpn tunnel and hope that vpn layer could survive the router interference <cnmne>leungbk: on line 102, is the strip phase removing the needed binary? <cnmne>actually I think this is over my head, sorry :( <leungbk>hmm, that does seem to be what's going on <leungbk>guix.texi mentions that strip removes debugging symbols; what other things does it generally do? <cnmne>i honestly don't know. i haven't made a package besides modifying hello.scm, but I'll use yours as an excuse to dig deeper ! <leungbk>there seem to be a lot of strip-foos, now that i look at it <nikola_i>11:23 am <nikola_i> hello. I use nixos right now. I wish to switch to guix. But my main concern is I want to use packages available in nixos channels. I knows I can use nix package manager in guix. But is it possible to use nixos channels and nix packages in guix? <YJSNPI>Hi, everyone. I recently installed GuixSD on my laptop, and I find something strange... (This is just an intro) someone put librime, which is an input method engine's core lib, into gnu/packages/ibus.scm, but I think it shouldn't be there since librime is somewhat universal, it also has a fcitx frontend. So is it possible to move it into another module...? <YJSNPI>By the way, in some cases fcitx actually works better than iBus, but seems nobody here is interested in it đ <YJSNPI>Other IMEs like anthy and libpinyin are also in the same situation <YJSNPI>And... both iBus and fcitx are lacking Qt plugins (at least for Qt4)... <cnmne>leungbk: just got to where you were in the build process :) looking again at the log, it mentions bin/sisugem might be unsupported as a symlink, and later (l 89) it mentions `gem pristine sisu` might fix something that seemed to go wrong in the install <leungbk>i noticed that too, but i wasn't sure how to put it into the build process. i guess i can try using the substitute macro <cnmne>nikola_i: not an expert, but i think the channels option for guix and `guix import` having nixos options give you a lot of options <leungbk>@nikola_i it is indeed possible, and can be figured out from a quick search of the nix documentation, but i don't personally remember the details offhand <leungbk>i'm using stack and ripgrep off nix right now, for example <cnmne>leungbk: I'm looking at other ruby packages from `guix edit ruby` (for inspiration i guess); where <cnmne>what is the substitute macro? <leungbk>(substitute* FILE ((MATCH) DESIRED-REPLACEMENT)) <nikola_i>cnmne: leungbk: okay, I think I will try guixsd first on a vm and see how it goes ***atw` is now known as atw
***rekado_ is now known as rekado
<rekado>nikola_i: you cannot use Nix package definitions directly with Guix. <rekado>leungbk: during the build phase thereâs a warning: âbin/sisugem is a symlinkâ. This looks like it could explain your problems. <cnmne>leungbk: look at the package definition for "ruby-tins"; it looks like they fixed a similar problem <leungbk>all right, i'll try to address that. thanks. <cnmne>sneek: later tell leungbk i got a ruby-sisu.scm file that builds successfully; i deleted the symlink and copied the linked file instead. not good enough at packaging yet to actually use the package, so i don't really know if it's really working, but it builds/installs ! i can paste you later if you haven't already figured it out :) <ngz>Hello. Something is wrong with my Guix installation on a foreign distro. Guix pull reports hundreds of new or updated packages since last update even though it was yesterday. I suspect my .guix-profile/bin/guix is pointing to the wrong place. <ngz>Where should it point to? /var/guix/profiles/per-user/user/current-guix/bin/guix ? Or some other place? <cnmne>i think that's correct; this week i pulled a lot of packages from texlive and r. were those the main offenders? <rekado>ngz: you shouldnât have .guix-profile/bin/guix at all. <rekado>having it means that you have the âguixâ package installed in your Guix profile. <ngz>Indeed. This is because I use emacs-guix on a foreign distro. <ngz>And it is required to have access to all features, AFAIR. <ngz>Also, for some reason, even if I add "$HOME/.config/guix/current/bin" in front of $PATH in my ".profile", "/home/ngz/.guix-profile/bin" happens to be in front of it eventually. <ngz>I assume this is because "guix" package also updated path, through .guix-profile/etc/profile <rekado>ngz: I still think itâs bad advice to tell people to install âguixâ into the default profile. âemacs-guixâ shouldnât require this. <rekado>ngz: you are probably sourcing .guix-profile/etc/profile after you update PATH. <ngz>Indeed. Should I do the opposite? <ngz>So, for the time being, I should remove "guix" from my user's profile. <rekado>I think it is rarely ever a good idea to have âguixâ in the user profile. <rekado>though you could probably just update PATH in the end. <ngz>Of course, but there might be other environment variables set in .guix-profile/.../profile <rekado>you source etc/profile and then you export PATH=$HOME/.config/guix/current/bin:$PATH <ngz>What I mean is if "./guix-profile/.../profile" sets FOO, and I happen to update FOO in my "~/.profile", it could lead to the same problem. So maybe a good hygiene would be to first source "./guix-profile/.../profile", then do other stuff in ".profile". <ngz>I.e., not limit this protection to PATH. <rekado>yes, thatâs what I wrote: first source etc/profile, then update PATH. <rekado>if you have other variables you may want to do the same. <ngz>OK. I misunderstood you, then :) <ngz>For the record, as expected, removing "guix" from my profile fixes my update problem. <ngz>Thank you for the help, BTW. <ngz>While I'm at it, I have another pretty annoying issue with Gnome (from foreign distro, i.e., Debian testing) and Guix. I have to "unset GI_TYPELIB_PATH; unset XDG_DATA_DIRS" in my ".profile", or I cannot get past GDM, i.e, the session crashes and I'm brought back to it. <ngz>I don't know how to get a backtrace from this crash to help debugging. <pkill9>hmm, guix said it would only download something small but it's also downloading qt4 which is a graft <ngz>It seems that texlive-texmf is failing to build. <ngz>(on a recent guix pull) <rekado>ngz: hmm, this is not supposed to happen. <ngz>rekado: /gnu/store/19nmhi9l54x3m6pzs2ng31xzjb46xdk1-texlive-texmf-20180414.drv <rekado>in the meantime you may have better luck with texlive-base and the other texlive-* packages. <ngz>This is an update. For now, I use --do-not-upgrade=texlive <tune>looks like spacefm fails to build still. I commented out a month or two ago <rekado>did you report the failure on bug-guix? <ngz>I.e., you only need to add "#include <sys/sysmacros.h>" to "main.c" <ngz>Hmm nevermind, this patch is already applied in master. <ngz>Then I assume it hasn't been built yet. <ngz>Patch was introduced yesterday, but last compilation, at least on hydra dates back from the day before.. <ngz>tune: did you try to build it with a recent guix pull ? <civodul>rekado: i'm looking at current-guix but that turned out to be a rabbit hole <tune>recent but not brand new pull <tune>I'll try again later maybe <ngz>Hmmm I cannot download any add-on from Kodi: "Could not connect to repository". Is it expected? <Blackbeard>building /gnu/store/vdlf9w6pn9shx566n3ghdmy78sz85830-icecat-60.3.0-gnu1.drv... <Blackbeard>note: build failure may have been caused by lack of free disk space <Blackbeard>why am I getting a lock of free disk space error :/ <ngz>Blackbeard: what does "df -lh" returns? <Blackbeard>basically it has a size of 2.0 gb using 111M and aviable 1.9G using onl 6% <Blackbeard>/dev/mapper/luks-6d759619-2acf-452c-b58f-899155873a05 226G 32G 183G 15% / <Blackbeard>size 226G using 32G available 183G so only using 15% <Blackbeard>and other things work, like audacity or racket or vlc <Blackbeard>but icecat I can't install, neitther was I able to update it <ngz>Hmm I don't know then. <leungbk>is there any way of halting a %standard-phase? <sneek>leungbk, you have 1 message. <sneek>leungbk, cnmne says: i got a ruby-sisu.scm file that builds successfully; i deleted the symlink and copied the linked file instead. not good enough at packaging yet to actually use the package, so i don't really know if it's really working, but it builds/installs ! i can paste you later if you haven't already figured it out :) <leungbk>i have this right now, and am trying to run `gem pristine` on something that the computer asserts is not there: <eubarbosa>whoever recommend SICP as Scheme resource, Ive to thanks, what a superb book! <rekado>ngz: texlive-texmf should be okay now. <rekado>leungbk: you can replace phases with (modify-phases %standard-phases (replace 'name-of-the-phase (lambda âŠ))) <ngz>rekado: Thank you (for texlive). <rekado>eubarbosa: would you like to send a patch? <roptat>one of the packages I want depends on boringssl, which I can build, but it doesn't have an install target <roptat>even installing it by hand fails because the libraries don't have the right runpath <leungbk>@rekado thanks, i had tried that last night as well. i think i'll have to study the innards of ruby-build-system.scm to make that work. <eubarbosa>would be great if the Emacs guix package have a `guix-find-function` `guix-find-library`... <apteryx>eubarbosa: I've only watched the SICP videos <apteryx>eubarbosa: that would require building and distributing a cache of those things on the build server. <ArneBab>I get problems with cups trying to write its config file: E [19/Jan/2019:22:00:23 +0100] Unable to create "/gnu/store/qv49wv0b5s2jmn4ln3r14v4bks18wqp0-cupsd.conf.N": Read-only file system <ArneBab>my problem use-case is that when I try to print, the job fails with "Processing - "Color Manager: no profiles specified in PPD"" <ArneBab>do you have an idea how I can get that working? <civodul>ArneBab: i don't know but you could try asking on help-guix <ArneBab>civodul: did that now â thank you! <eubarbosa>apteryx:I am a total noobie so I need more than those videos! hehe <eubarbosa>apteryx: so that is achievable, just not easy to implement. Soon! <sneek>I could be wrong, but apteryx: so that is achievable, just not easy to implement. Soon!