IRC channel logs


back to list of logs

<vldn>maybe some kind of hook bot updated it's usernames because of reconnect duplicates?
<vldn>how are you connect to the irc channel @ jpoiret? :D
<nckx>vldn: Just a server having a small stroke, and the subsequent reconnection/auth indeed.
<nckx>Nothing to do with bots or bridges.
***ChanServ sets mode: +o litharge
***litharge sets mode: +b *!*@2001:470:69fc:105::2:75fc
***becky12345[m] was kicked by litharge (You are banned from this channel)
***litharge sets mode: -o litharge
*apteryx says farewell to mono in guix
*nckx shouldn't have said anything about bots.
<vldn>how to include modules written in wisp in the normale guile load path?
***sebbu is now known as sebbu2
***sebbu2 is now known as sebbu
<ytc>hello. i want to update the sbcl package to the latest version. but i don't know much about common lisp currently. i'm in the learning process. so i wonder if updated sbcl could break compatibility with the other sbcl related extensions and libraries. i would be glad if you can help me with that. thanks in advance.
***aeka` is now known as aeka
<dirtcastle>what is the shepherd equivalent of serial-getty@ttyS0.service
<dirtcastle>how to make shepherd list all services. looks like there is no command for that
<dirtcastle>what is the shepherd equivalent of serial-getty@ttyS0.service
<dirtcastle>how to make shepherd list all services. looks like there is no command for that
<podiki[m]>herd status
<podiki[m]>or sudo herd status for system shepherd
<podiki[m]>dirtcastle: ^
<podiki[m]>otherwise explore the guix docs to see what services are available
<dirtcastle>okkk. thanks.
<dirtcastle>does herd status list both active and inactive services?
<dirtcastle>I need to be able to use virsh console and use guix qemu image's console from host machine
<dirtcastle>I need to know the what serial getty's alternative in guix
<dirtcastle>mingetty seems to be the alternative
<dirtcastle>guys. guix qemu has a config at /run/current-system/
<dirtcastle> /run/current-system/configuration.scm
<dirtcastle>it's not updated.
<dirtcastle>it still has target and not targets
<dirtcastle>I just tried to run sudo system reconfigure using that configuration and it shows errors. I'm a noob. I thought I'll take this file , tweak it over and over and get a config I want in a vm and then do the installation on bare metal. it would be nice if some get this configuration working.
<dirtcastle>ok I fixed the errors and it's updating now.
<koleesch>Since yesterday i own a new I want to start with guix, but the wifi isn't working during the installation. Is it possible to get new iso images with a kernel with 5.15 or above?
<jpoiret>the latest installation image should have a 5.18 or 5.19 kernel
<koleesch>I used it  but it doesn't work.
<jpoiret>are you sure it isn't because your wifi chip requires some proprietary code/firmware/blobs?
<koleesch>here are the hardware specs...
<jpoiret>iiuc isn't like pine or purism
<jpoiret>the network card using iwlwifi is a sign of proprietary firmware
<koleesch>is it possible to activate wifi during the installation?
<jpoiret>koleesch: check your DMs
<jpoiret>ah, maybe you don't have access to those since you're on a web IRC client
<koleesch>now i'm writing from the laptop, but i've installed nixos
<koleesch>what do you mean with DMs
<jpoiret>direct messages
<jpoiret>sneek, later tell KE0VVT: Thanks to the core file you sent me, the segfault should now be fixed on master!
<sneek>Got it.
<koleesch>jpoiret; what should i find in DMs?
<jpoiret>my onw messages, but since you're on the web client I don't think you have access to those
<jpoiret>own *
***fps_ is now known as fps
<civodul>Hello Guix!
<sneek>civodul, you have 4 messages!
<sneek>civodul, wingo says: i believe it is a bug. in a pre-unwind handler, you want to make it so that if it throws again, it is handled by the handler that was current when the with-exception-handler was invoked (the dynamically outer handler)
<sneek>civodul, wingo says: to implement this, when a throw first starts, we capture the current handler stack, and arrange for subsequent throws to resolve via that captured handler stack
<sneek>civodul, wingo says: but if there is an additional with-exception-handler within the exception handler, that doesn't augment the captured exception handler stack to which inner exceptions will dispatch
<sneek>civodul, wingo says: the fix will be small but also gnarly :P will take some thinking. there are funny interactions with delimited continuations
<vldn>oh big merge?
<vldn>wrong channel sry :D
<jonscoresby>How does one build the documentation for guix?
<jpoiret>jonscoresby: the guix info documentation is installed alongside guix in ~/.config/guix/current/share/info/
<jpoiret>if everything works properly, you should be able to do `info Guix` to access the documentation
***taylan2 is now known as taylan
<jpoiret>don't we have gnulib-tool packaged?
<florhizome[m]>the webkitgtk-libsoup2 package is lacking substitutes for a day or so on staging is there hope for them to come soon :)
<jpoiret>ah, "gnulib does not make releases"
<ss2>hello, is cheese failing to build for anyone?
<ss2>, the source uri is throwing a 404.
<civodul>ss2: same here; mbakke: looks like the URL for Cheese is wrong :-)
<ss2>It should be 41 instead of 41.0.
<civodul>ah, i can fix it then
<civodul>lemme see
<civodul>ss2: fixed!
<ss2>okay, cool, thanks!
<florhizome[m]><florhizome[m]> "the webkitgtk-libsoup2 package..." <- I have written a few new Services for my Hardware that i would Like to test further
<vldn>somebody using a guix mail-server?
<vldn>what services do i need? Dovecot isn't enough is it? didn't i need postfix or something?
<vldn>ah i see opensmtp
***wielaard is now known as mjw
<Kabouik>Anyone using Wayland and Sway on Guix? I was thinking changing distribution was a good occasion to transition to a more modern compositor as well. I have many issues with my install remaining that I don't know how to solve yet, so it can only marginally get worse anyway. :>
<Kabouik>I have Sway and Wayland install already, but GDM only offers Xorg DE/WMs
<Cairn>Hey guix. How would you specify a commit for only a particular channel on a guix pull (when using multiple channels)
<vldn>mybe try launching it from the tty? via "sway" @kabouik
<Kabouik>Will try that, thanks
<vldn>i think there are somewhere custom definitions for a startup herd service
<vldn>i launch my dwm on login to my tty automatically via [ "$(tty)" = "/dev/tty1" ] && sx dwm
<Kabouik>I found something about "sddm" but that sounded complicated (for me) and I was afraid that if it failed, it could just be me failing to implement it correctly in my config.
<ss2>Kabouik: not quite, once you've enabled wayland you can load sway with it too.
<ss2>which is what I've set up now.
<vldn>ah that would maybe easier for your setup Kabouik - ss2
<Kabouik>How do I "enable" wayland?
<ss2>something like that.
<Kabouik>Thank you ss2
<ss2>Then sway becomes loadable. But it got a bit confusing for me, since .xsession isn't read as expected anymore and have since put all my nice startup programmes into sway/config. Not sure if it is the best option, though it works for me.
<Kabouik>I never used Wayland and am expecting many issues until I learn alternatives to setxkbmap, .Xresources and so on, anyway
<ss2>I had surprisingly few problems. Though I moved from i3 and had brought the kitchen sink with me. Which kind of helped.
<mbakke>civodul: thanks for taking care of Cheese! I wonder how the updater managed to construct the correct URL
<Kabouik>I'm coming from i3 too but am having many problems with my install on this machine so far (some of which I don't have in xfce4). Trying to change to Wayland at the same time may be a bit bold since I don't know my way in Guix and Scheme, but if it pays off, at least I'll be happy with this setup for years to come.
<abrenon>hi guix
***Dynom_ is now known as Guest1767
<civodul>mbakke: yw! the updater got the right URL but it didn't see that it doesn't match the code
<pkill9>is (eval) considered a dirty hack in guile world like it is for other languages?
<notantipode>pkill9: depends on what you are using it for
<notantipode>If you are thinking of JavaScript and json, read is your friend
<pkill9>i want to make scripts you can run directly, or import
<pkill9>currently if you import the script into a guile repl it runs it
<pkill9>because it evaluates the thing, but i probably need to do some more fun with the shebang to get it o only run when called
<notantipode>Import = import module?
<notantipode>Or: read into memory first, run later?
<muradm>hello guix
***sajith- is now known as sajith
<Kabouik>Awesome, I just tested and I'm in Wayland now ss2. I had to disable autologin or it would default to X11 and i3, maybe because it uses the last used session by default. I had a few adaptations to do in my Sway config but not many, it's running fine now. Glad to see Sway has gaps like i3-gaps, I didn't know.
<Kabouik>Now I'll still need to find alternatives to setxkbmap, xrandr, and a bunch of other things, in hope it will resolve part of the issue I had before.
<muradm>Kabouik: going over this list was helpful for me in migrating from from i3 to sway
<Kabouik>Apparently I starred it in the past, but forgot about it! Perfect, thanks.
<Kabouik>I'm surprised that some X programs still work fine, like dmenu and rofi (although I already had sway-launcher confifugred, so I'll use that isntead for now, to stay as Wayland as possible)
<Kabouik>They probably work through Xwayland.
<Kabouik>How is one supposed to alter udev rules in Guix, like for the `light` utility that alters screen backlight? Their README doesn't seem guix-compliant:
<abcdw>muradm: o/ Kabouik: Just add a service like this one:
<abcdw>This system service automatically picks udev rules from package and adds them to /etc/udev/rules.d/
<Kabouik>Thanks, will look into it. Meanwhile I noticed that Wayland fixes my screen tearing issues. It's still choppy in some tearing tests (that one in particular: but not tearing, so I'll be able to simplify my config.scm that contained some Xorg extra config.
<Kabouik>So when I add this service, I just have to reinstall `light` so that it picks the udev rules shipped with it?
<abcdw>You don't have to reinstall anything, just guix system reconfigure ./your-config.scm
<abcdw>Also, make sure to change brigthnessctl to light in the service definition.
<apteryx>GNUtoo: I've fixed the FSDG problems in phoronix-test-suite. You might want to put it to the test :-)
<apteryx>I've eye-balled the complete list of tests marked as Free, and could find many non-free games in there, but with thousands of tests, there's a chance I may have missed something
<ss2>Kabouik: ah, reasons to use auto login is obviously your choice. I'd just had mine open and copied it out for you. Good luck, yeah, there are not so many, but good lists on migrating and setting things up. If you know what you need and are already coming from i3, the transition will be smooth.
<pkill9>notzmv: import module I mean
<Kabouik>Yes, I think I'm already closer to functional state in Sway than I was in i3 an hour ago. Now the volume buttons are recognized (although I still have no audio device identified by pulseaudio) and I can change brightness from commandline, I just haven't managed to use abcdw's solution above for udev ruls yet (I thought I would just need to use the three lines from the line-specific URL they posted, but that is not enough and I might need some of the
<Kabouik>code around it.
<reyman_aw>Hi ! \o/
***reyman_aw is now known as reyman
***reyman is now known as reyman_aw
<Kabouik>I have gotten sidetracked looking at other abcdw's rde files and I find it fascinating that there are so many .scm files, I had no idea people were setting up their environment that way. I'm no discovering that there are no dotfiles whatsoever, just scm files defining base dotfiles when deployed. I don't think I can achieve that level of feng shui yet though, I'm stuck with old fasioned dotfiles for now (which are a pain to deploy and adapt to new
***reyman_aw is now known as reyman
<Kabouik>Sorry for the nuerous typos, hope it's legible still.
<Kabouik>numerous*. Damn.
<reyman>I'm searching example of someone that use zsh + ohmyzsh in guix ? I use that in Ubuntu, not sure it's possible with guix
***rgherdt_ is now known as rgherdt
<Kabouik>I'm tempted to try abcdw's whole rde tree while I'm still early in configuring my WM and other base functions, that way if that works I can just try editing scm files instead of feeling the burden of my regular dotfiles not written in scm features…
<abrenon>how can I simply package two local files ? they're not versioned, they're scripts I just wanna copy over to another machine with their dependencies (interpreter + libs)
<nanami>source can be local-fille
<abrenon>I've been fighting with (local-file …) but get errors I don't understand, even with #:recursive? #f I get errors for files deep-down in the subtree where my scripts are
<apteryx>is there a way to prevent guix lint from looking at the CVE database; it has a matching entry but it's not the same package
<abrenon>trouble is: if I pass only one file-name it works, but it doesn't seem to like two of them
<abrenon>passing doesn't work either
<abrenon>I don't want it to recurse (it's a huge folder with many large files), I know the two files I need to pack
<nanami>try file-union?
<abrenon>looks promising, thanks !
<abrenon>(is there any reason why local-file with the #:recursive? flag set to #f wouldn't work though ?)
<nckx>apteryx: Does setting the 'correct' (assuming one exists) CPE name not suffice?
<apteryx>it doesn't exist, that I can see. I'm now trying (properties `((cpe-name . "bower-cpe-refers-to-a-different-bower")))
<apteryx>hoping in never finds anything
<apteryx>seem to work
<abrenon>ok, seems to work for source… now to see how that behaves with install-plan
<civodul>apteryx: hi! thanks for the fwupd followup, i suspected the default behavior wasn't great
<apteryx>I couldn't really test if it (prevents me from installing non-free firmware by default) because the only UEFI system I have available doesn't seem to request any.
<apteryx>but you're welcome!
<abrenon>no really, I can't make head nor tail of the error messages
<abrenon>saying that a pattern matching fails without saying where is about as useful to me as saying "it doesn't work"
<abrenon>it's really terrifying a feeling to try and plug something into a tree without knowing clearly what is acceptable there
<abrenon>a function ? a string ? should it be in the gexp monad ?
<abrenon>no idea
***reyman is now known as reyman_aw
<abrenon>so I have managed to generate a folder inside the store with only the two files I'm interested in there, but get this weird matching error
<abrenon>I've commented the arguments and everything, so I have no idea what guix is doing, I wonder if maybe my origin doesn't work after all
<abrenon>the scheme file (with all the ugly commented-out previous attempts)
<abrenon>is maybe the copy-build-system at fault here ? would it be simpler to read about the trivial build system and simply copy files one by one ?
<nckx>Why was fwupd packaged? To get to something else?
<nckx>abrenon: But it's mapping over systems (IIRC), so weird that it would be file tree related.
<nckx>Or was that not your point.
<abrenon>my point was that I'm attempting things being my ability ^^
<abrenon>now I think I follow what you mean by "mapping over systems", but I have no idea if that should apply to origin, install-plan or anything
<abrenon>I'm just ready to bet there's a very simple way of putting two simple scripts with their dependencies into a package, and I realize I have no idea how : (
<ss2>It'd be so nice to see emacs-guix receiving a bit more love. :/ It appears geisers recent update brakes emacs-guix now for good.
<apteryx>nckx: I don't know; I guess the submitter had a need for it
<ss2>I just tried this repo, that had been recently discussed in devel:, which builds fine, but fails.
*apteryx only uses emacs-guix for guix-scheme-mode and guix-log-mode
<nckx>That seems very likely. I just wondered if it was useful on its own, or if any real value in practice was provided by the LVFS.
<apteryx>probably the later for now, until someone offers their own service packed with free software only firmwares
<apteryx>it also seems to be built to run as a service, although that's perhaps optional
<ss2>Its so handy to use it with several profiles. Anyway: Geiser removed calls to company, while emacs-guix still calls it while loading the repl. Will see if removing it will solve it.
<Luk6655>Could anyone show a working example of "modify-services" , please? One in the manual fails with "invalid field specifier"
<nckx>Luk6655: Please share the failing file.
<Luk6655>ss2: I'm getting the same error, just a sec I'll review the entire file sent by tricon, perhaps I'm not using it right
<ss2>What are you trying to modify?
<nckx>It is unlikely that the exmalpe in the manual is wrong.
<Luk6655>well, the example in the manual is _definitely_ wrong
<nckx>No we don't make imtaskes.
<Luk6655>for one, it looks for guix-service-type in %desktop-services
<Luk6655>while it is in %base-services
<Luk6655>however, this is correctd in modify-services source code
<Luk6655>here is the failing file I just tried to paste the example given by ss2
<ss2>uh, please don't just copy it. That example is mine, and will most likely fail.
<Luk6655>why would it "just fail"? It'n not like I'm going to run my system with it.
<ss2>which is not part of the manual. ;)
<Luk6655>I'm just trying to get _any_ modify-service that works
<Luk6655>does your example work in your config file?
<ss2>sure, but that because I'm loading it with %desktop-services and using gdm too.
<Luk6655>I too use %desktop-services and gdm
<Luk6655>still, I've tried the example given in guix/gnu/services.scm (which is the corrected version of what we have in the manual) and even that fails with the same error
<Luk6655>so I'm either using it wrong? Or there is some bug.
<nckx>You're using it wrong.
<Luk6655>please elaborate
<ss2>Is it the first time you are trying to modify a loaded service?
<nckx>This is what your services field looks like: (services (append stuff %desktop-services) (modify-services %desktop-services => stuff))
<nckx>Hence the invalid field specifier; fields can't have 2 values.
<nckx>Luk6655: Now please elaborate on which manual example is wrong, and why 😛
<Luk6655>fine ;-) this one
<nckx>What error do you get?
<Luk6655>go to bit that starts with: (modify-services %desktop-services
<nckx>Yep, thanks, that's what I'm looking at. Looks good at first glance.
<Luk6655>it should read %base-services, as guix-service-type is in %base-service
<nckx>Or %desktop-services.
<ss2>%base-service is part of %desktop-services.
<Luk6655>ok, lets briefly go back to my config
<nckx>It would help if you shared the actual error message.
<ss2>If you're loading desktop, you've automatically pulled base in too.
<Luk6655>so what would be the way to modify a service that is in %desktop-services?
<nckx>That's what the example in the manual does: (modify-services %desktop-services …)
<Luk6655>it was mentioned this: (services (append stuff %desktop-services) (modify-services %desktop-services => stuff)) is wrong
<nckx>Yes. Should be (services (append stuff (modify-services %desktop-services => stuff)))
<nckx>You were appending your services to %desktop-services and then appending that to a modified set of %desktop-services.
<Luk6655>wait.... so is (modify-services %desktop-services => stuff) returning modified %desktop-services
<nckx>You can't say it's not aptly named :)
<Luk6655>well, I assumed it modifies a global variable called %desktop-services and returns nothing
<abrenon>it's pure
<abrenon>it computes a new value
<nckx>No, we don't do or speak of such horrors here.
<nckx>Also, not sure how that would even work.
<Luk6655>not that return values are described anywhere.... but I might be missing something obvious to everyone
<Luk6655>ok, this does explain it
<Luk6655>I'll try it
<nckx>I'm going to assume the example in the manual is fine then?
<abrenon>: )
*nckx closing… in brain.
<Luk6655>that's what I was going to clarify next , does %desktop-services contain everything from %base-services?
<Luk6655>ok, so the example is fine...
<nckx>It is a superset.
<Luk6655>nevermind :-)
<Luk6655>thanks for clarifying it
*nckx , aside, is being remarkably productive on a rootfs that went emergency-ro 2 hours ago.
<nckx>Thanks for reporting it nonetheless! :)
<tricon>nckx: ya fixing it with one hand and eye while chatting on IRC with the others?
<nckx>tricon: Not even that, more important work to do first (no, not the IRC part, that's during busywaits :)
<nckx>Luk6655: Not related, but I noticed ‘dr.who’. I guess that's a placeholder for a real user?
<ss2>nckx: that came from my example posted higher up.
<Luk6655>nckx: yes, this entire snippet was copied from an example shown just to text if the syntax is correct
<Luk6655>my actual config modifies elogind
<nckx>If it works in practice, all the better, but note that it's not guaranteed to.
<nckx>Luk6655: I see. Sorry for the confusion.
<nckx>(See man useradd(8) for why ‘.’ isn't technically valid, but I'm almost certain there are ill-advised custom regexen out there.)
<nckx>‘They can end with a dollar sign.’ — of course, because disallowing *that* would be madness‽
<tricon>nckx: hey, now, some of us are nostalgic for C64 BASIC.
<nckx>Soundcloud rappers rejoice.
<nckx>abrenon: I don't have time to dive in, but you can use --disable-grafts for now. It does look like a legitimate bug, even if this is a very unusual way to write a package.
<Luk6655>btw, is there some way to check the system config.scm syntax without running guix system reconfigure? Something like --dry-run for reconfigure?
<nckx>abrenon: It seems to be some interaction of file-union ←→ grafts ←→ propagated-inputs, because removing any of those ‘fixes’ it.
<nckx>Luk6655: …yes?
<Luk6655>is it actually --dry-run? (There doesn't seem to be a man page for guix system reconfigure)
<abrenon>really ? I was just playing with forces I don't understand : S
<nckx>Luk6655: Yes, it is. The full documentation for guix is maintained as a Texinfo manual, not man pages.
<abrenon>I have no idea if that triggered a bug
<nckx>Luk6655: There's also ‘guix SUBCOMMAND --help’.
<abrenon>I'll keep the file around if you want to dive in some day then (and I'd gladly take the explanations of what would be expected or not)
<nckx>abrenon: I suggest using git-checkout to continue, even if this means you'll have to commit all changes.
<nckx>That seemed to work anyway.
<abrenon>yeah, what I eventually did, I thought I was really stupid to spend so much time to try and understand gexps when I really needed was the script to run on another machine
<Luk6655>nckx: good stuff, thanks
<abrenon>which it has been doing for approximately half an hour now : )
<nckx>abrenon: You could post it to the mailing list as a puzzle. I'd be curious to know if there's an obvious reason behind the madness.
<abrenon>(you mean appart from my not understanding what there is expected to be a gexp and what isn't ? ^^)
<abrenon>I'll try it then when I have SSh access to that second machine and I can watch over the script during the weekend
<nckx>Luk6655: I interrupted myself whilst typing: there's also ‘guix system build’, which builds the entire system just like reconfigure would, but does nothing with the end result. A very thorough test of sorts.
<nckx>Any reconfigure after that will be close to instant, modulo building the bootloader and some minor derivations.
<Luk6655>This is even better, I can definitely see myself using it a lot.
<nckx>Great! It's overkill for checking your basic system.scm syntax.
<nckx>(While it has a few rough edges, like ignoring your file-systems for technical reasons, and of course missing hardware, I recommend checking out ‘vm’ and the other subcommands, there is more coolness beyond ‘reconfigure’.)
*nckx reboots. Apparently /run being read-only is a huge deal that needs to be whined about. o/
<podiki[m]>abrenon nckx what bug? i saw something grafts, union, propagated inputs and I got flashes of
<abrenon>hmm unsure (and a bit clueless to be honest) but my union was a file-union, advised by nanami
***cwebber` is now known as cwebber
<abrenon>also nckx mentioned grafts but I had no idea my snippet was doing grafts (if it is indeed doing so)
<podiki[m]>what snippet? (sorry jumping in late)
<abrenon>it's really ugly because I was in the middle of things, it shows my multiple attempts
<abrenon>I shared it because I thought like last time it would only be my trying and re-trying things I don't properly understand, and having made an obvious syntactic blunder
<podiki[m]>not that it is grafting, but that some input in the tree has grafts probably
<podiki[m]>this does seem a bit similar to 53406, but not sure
<ss2>I posted a quick fix regarding emacs-guix. Would be nice to get it rolling again.
<tricon>what seems more popular these days for dev envs? guix.scm or manifest.scm?
<Luk6655>Could someone have a look why my modify-services fails this time, please? Here is the config (relevant part around line 89) here is the error:
<Arya[verncc][m]>use what you prefer :P
<tricon>Arya[verncc][m]: ^_^
<rekado_>Luk6655: line 90
<rekado_>you need to wrap that expression in parentheses
<Luk6655>rekado_: ok, let me try that
<rekado_>modify-services takes a list of services and then any number of clauses; each clause is an s-expression
<abrenon>podiki[m]: nckx that'll be for another day unfortunately sorry : )
<rekado_>mbakke: okay, I’ll rebase on top of staging
<dthompson>tricon: I'll be opionated: I recommend guix.scm
<dthompson>all of my projects have a guix.scm file that evaluates to a package expression that can be used with 'guix shell', built with 'guix build -f guix.scm', and installed with 'guix package -f guix.scm'
<Luk6655>rekado_: I did that, now I have a different error: extraneous field initializers (hibernate-key-ignore-inhibited)
<Luk6655>this is the updated config
<dthompson>tricon: here's an example:
<wdkrnls>Dear Guix, suppose I'm developing a script that might have a dangerous bug in it. How can I call the script so that any changes it makes to the file system outside of a set of provided paths are rejected and immediately cause the script to stop executing?
<ss2>how did I manage this:
<ss2>I messed up a generation and can't delete it now.
<podiki[m]>abrenon: no worries, it is a weird bug
<oriansj>ss2: does guix package --roll-back not work?
<podiki[m]>could just manually rename that bad link name? not sure what happened, maybe a wrong profile argument or something?
<ss2>fancy stuff, no idea how I created this though. Must have misstyped something in a shell somewhere.
<podiki[m]>yeah, could have used the wrong name or how you parsed something, I had some weird profiles when I was first getting my profile looping script working
<wdkrnls>I'm thinking maybe I can do this with guix shell --expose, but the examples in the manual only refer to containers... which is probably what I am looking for anyways.
<ss2>good call, that worked.
<tricon>dthompson: yeah, that's what i've decided to go with. i was curious as to what was popular these days. i like the convenience guix.scm provides with the ease of publishing the package definition.
<tricon>(haunt is great, btw.)
<podiki[m]>wdkrnls: yes, sounds like you want to run in a container, that will isolate it completely except for what you choose to give the container
<podiki[m]>ss2: great!
<rekado_>Luk6655: need a trailing question mark
<rekado_>the option is called hibernate-key-ignore-inhibited? not hibernate-key-ignore-inhibited
<Luk6655>I believe I tried that before, but let me try again
<wdkrnls>what is "resp." short for?
<Luk6655>well, would youlook at that..... it works :-)
<Luk6655>I must have had a typo in it when I tried it with the question mark
<wdkrnls>In the context of the guix shell documentation, I am guessing it means something like "another name for command X".
<wdkrnls>But I have no idea what actual word it refers to.
<podiki[m]>yes, respectively, as in it is documenting two options at the same time, noting the differences
<podiki[m]>expose vs share (expose read only, share write allowed)
<rekado_>mbakke: is it fine to push the update-vulkan-headers commits to staging? Or would you prefer to replace the update-vulkan-headers branch with the one that’s been rebased onto staging?
<wdkrnls>thanks, `podiki[m]`. I struggle a lot with remembering abbreviations and acronyms. I'll add this one to my lookup table :)
<podiki[m]>I don't know if that is common necessarily, mostly from context I would get that one
<wdkrnls>My "parsing intuition" is notoriously bad as my brain seems more interested in sorting by novelty than by plausibility. I might hear Funny Elephant Analysis rather than Finite Element Analysis for example in a noisy room.
<wdkrnls>But your interpretation makes tons of sense "now that I know".
<podiki[m]>I often find respectively not very necessary, but sometimes it does make things nice and compact, as in this example (probably would read okay without the "resp." but might be less clear)
<wdkrnls>For me it's not clear, but it is definitely precise.
<wdkrnls>I have seen a sort of "or" syntax used instead.
<nckx>Respectively is fine, but ‘resp.’ is the kind of needless abbreviation that Guix normally avoids.
<nckx>It's now officially led to more innocent bytes being wasted than were ever saved.
<unmatched-paren>nckx: I agr.
<wdkrnls>Presumably multiple --expose and --share are possible as well?
<wdkrnls>I didn't see it documented and haven't tried it, but my intution would say it exists.
<wdkrnls>Unfortunately, that intuition doesn't always work well.
<nckx>The first ‘Invoking guix shell’ example has 2.
<wdkrnls>Oh, it does at that. I came at it from below and missed them.
<wdkrnls>I must have read it earlier and forgot.
<wdkrnls>That makes more sense to me.
<wdkrnls>It gets rid of resp. and respectively as well.
<nckx>Heh, I saw the ‘resp.’ when I looked up ‘--expose’ and was thinking the same thing. Would you like to submit a patch?
<nckx>One could also just document --expose in full, then have --share as its own thing with ‘literally exactly --expose, but the mount is writable’ or something.
<nckx>Or the other way 'round, respectively.
<nckx>Juggling them both was an admirable attempt but I don't think it worked, it just adds mental overhead.
<davidl>nckx: hi, do you have any estimate on when you think you will get around to merge the bash-bcu patch?
<nckx>I said this weekend, that hasn't changed.
<nckx>I do not have access to my signing keys.
<davidl>oh, I must have missed that, then Im sorry to nag you
<nckx>I thought you replied with something about a PM with changes? (Which I'll also see then; mobile client no do PM gud.)
<lispmacs[work]>Hi, I am curious about that new "cardpeek" package, i.e., if anybody here has used that successfully with any particular brand/model of cardreader?
<wdkrnls>I would try to submit a patch, but I haven't bothered to figure out how to make sure my change doesn't break things. That, and at the moment I am procrastinating on an urgent deadline.
<wdkrnls>I was hoping there would be a `rebuild-the-docs` script that didn't involve the intimidating `pre-inst-env` stuff.
<nckx>‘make doc/’?
<nckx>Please tend to your deadline first.
<nckx>There is no need for ./pre-inst-env. At most a ‘guix shell -D guix’, and that should really not be intimidating.
<nckx>lispmacs[work]: I'd never heard of it, but wow, this looks nifty! Even comes with a built-in BE ID schema; nice. Works fine here, with an ancient ACS ACR 38 I got for free. I usually use a Digipass 875 which I can test once at home, if you want.
***owsley is now known as Guest7666
<lispmacs[work]>nckx: if you could test that Digipass, that would be great. Maybe I'd by one just to play around with it
<nckx>Well, it's just a card reader, not the most exciting toy, but sure, will do.
<nckx>Are you having problems?
<nckx>If so it might be a matter of tweaking your system configuration. I already had this machine set up to read my ID card through Firefox.
<lispmacs[work]>nckx: I just don't have any card reader yet to play around with
<tricon>i need to setup my unlocked Solo. been meaning to do that.
<nckx>Card reader's practically mandatory here.
<nckx>I don't have one of them contactless gizmos, though.
<nckx>tricon: What about it needs unlocking?
<tricon>nckx: With the developer version, you can reflash the firmware.
<tricon>i like that they offer both options. same price too.
<nckx>Yes, that is surprisingly enlightened.
<nckx>Is it Free?
<nckx>(Not going to kick you out! Just curious.)
<tricon>it is!
<tricon>(sorry for the GH link, but it is what it is.)
<nckx>It is, and it's fine.
<nckx>tricon: So this was very confusing: because of the context, I was looking at <>.
<tricon>that _is_ confusing.
<nckx>Had that thing been Free software I literally would be buying one right now.
<nckx>But! The key looks cool too!
<nckx>(So ignore my ‘contactless’ confusion above. The key is very contactful.)
<drakonis>its mit or apache 2.0 licensed
<nckx>I lies:
<tricon>i just have the SoloKey1 Developer. i don't think the NFC version was out when i purchased it; either that or i opted for no NFC as i would like to rid myself of a cellphone some day.
<tricon>still, i definitely like the idea of Bring-Your-Own-*-Factor
<tricon>related: Dashlane is starting to push passkeys, which iOS 16 will support. i don't know all the impl details, but i cringe at trusting other parties with my keys.
<nckx>I'm not familiar with Dashlane.
<nckx>(Or iOS, but that's rather less surprising :)
<tricon>it's not free software. i only bring it up as an interest in what is going on in the tech sector as these things can have reaching implications.
<tricon>yeah, i just use a work phone at the moment. hoping to build out my Pine64s, one as an audiobook player, the other as a severely limited phone (basically a home handset).
<nckx>I want to clarify that it's not forbidden to mention non-free software here, as long as it's not simply off-topic. There's a thick grey line between promoting and discussing, but most people can tell the difference.
<tricon>yeah. i figured that was the line, and appreciate the clarification.
<nckx>Hey check out this cool Dashlane < | curl bash> is a different matter.
*nckx notices ur cloak. Ah. Clarification was probably not needed. Still.
<vagrantc>lispmacs[work]: i only know cardpeek doesn't build reproducibly and i couldn't come up with a workaround or fix :(
<vagrantc>i have a couple solokeys that were rejected because whatever wireless technology didn't work (so much the better, in my opinion) but i never managed to carve out the time to use them
<vagrantc>pretty happy with my gnuk devices anyways :)
<vagrantc>parts shortages aside, you ought to be able to build yourself a gnuk...
<dthompson>I usually only peek at the nix community through the lens of HN threads but I've noticed several times that they believe a lazily evaluated language is core requirement, and sometimes they try to square that with the fact that guix despite using an eagerly evaluated language.
<dthompson>that guix works*
<nckx>My first-ever message to this channel was literally (well, paraphrased) ‘how does this guix thing even work if u no hav lazy language lol?’.
<dthompson>makes me wonder what communal truths we may have that don't really hold up
<dthompson>nckx: lol
<dthompson>"I don't even understand how Guix manages to work without laziness, though clearly it does somehow."
<dthompson>I wish I knew a little more about nix the language so I knew how to explain why guix works.
<nckx>Well, obviously I saw it worked and was willing to give it the benefit of the doubt, and look what happened. Let this be a warning.
<singpolyma>dthompson: what do you mean "without laziness"?
<singpolyma>Guile has as much laziness as you want
<singpolyma>Oh, I see I didn't scroll up far enough
<nckx>Yes, but it's managed explicitly where needed, not magically everywhere.
<nckx>I'm not saying magic is good or evil, but it sure is magic.
<nckx>Hi two[m].
<two[m]>getting this error on guix push
<SubberUser>Hi people
<two[m]>substitute: updating substitutes from ''... 0.0%guix substitute: error: TLS error in procedure 'write_to_session_record_port': Помилка у функції push.
<nckx>(I'm guessing ‘guix pull’, but ‘error in the pull function’.)
<singpolyma>Default to lazy is nicer for sure, but that's just ergonomics the feature is equivalent either way
<nckx>Dammit, now I did it. *push function.
<nckx>singpolyma: It's not always nice.
<nckx>two[m]: Does it keep happening each time?
<two[m]>it happened twice
<nckx>Very unsatisfying answer: try a few more times, if it still happens then, try to share the full output (if there's more).
<two[m]>yes it's error in function push
<antipode>on non-niceness, if you have lazyness by default, you need to be extra careful with exception handlers and other dynamic environment things (stack limits, parametrisations/fluids, call-with-blocked-asyncs, prompts, ...?)
<antipode>with 'lazyness', Nix folks don't seem to be just referring to lazyness, but also to self-modifying code (overlays overriding the things they are overriding?)
<nckx>By the way, two[m], I can't verify it now but I think running ‘LC_ALL=C guix pull’ should print English error messages, which will help others debug the problem.
<nckx>It's likely already known, but still worth reporting IMO.
*nckx AFK.
<antipode>Which Guix also kind of has, in the form of package variants inheriting from the original and adjusting the inputs (modify-inputs) or arguments (substitute-keyword-arguments), but it's more separated from the 'lazyness' concept and it appears to be different.
<two[m]>thank you
<two[m]>it's working now though
***mark__ is now known as mjw