IRC channel logs

2016-04-12.log

back to list of logs

<lfam>Never copy and paste command lines from the net! :)
<suitsmeveryfine>lfam: you haven't played the game? It's a classic
<suitsmeveryfine>good point
<lfam>There are so many classic games, and only so many years :)
<suitsmeveryfine>In OpenTTD you begin in 1950 and you can continue far into the future.
<suitsmeveryfine>So you save a lot of time ;)
<lfam>Lol
<lfam>I began in 1986, so I saved even more time ;)
<lfam>I've reconfigured my GNOME system with paroneayea's xscreensaver. Anybody know how to invoke it on GNOME? Or, how to invoke any screen locker on GNOME?
<suitsmeveryfine>lfam: I would try it in Xfce instead
<lfam>Hm, another environment I'm totally unfamiliar with ;)
<suitsmeveryfine>Really?
<lfam>Yeah, when I discovered tiling window managers I was very glad I'd never again need to fiddle with icon and window placement
<suitsmeveryfine>It's really cool to generate graphs like this
<suitsmeveryfine>lfam: Xfce has fewer bugs in GuixSD than GNOME
<lfam>Check the options to `guix graph`. The default graph is not very complete. You can really overwhelm yourself with the complexity of software systems with some of those options
<suitsmeveryfine>Xfce is a DE where Xscreensaver belongs while GNOME 3 isn't
<lfam>Okay, I'll give it a try
<lfam>On XFCE, how do I invoke a screen locker?
<suitsmeveryfine>Check the system settings
<suitsmeveryfine>You'll find a list of key bindings
<suitsmeveryfine>I'm not in Xfce currently so I can't look
<lfam>Okay, I'll dig in
<mthl>lfam: check xscreensaver-command man page
<sippican>greetings all. Longtime Open *nix, pondering going free but wondering if graphic performance is still going to be a significant trade-off. Son games on the machine. Not happy with the blobberfest that is OpenSuse, even the kernel/tcp tuning/timeouts were utterly ridiculous,.
<mthl>lfam: xscreensaver commands are not bound to any desktop environment interface.
<lfam>sippican: Welcome!
<lfam>mthl: Thanks for the advice
<mthl>lfam: np
<sippican>lfam: happy to be here.
<Jookia>Does guix lint check code formatting?
<suitsmeveryfine>Jookia: yes
<suitsmeveryfine>It just complained on me mixing tabs and spaces
<Jookia>ok
<mthl>Jookia: it checks about tabs but not the number of spaces
<suitsmeveryfine>Jookia: I suggest that you edit in Emacs to get indentation and all that right
<Jookia>i don't use emacs
<sippican>ruh roh
<sippican>;-P
<suitsmeveryfine>You might want to try it just for Guix code?
<Jookia>no
<suitsmeveryfine>ok :)
<suitsmeveryfine>Anyway, you get automatic indentation then.
<lfam>I use Vim with the paredit plugin
<lfam>It's another option, although it's probably not as good as the functionality in Emacs
<Jookia>do i have to use emacs to work on guix?
<suitsmeveryfine>In Emacs I press C-M-a + C-M-q and the correct indentation is automatically set.
<mthl>Jookia: no you don't, but before submitting a patch launching emacs for fixing the indentation could help you and the reviewers
<Jookia>mthl: that's okay, i don't plan on submitting patches
<mthl>Jookia: So no need to use it then
<suitsmeveryfine>Jookia: or you can send the patch to somebody who pastes it in Emacs for you :)
<reepca>So I'm getting the error "error: cannot bind to socket `/var/guix/daemon-socket/socket': Address already in use" when trying to run guix-daemon on Ubuntu 15.10, and ps aux | grep "guix" yields nothing except something I suspect is unrelated in ~/.guix/bin/
<reepca>And whatever currently is using it isn't talking to the guix client used via emacs ("connection refused")
<lfam>reepca: I haven't seen a ~/.guix directory before. How did you install Guix?
<reepca>errr my bad I meant .guix-profile
<lfam>And what is returned by `ps aux | grep guix-daemon`
<lfam>?
<reepca>nothing
<lfam>Also, how did you install Guix?
<reepca>binary tarball I think, did it awhile ago, haven't used it too much since, trying to get it working now
<lfam>Are you trying to start the daemon as an unprivileged user? I get that message when trying that. You should run it as root
<lfam>reepca ^
<reepca>sudo just says "guix-daemon: command not found"
<reepca>but it finds it when I don't try it as root
<lfam>For debugging purposes, try running it with the full path, as root: /var/guix/profiles/per-user/root/guix-profile/bin/guix-daemon
<lfam>In general, `sudo foo` won't give you the same environment as when logging in as root. You'd need `sudo -i` to get the login shell. It's relevant because your $PATH should be set when logging in
<lfam>That's `sudo -i foo`
<lfam>By the way, it's a daemon, so expect it to do nothing when you run it
<reepca>huh, alright. Well using the full path did the trick, thanks, successfully ran (guix-command "package" "--upgrade") :~)
<reepca>Now I thought I remembered configuring it to start the daemon on startup, so next on my to-do list is figure out why that didn't happen / if I forgot to do it in the first place
<lfam>Cool. Since you say you haven't used Guix in a while, be sure to update to the latest version. There are of course many security and bug fixes.
<lfam>And remember that each user has their own version of Guix, so at the minimum you need to udpate with `guix pull` for both root and your regular user
<lfam>reepca: I've filed a bug report regarding the misleading error message
<lfam>ACTION afk
<reepca>Huh, okay, so I've got guix-daemon.service in /etc/systemd/system/ but it isn't getting run when I start up the system. Or at least, it hasn't in the past.
<floating_dinghy>is recovering from "guix substitute: error: TLS error in procefure 'handshake': The TLS connection was non-properly terminated." by executing "guix system init /mnt/etc/config.scm /mnt" again ok?
<floating_dinghy>I hope I'm not DDOSing you guys... :D
<iyzsong>yes, I think it's always safe to re-run any guix commands.
<floating_dinghy>cool. I'm going through the install and this is the third time a download failed like that...I was worried it restarted the whole deployment again...
<iyzsong>the deployment only occurs after download all the files (to store). but if you reboot, those files will be downloaded again.
<floating_dinghy>aww... I did reboot once :/ but I had a different reason for that... I didn't label my partition while setting config.scm as "" thinking it would work using the /dev/sda1 device... well, it didn't find kernel. I didn't think it will redownload :/
<paroneayea>sneek: later tell lfam I tested xscreensaver with xfce
<sneek>Okay.
<floating_dinghy>I should probably just have modified the grub config manually... but I thought the installer will use the packages... rats :D
<floating_dinghy>is there a way to restart the process without redownloading ? for future mishaps...
<iyzsong>floating_dinghy: um, you can modify the "root=.." argument in the grub menu (press 'e'). 'guix system reconfigure' is much easier than the 'system init' process we have currently.
<floating_dinghy>iyzsong: ok. then that's what I'll do if I'll mess up grub again.
<floating_dinghy>iyzsong: thanks :)
<iyzsong>I'm not sure if it's possible to avoid redownload. I think the state for packages avaliable are in an sqlite database, and it's fresh in the live usb...
<iyzsong>floating_dinghy: welcome :-)
<floating_dinghy>what are the ram constraints on the system btw? I tried nixos but saw the livecd with it's kde wm taking 1gb of ram... I'm messing around with a baytrail hdmi stick that came with 2gb so it's kinda tight on this little atom.
<floating_dinghy>like, I understand there's a store database always running in the background? or was it kde alone taking 1gb? 0_o
<iyzsong>it depends.. you can choice not to use heavy desktops. also nixos have cli only minimal install cd. I guess guix require more ram than nix, since it's a whole language vm (guile). but I think 256M is minimal.
<floating_dinghy>good. I use i3 and vi (might move to zile... emacs a bit too much for me) so it's not a problem so long as the store overhead doesn't start swapping on the emmc.
<lfam>reepca: Regarding starting the daemon at startup, did you do `systemctl enable guix-daemon`? That's how you register services to run at boot on systemd-based systems
<sneek>Welcome back lfam, you have 1 message.
<sneek>lfam, paroneayea says: I tested xscreensaver with xfce
<lfam>reepca: The output of `systemctl status guix-daemon` should have useful info in the "Loaded:" line
<reepca>lfam: Oh, okay, thanks. Did it and systemctl now says it is enabled.
***anon is now known as Guest70097
<rekado>davexunit: you may have to specify an elements path.
<rekado>I have a project file containing "elements-dir lib/footprints" with custom footprints.
<rekado>you may also need to specify the default footprint directory with "-d ~/.guix-profile/share/pcb/newlib" or similar
<rekado>we should patch these tools to look at the correct paths by default.
<lfam>Did anyone see my question in the btrfs-progs thread about updating util-linux? Should it be reverted (due to too many rebuilds — 729)?
<lfam>Feel free to revert if necessary
<ng0>what's the current dualboot status, has someone written about this somewhere? I plan on testing/investigating a system of cryptdisk section by key choice.
<davexunit>rekado: I tried that but it didn't work :(
<davexunit>not sure what's wrong.
<phant0mas>davexunit: did you have any progress with the avr toolchain?
<davexunit>phant0mas: no I stopped when you told me to wait.
<phant0mas>well we still can't link :-(
<davexunit>could you share your changes?
<phant0mas>I have an idea though, will try it out
<davexunit>I will try again
<phant0mas>yes
<phant0mas>I will send you the changes
<davexunit>thanks
<phant0mas> http://paste.lisp.org/display/313284
<phant0mas>here you are davexunit
<phant0mas>if you find anything tell me :-)
<phant0mas>we are now closer than ever
<phant0mas>to get a working toolchain
<janneke>phant0mas: you saw my guile.exe link? could my mingw fixes be related?
<phant0mas>janneke: I was just about to mail you :P
<phant0mas>yes, it could be
<phant0mas>it's similar problems, just different targets
<janneke>i have cleaned-up a lot, just pushed to wip-hurd+mingw
<jordila>sorry
<jordila>just to say that http://audio-video.gnu.org/video/misc/2015-01__GNU_Guix__The_Emacs_of_Distros.webm
<jordila>video link or something is broken... :-(
<jordila>:-)
<jordila>thanks !
<ng0>how long does it normally take for someone on bug-womb to reply? I know womb is no focus project, but at least some "read it, will take time to reply" is great to see at least someone noticed it.
<taylan>jordila: it works for me
<davexunit>works for me, too
<davexunit>phant0mas: thanks, trying out your changes.
<davexunit>last night I got some prototyping pcbs in the mail, the ones that just have a ton of holes, and soldered the ps2 analog sticks onto them with jumper wires to make them breadboard compatible.
<phant0mas>davexunit: I made it
<phant0mas>it links
<davexunit>yay!
<davexunit>what did you add!
<davexunit>?*
<jordila>ah... i'll review ^^ thanks
<phant0mas> (search-paths
<phant0mas> (list (search-path-specification
<phant0mas> (variable "CROSS_CPATH")
<phant0mas> (files '("avr/include")))
<phant0mas> (search-path-specification
<phant0mas> (variable "CROSS_LIBRARY_PATH")
<phant0mas> (files '("avr/lib")))))
<phant0mas>to avr-gcc
<phant0mas>and before building a avr binary export CROSS_CPATH=/home/manolis/.guix-profile/avr/include and export CROSS_LIBRARY_PATH=/home/manolis/.guix-profile/avr/lib/
<phant0mas>:-D
<davexunit>phant0mas: oh damn I totally had those in an old avr-gcc recipe of mine :)
<davexunit>thanks!
<davexunit>phant0mas: do we still need the i686 libc input to avr-libc?
<phant0mas>yes
<phant0mas>we need to find a way to remove that
<davexunit>phant0mas: okay
<davexunit>phant0mas: why?
<phant0mas>well I can't undestand why we need it :-)
<davexunit>:)
<phant0mas>ACTION feels relieved that the avr toolchain works now
<davexunit>:)
<davexunit>me too
<davexunit>I'm excited
<davexunit>phant0mas: any particular reason why you removed the --disable-libssp flag from the avr-gcc build?
<davexunit>the official avr-gcc docs include that flag
<phant0mas>non that I can remember
<phant0mas>add it
<phant0mas>probably I removed by mistake
<phant0mas>s/removed/removed it
<phant0mas>rekado: your toolchain is next :-)
<phant0mas>davexunit rekado maybe we should add 'fix-genmultilib to one of the existing phases
<phant0mas>because this problem exists in both the avr and arm-none toolchains
<davexunit>phant0mas: yeah it could benefit all gcc cross compilers
<davexunit>phant0mas: hmm my avr-libc doesn't build...
<davexunit>ACTION compares
<phant0mas>davexunit: http://paste.lisp.org/display/313285
<davexunit>phant0mas: btw, instead of doing `(,@(foo)), you can just write foo
<davexunit>in this case: (arguments (substitute-keyword-arguments ...))
<phant0mas>davexunit: didn't know that
<phant0mas>thanks :-)
<davexunit>it's like putting an equation into simplest terms :)
<phant0mas>ACTION still learning scheme :-)
<phant0mas>sometimes I think that the greek university system is really broken
<phant0mas>I was never taught things that are supposed to be cs101
<davexunit>my university had similar issues
<davexunit>really depends on the school
<davexunit>too much Java in mine.
<htgoebel1>I have a source package, which contains two parts, eg. libblabla and libbarfoo. I want to build two independent packages but use the same source. How can I reuse the source declaration?
<davexunit>htgoebel1: it's just an object like anything else.
<davexunit>give it a name i.e. store it in a variable
<davexunit>refer to it where needed
<htgoebel1>davexunit: Me scheme knowledge is still quite limited. I tried (define my-lalala-source (source ....)) and then (define-public libblabla (package ... (my-lalala-source) ...)
<htgoebel1>This fails: "unbound variable source".
<davexunit>htgoebel1: 'source' is syntax for packages
<davexunit>'origin' is the constructor
<davexunit>(define foo (origin ...))
<htgoebel1>devxunit: Thanks. It took me some tries to get the details right, but this did the trick.
<phant0mas>davexunit: does it work now?
<davexunit>phant0mas: no
<phant0mas>what does it say?
<davexunit>configure: error: C compiler cannot create executables
<phant0mas>did you use the patch from here? http://paste.lisp.org/display/313285
<davexunit>phant0mas: yes, I am using pretty much the exact same thing
<davexunit>I did a very small refactoring of the avr-gcc recipe
<davexunit>but it's equivalent
<phant0mas>can you show me?
<davexunit>phant0mas: http://paste.lisp.org/display/313288
<davexunit>I tried the build with and without the --disable-libssp flag, so that's not the problem.
<phant0mas>you are right it's essentially the same
<phant0mas>hm..
<phant0mas>let me try it
<davexunit>what the... it is building right now
<davexunit>wait nvm, it's the cross-libc
<davexunit>which I thought I built...
<davexunit>phant0mas: sorry, it's actually the i686 cross-libc that fails
<davexunit>I misread a build log
<phant0mas>aaa :P
<phant0mas>phew.. got worried there for a sec
<davexunit>so there's some nondeterminism
<davexunit>...
<davexunit>grrrr
<davexunit>no idea how to fix this.
<phant0mas>hm..we need someone else to try it as well
<phant0mas>anyone? :-D
<htgoebel1>Has guile something like hashes or associative arrays? What I've found in the manual is terrible complicated.
<ijp>there are hashtables, but the default api is more complicated than necessary
<davexunit>htgoebel1: use an alist
<davexunit>'(("key1" . 1) ("key2" . 2))
<davexunit>mutable hash tables should almost certainly be avoided in Guix code.
<ijp>but for future reference, the relevant manual section is 6.7.14
<htgoebel1>davexunit: Thanks.
<davexunit>yw
<optikalmouse>I've decided to force guix into our company. I had to create a VM + scripts to be able to setup and tie all our microservices together so we can have per-branch/per-fork environments to QA things easily
<optikalmouse>and I need a web interface to manage the box so I'm going with Guile Scheme which opens up a nice path for guix ;)
<vekth>Plop
<vekth>by wish manual do i need to begining? Guix Manual or GuixSD Manual?
<davexunit>vekth: they are the same manual
<davexunit>to use Guix on your current GNU/Linux system, you do not need to read the GuixSD section
<vekth>oh yeah. I didn't see that it was the same but different part ^^'
<vekth>i want to install, (try to re-install), GuixSD
<davexunit>then reading over the GuixSD section would be good
<davexunit>some sections will be more immediately interesting than others
<davexunit>for example, you don't need to know all the services to make a basic config to install
<rekado>optikalmouse: cool. Do you plan on sharing the web interface with Guix upstream?
<rekado>ACTION leaves again
<vekth>davexunit, okay. I will try to install it, boot on it and after finish to read the configuration file wiki : )
<davexunit>vekth: my recommendation to you is to use the most minimal configuration that could possibly work for you
<vekth>okay
<davexunit>once you have the system up and running, it becomes very easy to experiment with changes
<davexunit>because if a change goes wrong, you can just reboot the last known working config. a luxury you don't have when installing for the first time.
<optikalmouse>rekado: the web interface is just for managing git repos and some database updates that are specific to our work environment, not really directly related to guix
<vekth>Another question, i saw that lvm is not available for the installation process, but i saw the lvm package, i will be able to use my lvm partition after the installation?
<rekado>vekth: yes
<rekado>optikalmouse: ah, ok.
<optikalmouse>rekado: the main use is as a stepping stone to scheme
<optikalmouse>we're using docker everywhere and the complexity is kinda nuts with all the extra tooling surrounding docker
<davexunit>I hear ya.
<davexunit>ACTION watches work environment grow to be a gelatinous mass of a Docker, Chef, Semaphore, AWS scripts
<davexunit>s/of a/of/
<paroneayea>davexunit: heh
<vekth>somebody knows when we'll be able to install on a lvm device?
<davexunit>no.
<davexunit>it depends on when someone who is interested in this writes the code to do it.
<vekth>okay.
<vekth>i will maybe change my file-system scheme then : /
<davexunit>I have seen numerous people talk about our lack of LVM support
<davexunit>but talk doesn't make it happen
<vekth>well, it could be cool. 'Cause some people can be afraid about needed a bigger / partition than at the installation process.
<davexunit>yes, it would be cool.
<davexunit>but does anyone want to step up and actually do it?
<datagrok>davexunit: what's the current state? the patch posted to the discussion list mid-2015?
<davexunit>I don't know.
<davexunit>I don't really follow it closely.
<davexunit>I just want people to understand that they shouldn't wait around for something to happen. we need people to do things.
<vekth>i don't think i have enough of knowledges to help, or even try, GuixSD : /
<davexunit>I personally don't use LVM so it's not a feature I'm going to spend time on.
<davexunit>vekth: you could always try GuixSD in a VM
<vekth>yep
<davexunit>or just try Guix on your current system
<datagrok>davexunit: totally understand. i hope to contribute, but i'm still getting started with scheme. feels like a long way to go :)
<ng0>davexunit: i agree with the last sentence. Part of why I am now officially stepping into gentoo packaging in addition to everything else i do.. nobody else does it. you have to when you want to see something get done.
<davexunit>right, exactly.
<davexunit>and we're here to help people along the way when they try to do something new.
<ng0>today I ran a search on something cryptdisk related, a guide or experience/experiment, and even ahmia.fi or google or others found nothing.. so the internet told me again: just do it, document it and let others find it.
<davexunit>I'm in completely unfamiliar territory with this AVR toolchain thing, but it wouldn't get done if I didn't try.
<ng0>i had next to zero usable results and in addition I found a malware/virus/exploit download DB xD
<davexunit>and once I tried, some people came to help.
<davexunit>ng0: heh, well that's something ;)
<datagrok>vekth: fwiw here's the LVM patch and discussion I mentioned: http://lists.nongnu.org/archive/html/guix-devel/2015-04/msg00317.html
<janneke>why do i suddenly get guix/ui.scm:252:7: source expression failed to match any pattern in form (report-error (_ "failed to load '~a': ~a~%") file (strerror err))
<vekth>datagrok, it is a topic created one year ago?
<datagrok>vekth: that conversation thread was started April 2015 but I don't know if there are other discussions that are older or more recent
<datagrok>vekth: that particular link is where I'm starting, because it contains a patch that I can study
<janneke>i'm using guile-next :-( i thought guix environment guix --ad-hoc emacs git installs guile-2.0? hmm
<davexunit>janneke: guile-2.0 is part of that environment, yes.
<davexunit>one possibility is ABI breakage, which will mean that you should rebuild
<janneke>davexunit: it used to work
<janneke>now my guile-next remains after guix environment guix --ad-hoc git guix emacs
<janneke>i am rebuilding world all the time, could have tricky setup somewhere...
<vekth>datagrok, okay : )à
<davexunit>janneke: what do you mean "guile-next remains"
<davexunit>the command you sent doesn't create an isolated environment
<davexunit>it augments your environment with those tings
<davexunit>things*
<davexunit>and 'guix environment' doesn't install anything to your profile, so it's not going to install guile-2.0
<janneke>davexunit: i have guile-next in my profile, then run the guix environment guix .. command
<janneke>and i expect `guile' to be guile-2.0 then...
<janneke>apparently i'm mistaken? i think that used to work...
<davexunit>yes, it should be
<davexunit>inspect PATH and see what's going on
<janneke>duh... :-)
<janneke>PATH still starts with /home/janneke/.guix-profile/bin
<davexunit>janneke: your .bashrc is broken
<davexunit>most likely
<davexunit>it probably has something like:
<davexunit>export PATH=...
<janneke>arg, yes that could be
<janneke>oops, thanks!
<davexunit>.bashrc should pretty much *never* set environment variables, because it is used for non-login interactive shells
<davexunit>.bash_profile is where environment variables should be set
<htgoebel1>I'd like to contribute a first package of python-django. Since django itself is a framework with many many packages, I'd suggest putting these into either "gnu packages django" or "gnu packages python django". Which one should I use?
<davexunit>htgoebel1: (gnu packages django)
<davexunit>we don't make additional subdirectories for package modules
<ng0>python is already long enough
<ng0>python.scm i mean
<davexunit>the size of the file is unimportant
<ng0>s/enough//
<davexunit>I think a new module makes sense because django is this huge collection of stuff
<ng0>pointing out that it is long. size, yes i agree
<htgoebel1>ack. gnu packages django
<keverets>hmmm; "guix package -I" shows guix 0.10.0-0.7611 installed, but "guix --version" says 0.9.0
<keverets>ls -l ~/.guix_profile/bin/guix shows a symlink to the 0.10.0-0.7611/bin/guix
<davexunit>do you use 'guix pull'?
<davexunit>that's an expected bug currently
<davexunit>guix always uses the modules in ~/.config/guix/latest if its there
<kyamashita>Good news! Red Eclipse builds, but it needs its data. How can I enable the data to be downloaded? I know that many other distros package the engine and the data seperately.
<htgoebel1>Should the packages and variables by prefixed with "django" or "python-django"?
<kyamashita>*separately
<keverets>davexunit: I do. Is there an alternative?
<bavier>kyamashita: look at other package in gnu/packages/games.scm to see how they handle "data" packages
<kyamashita>I'll play with it a bit.
<davexunit>keverets: so it's a known bug, is what I'm saying.
<davexunit>don't fret over it.
<davexunit>htgoebel1: python libraries are prefixed with "python-" typically
<davexunit>I would keep with that convention, but don't let any uncertainty in naming block you from packaging things.
<davexunit>we can adjust names later if need be.
<htgoebel1>I know. I just wanted to know if for frameworks this is the same convention.
<htgoebel1>:-)
<davexunit>yeah, it's the same.
<davexunit>a framework is a library
<davexunit>we don't really draw such fine distinctions like that
<davexunit>we can always deviate if there's good reason
<htgoebel1>davexunit: Do I have to add the new file to some Makefile.am? Some log messages say so.
<davexunit>htgoebel1: yes, otherwise it won't build.
<keverets>davexunit: ok, thanks for letting me know
<davexunit>it also wouldn't be installed by 'make install'
<htgoebel1>Found it. It's hidden in gnu-system.am :-)
<davexunit>I feel like avr-libc shouldn't need a 32-bit host-arch libc in order to build. I've built avr-libc in the past just fine.
<davexunit>but then it broke sometime, and we even have a bug report from someone last month about it
***kelsoo2 is now known as kelsoo
<htgoebel>Hmm, the license of one of the packages is "MIT Licence". There is no MIT-licence in guix/licences.scm. How should I add this to the package?
<davexunit>htgoebel: the reason we don't have such a variable is because "MIT license" is ambiguous
<davexunit>it often means either Expat or X11
<htgoebel>IC.
<davexunit>read the license text for them and you'll see if either fits
<htgoebel>davexunit: Thanks. I'll keep in mind
<bavier>is there no "what software license is this?" tool?
<davexunit>I find that it is usually Expat in these cases
<davexunit>I don't know of one.
<davexunit>but in general I'm wary of automated tools for this
<bavier>might be a fun hack for 'guix import'
<davexunit>copyright law is not something a machine can do
<bavier>davexunit: no, but in the spirit of 'guix import' it should be able to give a start
<davexunit>packagers *should* make an attempt to understand the licenses
<davexunit>yeah perhaps it could make a best guess
<alezost>there is a kinda helper tool: "M-x guix-licenses"
<bavier>alezost: nice
<bavier>haha, browsing the list, I see that our boost package maybe should have the "Boost 1.0" license instead of an x11-style
<davexunit>omg I just got avr-libc to build with a nasty hack!
<bavier>yay!
<davexunit>now to see if it can actually compile firmware
<davexunit>CFLAGS=-D__x86_64__
<davexunit>'(._. )
<lfam>Is the ' a celebratory raised fist?
<davexunit>it's a bead of sweat
<lfam>Ah :)
<ijp>✊(._.)
<lfam>That's more clear
<davexunit>that first character doesn't render for me
<davexunit>holy crap that compiler works!
<lfam>I think I see the first from the unifont
<lfam>Nice!
<lfam>fist
<ijp>davexunit: fwiw it is the unicode codepoint RAISED FIST
<davexunit>is there seriously a codepoint with that name?
<ijp>yes
<davexunit>well I'll be
<lfam>What *isn't* there a codepoint for?
<ijp>there is also FACE WITH OPEN MOUTH AND COLD SWEAT
<davexunit>fish in a bear suit
<lfam>I read through all the face / emotion codepoints recently, to pick the very best one
<lfam>It read like a poem
<ijp>I've been complaining about unicode bloat since before PILE OF POO made it cool :)
<lfam>You should send U+1F595 to the Unicode Consortium
<anthk_>how can I use "--fallback" with "guix package -u "?
<bavier>anthk_: place it before the -u
<anthk_>bavier thanks
<slim404>hi
<bavier>hello slim404
<slim404>i'm using guixsd gnome desktop
<slim404>bavier: hi
<slim404>I managed to set up the french keyboard on the console and in gnome..
<bavier>cool
<slim404>..but not in the display manager
<slim404>so my question is : what's the display manager?
<slim404>it seems to be SLiM
<slim404>but when I do "guix package -I slim" it displays nothing
<bavier>slim404: SLiM is typically installed in the system profile
<slim404>bavier: how can I setup a french keyboard in slim?
<lfam>I want to package https://github.com/letsencrypt/boulder
<lfam>We need Go
<bavier>lfam: what are the advantages over the referent implementation in python?
<bavier>*reference
<lfam>boulder is the client the reference client is talking to
<lfam>I mean, Boulder is the server
<bavier>oh, i see
<lfam>efraim: Should we add libidn to our Mutt package?
<efraim>lfam: I don't know of any email servers that need international characters in their address
<efraim>we probably should in the interest of supporting it, and it shouldn't add much if anything to the closure
<lfam>What would it be used for if there are no email servers that need it?
<efraim>internationisation is there on the internets so I think its more of not enough servers support it correctly
<efraim>but I assume it would also be useful if you had a username with UTF8 characters in your name
<lfam>So we'd be supporting the standard before it is widely adopted?
<efraim>or in your hostname
<efraim>I haven't looked at it too closely
<lfam>Okay, just wondering
<efraim>I can try setting up אפרים@flashner.co.il and see if it goes through
<lfam>Cool!
<efraim>it got forwarded to me correctly, but I think libidn is more for if I had efraim@פלשנר.co.il
<lfam>Right, the domain name
<lfam>Debian enables it