IRC channel logs


back to list of logs

<unmatched-paren>lunabee: only for tools
<nckhexen>The ambiguity isn't really Go-specific. The ‘tool’/library distinction *is* fuzzy and somewhat arbitrary.
<lunabee>ahhh okay
<lunabee>thanks everyone!
<mohamed>I have a problem, I don't have linux headers, For exemple when I try to compile a c program usung limits, I have an error, it tells me that I don't have <linux/limits.h>
<nckhexen>Got linux-libre-headers?
<mohamed>(I'm using nonguix linux)
<unmatched-paren>mohamed: it'll be called linux-headers then
<unmatched-paren>sorry, it won't
<unmatched-paren>the forbidden channel uses the same headers apparently
<unmatched-paren>so just install linux-libre-headers
<mohamed>I tried with linux-libre-headers from (gnu packages linux) inside my home.scm but I didn't change the result. Should I put it in my system config ?
<unmatched-paren>it should work inside your home.scm...
<unmatched-paren>that's very odd. maybe you need to reload the profile?
<nckhexen>Is it in the same profile as gcc-toolchain?
<mohamed>nckhexen : yes I pu gcc-toolchain also in home.scm
<mohamed>I put*
<mohamed>I tried to reload my zsh profile but no change
<nckhexen>If you're using a custom shell, perhaps it's messing with your environment?
<nckhexen>Check $C_INCLUDE_PATH/linux/limits.h
<nckhexen>Hullo lampperson.
<lamp1400>i am trying to communicate with mariadb, but the DB does not respond. Wich is the standard service operating port?
<mohamed>$C_INCLUDE_PATH is empty
<nckhexen>lamp1400: 3306, apparently.
<nckhexen>mohamed: It shouldn't be. Check without zsh etc. For example, if you use ‘guix shell --pure gcc-toolchain linux-libre-headers’, you'll see that it should be set and to what.
<nckhexen>When packages share a profile, they can request ‘search paths’, which are just environment variables that are provided by other packages. Looks like something is clobbering or preventing that.
<nckhexen>If you're sure that gcc-toolchain & linux-libre-headers share a profile, it must be the clobberin'.
<nckhexen>(Or there's a bug in guix home, which seems unlikely.)
<lamp1400>nckhexen: but i can't connect to the db by this port
<nckhexen>Is it running?
<lamp1400>nckhexen: the service is running
<nckhexen>lamp1400: Does ‘sudo lsof -i’ show anything relevant?
*unmatched-paren away
<unmatched-paren>night \o
*nckhexen also away, actually. o7
<mohamed>nchhexen: I will try your suggestions, thanks guys
<lamp1400>nckhexen: it's running mysql
*nckhexen ret, but not for long.
<nckhexen>lamp1400: What do you mean? I thought you said mariadb. The service is called mysql, but the process should be called mariadb.
<nckhexen>Do you mean it's listening on the ‘mysql’ port (3306/tcp/udp)?
<nckhexen>You can use -n to ask lsof not to look up service names.
<nckhexen>You can also share the output on
<lamp1400>the name of service is mysql
<nckhexen>But it runs mariadb by default, unless you ask it not to.
<lamp1400>nckhexen: output "mysqld 4889 mysql 19u inet 53529 0t0 TCP localhost:mysql (LISTEN"
<nckhexen>So it's listening or port 3306 (see /etc/services).
<lamp1400>nckhexen: the nextcloud is not connecting
<nckhexen>We keep moving the question.
<nckhexen>What error message(s) do you *actually* get?
<nckhexen>I'm not familiar with either NextCloud or MariaDB, so the more info you can give, the less unlikely it is I'll be able to be useful…
<nckhexen>If there's a /var/log/mysql, it might contain application-level information detailing what went wrong.
<nckhexen>Ah, it's /var/log/mysqld.log, says databases.scm.
<nckhexen>Worth a try. Same goes for NextCloud, assuming it logs at all.
<lamp1400>nckhexen: output nextcloud ""
<lamp1400>Error while trying to create admin user: Failed to connect to the database: An exception occurred in the driver: SQLSTATE[HY000] [2002] No such file or directory
<lamp1400>Create an admin account"
<lamp1400>nckhexen: my config.scm of system:
<nckhexen>‘No such file or directory’ (which???) is the bane of my existence.
<nckhexen>Thanks for the system configuration, but I think the NextCloud settings need to be inspected. Like, how is it trying to connect (port or socket?) and which database file is it trying to open (does it exist?) etc.
<nckhexen>I recommend connecting to mysql (mariadb) using the command-line client first, to make sure the server is up and running, which seems to be the case. But I can't give you more specific instructions, sorry, I'm a Postgres person…
<nckhexen>Do you know your way around a MySQL command line?
<nckhexen>When I search for ‘error while trying to create admin user’, I get lots of NextCloud configuration issues like <> and <> (which has a checklist, of sorts).
<nckhexen>I'm no NextCloud expert, but I don't *think* your error is Guix-specific at this point.
<nckhexen>More like you need to manually make them find each other, by pointing NextCloud to the right DB settings, and possibly creating these DBs first. This is all ‘stateful’ stuff that isn't (yet) handled by Guix at all, so the steps should be similar on any distro.
<nckhexen>I've got to go, sorry I couldn't be of service, but good luck.
<lamp1400>nckhexen: i'm connecting by port and the data folder of nextcloud exists
<lamp1400>nckhexen: thank you so much
<lamp1400>last thing, log of nextcloud: Operation not permitted at /srv/http/
***daviid` is now known as daviid
<podiki[m]>sneek: later tell civodul looks like a bug in fhs container after as the glibc-for-fhs package may not appear if another package propagates (I think) glibc
<sneek>Got it.
<podiki[m]>sneek: later tell civodul I think because of glibc-for-fhs being added to the ad-hoc packages before/after what the user specifies? not sure, will see if I can come up with a fix but can confirm glibc (regular guix) can appear instead of the fhs version
<linj>is it possible to use /boot on btrfs raid1 on luks on 2 devices?
<podiki[m]>sneek: later tell civodul throwing in a reverse for options/resolve-packages in (guix scripts environment) would fix it... so maybe just make sure glibc-for-fhs appears last?
<sneek>Got it.
<podiki[m]>sneek: botsnack
<vivien>I have an error doing pull with the following channel (save it in a.scm, add it to a guix repo, use it in a channels.scm):
<vivien>(define-module (a) #:use-module (guix store) #:use-module (guix monads))
<vivien>(define b (run-with-store (open-connection) (mbegin %store-monad (return #t))))
<vivien>Do you have the error too?
<vivien>(the define is not required, (run-with-store (open-connection) (mbegin %store-monad (return #t))) does the same)
<vivien>Could someone tell me what I did wrong to deserve the error?
<johnabs[m]>Hi all, is there an easy way to remap my caps lock to esc in config.scm? I saw the keyboard layout section in the manual, but I don't think that's wha tI'm looking for. Also, does anyone else experience applications like element-desktop or signal-desktop being unreasonably slow, while other gui apps are quite snappy? I've never experienced this on other distros.
<johnabs[m]>Another quick one hopefully: is the "commit" field required for the channels.scm entries? If I remove it, does it just use the most recent commits when I run "guix pull"?
<unwox>johnabs[m]: keyboard layout section is the answer: see #:options and this thread
<unwox>signal and element are electron apps so they are usually slower than qt/gtk apps
<sneek>yewscion: Greetings :D
<thecom>hey, i can't figure out how to use guix in ~guix system vm~
<thecom>the qemu vm boots fine, but if i run any guix command inside the vm, i get /gnu/store read-only filesystem errors
<thecom>what am i missing?
<thecom>i'm on a foreign distro (arch), maybe that's relevant?
<johnabs[m]><unwox> "signal and element are electron..." <- Right, but my ancient thinkpad handles them flawlessly whereas my shiny new 12thgen i7 machine shows an embarassing amount of (what appears to be) input lag for some reason :/
<lilyp>perhaps some performance vs. efficiency core thing?
<johnabs[m]>Maybe, but I'm skeptical; are there other electron apps you'd recommend I try first? If I understand correctly, electron apps are basically just webapps packaged with chromium. So if I install chromium and observe the same/different behavior, would that indicate something more useful?
<johnabs[m]>Strangely, it does appear to be input related, as both keyboard entry, and moving my mouse over areas that are highlighted slows things down a lot, but I can build packages and surf the web in the background no problem
<lilyp>hmm, I'd try google maps in chromium – or any browser really – that caused me some problems with my nvidia card for example
<johnabs[m]>Ah, this may be an issue: I got this when I ran chromium from the terminal: [17091:17091:1022/] dri3 extension not supported. Any ideas?
<lilyp>what graphics do you have?
<lilyp>integrated intel?
<johnabs[m]>I'm on a framework laptop if it helps
<lilyp>hmm, try something that makes extensive use of 3d and report your framerate
<lilyp>3d accel should be supported on most intel cards, but if it's a shiny shiny new one you might have found one without
<johnabs[m]>Any suggestions there? And should I try them in chromium/firefox, or should it be an independent program?
<lilyp>any program is fine, but stuff that needs a browser or wine exacerbates this
<lilyp>try opening for example, it hated my nvidia card
<johnabs[m]>It's instant on firefox
<johnabs[m]>Same on chromium too, but the input lag is BAD, same with the other apps
<johnabs[m]>Also, here's my processor: Intel® Core™ i7-1260P Processor
<johnabs[m]>Oh and the little globe animation is CHUGGING on chromium
<johnabs[m]>smooth on figrefox
<johnabs[m]>Semi-related bug, but not the problem in my case:
<lilyp>okay, so IOW chromium sucks and electron by extension
<johnabs[m]>sorry, what is IOW?
<johnabs[m]>Okay, so just passing --disabl-gpu fixes it.....
<johnabs[m]>It even works on element xD
<johnabs[m]>Although, not on signal? Hmm...
<lilyp>johnabs[m]: "in other words"
<lilyp>perhaps signal doesn't honour disable-gpu; try to find out why
<johnabs[m]>lilyp: Oh, I thought it was some sort of jargon xD Thanks for the clarification
<johnabs[m]>And I'll try :)
<johnabs[m]>I figured it out: it doesn't honor typos 🤦‍♂️
<johnabs[m]>"diable-gpu" != "disable-gpu"
<johnabs[m]>I do find this quite strange that the performance is so dramatically improved. Even the video refresh-rate of signal improved to probably 50fps minimum from around 15-20 previously. Something is wack with that dri3 thingy
<unmatched-paren>morning guix! :)
<unmatched-paren>johnabs[m]: is this a gen 12 framework, by any chance?
<unmatched-paren>because i found that the gpu driver for its GPU is the nonfree i915
<vivien>Hello guix, can someone try and pull this channel? I’m getting an error that I don’t understand:
<vivien>(define-module (a) #:use-module (guix store) #:use-module (guix monads))
<vivien>(run-with-store (open-connection) (mbegin %store-monad (return #t)))
<vivien>There is no problem loading this module in guix repl, but I can’t seem to have guix pull compile it
<vivien>I get: (exception %exception (non-self-quoting 140737175834880 "#<&store-connection-error file: \"/var/guix/daemon-socket/socket\" errno: 2>"))
<vivien>Is anyone having the same issue? I don’t want to open an issue if it’s just something on my end
<unmatched-paren>vivien: errno 2 is apparently ENOFILE
<vivien>If I do: $ file /var/guix/daemon-socket/socket
<rekado_>vivien: are you trying to connect to the Guix daemon while in a build container?
<vivien>I get /var/guix/daemon-socket/socket: socket
<vivien>Oh I see so the problem is the guile compiler wants to execute run-with-store while it should not
<rekado_>thunk it
<vivien>If I enclose it with (eval-when (load) …) it seems to work
<vivien>Also maybe I should add eval?
<vivien>I need to run it for a global definition, so I can’t just make it a thunk
<vivien>I’m not sure eval-when does what I want though
<vivien>Well it does not, and I understand that I won’t be able to get away with it anyway
<vivien>If self-contained-tarball in (guix scripts pack) used computed-file instead of gexp->derivation, I could stay with gexps the whole time and I would have no problems at all
<vivien>Now another question: can I extend the guix command-line with a plug-in in my channel? I would like to introduce "guix ostree" that would create a binary pack of a set of packages and commit it to an ostree channel
<vivien>If I can do it in my channel it would be easier than changing the guix code base
<vivien>(not an ostree channel, an ostree repo)
<abralek>Hi Guix!
<lilyp>vivien: I think you can, guix home was/is distributed via rde's channel
<jpoiret>iirc it uses the guix extensions mechanism that is not documented :)
<vivien>Do you have an URI for this channel?
<jpoiret>zimoun used it in a lightning talk at the 10years event
<unmatched-paren>vivien: To do a Guix extension, you need to install a module to share/guix/extensions/*.scm
<unmatched-paren>and the module needs to define a command called guix-* using define-command from (guix scripts)
<vivien>unmatched-paren, can I do this in a channel, or do I have to package the extension?
<unmatched-paren>vivien: you should package it
<unmatched-paren>as a guile project
<vivien>That’s a bit awkward because hall won’t install my files inside share/guix/
<jpoiret>i think if you just add a module to the guix/scripts/ subdir in a channel it will also work
<lilyp>that talk will be a cookbook entry once things have been ironed out
<unmatched-paren>jpoiret: it will? nice
<jpoiret>since commands are resolved by just doing (resolve-interface `(guix scripts ,command))
<vivien>Oh cool
<jpoiret>not guaranteed that it will work with -L though
<vivien>Nevermind I can still do "guix time-machine -C <file with guix and my channel> -- command"
<abralek>Is there a way to add service user (the one from activation step) to some extra supplementary-groups?
<vivien>I’ll plagiarize (guix scripts show)
<lilyp>abralek: if it's a new user account, you can just customize it however you want
<lilyp>if you want to modify an existing one, no
<abralek>Yeah, it is not new. It is the one which creates during service activation step.
<abralek>I could write yet another service to run usermod thought, but ti doesn't look right
***Dynom_ is now known as Guest9557
<lilyp>don't do that; instead discuss on the ml which group to add and why
<lilyp>or add an escape hatch to add more groups to this account
<abralek>adding a hatch is possible, but that would not rely on /etc/groups. In the first case it is also wired. In my case i want to run git-daemon and gitolite. They use different accounts. So I need git-daemon be in group git. But someone would just want to have git-daemon only. No need for git group in that case
<civodul>podiki[m]: hi! could you report a bug for the --emulate-fhs issue?
<abralek>lilyp: why do you think that service is a bad idea?
<abralek>hmm probably because of the lack of order during the start..
<lilyp>imperative sux, declarative rocks
<lilyp>why can't gitolite run its git whatever under the git user?
<lilyp>assuming it needs to do so in the first place
<unmatched-paren>lilyp: hi, may i ask some questions about your review of my home-emacs-service-type?
<lilyp>go ahead
<unmatched-paren>what's a systemd-style constructor/destructor pair?
<lilyp>I'll just paste you my handwritten shepherd.scm (a fragment of it anyway)
<lilyp> here, this uses make-systemd-constructor/destructor
<unmatched-paren>cool, thanks
<unmatched-paren>why should i use one, though? what advantage does it provide over a forkexec constructor?
<lilyp>it starts as-needed, meaning once you start your first emacsclient
<lilyp>disadvantage is that shepherd doesn't currently clean up the socket nicely, but that's a bug
<unmatched-paren>also, i think we should stick with ~/.local/var until we collectively agree that it should be replaced, and replace all instances of it at once
<unmatched-paren>since it seems like every home service that uses logs uses it
<unmatched-paren>with "you should collect the emacs package plus lisp packages into a profile": do you mean add the lisp packages to the home-profile-service-type? i was explicitly trying to avoid that with the manipulation of the environment variables
<unmatched-paren>and for an "escape hatch for additional elisp files": do you mean something like an ``extra-lisp-files'' list of extra files to be ``load''ed on startup?
<unmatched-paren>lilyp: also, since i use make-forkexec-constructor for both start and stop, so that kill-emacs-hook is run when emacs is stopped, should i retain the stop make-forkexec-constructor and just change start to make-systemd-constructor?
<unmatched-paren>or should i change both to make-systemd-constructor, or something else?
<lilyp>no, a separate profile
<lilyp>not loaded on startup, just put in the directory
<lilyp>i don't think kill-emacs-hook changes between constructors, does it?
<unmatched-paren>okay, cool. thanks. i'll try to figure out how to generate a separate profile.
<lilyp>and I very much think we shouldn't proliferate the status quo, it will only make the switch to xdg conformance more painful
<unmatched-paren>lilyp: hmm, okay. so, ${XDG_STATE_HOME:-$HOME/.local/state}/log as the log directory?
<lilyp>note that shepherd itself already does this
<lilyp>so it should probably be ~/.local/state/emacs/log or something like that with log being a regular file
<unmatched-paren>i might send another patchset to convert the rest of the home services to XDG conforming state directories...
<lilyp>or ~/.local/state/emacs.log
<unmatched-paren>i see.
<unmatched-paren>thank you for the review! :)
***fanta1 is now known as fanta1|away
***fanta1|away is now known as fanta1
***phonymontana is now known as TopExpert
<abralek>lilyp: Technically we can, but for example gitolite has to have a shell and git-daemon doesn't
<abralek>maybe it would be better to define git group in both services at the same time..
<unmatched-paren>lilyp: i'm not really clear on how i would apply a profile's search paths to the service
<unmatched-paren>would i just write a function to convert a search path into a key=value string?
<vivien>>:( guix pull won’t care about any modules in (guix …) in my channel
<vivien>It just pretends share/guile/site/3.0/guix is just the main guix channel
<csepp>just to gauge the state of things, is someone working on upgrading gtk+? there are a few irritating bugs that I hope would be fixed by an upgrade.
<unmatched-paren>csepp: it *might* need to be updated in staging or (worse) core-updates
<csepp>unmatched-paren: i looked at staging on cuirass and it has the same version.
<csepp>guix refresh -l says it has 2811 dependents? not sure how to interpret its output.
<unmatched-paren>in that case it may need to go in one of the other branches...
<jpoiret>csepp: the output is saying that the package has 2811 indirect dependents, but that you can check that everything builds by only rebuilding the n leaf packages listed
<lilyp>unmatched-paren: see set-search-paths, but more easily you can source that profile's etc/profile
<lilyp>csepp: if you can fix it with a graft, that can go to master
<unmatched-paren>lilyp: yes, but i'm not sure how i'd possibly apply that to a service
<unmatched-paren>launch ``sh -c "source (profile)/etc/profile; (emacs) ..."''?
<lilyp>there could be some super cool env hack that I don't know of, but my thought was more like using guix's apis inside shepherd or doing something similarly clever
<lilyp>but since you're trying to pull this into #:environment-variables, try utilizing evaluate-search-paths
<pkill9>hello guixers
<unmatched-paren>pkill9: hi!
<AwesomeAdam54321>pkill9: hello
<unmatched-paren>lilyp: thanks!
<nckhexen>sneek: later ask antipode: Have you successfully used mosh on the Telenets recently?
<Lumine>sneek: botsnack
<Lumine>Happy bots make for a better world!
<nckhexen>Chonky bot.
<antipode>nckxhexen: I haven't ever used mosh at all.
<sneek>Welcome back antipode, you have 1 message!
<sneek>antipode, nckhexen says: Have you successfully used mosh on the Telenets recently?
<antipode>In case this is about network trouble: I often can't connect to on '4G', don't know if that's related.
<nckhexen>Specifically UDP (the SSH part connects fine, so does regular SSH), and on their Wi-Fi (although I expect it not to matter if it were wired).
<nckhexen>It's new since a few weeks. I can't imagine them enraging gamers by blocking much of any UDP, so I wonder what's going on.
<nckhexen>Anyway, thanks for surfacing to reply.
<xd1le>goodnight guix
<xd1le>(for me)
<Lumine>Anyone here uses zram? I'm wondering what the optimal priority should be set as, right now I have it on 10, which seems very low looking at the maximum size
<apteryx>Lumine: it's not very important unless you also use a normal swap
<apteryx>it's just a crude way to give priority to one above the other, IIRC
<Lumine>Well I do have a normal swap space, which is smaller in size than the zram
<Lumine>Guess I'm wondering which takes takes the lead currently
<lilyp>I think I solved my spooky input device case
<nckhexen>Lumine: Whichever has the higher priority.
<nckhexen>(See swapon.)
<nckhexen>Equal priorities will stripe.
<Lumine>nckhexen: ah, thanks
<Lumine>That's odd, swap space's priority is -2
<nckhexen>It starts counting down at -2.
<johnabs[m]><unmatched-paren> "because i found that the gpu..." <- Hey paren! Yeah, it's the 12th gen, but I'm also using "full" linux from nonguix, but maybe I need to specify that driver somewhere in my config?
<nckhexen>So each swap space added later will have a lower priority than the previous one.
<Lumine>Ah, I see
<nckhexen>Lumine: This leaves the positive range for human-specified priorities.
<nckhexen>See (guix)Swap Space for how to do that.
<unmatched-paren>johnabs[m]: Yeah, you do. I'll PM you:
<johnabs[m]>unmatched-paren: Thank you so much for your help!! 😄
<Lumine>nckhexen: got it
<nckhexen>Happy swappin'.
<nckhexen>antipode: Nope, they just firewall UDP port 60001. God knows why.
<nckhexen>Easy to work around, at least.
<apteryx>what was the bug of the PID in the builder not reaping processes?
<apteryx>PID1 in the builder chroot
<apteryx>(which is the guile script)
<apteryx>can't find it
<apteryx>ah, it's a patch
<apteryx>thanks MUMI
<abralek>Is anyone running gitolite with http access (cgit, http)?
<apteryx>civodul: heya! should we have substitutable grafts, that would remove the need to download all outputs (required to do the graft locally), right?
<apteryx>If so, it seems it'd be worth it. it's annoying to download hundreds of MiBs of debug data just to get a binary.
<podiki[m]>civodul: yes, I'll report it later today. I was hoping to have a simple patch too, but I'm not sure where best to do the change (at least I think I know what we need)
<AwesomeAdam54321>How do I refer to a different package output when using #$(this-package-input ...)?
<unmatched-paren>AwesomeAdam54321: Try #$(this-package-output "foo:bin") maybe.
<unmatched-paren>Hmm, never mind. It looks like no extra string is appended when using an alternative output...
<sneek>Welcome back Lumine
<johnabs[m]><unmatched-paren> "johnabs: Yeah, you do. I'll PM..." <- Hey, I'm a bit new to IRC/LiberaChat/Matrix, I wanted to check if you sent that PM yet, and if so, where do I find it? 😅 Not trying to rush you or anything, just trying to deal with minor ignorance on my part, lol
<unmatched-paren>johnabs[m]: Ah, you're on Matrix. I'm not sure how you access PMs on Matrix...
<unmatched-paren>I have sent it though.
<unmatched-paren>Eh, I'll just put the links here:
<johnabs[m]>unmatched-paren: I can hop on libera chat super quick, would that work?
<unmatched-paren>yeah, sure
<unmatched-paren>it should appear there, i think...?
<johnabs[m]>Oh okay, thanks!
<johnabs[m]>Yup, I see 2 links!
<unmatched-paren>johnabs: in my PMs?
<johnabs[m]><unmatched-paren> "johnabs: in my PMs?" <- I am so sorry, but I'm so confused right now, along with the s/my/your/ thing?
<unmatched-paren>johnabs[m]: Never mind, just use those links I posted above :)
<Lumine>s/ means they are correcting a word, johnabs[m]
<johnabs[m]>Oh okay, will do, thanks!
<johnabs[m]>Lumine: OH, like a sed command, lol. Thank you for the clarification lol
<Lumine>No problem, if this is really your first time to IRC, it can be a bit disorienting seeing all the little conventions
<johnabs[m]>Yeah, I'm quite new to IRC, and I suck at acronyms already since I usually need a little extra context to deduce what specifically they mean, so when people start throwing regex replace statements in where I don't expect them, I start questioning my sanity 😂
***LispyLights is now known as Aurora_v_kosmose
<nckhexen>Oh. Expect them.
<blake2b>getting a strange error using 'guix repl' (version 3.0.8) while working on a package that doesn't occur when i launch geiser without a repl server:
<nckhexen>johnabs[m]: ‘Fun‘ fact: the Matrix <-> IRC bridge automatically generates them nowadays. So if you edit your message on Matrix (please don't), it will try to diff it as s/foo/bar/.
<blake2b>unknown file:#f:#f: encountered raw symbol in macro output in subform socket of (current-location-vector)
<blake2b>does anyone know what this is about?
<nckhexen>It used to send the entire message again, which was worse.
<nckhexen>I'm not really following the steps, blake2b. Is this also within Geiser?
<johnabs[m]><nckhexen> "johnabs: ‘Fun‘ fact: the Matrix..." <- That's pretty sick actually, thanks for the tip! :)
<kaelyn>Hi #guix, I just reported the i686-linux build failure for file that I mentioned here the other day:
<nckhexen>I think h_errno should now be __h_errno. Did you ping jan·ekke about it?
<kaelyn>Not yet, because other than narrowing it down to a glibc upgrade I wasn't sure how to approach debugging or fixing it
<nckhexen>(Assuming Mes is to blame, you might know better.)
<nckhexen>Hm, no, never mind, I think.
<blake2b>nckhexen: yeah, its with geiser, but it only occurs when I'm using a repl server for whatever reason, ie $ guix repl --listen=tcp:8081
<blake2b>I can confirm that it also occurs with a repl server from guile ala $ guile --listen
<blake2b>If i simply launch a repl with M-x geiser, the package compiles
<nckhexen>Hm. I don't see any differences here, but to answer your original question: <does anyone know what this is about?> Not me, it's not familiar.
*blake2b cracks knuckles, M-x magit-blame
***rgherdt_ is now known as rgherdt
<nckhexen>$ egrep 'p(pc64|9)' && echo yay
*nckhexen TIL upstream considers [ef]grep dead & will remove it soonish. Clutch your pearls.
<vagrantc>so those recommendations to change my habits some 15 years ago might actually pay off
<efraim>cbaines: regarding grub and system installs, Guix calculates 'grub' and 'grub-32' and then figures out what to use. works for x86_64, not so much for other architectures
<singpolyma>nckhexen: I expect most downstreams will still ship them for at least another decade
<nckhexen>Debian will.
<nckhexen>But then Debian does everything for at least a decade, so that's not saying much.
<johnabs[m]><unmatched-paren> "Eh, I'll just put the links here..." <- Hey paren, I was finally getting around to testing your config suggestions, but I realized I had the following line in my config: (firmware (list linux-firmware)), do you know if this would this provide the i915 firmware by default, or are they separate?
<mroh>current debian sid/unstable prints an annoying warning on [ef]grep.
<unmatched-paren>johnabs[m]: Hmm. That should provide it, yes. But we're not supposed to provide help with nonfree software on GNU channels.
<johnabs[m]>(P.S. Sorry for the delay, my wife and I are both fighting covid; I really do appreciate your help so far, there's no way I'd have enough expertise to solve this on my own! )
<johnabs[m]>unmatched-paren: Oh okay, thanks for the input, I'll try a few more things I found from your config suggestions and see what I can come up with :)
<johnabs[m]>Oh boy, input lag still an issue :(
<johnabs[m]>Fully updated system, added microcode and such, but chromium/electron is still "the suck" :/
<PotentialUser-33>hola ""Debian incluye un cargador “shim” firmado por Microsoft, por lo que funciona correctamente en sistemas con el arranque seguro activado.""
<PotentialUser-33>saben si fuciona Guix con el Arranque seguro activado
<unmatched-paren>PotentialUser-33: (traducido automáticamente a través de <>) lo siento, no lo hace
<unmatched-paren>PotentialUser-33: Deberá deshabilitar el Arranque seguro en su BIOS :)
<PotentialUser-33>imposible por que estoy usando una pantalla secundaria
<PotentialUser-33>la pantalla principal está rota (por ahora estoy usando Debian)
<lilyp>win up to a ban by investing the minimum of a kick
***ChanServ sets mode: +o litharge
***litharge sets mode: +b *!*@2001:470:69fc:105::2:a756
***philip_morgan[m] was kicked by litharge (You are banned from this channel (by nckhexen))
***litharge sets mode: -o litharge
<nckhexen>litharge: You have disappointed me.
<nckhexen>You know what happens to bots who disappoint me.
<nckhexen>They get… reconfigured.
<lilyp>Garbage collected
<johnabs[m]>Remind me not to piss off the mods, that was brutal nckhexen, lilyp 😂
<nckhexen>The more these dorks force me to add new regexen, the less I advise you to chat about your crypto investments in this channel.
<nckhexen>☝ even sending that was a gamble.
<nckhexen>Technically, I think I could remove that ‘full message’ but… ugh. Effort. And Matrix. Matrix and effort! Me no like.
<nckhexen>johnabs[m]: Fleshbots—I mean, normal humans have close to 0% chance of pissing anyone off, don't worry. You have to go pretty far overboard.
<Lumine>It's true, I once delibaretly tried to piss people off but they just laughed
<nckhexen>Get well soon, by the way.
<johnabs[m]>nckhexen: Fair enough, though I'm an active member on mastodon and legit had someone convinced I was a bot because I liked sharing links to youtube videos 😂 And if you were referring to me with the "get well soon", I will :)
<nckhexen>Good. Obey. You both, BTW.
*nckhexen meant your wife, not Lumine. Sorry to be a bit rambly; I've been staring at backtraces for longer than should be legal.
<Lumine>Right, was wondering for a few seconds :)
<johnabs[m]>> * <> meant your wife, not Lumine. Sorry to be a bit rambly; I've been staring at backtraces for longer than should be legal.
<johnabs[m]>Thanks for the clarification, it was quite helpful because I totally thought you meant Lumine too xD And let me tell you, I understand the feeling, I was prepping for a conference and running I don't even know how many hours of simulations to try to collect data...when I found a bug
<johnabs[m]>I wanted to cry at that point
<johnabs[m]>And then I got sick and missed the conference anyway 😭
<nckhexen>Argh. Did you at least manage to fix the bug?
<jeko>Hey Guix ! Anyone using Guix Home to (1) configure Emacs ? (2) clone/pull a bunch of git repo ?
<unmatched-paren>jeko: I'm using it to configure Emacs with an Emacs home service I wrote.
<unmatched-paren>The emacs service is currently sitting in the mailing list, pending a few changes I haven't made yet.
<efraim>for the second I'm trying to decide if using package transormations to use --branch=master is better or to use the guile-git bindings to keep a repo up-to-date
<jeko>unmatched-paren: ohh I will make a quich search for that then
<johnabs[m]>I fixed the bug and collected the data, and almost finished the slides. On the bright side, we're gonna publish it soon, but I was looking forward to giving my first invited conference talk q.q
<unmatched-paren>jeko: I also added it to the guixrus channel so that I can use it before it's merged, so here it is:
<unmatched-paren>There *are* a bunch of improvements I should be able to make.
<unmatched-paren>lilyp: By the way, I'm not sure I'll be able to use search paths like you suggest to replace the (env-var ...) procedure, because of the seeming special-casing of Emacs directories with subdirs.el
<unmatched-paren>if i use search-paths, i'm pretty sure it'll return the #$output/share/emacs/site-lisp directories
<unmatched-paren>whereas it needs to return the #$output/share/emacs/site-lisp/$NAME-$VERSION directories
<unmatched-paren>and of course since those vary from package to package, i don't think search-paths will be able to handle them
<unmatched-paren>(hence their special casing to generate subdirs.el in profiles...)
<jeko>efraim: do you see those options to be part of a guix home service or do you think of something else ?
<jeko>unmatched-paren: thank you for the link !
<unmatched-paren>jeko: here's my home-emacs-configuration:
<efraim>jeko: I'm not sure how I would implement it, but it's not part of guix home ATM
<efraim>the package transformations to use a different commit/tag for git is already part of package transformations though
<unmatched-paren>jeko: Do you envision such a Guix Home feature as "if this directory does not exist, clone such-and-such git repo there?"
<unmatched-paren>i'm not quite sure what you hope to achieve with such a thing
<jeko>efraim: alright! as soon as i get my emacs config managed though guix home, I will do some experiments then
<jeko>unmatched-paren: thats what I was thinking about.
<unmatched-paren>how would that be useful, though?#
<jeko>when I setup a new PC I would like to have all the git repo i am working on at the right place after a guix home reconfigure
<unmatched-paren>sounds like you want a syncing service...
<unmatched-paren>because if you were to do it like you're suggesting, you'd lose all your unpushed work, and you'd lose your HEAD position
<jeko>the unpushed work is not what is bothering me. it's more the several and manual `git clone` I had to do
<unmatched-paren>well, okay then
<unmatched-paren>you'll probably want to do something with activation-service-type in a simple-service
<jeko>unmatched-paren: interesting !
<unmatched-paren>jeko: Basically, a ``simple-service'' looks like this:
<unmatched-paren>(simple-service 'name foo-service-type VALUE)
<unmatched-paren>this is equivalent to
<unmatched-paren>(service (service-type (name 'name) (service-extension foo-service-type (lambda (config) VALUE)) (default-value #f)))
<unmatched-paren>i think
<unmatched-paren>so, basically, it lets you extend a service without writing a whole new service-type
<jeko>yeah I find examples of this in other guixters'repo haha
<jeko>i ve found*
<jeko>I will then look for example of the activation service you mentionned !
<jeko>I m gonna sleep haha.
<unmatched-paren>bye \o
<jeko>unmatched-paren thank you for your advises ! I now what to do tomorrow ;)
<unmatched-paren>jeko: I *think* the VALUE for home-activation is simply a gexp to be run on reconfigure
<unmatched-paren>so you could do something like #~(for-each (lambda (repo) ...) (list "" ""))
<lilyp>unmatched-paren: you could use search-paths with the profile
<unmatched-paren>lilyp: so build an actual profile and use its search-paths?
<lilyp>build an actual profile and evaluate emacs' search-paths on it
<lilyp>or evaluate the profile's search paths, same thing really
<dgcampea>with, is it possible to use that for the certbot deploy hook? (
<Andronikos>Is it possible to see the actual logs instead of just build phase if I install packages via guix install <package>?
<nckhexen>dgcampea: Likely so, since that case doesn't involve changing the configuration.
<nckhexen>Andronikos: -v3 or higher.
<nckhexen>You can always—well, no often—view previous build logs with ‘guix build --log-file PACKAGE’.