IRC channel logs

<oriansj>will someone please fix lightdm-gtk-greeter http://paste.lisp.org/display/355455

<wigust>oriansj: Could you try to download lightdm-gtk-greeter manually? Like “guix download https://launchpad.net/…” (first URL above ERROR in log). Or just try later. Seems launchpad is not available for you now for some reason.

<oriansj>wigust: launchpad is available but the url is requesting a file that does not exist (it post pends the hash (it is literally attempting to download https://launchpad.net/lightdm-gtk-greeter/2.0/2.0.2/+download/lightdm-gtk-greeter-2.0.2.tar.gz/gnu/store/iaphvvldvzmg2j3r5xa6pjs54bf4a854-lightdm-gtk-greeter-2.0.2.tar.gz which does NOT exist)

<oriansj>if the url is truncated to just https://launchpad.net/lightdm-gtk-greeter/2.0/2.0.2/+download/lightdm-gtk-greeter-2.0.2.tar.gz it will download without issue

<oriansj>has anyone seen this before? http://paste.lisp.org/display/355466

<oriansj>oh and guix system switch-generation broke with it

<jak3b>hello

<jak3b>are there sample config.scms other then in /etc/configuration, for instance id like to have a gnome desktop environment no crypto and a dual boot with archlinux

<efraim>sneek: later tell civodul you forgot the 2 foomatic-filters patches in your commit

<sneek>Got it.

<efraim>sneek: botsnack

<sneek>:)

<wigust>oriansj: I failed to build lightdm-gtk-greeter, too. It's strange for me, because when I manually copy failed URL and paste to “guix download” then it works. Did you open a bug report? Maybe I'll try to fix it, if I could.

<wigust>(package-source lightdm-gtk-greeter) returns valid URL, hm.

<ng0>what's the closest thing to wildcard package building? I want to check the last half of the google fonts I did not build.

<ng0>like guix build --quiet font-google-noto-* but in a way that works

<ng0>I could also work with building all of (gnu packages fonts).

<ng0>I wrote a shellscript. But iirc there is a way with guix.

<efraim>i do it with `guix package -A' and cut or with grep and cut

<ng0>ok

<ng0>thanks

<buenouanq>is there a reason why guix can't do globbing/wildcards like that?

<buenouanq>funnily enough, the main thing I want it for also involves font packages

<ng0>86 patches incoming now :)

<janneke>ACTION just compiled bit-for-bit identical `int main (int argc, char *argv[]) { return 42; }' with mes-tcc and i686-unknown-linux-gnu-tcc

<oriansj>Someone really should do some chaos monkey style testing with guix

<efraim>oriansj: I haven't installed guixsd on that new computer yet. Pentium e2140, 64 bit, so it looks like a nice upgrade from some of my systems

<oriansj>efraim: I'm personally elbow deep in trying to fix guixsd on my test system (it broke hard)

<oriansj>efraim: out of curiousity where are you located?

<efraim>Israel

<oriansj>I'm in Michigan myself

<oriansj>I have an old server (Sun SPARC M4000) that needs a new home but its a bit too big for shipping overseas

<efraim>Is that SPARC or SPARC64?

<efraim>I heard they were noisy

<efraim>According to Wikipedia its 64 bit

<oriansj>yep

<efraim>6U? that's a beast

<efraim>Anyway, I need to find all my spare old hard drives to stick in this machine, and then figure out if guixsd can boot from a btrfs raid 10

<efraim>If I understood the BIOS right I can stick in 4 IDE drives and another 4 sata drives

<archie123412>Hello, is it possible to install Guix in home folder?

<jonathan_>oriansj: so you could work on a sparc64 port :P

<ng0>what could be an autotools way to check for guix? I'm just getting a bit deeper into autotools.. Maybe simply check for the existence of binary "guix"?

<ng0>with autoconf of autotools in this case

<ng0>otoh maybe I don't need to check at all if what I'm doing implicitly requires guix and can not be used without it.

<ng0>experimenting .. eventually it will work out

<civodul>Hello Guix!

<sneek>Welcome back civodul, you have 1 message.

<sneek>civodul, efraim says: you forgot the 2 foomatic-filters patches in your commit

<civodul>efraim: yep, sorry!

<civodul>fixed now

<efraim>now I can try building it on aarch64

<civodul>anyone would like to provide feedback for https://bugs.gnu.org/28324 ? :-)

<ng0>the idea sounds solid

<efraim>i'm going to need to think about the trusting people part :)

<efraim>before I go and experiment, if I set up a raid10 btrfs, can guixsd boot from that?

<civodul>woow, no idea

<civodul>it can boot from btrfs, see the test in (gnu tests install)

<civodul>i dunno whether the "raid10" part makes a difference

<civodul>ng0: there are several items opened at bugs.gnu.org/guix-patches about Google fonts

<civodul>ng0: could you merge them if needed?

<ng0>I closed all but one

<civodul>i'm afraid of getting lost, or applying the wrong one, etc. :-)

<civodul>ok

<ng0>just look for the one with 86 patches in the first message

<efraim>civodul: I'll probably have to play with mapped-devices like bayfront.scm to make them appear as one device

<civodul>ng0: i still see 4 font-related items in total, but maybe only the latest one is Google?

<ng0>yes

<ng0>the other font ones are different fonts

<ng0>lato, etc

<civodul>awesome, thanks for cleaning it up!

<civodul>right

<civodul>efraim: doesn't btrfs have built-in RAID support?

<civodul>ACTION is ignorant

<efraim>i'm pretty sure thats how it works

<civodul>the thing used for bayfront is the whatever-it's-called in-kernel, file-system-independent thing

<efraim>mkfs.btrfs [options] dev [dev ...]

<efraim>so its part of btrfs, but I want it to load all the drives at once, so i'll have to see how it apears when I'm formatting them

<mekeor>hellou guix :)

<janneke>civodul: wouldn't it be nice if store items started with their proper names and end with hash?

<janneke>instead /gnu/store/deadbeaf0123-bash-4.2, have /gnu/store/bash-4.2-deadbeef0123?

<mekeor>yeah, we already discussed that at some point here in the channel. there were arguments against it but i forgot them -.-"

<janneke>mekeor, okay good. i can imagine but itbn if that was documented somewhere. just wanted to be sure we didn't overlook this simplification.

<ng0>eaier to grep etc

<ng0>the binary "convert" is part of graphviz, right?

<ng0>oh. imagemagick

<efraim>Yep

<ng0>ah I want ghostscript. what I need to achieve is to take a pdf and convert it to a png. at least I believe png is more suite for texinfo pdf/html for this image which is no .svg

<ng0>no success. I guess imagemagick should work

<ng0> https://www.gnu.org/software/texinfo/manual/texinfo/html_node/Image-Syntax.html: "If you want to install image files for use by Info readers too, " there are info readers with image support?

<xkapastel>ng0: i would guess emacs is one of them

<ng0>I read so down that page

<ng0>now this page confused me about image extensions.. it used to work. now I get extensions not found.

<ng0>hm

<ng0>what's even more funny, the "can not be found, error" images are in the pdf version

<ng0>oh.

<catern>dear #guix, what are the prospects for interoperability between Guix and Nix? using the same daemon? depending on packages produced by one, from the other?

<catern>unfortunately i've been forced into pushing for Nix in my group rather than Guix (which I'd prefer) 'cause it seemed like people were more scared of Guix and I wanted to make my work as light as possible

<catern>but I want to keep my options open for the future :)

<catern>(or more specifically use guix in the future)

<catonano>why are they scared ? Because of scheme ?

<catern>yes, also GNU and GPL and stuff

<catern>i can't understand the reasoning of corporates ¯\\_(ツ)_/¯

<catonano>ah

<catonano>you should ask Eben Moglen for advice ;-)

<Apteryx>efraim: I see you are interested in btrfs raid

<efraim>Apteryx: yeah, it goes nicely in my free desktop with muliple leftover drives :)

<Apteryx>I'm looking at it also. I'm under the impression that it is unnecessary to define explicitly the mapped device. btrfs seems to be able to scan and assemble the devices itself. We only need our initrd to run the command `btrfs device scan`. We can then mount the assembled raid with any of the drive which is part of it (so we can simply use /dev/sda say in our config)... Seems easy enough!

<Apteryx>"Once you create a multi-device filesystem, you can use any device in the FS for the mount command [...]" from https://btrfs.wiki.kernel.org/index.php/Using_Btrfs_with_Multiple_Devices

<Apteryx>And also: "After a reboot or reloading the btrfs module, you'll need to use btrfs device scan to discover all multi-device filesystems on the machine [...]"

<civodul>janneke: it's a common suggestion :-)

<civodul>it's too late to change, and it wouldn't necessarily be helpful

<civodul>currently, on a reasonable store, a couple of characters are enough to discriminate a store item

<civodul>which is pretty convenient

<janneke>civodul: hmm, so many aspects to this

<janneke>i just installed GuixSD from USB and used the remote ssh option

<janneke>it had a hint about using passwd to enable ssh login

<janneke>passwd is not in PATH by default... /gnu/store/shadowTAB would be much friendlier than find /gnu | grep etc

<janneke>esp when stuck in a bournish...but you know all that...

<civodul>janneke: we should add shadow to the packages of the install image

<civodul>will do right away

<civodul>one shouldn't have to type hashes in a terminal :-)

<janneke>great -- of course, that's part of why the `common' suggestion ;-)

<jak3b>hello

<janneke>hey jak3b

<jak3b>I am installing Guixsd on my desktop, I want to dual boot with Void linux

<jak3b>where in the config.scm do I add the menu entry?

<jak3b>I keep getting an invalid field error

<civodul>jak3b: could you paste the exact error message?

<civodul>janneke: did you need just 'passwd', or also something else?

<civodul>while we're at it :-)

<janneke>civodul: i could not scp to the box, `scp' was missing

<janneke>so i had to cat | ssh 'cat > .ssh/authorized_keys' etc

<civodul>ok, we can add it too

<civodul>what else? :-)

<janneke>i only have one more wish: that master always be A-OK whenever I decide to run guix pull ;-)

<janneke>other than that...compiling .go files yadda,yadda, it was beautiful!

<civodul>:-)

<jak3b>hello?

<jak3b>ah

<jak3b>the message is error invalid field specifier

<civodul>the message certainly contains the name of the field :-)

<civodul>can you copy/paste it?

<janneke>i managed to find this ~12h window of master lcap breakage in weeks of being OK...but that allowed me to experience our great support by efraim, pmikkelsen, lfam

<jak3b>Im on my lap top

<civodul>oh, and i guess i was the one responsible for the breakage :-/

<civodul>(the missing files)

<janneke>oh, good

<jak3b>I am using the lightweight-desktop.scm as a template

<jak3b>I use the format from the grub docs for menu entry

<jak3b>but I keep getting the error

<civodul>jak3b: the error message normally mentions the name of the "invalid field"

<civodul>could you please check it?

<jak3b>yes

<jak3b>"config.scm:17:2: error invalid field specifier"

<jak3b>I put the menu entry right after the file systems entry

<jak3b>then put it right after the initial grub configuration lines

<jak3b>either way its not correct

<jak3b>I am not sure from the docs exactly where to make the entry

<civodul>jak3b: the message looks like this: "error: extraneous field initializers (sourcex)"

<civodul>where 'sourcex' is the invalid field in question

<civodul>what's the invalid field in your case?

<civodul>also, can you paste config.scm to paste.lisp.org or similar?

<jak3b>not right now

<jak3b>can I just install and edit grub.cfg later?

<jak3b>or will it get over written on evrey update?

<jak3b>Im pretty sure (menu-entry is the invalid field specifier

<jak3b>is the format (menu-entry (cons (menu-entry?

<jak3b>is the format "(menu-entry (cons (menu-entry"?

<jak3b>sorry

<jak3b>anybody dual boot?

<civodul>too late :-/

IRC channel logs

2017-09-10.log