<gchristensen>
I don't buy it, that XML is preventing docs. I think docs being hard to write is preventing docs.
alexteves has quit [Remote host closed the connection]
<gchristensen>
when I write docs, even when I have to look up syntax, the hardest and most exhausting part is figuring out how to describe the thing.
<infinisil>
Well I can only tell you about one datapoint, and that datapoint is me. I always get annoyed having to figure out how to insert these xlinks again, or to remember that I can't just make paragraphs in options
<worldofpeace>
Is there such a thing as `buildTools = [];` in nix?
<infinisil>
I've never written more docs than a forced release note, and I don't even want to write a full document
<gchristensen>
infinisil: then write it in markdown and I'll convert it for you
<infinisil>
On the other hand I'm using markdown-like stuff all the time for my local things, wiki entries, and co.
<aszlig>
infinisil/gchristensen: well, it's just a matter of learning the basic elements
<infinisil>
gchristensen: That's not the same. Converting to xml just means it's "write once" and everybody not being familiar with xml won't ever touch it
<aszlig>
i mean, when you don't know anything about markdown, you probably to look up the syntax as well
<gchristensen>
infinisil: dude, you're smart. every single .xml document in the nixpkgs repo has sufficient examples to copy-paste your way to success. don't give me that.
<infinisil>
Of course, but having to constantly copy paste my way through writing docs is annoying when it can be so much easier
<aszlig>
s/bly/& need/
<aszlig>
infinisil: you could also write it in markdown and use pandoc to convert it to docbook
<infinisil>
markdown like syntax is so universal, we write it in github comments all the time. Almost every dev will be familiar with it and won't have to learn anything
* gchristensen
is no longer participating in this conversation.
hellrazor has joined #nixos
<aszlig>
gchristensen: ah, don't give up so fast =)
<emily>
FWIW I think reStructuredText reuses DocBook's object model or something
joehh has joined #nixos
<joehh>
Hi, is there a simple way to package a single shell script?
<infinisil>
It would be cool to make a poll to get the opinion of lots of people on that
<gchristensen>
aszlig: I'll converse with you about docbook. I have no patience for people whose primary argument is "typing is the hardest part of the task"
IRCsum has quit [Ping timeout: 250 seconds]
<joehh>
I'm aware of writeShellScriptBin, but want to supply the contents from a file rather than
* emily
doesn't think typing is the hardest part of the task, but thinks it is the kind of thing that makes people give up quickly
<emily>
I also think Markdown is a bad format.
<aszlig>
emily: yeah, as said, rest is certainly in between markdown (very much inflexible) and docbook (xml - you know - scary, but quite powerful and flexible)
<gchristensen>
and it doesn't have to be scary
__monty__ has quit [Quit: leaving]
<aszlig>
emily: dito, if you for example take a look at how newline formatting is handled, it becomes even harder to read than docbook when you need to cramp a whole load of characters into a single line just to prevent breaks
<aszlig>
gchristensen: to me it isn't scary, but i do see the network effect here
<aszlig>
gchristensen: people are so used to markdown, they want it to be everywhere
<emily>
Sphinx would produce output with fewer weird HTML reference names and nicer navigation, probably.
<emily>
I think the end-user experience in the long run would be nicer.
<gchristensen>
emily: reference names?
<hodapp>
to me markdown starts to break down as soon as any more complicated structuring or formatting is needed :|
<emily>
gchristensen: you were complaining about something like that just before :p
<gchristensen>
emily: ah, idm/idp links :) I think they're all gone now!
carlosdagos has joined #nixos
<JonReed>
joehh: If you do "${./myScript.sh}" the file will be automatically added to nix store and referenced from there
thc202 has quit [Ping timeout: 250 seconds]
<JonReed>
joehh: Ah, you can actually do `writeShellScriptBin myScript (builtins.readFile ./myScript.sh)`. That will be even simpler
<lostman>
how can I pass `--arg` to one attribute of a derivation? I have `shell.nix` whose output is more than one derivation. I would like to invoke `nix-shell -A someAttr --arg enableThisOrThat true`. Is that possible? Or do I need to modify my `shell.nix` to pass arguments down explicitly?
Dedalo has joined #nixos
dmj` has joined #nixos
IRCsum has joined #nixos
rprije has quit [Remote host closed the connection]
astrofog has quit [Remote host closed the connection]
sigmundv__ has quit [Ping timeout: 245 seconds]
fzer00 has quit [Quit: Leaving]
<lostman>
I have something like this in `shell.nix` `{ myDerivation = wrapper imported.myDerivation }` and then `nix-shell -A myDerivation --arg someArg true` is not passed to `myDerivation`
<gchristensen>
18.09 is beta as of soon, and released at the end of the month, emily :)
lassulus_ has joined #nixos
lassulus has quit [Ping timeout: 240 seconds]
lassulus_ is now known as lassulus
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos
<MarkRBM_>
is there a way to find the path to an installed nix pkg? for example my old system .zshrc referenced the install path of oh-my-zsh as /home/mark/.oh-my-zsh/oh-my-zsh but nixos installed it for me this time and I dont know where to
<MarkRBM_>
I can see pathsToLink in the oh-my-zsh.nix but not sure that that is what I should point to in user land
<lostman>
clever: it returns a set
<clever>
lostman: --arg foo value, only works for functions in the form of { foo }: ...
<lostman>
clever: yeah, `wrapper` does overrideAttrs on a derivation. Maybe instead trying to push `--arg`s down I can use environment variables to control my build script. Less pretty but will work
<clever>
lostman: what i mean, is
<{^_^}>
[nixpkgs] @romildo opened pull request #45966 → qt5dxcb-plugin: init at 1.1.11 → https://git.io/fA8vg
<clever>
lostman: this will accept --arg someArg true
<clever>
lostman: and if wrapper returned such a function (and not a set), it would also work
<clever>
this applies at every attribute in -A foo.bar.baz
JonReed has quit [Ping timeout: 252 seconds]
drp has joined #nixos
Denommus` has quit [Ping timeout: 246 seconds]
<colemickens>
if I am booting and instance in Packet, and I pass it config with nix.nixPath=["nixpkgs=<some-path-to-nixpkgs>"], but I also pass it configuration options, how does this work?
<colemickens>
Does it build the nixPath first, and then re-parse the config?
<colemickens>
I'm having a hard time phrasing this question well, I've got myself confused.
<gchristensen>
nix.nixPath doesn't take effect until after the nixos-rebuild switch
<gchristensen>
which is VERY confusing at first thought.
<colemickens>
Hm. Is there a "trick" I can use to boot an instance with a custom nixpkgs and custom config then have it switch twice basically?
<colemickens>
(or a better way to achieve a similar end?)
<clever>
colemickens: note that nixpkgs.pkgs only effects the pkgs tree, and the nixos modules will not be effected
sir_guy_carleton has quit [Quit: WeeChat 2.0]
rauno has joined #nixos
<colemickens>
I've been fguring that out over the last hour :(
<colemickens>
Is there any way to achieve this without setting NIX_PATH and running nixos-rebuild again, clever?
hamishmack has quit [Ping timeout: 246 seconds]
<clever>
colemickens: the only way to effect the nixos modules is to set NIX_PATH before you run nixos-rebuild
endformationage has quit [Ping timeout: 252 seconds]
shreyansh_k has joined #nixos
<clever>
colemickens: or to do both at once, either with `NIX_PATH=.... nixos-rebuild ...` or `nixos-rebuild -I`
vandenoever has quit [Ping timeout: 240 seconds]
<colemickens>
Is that an intentional design decision?
<clever>
more of a limitation of how it works
<clever>
its the nixos modules that are responsible for reading configuration.nix and loading everything
<clever>
and its too late to change them by the time the config has been loaded
<colemickens>
Oh interesting.
jasongrossman has quit [Ping timeout: 240 seconds]
<colemickens>
I had assumed it somewhat the other direction in my head. Thanks for letting me know so I don't keep Googling the same 4-6 keywords for another hour.
<clever>
and in the case of the amazon AMI's, it will use the user-data to create /etc/nixos/configuration.nix
<clever>
and optionally change your channels
<clever>
which could deal with what you want
<clever>
a custom channel you control the updating of would effectively be pinned
<colemickens>
sure, but that's just cheating and running as an init script or something, if I had to guess?
<clever>
yeah, thats in the amazon-init.service within systemd
<clever>
after the first boot, it will read your user-data, change the channels and config, and run nixos-rebuild switch
<colemickens>
yeah, I had thought about something similar. I'm playing with Packet right now and actually can't get userdata working at all, so I'll have a chat when gchristensen is available again about supporting something similar.
<colemickens>
eventually I'd like to have everything automated such that there's a channel, so that seems like a fine approach to adopt, and there's precedent
<clever>
have you looked into nixops?
<colemickens>
Yeah. I was just mentally tossing it around more. I have been on and off again with NixOps a few times... I don't know if there's any integration with Packet at the moment, or how hard it would be to add. (and I'm a bit unclear on the way nixops works with adopted machines via ssh when the system already has configuration that I can't override)
<clever>
colemickens: as long as you can reproduce the filesystem and boot config locally (just copy in over), nixops can adopt anything it can ssh into
<clever>
the configuration.nix on the remote machine will just be ignored entirely
<colemickens>
that's what I'm afraid of though, there's networking/identity/machine-optimization stuff that I want to keep.
<clever>
you will need to properly keep that unique, when defining each machine in nixops
<clever>
then tar example up and ship it to somebody
<infinisil>
./run takes a list of packages to do it with, and the slowness of HIE's seems to depend on it
<clever>
after they unpack, change --to into --from, to cpy it back out into /nix/store/
<clever>
though your previous method saves time by reusing anything the user has in /nix/store, and cache.nixos.org helps out
<infinisil>
Yeah, and I don't need to host anything myself
<clever>
oh, i just noticed, your already using cachix to host it, lol
spacefrogg has joined #nixos
<infinisil>
clever: Let me know if the command works, it should take not too long, start an emacs instance, automatically kill it once the startup is complete, and output the result in a ./result file
<clever>
run is running
<infinisil>
However I wouldn't be surprised if there were problems when you run it..
<clever>
running auto-GC to free 3398696960 bytes
<clever>
waiting for the big garbage collector lock...
<clever>
infinisil: the nix-shell started a GC, and thats going to eat hie, lol
<clever>
infinisil: i also added $(date) to both echo's
<infinisil>
Darn
<infinisil>
clever: Okay one last test: When emacs started, go to the `putStrLn` identifier. Before it gets shut down again, does that identifier get highlighted red or not?
<clever>
infinisil: nope
<clever>
something flashed red in the status bar momentarily, but it vanished and wont recreate
<infinisil>
clever: Okay how about this: put an exit after the nix-shell call at line 30, then it shouldn't get killed
<clever>
i had bumped the sleep up to 45, but i can do that too
<infinisil>
clever: Then you can click the Main.hs on the bottom left to cycle through buffers, which should display some errors
lostman has quit [Quit: Connection closed for inactivity]
<infinisil>
Heh
<clever>
infinisil: emacs is open once more
<clever>
and it quit
<clever>
sleep raised to 30
<infinisil>
Darn, what he hell
<infinisil>
I should check it for errors somehow..
<clever>
i see the hie process consuming 93% cpu
<clever>
hie --lsp -d -l /tmp/hie.log
<infinisil>
Yeah that's correct, that's indeed the problem it's having right now
<clever>
its also using 428mb of ram
<clever>
474
<infinisil>
It's very weird, it only happens when the file you're using uses TemplateHaskell (or a dependency in your package) *and* only when you have certain packages as dependencies
<clever>
does it happen if you just run hie directly, without emacs?
<infinisil>
Some packages apparently add delay, while others don't add any delay
<infinisil>
clever: HIE needs a client to tell it to do stuff, which isn't very practical from the command line
<clever>
ah
<clever>
write a custom client that runs hie and gives it a simple cmd?
<infinisil>
I think I saw something like that recently
<lorilan>
Hi, is there a way to increase stack size ? I have an `error: stack overflow (possible infinite recursion)`. It's definitively not an infinite recursion: I have an expression to build scala project which is recursive but it works if I evalute it at "step" N and fails with the error when invoked at "step" N+1
vcunat has quit [Ping timeout: 252 seconds]
<clever>
lorilan: does --show-trace show any details?
carlosdagos has quit [Quit: Connection closed for inactivity]
Ridout has quit [Quit: Lost terminal]
TweyII has joined #nixos
trcc has quit [Remote host closed the connection]
trcc has joined #nixos
trcc has quit [Ping timeout: 246 seconds]
<{^_^}>
[cabal2nix] @peti pushed to master « cabal2nix: add post-process hook to configure the test environment properly »: https://git.io/fA8Cr
<cryptix>
hey - i want to add this package to nixpkgs but would also like to be able to nix-shell and build from the repo. is there a trick for that? another way i thought about this was some kind of 'someNewTool = callPackage (importFromGit //myUrl)' in nixpkgs but i dont know if that is a thing
<inquisitiv3_>
Could be that. I've verified the file with SHA256 sums. Coulnd't find any open issues @ Github.
inquisitiv3_ has quit [Read error: Connection reset by peer]
simukis has joined #nixos
trcc has joined #nixos
<tilpner>
I found the issue, now come back :c
lorilan has quit [Read error: Connection reset by peer]
<symphorien>
tilpner: you can use the ,tell functionnality of the bot
<tilpner>
,tell inquisitiv3_ It's a type-1 appimage. Try sudo mount -o loop, then steam-run bash and executing it
<{^_^}>
tilpner: I'll pass that on to inquisitiv3_
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-Cabal: keep a copy of the 2.2.x version around in the package set »: https://git.io/fA8RL
tehidiot has joined #nixos
janiczek has joined #nixos
ersran9 has joined #nixos
<janiczek>
Hello folks! I've had nix 2.0 working today on Ubuntu, ran `nix upgrade-nix` to get 2.1 and it basically broke my setup. Can you help me fix it?
sb0 has quit [Quit: Leaving]
<aminechikhaoui>
what's broken exactly ?
<janiczek>
I can't see `nix-*` binaries. When I looked deeper, links in my home directory are broken, ie. ~/.nix-profile links to ~/.nix-profile-1-link which links to ~/.nix-profile-1-link-1-link which links to /nix/store/...-user-environment which doesn't exist
<janiczek>
I looked into /nix/store/*-user-environment and there are two (with different hashes than what's in the link)
<aminechikhaoui>
janiczek: was there any issue while running `nix upgrade-nix` ?
<janiczek>
I don't think so, it seemed to go well, without any warnings
<LnL>
did you run a gc after upgrading? if not don't :)
<janiczek>
yeah I might have
<LnL>
yeah, I assume that's when you noticed something wasn't right
<janiczek>
so, to clarify, I probably did something like `nix-channel --update && nix-env -u && nix-collect-garbage -d` (stupid in retrospect) after the upgrade
<{^_^}>
nix#2070 (by chris-martin, 20 weeks ago, open): nix-collect-garbage deletes my user profile
<janiczek>
OK so I deleted the ~/.nix-profile* and did `ln -s /nix/var/nix/profiles/per-user/martin/profile /home/martin/.nix-profile` and it seems it has worked
<janiczek>
when I look at the symlinks in my home directory there is still a non-functional link in ~/.nix-defexpr/channels_root pointing at /nix/var/nix/profiles/per-user/root/channels
<janiczek>
is that important?
<symphorien>
I'd say no. the .nix-defexpr mechanism is legacy, you are unlikely to use it anyway
<janiczek>
Alright. Thank you very much!
janiczek has quit [Quit: Page closed]
fendor_ has joined #nixos
lorilan has joined #nixos
<lorilan>
clever, Hi, hope you're still there. The meeting was longer than expected, I set the stack size to 256MB and when I ran the expression it froze my computer ( which has 32GB of ram … ). I'll create the gist in the following minutes
<aiwakura>
I'm trying to get systemd-timesyncd not to stop the boot process
<booglewoogle>
System.Console.ANSI*
<aiwakura>
I think the reason why systemd-timesyncd is stopping the boot process is because it's set as Before=sysinit.target. Is there any way to change that from configuration.nix?
<Lears>
booglewoogle: your ghc and your ansi-terminal package don't know about each other. Try -p 'haskellPackages.ghcWithPackages (hspkgs: [ hspkgs.ansi-terminal ])'
<tilpner>
Putting them in boot.kernelModules, probably
<timokau[m]>
But in hardware-configuration.nix that is a list of simple module names, not packages. Don't I have to add the module package first somehow when it doesn't come with the kernel?
<shlevy>
samueldr: Do you know if we'll merge haskell-updates before branch-off? Just merged two important bugfixes and have another GHC patch to get in soon
<shlevy>
I think they're all backport worthy, I can also prepare the backports if it won't be merged yet
<gchristensen>
already branched
juhe has joined #nixos
rauno has quit [Ping timeout: 244 seconds]
<gchristensen>
better backport :)
<shlevy>
Ah, for some reason I thought it was happening tomorrow. OK!
<LnL>
most changes are generally still synced to the release branch until the release
<etu>
gchristensen: oh, has 18.09 already been branched?
<adamt>
Is there any documentation about the nix daemon protocol somewhere (the one the nix client uses when talking to the daemon)? And is it even expected to be sorta stable?
joehh has joined #nixos
<joehh>
hi, what is the best way to get cached haskell packages?
<joehh>
should I aim for particular releases of particular channels?
<joehh>
or can I tell from some sort of hydra status page that one git ref will have better coverage than another?
<johanot>
how to link to a manual section from a module option description? does it have to be an external link to nixos.org, or can it be an internal reference?
<joehh>
interesting.... so I can see when each channel has been updated
periklis has quit [Ping timeout: 240 seconds]
<lorilan>
Hi, I'm using nix to drive sbt builds but I'm at a point where I encounter an `error: stack overflow (possible infinite recursion)`. It is not an infinite recursion (the build is OK at rank N, fail at N+1 and then succeed if I increase the stack size) However to build "world", I need to set the stack size to 256MB, but then all my ram is used and my computer freeze … is this maybe normal (the build is quite big) or does my expression clearly have a bug ?
<{^_^}>
[nixpkgs] @thefloweringash opened pull request #45988 → emacs: allow X and cli-only builds on Darwin → https://git.io/fA8SY
<srhb>
jtojnar: You decide. If you want it to be able to run eg. kernel builds, give it big-parallel (or whatever it's called)
<srhb>
jtojnar: Basically grep through requiredSystemFeatures and select them all unless you explicitly want that machine to not be able to build that derivation.
<srhb>
(kvm actually does require kvm support on the host, though)
<srhb>
At least I think so. If not, it will be painfully slow at least.
jtojnar has quit [Remote host closed the connection]
<ilikeheaps>
I have a problem with my configuration and I found the generation on which the problem is introduced. Is it possible to see configuration.nix of an old generation?
lorilan has quit [Ping timeout: 272 seconds]
<srhb>
ilikeheaps: Only if you had system.copySystemConfiguration true
<ilikeheaps>
srhb: Alright, thanks. Now I wish it was true by default
<makefu>
ilikeheaps: keep in mind that this will only copy the configuration.nix, no included files. better keep your configuration.nix and dependencies under version control
<Lears>
I forget how, but you can inspect individual options in the running config. So presumably you could coax nix into dumping something like a config.nix.
<symphorien>
Lears: if you mean using nixos-option, it just evaluates your current configuration.nix
<Lears>
Oh..
<symphorien>
ilikeheaps: if the difference between the good and bad generations is small you may use nix-diff to compare them
<{^_^}>
[nixpkgs] @vcunat pushed to master « fixup! treewide: fix various links (homepage, src) »: https://git.io/fA85M
<{^_^}>
[nixpkgs] @vcunat pushed to release-18.09 « treewide: fix various links (homepage, src) »: https://git.io/fA85Q
<tos9>
LnL: A ha... thanks. That does seem to have finally unhosed me.
<tos9>
I'm sure this is a dumb question, but why does nix-env -i '*' not imply `nix-env -iA nixpkgs.nix`
<tos9>
Oh, whoops, nope
<tos9>
Spoke too soon
<tos9>
Still hosed.
<{^_^}>
[nixpkgs] @vcunat pushed to release-18.03 « treewide: fix various links (homepage, src) »: https://git.io/fA85A
<LnL>
what did it upgrade to?
<tos9>
LnL: looks like it installed 2.0.4 again
<LnL>
then your channel doesn't have 2.1 yet
<LnL>
you probably have 2 versions of nix in PATH, one in ~/.nix-profile/bin and one in /nix/var/nix/profiles/default/bin
dbmikus has joined #nixos
<tos9>
I do
<tos9>
The latter saying it's a 2.0.4
hyper_ch2 has quit [Quit: Page closed]
<tos9>
LnL: How do I uninstall enough to get back to a working env -- running `sudo rm -rf /etc/nix /nix /var/root/.nix-profile /var/root/.nix-defexpr /var/root/.nix-channels /Users/Julian/.nix-profile /Users/Julian/.nix-defexpr /Users/Julian/.nix-channels` was apparently not enough
<tos9>
(After reinstalling, the problem was back again)
nbp has joined #nixos
<LnL>
you'll also have to unload the daemon
<LnL>
there's an uninstall section in the manual
grp has joined #nixos
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @edolstra pushed to master « Manual: Random indentation fixes »: https://git.io/fA8Fr
<xok>
I now have a fully working NixOS but I don't know how to enable networking...
<xok>
the system has no internet connection, I am connected to it through a "console"...
<xok>
can anyone help me confiugre the networking for it?..
Orbstheorem has quit [Ping timeout: 240 seconds]
Dedalo has joined #nixos
Orbstheorem has joined #nixos
__Sander__ has quit [Quit: Konversation terminated!]
kandinski has joined #nixos
<kandinski>
hi, does nixos-install wipe /home?
<kandinski>
I managed to mess up my install and I'd like to rerun the install process on the existing disk, but not if it will delete data
<gchristensen>
no
<kandinski>
thanks
oscarvarto has joined #nixos
<oscarvarto>
Hi! I am following installation guide and need some help
<oscarvarto>
I am using fdisk and want to "make partition bootable"
<oscarvarto>
Guide says to use `a`
<oscarvarto>
But there is no such command
<thefloweringash>
@xok I have no particular experience, but if you want to muddle through it with me: can you verify that there's an /etc/network/interfaces, that /etc/nixos/nixos-in-place.nix has a "systemd.services.setup-network", and find the names of the network interfaces by running `ip link`.
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<bigvalen>
Seems that it's because my HP has a flaky power sensor of some sort that's not ACPI compliant.
<bigvalen>
if I'd lm-sensors on my box, i'd tell it to ignore the sensor. Is there anything else installed with a base nixos install that'd fiddle with that ?
<bigvalen>
Oooh. I wonder does prometheus poke it.
mayhewluke has quit [Ping timeout: 250 seconds]
simukis has quit [Quit: simukis]
mayhewluke has joined #nixos
teej has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
Dedalo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sir_guy_carleton has quit [Quit: WeeChat 2.0]
sigmundv__ has quit [Ping timeout: 252 seconds]
kenshinC1 has joined #nixos
<andi->
teto: findFirst?
tos9 has left #nixos [#nixos]
Tucky has quit [Quit: WeeChat 2.2]
Orbstheorem has quit [Ping timeout: 246 seconds]
chreekat has quit [Ping timeout: 240 seconds]
Thra11 has joined #nixos
Dedalo has joined #nixos
aarvar has left #nixos [#nixos]
chreekat has joined #nixos
tehidiot has joined #nixos
xok has joined #nixos
<xok>
Lears: I have destroyed the old CentOS system and built a new Debian 8...
<xok>
it still has a network problem...
<xok>
I am not able to connect to the outside world...
justanotheruser has joined #nixos
kiloreux has joined #nixos
<Lears>
xok: I think you may have meant to ping someone else.
<xok>
network doesn't work, I can't neither ping, nor dig or do anything network related...
<drakonis>
xok, do you have the firmware package?
<xok>
what firmware?...
<drakonis>
debian by default doesn't install the linux kernel firmware
<drakonis>
centos didn't do that until recently
<justanotheruser>
how can I ensure that any internet connection goes through openvpn?
<Lears>
So I wrote a thing for opening local html files (particularly haddocks) through a static web server; it's packaged through a nixos module. I'd be interested in feedback if anyone wants to take a look at it or try it out: https://github.com/LSLeary/lsfws
<bigvalen>
I've a scoping question; at the very end of https://gist.github.com/BigValen/7266d80a44a02227f631e844621e1354 I'm trying to re-write a prometheus node-exporter startup script. But I get told "Can't find 'prometheus'". How do I reference it properly ?
<bigvalen>
I tried changing it to 'services.prometheus', but then it complained it couldn't find 'services' either.
<symphorien>
use config.services.prometheus instead
<symphorien>
config is an argument to the whole file and enables you to read the current values of options
<bigvalen>
Sound
<teto>
andi-: findFirst + stringToCharacters got me there thanks
worldofpeace has joined #nixos
<worldofpeace>
Anyone want to test #45967 ? the borg isn't feeling so great today
<worldofpeace>
Wow, that's quiet cautious of you. I wouldn't have seen that
<{^_^}>
[nixpkgs] @jwiegley pushed to master « ledger: Enable building of tests, and concurrent make »: https://git.io/fA4Ue
<gchristensen>
I have pretty good monitoring in place
<gchristensen>
:)
<gchristensen>
so when nothing seemed broken and this was broken, I started with the assumption it was something in the command itself
Ariakenom has joined #nixos
<hodapp>
anyone know of examples in nixpkgs of something like a CMake build calling a secondary CMake build in the same repo? are there cleaner solutions than just adding some project(...) calls to the top-level CMakeLists.txt?
<{^_^}>
[nixops] @AmineChikhaoui merged pull request #998 → Documentation about special NixOS module inputs → https://git.io/fAcIx
<{^_^}>
[nixops] @AmineChikhaoui pushed 2 commits to master: https://git.io/fA4UO
<infinisil>
hodapp: Calling cmake in the configure phase twice doesn't work?
<{^_^}>
[nixpkgs] @jwiegley pushed to master « Revert "ledger: Enable building of tests, and concurrent make" »: https://git.io/fA4Uc
sigmundv__ has joined #nixos
<worldofpeace>
hodapp: Is it like a subdirectory?
johanot has joined #nixos
<hodapp>
worldofpeace: yes, but it needs the results of the first build
<hodapp>
it is some plugins/samples that use a library just built
johanot has quit [Remote host closed the connection]
IRCsum has quit [Remote host closed the connection]
rfold has joined #nixos
<worldofpeace>
hodapp: I'm not a cmake wizard but, you shouldn't need to do that. cmake has commands for that
drakonis has quit [Remote host closed the connection]
<hodapp>
worldofpeace: commands for what?
IRCsum has joined #nixos
nikivi has quit [Quit: ZNC is awesome]
<hodapp>
the samples/plugins, AFAIK, are separate builds that the main build doesn't include
johnny101 has joined #nixos
<worldofpeace>
Then I would think that the author would then put that under a build option
<{^_^}>
[nix] @vcunat pushed to master « docs: change expired bzip2.org to archive.org »: https://git.io/fA4T7
nikivi has joined #nixos
<kiloreux>
I am trying to install aws-cli from pythonPackages but the output is consistently failing with `warning: setlocale: LC_ALL: cannot change locale (en_US.UTF-8)`
<sphalerite>
the former approach is more versatile but complex, and the latter is a bit more limited but also "safer" (it won't touch anything outside ~/.nix-profile really) and simpler
redj has quit [Disconnected by services]
<Thra11>
sphalerite: Neat. I'll try the overlay approach. Thanks!
redj has joined #nixos
<{^_^}>
[nixpkgs] @cryptix opened pull request #45997 → systrayhelper: init at 0.0.2 → https://git.io/fA4kh
<elvishjerricco>
I'm trying to package something that expects to find some libs with pkgconfig, but those libs don't seem to ship with pkgconfig files in nixpkgs. Is there an automatic way to generate them?
<elvishjerricco>
In particular, libgcrypt, uriparser, and uthash are being searched for
<Neo-->
hey all, question - I would like to install virtualbox but it doesn't look liki it's installing? I've included linuxPackages.virtualbox, virtualboxHardened and virtualboxGuestAdditions in systempackages, and included the virtualisation.virtualbox.host.enable = true. Anything else that I should be doing?
<clever>
Neo--: the only thing you need to do is set virtualisation.virtualbox.host.enable = true and then nixos-rebuild switch
<clever>
Neo--: how are you checking to see if you have virtualbox?
<Neo-->
cli autocomlete of various vbox/virtualbox/etc :)
<srhb>
elvishjerricco: On thinkpads it's a fun process involving booting into a FreeDOS live disk :P
<{^_^}>
[nixpkgs] @xeji pushed commit from @aanderse to release-18.03 « kodi: fix nfs support (#46006) »:
<gchristensen>
same
<joehh>
how can I add the python3.withPackages into the function arguments at line 7? or do I just build my own python3withpackages in the let expression?
oscarvarto has joined #nixos
<elvishjerricco>
srhb: Lol hopefully a Ryzen board will be easier.
<srhb>
joehh: You can't have dots in the arguments, but this method won't work regardless
<clever>
elvishjerricco: interesting, i have a TPM visible on my laptop
ryanofsky has quit [Remote host closed the connection]
<srhb>
joehh: Oh, actually, it might... Hang on, let me try.
<elvishjerricco>
Oh here's some interesting logs:
<elvishjerricco>
kernel: tpm_crb MSFT0101:00: can't request region for resource [mem 0xb900e000-0xb900efff]
<elvishjerricco>
kernel: tpm_crb: probe of MSFT0101:00 failed with error -16
<clever>
elvishjerricco: any fun commands i can run to query mine?
<{^_^}>
[nixpkgs] @xeji merged pull request #46010 → ion: broken on darwin →
<{^_^}>
[nixpkgs] @xeji pushed commit from @dywedir to master « ion: broken on darwin (#46010) »:
<elvishjerricco>
clever: You may have some luck with the `tpm-tools` package, but it requires enabling the `tcsd` service. I'm trying to package the `tpm2-tools` package right now, which seems required for the TPM 2.0 spec (which is very likely what you have)
<{^_^}>
[nixpkgs] @xeji pushed commit from @dywedir to release-18.09 « ion: broken on darwin (#46010) »:
<elvishjerricco>
clever: But the very fact that you have /dev/tpm0 shows you're already further than me :P
<oscarvarto>
Hi! When I run `nixes-rebuild switch` I am getting "This account is currently not available"
<oscarvarto>
How can I fix that? I have googled but cannot fix it
<srhb>
joehh: Right, just grab python3 as an argument, then add (python3.withPackages (ps: with ps; [ matplotlib basemap numpy ] )) to librarySystemDepends
<srhb>
joehh: Note that it appears to be quite thoroughly broken with a modern python matplotlib and you need to set doCheck = false; in order for the tests to pass, even if you add the right packages (the same from before plus some texlive stuff)
<srhb>
joehh: But this will at least allow you to spin up a cabal new-repl in the project dir that has access to the relevant python stuff.
<joehh>
awesome - thanks for that srhb!
philippD has joined #nixos
<joehh>
not surprised it could be thoroughly broken, but trying will be the best way to fix it :)
<srhb>
joehh: Definitely. It might require either updating (best) or grabbing some older python packages (fastest :P)
<srhb>
joehh: fwiw it looks like only one test fails with the dependencies above in testSystemDepends plus texlive.combine { inherit (texlive) scheme-full; } -- and that can probably be trimmed way down
<symphorien>
try reverting this commit locally and rebuilding
oscarvarto has quit [Quit: This computer has gone to sleep]
<symphorien>
in any case, it is probably worth opening an issue
<LnL>
Is there a beginners “Homebrew to Nix”?
<srhb>
joehh: But I'm able to have file "test" $ plot [1,42] [1,42] save to test.png correctly... can't get onscreen to work though...
<LnL>
does anybody think of a good resource like that?
jtojnar has quit [Remote host closed the connection]
<elvishjerricco>
I may just have to update my bios...
jtojnar has joined #nixos
<joehh>
srhb: I suspect onscreen may need some wxpython or gtk packages added and maybe a little bit of config somewhere to tell it which backend to use
<srhb>
joehh: Ah, OK. Maybe less broken than suspected then. :)
tehidiot has joined #nixos
<{^_^}>
[nixpkgs] @romildo opened pull request #46012 → qt5integration: init at 0.3.5 →
baimafeima has quit [Remote host closed the connection]
trcc has quit [Ping timeout: 244 seconds]
v0|d has joined #nixos
kenshinC1 has quit [Ping timeout: 246 seconds]
<v0|d>
hello, i'm trying to migrate to a new machine. I'd like to have access to my old stuff. Lets say I mount my old nixos root and do chroot. How can I find the script thats sets the proper environemnt?
<v0|d>
I thought its something like ~/.nix-profile/bin/env.sh yet faild.
__monty__ has quit [Quit: leaving]
<clever>
v0|d: you want nixos-enter
<clever>
for example, `nixos-enter --root /mnt`
<v0|d>
clever: thnx, much appreciated.
<oscarvarto>
symphorien, Thanks a lot for replying! I will take a look asap
<oscarvarto>
By the way, I just triggered a `nixos-rebuild switch` and now it seems to be recompiling a lot of stuff
<oscarvarto>
Is that supposed to happen?
<oscarvarto>
I am trying to use zsh as my default shell
<v0|d>
is there a way to convince nix-store to compile everything from scratch instead of dlng?