<jared-w>
It's not necessarily bad, it's more that it doesn't always align well with the nix philosophy
wildtrees has joined #nixos
<jared-w>
For example, my zsh configuration uses a zsh package manager that installs programs like diff-so-fancy which is referenced in my configuration.nix file. So my configuration.nix file has hidden dependencies on a private git repo which has hidden dependencies on some packages installed by configuration.nix
<mizmor>
ok i see!
shibboleth has quit [Quit: shibboleth]
<jared-w>
Since nix is purely lazy, it's not any less efficient to have every single aspect of your system require every single thing it needs (requiring 20 packages 40+ times is the same as requiring them only once), but has the benefit that each aspect of the system is now fully self-documenting. Which is the point of nix; reproducible builds :p
<ddima>
mizmor: people attempt to have fully reproducible systems, as far as possible (with some more "extreme" configs where people use zfs snapshots or lvm to reset the filesystem etc). pick your degree :)
<jared-w>
A lot of the complexity in how people use nix comes from wanting fully reproducible builds but also the convenience of "implicit connections" so people build modules, pin things, add abstractions, etc., to prevent repetition that comes from exact specification where 90% of that specification is gonna be the same every time
cosimone has quit [Ping timeout: 250 seconds]
<jared-w>
Things get trickier when you want to fully specify things in an OS agnostic way, which adds more complexity
<ddima>
it's nice if your whole dev setups, editor configs, firefox+plugins etc just falls out with as few tweaks as possible - esp if you for instance have a desktop and a laptop where you'd want to have the same experience with the exception of minor differences due to hardware/screens what not.
aither has quit [Ping timeout: 276 seconds]
<mizmor>
Yeah i like the idea of a reproducible system but i'm not at that level of knowledge yet! I really like nixos for the dev setup/editor config part tbh
<jared-w>
That's the part that's actually most annoying for me at the moment lol
<mizmor>
hahaha oof
ng0 has quit [Quit: Alexa, when is the end of world?]
dolanbatar has quit [Excess Flood]
<ddima>
mizmor: if I may joke for a second - in my experience there is a rite of passage of "going long enough until stockholm syndrome sets in". so dont worry about how you start, just make it your own.
<jared-w>
I have a decent chunk of stuff in configuration.nix dedicated to getting HIE and ghcide working with a cache so I don't have to build everything, and it sorta kinda not really barely works
hexo has joined #nixos
aither has joined #nixos
monsieurp has joined #nixos
<jared-w>
Really HIE is just a completely massive waste of CPU cycles and the haskell IDE experience _sucks_ compared to typescript. So you just kinda have to figure out what you're willing to tolerate lol
<mizmor>
ddima: thanks, i'll figure things with this in mind!
<clever>
jared-w: most of the time, you can get away with just: runhaskell Setup.hs configure ; ghcid -c "runhaskell Setup.hs repl"
<jared-w>
yeah I need to just bite the bullet and setup ghcid and just use that + ghcide (which works great)
<clever>
there is nothing to even configure with ghcid, you just run it on a repl/ghci command
<jared-w>
I got used to having perfect "go to definition", "rename symbol", instant docs, etc. but having ghcid and ghcide is really all I end up _really_ wanting in haskell dev
<jared-w>
(which basically just amounts to some types in the editor, and pretty decent compile-checking on type)
halfbit has joined #nixos
leungbk has quit [Ping timeout: 245 seconds]
<jared-w>
really miss autocomplete + autoimport though. Fuck that was a huge time saver when using typescript. _sigh_
werner291 has quit [Read error: Connection reset by peer]
mexisme has quit [Ping timeout: 265 seconds]
<clever>
jared-w: i use youcompleteme in vim, and in the default mode, it just auto-complete any symbol found in any open file
<clever>
jared-w: so if i open the other files in tabs in vim, it can complete the functions they export
bvdw has quit [Read error: Connection reset by peer]
jgeerds has quit [Ping timeout: 246 seconds]
<jared-w>
Yeah I can do that too; there's just a world of difference when you get type aware, language aware autocomplete with inline docs/types/etc instantly on hover and hitting tab will complete things, fill in the parenthesis etc., and autoimport if needed
bvdw has joined #nixos
<clever>
yeah, ive had some of those with java before in eclipse
werner291 has joined #nixos
<clever>
and youcompleteme gives some of that for python
<jared-w>
One means you need to have a vague idea of what you're looking for and the other one means you can literally install a library and figure out how to use it just by hitting "what completions do I have right here?" It's awsum
<pbb>
if anyone is interested I packaged u-root and made a module that generates a simplified grub config that it can parse, now using it as my main bootloader
<jared-w>
coc.nvim does it very well; but it's entirely dependent on the quality of the language server. Typescript's is the best out there by far, but rust's rust_analyzer isn't terrible. Need to play around with some others
<jared-w>
pbb: any particular reason to use that over systemd-boot?
<pbb>
jared-w: I am using coreboot and systemd-boot requires some EFI implementation and I don't want/need that
kaba has quit [Quit: Leaving]
<pbb>
I think it's just cool to have a small linux kernel in my flash and use that to kexec my main kernel, because it has endless possibilities
<jared-w>
ahh gotcha
<pbb>
it can do fast dm-crypt and stuff
<jared-w>
yeah I see now it's for way more embedded or smaller environments where systemd is a nogo
<clever>
pbb: ive wanted similar to get ssh access to the bootloader, so you could access the rollback menu
<jared-w>
looks neat
<clever>
pbb: then you could unbrick things in the cloud
<pbb>
yeah x)
<pbb>
I have managed to squeeze in almost a full kernel now, including wireless drivers and stuff
<clever>
samueldr: ^^ and it could be used as a bootloader menu on android devices
<pbb>
you could also use u-root on systems that have a traditional bootloader
jlv has quit [Remote host closed the connection]
<samueldr>
android-based devices will need something that can deal with touch input, or very least volume up/down + power
<pbb>
for example a cloud vm could first boot u-root, and then kexec NixOS from there, and provide ssh access in the u-root stage
<clever>
samueldr: it would have to be a kernel that can work on that device, with a userland gui that can deal with the drivers, then kexec things based on user input
<clever>
pbb: yep
<samueldr>
though u-root was part of the tools I wanted to look into
levdub has joined #nixos
<samueldr>
clever: yes, we already have the kernel part well taken care of in Mobile NixOS :)
<pbb>
I think building a gui with u-root might be challenging since it only contains a very minimal userland in general
<samueldr>
yeah
<pbb>
but a console thingy with the volume buttons might work
<samueldr>
I haven't started actually researching the final solution
<samueldr>
pbb: you assume there is a console :)
<pbb>
x)
<clever>
samueldr: if you have framebuffer, then fbcon should work?
<samueldr>
all of qualcomm devices have framebuffer working, but no VT
<samueldr>
well, they have VT, but no console on them
<samueldr>
I do need to re-check what was up with all that
<pbb>
I also thought of putting the userland of the u-root stage on my boot partition
<samueldr>
but in my experience it looks like a case of qualcomm not bothering with something that's not used by android
<pbb>
that would make it possible to put a bigger environment with gui in there
<clever>
samueldr: try to echo to both /dev/tty and /dev/console and see what happens? and if you booted with console= then /dev/console goes there
<samueldr>
clever: echo using what?
<samueldr>
:D
<pbb>
ssh?
<clever>
samueldr: `echo foo > /dev/tty1` before you start X up
<samueldr>
oh, right, I was about to say "no wifi"... but I do have RNDIS
<pbb>
samueldr: if you are the person doing some of the mobile-nixos, may I message you about a problem I ran into when trying to add support for my phone, the Xiaomi Mi 9T / Redmi K20 ?
<clever>
pbb: this generates an initrd with a haskell binary at /init, and absolutely nothing else in the initrd!
<samueldr>
though, not sure if I'll end-up using u-root, for one "dumb" reason, I might end up preferring relying on rust rather than go...
<samueldr>
pbb: what "if you are"? :D
<clever>
pbb: though due to the complications within kexec, you would want to add a kexec binary to things at a minimum
<samueldr>
pbb: unless personal, in the open, on #nixos-aarch64 please :)
<fresheyeball>
I want to make a derivation for this
drakonis has joined #nixos
<jlv>
Anyone have any idea why the amdgpu module for my new AMD 5700 XT won't load on NixOS 19.09, but works on unstable? Using the same kernel on both.
<fresheyeball>
is there a way to package a python script as an exe with nix?
<ogle>
has anyone here successfully gotten a lustre filesystem mounted on nixos? I have been spinning my wheels the last week trying to get the thing working
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « grabserial: 1.9.8 -> 1.9.9 »: https://git.io/JePXg
<nh2>
how do I debug `cycle detected in the references of '/nix/store/drrpq7xcifkdimqgcb32mh4rwhhjzwxr-mesa-19.1.5-drivers' from '/nix/store/gmx041hj1b07c3hcgzz4jmi5bpx57jaf-mesa-19.1.5'`
<clever>
nh2: grep each dir for references to the other one
<clever>
zeta_0: there are no nix files in that repo, so you need to start by writing an expression that will build etcher, likely using yarn2nix
horner has quit [Quit: zzZzZzZzZZZzzz…]
<clever>
zeta_0: how did you mess up the SD card using dd?
<zeta_0>
clever: well if you could instead help me use dd correctly, i downloaded the raspbian os zip file and need to burn it onto my sd card, do i first need to run mkfs.vfat to format my drive /dev/sda before running dd, or does dd just overwrite everything ?
<clever>
zeta_0: what are the contents of the zip file?
<jlv>
I think I need help from someone with more linux experience than me. The amdgpu drivers won't load in NixOS 19.09. They do load in NixOS unstable, but even using the kernel from unstable doesn't fix it. I thought the drivers were in the kernel, but as far as I can tell, the kernel isn't the issue here.
<zeta_0>
clever: i couldn't get nixos to install on my rasberry pi so i am trying to install raspbian os instead: 2019-09-26-raspbian-buster-full.zip
<clever>
zeta_0: what does `unzip -l 2019-09-26-raspbian-buster-full.zip` report?
<zeta_0>
clever: i couldn't get that noobs to work either
<zeta_0>
clever: one second
<jlv>
The specific error is "amdgpu 0000:03:00.0: Direct firmware load for amdgpu/navi10_gpu_info.bin failed with error -2"
<clever>
jlv: the driver loaded fine, but it cant find the firmware for the card
<clever>
jlv: the firmware is usually in /run/current-system/firmware
<clever>
$ vi nixos/modules/system/activation/top-level.nix
<clever>
List of packages containing firmware files. Such files will be loaded automatically if the kernel asks for them (i.e., when it has detected specific hardware that requires firmware to function). If multiple packages contain firmware files with the same name,
<clever>
jlv: so you need to shove some unstable package (that contains a navi10_gpu_info.bin file) into hardware.firmware
<notgne2>
zeta_0: node2nix will generate Nix files from pretty much any node project (but requires the command to be ran to generate the files), pnpm2nix will expect a pnpm lock in the repo, yarn2nix is the same but for yarn
<clever>
jlv: i can see that the file exists on my machine, and is part of the linux firmware non-free package
<notgne2>
as a simple hack you can probably fork the repo, run `pnpm install`, then push up the changes and reference that repo in your config
<clever>
jlv: so hardware.firmware = [ unstable.firmwareLinuxNonfree ]; may do it
<zeta_0>
notgne2: oh, i think i will try dd one more time, i hope my sd card is not broken from the time before
<clever>
zeta_0: its pretty hard to break an SD card with dd
<notgne2>
if you always ctrl+c the `dd` command or wait for it to finish instead of ripping it out of your PC it should probably be safe
misterwhatever has quit [Quit: misterwhatever]
<notgne2>
it will leave your SD card looking very broken however, as it won't have a complete filesystem
<jlv>
clever: huh, I thought the amd drivers were free now.
steell has quit [Ping timeout: 246 seconds]
<zeta_0>
clever: dd froze my sd card and i had to unplug it to kill the process so i don't know what happened
<notgne2>
but you can restore it to normal by partitioning it, or simply run `dd` again as `dd` just writes directly and doesn't care what the FS is
<clever>
jlv: there are ~3 components, the driver itself (c code in the linux kernel), the firmware, and then the opengl libraries that talk to the driver
<clever>
zeta_0: was it actually frozen? or was it just busy writing the image and you didnt wait for it to finish?
<zeta_0>
notgne2: that's good too know
vonfry has quit [Ping timeout: 240 seconds]
<zeta_0>
clever: it was actually frozen
vonfry has joined #nixos
<zeta_0>
clever: dd normally just takes a few minutes
o1lo01ol1o has joined #nixos
<clever>
zeta_0: how long it takes depends on how big the .img file is
<notgne2>
a silly way to tell if something is stil writing, is running `sync` and seeing how long it takes to finish
<notgne2>
if it takes a while every time, I/O is probably just slow, not stuck
<clever>
notgne2: you can also `grep Dirty /proc/meminfo`
<zeta_0>
well i'll show you guys the new dd command before i try running it again
<clever>
that tells you how much is still pending to be written
<notgne2>
clever: that's pretty neat I think I'll alias that
<zeta_0>
clever: oh so i can run that grep command in another terminal while dd is executing ?`
<clever>
zeta_0: yep
<clever>
zeta_0: you can also use `pv` to get better feedback, one minute
<notgne2>
zeta_0: basically the kernel won't write everything to disk (or in this case, SD card) right away, as it's faster to process than it is to physically write it, and that command will give you an idea of how much is still in the "to-be-written" queue
<notgne2>
bit of a random question, would it make sense to create an issue or pull request to add Nix toggles for `[ "radeon.si_support=0" "amdgpu.si_support=1" "radeon.cik_support=0" "amdgpu.cik_support=1" ]`, I've needed those kernel params to make nearly every AMD GPU I own run Vulkan (or work at all in some cases)
<clever>
zeta_0: maybe, just make sure sda is the SD card
<clever>
zeta_0: /run/media/wolf only exists if you have a drive labeled wolf
<zeta_0>
clever: i place my sd card in the usb then place the usb in my laptop
<clever>
zeta_0: but dd erased everything, so its not labeled anymore
<notgne2>
I don't know if there's some special rules regarding raspi filesystems but generally at this point I'd honestly "cheat" and use clever's kexec boot config to install NixOS if it's easier to get another Linux OS running
<clever>
zeta_0: looks like a freshly partitioned disk, with zero partitions, you should be safe to try `unzip -p 2019-09-26-raspbian-buster-full.zip | sudo dd of=/dev/sda bs=4M status=progress conv=fsync`
<hodapp>
weird... I didn't add a certain port to networking.firewall.allowedTCPPorts yet the port is still open
<hodapp>
firewell is enabled...
<clever>
hodapp: ssh?
orivej has joined #nixos
<hodapp>
no, it's just a port number picked more or less at random
<zeta_0>
clever: i want to use raspbian os on my pi for the next couple weeks, then i'll check to see if the build errors for nixos for the pi 4 are fixed
<clever>
hodapp: how did you test that its open?
levdub_ has quit [Remote host closed the connection]
levdub has joined #nixos
<hodapp>
clever: using my browser and curl, because there's a Docker container listening on the port and serving up a page because I created a systemd service with 'docker run blahblah -p 12345:80 blahblah'
<clever>
hodapp: localhost isnt blocked by the firewall
<notgne2>
I think docker has a habit of awkwardly punching through some types of firewalls
<hodapp>
clever: but this isn't from localhost, it's from the external address
<clever>
notgne2: that sounds likely
<clever>
hodapp: is the port listed in the output of `iptables-save` ?
vonfry has quit [Ping timeout: 265 seconds]
<hpfr[m]>
is `boot.extraModulePackages = [ config.boot.kernelPackages.exfat-nofuse ];` necessary with kernel 5.4? what about `pkgs.exfat`?
<zeta_0>
clever: also, i think i need to use raspbian os in order to upgrade/upgrade the firmware
<hodapp>
clever: yes
<clever>
hodapp: and if you stop the docker container, does it disapear?
<clever>
zeta_0: is it an rpi4?
<hodapp>
clever: yes
<clever>
hodapp: then you can thank docker for opening the service to the world without asking
vonfry has joined #nixos
<notgne2>
hodapp: just googled this, the `-p` flag uses iptables to route the port, and does so in a way that just breaks past any firewall
<hodapp>
ugh
<clever>
notgne2: it might even be broken by restarting the nixos firewall service!
<hodapp>
the only reason it is 12345:80 and not localhost:12345:80 is because, for some unknown reason, I started getting 502 Bad Gateway after a NixOS upgrade; I was using nginx to set up an SSL proxy to localhost:12345
<zeta_0>
clever: yes, it's a rpi4
<notgne2>
apparently you can put `{ "iptables" : false }` in "/etc/docker/daemon.json" (I'm not sure where this goes on NixOS)
levdub has quit [Ping timeout: 240 seconds]
<zeta_0>
so the rpi4 is still not fully supported yet ?
<clever>
zeta_0: the only part of the firmware you can update on the rpi4 is fairly small, and i dont know of any world-breaking bugs with it yet
<clever>
zeta_0: and i'm pretty sure you can update the firmware from nixos as well
<hodapp>
clever: errr, wait. so if I had localhost:12345:80, and it made an iptables rule, might this rule also have been broken by nixos firewall being restarted?
<hodapp>
hence my 502?
<clever>
hodapp: that sounds likely
* hodapp
stabs Docker in the face
<clever>
hodapp: nixos isnt aware of docker messing with things, and just undid it
<notgne2>
docker will just paste things into the iptables rules, nix will probably erase and rewrite it all to ensure purity
levdub has joined #nixos
<hodapp>
blargh. is there any "correct" way to handle networking like this?
<clever>
hodapp: either tell docker to stop touching iptables (see above) or find a way to tell docker "i broke it, re-do whatever you did to the firewall"
<notgne2>
in my case I'm just trying to deprecate every last docker container at my company as a solution
<zeta_0>
clever: i was told you couldn't do it with nix, anyways i wasn't able to install nixos on my pi because of that jackd error
<hodapp>
clever: but if docker doesn't touch iptables, how can I actually access the server the container is running?
<zeta_0>
clever: i couldn't get it to build
<notgne2>
hodapp: I think it has other ways to work as well, their userland-proxy stuff which always breaks is probably related to this in some way too
<notgne2>
also without ports you can still use the assigned IP address of the container to access it
<hodapp>
notgne2: can I get that IP address in a sane way?
<clever>
zeta_0: the script to update the rpi4 firmware is just a regular bash script that runs a static elf binary
<notgne2>
it should be on the `docker inspect` output
<clever>
zeta_0: it should work on nixos just fine
<hodapp>
notgne2: but this has to somehow make it into my configuration.nix for my nginx config
<hodapp>
preferably without being hard-coded
levdub has quit [Ping timeout: 250 seconds]
<clever>
hodapp: docker has other network modes, like bridging, which will just give the container its own ip on the real lan
<clever>
hodapp: then it will just give a shout out to the nearby dhcp server
<zeta_0>
clever: oh that's good, well i am having problems installing nixos-unstable on rpi4, i don't know if it is because of the image or i don't know what i am doing, i couldn't get it to build
<hodapp>
clever: this is on a dedicated server where I'm not sure that's permissible nor desireable (I really don't want to just expose port 80 to the world)
<notgne2>
there's probably some way to force the container to use a specific IP so you don't have to worry about updating it everywhere all the time, but I'm not sure
<hodapp>
I feel like I looked this up once before and the answer was a combination of unenlightening, stupid, and useless
<hodapp>
like many things in Docker
<zeta_0>
have any of you guys got nixos-unstable to work on rpi4 ?
<clever>
zeta_0: i still need to order a 4
<hodapp>
my Pi 4 is still in its box, but it is progres
<notgne2>
talking of hardware support I wonder if I can get NixOS to have working sound on my Skylake laptop one day
<notgne2>
I heard the guys over in the Debian world figured out some way to make it work
ddellacosta has quit [Ping timeout: 240 seconds]
<notgne2>
though I'm not sure where to start with adding it to my Nix config, as so much of it is writing random etc files and I'm not sure how many of them I can just write to and how many I will have to use some other Nix config option for
mexisme has joined #nixos
<clever>
notgne2: maybe start by undoing changes on debian, to see what breaks it and what is actually needed, then document what is needed first
<hodapp>
hm, I guess I can 'docker network create' and then point 'docker run' to it
boxscape has quit [Remote host closed the connection]
<notgne2>
clever: some of that might be figured out already, last time I checked in on this issue was about a year ago and my laptop almost caught fire, so I decided to just disable the sound card until it was a bit more mature
<zeta_0>
do rpi's work well with tvs ?
<clever>
notgne2: almost caught fire?
<hodapp>
zeta_0: mine always have
wildtrees has quit [Quit: Leaving]
<notgne2>
yes, skylake seems to be a bit balsy with their sound support, it will happily drive the speakers at 1000x volume and get insanely hot
<zeta_0>
hodapp: cool, this is my 1st pi, so i don't know what to expect
<hodapp>
notgne2: yikes, I've not seen sound cards like that for quite awhile
<zeta_0>
i'm excited
<notgne2>
most people just have crackling sounds or it not working, but I've found a few more posts like mine where the laptop gets hot enough to melt the stickers off around the touchpad despite the CPU being insanely cool every time there's music trying to play
<hodapp>
that sounds awful
<notgne2>
I guess in 2019 hackers can turn your computer into a bomb
jlv has joined #nixos
<zeta_0>
i have had some of those laptop speakers, i am surprised my ear drum didn't bust
<zeta_0>
i remember back when i had windows, i virus that my laptop was infected with kept making a random beeping sound
<hodapp>
o_O
<zeta_0>
i am never going back to a proprietary os, i love linux
<notgne2>
zeta_0: that could've been going through your pcspkr which can be exceptionally loud
<hodapp>
some of the most insanely obnoxious things I have ever had to listen to came from the PC speaker
<notgne2>
lots of computers have these small speakers entirely dedicated to either making ear piercing beeps whenever something breaks, or wasting your time when you try to use weird old kernel modules to play music through it for fun
<notgne2>
yet I have still only met 1 or 2 people that have memorized 2 or more beep codes
<clever>
notgne2: one of my laptops lacked that speaker, and routed it thru the sound card
<clever>
notgne2: it produced ear-splitting noises if you happened to be wearing headphones
<hodapp>
>S<
<clever>
and it ignores all volume controls
spacefrogg has quit [Quit: Gone.]
aw has quit [Quit: Quitting.]
aw has joined #nixos
<o1lo01ol1o>
Is there a recommended way of getting a recent Xcode to derivation's build inputs?
<notgne2>
I guess you have to be extra careful not to backspace too many times when stopping terminal bel from triggering it
spacefrogg has joined #nixos
<hodapp>
clever & notgne2: thanks for help. my box is now no longer opening up ports to the world, nor getting nonsensical 502 Bad Gateway errors.
* hodapp
stabs Docker in the face one more time
<zeta_0>
so far nix seems pretty similar to haskell, i think i'll be able to get the hang of it eventually
<clever>
:D
<hodapp>
zeta_0: there is certainly some overlap
<jlv>
clever: your suggestion worked btw :) Thanks a ton!
<zeta_0>
dd takes to long, i have no patience
<clever>
zeta_0: other programs will take just as long to write the image
<notgne2>
violence towards computers always seems so much more enticing after your first "docker: Error response from daemon: OCI runtime create failed: container_linux.g"
<hodapp>
notgne2: the silly thing with Docker is how it's a very useful diaper for software that was never properly toilet-trained... and then a bunch of people ran around singing the praises of designing software that way because then you have the benefit of diapers
<zeta_0>
sorry for being off topic
<hodapp>
zeta_0: I know a bit but am out of practice
<clever>
hodapp: who needs to potty train your software when you can just slap a daiper on it and call it done!!
<zeta_0>
monads still confuse me
<clever>
hodapp: just ignore the smell of poo coming from it!
<notgne2>
hodapp: 90% of times I've seen people use docker is because they want a container for some security which it doesn't actually provide, or because they haven't discovered Nix and therefore can't easily make their code run properly on different operating systems
<zeta_0>
i have never tried docker, i heard nix is much better
<clever>
notgne2: i once saw somebody showing off a docker image that can act as a steam content server cache, so you can setup a lan party and download games at gigabit speeds
<clever>
notgne2: i looked into it, and it was literally just bind+nginx with 2 config files
<zeta_0>
i looked up best linux package managers on that slant website, that's how i found out about nix(os)
<notgne2>
it's generally better than docker build recipes, plus you can reference the module from your NixOS config if you ever don't want to use Docker
<zeta_0>
i don't miss all the problems i had with debian/ubuntu
<zeta_0>
they crashed too many times to count
<notgne2>
most problems you could almost call Docker a "solution" to, NixOS doesn't have, so I just wanted to make the glue back to Docker so migrating stuff is easier
<hodapp>
stupid question I can't find the answer to: how do I turn an integer to a string?
<notgne2>
`toString` iirc
<notgne2>
> toString 1
<{^_^}>
"1"
<zeta_0>
clever: dd looks frozen, but i'll keep waiting
<clever>
zeta_0: `grep Dirty /proc/meminfo` again, what does it say?
<clever>
zeta_0: you will want to `watch -d 'grep Dirty /proc/meminfo'` and wait until it gets near zero
eraserhd has quit [Quit: WeeChat 2.6]
vonfry has quit [Quit: WeeChat 2.6]
<zeta_0>
clever: that command isn't doing anything
<clever>
zeta_0: did you get the quotes right?
<zeta_0>
clever: probably not
felixfoertsch has joined #nixos
<zeta_0>
how many years have you guys been using nix(os) (i have been using it for a couple months)?
<zeta_0>
the learning curve was rough for me
<zeta_0>
it still is rough, lol
felixfoertsch23 has quit [Ping timeout: 252 seconds]
mexisme has quit [Ping timeout: 240 seconds]
<clever>
zeta_0: ive been using nixos for ~3-4 years
eraserhd has joined #nixos
<notgne2>
just checked chat logs, apparently since june
<zeta_0>
notgne2: so your pretty new to nix too ?
<clever>
--- Log opened Thu Jun 25 11:13:00 2015
<clever>
$ head ~/irclogs/freenode/#nixos.log
<notgne2>
I had some stupid dream and got really really angry at files, so decided I would make my computer perfect and pure by using TMSU for my files and NixOS for my system
paprika has joined #nixos
<zeta_0>
do you guys think purely functional programming will go mainstream anytime soon, languages like haskell and nix still seem very niche
<zeta_0>
most people love python
<notgne2>
I'm not sure, I actually don't know haskell very well, the only time I do anything function is either with Nix or Elm
jlv has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @marsam opened pull request #74246 → protobuf3_11: init at 3.11.0 → https://git.io/JePMn
<notgne2>
for me at least it's seemed a little more difficult to learn then a lot easier afterwards, but that might be because I'm so used to imperative approaches to programming
paprika has quit [Remote host closed the connection]
<zeta_0>
notgne2: same here, i did java for 3 years, then tried haskell and it was so different
<notgne2>
I'd really like to learn Haskell soon though, as if Nix and Elm are anything to go by I think I'd be more comfortable with it than I am with anything else after I have a better understanding
<zeta_0>
notgne2: i'd recommend first learning some category theory and lambda calculus before learning haskell
<zeta_0>
notgne2: things like monads will make much more sense
<zeta_0>
clever: ok, dd finally finished, i hope it works
<zeta_0>
clever: since you have been using nix for years, what are some of the difficult concepts of the language to understand ?
<clever>
zeta_0: i think the most complicated to try and figure out is the --show-trace output
<clever>
its not the reference call-chain, but the force callchain
<zeta_0>
clever: so what's the difference between the 2 of them ?
<clever>
so if you do `let a = 5*5;` and then pass `a` to a function like `a_in: { b = a_in; }`, then try to read b, the --show-trace may point towards b, rather then a
<clever>
and you have no idea where b came from
<clever>
the problem is that nix is tossing lazy values all over the place
<notgne2>
zeta_0: I'm not sure what would be a good resource for learning that besides the language itself is the problem, I generally suck at learning this kind of thing from books
<clever>
and only when its recursively making things non-lazy, does it have a backtrace
<zeta_0>
clever: oh i see how that could be confusing
<{^_^}>
[nixpkgs] @matthewbauer pushed 3 commits to master: https://git.io/JePMz
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
sb0 has joined #nixos
paprika has joined #nixos
paprika has quit [Client Quit]
<{^_^}>
[nixpkgs] @bhipple opened pull request #74247 → par: use pname → https://git.io/JePMa
paprika has joined #nixos
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
paprika has quit [Client Quit]
jlv has joined #nixos
<jlv>
clever: I may have celebrated a little prematurely. The amdgpu module is loading now, but the system doesn't appear to be using my graphics card. Performance is very poor, and my OpenGL renderer is "llvmpipe".
<clever>
jlv: sounds like hardware.opengl isnt linking to it properly
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jlv>
clever: hmm. I could try overlaying the unstable hardware.opengl. I'm not really sure what else to do. I don't have much experience with hardware on Linux.
jlv has quit [Remote host closed the connection]
jlv has joined #nixos
<jlv>
clever: doesn't look like overlaying the unstable firmware.opengl did anything. I know the gpu works fine on the unstable branch. I just can't figure out what changed to make it work.
o1lo01ol1o has quit [Remote host closed the connection]
endformationage has quit [Ping timeout: 265 seconds]
domogled has joined #nixos
zeta_0 has joined #nixos
<zeta_0>
clever: finally got raspbian os working on my pi, thanks for the help
<clever>
yep
<zeta_0>
clever++
<{^_^}>
clever's karma got increased to 255
<clever>
0xff!!
gagbo has quit [Ping timeout: 276 seconds]
<hpfr[m]>
where can I find the channel history for nixos? rollback says switching to generation 17, but where do I find the associated nixpkgs revision
<wedens[m]>
hpfr: file name in /nix/var/nix/profiles/ contains nixpkgs revision (if you use channels and not git revision)
<hpfr[m]>
wedens: but how do I make sure my nix-channel is aligned with one of the system-*-link revisions in that directory? the number given by --rollback doesn't match up
<hpfr[m]>
wedens: I mean nix-channel doesn't give me a revision to work with, and what nix-channel you're on has nothing to do with what system-*-link revisions you might have
<hpfr[m]>
ok I'll try that
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
domogled has quit [Remote host closed the connection]
<hpfr[m]>
what nix-channel revision* you're on
o1lo01ol1o has joined #nixos
rople has quit [Ping timeout: 268 seconds]
Rusty1 has quit [Quit: Konversation terminated!]
o1lo01ol1o has quit [Ping timeout: 268 seconds]
rople has joined #nixos
jmeredith has joined #nixos
kolbycrouch has quit [Ping timeout: 245 seconds]
horner has joined #nixos
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
felixfoertsch has quit [Quit: ZNC 1.7.3 - https://znc.in]
felixfoertsch has joined #nixos
marcus7070 has joined #nixos
[Leary] has quit [Remote host closed the connection]
Lears has joined #nixos
zeta_0 has joined #nixos
<marcus7070>
When making an issue (not a PR) on GitHub/nixos/nixpkgs , can I do "@GrahamcOfBorg build slic3r" or something similar to get the bot to build from master?
<zeta_0>
after connected my pi to my tv, i went to the tvs build in settings and lowerd the brightness and backlight but it is still to bright, is there another way to lower my tvs brightness ?
<zeta_0>
sorry for this question being off topic
<jared-w>
notgne2: your continix repo looks super nice
gxt has quit [Remote host closed the connection]
gxt has joined #nixos
levdub has joined #nixos
palo1 has joined #nixos
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
bahamas has joined #nixos
levdub has quit [Ping timeout: 276 seconds]
palo has quit [Ping timeout: 276 seconds]
palo1 is now known as palo
bahamas has quit [Read error: Connection reset by peer]
horner has quit [Quit: zzZzZzZzZZZzzz…]
horner has joined #nixos
levdub has joined #nixos
marcus7070 has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @marcus7070 opened pull request #74251 → [WIP] slic3r: patch to fix compile error → https://git.io/JePyL
bahamas has joined #nixos
hmpffff has joined #nixos
jlv has joined #nixos
<bahamas>
I don't know if my question went through, so I'll repeat it
<bahamas>
how can I apply an overlay to a set of packages within nix packages? in this case, I have a file generated by pip2nix and I want to apply it to python37Packages
mexisme has joined #nixos
<jlv>
Can I set my nixos channel to a specific nixpkgs-channels revision?
<jlv>
dminuoso: I'm looking to set the root nixos nix-channel to a specific revision. I'm trying to diagnose a hardware issue by doing a bisect on nixpkgs.
tbenst has quit [Remote host closed the connection]
<dminuoso>
jlv: But you can do this by hand as well.
mexisme has quit [Ping timeout: 246 seconds]
noudle has joined #nixos
<dminuoso>
jlv: A nix-channel is just a set of files/directories at the end.
<dminuoso>
jlv: You could have NIX_PATH point into a git repository of yours, and then `git bisect`
<dminuoso>
At least as far as things like nix-env/nixos-* are concerned
<dminuoso>
jlv: That is, a channel `nixos` is just some path `/foo/bar` containing the channel data, and your NIX_PATH containing an entry `nixos=/foo/bar`
<dminuoso>
jlv: gah what was my last message you read?
<dminuoso>
Internet is quite buggy here on the train
acarrico has quit [Remote host closed the connection]
<bahamas>
is there any downside to doing myPythonPackages = python37Packages // projectRequirements; ?
<dminuoso>
jlv: (If we are being very pedantic, then channels are realized by NIX_PATH containing a path `/nix/var/nix/profiles/per-user/root/channels` which in turn contains sub-directories representing these channels.
<jlv>
dmiuoso: I think I got all your messages. I'm just a little distracted :)
<marcus7070>
Anyone able to test a nixpkgs attribute from master for me? I think `slic3r` fails to compile (due to a new version of Boost libraries), but I'd like someone else to confirm.
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « gource: 0.49 -> 0.51 »: https://git.io/JePSM
horner has quit [Quit: zzZzZzZzZZZzzz…]
levdub has quit [Remote host closed the connection]
levdub has joined #nixos
<jared-w>
wrote me a nice little aws lambda function to hit the irccloud API every hour to keep me logged in for free forever. \m/
FRidh has joined #nixos
<jared-w>
was gonna write in haskell but I couldn't figure out how to get nix to build a haskell binary nicely for whatever reason so I gave up and used rust + serverless api. Worked out pretty nicely although, annoyingly, I needed to duplicate the entire build all over again in docker because their CLI forces its use. *sigh*
<jared-w>
either way, the rust setup was very pleasant. It's really nice to have a shell nix + lorri setup so that if I ever run into something not being there in the environment I just... declare it into existence. Yehh
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
philr has quit [Ping timeout: 245 seconds]
cfricke has quit [Quit: WeeChat 2.6]
<srhb>
jared-w: Poor irccloud. :)
<jared-w>
heh, there's very easily findable documentation and step by step explanations on how to do it, I just went the nerdy #hype route with rust + lambda functions vs python and heroku
<jared-w>
I mainly have the setup to communicate with hvr on #haskell about the stuff I'm working on him with regarding the github actions for installing ghc+cabal
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « gitfs: 0.5.1 -> 0.5.2 »: https://git.io/JeP9k
levdub has quit [Remote host closed the connection]
civodul has joined #nixos
<`_>
It seems like half the things I used to have installed within haskellPackages are broken anymore
<`_>
Admittedly I don't have _much_ installed, jenkinsPlugins2nix, stack, a few others
<`_>
But even those are marked broken
<jared-w>
yeah I noticed that some haskellPackages stuff was breaking recently. stack2nix was one I noticed and gave up installing. Isn't nixpkgs going through a haskellPackages refactoring to use iohk's haskell.nix stuff?
<`_>
Ah also dhall2nix
<`_>
I'm honestly not sure what's going on with it, I'm just trying to upgrade my jenkins plugins lol
<{^_^}>
[nixpkgs] @andir opened pull request #74253 → stdenv: make symlinks that refer to the same output relative → https://git.io/JeP9m
<`_>
But jenkinsPlugins2nix hasn't been building for a while now, so I'm trying to manually build it and am even running into problems there
<{^_^}>
error: syntax error, unexpected ':', expecting ')', at (string):271:21
<ptrcmd>
this is what I got from the build log
Chiliparrot has joined #nixos
magnetophon has quit [Remote host closed the connection]
joshuagl has joined #nixos
zupo has joined #nixos
zupo has quit [Client Quit]
<srk>
ptrcmd: try without CONFIG_ prefix
<ptrcmd>
srk: that doesn't work
<ptrcmd>
now I get > error: unused option: SND_SOC_SOF_INTEL_TOPLEVEL
alexherbo2 has joined #nixos
<nh2>
`_ jared-w: IOHK seems to have abandoned stack2nix to a large extent, but I'm still using it. Where is it broken? I may want to look into fixing it
<ptrcmd>
ddima: I thought that just means that it's not enabled by default?
<ptrcmd>
ddima: that's why I am trying to enable it manually :)
<`_>
nh2 `for i in dhall2nix jenkinsPlugins2nix stack2nix stack; do nix-env -i $i; done`
<`_>
All of those should error out. and that's just the bare minimum I can come up with for now, as those are the ones I most commonly interact with.
<ddima>
ptrcmd: Well, don't know, but since some of the supposedly working configs floating around the web dont even succeed in the configure phase (I tried a few versions, with the same result of unused option) I'd question the readyness. You could ofc try 5.4 from unstable or so.
lux4 has quit [Ping timeout: 276 seconds]
lux4 has joined #nixos
<nh2>
`_: on 19.09 or master, and failing at evaluation time or build time?
<`_>
nh2: I even tried with jenkinsPlugins2nix `while ! nix-build; do git reset --hard @~; done` and it never properly built
<`_>
19.09
<`_>
Build time
<ptrcmd>
ddima: I *am* trying 5.4
<`_>
And jenkinsPlugins2nix used to build perfectly fine.
<ddima>
ptrcmd: oh, sorry, I dropped a frame there. Hmm.
<ddima>
(The option in question was introduced in 5.3.something, so I gave that a go)
sb0 has quit [Quit: Leaving]
<`_>
Sorry, didn't tag you: Respectively, nh2, 19.09 and build time.
<`_>
If I recall correctly hnix is failing to build properly.
<`_>
I finally got jenkinsPlugins2nix going with stack so may we rejoice
<`_>
`nix-env -iA nixos.haskellPackages.stack` (if I recall correctly) failed unless I entered a nix-shell
<nh2>
`_: ah, I see what the error is now. This is stack2nix not being upgraded to the newer version of stack. The stack changes were announced a long time ahead, but iohk cancelled the efforts of making stack2nix compatible with them
thc202 has joined #nixos
<`_>
I honestly can't say that I know much about that.
rople has quit [Ping timeout: 268 seconds]
oscarvarto has joined #nixos
<nh2>
`_: stack should install on both 19.09 and master though
<`_>
I'm on 19.09 and stack is marked as broken unless I'm inside of nix-shell
<`_>
nh2 ^
rople has joined #nixos
<nh2>
`_: that's odd, why would the being-inside-nix-shell change the brokenness label? Also I'm pretty sure I installed stack the normal way on my 19.09 laptop a few days ago, let me confirm
<oscarvarto>
Hi! Anybody nows why stack build fails with an error related to lindo
<oscarvarto>
Maybe the error is related to the one mentioned by nh2
<oscarvarto>
Libffi*
<nh2>
`_ oscarvarto: so right now I have /run/current-system/sw/bin/stack -> /nix/store/6lzdpxshx78281vy056lbk553ijsdr44-stack-2.1.3.1/bin/stack on that machine, and `nixos-rebuild switch` suggests I'm up to date
oscarvarto has quit [Remote host closed the connection]
ThatDocsLady has joined #nixos
<nh2>
so I can't see any error currently on 19.09
<ddima>
ptrcmd: oh. just waking up - probably no CONFIG_ prefix needed.
<`_>
nh2: let me check
<ptrcmd>
ddima: I also tried that
<ptrcmd>
ddima: and that didn't work
captn3m0- has joined #nixos
m0rphism1 has joined #nixos
captn3m0 has quit [Ping timeout: 265 seconds]
oscarvarto has joined #nixos
<ddima>
hm, sad.
<`_>
nh2: Stack seemed to properly build outside of a nix-shell this time. Perhaps what I was experiencing earlier was a red herring, so I wouldn't pay it much mind. I'll update if I find anything more.
<nh2>
`_: great, let me know if you find any problem
<`_>
nh2: Will do. I appreciate it.
lux4 has quit [Ping timeout: 268 seconds]
cfricke has joined #nixos
lux4 has joined #nixos
<ddima>
ptrcmd: `SND_SOC_SOF_TOPLEVEL` seems to be the one
<ddima>
thats one simple to configure subsystem ;)
<ptrcmd>
ddima: it worked :)
<ddima>
well, good luck with the rest ;)
<ptrcmd>
:)
<ptrcmd>
thanks!
rydnr has joined #nixos
oscarvarto has quit [Ping timeout: 265 seconds]
<rydnr>
Hi. I'm trying to do a reinstallation of my NixOS using latest 20.03 livecd, and nixos-install. But my configuration is trying to build perl5.30.0-po4a-0.55 and gets stuck.
<rydnr>
I don't know which package is depending on po4a. It gets stuck with: [1/1/196 built] building perl5.30.0-po4a-0.55 (buildPhase): Created META.yml and META.json
<abbec>
hello, is it possible to control sandboxing for nix-build with an env-var?
<LnL>
rydnr: I have a feeling your install cd might be running out of ram/disk space
<flokli>
nh2: wow, very nice :-)
mexisme has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @adisbladis pushed to master « openrazer: Fix build with Linux 5.4 »: https://git.io/JeP7s
werner291 has quit [Quit: werner291]
bahamas has quit [Ping timeout: 276 seconds]
<rydnr>
It was the autofs package.
mananamenos has joined #nixos
oscarvarto has joined #nixos
dema has joined #nixos
oscarvarto has quit [Remote host closed the connection]
zupo has joined #nixos
<Twey>
Discovered something weird today… as a NixOS user, my `nixbld` user in the sandbox is only a member of the `nixbld` group. But for my coworkers using multi-user Nix on other distros, build processes also run with `nogroup` in their supplementary groups, even though the `/etc/group` inside the sandbox doesn't have `nixbld` as a member of `nogroup`
<Twey>
This breaks `go_bootstrap` builds, because it has a test where it tries to chown a file to every supplementary group of the process, and the `nogroup` chown fails
romildo has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
<ddima>
raboof: uff. I was just briefly contemplating if I know a single person who still has a working CD-burning setup and that probably quite tricky nowadays :/
<ddima>
(ofc not referring to testing ISO with a VM)
<{^_^}>
[nixpkgs] @Evils-Devils opened pull request #74259 → Kicad cleanup, fix and update → https://git.io/JeP55
<raboof>
I can probably find one but it might take a while :)
<ddima>
ptrcmd: Ie dont have such hardware though, so don't really know much about it - but probably somebody else here has such a new device (I suspect 7th gen X1 carbon?).
<jonge[m]>
hey there. i have a pretty much standard nixos with GDM setup. when i boot the machine, i often end up in tty1 and need to press "alt+left" manually in order to get to the GDM login. do my tty1 and gdm maybe race or something? how to fix that? i am a bit surprised that this happens with a nearly empty nixos config, too.
<jonge[m]>
Shoubit: ok... i had this problem with some minimal config before, but it seems like i got rid of it with a new minimal config. must be my own fault somewhere. i will ping you back if i found the setting that bricks my setup.
mexisme has joined #nixos
rauno has quit [Ping timeout: 245 seconds]
<fling>
ddima: 5.3 version applies on 5.4 just fine
<Ariakenom>
I followed https://nixos.wiki/wiki/I3 to set up i3 which worked quite smooth. But when I try to enable autologin it doesnt like desktopManager.default = "none"
<ddima>
fling: if you can test it and it works fine, maybe worth adjusting the kernel expressions in nixpkgs and making a PR out of it.
<dredozubov>
Can anyone clarify how `nix build ... --builders ...` work? Does it copy all the nix store data used in the build locally or just builds remotely?
mexisme has quit [Ping timeout: 252 seconds]
bahamas has quit [Ping timeout: 240 seconds]
<clever>
dredozubov: it will download the inputs to the local machine, then copy those inputs to a remote machine, build the product there, then copy the result back
<dredozubov>
so it'll copy the whole /nix/store tree relevant for this build back, hmm
<dredozubov>
That's not what I want. I want to start remote jobs on CI and not copy anything unnecessary.
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<dredozubov>
Building a docker container remotely and copying just that would be better, I guess.
<clever>
dredozubov: if you `--option builders-use-substitutes true`, the remote machine will be allowed to use cache.nixos.org
<clever>
dredozubov: so there will be less things you have to copy from the local machine
mexisme has joined #nixos
<dredozubov>
that helps, but I was wondering about the copying between CI box and remote builder
ddellacosta has joined #nixos
<pbb>
How do I check if a list contains a specific string in the Nix language?
<dredozubov>
I mean, I want to build the test suite and copy only it back to the CI box
<clever>
pbb: builtins.elem
<pbb>
thanks :)
<pbb>
clever++
<{^_^}>
clever's karma got increased to 256
<clever>
dredozubov: after building any given derivation, only the result of that derivation is copied back
<clever>
dredozubov: and the machine triggering the build already has all of the inputs to the drv
<jared-w>
nh2: nice work with gtk and static nix! Sorry I didn't reply earlier, I was asleep. Looks like you were able to get to the root of the issue, though.
<jared-w>
nh2: on a static Haskell related issue, do you think you might be able to help me troubleshoot my efforts to setup an aws-lambda-haskell-runtime project to be built statically? It works with their stack template but I've been unsuccessful in converting it to nix and I'm not sure if I'm misunderstanding something or if it's an edgecase.
<jollyjester>
it's been sitting there for quite a long time
<{^_^}>
[nixpkgs] @dasJ opened pull request #74267 → nixos/nextcloud: Do not run sudo if not needed → https://git.io/JePNd
<jollyjester>
if i knew how to build packages i'd do it myself but i don't
__monty__ has joined #nixos
m0rphism1 has joined #nixos
mexisme has joined #nixos
jollyjester has quit [Quit: leaving]
<ddima>
jollyjester: people are a bit wary of adding too many random new packages, esp when they are very niche. NUR (nix user repo) or just own repo seems like a better place to start. If you'd be willing to try to package, I'm sure people here or on discourse would help you along the way.
<dredozubov>
clever: concerning `--option builders-use-substitutes true`: will it use a list of substituters configured on the machine starting the build or the one actually building it?
<clever>
dredozubov: the caches configured on the remote machine
<dredozubov>
cool, thanks
ng0 has quit [Ping timeout: 260 seconds]
ng0_ is now known as ng0
<aswanson>
anyone have experience running a flask development server within a nix-shell environment? I'm able to get it running in production mode without issue, but when I export `FLASK_ENV=development` I get a syntax error from the script in the nix store that actually executes flask
mexisme has quit [Ping timeout: 246 seconds]
<aswanson>
specifically, an invalid syntax error when the flask script attempts to modify PATH before actually executing flask
_Lemon_ has joined #nixos
endformationage has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
mexisme has joined #nixos
<vaibhavsagar>
aswanson: how are you exporting FLASK_ENV?
<aswanson>
just manually within my shell right now
<dredozubov>
I was under the impression that `nix build ... -builders builder --max-jobs 0` will just force all the work to `builder`, but I'm getting "error: unable to start any build; either increase '--max-jobs' or enable remote builds"
<vaibhavsagar>
aswanson: super weird, it looks like it's trying to run bash as Python and failing for that reason
<clever>
dredozubov: are you running it as root or a trusted user?
<dredozubov>
I have no idea what trusted user is. I'm not running nixos, to be clear.
waleee-cl has joined #nixos
<clever>
dredozubov: single or multiuser mode?
<aswanson>
vaibhavsagar: if it just failed in both cases I would be less confused but such a strange error after setting an environment variable just has me confused
<clever>
dredozubov: then your on a singleuser mode and it should just work
<aswanson>
vaibhavsagar: I will confess, I am running this within Windows Subsystem for Linux on a corporate network, but still, things have been functioning until I started trying this
<clever>
dredozubov: but ive notices some weirdness with people trying to force all remote jobs, try `--max-jobs 1` ?
<vaibhavsagar>
aswanson: I don't suspect that's related
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen1 has joined #nixos
<dredozubov>
clever: trying it, but I would really like no CPU action though
<clever>
dredozubov: why not?
werner291 has joined #nixos
<dredozubov>
I expect to run it from CI and I expect remote workers to be much more powerful in this scenario
<clever>
dredozubov: sounds like you want something like hydra, which is meant to do CI jobs
m0rphism1 has joined #nixos
<dredozubov>
I don't find it suitable for application CI
dingenskirchen1 is now known as dingenskirchen
<dredozubov>
I'm actually trying to replace it with a different setup
<Ariakenom>
looks like "windowManager.default = "i3";" fixed my issue, if anyone's interested
mexisme has quit [Ping timeout: 250 seconds]
<clever>
dredozubov: you can also use `nix-copy-closure` to copy a `.drv` file to a remote machine, and then run `nix-store -r` on the drv to build it
<clever>
dredozubov: and then nix-copy-closure the get the products back
<dredozubov>
Isn't nix-copy-closure bring all the dependencies to the target nix store as well?
<clever>
dredozubov: if ran on an output, yes
<clever>
dredozubov: but if ran on a .drv file, no
<dredozubov>
my current plan is to build a docker image on the remote build machine and copy it back without calling nix-copy-closure, I don't need to build a local nix store on these throwaway machines
<dredozubov>
--max-jobs 1 started a local build and remote builder is idling
<dredozubov>
this is exactly the opposite of what I expect :(
<vaibhavsagar>
sounds like it isn't properly set up
<clever>
dredozubov: can you pastebin the entire output, up to when it started to build
<clever>
warning: substituter 'ssh://b2builder' does not have a valid signature for path '/nix/store/m80snlingdn5qgnklzaixscsvw2gsamq-ghc-8.6.5'
<clever>
warning: substituter 'https://cache.nixos.org' does not have a valid signature for path '/nix/store/m80snlingdn5qgnklzaixscsvw2gsamq-ghc-8.6.5'
<dredozubov>
vaibhavsagar: it was working fine until some point though - no compilation locally, remote builder was busy
<zeta_0>
vaibhavsagar: apologies, it was just a quick question, when i get a little bit more comfortable with the nix language i can start contributing to the nixos
mexisme has quit [Ping timeout: 276 seconds]
<ptrcmd>
vaibhavsagar: Is there a way to do that without root?
<{^_^}>
[nixpkgs] @bcdarwin opened pull request #74274 → ocamlPackages.owl: init at 0.7.0 → https://git.io/JePpE
<elvishjerricco>
ptrcmd: I always forget the functions of the various bash dot files in ~/, but at least one of them should be loaded by the SSH session, which you can use to add stuff to PATH
<elvishjerricco>
ptrcmd: But what command are you wanting to point it to? You'll have to point it to a wrapper that invokes the chroot, right?
hmpffff has joined #nixos
<ptrcmd>
elvishjerricco: yes
<ptrcmd>
elvishjerricco: exactly, a wrapper that invokes the chroot
<ptrcmd>
elvishjerricco: and pass the necessary arguments to nix*
hmpffff has quit [Client Quit]
<ptrcmd>
elvishjerricco: the .bashrc file is not loaded in the non-interactive login shell
<elvishjerricco>
ptrcmd: What about .profile?
roconnor has joined #nixos
<roconnor>
ryantm: Regarding semi-automatic updates like https://github.com/NixOS/nixpkgs/pull/74128, how should I go about ammending the changes, or should I just close it and open my own?
lnadav has quit [Read error: Connection reset by peer]
o1lo01ol1o has quit [Remote host closed the connection]
lukash_ is now known as lukash_|away
<gchristensen>
I'm using gpg as my SSH agent, and I have some keys registered in there that aren't actually on disk. `ssh-add -l` lists some keys, but I can't delete any of them with `ssh-add -d` because the file doesn't exist. is there a way to remove a specific key without removing all of them?
<{^_^}>
[nixpkgs] @WilliButz pushed 2 commits to release-19.09: https://git.io/JePjL
<ddima>
gchristensen: you can use gpg-connect-agent and then run DELETEKEY iirc
<gchristensen>
oh cool
<gchristensen>
I should really delete all my keys and start over :)
Henson has joined #nixos
<gchristensen>
thank you, ddima
<Henson>
what virtualization mechanism do NixOS containers use, if any?
<gchristensen>
systemd-nspawn
halfbit has joined #nixos
drakonis has joined #nixos
drakonis1 has quit [Ping timeout: 265 seconds]
<Henson>
gchristensen: are the networks and process trees pretty much separate? I want to run two resource-hungry programs on one powerful computer as if I were running each of them on two less powerful computers, and am wondering if NixOS containers would be suitable for this.
<gchristensen>
I'm looking in to what the KEYINFO --list columns mean and wow
<ddima>
or something there-like. hmpf. maybe it also needs a RESET command or sth. it's quite obscure, but I remember that I had a similar thing a while back.
<gchristensen>
all of a sudden my computer randomly decided to put the keys in a different order or something, and I can't SSH to a bunch of boxes now :)
ckauhaus has quit [Quit: WeeChat 2.6]
knupfer has quit [Quit: knupfer]
knupfer has joined #nixos
knupfer has quit [Client Quit]
knupfer has joined #nixos
<ddima>
gchristensen: because too many keys to try? I usually use ssh_config and just stupidly declare which is for which and have had an OK life for a while with that.
<gchristensen>
yeah
<gchristensen>
I have done that, but also I should also delete and regenerate a bunch of these keys
ThatDocsLady has quit [Ping timeout: 246 seconds]
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen1 has joined #nixos
<gchristensen>
if I just copy my ~/.gnupg elsewhere, that'll be a sufficient backup of my current GPG agent, right?
<gchristensen>
I think my best bet is to just delete all these keys and start fresh
dingenskirchen1 is now known as dingenskirchen
tilpner_ is now known as tilpner
<_Lemon_>
Should $out always be an absolute path? For example I'm trying to build the serving path for a tftp server. I have a package that builds the contents of the pxelinux.cfg; should this package output to something like $out/srv/pxelinux.cfg/$files or it could just do $out/$files? (I realise I'll then need some other glue to pull it altogether to point tftp.path at it)
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
justanotheruser has quit [Ping timeout: 252 seconds]
<ptrcmd>
tilpner: SetEnv would require sshd config changes
<ptrcmd>
tilpner: I am passing argument to bash to ssh
<ptrcmd>
tilpner: which is meant for bash
<exarkun>
Can I read out the static address assigned to a node nixpkgs/nixos/tests/make-test.nix somehow? So that I can put it in another node's hosts file.
<DigitalKiwi>
are there any stats available about that could indicate how popular a package is? like number of downloads or something
<tilpner>
ptrcmd: Then I can't help you
<ptrcmd>
tilpner: looks like removing the keyfile worked
<clever>
exarkun: this is how the hosts file is generated
<clever>
exarkun: and i think you can get a reference to the other machines, much like nixops
halfbit has joined #nixos
<clever>
exarkun: yeah, just { nodes, pkgs, config, ... }:
<exarkun>
reading that, I wonder if I should just set hostName and domainName on this node properly and have the existing logic write what I want to the hosts file
<clever>
that also sounds simple and easy
aveltras has joined #nixos
<exarkun>
That `nodes` argument looks like it's really good to know about too, though. Getting a reference to the rest of the nodes is pretty much where I was stuck, so thanks for that also.
<exarkun>
clever++
<{^_^}>
clever's karma got increased to 257
<lukego>
My Nvidia GPU is not providing any signal on its DVI output. Seems like this used to work and I don't know what's changed. Any tips for troubleshooting this on nixos master/17.09/17.03 (tried all)?
<hodapp>
lukego: have you poked around at xrandr to see if it's some kind of transient glitch?
<lukego>
hodapp: Hm. I don't have a working X display so I suppose I need xrandr to connect to the local X server when I'm running that over ssh. not sure how to accompliosh that. "DISPLAY=:0.0 xrandr" doesn't seem to work
<{^_^}>
[nixpkgs] @Lassulus pushed commit from @d-goldin to master « samba3, fusesmb: removing »: https://git.io/JeXfY
<nDuff>
Howdy -- I'm trying to use node2nix, but hitting what looks like a file-descriptor-exhaustion issue when building a very large project. Trying to fix it without modifying any generated code, though, is something that I'm needing to think about how to do. One approach that comes to mind is modifying the copy of stdenv I pass in on the argument list to node-env.nix to have a mkDerivation that sets ''ulimit -n hard'' -- but is there something easier/saner?
<jared-w>
node2nix can use a package lock file. Are you passing it in with -l?
<nDuff>
jared-w, yes; I don't have any problem with the selection of dependencies; the problem I have is that it's running out of file descriptors during operation.
<nDuff>
err, during *actual build time*, after all the code generation was successfully completed.
<jared-w>
right. It's probably implemented in an inefficient way unfortunately. Have you tried flattening the dependencies and then running node2nix?
<jared-w>
the larger the lockfile the more hideous the nix expressions and the larger the paths and at some point nix explodes, apparently
<nDuff>
What do you mean by "flattening the dependencies", in this context? Building one javascript package that incorporates the full dependency tree?
drakonis1 has joined #nixos
<nDuff>
...assuming I wanted Nix to build that one package, that seems like it would be moving the problem; I'm a little unclear on why/how it wouldn't just happen at that earlier point.
<jared-w>
npm dedupe is what i was thinking of. My thought process was that it seems like your nix expressions are getting too large if it's trying to create too many files at once. So if you reduced the amount of redundant packages in the package.lock, nix wouldn't have as much work to do
<jared-w>
But, not sure if that's really the underlying issue. Do you need to run something like nix-collect-garbage? or nix-store --optimize?
<nDuff>
I very intentionally have a huge Nix store -- multi-TB storage with a deduplicated filesystem underlying.
<nDuff>
Current attempt was to have a preInstallPhase of ''ulimit -n hard'', though that seems not to have done anything.
drakonis_ has quit [Ping timeout: 245 seconds]
<clever>
nDuff: is it even getting to the preInstall? add an echo to it?
<nDuff>
did that, looks like it isn't, prepending to installPhase instead (as that's where the failure is).
alexherbo2 has joined #nixos
<clever>
nDuff: i think its just preInstall, not preInstallPhase
<tilpner>
Is it an option to just globally raise the limit?
<nDuff>
@tilpner, ...change the whole system to make one package build, and then require any other build system to have the same change made before that same package will build there? That seems like it would be unfortunate. Yes, on Linux; yes, using nix-daemon.
<jared-w>
I'm not sure it would be required on any other build system. Your computer might just be running out of file descriptors due to the massive size of your nix store
<nDuff>
jared-w, why would it be opening every file or directory in the store?
<{^_^}>
[nix] @edolstra merged pull request #3238 → Ensure enough space in attrset bindings → https://git.io/JePl1
<tilpner>
nDuff: I suspect ulimit is working, but the systemd hard limits are different than what you expected
afics has joined #nixos
<nDuff>
tilpner, ...*shrug*. For my user, they're plenty high, and far higher than the default/soft limit. For the build user, it's a fair question, haven't checked yet.
<jared-w>
nDuff: mm, good point.
<nDuff>
that said, it looks like fixing the bug clever identified did indeed resolve my build :)
<nDuff>
...damnit, spoke too soon.
<nDuff>
ah well, need to run to lunch anyhow.
Rusty1 has quit [Remote host closed the connection]
levdub_ has joined #nixos
dingenskirchen has quit [Remote host closed the connection]
dingenskirchen has joined #nixos
Rusty1 has joined #nixos
<selfsymmetric-pa>
Trying out `optimize` for the first time.
<selfsymmetric-pa>
`error: cannot link '/nix/store/.tmp-link-10027-321231767' to '/nix/store/.links/00pi6zxnz2qnsw7l0drh3py7xr28xj97xxprb597iw1d2bdzl3vp': File exists`
<selfsymmetric-pa>
What's that about?
<selfsymmetric-pa>
Also some other questions to test my assumptions:
<selfsymmetric-pa>
- if I want to clear up some space, the right thing to do is probably collect-garbage followed by optimise, right?
levdub has quit [Ping timeout: 276 seconds]
<selfsymmetric-pa>
- if I use a deduplicating file store like zfs, then I don't need optimise, right?
<tilpner>
selfsymmetric-pa: 1. zfs doesn't dedup by default
__Myst__ has quit [Ping timeout: 265 seconds]
<tilpner>
selfsymmetric-pa: 2. nix-store file-level should be cheaper than block-level dedup
<{^_^}>
#20548 (by ToxicFrog, 3 years ago, open): zsh does not source set-environment when run as a non-login shell
__Myst__ has joined #nixos
<tilpner>
selfsymmetric-pa: 3. I don't expect there to be lots of block-level deduplication on files that wouldn't dedup with store optimisation (but that's a guess without data)
<jared-w>
yup that one is very non obvious. Luckily I managed to find that before I ran into the bug
Rusty1 has quit [Remote host closed the connection]
<tilpner>
selfsymmetric-pa: 4. You can set auto-optimise-store in nix.conf, and you shouldn't have to manually call --optimise
<selfsymmetric-pa>
tipner <3
<tilpner>
selfsymmetric-pa: 5. A GC alone would free space too, you don't need to couple every GC with an optimise
<{^_^}>
[nix] @edolstra merged pull request #3230 → Check for and repair bad .links entries → https://git.io/JeoYu
<selfsymmetric-pa>
Ah well. Not a huge issue. :)
<tilpner>
selfsymmetric-pa: zfs get dedup -t filesystem
gentauro has quit [Read error: Connection reset by peer]
<tilpner>
Is that on for your store?
gentauro has joined #nixos
<selfsymmetric-pa>
tilpner: I don't use zfs personally but I was curious. :)
<tilpner>
Oh
<selfsymmetric-pa>
haha sorry
<DigitalKiwi>
make sure your data will actually benefit from dedup if you're going to enable it
<DigitalKiwi>
there's a lot of extra overhead and depending on your data it might not be worth it (my data didn't benefit at all)
steell has joined #nixos
<DigitalKiwi>
ime enabling compression is almost always a good idea
<elvishjerricco>
selfsymmetric-pa: Also, having dedup on with ZFS isn't guaranteed to dedup identical files; only identical blocks
<clever>
DigitalKiwi: i turned on gzip-9 when i was fixing my nix store on the nas (moving it from / to /nix)
<clever>
DigitalKiwi: and i got a 2x ratio after re-writing the entire store back to disk on zfs
<elvishjerricco>
If you copied a file one byte at a time, syncing after every byte, you'd probably end up with tons of blocks, not identical to the larger blocks from the original file
<tilpner>
elvishjerricco: Only if you forced a sync each time, right?
<elvishjerricco>
tilpner: *probably*
<clever>
elvishjerricco: i think it will replace the tail block each time, until it hits some min block size
<elvishjerricco>
ZFS is free to block it up however it wants
<elvishjerricco>
clever: Oh really?
<tilpner>
That would be terrible space efficiency
<clever>
elvishjerricco: it it wont make a ton of 1 byte blocks
<clever>
elvishjerricco: i believe it will replace the tail block until it hits some min size, but the blocks will be larger if you wrote more without a sync
psyanticy has quit [Quit: Connection closed for inactivity]
<clever>
ah, uncompressed data was added, making the ratio worse
<elvishjerricco>
clever: Why not use lz4 rather than turning it off completely? I still get 2.22x compressratio on my nix store, and lz4 is basically free perf-wise
<lordcirth>
elvishjerricco, 2.22x? Nice, I should enable compression
<clever>
elvishjerricco: flipped over to lz4, and i'll leave it like that for a while
<elvishjerricco>
lordcirth: Yea, and that's on top of saving 10G from optimizing the store
<elvishjerricco>
Loooots of text files, and binaries with very similar contents in the store
<elvishjerricco>
s/similar/compressible/
<{^_^}>
[nixpkgs] @d-goldin opened pull request #74287 → nix-prefetch-git: list --branch-name in help → https://git.io/JeXUU
<rnhmjoj>
do you if it's possible to change the opengl driver (mesa) in nixos without rebuild the universe? there used to be a mesa_drivers attribute but it seems now it's only an alias for backward compatibility.
werner291 has quit [Remote host closed the connection]
alexherbo2 has joined #nixos
<{^_^}>
[nixpkgs] @Ma27 pushed to master « nixos/xss-lock: port test to python test-driver »: https://git.io/JeXk9
<tilpner>
rnhmjoj: I think that's the entire purpose of the /run/opengl-driver impurity
Neo-- has quit [Ping timeout: 250 seconds]
mananamenos_ has quit [Quit: Leaving]
<nDuff>
Hmm. When I set ''systemd.services.nix-daemon.serviceConfig.LimitNOFILE = 16384'' (overriding the value initially defined in nixos/modules/services/misc/nix-daemon.nix), it's coming back with: ''attribute 'systemd.services' at /etc/nixos/configuration.nix:196:3 already defined''
drakonis has quit [Ping timeout: 252 seconds]
ambro718 has quit [Quit: Konversation terminated!]
<rnhmjoj>
tilpner: uhm, right: but how to change it? i can't find any option to specify a package
philr has joined #nixos
<tilpner>
rnhmjoj: The option hardware.opengl.package specifies it
domogled has quit [Remote host closed the connection]
<rnhmjoj>
tilpner: ah, that's strange: i can see it exists with nixos-option but it's not listed in the configuration.nix man page
m1cr0man has joined #nixos
chloekek has quit [Quit: WeeChat 2.6]
<tilpner>
It's internal. See <nixpkgs/nixos/modules/hardware/opengl.nix>
<nDuff>
...hmm; I see systemd/boot/systemd-lib.nix providing a mechanism to populate an "overrides.conf" file that modifies a unit definition; just need to figure out what the path to get data into it looks like...
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #74295 → Remove lots of pygtk using software → https://git.io/JeXIP
shibboleth has quit [Quit: shibboleth]
<{^_^}>
[nixpkgs] @Ma27 pushed to master « nixos/iftop: port test to python test-driver »: https://git.io/JeXI1
domogled has joined #nixos
werner291 has joined #nixos
<m1cr0man>
Hi folks. I'm trying to build a netboot image from master but I'm getting an error while building append-initrd-secrets.drv, it says aarch64-linux is required to build it but im on x86_64-linux
<m1cr0man>
Obviously I want to build for x86_64
<clever>
m1cr0man: the netboot attribute is a set of netboot images for every arch
<clever>
m1cr0man: add .x86_64-linux to the end of the -A
aveltras has quit [Quit: Connection closed for inactivity]
virus_dave has joined #nixos
<virus_dave>
heyas! What’s the standard way to pass positional arguments to a command run via nix-shell? Specifically, how to workaround https://github.com/NixOS/nix/issues/534 ?
<{^_^}>
nix#534 (by Ericson2314, 4 years ago, open): nix-shell: argument redirection with --command
<{^_^}>
[nixpkgs] @rycee pushed commit from @dywedir to master « overpass: 3.0.3 -> 3.0.4 »: https://git.io/JeXLO
<nh2>
jared-w: yes, I can look into helping with that. Are you using static-haskell-nix for it already, and if yes, do you want to post your issue on my issue tracker?
<jared-w>
Yeah I'll post about it on the issue tracker. I suspect it's a misunderstanding of how to use it. But perhaps not. I've built an equivalent project using stack but this is the first time trying to build it with static Haskell nix. I can build it dynamically linked, just not statically
halfbit has joined #nixos
Jackneill has joined #nixos
<jasom>
So I ran into an issue with a terminal application not being able to find locale information when using nix on a non-NixOS host. Setting LOCALE_ARCHIVE to point to the host distributions locale-archive fixed it, as does installing glibc locales into nix and pointing it at that. How are locales supposed to work in this situation, and should this be in the manual?
philr has quit [Ping timeout: 268 seconds]
jco has quit [Quit: WeeChat 2.6]
fendor has quit [Quit: Leaving]
justanotheruser has joined #nixos
domogled has quit [Ping timeout: 250 seconds]
Jackneill has quit [Remote host closed the connection]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ennui>
m15k: if the resizing is limited to small sizes that don't fit your screen then it's the virtual VGA RAM size.
<ennui>
whats the best way to get the .out path of a .drv? is there something simpler than 'nix show-derivation mydrv | jq ...'?
o1lo01ol1o has quit [Remote host closed the connection]
<coderobe>
do i understand correctly that changes made with nix-env will be discarded on the next nixos rebuild?