IRC channel logs


back to list of logs

<civodul>lfam: beware of search results :-)
<civodul>the order is sometimes disconcerting
<lfam>Right... but still there is no substitute :/
<lfam>I want to make a union of alsa packages. I'm wondering how to specify the mutiple outputs of alsa-plugins in the list of packages to pass to my alsa-union procedure
<pkill9>i think there's a union build procedure now. as for specifying the output, I think you can just add it to the input like so: ("package" ,package "output")
<pkill9>oh, there's a gexp procedure for create directory unions, not onje to create a pakcage union
<lfam>I'm working with (guix build union)
<lfam>Copying from sdl-union
<lfam>So, I've got this:
<lfam>But, it only uses the primary "out" output of alsa-plugins
<lfam>It would be nice for the default to include all of them
<vagrantc> doesn't appear to be picking up new commits?
<sirgazil>I'm trying the system with core-updates now and so far things are working alright. No new bugs, except for some kind of button that I think is supposed to look like three vertical dots or square, but looks like a stain instead.
<sirgazil>In the wifi configuration, in Region and language (after pressing the + button), it looks similar.
<sirgazil>I'm hopping GNOME 3.34.2 fixes the gnome-shell leak. I'll keep testing this weekend.
<sirgazil>Oh, GNOME Web icon looks weird. It seems as if they forgot to clip the land in the globe...
***jonsger1 is now known as jonsger
***catonano_ is now known as catonano
<guix-ci>Problem: Too many processes on hydra-guix-125 Problem started at 02:58:21 on 2020.04.25
<guix-ci>Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on 2020.04.25
<vagrantc>hydra's back?
<dsmith-work>Heya guix people. Was there any messages around 21:09 utc? Maybe to the bot?
<dsmith-work>The logs don't have timestamps...
<dsmith-work>That would be less that 20 minutes ago, I think.
<dsmith-work>Like 19 minutes
<atw>dsmith-work: I think that on eg, if you click on someone's nick, you get a URL like …#hhmmss, ie hours, minutes, and seconds. I'm guessing though, I don't know for sure
<dsmith-work>Ok. Will check.
<apteryx>dsmith-work: yes, there was: 21:08:40 guix-ci | Problem: Too many processes on hydra-guix-125 Problem started at 02:58:21 on 2020.04.25
<apteryx>and 21:09:40 guix-ci | Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on │ catern
<apteryx> | 2020.04.25
<dsmith-work>When did jonsger and catonano change nicks?
<raghavgururajan>Hello Guix!
<apteryx>raghavgururajan: o/
<dsmith-work>Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on 2020.04.25
<guix-ci>Resolved: Too many processes running on hydra-guix-125 Problem has been resolved at 03:21:23 on 2020.04.25
<atw>dsmith-work: 02:38 and 02:59, I think? based on the <div>s id?
<raghavgururajan>apteryx Any luck with moving linphoneqt to master?
<dsmith-work>It's the message sform guix-ci.
<guix-ci>Resolved: Too many processes on hydra-guix-125 Problem has been resolved at 03:22:21 on 2020.04.25
<dsmith-work>apteryx: Thanks muchcly!
<apteryx>raghavgururajan: no! But I saw mbakke requesting to try out core-updates, so hopefully core-updates will be integrated into master soon
<apteryx>dsmith-work: thanks for working on that bot, or fixing whatever problem there was with hydra
<dsmith-work>Working on the bot.
<apteryx>OK :-)
<raghavgururajan>apteryx Cool!
<apteryx>raghavgururajan: by the way, I noticed that liblinphone should have its tester binaries and data moved to another output; just the test data weights 11 MiB
<apteryx>like I did for mediastreamer2
<raghavgururajan>apteryx I thought I disable tester binaries. No?
<dsmith-work>Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on 2020.04.25
<dsmith-work>sneek: botsnack
<raghavgururajan>apteryx Anyway, I will check on that.
<dsmith-work>sneek: botsnack
<dsmith-work>Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on 2020.04.25
<dsmith-work>sneek: seen civodul?
<sneek>civodul was in #guix 3 hours ago, saying: the order is sometimes disconcerting.
<apteryx>raghavgururajan: ah, perhaps it was enabled following review, seeing all the dependencies were already satisfied. I think it's nice to have them around, given the package has no test suite.
<raghavgururajan>apteryx Cool!
<dsmith-work>Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on 2020.04.25
<dsmith-work>sneek: seen civodul?
<sneek>civodul was in #guix 4 hours ago, saying: the order is sometimes disconcerting.
<dsmith-work>Problem: Too many processes running on hydra-guix-125 Problem started at 02:59:23 on 2020.04.25
<dsmith-work>sneek: botsnack
*raghavgururajan is in love with Xfe File Manager. (
<bandali>nice, looks handy
<apteryx>eh, updatedb uses more than 2.5 GiB of RAM
<apteryx>2.6 now. Is this normal?
<apteryx>3 GiB. The memory is used by the find process it spawns.
<apteryx>bah, terminated by earlyoom. seems a bug!
<guix-ci>Problem: Too many processes on hydra-guix-105 Problem started at 06:11:44 on 2020.04.25
<raghav-gururajan>bandali I recommend you to try it. :-)
<guix-ci>Problem: Too many processes on hydra-guix-113 Problem started at 06:17:28 on 2020.04.25
<guix-ci>Resolved: Too many processes on hydra-guix-105 Problem has been resolved at 06:17:44 on 2020.04.25
<apteryx>dsmith-work: looks like the number of process is oscillating around the limit
<guix-ci>Resolved: Too many processes on hydra-guix-113 Problem has been resolved at 06:21:28 on 2020.04.25
<apteryx>dsmith-work: also, is there a reason why the bot can't permanently stay connected instead of joining, messaging, then leaving?
<emys>hi, I bumped two versions of Python dependencies like so
<emys>when I try to contribute patches for these, would you rather have two mails with patch for each package individually or one mail with both patches?
<catonano>dsmith-work: did I change my nickname ? I didn't realize that
<catonano>emys: I need an account in order to see your link
<emys>catonano, hmm, can you see and
<catonano>emys: yes, I can see those, thanks
<catonano>emys: I think you can send one email
<catonano>emys: just one minor comment:
<catonano>on the pylonsproject patch, on line 1726
<catonano>emys: there's a cr before the actual hash. I would leave the same formatting as before
<catonano>it's a minor thing but some people appreciate uniformity in formatting
<emys>I see, yes
<emys>no problem to preserve formatting
<emys>now I only need to get git send-email to work .. *sigh*
<guix-ci>Problem: Too many processes on hydra-guix-126 Problem started at 09:34:04 on 2020.04.25
<catonano>emys: that's not strictly necessary. You can attach your patches to an email message manually
<catonano>emys: I did so several times
<guix-ci>Problem: Too many processes running on hydra-guix-126 Problem started at 09:35:06 on 2020.04.25
<catonano>emys: you can fight with git-sendmail later 🤷‍♂️️
<emys>catonano, yeah I know but I'd like to have that set up once and for all :)
<emys>catonano, thanks for your help, patches are out via my normal email client now.
<catonano>emys: 😊️
<guix-ci>Resolved: Too many processes on hydra-guix-126 Problem has been resolved at 10:03:04 on 2020.04.25
<guix-ci>Resolved: Too many processes running on hydra-guix-126 Problem has been resolved at 10:03:06 on 2020.04.25
<jonsger>dsmith-work: what?
<emys>is guix-patches moderated? Wondering whether moderation approval is pending or whether there was a problem with my email
<TZander>emys: did you check ?
<jonsger>emys: what is your mail address?
***wxie1 is now known as wxie
<ecbrown>emys: i seem to remember my first patch taking a while to appear, but subsequent patches were registered almost immediately
<rekado_>…why does dsmith-work speak like guix-ci?
<rekado_>sneek: later tell vagrantc no, hydra-guix-* is the name of those build nodes behind that are hosted at the MDC
<sneek>Will do.
<PotentialUser-91>Has anyone setup guix with StumpWM and if so is he willing to share config files?
<dissoc3>PotentialUser-91: i got stumpwm to work but i dont really know what im doing
<dissoc3>i just added stumpwm to packages in /etc/config.scm
<alextee[m]>which package provides "dot"?
<ecbrown>graphviz i believe
<ecbrown>i wonder if i've gotten "off the rails" with my package manager. i'm guix pull'ing from master, e.g. substitutes seem available, but guix pull && guix package -u seems to be rebuilding everything from source. (it does work as expected, though.)
<ecbrown>(this machine is running 5.6.7 with non-free radeon)
<ecbrown>odd my very similar laptop (without the radeon) finds substitutes just fine
<ecbrown>full disclosure: i did visit the core-updates branch, but coming back to master
<leoprikler>ecbrown: even on master substitutes are not always available on fresh `guix pull`s
<alextee[m]>ecbrown: thanks
<leoprikler>the non-free stuff also has no substitutes anyway
<ecbrown>sure, just annoyed at having to compile all the rust's
<ecbrown>maybe it will be back on track once this particular update completes
<dsmith-work>rekado_: I was using that as a test input to the bot. For some reason, it was freaking out on it.
<janneke>with a backtrace like this:
<janneke>is there anything i can do except start guessing and adding print statements?
<dsmith-work>apteryx: The the bot is joining/leaving all the time becuse I'm starting it.
<dsmith-work>Sorry for the noise.
<cbaines>janneke, in my limited experience at least, you have to avoid Guile evaluating code on the fly, as that seems to get it thinking that all errors occur in the evaluator
<cbaines>assuming you've run make to generate the .go files for Guix, I guess that error is occuring when evaluating the bare-hurd.tmpl file
<janneke>cbaines: thanks, at least that helps to change my perspective on it -- hmm
<janneke>it's definately related to my bare-hurd effort -- ... but where :-)
<janneke>i'll see if i can somehow process/invoke/run smaller chunks of the opaque guix system call
<cbaines>it does seem odd to have an interpreted language (Guile), where the first rule is to avoid interpreting code at runtime, as it means the stack traces are unintelligible!
<cbaines>janneke, have you tried just running the bare-hurd.tmpl file with Guile?
<janneke>:-) oh, /me goes to try
<janneke>that works...adding PK to the resulting os
<janneke>ah, bootloader #f -- i'm running with --no-bootloader, but forgot about that
<janneke>and also linux firmware :-/
<dsmith-work>Testing 😊️
<thvk-ivgf>query sneek
<reepca`>so today I learned that there are two completely different software projects called "libtorrent", both providing similar features and used by similar programs. And neither is a fork of the other AFAIK. I only found this out at the bottom of a github wiki page after 30 minutes of confused searching.
*reepca` gained a newfound respect for projects renaming to avoid confusion
<stikonas_>well, gentoo calls one package libtorrent-rasterbar and the other libtorrent
***stikonas_ is now known as stikonas
*ecbrown memorializes in search engines: guix offload problem? make sure user on remote machine is not using non-default shell, e.g. not using zsh. create another user for offload build
<rekado_>ecbrown: does “guix offload” print a good error message in that case? If not, could you please submit a bug report?
<ecbrown>rekado_: happy to do so. it's a bit of a cryptic message
<dlowe>if I guix install guile3.0-guix will it work?
<cbaines>dlowe, that's only really useful if you're using guix as a Guile library
<cbaines>in which case, yes, it will work
<cbaines>(well, or if you have a package that depends on guix)
*ecbrown busts out the second t420 in order to confirm gnome wifi connection password bug in core-updates
<ecbrown>i accumulate t420's at the earth day dumpster
<dlowe>cbaines: hm, okay. thanks
<ecbrown>it's a shame they require non-free firmware for wifi
<Bronzu>Hi, I'm having issues running wayland.
<Bronzu>It says "User has no sessions"
<Bronzu>Couldnt find active session or greeter session
<Bronzu>cannot become drm master, do not have cap_sys_admin
<Bronzu>failed to load backe d
<rekado_>tried to upgrade R, but there are new reproducibility problems :(
<rekado_>Bronzu: how do you run wayland?
<Bronzu>Well I don't... I'm trying to get sway working.
<Bronzu>I don't even have any graphical ui installed (guess im regretting that now huh)
<guix-vits>Bronzu: `sudo herd status |grep login` -- elogind?
<Bronzu>nothing, guess I'll have to start a service?
<rekado_>Bronzu: are you trying to start it via gdm? AFAIU you should just start sway without gdm.
<rekado_>(I’m guessing based on you mentioning “greeter”)
<Bronzu>i dont have gdm, just thr shell
<guix-vits>Bronzu: elogind should be in your sys. config, i think; do you've the %desktop-services, or the %base-services?
<Bronzu>In config.scm i have base services
<rekado_>I think it would be good to have a cookbook for using sway here:
<Bronzu>i installed elogind
<Bronzu>is this where i put it?
<guix-vits>Bronzu: you need to add elogind to services
<rekado_>yeah, don’t install it into your user profile and run it from there. You’ll need the service.
*rekado_ needs to go
<Bronzu>ok ill try
<Bronzu>I did include it in services but herd still says it cant find it
<Bronzu>elogind that is
***benny is now known as Guest80763
<mbakke>Bronzu: did you reconfigure after adding (elogind-service) to your config?
<mbakke>Bronzu: the SDDM display manager is able to find and start Sway
*jonsger builds thunderbird yet another time :P
<civodul>jonsger: oh you have a Thunderbird package?
<jonsger>civodul: reviving some old patches (not from me). I think I'll send a first version the next days
<TZander>yeah, lots of unfinished packaging on the bug-tracker.
*ecbrown makes plug for sypheed patch, another gui email program
<mbakke>I want to use (package-with-relocatable-glibc ...) outside of (gnu packages make-bootstrap), but can't make up my mind whether to just export it, or move the relevant procedures to a new module (where?). Ideas?
<mbakke>exporting is suboptimal, because I want to use it in one of the modules imported by (gnu packages make-bootstrap), creating a circular dependency
<TZander>bootstrap stuff we typically just include in C++, which effectively means create two compiled copies. Even if the sources only exist once
<TZander>probably best to not change the main app design too much to adjust for singular exceptions like this
<kraai>Hi. I'm using GNOME. When I try to start Geary by selecting Activities > Show Applications > Geary, it doesn't appear. If I run geary in a terminal, it does. Does anyone know how to fix or investigate this?
<mroh>where can I put .patch files in a channel, so that (package (source (origin (patches (search-patches)))) finds them?
*kmicu ಠ_ಠ at new Nix(OS) Marketing Team.
<drakonis>its working out so far
<mbakke>mroh: if your channel uses namespace (foo packages), with a subdirectory "patches", using (search-patches "foo/packages/patches/bar.patch") should work
<mroh>mbakke: ty!
<jonsger>kmicu: hm?
<cbaines>kraai, first you need to find out what the GNOME Shell is trying to launch. Do you have a file for Geary in ~/.guix-profile/share/applications ?
<cbaines>I wonder what the costs are of using grafts without substitutes are...? I think I'm seeing some wierd behaviour when I don't have a source of substitutes. Guix builds some seemingly unnecessary stuff, and I think it might just be to figure out something about grafts. If I disable grafts, it doesn't think anything needs building...
<emys>Trying to package some more python modules I get the following error: "ValueError: Unable to create the compiledir directory '/homeless-shelter/.theano/...".
<emys>it seems homeless-shelter is a sandbox $HOME by guix
<emys>is it write-protected?
<cbaines>No, but I don't think it exists in the build environment
<cbaines>This is helpful to highlight when packages expect it to be there
<cbaines>It seems wrong to me for a python module to be trying to use a users home directory when building/installing...
<emys>cbaines, it seems wrong to me, too, it is not my module though.
<cbaines>To work around wrong things like this, you can do thinks like (setenv "HOME" "/tmp")
<emys>cbaines, I already did, I just put it into the wrong package definition and didn't realize why it had no effect
<kmicu>(FYI homeless-shelter is a thing inherited from Nix.)
***Intensity4 is now known as Intensity
*kmicu would prefer a descriptive term like mathy ‘imaginary-directory’ more than a confusing pun 😺
<kmicu>(even ‘impurity-detected’ should be more helpful)
<catonano>cbaines: as for grafting inducing (apparently) pointless building, that's a debuggign problem. It's be important to see exactly how it visits the graph of packages and how it decides to build things. It's a tooling problem
<catonano>I not this because tooling insufficiencies usually impress me
<kraai>cbaines: There's a geary-autostart.desktop in ~/.guix-profile/share/applications.
<kraai>cbaines: There's also an org.gnome.Geary.desktop there.
<cbaines>kraai, I guess you could look in the org.gnome.Geary.desktop one?
<cbaines>I'm not sure how to launch a .desktop file from the command line...
<kraai>cbaines: There are two sections in that file, [Desktop Entry] and [Desktop Entry Compose].
<kraai>The Exec line in the Desktop Entry section is /gnu/store/x97linxpzknsjaypy7x3i6r078agb67j-geary-3.34.1/bin/geary %U.
<kraai>According to the spec, it looks like the %U should be stripped if no URL is specified.
<efraim>datefudge failed to build for me on core-updates (on powerpc)
<efraim>nvm, it seems to be a conflict with glibc-2.30
<guix-ci>Problem: Too many processes on Zabbix server Problem started at 21:21:12 on 2020.04.25
<guix-ci>Resolved: Too many processes on Zabbix server Problem has been resolved at 21:24:12 on 2020.04.25
<civodul>mbakke: package-with-relocatable-glibc could be in (gnu packages base), next to glibc
<civodul>you sure you want to use it though? :-)
<civodul>it's a bit of a strange hack
<civodul>heavyweight and all
<mihi>Hello. Trying to follow, but always during the build (my machine is not very fast) the wip-hurd-vm branch seems to get rebased so that the process dies in the middle with "reference is not a tree: cd3193f..."
<sirmacik>does anyone have working config.scm with current kernel and wireguard?
<mihi>is there a branch that is more stable to do so?
<mihi>also, I guess it is normal that after updating from git I always have to run make before pre-inst-env guix build is able to give it another shot?
<vagrantc>mihi: it will usually produce some warnings about, but it should work as long as you haven't run "guix gc" clearing out the version that was used to run the initial make
<sneek>Welcome back vagrantc, you have 1 message!
<sneek>vagrantc, rekado_ says: no, hydra-guix-* is the name of those build nodes behind that are hosted at the MDC
<vagrantc>sneek: thanks
<vagrantc>sneek: botsnack
<vagrantc>so particular.
<mbakke>civodul: I needed package-with-relocatable-glibc to create a 'guile-static@3.0' for the initrd... I think I have it working now, will clean up and submit the patches tomorrow. :-)
<mbakke>I first tried without it, but the initrd failed to look up 'getpw'.
<bricewge>sirmacik: Since you asked about it, I have. There is a patch waiting to be merged about simplifying the process (ie. using lodable module).
<seepel>Hi guix! I'm trying to figure out how to compile a rust project I found on github, but I cannot for the life of me figure out how to get cargo in my path.
<raghavgururajan>Hello Guix!
<raghavgururajan>sneek, later tell lfam: Any update regarding #40659 (v3) please? Thanks!
<sneek>Got it.
<raghavgururajan> sneek, later ask civodul: Would you be able to push #40785 please? This is regarding `guix edit` and nano. Thanks!
<sneek>Will do.
<vagrantc>anyone else seeing some alarming messages during boot on core-updates?
<vagrantc>things like: WARNING: (#{ g115}#): imported module (guix build utils) overrides core binding `delete'
<vagrantc>;;; ERROR: In procedure make_objcode_from_file: bad header on object file: "\x7fELF\x02\x01\x01ÿ\x00\x00\x00\x00\x00\x00\x00\x00"
<vagrantc>saw on both aarch64 and x86_64
<vagrantc>(more-or-less similar, at least, only have details for aarch64)
<jackhill>vagrantc: yes, I think I saw those, but they scrolled fo the screen before I could read them.
<vagrantc>serial console, FTW!
<vagrantc>hence, why i only have them on aarch64
<sirmacik>bricewge: thx i'll test it tomorrow
<rekado_>the second one looks like a mismatch in Guile versions
<rekado_>the warning is new with Guile 3, I think
<vagrantc>i guess i could comment on the thread with a little more detail?