IRC channel logs

2019-09-16.log

back to list of logs

*civodul posts a news distribution mechanism for 'guix pull': https://issues.guix.gnu.org/issue/37413
<civodul>oh, internal error
<civodul>anyway it's pretty fun
*civodul can now go to bed and sleep well
<civodul>good night!
***Guest80997 is now known as cap
***cap is now known as Guest12753
<quiliro>Hello. If a person has IRC access blocked by the ISP, is there a way to circumvent this censorship? Tor? Proxy?
<erudition>Yup
***Guest12753 is now known as cap
***cap is now known as Guest58982
<quiliro>erudition: both or either?
<erudition>Yeah, either can do the trick
<quiliro>erudition: thanks
***Guest58982 is now known as cap
***cap is now known as Guest33021
***Guest33021 is now known as cap
***cap is now known as Guest51317
***atw` is now known as atw
***Guest51317 is now known as cap
***cap is now known as Guest28879
***Guest28879 is now known as cap
***cap is now known as Guest32353
***MinceR_ is now known as MinceR
***Guest32353 is now known as cap
***cap is now known as Guest90683
<nee`>PotentialUser-25 kmicu the last php-fpm news is probably from 2011, because it was incorperated into php itself. News about php-fpm are in the php release NEWS. php-fpm It keeps warmed up php processes around to make quicker responses when php pages are requested.
<nee`>The php-fpm executable is in the php package
***Guest90683 is now known as cap
***cap is now known as Guest83429
<civodul>Hello Guix!
<wigust->hello
<roptat>hi guix!
<jlicht>hey guix!
<civodul>hey hey!
<civodul>kudos for http://data.guix.gnu.org !
<ng0>hi
***Guest83429 is now known as cap
***cap is now known as Guest55852
<ng0>a question to you as distributors of linux-libre: do you ship with net80211/ieee80211_radiotap.h or a file which defines at least the struct ieee80211_radiotap_header?
*civodul has no idea
<civodul>i'd recommend grep -r ieee80211_radiotap_header $(guix build -S linux-libre) :-)
<ng0>maybe someone can grep in what is /usr/include on my system, the most I have is linux somewhere out there
<ng0>I'm writing this from NetBSD, and my server which has Guix in a VM is crashed right now
<civodul>ah well, you're on #guix tho :-)
<ng0>yes :) but I'm asking as a developer, to see what my approach to the file someone in our project bundled as the linux version ~10 years ago could be
<ng0>I have several ideas, but first I want to check how common the file is now
<ng0>I know the file exists on NetBSD, and that is also in FreeBSD (because of the check in line in the file) and that except for the copyright there is almost no significant difference to the one we bundle. which made me look into this, because I don't like bundling stuff
<ng0>it's in the sourcecode of main linux in net/, so I think it depends on a specific kernel configuration
<ng0>I'll know more once I can use a browser again
<civodul>ng0: https://paste.debian.net/1101005/
<civodul>it's GaaS (grep as a service)
<ng0>thanks
<ng0>what I learned only now is that this originated in NetBSD and was imported into Linux radio stack a while back, so the similarity makes sense :)
<civodul>heh
***Guest55852 is now known as cap-
<gnu_srs>Hello, when building guix from source is the tarball 1.0.1 or git to be preferred?
***cap- is now known as cap
***cap is now known as Guest48036
<civodul>gnu_srs: i'd recommend Git
<gnu_srs>k!
<g_bor[m]>civodul: i have two questions. 1. How should we coordinate on the issue we are mailing with Outreachy? Should we move the thread to some public ml, or should the maintainers for an appropriate plan? 2. The mail abot the RB summit just fell into my mailbox. Should we start some coordination in the guix community, or everyone participating should do everything individually? I ask this because the location is a bit further away
<g_bor[m]>than usual for most of us.
<ng0>where is it this time?
***Guest48036 is now known as cap
***cap is now known as Guest81265
<iyzsong>is ci.guix.gnu.org down?
<Gamayun_>Hm, looks like it.
<iyzsong>okay, but the substitutes works, maybe only curiass is down.
<rekado>I restarted the web interface
<rekado>the web interface is separate from cuirass itself, and also separate from “guix publish”
<efraim>'guix pull' failing for others?
<tune>I'll check
<tune>ah yes after a few seconds it fails
<iyzsong>rekado: good to know, thank you!
<efraim>rekado: looks like 41ca406fa54e69f61c55b11ffe5cf465192a907c is the culprit
<civodul>efraim: yes, argh!
<civodul>oh you just fixed it, right?
<efraim>i pushed a revert
<civodul>awesome
<civodul>thanks!
<civodul>so the problem is that this commit pulled in (git) indirectly
<efraim>without tracking it down, i'd guess that (guix import cran) uses guile-git and (guix build-system r) doesn't
<roelj>Why does Guix install the systemd service file in the lib/ dir instead of in etc/?
<roelj>(the systemd service file for guix-daemon*)
<civodul>efraim: yes, exactly
<civodul>so the solution is to do it the other way around: have the importer depend on the build system
<civodul>or have an extra module used by both the importer and the build system
<civodul>roelj: it installs it under /etc/systemd/system, doesn't it?
<efraim>/gnu/store/2bzgj0wmndqiirz1mamnmhyfff68ki2z-guix-1.0.1-5.cc98b00/lib/systemd/system/guix-daemon.service
<civodul>ah yes, but that's not where guix-install.sh puts it
<civodul>lib/systemd is because we're under $prefix i think
<civodul>i think it's a widespread convention
<civodul>dunno
<efraim>disapointing failure, rust-rayon depends on rust-rayon-core, and explicitly linking rayon fails to find the (probably) embedded rayon_core code
<efraim>not exactly the same, but 'apt-file show syncthing' on debian includes: syncthing: /usr/lib/systemd/user/syncthing.service
<roelj>civodul: efraim: Thanks.
<efraim>and a non-user one: dovecot-core: /lib/systemd/system/dovecot.service
<rekado>efraim, civodul: oof! I’m sorry for breaking things!
<efraim>we caught it in under an hour, better than when I broke openssh-service :)
<civodul>interestingly, a colleague of mine was getting started with Guix precisely at that moment :-)
<civodul>but yeah, you were very quick efraim!
<civodul>i think we should have some server-side automation at some point to check commits on push
<civodul>we should do that with signatures
***Guest81265 is now known as cap-
<civodul>otherwise, when 'guix pull' checks those sigs, users might end up with unrecoverable errors
<efraim>I bet guix-data-service would've caught it, the pull would fail
<civodul>yeah
<efraim>and curiass may be down, same reason
<efraim> http://data.guix.gnu.org/revision/518fd3b6592203df73fec347d30016d4260ee127
***cap- is now known as cap
<spk121>hello guix people. So I noticed that in the Automake repo, it comes with a test-driver.scm. In a comment it says it is Guix's srfi-64 test harness. Is there a repo you can point me to that uses this, so I can see how it works in practice?
<civodul>spk121: you can look at Guix or Guile-Gcrypt, for instance
<civodul>oh, and hi! :-)
<civodul> https://notabug.org/cwebber/guile-gcrypt
<spk121>civodul: thanks!
<civodul>yw!
<civodul>rekado, g_bor[m]: ant-bootstrap fails on i686: https://ci.guix.gnu.org/log/yfjlb6b42gzn39vgmhj3gigdf7amw7a2-ant-bootstrap-1.8.4
<civodul>is it the same issue as before?
<efraim>it hasn't not failed in forever
<raingloom>hrmmm, could someone help with installing a ghc package? i imported it from hackage, it had a filing test that i disabled for now (just need it for a quick script) and installed it, but ghc(i) can't find it.
<raingloom>(it's listed in my profile by guix package --list-packages)
<rekado>civodul: it’s not the same issue as before.
<rekado>it segfaults.
<civodul>ah, good
<rekado>heh, “good” :)
<civodul>well, yeah :-)
<civodul>so we don't have any lead?
<roelj>raingloom: If it's listed in --list-packages you may still need to install it into your profile with “guix package -i <the-package-name>”
<civodul>g_bor[m]: for the R-B summit, we could coordinate, that's a good idea
<raingloom>roelj, that's how i installed it, from a local branch of guix
<rekado>civodul: it’s probably jamvm-1
<civodul>ah ha! was it upgraded?
<raingloom>(well, i used guix install, but that's the same thing)
<rekado>civodul: no, but it’s part of the brittle bootstrap JVM
<rekado>is this on the master branch?
<roelj>raingloom: Ah okay. Can you find its files in $HOME/.guix-profile somewhere?
<rekado>(because we have changed it on core-updates to fix x86_64)
<rekado>looks like core-updates because of glibc 2.29
<raingloom>roelj, find -iname '*unicode-show*' doesn't return anything...
<sirmacik1>hey guix. Quick question: what's your strategy for installing packages? Do you keep a list of all of them in config.scm or you rather keep the minimal valuable system packages there and install all others on your user profile?
<raingloom>...maybe i skipped a step when entering the environment for the local branch...... hmm.
<rekado>sirmacik1: I never add user packages to the system config.
<raingloom>roelj, oh, gotta pass -follow to find(1)...
<roptat>sirmacik1, I always try to keep the system packages to a minimum, and use a manifest for user packages
<sirmacik1>is there any „default” way to keep user environment packages in some manner similar to config.scm? So when I'm moving my system to new ssd I can just input file and have them all installed?
<roptat>sirmacik1, manifests :)
<raingloom>roptat, found this: ./lib/ghc-8.4.3/ghc-unicode-show-0.1.0.3.conf.d
<sirmacik1>oh, duckducking it now
<roelj>raingloom: You could build the package separately first to see what's in the build outputs and try to figure it out from there.
<roptat>sirmacik1, http://guix.gnu.org/manual/de/html_node/Aufruf-von-guix-package.html#profile_002dmanifest
<roelj>raingloom: Ah okay. And does GHC need an environment variable to find its packages?
<roptat>(I suppose you're german because of the quotes :))
<raingloom>roelj, yeah, i think so, there is a $GHC_PACKAGE_PATH
<roptat>raingloom, maybe try to install ghc in the same profile as your package
<raingloom>roptat, it's installed
<sirmacik1>is there a way to generate such manifest file or it has to be done by hand?
<raingloom>roptat, or at least it should be. but let's see what a guix install ghc does...
<roelj>raingloom: Can you check whether GHC_PACKAGE_PATH is set to the most recent version of your profile?
<roptat>raingloom, so try to source ~/etc/profile, $GHC_PACKAGE_PATH should be defined in there
<raingloom>roptat, oh.... wait. i think i figured it out. it seems to be using a different version of ghc... :facepalm:
<raingloom>8.6.3 vs 8.4.3
<raingloom>s/6.3/6.5/ but whatever
<sirmacik1>roptat: rekado do you guys host your config.scm files somewhere in clearnet?
<roptat>sirmacik1, https://framagit.org/tyreunom/system-configuration
<roptat>it's not the clearest because I have multiple machines
<sirmacik1>thx (:
<roptat>for the guix-home-manager I build a derivation from a configuration, but how can I make it into a generation of a profile?
<roptat>I see in (guix profiles) I can use a manifest and make it a profile, but I don't think I want a manifest, do I?
<civodul>rekado: actually ant-bootstrap/i686 is also broken on master! https://ci.guix.gnu.org/log/ls96d9cig605058dsg2a4q96564y6bnx-ant-bootstrap-1.8.4
<civodul>in a way, that's good news in that it's not a regression
<rekado>oof.
<rekado>I don’t think I can fix this now, but if nobody beats me to it I’ll look into this later this week.
<civodul>well, no rush i guess!
<civodul>htslib has test failures on i686: https://ci.guix.gnu.org/log/ndgnzmd2vmxi8srf1whsgadf79mn57b5-htslib-1.9
<civodul>that has to do with floating point rounding, as usual
<civodul>(maybe not a regression either)
<roelj>If I have a package that depends on Java libraries, and I'd like to add a Java library to the CLASSPATH of the program. How can I do that (in the package definition)?
<roptat>I think there's a wrapper (wrap-program or something similar)
<roelj>Can I do it with "search-paths"?
<roptat>no, that sets search paths at runtime
<roptat>it won't work if you want to run from the store path for instance
<rekado>civodul: I just built ant-bootstrap for i686 without segfault. Sent a patch for discussion to guix-devel. It was the first thing I tried.
<civodul>rekado: awesome, thank you!
<rain>roptat, ok, i still don't get it after all. i guix pulled in my profile, i merged master in the local checkout, i ran an upgrade from the local checkout, but ghc-unicode-show is still installed with the wrong ghc version :/
<rain>but i gotta go, so i'll have to debug this later
<roptat>rain, probably "ghc" the variable is not the same as "the latest version of anything named ghc"
<roptat>mh... that's a weird one, I have defined a record with define-record-type* with a mandatory field, data-directory
<roptat>if I try to create an object of that record type, it tells me data-directory is mandatory
<roptat>"missing field initializers (data-directory) in form (home)"
<roptat>but if I try to use "(home (data-directory ""))", it tells me data-directory is not bound
<roptat>am I missing something?
<roptat>home is the name of the record type
<civodul>roptat: can you try: ,expand (home (data-directory ""))
<civodul>?
<roptat>ah!
<roptat>it's caused by another field, whose default value I set to something that depends on the value of data-directory
<roptat>I guess I can't do that :/
<civodul>you can do something like that, see <user-account>
<civodul>the 'home-directory' field is a function of the 'name' field
<civodul>beware, this is high technology
<roptat>oh I see, thank you!
<Minall>Hello guix!
<joshuaBPMan>can someone send me a link to the guix documentation videos that the outready gal did?
<joshuaBPMan>nevermind I found it.
<joshuaBPMan>hey guix, I can't get the guix website to locally build. It is telling me guix-url is an unbound variable
<joshuaBPMan>and the error is not telling me what file the error is in.
<Tirifto>Hello! Is there documentation on setting up Guix for a user that has been created after Guix has already been installed?
<joshuaBPMan>Tirifto: what do you mean?
<joshuaBPMan>What issues are you having?
<Minall>Are deepin packages free?
<Tirifto>joshuaBPMan: I've been using Guix with my main user on this computer, but I have recently created a new user, and can't figure out how to use Guix there. The new user seems to have no Guix profile, and when I try to run Guix, it advises me to run ‘guix pull’ followed by ‘guix package -u’ (I think), even though I already ran those. I figure I need to take some steps to set Guix up so that the new user may use it…
<Tirifto>Oh, and I'm running Parabola GNU+Linux-libre, not Guix System.
<joshuaBPMan>Tirifto: I thunk you need to run "guix pull" as the new user.
<joshuaBPMan>I would imagine that that would create your profile for you new user.
<Minall>Hello guix!
<Tirifto>Hello Minall!
<Minall>Tirifto: o/
<efraim>or at least installing a first program
<Tirifto>efraim: Alright; ‘guix pull’ didn't do the trick, so I'll give that a try… actually I guess GNU Hello is perfect for this. :P
<Tirifto>Minall: I've read in various places that Deepin is free software, but I have not checked the project myself.
<Tirifto>Deepin Desktop Environment*, sorry. I guess the Deepin OS does have non-free bits. :)
<roptat>Tirifto, just use "guix install" and friends, no?
<roptat>if you've already run "guix pull" (as user, not root or sudo), you should have the latest version running (make sure ~/.config/guix/current/bin is first in your $PATH, and run "hash guix" (no output))
<Tirifto>roptat: I'll see if that works; I was just unsure of whether it's working as intended when it kept asking me to do ‘guix pull’ and ‘guix package -u’ even though I kept doing them.
<Tirifto>I don't have a profile on that user yet, but I'll add it there if it's not in the path yet!
<roptat>I guess it's a UI issue
<roptat>guix package should add the profile by itself when you first run it
<roptat>you should have a .config/guix/current symlink if you ran guix pull though
<roptat>it's not the same as the user profile, which is ~/.guix-profile
<alloy>When I use offloading, why isn't my local machine used for building anymore?
<roptat>I'm having some troubles with a derivation. I created one with gexp->derivation, and tried to build it, but in the builder #$output is replaced by a read of the out environment variable, which is evaluated to #f when building the derivation
<roptat>(I use (build-derivations store drv) to build my derivation)
<roptat>the drv file has references to the output and seems to specify the out environment variable correctly
<roptat>ha, nevermind, it's my fault
<Tirifto>roptat, efraim, joshuaBPMan: Thank you for your help; I'll go see if things are going to work out! :)
<gnu_srs>Hi again, when configuring guix guile-git is not found:checking if (git) is available... no. I've built from git and installed. What is missing?
<rekado>is it found on the GUILE_LOAD_PATH and GUILE_LOAD_COMPILED_PATH?
<gnu_srs>I have: export GUILE_LOAD_PATH=/.../scheme-bytestructures/scheme-bytestructures:/usr/share/guile/site/2.2/
<rekado>what about GUILE_LOAD_COMPILED_PATH? What’s the check performed by the configure script and what does the log say?
<gnu_srs>I've configured with --prefix=/usr
<gnu_srs>configure: $GUILE -c "(use-modules (git)) (exit ((lambda () 0)))" > /dev/null 2>&1
<kmicu>Thank you nee`
<gnu_srs>I have not set GUILE_LOAD_COMPILED_PATH
<gnu_srs>gnutls is found however. ls /usr/share/guile/site/; 2.2 gnutls gnutls.scm. ls /usr/share/guile/site/2.2; gcrypt git git.scm json json.scm sqlite3.scm ssh
<gnu_srs>I'd like to report a bug in guix git too: doc/local.mk: info_TEXINFOS = %D%/guix.texi , All other guix.de.texi, etc. are not available!
<Minall>Hello guix!
<Minall>I'm unable to reconfigure
***atw` is now known as atw
<Minall>| fase 'configure'builder for `/gnu/store/fyxycavvagxi2aj0r2hbrnav5pj25aj2-mate-applets-1.22.0.drv' failed with exit code 1
<Minall>la construcción de /gnu/store/fyxycavvagxi2aj0r2hbrnav5pj25aj2-mate-applets-1.22.0.drv falló
<Minall>Muestra el registro de construcción en '/var/log/guix/drvs/fy/xycavvagxi2aj0r2hbrnav5pj25aj2-mate-applets-1.22.0.drv.bz2'.
<Minall>cannot build derivation `/gnu/store/xv3gg281mdlpv9r22ms9n7v0f3dwa5jg-mate-1.22.0.drv': 1 dependencies couldn't be built
<Minall>guix system: error: build of `/gnu/store/xv3gg281mdlpv9r22ms9n7v0f3dwa5jg-mate-1.22.0.drv' failed
<Minall>
<rekado>Minall: what does the log at /var/log/guix/drvs/fy/xycavvagxi2aj0r2hbrnav5pj25aj2-mate-applets-1.22.0.drv.bz2 tell you?
<Minall>How do I check that log?
<rekado>unzip it with bunzip and then use “less” or another pager.
<gnu_srs>Is the configure script looking for git.scm or git.go?
<rekado>gnu_srs: when you run the Guile REPL and (use-modules (git)) in it — do you get an error?
<eric-freedombox>Hi, Someone Know how downgrade a package with guix?
<rekado>eric-freedombox: you can either downgrade to a previous profile generation, or you can install a package via an inferior linked to a previous version of Guix, or build a package variant using the previous version string and hash.
<gnu_srs>rekado: I'm building from git.
<eric-freedombox>Ok, bat how?
<eric-freedombox>but*
<pinoaffe_>Hi, a guix system of mine died, the configuration files I used for input were lost, is there a way to recover the list of packages in an active profile and/or the system config for the current generation?
<rekado>gnu_srs: that doesn’t answer my question :) You should try to reproduce this outside of the configure script.
<rekado>ng0: it’s in Marrakesh: https://reproducible-builds.org/events/Marrakesh2019
<OriansJ>anyone else notice an issue with /gnu/store/2qxdvq5mckri0p4wf9a7w7cicrl5ijw2-file-roller-3.28.0 ; To be specific the following error message "GLib-GIO-ERROR **: 14:43:21.542: Settings schema 'org.gnome.FileRoller.Listing' is not installed"
<gnu_srs>rekado: guile repl; No such file or directory: "/.../guix/guix/repl"
<rekado>gnu_srs: just “guile”
<rekado>OriansJ: this class of problems has affected other packages before. The solution is usually to also install some schema file.
<gnu_srs>(use-modules (git)) While compiling expression: no code for module (bytestructures guile)
<OriansJ>rekado: any recommendation for how to find that schema or is it more of a brute force sort of issue to resolve?
<rekado>gnu_srs: make sure that the directory where bytestructures was installed to is on GUILE_LOAD_PATH / GUILE_LOAD_COMPILED_PATH.
<rekado>OriansJ: I’m not sure. You’re welcome to send a report to bug-guix@gnu.org. I bet someone is going to be able to suggest a fix.
<rekado>(there is a mapping between these schema names and files, and essentially that file needs to appear in a directory that is mentioned in some env variable. But I don’t know the details.)
<gnu_srs>rekado: export GUILE_LOAD_PATH=/.../scheme-bytestructures/scheme-bytestructures:/usr/share/guile/site/2.2/
<rekado>what’s “...”?
<gnu_srs>OK; I had scheme-bytestructures/scheme-bytestructures but the package was at .../scheme-bytestructures.
<gnu_srs>Now it is complaining on Guile-JSON, which is installed
<rekado>same here: try to load (json) in a REPL and see what it says; then check the load path.
<gnu_srs>guile-json 1.2.0-2: Is that too old? checking whether Guile-JSON is available and recent enough... no
<rekado>latest Guix uses guile-json version 3
<gnu_srs>(use-modules (json)) within guile does not complain
<rekado>then it’s the version
<rekado>use Guile-JSON 3
<roptat>just pushed an update to the home manager to have a "guix home" command, with "guix home build" and "guix home reconfigure"
<efraim>roptat: nice
<gnu_srs>OK, installed guile-json 3.1.0-1. Now it is guile-sqlite3: I have 0.1.0-2 installed.
<roptat>I reconfigured my home, and it works as expected :)
<Parra>Hello everyone, I have discovered Guix recently, I plant to use it as a packer for distributing my software as self contained package
<Parra>plan*
<Parra>I think it's an awesome project and I found compiling is the best way for distributing binaries, I tried to do this before with patchelf but I had problems with it
<rekado>Parra: welcome!
<rekado>Parra: in what format do you plan to distribute the binaries
<rekado>?
<nickey>hi guys! is `pass` password manager unwanted on guix, or just nobody still needet it?
<pkill9>nickey: `pass` is packaged in guix as `password-store`
<nickey>pkill9: oh... thank you... I didn't found it
<Parra>rekado: what do you mean? they are .so files
<Parra>basically it's a c/c++ library that will be used from pip, npm, nuguet, etc.. so basically I want to achieve portable distributables for all architectures in order to provision this package managers
<Parra>nuget*
<Parra>nowadays it's difficult to do properly this, and for me guix seems very promising to solve this problem
<Parra>I just want a zip with all files self contained, with relative path to rpath and the loader (also self contained in that zip)
<Parra>I think it's doable after what I read in the doc
<roptat>Parra, it's possible, but the archive will probably be very big
<roptat>if you use guix pack
<roptat>and the relocatable option might also slow things down a little
<Parra>at runtime?
<roptat>a bitcoin dev uses guix to distribute binaries though, and they don't use guix pack it seems
<Parra>do you have a link of their solution?
<Parra>it will be interesting to follow it
<roptat>so maybe their method is going to work better
<ng0>oh, cool
<Parra>well but it's possible, right? what's the problem with pack?
<guixgoldfish>Hi all I've installed GNU Guix few day ago in a virtual machine and had to learn how to pull, install and delete with guix. GNU IceCat and LibreOffice are not in my first language but in english. Is there any work around to get them locale? Thanks in advance.
<ng0>I mean the reproducible builds location
<roptat>Parra, https://github.com/bitcoin/bitcoin/pull/15277
<roptat>I think that's it
<roptat>Parra, nothing, but it creates huge packs and small runtime performance issues
<roptat>guixgoldfish, for icecat, you'll have to download a language pack
<Parra>thank you so much roptat
<Parra>and a question, I've been told that netcore isn't ready in guix, right?
<Parra>maybe I will implement the guile spec for it, so I would like to contribute with it to Guix, where can I push new guile specs?
<roptat>I think that's a question for #guile, I have no idea what you're talking about ^^'
<rekado>I think “guile spec” here means “package definition”
<rekado>Parra: you can send patches via email to guix-patches@gnu.org. Please see the Contributing section in the manual first.
<Parra>exactly, I just want to contribute with new packages
<rekado>Parra: the Contributing section in the manual will tell you everything you need to know to contribute.
<rekado>welcome!
<Parra>thank you so much, I will keep you informed about my adventures
<Parra>I'm also using guix inside docker, in order to automate the CI/CD, not sure if it is a good practice or not
<Parra>I will opensource all code anyway
<rekado>I think there’s no advantage to using Guix inside of Docker compared to using Guix on its own.
<rekado>Guix doesn’t use any libraries or tools of the host system, so the host system doesn’t really matter
<rekado>(as long as the kernel is recent enough)
<Parra>but how can I run it inside travis or gitlab otherwise?
<Parra>I assume that I can decompress the 1.0.1 version in a docker container from scratch, and guix will be able to run by itself, right?
<Parra>it's just a standard way of automating it, nothing more
<Parra>I can move easily later on from different CI servers
<Parra>I asked in a telegram group and most guys said the same, but I like to work like that
<civodul>roptat: re "guix home", nice!
<civodul>roptat: it'd be great to work towards integrating it in Guix
<roptat>civodul, I'd like to have a mechanism close to the service mechanism before that
<kmicu>Thank you roptat You're breathtaking.
<nickey>hm... dwm don't show unicode symbols in statusbar... but in arch dwm buit from same code and using the same font all fine... what can be different? X?
<dongcarl>no kmicu _you're_ breathtaking!
*rekado sees keanu everywhere
<dongcarl>across space _and_ time, I hear
<kmicu>no dongcarl you’re breathtaking!
*rekado looks for the inhaler
*rekado takes a deep breath
<dongcarl>question about general Guix debugging... I'm building a manifest, and I see this: https://pastebin.com/JjpEjxaG
<rekado>dongcarl: I’ve bumped into this very same bug just a day ago
<dongcarl>It's hard to know where to look... :-/
<rekado>I’m not sure what’s going on here.
<dongcarl>rekado: REALLY?!
<rekado>a user at the institute, actually
<dongcarl>Where did you bump into it?
<dongcarl>Sounds like it needs to be fixed
<rekado>we’re using inferiors with more than one channel
<dongcarl>rekado: Right exactly!
<rekado>the user managed to avoid the problem for now by pinning the 'guix channel to commit 2f43e5db1c36ec93a80ed1e3cbe763a8d64adcb5.
<rekado>this sounds like a meta error, really
<dongcarl>Hmm... Well let me describe what I have
<rekado>there’s something triggering an error in an expression that is wrapped in “call-with-error-handling”
<dongcarl>Perhaps the info will be helpful for debugging
<rekado>but the error is not a condition, so the guard fails
<rekado>or something like that
<dongcarl>my manifest only has one channel in it, called `'bitcoin-guix`, and that channel's repo has a `.guix-channel` file that specifies a `'guix` channel
<rekado>actually, the error says that it got a “condition” but wrapped inside an “inferior-object”
<rekado>that’s the problem, I think
<rekado>it should just be a condition.
<rekado>the fact that we’re using an inferior results in a wrapped value
<rekado>it should be unwrapped before raising the condition.
<dongcarl>another piece of info, this is only happens for one of the packages in the manifest (I narrowed it down), all others are fine
*dongcarl reading
<rekado>same here
<rekado>it happened for python-scanpy
<rekado>others were ine
<rekado>*fine
<rekado>I’m guessing that this may be related to sexp->object in read-repl-response in guix/inferior.scm
<dongcarl>rekado: Huh...
<rekado>conditions don’t have a self-quoting sexp representation
<rekado>so they end up being wrapped in inferior-object
<dongcarl>rekado: Is this easy to fix?
<rekado>I don’t know.
<dongcarl>rekado: what should I look for in guix/inferior.scm? I can take a look
<rekado>read-repl-response expects to read either (values …) or (exception …) from the port
<rekado>I don’t know what a condition value would look like
<rekado>we probably need to a) make sure that conditions are serialized properly and b) handle them in read-repl-response.
<dongcarl>rekado: Why does this error only happen for _some_ objects and but not all?
<rekado>I don’t know
<rekado>we don’t even see the actual error.
<rekado>I guess (a) needs to be done in guix/repl.scm in send-repl-response
<rekado>the problem seems to be that there we only catch plain exceptions, but not conditions
*dongcarl needs to read up on exceptions and conditions
<rekado>(error 'foo) is an exception
<rekado>conditions are more complex and are defined in srfi-35
<rekado>conditions have proper types whereas plain errors really only have a key and optional arguments
<rekado>errors can be caught with (catch 'the-key …) or (catch #t …) for any error
<civodul>dongcarl: could you paste the error on, say, paste.debian.net? (patebin.com cannot be accessed over Tor)
<dongcarl>civodul: Yes, would love to, one second
<dongcarl>rekado: Ah the type distinction is good to know
<civodul>debugging exceptions from an inferior is... inconvenient
<dongcarl>civodul: http://paste.debian.net/hidden/03c69d6b/
<dongcarl>civodul: agreed
<dongcarl>I was actually thinking that perhaps inferiors can be a first-class `guix` cli construct... Not sure if that's possible
<dongcarl>As in, run the CLI as if `guix` was at this commit
<dongcarl>Maybe that's dumb :man-shrugging:
<dongcarl>gunna be back online in 30 mins if you guys need more logs or wtv
<civodul>dongcarl: oh, i see
<civodul>actually i don't see that much, mostly ellipses
<civodul>ah the re-throw for SRFI-34/35 is interesting
<civodul>it cannot work
<civodul>we should maybe serialize &store-error conditions over the wire
<bavier>hello guix
<Minall>Hello guix!
<Minall>I'm unable to reconfigure, opening the log, it gives me this:
<Minall>configure: error: Stickynotes explicitly requested but gtksourceview not found
<Minall>Backtrace:
<Minall> 4 (primitive-load "/gnu/store/w4wxl1wnwm44wsf14igsykfim02…")
<Minall>In ice-9/eval.scm:
<Minall> 191:35 3 (_ _)
<Minall>In srfi/srfi-1.scm:
<Minall> 863:16 2 (every1 #<procedure 7575c0 at /gnu/store/0npakqh7q9kfi…> …)
<Minall>In /gnu/store/0npakqh7q9kfik8y0cc0vjqqpvnyv2h1-module-import/guix/build/gnu-build-system.scm:
<Minall> 799:28 1 (_ _)
<Minall>In /gnu/store/0npakqh7q9kfik8y0cc0vjqqpvnyv2h1-module-import/guix/build/utils.scm:
<Minall> 616:6 0 (invoke _ . _)
<Minall>
<Minall>
<bavier>Minall: that looks like a package build failure
<Minall>bavier: What can I do?, I'm trying to reconfigure...
<civodul>hi bavier!
<civodul>Minall: what did "guix system reconfigure" show before "View build log ..."?
<bavier>civodul: o/
<rekado>Minall: is this the log for “mate-applets”?
<rekado>Minall: the build for mate-applets would need to be fixed.
<rekado>I guess the problem might be the gtksourceview version; we have two of them.
<Minall`>Yes,
<Minall`>So... can't I reconfigure with mate?
<rekado>Minall`: I can’t reproduce the problem.
<rekado>what version of Guix are you using? Please show the output of “guix describe”.
<Minall`>Ok
<Minall`>Generación 7 16 sep 2019 06:25:36 (actual)
<Minall`> guix c34671a
<Minall`> URL del repositorio: https://git.savannah.gnu.org/git/guix.git
<Minall`> rama: master
<Minall`> revisión: c34671a69dcc6fc508c54822580e68095c6e719c
<rekado>is this the same Guix that you’re using to reconfigure?
<Minall`>Yes,
<rekado>because with that version mate-applets can in fact be built.
<rekado>what command do you use to reconfigure?
<rekado>do you run this as the root user or do you use sudo?
<Minall`>I do this as the root user: guix system reconfigure mate.scm
<Minall`>I removed my user from the wheel group, so I can only use use su
<rekado>what does “guix describe” say for the root user?
<Minall`>Wow
<Minall`>error:
<Minall`>the origin couldn't been determined
<Minall`>Perhaps this `guix' command was not obtained with `guix pull'? Its version string is 1.0.0-1.326dcbf.
<rekado>the root user’s Guix is probably too old. What does “type guix” say?
<Minall`>guix está asociado (/run/current-system/profile/bin/guix)
<Minall`>
<Minall`>Should I run guix pull?
<rekado>in that case either use /home/minall/.config/guix/current/bin/guix or run guix pull
<Minall`>I'll guix pull
<gnu_srs>(09:34:14 PM) srs: OK, installed guile-json 3.1.0-1. Now it is guile-sqlite3: I have 0.1.0-2 installed. Too old?
<gnu_srs>Again: Is it the .go file or the .scm file configure is looking for? /usr/lib/x86_64-linux-gnu/guile/2.2/site-ccache/sqlite3.go, /usr/share/guile/site/2.2/sqlite3.scm