IRC channel logs

2017-02-10.log

back to list of logs

<enderby>o
***Piece_Maker is now known as Acou_Bass
***reepca` is now known as reepca
<civodul>Hello Guix!
<effa>civodul: Hey civodul, are you around?
<civodul>hey effa, i'm around :-)
<effa>Yesterday I pushed the Stackage patch. For the error message I used '(leave (_ "package not found: ~A~%") package-name)'.
<effa>Could you explain me why you prefer not to use the (guix ui) module there?
<effa>OK, got it. Thanks! I will change the code to use '(raise ...)' as per your email during the WE.
<slyfox>civodul: does 'guix' happen to be a derivative of 'gui' french word or something like that? :)
<civodul>effa: cool, i hope that made sense!
<thomasd>Hi Guix
<civodul>slyfox: no :-) initially it was simply the contraction of Guile + Nix
<civodul>i'm surprised few people think of it
<civodul>hey thomasd!
<thomasd>I'd like to try and fix bug 25667 which I just reported (store references in ${HOME}/.config for xfce4), but haven't actually found out where this config gets created.
<slyfox>oh, i see :)
<civodul>thomasd: i guess it could be in any of the Xfce packages, maybe in a core library of Xfce
<thomasd>I've found something about skeletons in "activation.scm", but that only deals with bashrc etc, not the desktop-specific items. Maybe someone here can point me in the right direction?
<thomasd>ah yes, likely xfce itself creates that stuff on first login..
<civodul>i'd grep the code of Xfce package for references to XDG_CONFIG_HOME or similar
<civodul>yeah
<civodul>ACTION just posted a graft for the latest Bash CVE, which is quite scary
<catonano>thomasd: Hi !
<thomasd>hello catonano ;)
<thomasd>by the way, I decided to try out gnome-shell because it was available in GuixSD, and find it pretty smooth, and *much* nicer than unity.
<civodul>cool, i've never used it for real
<catonano>thomasd: I agree
<catonano>I also like its extensions
<thomasd>it's also much more snappy, but part of that might also be due to different disk encryption (?). On ubuntu, it was ecryptfs, while I'm using luks now.
<catonano>ah I can t really say. I know nothing about disk encryption
***snape` is now known as snape
<catonano>has anyone installed GuixSD on a Dell xps ? Can anybody confirm the hurdles described here ? https://lists.gnu.org/archive/html/guix-devel/2016-07/msg00550.html
<lfam>Is anyone able to access our Git repo at the moment? I'm getting 502 errors
<catonano>sneek later tell Ifraim: yes, Bad gateway, here
<sneek>Got it.
<catonano>sneek later tell lfam: yes Bad gateway here
<sneek>Got it.
<lfam>catonano: Thank you. Sorry for the intermittent connectivity. I'm on an airplane (!)
<sneek>Welcome back lfam, you have 1 message.
<sneek>lfam, catonano says: yes Bad gateway here
<catonano>lfam: be safe ;-)
<lfam>catonano: It's out of my control ;)
<lfam>Ah, there is a webkitgtk security update. Bad timing for the central repo to be offline
<civodul>hey lfam!
<civodul>lfam: is Savannah off-line?
<lfam>Hi civodul! The git repositories appear to be offline
<civodul>bah!
<civodul>nothing at https://quitter.se/fsfstatus
<lfam>I gave a message on #savannah
<civodul>let's see
<mekeor>Hello Guix <3
<sneek>mekeor, you have 1 message.
<sneek>mekeor, alezost says: I also prefer to blacklist pcspkr; you can do it like this: (kernel-arguments (list "modprobe.blacklist=pcspkr,snd_pcsp"))
<mekeor>nice
<civodul>/,
<civodul>'lo mekeor!
<mekeor>hi civodul (:
<mekeor>hm, there seems to be an issue with http://git.savannah.gnu.org/cgit/ , right? `guix pull` fails because that page is down
<sankey>mekeor: i'm getting "502 Bad Gateway" on that url
<mekeor>yeah
<jmi2k>I'm getting 502 Bad Gateway for git pull. Is there any problem with savannah?
<mekeor>jmi2k: i had the same issue. i reported a bug: https://savannah.gnu.org/support/index.php?109254
<jmi2k>mekeor: oh ok. At least it's not a problem of my connection. Thanks ;)
<civodul>apparently the FSF admins are working on it
<civodul>at least they should be aware of it
<civodul>SSH access works though
<mekeor>civodul: how do you know the admins are working on it?
<mekeor>oh well, it has been fixed actually
<mekeor>see https://savannah.gnu.org/support/index.php?109254#comment1
<civodul>good :-)
<jmi2k>That was fast!
<civodul>all i knew was that lfam reported the problem on #savannah
<snape>I'm not sure what is the best way to share scheme code between marionette-vm and host.
<snape>Like, I would run the code run by marionette-eval on host too, but without repetition
<snape>should I use eval? I heard it is bad programming style
<snape>or should I put the code in the store?
<snape>but then, I have to use 'system' to execute the code, and system only returns a number, which is not practical
<snape>Or a macro?..
<snape>Oh, I think I know
<snape>I could put the code in a module
<snape>... well I don't know
<snape>oh yes, I put the code in a module, and I do (use module ...)
<snape>I think that'll work
<civodul>"eval is evil"
<civodul>snape: the marionette service takes a list of modules to import IIRC
<snape>civodul: actually, the code I want to share is 'wait-for-file'
<snape>since I want to wait for a file in the host
<snape>I'd like it to be in a module, but the problem is: I should change it because it should be generic and not call marionette-eval, I think
<mvb>civodul: the bash update seems to cause a full rebuild
<civodul>mvb: i don't think so; what makes you say so
<civodul>?
<mvb>it downloads the bootstrap binaries and then builds guile..but maybe not much else is affected
<mvb>oops, false alarm..the bootstrap stuff was only needed to build the grafted bash
<davexunit_>I see we've started encouraging /usr/bin/env :(
<davexunit_>reproducibility will suffer as a result.
<slyfox>why?
<bavier`>encouraging?
***kelsoo1 is now known as kelsoo
***kelsoo1 is now known as kelsoo
<roptat>hi :)
<civodul>hey hey!
<civodul> https://debbugs.gnu.org/cgi/pkgreport.cgi?pkg=guix-patches
<civodul>who wants to be the first person to send a patch to guix-patches@gnu.org? :-)
<paroneayea>civodul: out of curoisity, why not guix-packages?
<paroneayea>is this meant for both features and packages?
<paroneayea>civodul: glad to see the new debbugs though :)
<civodul>paroneayea: good point! i don't know
<civodul>that's not a great answer
<civodul>oh well!
<paroneayea>civodul: we should be sure probably
<paroneayea>civodul: I think it can go one of two ways:
<paroneayea> - bug-guix for bugs only, not features; guix-patches for features and packages
<paroneayea> - bug-guix for bugs and features; guix-packages for packages only
<paroneayea>personally, the latter makes a big more sense to me, because it gives a constrained scope to the new list, and bug-guile handles both bugs and features I think
<bavier`>it seems useful to me to separate features out from bugs
<paroneayea>but as long as we make a distinctoin
<paroneayea>bavier`: ok! well I think it's fine as long as we know and are clear
<civodul>bugs should definitely go to bug-guix only
<civodul>the other list should definitely receive packages, but maybe that's fine if it receive features too
<paroneayea>ok :)
<paroneayea>I think it probably doesn't make sense to have three lists :)
<civodul>i think features can go to guix-devel or guix-patches, either way is ok
<paroneayea>so why not features and packages to guix-patches
<paroneayea>yes
<paroneayea>makes sense to me
<civodul>yeah
<paroneayea>things that are short and discussion oriented can go to guix-devle
<paroneayea>-devel
<civodul>right, that's what i thought
<paroneayea>and things that need to be tracked
<paroneayea>to guix-patches
<civodul>yes
<paroneayea>it sounds good to me.
<civodul>cool!
<civodul>well if that creates confusion or is inappropriate, we can always adjust
<civodul>but i think that should be fine :-)
<paroneayea>sounds good :)
<paroneayea>yay!
<civodul>anyway, someone needs to send a patch now!
<paroneayea>maybe I can see if angband compiles with the suggestion someone made
<bavier`>I might have some patches to send in a few hours
<paroneayea>ACTION patches up angband
<slyfox>\\o/
<paroneayea>goodbye productivity!
<paroneayea>assuming it compiles this time, that is :)
<paroneayea>okay, yup
<paroneayea>seems to work
<paroneayea>I'm doomed!
<slyfox>when guix fetches substitutions from internet it tells you current speed, unpacked file size but not actual size of downlaoded archive(s). why? is it hard to know upfront or it's easy to get and just not printed?
***jonsger1 is now known as jonsger
<slyfox> http://dpaste.com/0GBNM6N.txt to the very end it was not clear if it was 90MB archive or 30MB one
<bavier`>slyfox: iirc, the compressed archive size is not really know upfront, since compression is done on-the-fly
<civodul>exactly
<bavier`>slyfox: in this case, 36MiB was transferred, but occupies 91.9MiB once unpacked into the store
<slyfox>oh, ok that makes sense
<paroneayea>nice
<paroneayea>okay yep
<paroneayea>working angband package.
<paroneayea>yikes! :)
<paroneayea>is there something like modify-phases for arguments?
<paroneayea>in case I want to change one argument for a variant
<bavier`>paroneayea: substitute-keyword-arguments
<bavier`>paroneayea: in (guix utils) I believe
<paroneayea>right!
<paroneayea>I knew there was something.
<paroneayea>thanks bavier`
<paroneayea>ok!
<paroneayea>civodul: first package added to guix-patches@gnu.org, hopefully :)
<paroneayea>civodul: will guix-patches be subscribable via mailman similar to bug-guix?
<paroneayea>I don't see a mailman list
<slyfox>for some reason after 'guix pull' 'guix environment --dry-run ghc' tries to build guile instead of printing pretend list
<civodul>paroneayea: seems like the list doesn't exist yet, although Savannah thinks it does
<civodul>not sure why
<paroneayea>civodul: hm!
<civodul>slyfox: could be due to the Bash fix and the fact that hydra is not up-to-date yet
<slyfox>yeah. it built bash right before that
<paroneayea><guix-patches@gnu.org>: host eggs.gnu.org[208.118.235.92] said: 550-Callout
<paroneayea> verification failed: 550 550 Unrouteable address (in reply to RCPT TO
<paroneayea> command)
<paroneayea>civodul: ^^^ in reply to my patch
<paroneayea>so I guess the mailing list is not really set up right