IRC channel logs

2019-01-12.log

back to list of logs

<ArneBab>while I’m talking packaging: I have a working package for openfortivpn: https://paste.debian.net/1059840/
<ArneBab>(but no fully functioning email setup yet)
<ArneBab>… and the libaom build worked …
<bavier>nice
<ArneBab>but when building ffmpeg, I get aom >= 1.0.0 not found using pkg-config
<ArneBab>ok, found it, libaom v1.0.0-errata1 actually claims to be version 0.1.0
<ArneBab>yay!
<ArneBab>YESS, ffmpeg with AV1 support! — now to read up on the contribution guidelines :-)
<ArneBab>guix lint does not find my new package, even though I’m in guix environment guix …
<vagrantc>./pre-inst-env guix lint PACKAGE ?
<ArneBab>arg, yes, missed the pre-inst-env — thank you!
<ArneBab>patch sent — I hope i did everything right!
<ArneBab>… time to sleep :-)
<ArneBab>happy hacking!
***jonsger1 is now known as jonsger
<apteryx>ArneBab: cool :-)
<jonsger>ArneBab: I think https://code.videolan.org/videolan/dav1d could be even better then libaom, as it's really created for real usecases. Not only as implementation to show that it works...
<marusich>Hi Guix!
<sneek>marusich, you have 1 message.
<sneek>marusich, roptat says: I'd like to reply to you but since I'm abroad gmail rejects my replies :/
<marusich>roptat, do you still need anything?
<nly>i can build i686-linux system drvs on my machine using `guix system build --system="i686-linux"" config.scm`, but `guix system init` has no `--system=` flag, how do I install this system to a disk?
<nly>nope sorry
<nly>it was my fault
<ArneBab>my git complains that it is missing git send-email, do I have to add something?
<ArneBab>apteryx: :-)
<ArneBab>sneek: later tell jonsger: dav1d is a pure decoder, but I’m interested in *encoding* performance. Though I’ll need a decoder to actually see the quality :-)
<sneek>Got it.
<ArneBab>sneek botnsack
<ArneBab>sneek: botnsack
<ArneBab>sneek: botsnack
<sneek>:)
<janneke>ArneBab: install the git:email along with git
<janneke>git:email *output
<ArneBab>sneek: later tell jonsger: just found that dav1d is already packaged
<sneek>Okay.
<jonsger>oh fine ArneBab :)
<sneek>jonsger, you have 2 messages.
<sneek>jonsger, ArneBab says: dav1d is a pure decoder, but I’m interested in *encoding* performance. Though I’ll need a decoder to actually see the quality :-)
<sneek>jonsger, ArneBab says: just found that dav1d is already packaged
<ArneBab>jonsger: ah, sorry, did not see that you came in :-)
<jonsger>:)
<janas>When I start my laptop I need to manually type in 'sudo modprobe wacom' to activate the wacom driver - can that be automated in my config.scm?
<janas>actually nevermind I don't think that will actually solve my problem
<janas>dmesg is still not showing anything
<ArneBab>ffmpeg-3.4 currently inherits ffmpeg, but with an added configure flag for ffmpeg it fails. Can I remove the configure flag?
<ArneBab>(essentially: inherit ffmpeg, and remove one of the inherited flags)
<g_bor>hello guix!
<OriansJ>Oh I found a bug in Guix, where if you run guix package guix environment --ad-hoc icecat --fallback in one shell and guix package -r xfce4-terminal in a second shell; both fail with unable to obtain lock after about 6 hours.
<adfeno>Hi there
<janneke>OriansJ: does it read like #31785?
<OriansJ>janneke: why yes, yes it does
<janneke>OK...still bad tho
<janneke>otherwise i'd have asked you to mail bug-guix :-)
<OriansJ>yet, what would they have to lock on? as one is creating an ad-hoc environment and the other just needs to make a symlink forest
<janneke>OriansJ: i hope that civodul has a clue...
<adfeno>sneek: laterl tell civodul: Re: bug #33993: Missing Texinfo documentation for Guix itself: I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in .GZ format.
<sneek>civodul:, adfeno says: Re: bug #33993: Missing Texinfo documentation for Guix itself: I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in .GZ format.
<adfeno>?
<adfeno>sneek: laterl tell civodul Re: bug #33993: Missing Texinfo documentation for Guix itself: I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in .GZ format.
<sneek>civodul, adfeno says: Re: bug #33993: Missing Texinfo documentation for Guix itself: I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in .GZ format.
<adfeno>Just a test now.
<sneek>adfeno, you have 1 message.
<sneek>adfeno, adfeno says: Test
<adfeno>sneek: laterl tell civodul bug 33993 Missing Texinfo documentation for Guix itself - I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in GZ format.
<sneek>civodul, adfeno says: bug 33993 Missing Texinfo documentation for Guix itself - I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in GZ format.
***rekado_ is now known as rekado
<rekado>adfeno: “laterl” should be “later”
<adfeno>Oh.
<sneek>Welcome back adfeno, you have 1 message.
<sneek>adfeno, adfeno says: bug 33993 Missing Texinfo documentation for Guix itself - I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in GZ format.
<adfeno>sneek: later tell civodul Re: bug #33993: Missing Texinfo documentation for Guix itself: I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in .GZ format.
<sneek>Will do.
<adfeno>Thanks
<adfeno>Gess I hate tiny font sizes
<adfeno>Oh, on time :D
<adfeno>Welcome civodul :D
<civodul>hello!
<sneek>Welcome back civodul, you have 1 message.
<sneek>civodul, adfeno says: Re: bug #33993: Missing Texinfo documentation for Guix itself: I made the test ina QEMU VM as I described, I have the work flow writen down and the results which confirm that at least if one downloads using guix-install.sh (without pulling), the Texinfo files will be in .GZ format.
<adfeno>civodul: I didn't test the commit you referenced yet
<adfeno>If you want, I can reopen the bug and psot the workflow just so that this is at least preliminarly documented in the mailing list
<adfeno>and bug tracker
<civodul>adfeno: i'm not sure i follow, but please write your feedback in the bug report, that's easier to keep track of!
<civodul>the patch queue is growing worryingly, gentlefolks!
<roptat>hi guix!
<roptat>I think I'm very close to having sbt built from a binary scala
<civodul>yay roptat
<joshuaBPMan>Hello, how can I turn down the volume of my computer via the command line? alsamixer is not installed or packaged?
<pkill9>joshuaBPMan: alsamixer is in alsa-utils
<rekado>joshuaBPMan: it’s provided by alsa-utils-
<rekado>oops
<joshuaBPMan>pkill9: thanks
<janneke>i'm getting: sha256 hash mismatch for /gnu/store/ga61vgzqpcfz81d7lxi40ikrzffazm05-gcc-5.5.0
<janneke>expected hash: 0pbxid8shfcbwbl8fl6h4a4r674730l5rslmx5krxnklxhs7p6xh actual hash: 1929310w2cajgraii435frv5q763ajnnl5xk8hqp420s3y4c47q0
<davidl>Im tring to define a simple package but it errors with make: cc: command not found. Not sure what the correct way to solve it is.
<nly>amixer set Master 80%
<OriansJ>davidl: you need to include it in your dependencies
<OriansJ>or use (build-system gnu-build-system)
<OriansJ>and (gnu packages gcc)
<adfeno>civodul: Re: bug #33993: Missing Texinfo documentation for Guix itself: If I do send the mesage to the bug I'll re-open it.
*janneke just runs guix pull
<adfeno>civodul: I'll do guix pull in my copy, to see if the bug is fixed
<nixo_>anybody is using the mpd service on guixsd? man mpd.conf says that db_file is required, but it's not present in the guix-generated config. Running mpc update says "mpd error: no database"
<adfeno>On another subject: I noticed that `guix describe' behaves differently than the one available in the 0.16.0 release.
<rekado>nixo_: I used it but found it not flexible enough.
<rekado>when using mpc update you may need to tell it where the config file is.
<janneke>hmm, is there a friendlier way to get i686-unknown-linux-gnu-gcc than doing: guix environment --ad-hoc -e '(begin (use-modules (gnu packages cross-base)) (list (cross-binutils "i686-unknown-linux-gnu") (cross-gcc "i686-unknown-linux-gnu")))'
<davidl>OriansJ: I am using build-system gnu-build-system but there is no configure step in the source so I have removed the configure step and cc is not symlinked to gcc.
<OriansJ>davidl: in the (use-modules section you included (gnu packages gcc) right?
<janneke>rekado: the sha mismatch comes from berlin when I run: guix environment --system=i686-linux --ad-hoc gcc@5
<OriansJ>davidl: you can look at the mescc-tools package definition as it uses gcc but doesn't have a configure step
<nixo_>rekado: thanks, the NixOS version was complete enough, maybe I'll improve guix one
<davidl>OriansJ: yes I did. It's the same error when making from bash cli unless I first run export CC=gcc before running make.
<davidl>OriansJ: thx Ill check it out.
<OriansJ>davidl: does your makefile have the line: CC=gcc
<davidl>it has CC?=GCC
<davidl>OriansJ: I guess I can remove the '?' before the make phase.
<davidl>OriansJ: yeah that worked. See if I can figure out how to do it in the package definition.
<rekado>nixo_: the configuration is very limited and doesn’t have any knobs to configure the audio outputs, for example.
<rekado>nixo_: I would be very happy if you could improve it :)
<nixo_>is emacs-magit broken? I get Cannot open load file: No such file or directory, graphql
<rekado>yes, it was broken, but the offending commit has already been reverted.
<rekado>you may need to use guix pull.
<nixo_>rekado: thanks again
<adfeno>Re: `guix describe': Well the good thing is that `guix pull --list-generations' does the same thing
<adfeno>... as the 0.16.0 `guix describe' would do.
<adfeno>At least in regards to describing the commit a pull was based on.
<pkill9>something confuses me: qtox has as an input ("qmake" ,qttools) but qmake isn't in the qttools package
<bavier`>it may have been there at one point
<bavier`>does qtox fail to build as a result?
<pkill9>nope it builds
<pkill9>i forgot to mention that
<adfeno>Re: `guix describe': I found the issue: I did not add ~/".config/guix/current/bin" to $PATH.
<adfeno>Now it works :D
<bavier`>I noticed that on my gnome desktop, several apps don't show their icons in the gnome-shell
<bavier`>these apps seems to be those that provide only svg icons
<bavier`>handbrake, tint2, and gamine
<OriansJ>bavier: convert -density 1200 -resize 200x200 source.svg target.png ?
<civodul>janneke: regarding the hash mismatch: which server, commit, etc.?
<janneke>civodul: current master (146c6b62bf72abbe710ce29972a8db17c6285b54)
<davidl>finally managed to package my first package! :D
<davidl>was able to install it with guix package -f mypackage.scm
<davidl>will see if I can contribute it to guix. The package is prips.
<davidl>a tool for ip calculation.
*janneke runs the command again to get the url
<janneke>civodul: downloading from https://berlin.guixsd.org/nar/gzip/ga61vgzqpcfz81d7lxi40ikrzffazm05-gcc-5.5.0...
<janneke> http://paste.debian.net/1059973/
<civodul>janneke: should be fixed now, per "wget -q -O - https://ci.guix.info/nar/gzip/ga61vgzqpcfz81d7lxi40ikrzffazm05-gcc-5.5.0 |gunzip -c | guix hash -", which returns 0pbxid8shfcbwbl8fl6h4a4r674730l5rslmx5krxnklxhs7p6xh
<civodul>probably remnants of the discrepancies we got when migrating to the new storage device
<civodul>esp. since i686 is less used
<civodul>anyway, thanks for the heads-up!
<janneke>civodul: thank you!
<efraim>rekado, bavier`: sorry about core-updates and broken python2, I didn't test python[23] when I was pushing updates
<ArneBab>why are we using GCC 5.5 and not a modern version?
<bavier`>ArneBab: I think there's some work going on to move up to version 7, iirc
<ArneBab>nice!
<bavier`>some packages use more recent versions already, but moving the default version is difficult because of compatibility and such
*jonsger wishes that we have already gcc 6 or 7 as default...
<jackhill>speaking of which, how do I tell what the defualt gcc is?
<civodul>jackhill: you can run, say: guix gc -R $(guix build coreutils -d) | grep gcc
<jackhill>civodul: thanks!
<janneke>jonsger: we have a wip-gcc7 that mbakke has been working on
<janneke>*branch
<mbakke>Unfortunately cross-compiling is still broken in the GCC7 branch :/
***lostcoffee is now known as atw
<civodul>oh
<mbakke>I'll rebase and push the latest version in a day or three.
*rekado updates CRAN / bioconductor packages
<roptat>so I wasn't so close to build sbt after all
<roptat>I'm still closer, but I found a dependency that's going to be difficult I think
<rekado>who is the core-updates guardian this time?
<rekado>I tried rebasing wip-gnome3.30 on top of core-updates and wasn’t sure about some of the conflicts.
<rekado>I would be very grateful if someone could merge master into core-updates again.
<nixo_>how is the state of rust in guix? I wanted ripgrep (rg) but it requires lots of dependencies (and guix import crate does not support the recursive param)
<nixo_>same for ea
<nixo_>*exa
<nixo_>it's just that nobody is using them or there is some caveat with rust?
<bavier`>nixo_: rust in guix works
<bavier`>nixo_: there are several packages that use it successfully, notably icecat
<mbakke>Though I believe the Cargo build system is still rather primitive.
<nixo_>bavier`: thanks, so I guess I should just add all the dependencies until ripgrep compiles