IRC channel logs


back to list of logs

<marusich>Acou_Bass, what errors? If it's something you can fix by changing your operating system configuration file, then you should be able to just run "init" again to fix it. No need to run "reconfigure," I think.
<marusich>Basically, if you encountered a problem, then the safest solution is to do it over again from the start.
<Acou_Bass>yeah i figured as much ;D doing that now - and my grub didnt prompt me for my LUKS drive password (but i realised thats because i stupidly put the device as 'crypt' rather than /dev/mapper/crypt) so hopefully will be fixed
<ng0>in case a grub update breaks (which should not happen, but grub config broke for me one time) it would be nice to have a functional way to chroot. right now, chroot into guix from the install-image lacks some functionality
<Acou_Bass>yeah i noticed that
<paroneayea>hey mnd made a nice guix services tutorial
<marusich>Now I wish someone would do that for grafts :)
<marusich>paroneayea, read it. Seems clear. I'll have to play more with services sometime. Will probably learn more by trying, than simply by reading. BTW, I'll bet your article would be even more useful with a few carefully placed hyperlinks to the relevant Guix docs.
<marusich>Thanks for sharing it!
<marusich>Does it work for only GuixSD?
<Acou_Bass>does look very guixSD specific
<marusich>Since it depends on using the operating system configuration file, I think it is
<Acou_Bass>but i dont really know if shepherd is configured that way in other distros (if any other distros use shepherd) :D
<marusich>Well, to launch a service, you can't use the GuixSD stuff on a non-GuixSD system. You'd have to do whatever it is you do on those systems to start services.
<marusich>E.g. create upstart jobs or something
<marusich>I wonder how a nonprivileged user would create services for themselves using Shepherd or GuixSD
<Acou_Bass>you mean like how systemd has the systemd --user services?
<Acou_Bass>so rather than adding them to configuration.scm (which needs sudo to edit) just have them in your home dir like with guix package -i? thatd be awesome ;D
<paroneayea>marusich: it's not my article :)
<paroneayea>but yes it's GuixSD specific. We don't have a way to do user services yet...
<paroneayea>it's been vaguely discussed but does not exist and would require some careful thinking.
<Acou_Bass>in my experience at least, user services can be handy for desktop HOWEVER for servers i just do like what was done in the article and run it as a root service but make it run under X user
<paroneayea>yes it's desirable
<Acou_Bass>the only real difference (besides not needing root to run them) is that ones ran from the root session run as soon as you boot up whereas user ones dont run until you login... so i guess depends what youre wanting to do hehe, cant hurt to have both though
<marusich>I think shepherd might have a way to do it, but it would likely not be integrated into Guix, and I can't remember what it is (if it exists).
<Petter>lfam et al.: How should I prepare a redundancy fix in a bunch of files? Here's me winging it -->
<cynede>good morning
<cynede>where can I find brief info about package manager, is it binary and packages example :)
<Petter>I usually run: guix package --help
<cynede>Petter I didn't install it yet :)
<Petter>Oh :)
<cynede>Petter what's recent version of gnome available in distribution repositories?
<cynede>is it based on nix codebase or just same ideas?
<Petter>gnome-desktop is at 3.20.2. Does that seem right?
<Petter>Hm, 3.20.4 for guix package --show=gnome
<cynede>Petter yes, sadly actual version is 3.22.x
<cynede>Petter so is guix packages source based, e.g. how hard is to bump some package when you need it
<Petter>Have you seen this?
<cynede>not yet
<Petter>Packages are build from sources, yes.
<Petter>There's a build farm that does this, and the user can select to build everything from source as well.
<cynede>wow packages are on guile? it looks fun
<Petter>It is :)
<cynede>found that
<Petter>I actually just the other day package a new program (my first) based on that hello example.
<cynede>Petter distribution installation is ... semihow automated, some random script installs anything?
<cynede>e.g. you prepare disc and network then run some script and it's done?
<Petter>Yeah, pretty much.
<cynede>very confusing... you need to write guile script before installation?
<cynede>e.g. you install system wia guile script which defines what you want, users, drivers, etc, and if something will go wrong - correct script?
<Petter>You copy an example config and modify it.
<cynede>Petter can I at least typecheck it before getting in trouble with automation? :)
<cynede>Petter I think I need to run first installation on VM to see how that will work but so far it looks interesting for me
<Petter>Not sure, maybe you can run some lint program if that's what you mean.
<cynede>Petter yes I mean that, it will be ideal to get live CD with setup IDE which will check correctness for me :)
<Petter>There's "guix lint <file>" which I can run on an installed system. I would assume it's available during installation, but I don't know.
<cynede>Petter where is packages repository?
<cynede>Petter is there multilib support or is it multiarch and very complicated? :)
<cynede>is it legal to write non-gpl packages for myself, alike chromium instead of some gpl web browser :)
<Petter>GPL isn't a requirement. But it should be Free Software. Non-free software is not encouraged, and you won't get any help from the community. But it's not illegal as far as I know.
<cynede>Petter free or open source?
<cynede>Petter I mean google chrome is free and chromium is open source, not sure if either of them are welcome or only second one
<cynede>Petter what about nvidia?
<Petter>I don't know. They don't release Free Software drivers do they?
<cynede>Petter they release drivers for linux but well...
<cynede>Petter it's pointless to use linux distribuion without drivers support with their hardware XD
<Petter>I believe nouveau does this.
<cynede>Petter nouveau is terrible
<cynede>Petter I mean perfomance is much lower and window managers are buggy with that
<cynede>Petter will try talk more on monday, see you
<ng0>iyzsong: I don't understand the reply to the rfc git-service i've sent.. did you fix it? my intention with the service is to learn and correct until it works (again, used to work almost before I added many options)
<ng0>i have yet to read the entire patch, i've just seen that you added a co-authored by
<iyzsong>ng0: yes, I roughly adjust it to my taste..
<ng0>ah, i#ve seen your message before that now
<ng0>I'm okay with it. I have some further comments I'll add inline in the new patch, and ask questions about parts if I do not understnad them. Thanks
<iyzsong>ng0: sure, thanks!
<ng0>what's @asis in tex?
<iyzsong>well.. i don't know, it seems other tables are using this format, so I just follow. (will searching).
<iyzsong>it's: 'asis' - Do not change the existing indentation.
<rekado> You may use the '@asis' command as an argument to '@table'. '@asis'
<rekado>is a command that does nothing: if you use this command after '@table',
<rekado>the first column entries are output without added highlighting ("as
<rekado>that’s what the Texinfo manual says
<rekado>I just looked up ‘asis’ in the index.
<iyzsong>ok, thanks!
<ng0>ah, thanks
<Petter>Hi. lfam encouraged me to fix a redundancy in multiple package recipes. I have, but I'm not sure how I should send them. Is one patch like this ok? Or should there be one patch pr. file?
<Petter>Also, I don't know what to write in the commit message.
<iyzsong>Petter: yes, i think it's should be one patch, and the commit message can use the 'Likewise' pattern :-) (i failed to open the page with my slow net).
<Petter>I'm not familiar with the Likewise pattern. I'll see what I can dig up about it. Thanks!
<ng0>for "and=>", the => , is this in the gexp of guix documentation mentioned or which section of the guile documentation do I look at to understand it?
<ng0>> + (and=> #$base-path mkdir-p))))
<ng0>I understand what it will do, but I'd like to understand the => part
<ng0>read up on it I mean
<mange>Petter: look at 966a543 for an example of "Likewise"
<Petter>mange: Thanks!
<Petter>I see.
<iyzsong>ng0: it's a simple guile builtin, "When VALUE is`#f', return `#f'. Otherwise, return `(PROC VALUE'). -- from 'info guile'.
<ng0>ah, thanks
<ng0>i think my rewrite made the service too complicated for me. i should've used the initial revision I had
<ng0>yours is simple
<iyzsong>um... does return '#f' in the activation gexp harmful?
<ng0>i'Ve sent my comments
<iyzsong>ng0: re '/srv/git', I see that from 'man git-daemon', also i prefer '/srv' for serve files (/srv/http, /srv/ftp, etc.). while files under /var can be cache, incoming, etc.
<iyzsong>yeah, our nginx is changed from using '/var/www' to '/srv/http' commit 8c00b838 :-)
<ng0>that's the only /srv so far?
<iyzsong>yes, with git-daemon, two. we don't have a ftp service yet.
<ng0>what did you use for opensmtpd?
<ng0>/var/mail ?
<iyzsong>yes, it's the default mbox location. note it's not for serving to public.
<ng0>when was the meeting in berlin again, 12th?
<Petter>In a patch, should one add oneself to the Copyright list for any change, no matter how trivial?
<Acou_Bass>eey guys, im not sure if this is something guixSD supports, but is there any way to have a sort of... i dunno, 'private browsing' user account, ie. one that i can log into and the data gets wiped after a reboot?
<Acou_Bass>im sure i remember having such a thing on arch a long time ago but not sure if its even a thing that people do anymore XD
<Petter>Maybe put that users home account in /tmp. I don't know.
<Acou_Bass>yeah ive been having a poke around, seems like thats one way of doing it
<Acou_Bass>its not urgent, more of a curiosity than anything ;D used to always have one on my PC for if friends come round, have a nice locked-down account with none of my details on it and minimal access to things like web browser/VLC hehe
<ng0>create a disk-image with what you want to dd to an usb stick and use that?
<rekado>ng0: yes, 12th
<ng0>i have a couch for the weekend, i only need to find out if someone can come by for the cat
<rekado>I ran out of space again right before finishing a big upgrade. Is there a way to temporarily mark all those nice build artifacts that are already in the store that I’ll need for the upgrade and exclude them from GC cleaning?
<ng0>notmuch build is broken
<snape>Does anyone know if there is a way to know whether a define-record-type* getter uses the default value or not?
<snape>I want to do something depending of whether the value has been explicitely set by the user.
<Acou_Bass>ok errm... im trying to run guix pull (fresh install, so doing the usual first-time update), and its giving me a failed to download error, says ERROR: In procedure getaddrinfo: Name or service not known
<Petter>Have you set up the network?
<Acou_Bass>yeah i can ping the outside world just fine
<Petter>Using IP or domain name?
<Acou_Bass>ahh crap, yep, IP only cant do '' but i can do ''
<Acou_Bass>there we go
<Acou_Bass>for some reason dhclient wasnt run, my mistake ;D thanks!
<Petter>Great! :)
<Acou_Bass>also managed to get an encrypted root going, though for now i have to drop to the grub command line and manually decrypt it (though really thats no different from entering the password for it on boot anyway), so woop
<Petter>Is Grub installed in boot firmware or on disk?
<Acou_Bass>its on the HDD - two partitions
<Acou_Bass>/dev/sda1 for grub /dev/sda2 for crypted root
<Acou_Bass>so yeah grub isnt encrypted but oh well
<Petter>Hm, are you sure Grub's not installed in MBR, and /dev/sda1 is for /boot?
<Acou_Bass>yeah thats what i meant ;D
<Acou_Bass>i always thoght /boot was for grub? or is that just for the linux kernel
<Acou_Bass>(or what?)
<Petter>I'm not sure, I think the boot loader is installed in a special place on the disk (MBR), and each distribution has their boot files in their own /boot.
<jmd>Petter: If is complicated :( there are two ways machines can boot.
<jmd>The old fashioned way is the MBR. The modern way is UEFI.
<Acou_Bass>yeah my laptops definitely not new enough to use UEFI
<jmd>I think with UEFI, /boot/grub can be anywhere.
<jmd>With MBR it can be on any primary partition.
<jmd>That is my understanding, at least.
<Petter>Do you know if each distro on a multi-boot system uses the same /boot/grub, or how that works?
<jmd>I don't think it has to be in /boot/grub - in principle it can be anywhere. But /boot/grub is a common convention.
<Petter>I should read up on this... some day.
<snape>btw, the fact that grub is in MBR allows to have /boot encrypted
<snape>but could get it to work with Guix, unfortunately
<snape>s/but could get/but I could not get
<jmd>snape: Are you sure? the Stage1 would have to be unencrypted at least.
<snape>jmd: I heard so. stage1 is unencrypted in the mbr, I think.
<ng0>i think the problem for notmuch could be the new gnupg
<ng0>T350-crypto: Testing PGP/MIME signature verification and decryption <- that's one of the fialing test suites
<jmi2k__>Hi, I'm trying to package neovim, but it requires downloading dependencies when building. Any ideas?
<rekado>jmi2k__: add the dependencies to the inputs as origins and disable downloading
<jmi2k__>how can I put more than one origin?
<lfam>Fallout of the gnupg 2.1.16?
<lfam>jmi2k__: See the git package definition. Specifically, the 'git-manpages' input
<jmi2k__>Ok, I'll take a look
<ng0>rekado: I just secured a couch and a friend to feed our cat, I'm coming to the meeting on 12th :)
<lfam>It's possible, even likely, that neovim's build system can be persuaded to use locally available dependencies and not try to download anything. That's what our Python packages do. If the dependencies are missing locally, it will try and fail to download them
<lfam>Speaking of meetings... I bought airfare to FOSDEM :)
<ng0>lfam: possibly, otherwise I don't know how notmuch owuld break
<ng0>well, the tests break
<ng0>4 of them
<lfam>ng0: Good thing Guix doesn't let your installed notmuch break :)
<paroneayea>ooh I need to buy FOSDEM fare
<paroneayea>and figure out where I'm staying!
<lfam>Yeah, I need to figure that out too :)
<paroneayea>whee, I think dirvish is packaged. No service written yet.
<lfam>Nice. Back everything iup!
<ng0>it looks like supernodes of freifunk could almost be run by guixsd. it's mostly git and ansible.
<ng0>and vpn, etc
<jmd>Does anyone else constantly have problems cross building qemu?
<jmd>c1s6bsxz-qemu-2.7.0.drv' timed out after 3600 seconds of silence
<lfam>jmd: Can you give me a command to run to try reproducing it on x86_64 hardware?
<jmd>./pre-inst-env guix build qemu --system=i686-linux
<ng0>where does "autopoint" come from?
<lfam>jmd: Okay, I'll try to reproduce it
<jmd>ng0: It comes from automake or autoconf - one of the two.
<jmd>lfam: Thanks. It takes a while.
<jmd>ng0: Oh no. You are right. It comes from gettext. Wierd!!
<lfam>jmd: Looks like I already had the i686 QEMU substitute for current HEAD
<lfam>So, it's definitely possible :)
<jmd>I wonder why I don't have such a substitute?
<jmd>Which HEAD is for you "current" ?
<lfam>jmd: 25291aca748b8bf42bebb67134fa12a5c5980aeb
<albertoefg>you seem to have a lot of technical knowledge
<lfam>It's possible I built this package myself; not sure
<albertoefg>on this channel
<paroneayea>albertoefg: building distributions is a great way to pick up a bunch of technical knowledge :)
<lfam>I think that Guix is a good project from which to learn about computer systems
<lfam>You have to learn about all sorts of different programs and how they fit together
<albertoefg>I see
<albertoefg>Ok you have convinced we lfam i will install guixSD
<lfam>Okay :)
<albertoefg>but probably tomorrow
<albertoefg>first i have to back up
<lfam>That's a good plan
<jmd>On the other hand GuixSD deliberately prevents one from doing certain things which might break your system. Breaking a system, is a very good way to learn about it.
<lfam>That's true
<ng0>guix system: error: service 'file-system-/sys/fs/cgroup/elogind' requires 'file-system-/sys/fs/cgroup', which is not provided by any service what's needed in a dummy file-system based system needed to make this disappear?
<ng0>oh. base-file-systenms, right?
<ng0>or.. hm.
<efraim>got past the kodi ffmpeg failure with ffmpeg-2.8, but can't get the tests to pass
<efraim>giving up on that again for now
<efraim>have we worked out uefi support on guixsd?
<quiliro>how can i download all needed to install GuixSD offline?
<efraim>maybe `guix archive --export $(guix build os-config.scm --sources=transitive) | tar cf inputs sources.tar - `; and then `guix archive --import < tar xf sources.tar` once you're in the install medium?
<efraim>but we don't really have anything official for offline installs
<ng0>this is strange. guix system vm configgit.scm is just ending (or better: not starting, not doing anything) without a message.
<ng0>maybe I need to make clean
<fps>ng0: not at a guix system right now but does it maybe have a verbose switch?
<paroneayea>do any of our services expose the configuration file they generate in a reasonable place that you could find it?
<paroneayea>I ask because some with some programs, there is more than just "stop" and "start"
<paroneayea>you would feasibly want to run other operations than just what are managed by shepherd
<paroneayea>such as running schema migrations, or initializing some data store, etc
<paroneayea>but how do you find the config file that was configured by guix?
<jmd>paroneayea: In general, you can't.
<jmd>But you don't actually need it in order to run other operations.
<jmd>Although it might be useful for reference purposes.
<paroneayea>jmd: I think "you don't actually need it in order to run other operations" is making assumptions that can be false.
<paroneayea>eg, right now I'm trying to write a service for dirvish
<jmd>I suppose it depends what one means by "need".
<paroneayea>dirvish needs to be able to set up "vaults" for its backups
<paroneayea>they need to be initialized
<paroneayea>and, mediagoblin for instance needs to be able to run migrations. that's not just a start/stop operation.
<paroneayea>in order for those to work, you'll either need teh config file
<jmd>I don't think that is correct.
<paroneayea>or you'll need a script that's already wrapping passing in that config file for you
<paroneayea>jmd: why isn't it correct?
<jmd>Well why IS it?
<jmd>You need to define a service and then start it.
<jmd>It may well be easier to do that starting from an existing config.scm but it is not necessary.
<jmd>Note, that config.scm by itself is useless, except to pass as the subject of guix system reconfigure.
<jmd>(or init)
<jmd>But, anyway I agree with you that it would be nice to have configurations save the program which generated them.
<paroneayea>jmd: it's nto just starting a service
<paroneayea>it's running migrations or initializing data stores or some other operations (for example, some applications have a garbage collection operation that might be run manually, or a data integrity check)
<paroneayea>ACTION writes to the mailing list
<paroneayea>sent an email to the list.
<ng0>guix system: error: service 'file-system-/sys/fs/cgroup/elogind' requires 'file-system-/sys/fs/cgroup', which is not provided by any service
<ng0>how do I get rid of this for guix system vm?
<ng0>I used my systems config.scm and exchanged the file-systems section by a dummy
<jmd>paroneayea: I'll let someone else reply to your mail, but in the meantime I suggest that you look at the elogind service for inspiration.
<jmd>(I don't see that this has anything to do with preserving the config.scm file)
<paroneayea>jmd: it's not about preserving the config.scm file
<paroneayea>it's about the application-specific config files generated *for the daemon that runs*
<jmd>Oh then I misunderstood you.
<jmd>Those can also be generated by the service.
<paroneayea>jmd: right. it's just that right now, there's no way to access them afaict, so running a manual command that needs those configs for context can't be done
<rekado>I don’t get substitutes for linux-libre, doesn’t matter if 4.1 or 4.4.
<rekado>hydra doesn’t show queued builds
<jmd>paroneayea: Why is there no way to access them?
<paroneayea>jmd: they aren't exposed, are they?
<jmd>What do you mean "not exposed"? Do they exist or not? If so, then you could access them.
<paroneayea>jmd: they are somewhere in the store, but there's no obvious way for the user to find them. they're closer to being variables provided to the closure of the service than anything that a user can use.
<paroneayea>sure, I could go hunting in /gnu/store to try to find the file
<paroneayea>so I could do "dirvish --config=/gnu/store/..."
<paroneayea>but how would I know I found the right thing?
<jmd>I'm not familiar with dirvish, so I can't answer that.
<jmd>How do users normally know they've found 'the right thing'?
<paroneayea>I feel like I'm not getting through here. Normally on imperative distributions a user would point to something in /etc/dirvish/default.conf or something
<paroneayea>but since we're generating something that isn't exposed
<paroneayea>I think I've explained this a lot and if it's not clear I'm not going to do better.
<jmd>Normally, a user would not type in --config=/etc/dirvush/defualt.conf
<jmd>Presumably it's the default.
<paroneayea>jmd: right, but on guix it isn't.
<jmd>So you arrange for instead, /gnu/store/.../xxx.conf to be the default.
<paroneayea>right but ****I don't know where in the store I would find xxx.conf****
<jmd>Is this conf file generated at build time or later?
<paroneayea>jmd: it's not at build time of the package
<paroneayea>it's at service time
<paroneayea>so the package can't set it to be the default
<paroneayea>make sense?
<jmd>so you need a "gexp".
<paroneayea>if I was going to build a wrapper around it, sure
<jmd>There are a number of such services you can look at.
<paroneayea>but the services don't themselves build a wrapper script that set the default do they? they're just providing start/stop stuff to shepherd
<paroneayea>which is the very thing I said wasn't sufficient!
<jmd>You can write a service which provides a gexp in its start, and passes that the the daemon.
<paroneayea>jmd: right.. but what I'm saying is I need something the user can run
<paroneayea>I feel like I'm not getting through here, so I'm just going to stop for now.
<rekado>the only point where the location of the configuration is made explicit is probably at system reconfiguration.
<rekado>I don’t know if there’s another point where you could ask for the location of the configuration file.
<jmd>rekado: I think he isn't talking about that (although I thought he was at the start of the conversation).
<rekado>the derivation holds the location of the configuration file in the store.
<rekado>another crude way is to use ‘pgrep -fa conf’
<jmd>rekado: So the configuration file is copied to the store?
<rekado>the configuration files for all services is somewhere in the store.
<rekado>if I understood paroneayea correctly, they want to find the location of the configuration file generated by Guix to be able to pass it to a command the user may want to run manually.
<jmd>Anyway, one thing that is clear. There needs to be some better guide on writing services.
<rekado>e.g. to perform migrations that would not be handled by the service but for which the existing config file is needed.
<paroneayea>rekado: yeah, you have it right.
<paroneayea>so for example
<paroneayea>"foo-db --config=/path/to/config.cfg gc" to manually run garbage collection
<paroneayea>"foo-db --config=/path/to/config.cfg migrate" to manually run schema migrations
<paroneayea>since the path of the config is only passed to the service, I don't have a way to run manual commands
<paroneayea>jmd: make sense now?
<rekado>can you assume that the version of Guix stays constant for the period between installing the tool and running the extra commands?
<rekado>because then you *could* be able to provide extra commands (e.g. a script) using a gexp.
<paroneayea>rekado: yeah that would be fine using a gexp
<paroneayea>but the gexp would need to be able to have access to the same config used by the service I think, right?
<paroneayea>I think these kinds of operations will be pretty common on servers, so we'll need a reasonably nice way to do this
<rekado>this means that Guix may not be upgraded, which is something you cannot reasonably assume.
<paroneayea>rekado: hm, because gexps break on service upgrade?
<rekado>no, because the gexp would reference the latest version of the configuration file.
<paroneayea>rekado: maybe the right thing is to allow services to expose scripts that wrap the original command with the --config passed in explicitly or something.
<paroneayea>I'm not sure.
<rekado>say you installed the server a week ago, then updated Guix, and then build the gexpy script to migrate.
<rekado>a more general approach would be to allow services to define their own sub-commands
<paroneayea>rekado: sub-commands to shepherd?
<rekado>this would mean swapping the position of actions and services
<rekado>herd foo-db frobnicate
<paroneayea>rekado: right, so shepherd does have the "actions" slot
<rekado>oh, okay
<paroneayea>rekado: it would also mean that shepherd would have to be able to take arguments to those actions; I'm not sure if that's currently possible
<paroneayea>eg if I did
<rekado>it sure would be nice if shepherd services could retain ‘context’ (like a reader monad)
<paroneayea>"foo-db gc --aggressive"
<paroneayea> -- method: action (obj <symbol>) the-action . args
<paroneayea>maybe it is posssible to pass in args currently!
<paroneayea>I should test this
<paroneayea>rekado: maybe this is the right way forward.
<sirgazil>Hi, I'm using Guix on Debian. How can I make guile find guix modules? I'm getting this error when importing a module that uses (guix ui):
<paroneayea>hi sirgazil
<sirgazil>Hey, paroneayea :)
<sirgazil>This is guile 2.0.13 from Guix
<paroneayea>sirgazil: hm, so one way you can do it
<paroneayea>cd path/to/guix
<paroneayea>./pre-install-env guile # --listen=/tmp/guile-guix-socket <- in case you want to M-x geiser-connect-local in emacs
<paroneayea>sirgazil: I hope that helps
<paroneayea>sirgazil: also a long time ago I wrote
<paroneayea>maybe pretty out of date thoug.
<sirgazil>Oh, I really don't understand. I installed guix from the binary distribution, I'm not using from the git repo :)
<sirgazil>paroneayea: I mean, with cd to guix you mean to the git repo, right?
<paroneayea>sirgazil: ah yeah
<ng0>i wonder what i'm doing wrong with guix system vm suddenly
<paroneayea>sirgazil: I never installed from the binary distribution so I'm not sure :)
<sirgazil>paroneayea: Ok, thanks :)
<sirgazil>I was expecting to be able to use Guix modules in a REPL, but I don't know what should I set...
<ng0>but the system (not git checkout) guix system vm seems to work
<efraim>I'm thinking of putting together a shirt before FOSDEM
<efraim>GNU GUIX, next line the 3 raindrops, next line: 'define*' your system
<Acou_Bass>efraim: id wear the hell out of that :D
<ng0>no, i don't understand it
<ng0>why does this suddenly no longer work
<ng0>if I would have a message, but I get nothing
<paroneayea>sneek: later tell lfam talked with assword developer dkg:
<Acou_Bass>is there an alternative to the webpage? i find it bogs my browser down terribly ;(
<civodul>Acou_Bass: you can run "guix package -A"
<civodul>that web page is going to be split Real Soon
<sneek>Got it.
<Acou_Bass>cool civodul ;D and thanks for that command hehe
<paroneayea>what you talkin' bout sneek
<Acou_Bass>i also saw theres an emacs mode for it too, maybe thatll be better
<paroneayea>the emacs mode is *great*
<paroneayea>M-x guix-all-available-packages <- esp this :)
<Acou_Bass>hehe ill give it a bash then
<paroneayea>or an elisp!
<Acou_Bass>... thats it im going back to ubuntu
<paroneayea>hey civodul, how you doing
<Acou_Bass>hmm i wonder if itd be possible to run guix, inside the libertine container, on an ubuntu phone...
<civodul>would be fun
<civodul>paroneayea: i'm good, & you?
<paroneayea>you probably could run guix on such a device, but would probably want to push built packages to it
<paroneayea>rather than building from the phone :)
<Acou_Bass>does the guix build farm have ARM binaries?
<paroneayea>civodul: pretty good. Working on that dirvish stuff still
<civodul>your pocket could get warm :-)
<paroneayea>civodul: you may have seen my email to guix-devel :)
<civodul>not yet!
<civodul>i was trying to get my icecat fix done
<civodul>and a syscalls.scm hack
<Acou_Bass>but guix itself DOES have an ARM version doesnt it? id just have to setup a build farm for it ;D
<cheim>Acou_Bass: there are some packages built for the ARM architecture. Otherwise my beaglebone would be overwhelmed.
<cheim>I was really fortunate that the build farm had built libmicrohttpd. One step closer to installing gnunet.
<jmd>is libmicrohttpd a very costly thing to build?
<Acou_Bass>most things are costly to build on small ARM Soc's :P
<cheim>No, but one of the build tests wasn't passing for me, so I couldn't install it
<cheim>I wasn't sure how to go about fixing it myself either
<Acou_Bass>ahh the emacs interface is much better
<Acou_Bass>unsure yet if i want to copy my emacs config from my desktop over or just start afresh :D i dont need too much in emacs for my laptop just lilypond/tex/org