<Matthieu_>
If the version is updated, then everything works well.
H55434 has quit [Ping timeout: 248 seconds]
polman has joined #nixos
<Matthieu_>
I'm writing a mail to the maintainer (I'm not sure he'll be able to make something about this)
<zfnmxt>
Does anyone have an inkling why `fc-list | grep gohufont` only turns up the 11 px font? The package should include both 11 and 14, but I update my system today (after a number of months...) and now the 14 px font is just gone. Here's the nix expression: https://github.com/NixOS/nixpkgs/blob/master/pkgs/data/fonts/gohufont/default.nix
<simon_weber>
would it be blasphemous for me to manage app servers through nixos, but deploy their code manually? or is this something of an accepted pattern?
<samueldr>
known to be done
<simon_weber>
in my case I feel like I'd have better control over the deployment process (ie running migrations and whatnot) that might be annoying to do inside nix packaging
<samueldr>
and if you're not used to nixos, a good first step, at least the system is declarative and reproducible :)
<aanderse>
simon_weber: i'm pretty sure a few people use tomcat module in nixos, but deploy code themselves
<aanderse>
and as far as apacheHttpd goes i manage all virtual hosts via local config files debian style (ie. /etc/httpd/sites-enabled)
<aanderse>
samueldr: so line 11-12 checks if the mediawiki database is installed and returns a 1 or 0 exit code
<samueldr>
aanderse: anything particular to look for or you were just looking for [...]
<simon_weber>
cool, I'm relieved to hear that's not a ridiculous thing to do, haha. I'm already reasonably confident in the app-specific side of things, but nixos already seems like a big improvement over my previous ansible setup
<samueldr>
right
<aanderse>
i use that value to conditionally check if i should run the install script (line 13+)
<aanderse>
here is the helpful message systemd gives me: mediawiki-init.service: Main process exited, code=exited, status=1/FAILURE
<aanderse>
-_-
<samueldr>
the last command to run was a failure, 1, because the table exists?
<samueldr>
that would be my 10s assumption
<samueldr>
(false && true) # return value 1
<aanderse>
if db is not installed it ran fine and installed db
<aanderse>
if db installed it was crashy crashy on me
<aanderse>
oh...
<aanderse>
wait a sec
<aanderse>
yeah i'm bad at bash
<aanderse>
so...
<aanderse>
if the install script never returns
<samueldr>
in bash, `x && y` will return either success if x and y succeed, or the failure of the first one that fails
<aanderse>
that means the last value in the script...
<aanderse>
yeah
<aanderse>
>_<
<samueldr>
it's implicitely the last return value
<aanderse>
sometimes you just need to talk things out, right :)
<aanderse>
yup
<aanderse>
ha silly me
<catern>
simon_weber: depends on what you mean by "manually"; it's still very nice to package your code with nix, and I don't see a reason why you shouldn't do that; but deploying it with nixops or nixos is definitely not necessary. i myself don't use nixops or nixos :)
<aanderse>
thanks
<samueldr>
might want to set -e at the start, so it exits on first failure, and instead run a proper `if` rather than chaining with `&&`
<samueldr>
the if will convey your meaning
<samueldr>
or continue like that and ping me on the PR and I'll look at the bash more in details
<aanderse>
samueldr: definitely pinging you for testing :)
<aanderse>
did you have a specific db in mind?
<aanderse>
i was nearly mamed in the mediawiki channel for mentioning a db other than mysql
<samueldr>
I don't remember what the current wiki runs on
<aanderse>
they said "no... don't do that. just use mysql."
<samueldr>
hm?
<aanderse>
apparently other db support for mediawiki is not good they say
<clever>
aanderse: i once got banned from #debian for mentioning i was running ubuntu, lol
<aanderse>
clever: lol
<samueldr>
is it a channel with authoritative users or only end-users being kinda nooby together?
<aanderse>
i'm being sarcastic, but they really emphasized i should either only support mysql or highly suggest to only use mysql
<clever>
dont really remember, it was ages ago
<samueldr>
aanderse: likely to be mysql, from looking at my notes
<samueldr>
clever: the question was more for aanderse :)
<aanderse>
one of the wikimedia foundation guys was helping me out
<aanderse>
stellar guy
<samueldr>
right
<aanderse>
he walked me through everything i needed to know
<aanderse>
anywho, samueldr you have an existing mediawiki instance you want to migrate to nixos or something?
<samueldr>
aanderse: the current nixos.wiki isn't declaratively configured, and in the past wanted to do handle doing it, but never really ended up doing it
o1lo01ol1o has quit [Remote host closed the connection]
<samueldr>
it would greatly help in improving it by just sending PRs for the config
<samueldr>
(I am not the one hosting it)
o1lo01ol1o has joined #nixos
<aanderse>
ah
<aanderse>
ok
ryantrinkle has quit [Ping timeout: 268 seconds]
<samueldr>
the current iteration has declarative config up to the web server, but it a traditional php install like just previously asked "if anyone does that" :)
<aanderse>
well at least it isn't using the httpd.extraSubservices mediawiki module
<aanderse>
:)
<{^_^}>
[nixpkgs] @jonringer opened pull request #62746 → pythonPackages.azure-cli-core: init at 2.0.66 → https://git.io/fjuNK
<o1lo01ol1o>
i have some keys configured with deployment.keys that i'm able to pass to a hydra server and build machines in nixops. How can I configure the hydra evaluator to use one set of keys (to talk to github) and the hydra-queue-runner to use another (to talk to the build machines)?
<gchristensen>
you can give the hydra-queue-runner user an ssh config
<o1lo01ol1o>
gchristensen: can you point me to an example of that?
<gchristensen>
ehh just like sticking one in hydra-queue-runner's ~/.ssh/config
<o1lo01ol1o>
ah right
justan0theruser has joined #nixos
<o1lo01ol1o>
is there a way to get the key from en environment variable? Since they're not actually on the disk, it'd be nice to keep them that way
<clever>
i just run `ssh-keygen` with no args (and no passphrase) as the hydra-queue-runner user
<clever>
and then give it read on the github repo
<gchristensen>
you can use IdentityFile /run/keys/...
<clever>
that should also work
<o1lo01ol1o>
wait, which user sends the jobs to the build machines and which checks github?
<o1lo01ol1o>
gchristensen: cool, what's that called so I can lookup the docs?
justanotheruser has quit [Ping timeout: 244 seconds]
<gchristensen>
IdentityFile in `man 5 ssh_config`
<clever>
o1lo01ol1o: hydra-queue-runner is the user that connects to build machines via ssh
<o1lo01ol1o>
clever: ok, so if `hydra` has one set of keys to talk to github and hydra-queue-runner is pulling the nixops key to talk to the build machine, everything should be gravey
<o1lo01ol1o>
gchristensen: ah right, I thought that was a nixops tag.
thomasd has joined #nixos
<thomasd>
Hi all, I'm giving a talk tomorrow and I have never plugged an HDMI cable into my laptop running NixOS... am I going to have trouble tomorrow when I setup to present?
<jackdk>
find a screen with hdmi in and find out. Don't be that guy who fights AV gear just before the talk starts
<thomasd>
Don't worry, I have 2hrs at the venue before my talk to resolve any issues that crop up. I will be skipping the keynote if I need to. I was mainly asking just to know what to expect; if it's a "definitely" or a "maybe".
<jackdk>
my laptop with nixos has never had trouble over its miniDP connection, that's all I know
<thomasd>
NixOS is a lifestyle :'D
polman has quit [Ping timeout: 244 seconds]
<thomasd>
Q: Will it work on the first try?
<thomasd>
A: Probably not
ddellacosta has joined #nixos
drakonis1 has joined #nixos
polman has joined #nixos
srl295 has quit [Quit: Connection closed for inactivity]
aanderse has joined #nixos
hedning_ has quit [Remote host closed the connection]
thomasd has quit [Quit: Page closed]
<samueldr>
tbf, the question is more "will it work on Linux?"
<Mateon1>
Hi, I'm on nixos on the unstable channel and recently some things seem to have broken after an update, some tools like meld (graphical diff) and mousepad (xfce notepad) broke. The programs still sort of work, but it's impossible to toggle any configuration, like line wrapping, font or line numbers in either program
<Mateon1>
If I start the programs with --disable-server so I can examine the stdout, they keep complaining about dconf
<Mateon1>
(mousepad:75438): dconf-WARNING **: 04:41:31.985: failed to commit changes to dconf: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name ca.desrt.dconf was not provided by any .service files
<Mateon1>
(I'm unsure if dconf is related to the config not working)
<Mateon1>
I tried installing nixos.xfce.mousepad which is still mousepad-4.0 (instead of nixos.xfce4-13.mousepad - mousepad-4.1), but that immediately segfaults when I try to start it
<Mateon1>
Sorry, got the version numbers wrong, mousepad-0.4.0 and mousepad-0.4.1 respectively
Ralith_ has quit [Remote host closed the connection]
Ralith_ has joined #nixos
<colemickens>
"Will it work in NixOS?" => "how much Nix do you want to write today?"
<colemickens>
(but I mean it in a good way... I think?)
H55434 has joined #nixos
H55434 has quit [Client Quit]
hiroshi- has joined #nixos
hiroshi- is now known as hiroshi
schjetne has joined #nixos
<{^_^}>
[nixpkgs] @teto opened pull request #62751 → Luazlib renamed to lua-zlib + update + move to generated → https://git.io/fjuAQ
<mjrosenb>
can someone help me come up with a nix expr that will build two separate targets? I want two things to be visible inside of a nix-shell. (sorry, I ask this like once a week, and never have enough scrollback to see if anyone answered)
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<li_matrix>
mjrosenb: do you mean two different target system?
<li_matrix>
for cross compilation?
<li_matrix>
or two derivations?
<mjrosenb>
cross compilation I guess is a good analogy.
<li_matrix>
cause 'target' is usually used to talk about target architecture in nix
<li_matrix>
maybe you meant two derivations
<li_matrix>
its not so clear
<mjrosenb>
I have a project that I'm compiling with ghcjs, and I'd like to be able to use some emacs extensions that require all the libraries be build targeting x86
<li_matrix>
check out 'buildEnv'
<mjrosenb>
two different derivations would work.
<mjrosenb>
I'm used to make terminology.
<li_matrix>
ok so you want two derivations in a single expression, you can use the `buildEnv` function
<li_matrix>
it'll put all your derivations in a single 'environment'
<li_matrix>
thats now nixos and nixos containers are built
<li_matrix>
theres a builtin function by that name, but you most probably want to use the nixpkgs function of the same name
<li_matrix>
which will be `pkgs.buildEnv` if nixpkgs is imported as pkgs
<mjrosenb>
I'll grep through my checkout for buildEnv.
buffet_ has joined #nixos
buffet has quit [Ping timeout: 246 seconds]
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
<li_matrix>
pkgs/build-support/buildenv
reinhardt has joined #nixos
<{^_^}>
[nixpkgs] @teto merged pull request #62751 → Luazlib renamed to lua-zlib + update + move to generated → https://git.io/fjuAQ
<mjrosenb>
that's my default.nix, and it works. I want to be able to run "nix-shell -A env", it also grabs the build dependencies for nixpkgs.haskell.packages.ghc.callCabal2nix "verify" ./. {}
<clever>
mjrosenb: the .env is meant to be used with nix-shell, and cant be put into buildEnv
<clever>
mixing 2 ghc's like that may be difficult
<clever>
mjrosenb: have you thought of just doing most of the development on x86, and testing with ghcjs every now and then?
cantstanya has joined #nixos
<mjrosenb>
it is a game, and I do all of my rendering by manipulating html/svg
<clever>
ah
<clever>
mjrosenb: why does emacs need the x86 version of things?
schjetne has quit [Ping timeout: 248 seconds]
<mjrosenb>
I'm trying to get my old emacs/haskell environment set up, and it launches an inferior ghci process, runs imports, and tells me the types of expressions.
toppler`` has quit [Ping timeout: 258 seconds]
<clever>
mjrosenb: does ghci work on ghcjs?
<mjrosenb>
ghcjs: --interactive is not yet supported.
<clever>
ah
<mjrosenb>
at least that is what cabal new-repl tells me.
<clever>
what about having one shell for ghcjs, and then another shell for x86+emacs ?
<mjrosenb>
hrmm, I like being able to compile from within emacs, and have been doing that by just running emacs inside of my nix-shell that is configured for ghcjs.
<clever>
yeah, i can see the trouble
<clever>
let me try something...
dbmikus has joined #nixos
toppler`` has joined #nixos
<mjrosenb>
so, the miso website and server are both built with nix in a single nixexpr
cantstanya has quit [Remote host closed the connection]
cantstanya has joined #nixos
<li_matrix>
in my view nobody has yet developped a very good nix-based CI, which explains why nixos hasnt taken over debian. there are a few projects pursuing this from different direction and I have my own. I find that keeping an open mind and putting as much work as possible into nix is the right way to go. So maybe you can start rethinking your emacs testing flow, mjrosenb . That's the conclusion I came to, so Im looking at creating
<li_matrix>
a different interface than nix's commands, by wrapping them at first.
<mjrosenb>
well, the nix file that clever gave me is building now. I guess I'll find out how it worked sometime tomorrow.
<mjrosenb>
I suspect I'll have to add the same doctest exclusion
<mjrosenb>
or at least a similar one.
Glider_IRC_ has joined #nixos
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos
Glider_IRC__ has quit [Ping timeout: 248 seconds]
<mjrosenb>
well, that failed... but it failed in building network, which I don't actually need, so, trying that again without.
mexisme_ has quit [Ping timeout: 245 seconds]
<clever>
mjrosenb: was trying to think of something thats on both
orivej has joined #nixos
ddellacosta has quit [Ping timeout: 248 seconds]
emptyflask has joined #nixos
justan0theruser has quit [Ping timeout: 248 seconds]
Rusty1 has quit [Quit: Konversation terminated!]
justanotheruser has joined #nixos
orivej has quit [Ping timeout: 248 seconds]
mexisme_ has joined #nixos
endformationage has quit [Ping timeout: 258 seconds]
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #62752 → Fixes for x86_64-darwin failures on Hydra → https://git.io/fjuxG
palo1 has joined #nixos
palo has quit [Ping timeout: 258 seconds]
palo1 is now known as palo
mexisme_ has quit [Ping timeout: 248 seconds]
reinhardt has quit [Quit: Leaving]
jasongrossman has quit [Ping timeout: 248 seconds]
gyroninja has quit [Quit: WeeChat 2.4]
<gentauro>
I'm looking to install `LaTex` on my NixOS box
<etu>
gyroninja: And unless you added that config (that in that form doesn't tell us much at all) to an overlay on your configuration, I wouldn't know what it does
<gyroninja>
I don't seem to have the /python command though, maybe there's something else wrong
<etu>
Which version of weechat did you get?
<etu>
Or which channel are you on?
<gyroninja>
It installed weechat 2.4
<gyroninja>
I'm on unstable
<{^_^}>
[nixpkgs] @lightbulbjim opened pull request #62757 → libblockdev: add gptfdisk as a dependency → https://git.io/fjuxV
<etu>
gyroninja: Yeah, that should be fine then... I should check my weechat, just need to fetch a laptop (on the relay in my phone atm)
<gyroninja>
thanks
<etu>
I have the /python command
<etu>
Maybe plugins needs to be enabled or something... It was such a long time ago I set up weechat
dbmikus has quit [Remote host closed the connection]
<mjrosenb>
clever: as expected (sort of), doctest failed to build.
palo1 has joined #nixos
etbe has joined #nixos
palo has quit [Ping timeout: 248 seconds]
palo1 is now known as palo
<etbe>
In coreutils there is a selinuxSupport option which defaults to "false", how do I setup a configuration.nix to make coreutils be built with selinuxSupport=true and have that be the coreutils for the entire system?
schjetne has quit [Ping timeout: 244 seconds]
<gyroninja>
okay found something
<gyroninja>
the python plugin IS in my nix store
<gyroninja>
but in a different package than regular weechat
<gyroninja>
I think I can figure out the rest myself (will just need to figure out how to autoload it)
<pacman99>
etbe: in nixpkgs.config in your configuration.nix set coreutils.selinuxSupport=true
<ploupiboulba[m]>
I've enabled the nvidia driver, how can I get nvidia-xconfig ? I've lost audo output too
nschoe has joined #nixos
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #62758 → nixos/zoneminder: font files cannot be found → https://git.io/fjuxp
reinhardt has joined #nixos
erudition has quit [Ping timeout: 245 seconds]
alp has joined #nixos
linarcx has joined #nixos
schjetne has joined #nixos
<gyroninja>
If I wanted weechat to use python 3 would I need to make my own package?
emptyflask has quit [Ping timeout: 252 seconds]
<gyroninja>
actually I don't need python3, though a package I need for a plugin seems to be restricted to above python 3.3
<ar>
weechat scripts now work with python3? last time i tried, while weechat itsefl compiled and seemed to work fine with py3, scripts didn't :(
<gyroninja>
I don't know how to make it use python 3 on NixOS
schjetne has quit [Ping timeout: 248 seconds]
<gyroninja>
I don't need it though, I had misremebered what I had compiled against on my previous system
<etbe>
pacman99: sorry if I'm asking stupid questions, but I tried "nixpkgs.config = { coreutils.selinuxSupport = true; };" and it didn't change anything. Thanks for the suggestion but I'm a beginner at NixOS.
mighty_vee has quit [Ping timeout: 258 seconds]
wfranzini has joined #nixos
nschoe has quit [Ping timeout: 252 seconds]
<gyroninja>
ah it seems like I have to patch my scripts so they can find there libraries
jackdk has quit [Ping timeout: 246 seconds]
Soo_Slow has joined #nixos
alp has quit [Ping timeout: 258 seconds]
<gyroninja>
*their
Synthetica has joined #nixos
shabius has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #62759 → icr: do not wrap crystal so we can use one from the environment → https://git.io/fjupG
<colemickens>
I use Wayland and GTK3 apps do the right thing, I don't think I have a use for xsettingsd personally.
<colemickens>
although it is interesting you mention this, since I'm dealing with a issue that specifically deals with how GTK3 apps behave differently with regards to the mouse cursor because there isn't an xsettings daemon running.... but you're not in #sway where I mentioned that, so o_0.
<azazel>
ah, i thought you were looking for a gsettings replacement when you haven't installed gnome
<{^_^}>
[nixpkgs] @teto pushed commit from @rvolosatovs to master « vimPlugins: Update »: https://git.io/fjupb
<azazel>
i halted my Wayland/sway adoption because of how XWayland deals with HiDPI, which has yet to be decent
<colemickens>
Yeah. I wish sway would work around it but I don't think they want ot invest around xwayland and I don't blame them, it seems majorly annoying
<jD91mZM2>
Let's say I first enable `system.autoUpgrade.enable` (and rebuild my system), and then further change my configuration.nix without rebuilding. Which version would get rebuilt?
goibhniu1 has joined #nixos
gratto has quit [Read error: Connection reset by peer]
xok has joined #nixos
<{^_^}>
[nixpkgs] @chreekat opened pull request #62761 → keychain: Add long description → https://git.io/fjuhe
gratto has joined #nixos
xok has quit [Client Quit]
zupo has joined #nixos
<lewo>
Someone created a repository to quickly setup a development env for Hydra. But I'm not able to find it back:( Do you know what is the name of this repository?
<jD91mZM2>
colemickens: Right, thanks a lot! :) To clarify, that's by default /etc/nixos/configuration.nix, and will therefore indeed rebuild the *latest* changes?
<azazel>
I would be much more prone to enable such functionality if nixos had something like a "security updtates only" channel
<azazel>
but maybe I'm too strict
<jD91mZM2>
azazel: Isn't that basically the stable channels? They don't get all the latest versions
nwspk has joined #nixos
<azazel>
jD91mZM2: usually they don't get new major releases advancement, but I don't know if there is some kind of an "agreed upon" (read "written down") policy about what goes on stable updates
<azazel>
jD91mZM2: I any case, kernel doesn't get upgraded until the next system reboot, so it will get its security fixes applied only then
<jD91mZM2>
azazel: I suppose you're right. I was about to just enable autoUpgrading for my nixops-managed server, but I just got tired of manually updating (or rather, forget to update) my normal system
<ashkitten>
i wonder if it'd be possible to patch the live kernel, like the service canonical provides for ubuntu lts
<ashkitten>
i'm not clear on the specifics of such an operation, though
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ashkitten>
azazel: so essentially, "it requires manual intervention at every step"
emily has joined #nixos
knupfer has joined #nixos
<ashkitten>
which is not good for nixos
erudition has joined #nixos
erudition has quit [Client Quit]
Shoubit has joined #nixos
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mighty_vee has joined #nixos
mighty_vee has quit [Max SendQ exceeded]
mighty_vee has joined #nixos
mighty_vee has quit [Max SendQ exceeded]
mighty_vee has joined #nixos
mighty_vee has quit [Max SendQ exceeded]
mighty_vee has joined #nixos
nschoe has joined #nixos
mighty_vee has quit [Max SendQ exceeded]
mighty_vee has joined #nixos
hmpffff has joined #nixos
orivej has quit [Ping timeout: 244 seconds]
<Taneb>
How can I diagnose my hydra server failing to connect to a builder with "Aborted: cannot connect to ‘nix@dandy.myrtle’: Host key verification failed."?
mighty_vee has quit [Max SendQ exceeded]
mariel__ has joined #nixos
mighty_vee has joined #nixos
hmpffff has quit [Client Quit]
mighty_vee has quit [Max SendQ exceeded]
mighty_vee has joined #nixos
mighty_vee has quit [Max SendQ exceeded]
mighty_vee has joined #nixos
jasongrossman has joined #nixos
botayhard has joined #nixos
Matthieu_ has quit [Ping timeout: 248 seconds]
adamantium has joined #nixos
m0rphism has joined #nixos
polman has quit [Ping timeout: 244 seconds]
<azazel>
Taneb: probably it's an ssh issue with the host keys? have you checked the ~/.ssh folder of the user running that process
<azazel>
~/.ssh/known_hosts specifically
polman has joined #nixos
<Taneb>
azazel: it looks reasonable and I can connect manually
<turion>
Jun 06 09:22:07 cassandra systemd[1]: cassandra.service: Main process exited, code=exited, status=126/n/a
<turion>
Jun 06 09:22:07 cassandra systemd[1]: cassandra.service: Failed with result 'exit-code'.
schjetne has joined #nixos
<turion>
Jun 06 09:22:07 cassandra cassandra[334]: /nix/store/97gl5mlbfj91n1b64bp5lngnp8cig13n-cassandra-3.11.3/bin/cassandra: line 5: /nix/store/97gl5mlbfj91n1b64bp5lngnp8cig13n-cassandra-3.11.3/bin/cassandra: Argument list too long
ThatDocsLady_ has joined #nixos
mariel__ has quit [Read error: Connection reset by peer]
Soo_Slow has quit [Remote host closed the connection]
Soo_Slow has joined #nixos
<gentauro>
is it possible to write "open firewall port 42" in a `default.nix` file and then `nix-shell` it and open the port for a specific task. Once leaving the `nix-shell` the port will then be closed again?
__monty__ has joined #nixos
kvda has joined #nixos
kranium has quit [Remote host closed the connection]
toppler`` has quit [Ping timeout: 258 seconds]
<jbgi>
Hi, ellipsis pattern (...) does not works when providing arguments to a nix-shell. Is that a known issue? Is there a workaround?
CcxWrk has joined #nixos
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
evhan has joined #nixos
Anton-Latukha has joined #nixos
schjetne has quit [Ping timeout: 244 seconds]
toppler`` has joined #nixos
<jbgi>
or rather it's the @-pattern that does not work in shell.nix :(
evhan has quit [Read error: Connection reset by peer]
evhan has joined #nixos
gratto has quit [Read error: Connection reset by peer]
Soo_Slow has quit [Remote host closed the connection]
<__monty__>
@-patterns should work anywhere in nix code.
gratto has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
otulp has quit [Ping timeout: 259 seconds]
Matthieu_ has quit [Ping timeout: 245 seconds]
slyfox has quit [Ping timeout: 252 seconds]
kvda has joined #nixos
<jbgi>
__monty__: so the complete issue is given a shell.nix strating with {a ? 1, ...}args:
<__monty__>
Where's the @?
<jbgi>
__monty__: sorry {a ? 1, ...}@args:
<jbgi>
I cannot access either args.a and args.b, even when running nix-shell --arg b 2 for exemple.
<__monty__>
How about if you flip the record and `args`?
<jbgi>
__monty__: same "error: attribute 'b' missing, ..."
<jbgi>
I can however access args.a if I provide --arg a <value>
<mk_modrzew>
what's the best medium to ask questions about issues with nixos? forum, reddit or else?
<__monty__>
jbgi: Does nix-shell take --arg?
ee1943 has joined #nixos
evhan has quit [Quit: De IRC non curat rex...]
<__monty__>
mk_modrzew: This channel's pretty helpful/responsive when the experienced people are around.
evhan has joined #nixos
<mk_modrzew>
__monty__: i have no doubt but i have issue with X11 and nvidia and it could be hard to reproduce problem and describe it in the same moment xd
lordcirth__ has quit [Read error: Connection reset by peer]
lordcirth has joined #nixos
<jasongrossman>
mk_modrzew: You could post on the forum (or reddit) and also announce your post here.
<mk_modrzew>
okay, thanks
<jbgi>
__monty__: it does work in other cases (eg. 'a' get the value 2 if I provide --arg a 2)
<__monty__>
jbgi: As I read it --arg only passes arguments to a function if it the function needs them. So, since there's no b in the pattern maybe it just doesn't get passed.
dahirsch has joined #nixos
dne has left #nixos ["WeeChat 2.4"]
lordcirth__ has joined #nixos
akaWolf has joined #nixos
<yorick>
gentauro: no, the firewall opening is a nixos thing
jbgi has quit [Ping timeout: 245 seconds]
<yorick>
gentauro: you could probably make it with a shellHook
mexisme_ has quit [Ping timeout: 258 seconds]
<freddie_freeload>
Hey! I am on unstable and use networkmanager. Does anyone know what the best practice is for connecting with eduroam? 🙂
<clacke_movim>
Or, of course, simply including the value in the expression directly, depending on circumstances.
<freddie_freeload>
<alexarice[m] "freddie_freeloader: Mine just wo"> Okay, good old heisenbug. I wanted to check the error again - but it just works now. Well.... it take it :D
<alexarice[m]>
I sometimes have a problem with dns not being set correctly after a rebuild, but reconnecting fixes that
<alexarice[m]>
though I have no idea if that is eduroam related or just general networkmanager being odd
<freddie_freeload>
alexarice: Sometimes it seems wifi-tech is even more complicated and unpredictable than quantum physics. ;) Thanks anyway 😁
johnny101 has quit [Remote host closed the connection]
johnny101 has joined #nixos
<{^_^}>
[nixpkgs] @edolstra pushed to master « wireguard: Don't fail if modprobe fails »: https://git.io/fjzvK
<risson>
Yes I know I shouldn't but the network I'm on does not let me check those certificates, it does it for me. However I trust that network so it's ok I guess
<risson>
Can I change the cache server to something in plain http ?
schjetne has quit [Ping timeout: 258 seconds]
Avaq has joined #nixos
clacke_movim has left #nixos [#nixos]
<ajs124>
Are you in a corporate network? Maybe you can add their root to your trust store.
dahirsch has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #62771 → mpd: 0.20.23 -> 0.21.10 + nixos module support for standard users → https://git.io/fjzJd
<{^_^}>
[nixpkgs] @abbradar pushed to master « nfs-utils: fix format strings in nfsdcld »: https://git.io/fjzU5
fendor has joined #nixos
drakonis_ has joined #nixos
drakonis has quit [Read error: Connection reset by peer]
<Avaq>
Hi! I need help with an issue previously discussed with clever and symphorien (on Feb 05 between 20:45 and 00:00 CET, if you want to search your logs) related to my attempts to use NixOps to manage my Raspberry Pi.
<Avaq>
I have not been able to resolve the issue that after copying closures successfully, I am presented with an error: "Exception: unable to activate new configuration (exit code 2)". Full error: https://gist.github.com/0889642443a4465c14b616f76d2672b4
<Avaq>
I don't know where to start debugging this issue. I really want to get acquainted with NixOps, and eventually introduce it at work for managing some operations.
<{^_^}>
#53672 (by eadwu, 21 weeks ago, closed): switch-to-configuration not interpreted using perl
<gchristensen>
thaht is very likely it, still, what version kernel is on it?
<{^_^}>
[nixpkgs] @danieldk opened pull request #62773 → nixos/tests/docker-tools: check layer sharing with buildLayeredImage → https://git.io/fjzTL
<Avaq>
gchristensen: Linux 4.14.9 (see also the Nix Info gist I posted) - or is that not the kernel version?
<gchristensen>
4.14.9? not 4.14.99?
alex`` has joined #nixos
<Avaq>
Uh, copy paste off error, sorry. It's Linux 4.14.97, as taken from the gist. :P
<gchristensen>
hrm.
<gchristensen>
you downgraded and then rebooted to the older version?
<Avaq>
Yeah, downgraded, rebooted various times, gave up and left it in cold storage for 2 months, and booted this morning. :P
<Avaq>
Does it matter at all what software I'm using on the command machine (not sure what to call it - the machine I'm ussuing the nixops commands *from*)
<gchristensen>
hah
civodul has joined #nixos
<{^_^}>
[nixpkgs] @c0bw3b merged pull request #62759 → icr: do not wrap crystal so we can use one from the environment → https://git.io/fjupG
<Henson>
Hi everyone. Could somebody point me in the direction of understanding how NixOS packages install stuff in the /etc directory? I've done a lot of work with Nix packages, but never in the context of Nix OS. I have a custom package that needs to install stuff in /etc/udev and to patch the /usr/share/misc/usb.ids file.
<nschoe>
(disclaimer: I'm quite the newbie in nixOS)
<Henson>
I need to make the transition from understanding how stuff is done with "nix-env -i" to how things are done with nixos-rebuild and /etc/nixos/configuration.nix and environment.systemPackages
orivej has quit [Ping timeout: 248 seconds]
<stepcut>
I am trying to run `nixops delete` on a virtualbox instance that I never actually spun up, but it is saying, "Exception: cannot delete this deployment because it still has resources". How can I figure out what 'resources' it is talking about? I tried `nixops info` but it does not list anything under 'Resource Id' and nixops check has Exists=no
anton has joined #nixos
<stepcut>
I could just run destroy -- but I'd kind of like to know what it will destroy
<gchristensen>
did you erase it from your network configuration?
<stepcut>
gchristensen: no? I have a logical specification, and two physical specifications -- one for ec2 and one for virtualbox. I run nixops create for the virtualbox instance, but never ran deploy. I might choose to create the virtualbox instance again later so I am not sure I want to delete anything from the .nix files
<stepcut>
s/I run nixops/I ran nixops/
fendor has quit [Ping timeout: 248 seconds]
Jackneill has quit [Ping timeout: 248 seconds]
<nh2>
stepcut: given that it doesn't contain any important data, you could run `destroy` and see what things it prompts you to destroy (usually it prompts you for that)
<stepcut>
ah
<nh2>
alternatively you could also inspect its full state by opening the nixops db file in e.g. sqlitebrowser
<nh2>
and see what resources it has in there
<nh2>
(of course that shouldn't be necessary but is an option available)
<aminechikhaoui>
nh2 nixops export should dump the same thing
<nh2>
ah yes also a good idea
<{^_^}>
[nixpkgs] @vcunat pushed 2 commits to staging-next: https://git.io/fjzTx
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #62752 → Fixes for x86_64-darwin failures on Hydra → https://git.io/fjuxG
<{^_^}>
[nixpkgs] @matthewbauer pushed 16 commits to master: https://git.io/fjzkU
<stepcut>
I ran export -- this is the resources section, "resources": { "webserver": { "index": "0", "type": "virtualbox" }
Jackneill has joined #nixos
<stepcut>
anyway, I just renamed the deployment which is also an acceptable solution to my problem
Anton-Latukha has quit [Quit: Leaving.]
o1lo01ol_ has joined #nixos
Matthieu_ has quit [Read error: Connection reset by peer]
wzrd- has joined #nixos
o1lo01ol1o has quit [Ping timeout: 272 seconds]
orivej has joined #nixos
ottidmes has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ottidmes>
Is it possible to overwrite a mkOption definition? I guess I could resort to disabledModules, but is there a way for a single option definition as well?
<aminechikhaoui>
stepcut what was the state of the webserver in the export output
fendor has joined #nixos
<aminechikhaoui>
in theory when the state is "missing" which means destroyed, you won't even see the resource in the xport
<aminechikhaoui>
export *
<stepcut>
aminechikhaoui: what site do people use for pastes these days? The ones I used to use are offline
<azazel>
Henson: the difference is that the same packages you install using nix-env -i are installed into the system and configured using the configuration modules found in nixpkgs/nixos/modules
<stepcut>
I think it is the result of running nixops create but not nixops deploy
Avaq has quit [Ping timeout: 248 seconds]
<aminechikhaoui>
I think in any case you could run nixops delete -f
<stepcut>
yeah
jmeredith has joined #nixos
astrofog has joined #nixos
<Henson>
azazel: so is it the things in nixpkgs/nixos/modules that integrate regular nix packages into NixOS's system directories?
toppler`` has joined #nixos
Avaq has joined #nixos
rprije has quit [Remote host closed the connection]
rprije has joined #nixos
<Henson>
azazel: also, how do the configuration options in /etc/nixos/configuration.nix (for example the Apache configuration options) get transformed into a configuration file the Apache server is expecting to read?
<azazel>
Henson: yes, and the primary interface to them is to compile your desired configuration in /etc/nixos/configuration.nix which is the "main" module
<Henson>
azazel: so the files in nixpkgs/nixos/modules, are those just a static list of files, or can packages you install also put stuff in that directory to then be included in the new system? Another way is: can an individual package carry the instructions of how it should be integrated in to the system, or are all the integration instructions separate from the packages?
<azazel>
Henson: configuration is nothing more that another packages (or group of packages) containing configuration files built from derivations contained into the configuration modules in nixos/modules
<Henson>
aanderse: thank you for that link
<aanderse>
Henson: no problem. you can create your own modules and then include them into configuration.nix if you wish
ilmu has quit [Quit: statebox]
sigmundv has joined #nixos
ilmu has joined #nixos
<azazel>
Henson: usually configuration modules for services like apache produce the necessary configuration files in one package and augment the configuration of another service like systemd whose role is to manage the services
Soo_Slow has quit [Remote host closed the connection]
zupo has joined #nixos
Soo_Slow has joined #nixos
<Henson>
azazel, aanderse: so I still don't understand yet if the instructions needed to adjust things in the /etc (and other system) directories are present in the one package you install, or if the developer also has to make an adjustment in another package (for example creating a new derivation that contains the nixpkgs/nixos/modules files).
<pie__>
Henson, generally modules are written independently of a package because most people dont use nix
<pie__>
you _can_ include an external file into the configuration
<Henson>
azazel, aanderse: I would expect that for packages that need to adjust the system a new output could be create d(outputs like "out", "doc", "lib", etc) that would allow the NixOS configuration instructions to be composed with instructions from other packages.
<pie__>
generally system configuration is desired to be separated into a module, and package building is in a file in the rest of the tree
<Henson>
pie__: ok
<pie__>
and the module somehow refers to the package
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<azazel>
Henson: packages are just added to the nix store, it's the role of the configuration modules to hook them up
<Henson>
azazel: ok
<aanderse>
Henson: yes, simple applications which don't run as a service or anything *often* require no module
emptyflask has joined #nixos
reinhardt has quit [Quit: Leaving]
inkOne has joined #nixos
<pie__>
Henson, modules can also end up creating derivations, but thats just the normal mechanism, what actually ends up doing stuff in the rest of the system are i think "activation scripts" and some options that result in creating files
<Henson>
azazel: so what is the name of the derivation that contains the nixos/modules files? I'll need to create my own derivation of these files in order to incorporate my own configuration instructions.
<pie__>
so you could have something like the.file.option."/etc/wat" = generateConfigForThing {}
<pie__>
* not what it actually looks like
<Henson>
pie__: ok
ddellacosta has joined #nixos
<pie__>
a derivation is (usually) the result, not something containing the nix files
<Henson>
if I trying doing something like "nix-store -q -d ./2cai0ankbxbdhg0mvjdvr370gnfm2za9-nixos-19.03.172801.1a27474d474/nixos/pkgs/misc/vim-plugins/vim2nix/autoload/nix.vim" it says "unknown-deriver"
ottidmes has quit [Quit: WeeChat 2.4]
<pie__>
if you want to add your own module you either use an overlay (the point is it lets you add external stuff to the attrsets) or you modify a git checkout
<{^_^}>
[nixpkgs] @orivej-nixos pushed 2 commits to master: https://git.io/fjzIz
<pie__>
Henson, i dont know my way around all that stuff but you might try nix show-derivation
<Henson>
pie__: I'll probably use an overlay, but I need to know the name of the nix package that contains the nixos/modules files so I can edit it.
<pie__>
ah
Anton-Latukha has joined #nixos
Tucky has quit [Quit: WeeChat 2.4]
<pie__>
i havent tried figuring out yet if there's a way to do that. the usual fallbacks are grepping and https://nixos.org/nixos/options.html (i think theres a man page with the contents of this)
asheshambasta_m has joined #nixos
sigmundv has quit [Ping timeout: 272 seconds]
<Henson>
pie__: ok, I'll check that out
<Henson>
azazel, aanderse, pie__: thank you all for your help
<azazel>
Henson: why you want to modify nixos' modules? You can override it
<aanderse>
Henson: np
<Henson>
azazel: yes, that was my plan, but I need to know the name of the thing to override.
astrofog has quit [Remote host closed the connection]
<Henson>
azazel: presumably nix-build uses the files in the modules directory, and those files were put there by some nix package. It is that package I need to adjust in order to make my own files appear in the modules directory.
nschoe has quit [Ping timeout: 272 seconds]
hyper_ch2 has quit [Ping timeout: 256 seconds]
<azazel>
Henson: Instead of try to reason on stuff that you are just now trying to understand, why don't you describe what you want to achieve?
<pie__>
Henson, modules are not put there by packages
<Henson>
azazel: say I have a package that needs to be run at startup. In addition to the package binary files, I also have to make some changes to the files in /etc to make it started by systemd (or whatever init system is being used). How do I go about making those changes in NixOS?
<clever>
Henson: environment.etc and systemd.services, in your nixos module
<aanderse>
Henson: yes, as clever said. you would create 2 separate files. a package under the pkgs/ directory, and then a module under nixos/modules directory
<pie__>
also modules dont need to be in the nixpkgs tree, you can just include it in configuration.nix
<azazel>
Henson: man configuration.nix is your friend ;-)
asheshambasta_m has quit [Ping timeout: 272 seconds]
sigmundv has joined #nixos
<Henson>
ok, so do I need to create a new file and put it in the modules directory? If so, is there a package I need to edit or override to make that happen? Since /nix/store is mounted read-only I can't just copy my new module in there, which is contrary to Nix philosophy anyway. So I need to figure out how to get my own module file into the modules directory. Also (as pie__ hinted at) is it even...
<Henson>
necessary for me to add any new files?
<Henson>
I apologize if this is going on for a long time. It seems like I'm circling the understanding I need here, and it's just taking time for me to understand how it fits together.
<clever>
Henson: if you want to test a nixos module, put it in /etc/nixos/ and use imports = [ ./thatfile.nix ];
<Henson>
clever: I see. So I can include my own out-of-tree modules in, say, /root/custom.modules, and then patch those into the the configuration.nix file where they'll be included in whatever nix-build does?
silver has joined #nixos
<Henson>
clever: instead of "patch" I should have said "import"
mmlb has joined #nixos
<aanderse>
Henson: the module will take effect when you run nixos-rebuild test/switch
domogled has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
alex`` has quit [Ping timeout: 248 seconds]
<Henson>
azazel, aanderse, pie__, clever: thanks everybody for your help! I think I understand it now
<aanderse>
Henson: great! let us know if you have any issues when you get into it :)
<Henson>
aanderse: I'm sure I'll experience issues, so I'll be back. Just hopefully not too soon :-)
alex`` has joined #nixos
endformationage has joined #nixos
<{^_^}>
[nixpkgs] @vbgl opened pull request #62776 → [coq] simple-io: 1.0.0 -> 1.2.0; QuickChick: init at 1.1.0 for Coq 8.9 → https://git.io/fjzLl
stepcut has quit [Ping timeout: 248 seconds]
<pie__>
Henson, to be fair a lot of my answers werent able to directly answer the question you were circling around ;p
<pie__>
Henson, please take notes on points of difficulty while youre a beginner
<pie__>
by the time one gets around to trying to improve things theyve gotten better and it becomes hard to remember what the problems were
<Henson>
pie__: no worries, we got there in the end. It's difficult to get an answer to a question you don't really know how to ask :-)
<sphalerite>
Does anyone here use GNOME and have screen locking working right? If so, how?
<Shoubit>
sphalerite, what's wrong? Mine's working as it were on Ubuntu, haven't noticed anything wrong.
<sphalerite>
Shoubit: oh, haha, didn't realise gdm is still a thing :p will try that
fendor has quit [Read error: Connection reset by peer]
nschoe has joined #nixos
drakonis_ has quit [Ping timeout: 257 seconds]
<sphalerite>
Shoubit: yep, switching to gdm did the trick. Thanks!
npmccallum has joined #nixos
emptyflask has quit [Ping timeout: 246 seconds]
<Shoubit>
No worries, glad I could help!
pacman99 has quit [Quit: Konversation terminated!]
Matthieu_ has joined #nixos
dahirsch has quit [Remote host closed the connection]
kim0 has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @thefloweringash opened pull request #62777 → mariadb: fix clang and darwin build → https://git.io/fjzLx
ottidmes has joined #nixos
o1lo01ol_ has quit [Ping timeout: 248 seconds]
<zfnmxt>
I did nix-collect-garbage -d; nixos-rebuild switch and the nix-build command has been running for over 8 hours now and still isn't done...any idea if something's gone awry?
<zfnmxt>
I passed it the -v flag and the last output was `evaluating file '/nix/var/nix/profiles/per-user/root/channels/nixos/nixos/lib/from-env.nix'`
zupo has joined #nixos
<clever>
zfnmxt: is it building something or just chewing cpu?
<ottidmes>
If I have a bunch of paths outside the Nix store and I want to create a directory containing those paths in the Nix store, what is the best way to achieve this? In my particular case I could probably get away with using pkgs.linkFarm, but if I did not wanted to have a symlink per path, what would be my options?
<zfnmxt>
clever: How can I know?
<clever>
zfnmxt: if outputs things when building
<zfnmxt>
It's been stuck on the output above for 99.9% of the 8 hours.
<clever>
try restarting it?
<zfnmxt>
Will do
<zfnmxt>
Huh, that's weird. I still have all my old boot entries.
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schjetne has joined #nixos
<zfnmxt>
Do you have to manually remove boot entries? I thought `nix-collect-garbage -d` handled that.
<clever>
zfnmxt: nixos-rebuild will update the bootloader config
<zfnmxt>
Ah, right. Okay, I'll try the rebuild again.
wfranzini has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #62777 → mariadb: fix clang and darwin build → https://git.io/fjzLx
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fjztr
<wucke13_>
I have a rust bild which is ran in the nix-shell. When I try to compile statically, I get an error while linking, = note: impure path `/.../target/release/build/arrayvec-04d508924dca0069/build_script_build-04d508924dca0069' used in link. Is this caused by nix-shell setting some settings for the linker?
<Izorkin>
etu: Please check this moment php options --with-apxs2 - automatic enable options --enable-maintainer-zts and build php with zts?
<etu>
Izorkin: Do you want to enable more options by default?
<Izorkin>
etu: No, change ztsSupport ? config.php.zts or false to ztsSupport ? (config.php.zts or false) || (apxs2Support)
<Izorkin>
* ztsSupport ? (config.php.zts or true) || (apxs2Support)
<Avaq>
I've run into the problem that my disk space is full of /nix/store, and nix-collect-garbage doesn't manage to complete because of "error: committing transaction: database or disk is full".
<goibhniu1>
zfnmxt: is there high load on the machine when you run nixos-rebuild?
<etu>
Izorkin: Isn't that a yes to my question? :p
<zfnmxt>
goibhniu1: Well, one CPU thread goes to 100% the entire time.
<{^_^}>
[nixpkgs] @c0bw3b pushed commit from @r-ryantm to master « star: 2.7.0f -> 2.7.1a (#62577) »: https://git.io/fjztQ
<goibhniu1>
zfnmxt: can you see what it's doing e.g. with strace?
peacememories has joined #nixos
<Izorkin>
Correction parameter zts, it automatically turns on. It is to be seen that php is going to support zts. ./result/bin/php -v - PHP 7.2.19 (cli) (built: May 29 2019 07:14:43) ( ZTS )
<Avaq>
I actually don't see a clear way out. There is nothing on my system (yet) that is safe to remove, besides the nix garbage. I can't manually remove it because the fs is locked, and nix-collect-garbage apparently needs to write to a database file before it can start actually deleting things, which of course it can't because there's no space.
<sphalerite>
Avaq: you can remount the store read-write (`sudo mount -o remount,rw /nix`) then delete stuff (only /nix/store/.trash is safe I believe) manually
<sphalerite>
err, /nix/store I think
<{^_^}>
[nixpkgs] @grahamc opened pull request #62779 → Restart declarative containers when their host environment configuration changes → https://git.io/fjzth
<clever>
Avaq: nix-collect-garbage --max-freed 1, should delete .trash for you
<clever>
ahh, but its failing due to the disk being too full
<Avaq>
I think I managed by using `nix-store --gc`, which was mentioned in the nix-collect-garbage man page.
<clever>
Avaq: is /nix sharing the FS with something else?
<Avaq>
Ah, no -- it just ook longer to show me the same error. >.<
<{^_^}>
[nixpkgs] @lheckemann pushed 2 commits to release-19.03: https://git.io/fjzqO
<sphalerite>
neva: then yeah, boot it with copytoram, then repartition the stick and install as you would on a regular internal storage medium
arjen-jonathan has quit [Ping timeout: 272 seconds]
<zfnmxt>
goibhniu1: Sorry for the delay, had to read a bit about strace. Here's the trace, not sure if it's helpful at all: http://ix.io/1L55
<neva>
the regular way being just running the ISO?
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.1]
Glider_IRC_ has quit [Ping timeout: 258 seconds]
<Izorkin>
etu: Another option is to build a separate package for apache-httpd, without embed systemd cgi fpm. How in php-unit.
<goibhniu1>
zfnmxt: I barely know how to run it myself :D ... is that what you see after you let `nixos-rebuild switch` run until it appears to get stuck?
<zfnmxt>
I mean it gets stuck after 3 seconds or so.
<zfnmxt>
That's just `strace nixos-rebuild boot`
<zfnmxt>
Well, `strace -o trace.txt nixos-rebuild boot` to be specific
civodul has quit [Quit: ERC (IRC client for Emacs 26.2)]
<sphalerite>
neva: yes, as you already are AFAIU
<zfnmxt>
Do you think trying a different channel might help? Maybe I should just try 19.03?
<zfnmxt>
I'm a big nixos noob if it wasn't apparent so sorta floundering :D
<goibhniu1>
yeah sorry, I have no clue what that could be, trying a different channel sounds like a good idea
<zfnmxt>
No worries. Worst case I'll just re-install. Pretty painless with nix :)
<zfnmxt>
Thanks for your efforts, though.
neva has quit [Ping timeout: 256 seconds]
<goibhniu1>
zfnmxt: cool, BTW how much RAM have you got?
<zfnmxt>
16GB and it isn't filling up.
<goibhniu1>
great, and did you make any changes to your configuration.nix before running nixos-rebuild?
drakonis_ has joined #nixos
<zfnmxt>
Nope
<zfnmxt>
What's the difference between the nixos and nixpkg channels...by the way
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
Shoubit has quit [Quit: Leaving]
<gchristensen>
marek: it appears that people are taking the fun in that update script too seriously :P
schjetne has joined #nixos
<acowley>
Is it possible to make an installation image based on a running NixOS system?
<tilpner>
acowley: Can you specify what you mean by that?
<tilpner>
acowley: It could mean any of the following: 1. Build an ISO that behaves like a config you have. 2. Build a disk image that behaves like a config you have. 3. Build an ISO/disk image that supports installing a system that behaves like a config you have.
<acowley>
My original intention was to investigate backup / system restore options for NixOS, but then I thought a useful start would be an installation disk that included all the things I have installed on my system rather than a minimal store
<acowley>
I'm interested in all three :D
drakonis_ has quit [Ping timeout: 252 seconds]
<tilpner>
You can do all three, but in all cases you'll work off of your configuration files
<tilpner>
There is no feasible way to probe an existing system and synthesize a configuration that would build to the same
<acowley>
I was surprised there wasn't more about this in the manual. It only has a brief chapter on creating a stock installation image
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tilpner>
Hold on
zeta_0 has joined #nixos
dontobey has quit [Ping timeout: 252 seconds]
<acowley>
What I'll probably attempt first is migrating to a new drive, which I guess I can mostly do with just dd or the like. But then I thought it possible that NixOS would have some kind of media creation gizmo that could copy the whole system closure, so I could even get a new system up and running more quickly than following a from-scratch NixOS installation guide (e.g. Chris Martin's).
<tilpner>
And then you import that however you wish
<genesis>
with my local branches, need to look
zeta_0 has joined #nixos
<genesis>
i hope wxpython4 issue will be fix soon.
<tilpner>
genesis: You can try git worktrees
<genesis>
imagine i need to package beremiz, i need 5 unpackaged dependancies (matiec, canfestival, wxglade and so)
<genesis>
since reviewer prefers 1 commit by package, not easy to not lose time, even if you need to copy paste some derivation (for wxpython40 for exemple) from PR
<tilpner>
I don't package things in nixpkgs, I always use overlays first
knupfer has quit [Ping timeout: 248 seconds]
Matthieu_ has quit [Read error: Connection reset by peer]
<tilpner>
And then when it works, I integrate it into nixpkgs. Which is not that great, but it means I don't have that problem
<johnw>
tilpner: same here
<genesis>
temp1: +98.0°C (crit = +100.0°C) :/
<genesis>
i'll crash again.
<Izorkin>
etu: checked?
<tilpner>
(Although too often I just keep them in overlays now, I don't always have the motivation to polish and nag people about my PR)
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
alex`` has joined #nixos
ddellacosta has quit [Ping timeout: 248 seconds]
ddellacosta has joined #nixos
<genesis>
tilpner : a bit shame you don't contribute but i understand it's time consuming
<genesis>
sometimes i think some reviews are a bit too much
toppler`` has quit [Ping timeout: 248 seconds]
peacememories has joined #nixos
<genesis>
and it can be discouraging to ask too much time when anyway, a derivation could be improve time to time
<genesis>
s/to/after
copumpkin has quit [Read error: Connection reset by peer]
<joepie91[m]>
oh wow lol
<sphalerite>
Does anybody have a favourite SIP client that they would recommend to me? :)
<catern>
matthewbauer: btw I just noticed in your static Nix script your shebang is /bin/sh but you're doing "exec -a" - that's actually not valid POSIX sh, and /bin/sh on Debian at least (which is dash) rejects it
toppler`` has joined #nixos
rprije has quit [Remote host closed the connection]
<catern>
matthewbauer: unfortunately I don't know of a way to do exec -a in a POSIX-portable way
copumpkin has joined #nixos
<catern>
so you might want to explicitly use /usr/bin/env bash
rprije has joined #nixos
gml8 has joined #nixos
<{^_^}>
[nixpkgs] @abbradar pushed to master « nfs-utils: use upstream patch to fix i686 build »: https://git.io/fjzOe
<tilpner>
genesis: I don't mind time spent writing, discussing, implementing requested changes. I mind PRs becoming stale and having to be kept updated, and having to bug people to merge things. I'm not assigning blame to anyway, I'm part of the problem by not reviewing other PRs
<genesis>
true :)
<Henson>
I'm having a wierd problem trying to get a custom package's udev rules included in nixos-rebuild. I've specified the package path in /nix/store as one of the entries in services.udev.packages in configuration.nix. I see that the nixos/modules/services/hardware/udev.nix file loops over things in "$i/{etc,lib}/udev/rules.d/*" to copy into /etc. I've even edited this to print out the contents...
<tilpner>
genesis: I like going "Oh, I could improve this" or "Hey, I have this package, want to have it?", but that enthusiasm doesn't last the month it takes for things to get merged
<genesis>
because we don't have infinite time, and our process should understand that
<Henson>
of the directories and files it's processing. However it appears as though it's unable to see the contents of my derivation in /nix/store, even though it does exist. There's nothing different between my path in /nix/store and another one which it is able to process happily.
<Henson>
does anybody have any ideas?
<tilpner>
*blame to anyone
hmpffff has quit [Quit: nchrrrr…]
goibhniu1 has quit [Ping timeout: 252 seconds]
<tilpner>
genesis: On the other hand I don't want PRs to be merged review-less either
asheshambasta_m has quit [Quit: Leaving.]
orivej has quit [Ping timeout: 258 seconds]
rprije has quit [Read error: Connection reset by peer]
Copenhagen_Bram has quit [Read error: Connection reset by peer]
<o1lo01ol1o>
I need a user to be able to access a secret set via `deployment.keys.hqr-secret.user = "norooty"` and friends. However, the user doesn't have execute on /run/keys and so can't reach the key. How can I grant execute on the path leading to the key?
<slabity>
Typically you would use Nix to build another Nix
<hexo>
make cannot load docbook schema
<hexo>
i'm puzzled
<hexo>
:(
lockshaw_ has left #nixos ["WeeChat 2.4"]
<hexo>
i'm on powerpc64le
<hexo>
and on ubuntu
<hexo>
so... :/
Ariakenom has joined #nixos
<catern>
what kind of store url should I use for a read-only store in my local filesystem?
Copenhagen_Bram has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace pushed 8 commits to release-19.03: https://git.io/fjzs4
callahad6 is now known as callahad
<catern>
also, what's the most efficient way to get Nix to set up the namespaces to use a user-local store and run a binary inside them?
<catern>
going through "nix run nixpkgs.somepkg -c somepkg" requires evaluating Nix expressions
gml8 has joined #nixos
linarcx has quit [Quit: WeeChat 2.4]
sir_guy_carleton has quit [Quit: WeeChat 2.4]
gml8 has quit [Client Quit]
tmaekawa has joined #nixos
schjetne has quit [Ping timeout: 248 seconds]
__monty__ has quit [Quit: leaving]
<{^_^}>
[nixpkgs] @Izorkin opened pull request #62783 → php: update build config and php package → https://git.io/fjzsy
<catern>
matthewbauer: btw, maybe you've run into this; I'm trying to use nix-env with your static Nix, and I'm getting an error about not being able to find /nix/store/*-musl/bin/busybox ?
<catern>
oh, now that I think about it - does nix-env not have the auto-namespacing magic? is it only the "nix" command that has auto-namespacing?
mexisme_ has joined #nixos
talqu has joined #nixos
<noonien>
hello folks
<noonien>
how can i use patchelf on a cross-compiled binary?
neva has quit [Ping timeout: 256 seconds]
<catern>
also, does "nix build" have the ability to output the path of the result on stdout? or do I have to use nix-build if I want that?
shibboleth has joined #nixos
<manveru>
catern: no, it can't do that
ym555 has quit [Quit: leaving...]
<manveru>
really wish there was `nix build` that only outputs the path and `nix build-result` that only makes the symlink... or something like that :)
<manveru>
but atm that's `nix-build --no-out-link` and `nix build`
wfranzini has joined #nixos
<azazel>
nix build has --no-link
<manveru>
yeah, and then you don't know anything :P
tilpner has quit [Quit: WeeChat 2.4]
<gchristensen>
you know it built or not ... :P
<manveru>
:D
tilpner has joined #nixos
<noonien>
is there a way to get an unwrapped gcc for cross-compilation? this is what i've tried so far: http://ix.io/1L6F
<noonien>
(sorry, the second nativeBuildInputs should also be commented)
adamantium has quit [Remote host closed the connection]
talqu has quit [Remote host closed the connection]
drakonis has quit [Ping timeout: 248 seconds]
adamantium has joined #nixos
schjetne has joined #nixos
alex`` has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @jonringer opened pull request #62787 → vimPlugins.semshi: init at 2018-12-05 → https://git.io/fjzG7
<NemesisD>
is there a good way to protect my nix-based project from being garbage collected?
drakonis_ has quit [Read error: Connection reset by peer]
<o1lo01ol1o>
after much fiddling, i've finally got hydra-queue-runner able to ssh to my build machine (local .ssh/config for the user, dedicated user ssh key; home directories and shell access on the build machine). However, now: my hydra jobset fails with sh: nix-store: command not found
drakonis_ has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjzG5
<o1lo01ol1o>
Do I need to set some user rights or something?
<{^_^}>
[nixpkgs] @veprbl merged pull request #57947 → mkl: fix install_name on Darwin → https://git.io/fjfWN
<symphorien>
o1lo01ol1o: make sure ssh sources /etc/profile
<o1lo01ol1o>
symphorien: How can I do that?
<dhess>
simon_weber: I believe domenkozar was working on that for SnabbLabs a few years ago, by getting machines to update from a channel. Not sure he ever got it working, but there is a partial implementation in https://github.com/snabblab/snabblab-nixos/tree/customchannel
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-19.03: https://git.io/fjzGA
<symphorien>
Run as hydra : ssh foo nix-store --version
<o1lo01ol1o>
$ ssh foo nix-store --version ssh: Could not resolve hostname foo: Name or service not known
drakonis has joined #nixos
daGrevis has quit [Quit: daGrevis]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
troydm has joined #nixos
<o1lo01ol1o>
if you mean: ssh user@slave then the output is: nix-store (Nix) 2.2.2
daGrevis has joined #nixos
inquisitiv3 has joined #nixos
<symphorien>
Did you try in an interactive shell ?
<o1lo01ol1o>
that was an interactive shell
<o1lo01ol1o>
hydra doesn't have ssh apparenetly
Henson has joined #nixos
<symphorien>
Then try in a non interactive shell
<o1lo01ol1o>
hydra-queue-runner does
<symphorien>
ssh user@slave nix-store --version
<symphorien>
You can also try nix ping-store
<o1lo01ol1o>
right, doing that as hydra-queue-runner yields: nix-store (Nix) 2.2.2
drakonis_ has quit [Ping timeout: 272 seconds]
<symphorien>
Ah ok then this was a wrong hypothesis
<inquisitiv3>
I'm planning to install NixOS on ZFS. Do the ZFS partition need to be no. 1 as described in the wiki article? https://nixos.wiki/wiki/NixOS_on_ZFS
growpotkin has joined #nixos
AstroBadger has joined #nixos
drakonis1 has joined #nixos
<growpotkin>
We bumped into something a bit unexpected about Python packages. I wanted to see if it sounded odd to anyone else. When we list python modules in our build dependencies the order that the modules are listed will lead to wildly different derivations. It seems like modules that are listed first get to set version resolution priority for all later packages.
adamantium has quit [Ping timeout: 258 seconds]
peacememories has joined #nixos
Matthieu_ has quit [Ping timeout: 252 seconds]
schjetne has quit [Ping timeout: 248 seconds]
vidbina has quit [Ping timeout: 248 seconds]
ambro718 has quit [Ping timeout: 258 seconds]
growpotkin has quit [Remote host closed the connection]
<Henson>
If I want to specify an entry in services.udev.packages in /etc/nixos/configuration.nix, is it possible to specify something in /nix/store? I'm trying to deploy a custom package on a remote computer and it isn't present in nixpkgs, but all the contents are in /nix/store. If I specify the nix store path it doesn't work (the paths of the package aren't even present during nixos-rebuild. But...
<Henson>
if I specify the path as something like pkgs.nethack, then the package paths are present).
<Henson>
is there a way I can make this work without the package in question being in nixpkgs?
rprije has quit [Remote host closed the connection]
Anton-Latukha has quit [Quit: Leaving.]
<{^_^}>
[nixpkgs] @fpletz pushed 6 commits to release-19.03: https://git.io/fjzZ8
<Baughn>
Phew. Took me long enough, but I think I have a draft module for zrepl.
<Baughn>
Henson: It'll work with any package, not just the ones in github.com/nixos/nixpkgs
<Baughn>
Henson: So you should make your own. That can be fairly trivial, depending on what exactly you're trying to do.
npmccallum has quit [Quit: npmccallum]
Matthieu_ has joined #nixos
jackdk has joined #nixos
<catern>
Henson: how do expect the store path to get deployed to the remote computer?
o1lo01ol1o has quit [Remote host closed the connection]
<catern>
Henson: (btw, the direct answer to your question is "specify the path as something like pkgs.nethack" but I'm not sure why you think that isn't suitable)
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 248 seconds]
botayhard_ has joined #nixos
botayhard has quit [Ping timeout: 258 seconds]
thc202 has quit [Ping timeout: 248 seconds]
Havvy has quit [Read error: Connection reset by peer]
<Henson>
Baughn: if I specify pkgs.nethack, then it works. But if I specify the path of the .drv file or one of the outputs, then it doesn't work.
<Baughn>
Henson: Don't do that. Specify the derivation itself.
<Baughn>
The language construct, not the .drv file.
Havvy has joined #nixos
<Baughn>
pkgs.nethack evaluates to a mkDerivation call. You want something similar.
<Henson>
Baughn: ahh, ok
<Baughn>
.drv files are implementation details. I've never needed to touch one.
<Henson>
Baughn, catern: the issue is that my package is going to be deployed to a remote computer that is not online. I plan on copying the /nix/store paths directly to these remote machines through an intermediary that is sometimes connected to the machine, and sometimes to the internet