<__red__>
So my host is on nixos-unstable (which I think is master, right)?
<__red__>
Can you specify a different channel for your VM images generated by NixOps
<__red__>
seems like an obvious requirement for testing
ykaelig has quit [Ping timeout: 260 seconds]
fragamus has joined #nixos
<Lisanna>
I'm starting to suspect that I might be trying to do something impossible
<fpletz>
Lisanna: weird, that should work fine… I've wrote such services a lot
IrcL has joined #nixos
<Lisanna>
huh, ok
<bsima>
Is it acceptable to submit an expression generated by cabal2nix to the nixpkgs repo?
<__red__>
fpletz: I just pulled master nixops and am running it within a nix-shell
<Lisanna>
bsima the real question: will it be acceptable with 1.12 to submit code to nixpkgs that calls nix-build as part of the derivation
IrcL has quit [Read error: Connection reset by peer]
<__red__>
it's definately getting further
<Lisanna>
(and things like cabal2nix as well)
<fpletz>
__red__: oh, yeah, the dev-shell, right? was going to suggest that
<__red__>
yes
<fpletz>
was just looking into getting nixopsUnstable on master bumped
<fpletz>
so that works?
<__red__>
It;s getting there ... still building I blieve
<__red__>
believe
<__red__>
"filling Nix store..."
<fpletz>
ok, I'll prepare the update anyway :)
<infinisil>
Lisanna: I doubt it
<infinisil>
Lisanna: well it depends, if you mean import-from-derivation, that certainly won't get into nixpkgs. However if recursive nix were to be implemented, that might
<__red__>
deployment finished successfully :-D
<infinisil>
okay well maybe IFD would too, but most likely not
yann-kaelig has quit []
erasmas has quit [Quit: leaving]
binh_ has quit [Read error: Connection reset by peer]
randomstrangerb has quit [Ping timeout: 264 seconds]
randomstrangerb has joined #nixos
aswanson has joined #nixos
<aswanson>
hi all, what's the 'nix' method for editing `GRUB_CMDLINE_LINUX` parameters? I need to add something there to hopefully fix some problems with my graphics chip
astrofog has quit [Quit: Quite]
<joepie91>
initial attempt at running NixOS on OpenVZ: unpacking the containerTarball and symlinking /sbin/init does not work
<joepie91>
lol
<joepie91>
and with 'does not work', I mean 'totally breaks the control panel, no longer pings, console broken'
Lisanna has quit [Remote host closed the connection]
<chreekat[m]>
I just upgraded from 17.03 to 17.09, and now the KDE launcher (?) crashes every time I do an application search. If I search through the alt+space drop-down enough times, the whole windowing environment crashes, leaving me with a cursor on a black screen. If I search through the "start" menu, the menu dies and then comes right back. Any clues?
alexteves has quit [Read error: Connection reset by peer]
nkpart has joined #nixos
andersk has quit [Remote host closed the connection]
andersk has joined #nixos
magnetop` has quit [Remote host closed the connection]
magnetop` has joined #nixos
magnetop` is now known as magnetophon
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Shados opened pull request #34435: dante service: change to network-online.target dep (master...dante-service-improvement) https://git.io/vNQjh
<aswanson>
Rusty1: I've been poring over those to see if anything fits, closest I've seen is `grub.extraconfig` but the stackoverflow post I'm trying to work from references /etc/default/grub. I'm not sure if those two options end up in the same place
obadz has quit [Quit: WeeChat 1.9.1]
marusich has quit [Ping timeout: 252 seconds]
nuncanada has joined #nixos
thc202 has quit [Ping timeout: 248 seconds]
marusich has joined #nixos
aswanson has quit [Quit: WeeChat 2.0]
ryanartecona has quit [Quit: ryanartecona]
mkoenig has quit [Ping timeout: 240 seconds]
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mkoenig has joined #nixos
silver has quit [Read error: Connection reset by peer]
Supersonic112 has quit [Disconnected by services]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
randomstrangerb has quit [Ping timeout: 248 seconds]
<__red__>
Is there a way to create a directory or a file from inside my configuration.nix in nixos?
<__red__>
I have a strange situation where I set my documentRoot in my httpd service, but of course I need something to create the directory
<__red__>
and my certificate stores for crypto
<__red__>
shared between apps
<vaibhavsagar>
does anyone have an eGPU setup under NixOS?
<__red__>
also, things like log directories need to be created
<__red__>
/var/log/exim/domain.name/ for example
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 closed pull request #34400: lib: Better use the module type system in platform parsing (master...meta-check) https://git.io/vN9Td
NixOS_GitHub has left #nixos [#nixos]
<clever>
__red__: i always point the document root to a derivation
tmplt has quit [Read error: Connection reset by peer]
tmplt has joined #nixos
stepcut has quit [Remote host closed the connection]
nkpart has quit [Quit: Connection closed for inactivity]
<fearlessKim[m]>
I would like to support this printer on nixos, http://www.openprinting.org/printer/Fuji+Xerox/Fuji_Xerox-DocuPrint_CM305_df . I had a look at nixpkgs, I dont think it exists but I worry cups has some bundles I missed ? not sure how to package a printer driver, is there some doc or I just mimic other derivations ?
<sahabi>
hey, I'm having trouble using the right click in the shell.
<sahabi>
it doesn't work, I'm trying to copy paste into the command line and vim
<rnhmjoj[m]>
which terminal emulator?
<sahabi>
how can I find out?
<sahabi>
im also using xmonad
<sahabi>
but right clicking works in the browser
jb55 has quit [Ping timeout: 240 seconds]
<rnhmjoj[m]>
maybe the window title or the about menu
<sahabi>
I did echo $TERM
<sahabi>
i got screen
<rnhmjoj[m]>
did you not install it? in that case it's probably xterm
jb55 has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sahabi>
hmm
<sahabi>
xterm yes
<clever>
sahabi: hold control and then hold left, right, or middle click
<clever>
it will activate whatever option your pointing at when you release the mouse
<fearlessKim[m]>
rnhmjoj: thanks, thats what I had found, the dirver names are a bit cryptic
<sahabi>
clever: thanks, i tried that and it's really helpfull. it doesn't have paste though
<clever>
sahabi: middle click is paste
<clever>
and it uses the selection buffer, not the clipboard, so it will paste whatever you have selected
verite has quit [Ping timeout: 240 seconds]
<bhipple[m]>
I see a handful of blog posts showing how to just use NIX_PATH and a git clone entirely instead of channels. I'm setting up Nix for a bunch of people at work and I'm considering just going down that route and dumping channels entirely to keep things simpler and reduce the number of things to teach. Aside from having something that updates consistently with the binary cache, is there any real need for channels or are they
<bhipple[m]>
mostly a relic?
<sahabi>
ohh it works without ctrl
steshaw[m] has joined #nixos
<clever>
sahabi: yeah, control is the hotkey for the menu, normal middle click is paste
<sahabi>
thanks clever!
<clever>
left and right adjust the selection
hamishmack has quit [Quit: hamishmack]
<clever>
middle click also acts as paste in nearly every linux application
hamishmack has joined #nixos
<sahabi>
ok. clever: the middle click with ctrl doesnt remain because of my type of mouse
<sahabi>
it toggles quickly
<clever>
ah
<rnhmjoj[m]>
doesn't middle click use a different buffer than clipboard though?
<clever>
rnhmjoj[m]: yeah, the selection buffer
hamishmack has quit [Client Quit]
jb55 has quit [Quit: WeeChat 1.9]
pkill9 has quit [Ping timeout: 248 seconds]
schoppenhauer has quit [Ping timeout: 248 seconds]
magnetophon has quit [Remote host closed the connection]
orivej has joined #nixos
<infinisil>
Lisanna: the way mkIf works is that it just tells the module system to evaluate this thing later, but not bother about it currently, which is the reason it can evaluate modules at all (bc lazyness and fixed points). So when you use a function like trace*, it will try to force evaluation of cfg.enable right now (not lazily), which it can't when it's strict
<Lisanna>
infinisil okay, yeah I've been reading a bit more about the trickiness of writing NixOS modules
<infinisil>
have spent quiet a bit of time with this trickiness
<Lisanna>
okay, so I can get infinite recursion encountered if I accidentally put something in my config = ...; which forces the evaluation of cfg to happen right now
<Lisanna>
feels like a minefield ): how can I tell if doing one thing or another will force the evaluation to happen now, other than tracing?
<coconnor>
oof. I agree: a minefield!
<infinisil>
Lisanna: gut feeling :P
<coconnor>
even with tracing determining the source of infinite recursion is difficult
<Lisanna>
bleh
<coconnor>
in my experience: by making small changes and testing repeatedly you can at least know the source of infinite recursion quickly
<coconnor>
in cases like integration, tho, things definitely get tricky
cyraxjoe has joined #nixos
<infinisil>
Lisanna: i think a good rule is: if your config depends something from the config itself, you need to wrap it in a mkIf or such
<Lisanna>
infinisil but isn't that trivially true?
<infinisil>
what do you mean?
<Lisanna>
what good is having options if you don't use them in your resulting config = ...;?
<Lisanna>
like, what if I wanted to write the value of cfg.enable to a file in /etc with environment.etc? Would I need to use mkIf there?
<Lisanna>
or are things which are only needed at build time, i.e., don't affect the control flow of the nix expression, okay?
Itkovian has joined #nixos
<coconnor>
hmmm nix os is a fixed point of the module expressions. "build time" is a bit vague haha. Tho the later statement "doesn't effect control flow" is correct.
ThatDocsLady has joined #nixos
<Lisanna>
can the name of an entry in config = ...; be dependent on the value of one of your module's options?
<srhb>
Sure. Use antiquotation.
<Lisanna>
x_x
<Lisanna>
wut
<srhb>
${name}
<Lisanna>
systemd.services."${cfg.name}" = ...; is ok?
<Lisanna>
okay, what about making the number of systemd.services in a config dependent on the number of entries in a list option?
<srhb>
Just map over that list
<srhb>
map mkService services
<Lisanna>
and // that to config?
<srhb>
Where mkService is a function from the list elements to a systemd service.
<srhb>
Lisanna: Not necessarily. You might use one of the mkMergers or simply have systemd.services = ( ... expression that produces the final set of your services)
<tilpner>
hyper_ch2 - Sources can come in zips too :)
asuryawanshi has joined #nixos
<hyper_ch2>
tilpner: :)
<roundhouse>
fetchzip instead of fetchurl?
<tilpner>
roundhouse - Yes
<hyper_ch2>
damn, I get the unicorn on github :(
<roundhouse>
thanks
<roundhouse>
hyper_ch2, thanks for the link, I was unaware that there is a guide for this
roundhouse has left #nixos [#nixos]
<hyper_ch2>
roundhouse: this is for packaged binaries..... it's usually easier to compile from source and it's recommended to do so
roundhouse has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vbgl opened pull request #34445: mpd_clientlib, ncmpc: fix on Darwin (master...mpd_clientlib-darwin) https://git.io/vN7nc
NixOS_GitHub has left #nixos [#nixos]
rardiol1 has quit [Ping timeout: 240 seconds]
alunduil has quit [Ping timeout: 256 seconds]
alunduil has joined #nixos
leat has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nlewo opened pull request #34446: skopeo: remove the patch by setting a variable at build time (master...pr/skopeo) https://git.io/vN7ni
NixOS_GitHub has left #nixos [#nixos]
<roundhouse>
when trying to pack a prebuild binary I get the error "./result/bin/leela_0110_linux_x64: relocation error: ./result/bin/leela_0110_linux_x64: symbol , version GLIBC_2.2.5 not defined in file libc.so.6 with link time reference" when running the build binary. Anyone know why that is?\
<roundhouse>
a search on github reveals it could be related to misconfigured LD_LIBRARY paths, but mine seem to be fine
Enzime has quit [Ping timeout: 260 seconds]
<roundhouse>
the linked glib version is 2.26 btw
Ke has joined #nixos
<roundhouse>
it seems a similar error appeared on ubuntu after upgrading, so I guess it is the wrong version of glibc. How can I change the one used in a nix expression?
<Ke>
there is sourceByRexex but not sourceByNotRegex, even though I am not aware of a easy way to create complement regex or am I just stupid again?
nliadm has joined #nixos
Enzime has joined #nixos
rauno has joined #nixos
<Ke>
I would think people would want the complement even more often
fragamus has joined #nixos
zzamboni has quit [Ping timeout: 276 seconds]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zzamboni has joined #nixos
roundhouse has quit [Remote host closed the connection]
sud0x3 has joined #nixos
asuryawanshi has quit [Remote host closed the connection]
nschoe has quit [Ping timeout: 240 seconds]
fyuuri has joined #nixos
zzamboni has quit [Remote host closed the connection]
zzamboni1 has joined #nixos
lexappreciate has joined #nixos
<lexappreciate>
heyyyyyyyyyyyyyyyyy
<lexappreciate>
hey hey hey
knupfer has joined #nixos
goibhniu has joined #nixos
<lexappreciate>
how should I define a function that accepts list of arguments? [x y z]: x + y + z ? (that cant be evaluated)
thblt has joined #nixos
periklis has joined #nixos
dan_b has joined #nixos
<lexappreciate>
here are contents of my module3.nix : [x y z]: x + y + z , trying to import it , but nix-instantiate compains error: syntax error, unexpected ':', expecting $end
<LnL>
not sure if that's useful enough to add to lib
<lexappreciate>
maveru, that's mind blowing (not sure if that right word in context)
* lexappreciate
never expected to behave it so
<lexappreciate>
on plus2 = plus 2 I expect EXCEPTION NOT ENOUGH ARGUMETNS
<lexappreciate>
however... wait...
asuryawanshi has quit [Remote host closed the connection]
<lexappreciate>
yes
vidbina has joined #nixos
<lexappreciate>
i guess language should say when argument is optional, and when it is required
asuryawanshi has joined #nixos
<manveru>
i don't think there are optional arguments, but you can have default values for set arguments
ssmike has joined #nixos
<puffnfresh>
I just set up a server with NixOS
<puffnfresh>
some of my terminfo files have a "~nix~case~hack~1" suffix
<puffnfresh>
though my filesystem is not case insensitive, it's just standard ext4
<puffnfresh>
looking at the code, it looks like macOS uses the case hack
<lexappreciate>
I also discovered yesterday that language doesn't have static typing. I hear everyone around how they like statically typed languages for perfomance and bugfree reasons. I myself prefer optional typing like MYPY in python
a6a3uh has joined #nixos
<manveru>
lexappreciate: yeah, that's my biggest gripe with nix :(
<lexappreciate>
maveru you want optional type system?
<joko>
Hey, any way to override a NixOS service module?
<manveru>
lexappreciate: would be nice, proper static typing with inference would be even nicer :)
<manveru>
but i think it's slow as hell and doesn't support all of the nix syntax
<bennofs[m]>
Can I run a rustc build on ofborg? Or would that be considered a "massive" build?
<bennofs[m]>
x86_64 only would be ok
<lexappreciate>
I would like see translation of python to nix, or entire nix ecosystem re implemented in python (probably with pypy, or static typing + hot spot compilation cython like) .
<lexappreciate>
the reason I appreciate about python is core selling point - READABILITY and ease of understand.
<LnL>
bennofs[m]: I think the build timeouts are not long enough for that
<LnL>
bennofs[m]: I can build it, do you have a link?
<FRidh>
lexappreciate: if you want a package manager in Python, checkout spack
<lexappreciate>
this slide is easer to understand and read https://youtu.be/79dL7HgmW_k?t=18m20s . (and I want typechecker or language extension like that, easy to read for mortals)
<FRidh>
In our environment the following is set for Citrix `LD_PRELOAD=/usr/$LIB/libctxXrandrhook.so`. Unfortunately, this results in an error when running any Nix binary:
<FRidh>
ERROR: ld.so: object '/usr/$LIB/libctxXrandrhook.so' from LD_PRELOAD cannot be preloaded (wrong ELF class: ELFCLASS32): ignored.
<Dezgeg>
libctxXrandrhook.so sounds like a 32-bit binary, nothing you can do
<FRidh>
indeed 32 bit
<Dezgeg>
well I guess you could run/use 32-binaries in nix as well
<FRidh>
right
<Dezgeg>
in case unsetting LD_PRELOAD isn't an option
<FRidh>
unsetting is not really an option.
<FRidh>
we could use 32-bit builds otherwise
<bennofs[m]>
LnL: thanks, already started the build on my laptop, I hope it's not going to take tooo long (llvm is done) :) In general though it would be nice to have something like https://github.com/nix-community/aarch64-build-box for x86_64 is as well
<FRidh>
applications work fine regardless the error, so I ignoring it could be a solution as well
<Dezgeg>
yes, it's just telling that it ignored the preload
<LnL>
bennofs[m]: rust takes about ~1h30 IIRC
<LnL>
wish I could get rid of the extra llvm build
swayf has quit [Remote host closed the connection]
<lexappreciate>
thanks FRidh, looks nice. however wheels\pyinstaller could do the same, aren't they? I like whole set of features of nixOS - like generations, versioning, builds are done in isolation like pure functions, configuration as a code, dependency based on hash, active community, desktop friendliness, and so on. I doubt that we will see any reliable alternative in nearby future. (spack and ansible probably don't have all of that)
<bennofs[m]>
i agree, vendoring large dependencies is something I don't like but it seems that there is no pratical other way as lots of projects do it (and not vendoring it often leads to subtle, hard to debug issues later)
<LnL>
well we can still pull it out into a separate drv
LTannin has joined #nixos
LTannin has left #nixos [#nixos]
swayf has quit [Ping timeout: 265 seconds]
dan_b has quit [Ping timeout: 264 seconds]
zzamboni has joined #nixos
zzamboni1 has quit [Remote host closed the connection]
<Shados>
Is there a simple way in Nix to generate a list of strings of all single English letters, both upper- and lower-case?
<manveru>
lexappreciate: you can check out guix then?
<Ke>
I am trying to hack xfstests in nixpkgs by modifying code in many different stages, but $src does not seem to be writable anywhere
<Ke>
I don't see dontMakeSourcesWritable anywhere and I even tried to set it to 0
<puffnfresh>
Shados: yes, nixpkgs has lowerChars and upperChars
<LnL>
ah same case then, did you override the unpackPhase?
<Ke>
no
<LnL>
hmm
<LnL>
it's just a regular stdenv.mkDerivation?
<lexappreciate>
manveru dude, I checked but it is not clear what advantages it have over NixOS. afaik it's only fork. and lisp != /* is not */ python. I tried some dialect of lisp - HYLANG, no advantages over python, absolutely. even buggy XONSH have syntax adnavtages and doing better.
<Ke>
LnL: pkgs.xfstests.overrideAttrs where pkgs.xfstests is nixos package
<lexappreciate>
the reasoning for programmer, and people in general, they like to stick with their strong current languages, and do not like to switch to another. thats why we end up in the world when people re inventing wheels.
<manveru>
lexappreciate: well, there aren't a ton of other options around, that's why i suggested it
<tazjin>
lexappreciate: guix is "better" in some ways, in that they have a cleaner, more organised standard library and they actually have macros that properly assign names to entities
<tazjin>
however for actual OS usage it's a toy because it doesn't have systemd
<lexappreciate>
and I want to have my nvidia steam gaming out of the box!
<tazjin>
plus Scheme is a language that is ah bit, eh, easier to grasp maybe than Nix
<LnL>
Ke: weird, do you know at what point it's failing because the source is readonly?
<lexappreciate>
I sold my ass and soul for my 1070 GPU
fyuuri has quit [Ping timeout: 256 seconds]
<Ke>
LnL: I can't find any place where it's writable, from my gentoo/portage knowledge I tried postPatch and postUnpack and now preBuild
<manveru>
i'd prefer lisp/scheme to nix... but the issue is in the ecosystem, like with many other languages
a6a3uh has quit [Quit: Bye]
<manveru>
there's just nothing like nixpkgs for any other language than nix :)
<Ke>
LnL: postPatch does not seem to be run at all
<LnL>
the patchPhase is overwridden so the pre/post hooks won't work
<manveru>
lexappreciate: but what you could do is try binding to nixUnstable using swig or something... and call that from python
<LnL>
but as far as I can tell the sources should be writable after postUnpack
<Ke>
LnL: ie. if you override a stage post/pre gets ignored always?
randomstrangerb has quit [Ping timeout: 260 seconds]
randomstrangerb has joined #nixos
oida has joined #nixos
zzamboni has quit [Quit: Leaving.]
zzamboni has joined #nixos
<Ke>
LnL: my $sources tell me that I should use $PWD instead of $src
<Ke>
it seems to fail less
<Ke>
but I have to go eat now
<LnL>
yeah, that's why we generally discourage overriding an entire phase
<LnL>
that expression also breaks patches = [];
<LnL>
don't see any reason why that couldn't use the prePatch hook
eqyiel has quit [Remote host closed the connection]
<Ke>
thanks, anyway, I'll be back later
eqyiel has joined #nixos
vidbina has joined #nixos
Mateon3 has joined #nixos
orivej has quit [Ping timeout: 248 seconds]
Mateon1 has quit [Ping timeout: 240 seconds]
Mateon3 is now known as Mateon1
Myrl-saki has joined #nixos
<rauno>
Any idea if docker can be configured in nixos with custom registry ?
S0rin has quit [Quit: WeeChat 1.9.1]
S0rin has joined #nixos
TweyII has joined #nixos
TweyII is now known as Guest27207
<ij>
Can I use the new nix interface already (1.12, was it)? Is it not stable enough?
<infinisil>
ij: you totally can, just put `nix.package = pkgs.nixUnstable` in your configuration.nix
<makefu>
ij: nixUnstable is already very stable
hotfuzz has joined #nixos
hotfuzz_ has quit [Ping timeout: 246 seconds]
<ij>
Cool! Nr #2: Should I switch to system profile in root user's account? I would expect nix-env --list-generations to list system ones.
<gleber_>
I have a bizzare case. In a nixos configuration if I use `{ users.extraUsers = let ... in { ... }; users.extraUsers.foo.openssh.authorizedKeys.keys = fooKeys; }` I get an `attribute ‘users.extraUsers.foo.openssh.authorizedKeys.keys’ at ... already defined at ...`. If I move the let-defined stuff out into a top-level let it works. Any ideas what's happening? O_O
<rauno>
or another way would be to define in use home .docker but does configuration.nix allow creating such file automatically ?
<makefu>
gleber_: do you have the full snippet?
<gleber_>
no, let me try to create one
<infinisil>
gleber_: yeah you can't define the same attribute twice in a nix expression
<gleber_>
infinisil: but I do this all the time (or at least something seemingly the same) in my NixOS configuration
<infinisil>
the one you defined twice here is users.extraUsers (users.extraUsers = { ... } and users.extraUsers = { foo.openssh.... })
<infinisil>
gleber_: it's a semantic error in nix, you can't be using this since it's invalid nix
<infinisil>
you'll have to use mkMerge or a separate file if you want to do what you intend
<gleber_>
infinisil: let me create a snippet to show why I am surprised
<infinisil>
e.g. users.extraUsers = mkMerge [ (let ... in { ... }) { foo.openssh.authorizedKeys... } ];`
mint_ has joined #nixos
mint_ is now known as zol
zol is now known as zolen
zolen has quit [Client Quit]
a6a3uh has joined #nixos
<Ke>
LnL: in portage the convention would be to call the base stage from the custom stage explicitly
<gleber_>
I use `let-bizzare-good.nix` style code in my NixOS configuration all the time with various different fields / keys
<gleber_>
And yet if I put `let ... in { ... }` into `users.extraUsers` things stop working. Is there some magic happening at attrset evaluation time which fakes mutability of it to some extent?..
<infinisil>
gleber_: i see
<infinisil>
the reason is because `foo.bar.qux = 1` is just syntactic sugar for `foo = { bar = { qux = 1; }; };`
<Mic92>
gchristensen: was it you, who tested nvme on amazon aws?
rauno has quit [Remote host closed the connection]
<gleber_>
infinisil: but this means that `{ foo.bar.qux = 1; foo.xyz.zbc = 2; }` is also somehow magically merged into one statement, otherwise it actually does redefine top-level `foo` twice
<infinisil>
gleber_: in the top one, you're setting users.extraUsers to a let in statement, which it can't desugar
<infinisil>
gleber_: hmm, well it can merge attributes apparently
<gleber_>
s/one statement/one expression/
<gleber_>
Looks like it
<LnL>
Ke: yeah, that way overrideDerivation generally works as expected
<infinisil>
{ a.b = 1; a = { c = 2; }; } doesn't work
vidbina has quit [Remote host closed the connection]
<infinisil>
{ a.b = 1; a.c = 2; } does though
<infinisil>
so it seems to be syntactic sugar for the dot notation, all same attributes get put into 1
<infinisil>
I haven't noticed this myself yet, learned this myself just now :)
<therp>
hi gleber_
alasi has quit [Quit: Connection closed for inactivity]
jtojnar has quit [Read error: Connection reset by peer]
swayf has quit [Remote host closed the connection]
swayf has joined #nixos
<gleber_>
ij: it all boils down to `--add-root some/path --indirect`. It adds a gcroot for this shell. IIRC it also works with a simple nix-shell command
<ij>
So it adds a gc-root for every dependency?
<sphalerite>
ij: no, it adds a GC root for the shell derivation which transitively depends on all the deps
<sphalerite>
So they don't get gc'd because there's still something depending on them
vidbina has quit [Read error: Connection reset by peer]
grenade is now known as grenade|afk
<steveeJ>
any hints why pylint is not able to find the 'text' module? I've installed python with this package: (python35Full.withPackages (ps: with ps; [ autopep8 pylint pep8 ]))
magnetophon has quit [Remote host closed the connection]
Isorkin has joined #nixos
goibhniu has quit [Remote host closed the connection]
<gchristensen>
Mic92: wasn't me
goibhniu has joined #nixos
<mred>
I am new to NixOps, but I am a little bit worried about what the worst thing is that NixOps could do to the existing running services. Will it delete anything it doesn't know about for example?
<mred>
It also appears that many of the services are written with the idea that once a service is deployed that its resources never change afterwards, which in many cases is not a realistic assumption.
<mred>
In the case of Kubernetes, servers might come and go for example (didn't check the module), but the same holds for HAProxy.
<mred>
To my understanding most companies are creating cloud based infrastructure, but the people writing NixOS modules seem to be creating modules for the 1990s era, not 2015+.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] Profpatsch opened pull request #1827: Add docstrings to Derivation/SimpleDerivation WIP (1.11-maintenance...drv-docs-1.11) https://git.io/vN71W
NixOS_GitHub has left #nixos [#nixos]
magnetophon has joined #nixos
<sphalerite>
mred: it will switch to a fresh system configuration. That doesn't mean removing anything inherently, but if local stuff made and activated system configuration changes then yes that will be overridden
<infinisil>
mred: services should *never* delete any data themselves, and if you find such a module you can shout in this channel. Yes services don't handle changing data well, because nixos doesn't really have anything to handle state in an idiomatic way.
iqubic` has joined #nixos
<sphalerite>
mred: as for state management (which AFAIU is what your second question is about), that's essentially outside the scope of nixops
orivej has joined #nixos
<infinisil>
nixos needs a clear distinction of data that changes often, and configuration that changes seldomly (the former doesn't go in configuration.nix, the latter does)
<mred>
sphalerite: so, how should one manage a server that has multiple backend servers that can change a couple of times per week?
iqubic has quit [Ping timeout: 252 seconds]
<mred>
sphalerite: my first concern was mostly about whether NixOps would ever touch existing machines (I don't want to do anything bad to those, of course).
<mred>
I would also like to know if it is possible to use existing AWS instances for a NixOps deployment.
<mred>
I have the impression that the only operation there is, is the "create or update a NixOS machine which must have been created by NixOps".
<pmeunier>
Hi! Is there any way to provide a "dummy source" for a derivation? Some of my derivations are just tricks that don't have any source.
<sphalerite>
Yes, all nixops needs is root SSH access and for the target system to be nixos
<mred>
sphalerite: ok, and how do I teach nixops about such an existing system?
<sphalerite>
But note that it does make the assumption that it is solely responsible for the nixos config on the target machine so any configuration stuff already existing on it will be overridden
<sphalerite>
I don't actually know how 😅 I think there's a "none" back end or something
<sphalerite>
I haven't used nixops myself
<sphalerite>
As for managing a server with multiple back end servers changing on a regular basis: point it to them through DNS or something, unless you're managing them through nixops as well in which case you just configure them appropriately
<infinisil>
pmeunier: use runCommand, or writeText or something like that
<pmeunier>
infinisil: does that mean I can't make derivations?
<pmeunier>
for that purpose, I mean
<sphalerite>
As I said I haven't actually used nixops properly myself so I don't know the details, just the general mode of operation
rauno has joined #nixos
<infinisil>
pmeunier: you can, you could just set `unpackPhase = ":"` which i think will work without a src
<sphalerite>
pmeunier: unpackPhase = ":";
<sphalerite>
Yep
<pmeunier>
Another question is: how can I get gcc as "cc" on OSX? Whenever I add it to my buildInputs, cc is still clang.
<pmeunier>
infinisil, sphalerite: thanks!
<infinisil>
pmeunier: what's your usecase?
<infinisil>
i mean for the derivation
<pmeunier>
I want to compile a program that doesn't like being compiled with clang
<mred>
sphalerite: when machines can autolaunch, that doesn't work.
michiel_l has quit [Remote host closed the connection]
<pmeunier>
oh sorry, for the derivation, I just wish I could alias things sometimes.
<gchristensen>
mred: nixops isn't the tool for that then
<mred>
sphalerite: the ip address of the machines that have just lauched would have to be injected into the configuration file of the server.
<sphalerite>
mred: right so you're managing them outside nixops. In that case I'd just use DNS as I said
<gchristensen>
mred: I run dynamic services with NixOS without NixOps, using pre-generated AMIs.
<mred>
gchristensen: given a configuration.nix, how do I build an AMI from that?
<gchristensen>
an AMI doesn't have custom userdata, strictly speaking. what are you intending?
<mred>
gchristensen: on auto launch, I want to inject the IP address of the service (as well as the instance id) into the configuration of the high availability server.
<gchristensen>
you mean the one you just launched?
<mred>
I meant server, not service.
<mred>
Yes
<gchristensen>
why? :P the metadata API exists for good reasons
<mred>
gchristensen: you mean that I should poll the metadata API to discover the newly started server?
tertleeltret has quit [Quit: Connection closed for inactivity]
<mred>
The server itself only takes IP addresses, so the run-time injection of these IP addresses still needs to write to the server's configuration file and should then reload the server.
<gchristensen>
sorry, I guess my clarifying question wasn't clarifying enough. I thought you wanted to inject the instance ID / IP in to the server you just started. anyway, I'd use DNS for that instead and not deal with IPs wherever possible. Then, you can have instances add themselves to DNS on startup
<gchristensen>
but, this is fairly of topic for #nixos
<mred>
gchristensen: the server doesn't support resolving host names to ip addresses according to the man page.
<gchristensen>
then it'll need some extra help
<mred>
gchristensen: but yes, that would have been a good solution if that were possible (modifying the server is an option, but not realistic in this case).
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
xcmw has joined #nixos
Isorkin has joined #nixos
jtojnar has joined #nixos
waleee has joined #nixos
iyzsong has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
stphrolland has joined #nixos
xcmw has joined #nixos
Neo-- has joined #nixos
stepcut has joined #nixos
tmplt_ has quit [Ping timeout: 240 seconds]
asuryawanshi has quit [Ping timeout: 260 seconds]
FRidh has joined #nixos
coot has quit [Quit: coot]
__Sander__ has joined #nixos
xcmw has quit [Ping timeout: 248 seconds]
Rusty1 has joined #nixos
alexteves_ has quit [Ping timeout: 260 seconds]
ma27 has joined #nixos
ma27 has quit [Client Quit]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 3 new commits to master: https://git.io/vN75P
<NixOS_GitHub>
nixpkgs/master 6aa13b6 Tim Steinbach: linux: 4.9.78 -> 4.9.79
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 27f5262 Tim Steinbach: linux: 4.14.15 -> 4.14.16
<NixOS_GitHub>
nixpkgs/master 26e06c9 Tim Steinbach: linux: 4.4.113 -> 4.4.114
<bennofs[m]>
gchristensen: would it be possible to have something like the aarch64 build box for x86_64 as well? it would allow me to run some large builds while at university without draining my laptop battery so much 😅 not sure how quickly that would become overloaded though...
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.09: https://git.io/vN75M
<NixOS_GitHub>
nixpkgs/release-17.09 5cbd68b Tim Steinbach: linux: 4.4.113 -> 4.4.114...
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/release-17.09 6062ec0 Tim Steinbach: linux: 4.9.78 -> 4.9.79...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.09: https://git.io/vN75y
<gchristensen>
bennofs[m]: hmm I can look in to it, but it would be harder to get. we get the aarch64 community box because decently powerful aarch64 hw is hard to come by
<gchristensen>
so that box's bill is pretty much paid for by ARM itself
ma27 has joined #nixos
ma27 has quit [Client Quit]
ma27 has joined #nixos
ma27 has quit [Client Quit]
freeman42x[NixOS has quit [Ping timeout: 248 seconds]
<bennofs[m]>
is someone in charge of updating rustc in nixpkgs?
<bennofs[m]>
is a rustc update from 1.22.1 to 1.23 safe or is there a high chance things would break?
<gchristensen>
I know andi- and LnL hack on these things
<andi->
no idea, haven't read the changelog for 1.23 (yet)
<infinisil>
bennofs[m]: just see who did the last couple updates
<LnL>
bennofs[m]: firefox is one of the things to check, the rest is probably fine
<bennofs[m]>
andi-: I think updating to 1.23 should fix the jobserver "failed to acquire token" error that happens during large builds
<gchristensen>
omg finally
tmplt_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 5 new commits to master: https://git.io/vN7db
<lexappreciate>
heyyyyyyy, so I'm using nix-shell newenv.nix and inside it pkgs.buildFHSUserEnv with number of packages, how I can enforce usage of prebuilt binaries? cuz by default my nix-shell wasting lot time building them from source
coot has joined #nixos
<wacke>
hmmmm
<lexappreciate>
maybe there is some ENVIRONMENT VARIABLE like NIX_PLZ_USE_BINARIES ?
<lexappreciate>
i mean shell var
<manveru>
lexappreciate: you should probably make sure it's using a channel that has the packages built
spear2 has joined #nixos
spear2 has quit [Read error: Connection timed out]
{^_^} has quit [Remote host closed the connection]
{^_^} has joined #nixos
alexteves_ has quit [Ping timeout: 248 seconds]
FRidh has quit [Remote host closed the connection]
<fearlessKim[m]>
maybe 'build id' only exists in packages built by hydra ? which could explain why I don't have it
MercurialAlchemi has quit [Ping timeout: 264 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] ayyjayess opened pull request #34458: sacad: init at 2.0.6 (master...sacad) https://git.io/vN7hF
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 4 new commits to master: https://git.io/vN7hb
<NixOS_GitHub>
nixpkgs/master 94b6357 Peter Simons: hackage-packages.nix: automatic Haskell package set update...
<NixOS_GitHub>
nixpkgs/master 866a1ed Peter Simons: cabal2nix: build with hpack 0.23
<NixOS_GitHub>
nixpkgs/master b063a02 Peter Simons: LTS Haskell 10.4
NixOS_GitHub has left #nixos [#nixos]
<domenkozar>
aha!
<domenkozar>
good I'm not crazy
erasmas has joined #nixos
zzamboni has quit [Quit: Leaving.]
stepcut has quit [Remote host closed the connection]
wacke has quit [Quit: leaving]
<lexappreciate>
eureka, they usually say in the bath
<domenkozar>
gchristensen: fpletz: when did this start to happen?
<fpletz>
domenkozar: my estimate would be about 2 weeks ago
<gchristensen>
well for me it started this one time I plugged in my headphones and they just stopped working ... then for some reason they started again, then I tried to get fancy and add a pulseaudio bogus output which converted everything to mono, and that didn't work and nothing ever worked again
<gchristensen>
but I haven't rebooted ... so maybe it would work? about a week ago
<goibhniu>
does it work without pulseaudio?
<domenkozar>
even if I rollback to december and reboot, no sound
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/vN5vi
NixOS_GitHub has left #nixos [#nixos]
<domenkozar>
fearlessKim[m]: startup of linux?
<fearlessKim[m]>
domenkozar: I usually plug during kernel startup then once I reach my WM, I realize music is off and unplug/replug headsets and it fixes it
stepcut has quit [Remote host closed the connection]
<fearlessKim[m]>
even if I unmute prior to replug, doesn't seem to have effect. haven't looked into it.
leat has quit [Ping timeout: 240 seconds]
<domenkozar>
gchristensen: fpletz what kernel? 4.9.66 here
<domenkozar>
ah you said no reboot.
<fpletz>
domenkozar: I'm using 4.14.15
<fpletz>
also after reboots
stepcut has joined #nixos
a6a3uh has joined #nixos
stepcut has quit [Remote host closed the connection]
<ottidmes>
I am trying to get SSH to work within initrd, but I am getting a no such device errors, I tried with eth0 and with enp0s31f6, no success. I tried those with the ip kernel parameter and also just by calling ip link set <name> up in preLVMCommands. I had it working once a different machine, but I lost that config and I cannot seem to remember what I needed to do to make it work. Any suggestions?
raph_ael has joined #nixos
<pie_>
dtz, where can i find docs for that
macslayer has joined #nixos
<macslayer>
Good afternoon, all!
<raph_ael>
hello, is it possible to run nixos without another init than systemd ?
<dtz>
pie_: I think 'env' is sometimes used (not recently?) to be the nix-shell target? Seems only used with "buildEnv".. dunno :)
* dtz
gets some more coffee
<pie_>
i guesss that might be it
<pie_>
well... i want to be able to do nix-shell shell.nix
<pie_>
and get the stuff i need availible...
<pie_>
trying to set up a project template for my labs so i have LaTeX, R, python, and some other stuff
<pie_>
first time im trying to properly write a nix script
<raph_ael>
ottidmes: I read the suckless nixos church, but we'll probably never see if it's heading anywhere
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] avnik opened pull request #34461: [WIP] Proof of concept: Provide default libGL aliases (master...feature/default-libGL) https://git.io/vN5kU
NixOS_GitHub has left #nixos [#nixos]
<avn>
raph_ael: it because systemd haters won't write any code, they want hate
<gchristensen>
avn: let's not carry that hate in to #nixos :)
<raph_ael>
avn: you're calling for troll like a werewolf howling in the nigh
<raph_ael>
t
<ottidmes>
raph_ael: I only really delved deeper into Linux after Arch Linux adopted systemd, so the only init system I am really used to is systemd, so I am quite happy with using systemd (might be a case of not knowing better, i.e. ignorance is bliss)
tmplt_ has quit [Ping timeout: 268 seconds]
MichaelRaskin has joined #nixos
<hyper_ch2>
sys-v-init :)
<avn>
gchristensen: oh... sheeeit... sorry ;)
<hyper_ch2>
it worked for decades :)
mlin has joined #nixos
<avn>
better, let discuss PR which I made few secs ago ;)
<gchristensen>
great idea!
<raph_ael>
ottidmes: well I don't like the design of systemd, and almost each time I tried I got problems, so I prefer using something else :)
<mlin>
Hi all, does anyone know how to install polybar with i3 and volume support?
<hyper_ch2>
I do have to say systemd has some really nice stuff
<ottidmes>
mlin: What? Weren't you using bspwm?
<raph_ael>
avn: also Voidlinux went from systemd to runit, so your claim was a bit short :)
<tray>
Hello, what this error is nix-env -qa (error: cannot auto-call a function that has an argument without a default value (‘lib’))
Tucky has quit [Quit: WeeChat 2.0.1]
hyper_ch2 has quit [Quit: Page closed]
<pie_>
would this be correct?: buildInputs = [(with pkgs; somepackage anotherpackage)] , or what would be the correct way to do this?
<ottidmes>
mlin: I must have confused you with someone else. I have used i3 with polybar for a short amount of time, I believe volume just worked, what is it you have trouble with?
vidbina has joined #nixos
<mlin>
ottidmes: I see, I think that there is a certain package required to be present in order for it to build support for it
<mlin>
what audio related packages do you have?
<avn>
raph_ael: I am about concrete nixos ;)
<ottidmes>
mlin: Nothing special, only enabled pulseaudio in nixpkgs and nixos
<raph_ael>
avn: i don't see what you mean, for me concrete is for building, but I'm not a native english speaker
tray has quit [Ping timeout: 260 seconds]
<ottidmes>
raph_ael: Concrete as in, something actualized, not just something in theory.
<avn>
raph_ael: I am upvote for init system agnostic layer in nixos, and would like to join effort. But I am not motivated to do it myself.
<avn>
libGL hell is much more interesting problem for me ;)
<raph_ael>
ottidmes: I see :)
<raph_ael>
avn: I understand, well for now I have something of an answer, thanks all :)
<MichaelRaskin>
raph_ael: If you try, note that you already can convert (most of) NixOS services into runner scripts by NixOS built-in functionality.
<ottidmes>
mlin: I just checked, I had volume in my polybar config
<raph_ael>
MichaelRaskin: I haven't used nix for the moment, but I'll have a look thanks
<MichaelRaskin>
I am currently writing code that uses NixOS code as a black box, and uses both service script export and grabbing /etc/ entries
lexappreciate has quit [Ping timeout: 260 seconds]
stephenjudkins has joined #nixos
<ij>
Nope, I don't think so.
<pie_>
davidar, hey, i see you made the initial commit for cmdstan, do you know anything about stan for the other lanuage frontends? though ill probably just use cmdstan
<pie_>
davidak[m], do you maybe know how to work it with R?
iyzsong has quit [Ping timeout: 248 seconds]
bpa has joined #nixos
<ottidmes>
I am getting the message "2 blocks" after rebuilding my system, but each time I do, I get the message one more time, so after 4 rebuilds I am seeing 4x "2 blocks" messages being printed, any idea what that might be?
<raph_ael>
MichaelRaskin: thanks again
<ottidmes>
I solved my initrd SSH problem BTW, I just had to lookup my ethernet controller driver with `lspci -v` and then add it to `boot.initrd.kernelModules`
Arcaelyx has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
coot has joined #nixos
fendor has joined #nixos
<pie_>
is there a way to list how much space packages installed with nix-env are using?
stephenjudkins has quit [Remote host closed the connection]
<ottidmes>
clever: Could you tell me what the use is of serial tty? You use it in your kexec script, and I have machine with support of a serial tty, but I have no clue what use it is, other than being able to do some more complete logging.
<pie_>
of course i forgot about that one thing :P
<clever>
ottidmes: in some cases, the graphics drivers get confused during kexec, so you loose all console access
<clever>
ottidmes: and if it fails to boot, you wont get network
zzamboni has joined #nixos
<clever>
now how do you get logs?
<clever>
pie_: moved ycmd to my system packages, total is now 5.4gig
<ottidmes>
clever: So, you use it for better logging as well
<clever>
ottidmes: yeah, to debug when it fails to boot
<clever>
ottidmes: i think i had also used it when testing things under kexec
<pie_>
dude i just freed 80 gigs by removing some haskell stuff...jesus christ
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aszlig pushed 1 new commit to master: https://git.io/vN5YO
<NixOS_GitHub>
nixpkgs/master b3970e7 aszlig: mumble: Fix build with boost version 1.66...
NixOS_GitHub has left #nixos [#nixos]
<tokudan[m]>
ottidmes: it's also pretty common for remote control boards to support giving you serial access, so you can log into your broken system and figure out what exactly is wrong
<mlin>
ottidmes: I found the volume problem, but the internal/i3 module isn't working
<clever>
pie_: i prefer to have to the haskell stuff in a single expression, so they all upgrade at once and share a ghc, rather then having 20 ghc's
<pie_>
yeah
<pie_>
gonna work on losing a lot of my adhoc packages
<ottidmes>
A lot of VPS providers support VNC viewers that even if you system has no network connection still allow you to access your system, but I wonder if I could use serial access in those cases
<pie_>
wtf does trying to install texlive tarbomb my system or something
<tokudan[m]>
ottidmes: yes. it even allows a lot faster working
<tokudan[m]>
ottidmes: for example copy&paste works just fine
<tokudan[m]>
ottidmes: and there will never be any confusion about the keyboard layouts
<clever>
pie_: 5.3gig now
xcmw has joined #nixos
<pie_>
clever, how do i do this for the system store? just replace the path?
stepcut has quit [Remote host closed the connection]
Lisanna has joined #nixos
Isorkin has joined #nixos
<ottidmes>
mlin: What solved your volume problem? And my polybar was used with i3 and it does not contain any special tweaks to make the i3 module work, just the standard config taken from their wiki with some styling tweaks
<clever>
pie_: yeah, /run/current-system
<pie_>
clever, apparently my user store is 21 gigs of small things?
<pie_>
or i truncated the command
<pie_>
:P
mizu_no__ has joined #nixos
<raph_ael>
thanks to all, bye
raph_ael has left #nixos ["WeeChat 2.0.1"]
<ottidmes>
tokudan[m]: That is something I will have to look into than, because last time I had to come up with a very elaborate scheme to write data to the server and get the backup out (encoding the data as color coded console-like pixels and creating many automated screenshots that decoded it all)
<clever>
pie_: also, there will be paths shared between nix-env and nixos
<clever>
and when you du them seperately, those shared items get double-counted
vidbina has quit [Remote host closed the connection]
<tokudan[m]>
ottidmes: ouch...
zzamboni has quit [Quit: Leaving.]
jtojnar_ has joined #nixos
Sonarpulse has joined #nixos
strat has joined #nixos
<ottidmes>
tokudan[m]: It was a crazy idea, but it worked, although it took 2.5 days to complete...
<clever>
pie_: down to 5gig after removing the last thing that depended on ghc
<pie_>
ugh gonna run out of space trying to update agaain
* pie_
tries to clean moar
jtojnar has quit [Ping timeout: 240 seconds]
jtojnar_ is now known as jtojnar
<Lisanna>
how can I make a systemd.service restart when an arbitrary expression changes value? systemd.services.<name>.restartTriggers only works for derivations, despite the documentation heavily implying otherwise ):
mred has quit [Ping timeout: 260 seconds]
<clever>
pie_: and keep in mind, that profiles and generations will hold on to everything you think your removing
<clever>
pie_: nix-store --query --roots on a given storepath, to see what prevents a GC
<strat>
Can anybody known wthat is this error when I use nix-env ($nix-env -qa error: cannot auto-call a function that has an argument without a default value (‘lib’))
<tokudan[m]>
ottidmes: well, serial "only" supports ~115 kbit/s, but that's probably faster than your solution with the screenshots
<pie_>
oh
<pie_>
right
<clever>
Lisanna: use pkgs.writeText to write the result of an expression to a file
<clever>
Lisanna: then depend on the derivation that creates the file
<pie_>
well...if your commands say i have 8gigs used by system and 21 by user, does that mean im using 29 gigs or is that just this generation?
<pie_>
clever, ^
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti closed pull request #34134: Update our default version of OpenSSL from 1.0.x to 1.1.x (master...openssl-1.1.x-update) https://git.io/vN2dn
NixOS_GitHub has left #nixos [#nixos]
<clever>
pie_: some of the data is shared between those 2, and thats just the current generation
<macslayer>
Quick question: I love the feature that NixOS won't suspend when the laptop lid is closed and a monitor attached, but I'd like to also disable suspending after a certain amount of time. See, what's happening is that after about 10-15 minutes of inactivity, my laptop (closed and connected to a monitor) will suspend, and for some reason mouse or keyboard clicks don't wake it up.
<pie_>
clever, how do i remove generations again?
<clever>
pie_: `nix-store -qR /run/current-system ~/.nix-profile` would give a list without dups
asuryawanshi has joined #nixos
<clever>
pie_: low level: nix-env --list-generations and nix-env --delete-generations, high level, nix-collect-garbage --delete-older-than
<Lisanna>
clever if you use an expression in the systemd service itself, will it automatically restart if that expression changes?
<macslayer>
So my question is how I would disable that. Do I use `serverFlagsSection`, setting `SuspendTime` to `0`?
<pie_>
ok thanks
<clever>
Lisanna: yeah
<tokudan[m]>
macslayer: that "feature" of not sleeping you mention is probably a mixture of partially unconfigured hardware or another issue, unless you have explicitely configured that
hotfuzz_ has joined #nixos
<ottidmes>
tokudan[m]: I think I had it working at like 7kb/s, but it at least worked. For serial tty support, I should be able to see it in my journal somewhere right, at least I do one the one machine I own with support for it, but unfortunately my VPSs do not seem to have any
<clever>
Lisanna: if the contents of the generated .service file change for any reason, nixos will restart the service
<tokudan[m]>
macslayer: and the fact that it still suspends after some idle time is probably conigured in your window system (gnome, kde, or whatevery it's using)
peacememories has joined #nixos
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<macslayer>
tokudan[m]: I thought that it was caused by services.logind.lidSwitchDocked
<tokudan[m]>
macslayer: that's quite possible, if you configured it that way
<macslayer>
I ran KDE for a short time, but now I'm just using i3. I'm not sure how to undo any changes that KDE may have made to that option.
hotfuzz has quit [Ping timeout: 248 seconds]
<pie_>
lets see if baobab figures out whats taking so much space...
<tokudan[m]>
ottidmes: if you login to a running vps, check if you have /dev/ttyS*. if yes, the kernel has detected a serial device that you can configure. if you get access to the other side depends on the configuration of the hosting server
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] andir opened pull request #34463: firefox: enable official branding (master...firefox-enable-branding) https://git.io/vN5sg
NixOS_GitHub has left #nixos [#nixos]
<clever>
ottidmes: `cat /proc/tty/drivers` will also reveal what the kernel is using to power that tty
<tokudan[m]>
macslayer: kde shouldn't be able to make any lasting changes in nixos. basically you need to figure out how to configure i3 to not suspend your device then
swayf_ has quit [Remote host closed the connection]
<clever>
ottidmes: `cat /proc/tty/driver/serial` as well
<clever>
each driver has its own details page, and drivers lists them all
swayf has joined #nixos
<macslayer>
tokudan[m]: Oh, okay, thanks! I think I found the issue. Running `xset -q` shows that it should suspend after about 10 minutes. So I"m setting `services.xserver.serverFlagsSection`, and seeing if that works.
xcmw has joined #nixos
<macslayer>
I'm not 100% sure on the syntax of a multiline string in Nix though
<ottidmes>
tokudan[m] and clever: Thanks, I found them! And I will update my configuration right away, before I forget and find myself in a situation needing them
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
stephenjudkins has joined #nixos
<clever>
pie_: nix-channel --add again, and set the name to nixos, and the url pointing to maybe 17.09
<pie_>
hm
<clever>
the man page shows how to set the name
<ottidmes>
clever: I get serial /dev/ttyS 4 64-95 serial
<clever>
ottidmes: and also check `cat /proc/tty/driver/serial`
* pie_
upgrades again...
<hakujin>
I'm building a haskell project with nix (and am fairly new to nix) that includes ruby dependencies and pinned nixpkgs but I'm getting `warning: dumping very large path (> 256 MiB); this may run out of memory` when running `nix-build`: https://gist.github.com/hakujin/d1cb0a301b78bd83e4720251a0a71961
<hakujin>
I _think_ this has to do with copying nixpkgs into the nix store, but I'm unsure if that is 1) typical for pinned nixpkgs, and 2) correct
<hakujin>
someone mind glancing at that ^ gist?
<ottidmes>
clever: Then I get a list which probably matches those made available as /dev/ttyS{0,3}
<clever>
hakujin: line 28, how big is that directory it refers to?
<clever>
ottidmes: yep, and it also says what IO port each is on
ThatDocsLady has quit [Ping timeout: 252 seconds]
<clever>
ottidmes: in the case of my xen guest, the bulk of the 32 i saw had the same invalid port of 0
<hakujin>
clever: big in what dimension? MB?
<clever>
hakujin: yeah
gilligan_ has joined #nixos
<gilligan_>
Evening
throwboy has joined #nixos
stephenjudkins has quit [Read error: Connection reset by peer]
stephenjudkins has joined #nixos
<throwboy>
Question: is it generally ok to create files for importing to "configuration.nix" directly in "/etc/nixos"?
<hakujin>
clever: 2MB
leat has joined #nixos
<ottidmes>
clever: Mine all have different ports that seem valid as far as I can tell (zero experience, but they are not just 0 or something), only difference is that 2 have an irq of 3 and the other 2 an irq of 4
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] bzizou opened pull request #34464: irods: 4.2.1 -> 4.2.2 + server bugfixes (master...irods-4.2.2-bz) https://git.io/vN5nK
NixOS_GitHub has left #nixos [#nixos]
<clever>
hakujin: and thats everything, you didnt exclude anything?
<clever>
ottidmes: that sounds like normal hardware serial, or a VM trying to emulate normal hardware
<hakujin>
clever: ah, hmm. it looks like I previously had cabal's output directories included which balooned the size above 200MB
<gilligan_>
How can I configure a "oneshot" system service to only *ever* run once on a system? By using some sort of lockfile or something? I'm not that good with systemd stuff - any input appreciated ;-)
<hakujin>
clever: is there a way to `.ignore` stuff like cabal's `dist/` output directories?
<clever>
hakujin: builtins.filterSource, or just manually delete them
<clever>
ottidmes: what does `dmidecode -s system-product-name` say?
<hakujin>
clever: unrelated (I think..), have any idea why `nix-build` takes exponentially longer to run a haskell project's test suite than `nix-shell` + `cabal test`?
<pie_>
soooo...how do i log out of my desktop environment if the shell has crashed?
<hakujin>
like, 100x or more
digitus has joined #nixos
<pie_>
i can run a command in it though
<pie_>
alt-f2 still brings up a runner
<clever>
pie_: ctrl+alt+backspace is a hotkey in xorg, that just kills xorg
<pie_>
ah that works
<pie_>
o/
<pie_>
um
<clever>
hakujin: callCabal2nix involves import from derivation, so it has to run cabal2nix every time the source changes, to generate the nix expressions
<pie_>
didnt they change that key combo a while back or somethng
<pie_>
doesnt work for me
<clever>
pie_: it has been disabled on some systems
<clever>
pie_: if you can get root, systemctl restart display-manager
randomstrangerb has quit [Ping timeout: 260 seconds]
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
<clever>
hakujin: there is also some more layers of overhead in how ghc gets bootstrapped
throwboy has joined #nixos
<ottidmes>
clever: The output is: Standard PC (i440FX + PIIX, 1996)
asuryawanshi has quit [Remote host closed the connection]
<pie_>
well looks like that worked
fyuuri has joined #nixos
<pie_>
\o/
randomstrangerb has joined #nixos
<clever>
hakujin: nix has to build ghc1 with ghc2 which was built with ghc3, and it has to eval the closure of every ghc thru nixpkgs
<clever>
ottidmes: sounds like your just on baremetal, so it would be up to either the BMC/IPMI or a physical serial cable
pie_ has quit [Remote host closed the connection]
<throwboy>
Sorry if I a annoying any of you, but it would be really nice if someone could tell me if it's ok to create files in "etc/nixos".
<ottidmes>
clever: I doubt it is bare metal, I am using a VPS at transip.nl, which is KVM I believe
<throwboy>
*am
<hakujin>
clever: hmm. would that mean that actual compilation step (pre-tests, equivalent of `cabal build`) should also take exponentially longer?
pie_ has joined #nixos
asuryawanshi has joined #nixos
<clever>
hakujin: there is also some overhead in nix setting up sandboxes
<clever>
hakujin: as a test, you can compare how long it takes to nix-instantiate an expression, vs how long it takes to `nix-store -r /nix/store/foo.drv` the derivation it produced
<clever>
hakujin: that will tell you how much is in the nix language, and how much is just building
thblt has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
asuryawanshi has quit [Remote host closed the connection]
asuryawanshi has joined #nixos
<coconnor>
throwboy: definitely not annoying :) "ok" is always relative.
<coconnor>
throwboy: there are definitely modules that place files in NIX_CONF_DIR (aka /etc/nix)
<coconnor>
throwboy: they are all nix related. So, yes, it's OK if nix related.
<coconnor>
throwboy: what are you thinking of placing in nix conf dir?
<ottidmes>
clever: vendor: QEMU with `lshw -class system`
<throwboy>
coconnor: /etc/nixos not /etc/nix
<clever>
ottidmes: aha, so the system-product-name isnt always accurate
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<coconnor>
whoops!
<throwboy>
coconnor: i want to place files to import from my main configuration.nix there
<pie_>
is there any way to make this work? something = [with pkgs; lyx texmaker texlive.combined.scheme-full]
<coconnor>
throwboy: yes. that is fine
<clever>
pie_: with pkgs; [ lyx texmaker texlive.combined.scheme-full ]
<pie_>
oh
<pie_>
what will make that scoped to the assignment?
<clever>
yeah
<ottidmes>
clever: apparently so, which is always a shame, just like ECC checking, there are many ways to do it, but almost none are 100% certain (other than actually creating hardware failures by e.g. overclocking and seeing them fixed)
<clever>
pie_: both of the above, are equivelant to just a raw value or expression
<clever>
and can be nested wherever a value is expected
<pie_>
oh i thought the last ; would belong to value
<clever>
ottidmes: there is also the fun of finding the block size of the hard-drives
<pie_>
i mean like with foo; bar;
<clever>
ottidmes: when hardware started to go beyond 512 byte blocks, it broke many OS's, so the hardware lies and claims its still 512 bytes
xcmw has joined #nixos
<throwboy>
coconnor: is this generally where I should store files I want to import to the systemwide configuration or is there a better/more correct option?
<ottidmes>
clever: tell me about it, the official documentation is often lacking in that regard, or there are multiple version and they do document it, but continue by not telling which version you have
<pie_>
throwboy, wouldnt that just lead to even more breakage ;_;
<clever>
ottidmes: then if an OS starts a partition on block 3, and the FS deals with groups of 4 blocks, it will always overlap between 2 hardware blocks, which destroys performance
bpa has joined #nixos
<coconnor>
throwboy: I prefer having a minimal configuration.nix in /etc/nixos that only imports a config from a package plus any node specific options
ryanartecona has quit [Quit: ryanartecona]
<clever>
coconnor: similar here, my configuration.nix only contains what is required to make the machine boot, and imports = [ ./nixcfg/amd.nix ];
<clever>
then i just git clone nixcfg, and put a tree of modules in there
<coconnor>
clever: aye! works well IMO
<clever>
each machine has its own root that it starts at, and all of them include core.nix at some point
<coconnor>
clever: I use the git clone for my home machines. For work, the imports are from a nix package. Avoids some issues; adds some others
<coconnor>
clever: I think yours is better organized than mine haha.
<mbrock>
I'd like to have a sense of how much intelligence is in Hydra as a build server, disregarding the web interface... in particular, does it do any kind of load balancing? Or does it all just boil down to invoking `nix-build` and then the Nix daemon itself takes care of distribution?
<clever>
mbrock: it doesnt involve nix-build at all, but rather, directly calls ssh and talks to the remote nix-daemon
<coconnor>
clever: thanks! always nice to have more examples :)
<niksnut>
yes it does some load balancing
<clever>
mbrock: you can set a weight for each jobset, and there is a queue within hydra and you can bump some things up
goibhniu has quit [Ping timeout: 248 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] svanderburg pushed 1 new commit to master: https://git.io/vN58m
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 0d7a0d7 Sander van der Burg: titaniumenv: fix nasty IPA generation bug
<pie_>
ugh my plasma is all kinds of crashy
<mbrock>
okay, thanks. I also wonder whether Hydra is stateful with regards to the remote build nodes, or whether they can be treated as "cattle" and re-launched whenever. Like if build-1.example.com suddenly reboots into a fresh NixOS instance, it would be good if Hydra could continue to work with it seamlessly
<clever>
mbrock: hydra makes no attempts to track what a slave has previously downloaded and to optimize things around that
<clever>
mbrock: so complete loss of the store on a slave has little impact
<clever>
mbrock: only issue, would be the interuption of something that was in-progress at reboot
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mbrock>
yeah, some items in the queue would end up in some kind of limbo
<coconnor>
CrazedProgrammer: nice! I like the op level organization. will steal ;)
<coconnor>
s/op/top/
<clever>
mbrock: i have noticed poor scheduling before, where it will do 2 steps of a build on different slaves, and have to transfer the closure twice, once to each slave
fyuuri has quit [Ping timeout: 248 seconds]
<clever>
mbrock: and things getting stuck in limbo is either a case of the retry timeout not being visible, or require features not being met, and the GUI cant tell you what its waiting on
<pie_>
clever, something's fucky with my libraries, i get this from plasmashell and then it crashes: Cannot mix incompatible Qt library (version 0x50902) with this library (version 0x50901)
<clever>
mbrock: but if things are just interupted, it will restart on another slave
<pie_>
clever, this is after restarting xorg. is this something a reboot might fix?
<clever>
pie_: aha
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/release-17.09 2b89abe Sander van der Burg: titaniumenv: fix nasty IPA generation bug...
<NixOS_GitHub>
[nixpkgs] svanderburg pushed 1 new commit to release-17.09: https://git.io/vN58X
NixOS_GitHub has left #nixos [#nixos]
<clever>
pie_: reboot wont fix that, i have an issue open for it
<clever>
pie_: the problem is that you mixed QT versions inside nix-env
<pie_>
well i shouldnt have
<clever>
pie_: the only solution, is to upgrade every QT package in nix-env to the same nixpkgs
<clever>
and/or remove them and let systemPackages keep them in sync
<pie_>
i mean...this is a system package...
<clever>
the system plasma obeys the env vars that say to look in ~/.nix-profile/
<coconnor>
is it possible a user package is masking the system?
<clever>
and that has conflicting versions of the libraries
<pie_>
oh dear
<clever>
which just break everything :P
<pie_>
so whats the sane way to fix this
<clever>
coconnor: yeah, ~/.nix-profile has priority over systemPackages
a6a3uh has quit [Quit: Bye]
<coconnor>
makes sense
<pie_>
oh
__Sander__ has quit [Quit: Konversation terminated!]
<clever>
pie_: entire make sure every single QT program in nix-env is on the same QT version as nixos, or remove all QT programs from nix-env
<pie_>
ok i guess i need to upgrade my nix env stuff
<coconnor>
I've been appreciating the ability of users.users.X.packages to manage user env
<throwboy>
so I am really new to NixOS and hear about your minimal configuration.nix. If it is this minimal how do you people manage your system?
<clever>
pie_: if you do upgrade your nix-env stuff, it will cause the same problem 6 months down the road, when you update nixos again
<throwboy>
With what files/tools?
<pie_>
clever, well time to upgrade firefox and get all the dumbass changes mozilla came up with :c
<pie_>
srsly mozilla...wtf r u doin
<clever>
pie_: firefox doesnt use QT, so it shouldnt matter
<clever>
pie_: check nix-store --query --tree ~/.nix-profile/ for anything refering to QT 5.9.1, and upgrade those
<pie_>
i mean i think i have firefox from my nix-env, if i shotgun upgrade
<pie_>
oh
<pie_>
i guess that works
<clever>
pie_: it should result in you having only QT 5.9.2 in the tree
<Lisanna>
I don't think the documentation for restartTriggers and the implementation match up at all. Should I fix the documentation, or the implementation?
<clever>
or better, uninstall them, and add them to systemPackages so they remain in sync
jtojnar has quit [Quit: jtojnar]
<coconnor>
throwboy: only the /etc/nixos/configuration.nix is minimal. The repo imported by that config is not
<coconnor>
throwboy: that repo, tho, lets you collect configuration common to your systems. EG: all my systems reference the same "home" network config
<coconnor>
throwboy: git commands are used to update a nodes config. by pulling hte latest of hte shared repo
<coconnor>
throwboy: I use some scripts that assist in testing. They are only wrappers around nix-build and nixos-rebuild
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/vN544
vidbina has quit [Remote host closed the connection]
<throwboy>
coconnor: so its technically file imports, but those files are git/github-managed?
peacememories has joined #nixos
<catern>
haha, 1.12 is now 2.0
<catern>
very nice
<catern>
does that mean the release is coming soon?
<coconnor>
throwboy: right!
<coconnor>
throwboy: /etc/nixos/configuration.nix has an import, for example, "computers/agh/configuration.nix". Where that file is in a git repo and contains the actual config for the system "agh".
peacememories has quit [Client Quit]
<throwboy>
coconnor: Oh so its really just file imports, right?
<coconnor>
throwboy: where /admin/nix_configs is a checkout of a git repo
<throwboy>
coconnor: Ok, I think I will go this route then.
<clever>
i use imports = [ ./nixcfg/agh.nix ]; and keep the repo in /etc/nixos/nixcfg
<maurer>
u/go 22
<pie_>
clever, is qt4 also an issue?
<coconnor>
throwboy: you can also set NIX_PATH subvariable nix_config but I'd go with clever's scheme
<clever>
pie_: i dont think 5 can ever load 4 by mistake
<clever>
pie_: double-check in ~/.nix-profile/lib/ and you should see a qt directory, that is partially namespaced on version
<clever>
pie_: but the runtime assertion is more strict then that namespacing
<coconnor>
throwboy: the minimal config, even though it usually is only an import, does give an opportunity for very localized config. usually not needed, but nice to have when in trouble!
<pie_>
hm well i did the removals but i still have the issue
<pie_>
ok so it seems like qtbase is a "root" package not pulled in by anything...
thblt has quit [Ping timeout: 264 seconds]
<coconnor>
throwboy: I use a dedicated "admin" account for the checkout. User writable by that admin, but only readable for others
<pie_>
but its not in my nix-env -q list...
<clever>
pie_: propagated-user-env-packages
<clever>
pie_: something else you have is silently adding it
<pie_>
aha.
<clever>
keep searching the --tree for any other things depending on QT
<clever>
propagated-user-env-packages will also silently ignore collisions
alunduil has joined #nixos
<clever>
so when things ask for both 5.9.1 and 5.9.2, it just gives one and doesnt tell you
<coconnor>
throwboy: however, for my home systems, I add a "build" and "switch" scripts to the repo. WHich enables using a local checkout for config dev
<clever>
which is the oposite of what nix-env normally does
xcmw has quit [Ping timeout: 260 seconds]
Neo-- has joined #nixos
<domenkozar>
this audio bug is starting to be annoying
<pie_>
clever, ok i got it
<pie_>
thanks
<coconnor>
throwboy: I can always rely on the /etc/nixos/nixcfg checkout being "stable" but still iterate
<pie_>
clever, no idea what i would have done had you not been here 0.o
<clever>
domenkozar: next thing i was thinking of testing, disable pulseaudio, and then use an old alsa test application to see if the alsa drivers are even working
<clever>
pie_: this is also the second time ive seen this bug crash plasma
aj_ has joined #nixos
<Isorkin>
Help. How to need, to enable mount options lazytime?
<coconnor>
throwboy: anyways, don't worry about that level of fanciness yet. a user writable checkout is a fine start :)
<domenkozar>
clever: where is alsa test app?
<domenkozar>
ah
<domenkozar>
speaker-test
<pie_>
clever, so now im forced to use verything via nix-shell -p or does that also break? :P
<clever>
domenkozar: yep
nuncanada has quit [Ping timeout: 240 seconds]
ryanartecona has joined #nixos
<clever>
pie_: a lot of QT things assume that the propagated-user-env-packages land in $HOME/.nix-profile/ and configure their search paths to match that
<clever>
pie_: nix-shell works purely thru env variables, and wont set it up right
<throwboy>
coconnor: Yeah, I should probably keep it simple for now, because, i have to admit, I did not really understand most of your last 3 messages.
nuncanada2 has joined #nixos
<clever>
pie_: the QT stuff in nix is a bit of a mess, and the one thing ive found that breaks all the rules
<pie_>
clever, sooo....in short everything is broken?
<clever>
domenkozar: and without pulseaudio managing the difference between the speaker and headphone mutes, you may also want to manualy manage that in alsamixer
<pie_>
clever, actually arent qt minor versions supposed to be compativle?
knupfer has quit [Remote host closed the connection]
<clever>
pie_: it works best if you only install QT programs via systemPackages, then they always share the same nixpkgs version
knupfer has joined #nixos
<pie_>
meh :/
<pie_>
clever, im guessing there wont be a fix then any time soon
<clever>
pie_: i think the first step would be to rename the qt-5.9 directory in lib to qt-5.9.1, so that the versions that it complains about, cant conflict
<coconnor>
throwboy: those details assist in my continuous integration setup. Not terribly important for getting started haha
<clever>
pie_: but thats sort of the original problem nix was meant to solve, and relying on things being in a common directory goes against nix
<pie_>
ahh yup Cannot mix incompatible Qt library (version 0x50901) with this library (version 0x50902)
<pie_>
i get that with lyx
<pie_>
so gotta stick to stable
nuncanada3 has joined #nixos
<simpson>
Fucking Qt. So tired of it.
* pie_
pats si
<pie_>
* simpson
nuncanada has joined #nixos
knupfer has quit [Ping timeout: 246 seconds]
<clever>
domenkozar: oh, and do you have virtualbox enabled in nixos?
<clever>
domenkozar: the hda-intel hardware doesnt work with raw pcm samples, so the kernel has to use the right codec when streaming data to the card
<clever>
domenkozar: and some codecs have bugs that others dont
a6a3uh has quit [Quit: Bye]
liori has quit [Read error: Connection reset by peer]
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
m0rphism has quit [Quit: WeeChat 1.9.1]
philipwhite has joined #nixos
m0rphism has joined #nixos
<philipwhite>
I'm trying to install vim with python and perl support on NixOS, but it seems the stock ones don't come with many features. Can somebody help me?
<cinimod_>
How do install TeX? I tried `nix-env -i texlive.combined.scheme-basic`
<cinimod_>
`error: selector ‘texlive.combined.scheme-basic’ matches no derivations`
peacememories has joined #nixos
<clever>
cinimod_: -iA
<clever>
cinimod_: and probably a nixpkgs as well, or nixos
<clever>
cinimod_: only one of those can be installed at once
<cinimod_>
clever: I installed `nix-env -iA nixpkgs.texlive.combined.scheme-basic` but then realised I need some packages so I went for `nix-env -iA nixpkgs.texlive.combined.scheme-full`
<cinimod_>
I then uninstalled `nix-env -e nixpkgs.texlive.combined.scheme-basic`
<cinimod_>
But perhaps that doesn't work?
spear2 has quit [Quit: Leaving]
<clever>
cinimod_: -e only works on names, not attribute paths
<clever>
cinimod_: nix-env -q shows the names
throwboy has quit [Ping timeout: 240 seconds]
<cinimod_>
👍
rauno has quit [Remote host closed the connection]
<dtz>
hmm, looks like 17.09 (latest stable) is on 2016, you're right
aborsu has joined #nixos
<aj_>
What can cause `error: this derivation has bad 'meta.outputsToInstall'`? I get it when installing a makeWrapper'd Go package though the unwrapped version is fine
throwboy has joined #nixos
<gchristensen>
ofborg is incorrectly marking PRs as failing, I'm working on triage and a resolution. sorry
derchris has quit [Remote host closed the connection]
<coconnor>
nice thing about bots is that you can kill them without harm. kill -9 I mean
<coconnor>
sometimes I kill -9 the CI server just so it knows who's in charge
<gchristensen>
lol :)
knupfer has joined #nixos
<jsgrant>
Ha, ended up moving all but this box to NixOS anyways ... :^U
jb55 has quit [Quit: WeeChat 2.0]
<jsgrant>
Is there any trivial way to do hostname handling? Like set if nix detects it's on a certain box, you can load a certain module from your configuration?
<philipwhite>
I feel like I'm shadowing a shadowing a package in my configuration.nix; how can I tell?
<philipwhite>
I have a customized vim, but none of the customizations are showing up.
fendor has quit [Quit: Leaving]
<clever>
philipwhite: can you gist all of the nix files you have modified?
<philipwhite>
When I customize it so the executable is named something else, like `test-vim`, they do show up.
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<philipwhite>
I'm attempting to enable python; when I change the name of the customized vim, I can see python support showing up, but the way it is in the gist, there is no python support.
ryanartecona has quit [Quit: ryanartecona]
ma27 has quit [Ping timeout: 256 seconds]
ma27 has joined #nixos
<clever>
philipwhite: if i'm reading the expressions right, all 3 of those options default to on
<garbas>
magnetophon: yes, i'm on unstable. but lately have little time to fix stuff (parental leave)
<clever>
philipwhite: you installed vim as root with nix-env
<clever>
so you will want to remove that before you continue
<philipwhite>
ooohhhh, I was just trying to uninstall as user.
<philipwhite>
i knew it was some type of shadowing.
<clever>
oh, and that one also has priority over nixos
ryanartecona has joined #nixos
<clever>
so that would explain the shadowing
s33se has quit [Quit: WeeChat 2.0]
<clever>
you can see the priority if you just echo $PATH
fendor has quit [Read error: Connection reset by peer]
dan_b has joined #nixos
FRidh2 has quit [Quit: Konversation terminated!]
coot has joined #nixos
s33se has joined #nixos
swayf has joined #nixos
[0x4A6F]1 has joined #nixos
swayf_ has joined #nixos
erictape1 has joined #nixos
swayf has quit [Read error: Connection reset by peer]
[0x4A6F] has quit [Ping timeout: 256 seconds]
[0x4A6F]1 is now known as [0x4A6F]
swayf has joined #nixos
<philipwhite>
clever: mmhmm, that's a good point to echo $PATH. I probably should have thought of that.
<philipwhite>
well, thanks a bunch for the help.
throwboy has quit [Quit: WeeChat 2.0]
s33se has quit [Client Quit]
<clever>
yep
swayf_ has quit [Ping timeout: 265 seconds]
<plakband>
I'm having some issues with git; when I try to clone I get a permission error for ssh and a "I don't handle protocol https" for https. I have another machine with mostly the same installation where I don't have these issues. Any ideas what might have caused this?
<NixOS_GitHub>
[nixpkgs] edwtjo pushed 1 new commit to master: https://git.io/vN5PH
<NixOS_GitHub>
nixpkgs/master 0307a1a Edward Tjörnhammar: i2pd: 2.17.0 -> 2.18.0
NixOS_GitHub has left #nixos [#nixos]
plakband has quit [Quit: Lost terminal]
jmeredith has joined #nixos
<ottidmes>
This is getting annoying, any idea why every time a I do nixos-rebuild boot, I will get an additional message saying "2 blocks", as if some script that is being run only at rebuilds, keeps getting appended instead of being updated. I vaguely remember having had a similar bug, but that was a long time ago.
<clever>
ottidmes: can you screenshot the terminal when it happens?
<clever>
ottidmes: and gist your entire configuration.nix
<ottidmes>
clever: building path(s) ‘/nix/store/1cbwhw27agvjg05fxiv334krm8fqj8n9-nixos-system-...-17.09.git.0653b73’, and then I get those long list of "2 blocks" messages
<clever>
ottidmes: run nix-store -l on that path, does that show the same message?
nixer has joined #nixos
<ottidmes>
clever: And no, I cannot really gist my entire configuration, I would need to put my whole config repo somewhere online, because it not a simple configuration
<clever>
ottidmes: the importantance of the config depends on if that log shows the message or not
<ottidmes>
clever: nix-store -l with that path gives me no output, but no error either
<clever>
ottidmes: does the message happen every time you nixos-rebuild?
<radvendii>
How does one use win64 wine prefixes on nixos? From what I understand I have to compile wine with that option enabled, how do I turn that on in my configuration.nix?
<nixer>
clever: How does it know what to cache? I assume it used for protecting against outages?
<clever>
nixer: it only contains tar's that this script finds, and only for certain licenses
<ottidmes>
clever, it is for a local server in which I SSH, so I would just create SSH sessions, and call journalctl -f in one and run nixos-rebuild boot in the other?
<clever>
ottidmes: is boot.loader.grub.extraInitrd set?
<clever>
ottidmes: or other secret related things in the initrd?
<ertes>
is there a short github syntax for nix-prefetch-git?
<ottidmes>
clever: I do see a reference to /etc/ssh/ssh_host_* which I have in my initrd.secrets
<ottidmes>
clever: so, yeah
<clever>
ottidmes: that message is from cpio, as it generates an initrd on the fly, containing your secrets
<clever>
and it has to happen after the nix builds, so the secrets stay out of /nix/store/
<clever>
so nix cant cache things
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti force-pushed openssl-1.1 from 5791eff to b319766: https://git.io/vDOWO
<NixOS_GitHub>
nixpkgs/openssl-1.1 ecf3bc7 Robin Gloster: openssl: default to openssl_1_1_0
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/openssl-1.1 76c6978 Robin Gloster: uwimap: add patch to build with openssl 1.1.0
<NixOS_GitHub>
nixpkgs/openssl-1.1 a8ab22f Robin Gloster: postgresql: build < 9.5 with openssl 1.0.2
alunduil has quit [Ping timeout: 264 seconds]
alunduil has joined #nixos
MP2E has joined #nixos
<ottidmes>
clever: But it looks like it is doing this N times, where N is everytime I have called nixos-rebuild since adding that secrets option, surely that is not supposed to be happening? Or do you mean, for every still available generation it needs to rebuild it? So if I were to garbage collect them, they would be back to just the one?
<clever>
ottidmes: each generation has a unique initrd, so it has to re-make each of them
<ottidmes>
clever: Thanks for solving this mystery for me! You are living up to your nickname once again, thanks!
<clever>
ottidmes: there is also a nixos option that limits how many times it will happen, on grub
<NixOS_GitHub>
[nixpkgs] pSub pushed 1 new commit to master: https://git.io/vN5Du
<ottidmes>
clever: I know of that option, but I am OK with just waiting until the next garbage collection, but one annoying thing about the bootloader list of generation is that the date comes last, which is not visible
<clever>
grub and systemd generate the names differently, and ive always been a grub guy
<ottidmes>
So the most important data, which generation build at which time, should come first
s33se has quit [Quit: WeeChat 2.0]
hamishmack has quit [Quit: hamishmack]
hamishmack has joined #nixos
coot has quit [Quit: coot]
orivej has joined #nixos
<ottidmes>
I am still using BIOS on all my machines, so am using Grub as well, but on this server I am working on, it has UEFI, and I had problems with Grub and EFI on Arch Linux, but maybe this is not the case on NixOS
simendsjo has quit [Ping timeout: 240 seconds]
rardiol1 has quit [Remote host closed the connection]
<ottidmes>
Hmm apparently it should not be an issue since there is: boot.loader.grub.efiSupport
<NixOS_GitHub>
[nixpkgs] pSub pushed 1 new commit to master: https://git.io/vN5yq
NixOS_GitHub has left #nixos [#nixos]
rardiol1 has joined #nixos
hamishmack has quit [Ping timeout: 240 seconds]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<kini>
MichaelRaskin: I'm trying to use the quicklisp-to-nix stuff from behind an HTTP proxy -- any idea what I need to edit to get things to work? I tried editing the quicklisp-quickstart invocation in pkgs/development/lisp-modules/from-quicklisp/quicklisp-beta-env.sh but that doesn't seem to be right
s33se has joined #nixos
dan_b has quit [Ping timeout: 264 seconds]
bpa has quit [Remote host closed the connection]
a6a3uh has quit [Quit: Bye]
dan_b has joined #nixos
<MichaelRaskin>
kini: if you are editing source, I would recommend editing quicklisp-to-nix/quicklisp-bootstrap.lisp; after loading setup.lisp, you can set ql-http::*proxy-url*
MP2E has quit [Remote host closed the connection]
<kini>
MichaelRaskin: great, thanks. I'll give it a try.
Itkovian has joined #nixos
<MichaelRaskin>
If you set it to (uiop:getenv "http_proxy") whenever the variable is present and non-empty, I will probably merge such a PR…
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/vN57r
<NixOS_GitHub>
nix/master 88f4f02 Eelco Dolstra: Fix building without aws-sdk-cpp
NixOS_GitHub has left #nixos [#nixos]
mkoenig_ has joined #nixos
dan_b has joined #nixos
mkoenig has quit [Ping timeout: 240 seconds]
ryanartecona has quit [Quit: ryanartecona]
<ottidmes>
I am having trouble with the internet connection timing out if I use boot.initrd.network.enable = true; I think the last time I used it, I just wrote my own implementation and disabled the network connection before going to stage 2, but surely there is a better way? I am seeing: Error changing net interface name 'eth0' to 'enp0s31f6': Device or resource busy in my journal, causing the timeout
aj_ has quit [Quit: Page closed]
<ottidmes>
If it is just the renaming bit, maybe I should just try using enp0s31f6 instead (I am using udev rules to have fixed names for eth0 normally)
<symphorien>
maybe network.predictableInterfaceNames=false will prevent udev from trying to do that
hamishmack has joined #nixos
sigmundv_ has joined #nixos
aanderse has quit [Ping timeout: 260 seconds]
<ottidmes>
symphorien: Luckily it was indeed only the naming causing problems, using that setting solved my issue, thanks!
michaelpj_ has joined #nixos
<symphorien>
:)
ryanartecona has joined #nixos
hiratara has quit [Ping timeout: 252 seconds]
[0x4A6F] has quit [Quit: [0x4A6F]]
hiratara has joined #nixos
hamishmack has quit [Ping timeout: 240 seconds]
sigmundv_ has quit [Ping timeout: 264 seconds]
aanderse has joined #nixos
dan_b has quit [Ping timeout: 240 seconds]
hamishmack has joined #nixos
liori has joined #nixos
simukis has quit [Ping timeout: 248 seconds]
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Itkovian has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra pushed 2 new commits to master: https://git.io/vN5Nb
<NixOS_GitHub>
nix/master 746f8ae Spencer Baugh: remote_store: register for NIX_REMOTE=unix://path...
<almin>
I am trying to package openocd-esp32. I have preConfigure = "./bootstrap"; in my nix expression but I always get "./bootstrap: Error: libtool ist required" despite having libtool listed as buildInput. When I just start nix-shell with libtool I can run ./bootstrap without issues. Does anybody have an idea?
* jsgrant
will be more worried about it, he's sure, once he finishes migrating these 5 boxen either this or next weekend; Islanding on this Fedora box full-time till. Have about 5 little quirks though still.
michaelpj_ has quit [Read error: Connection reset by peer]
<viric>
I experienced a bitflip. Damn it! a sha256 caught it.
<samueldr>
spooky
asuryawanshi has quit [Ping timeout: 260 seconds]
<michaelpj>
a data point: I found an ubuntu bug where people had this issue. `systemctl poweroff` works *without* getting stuck, it's only `reboot` that doesn't work
<michaelpj>
go figure
<michaelpj>
(tried it and it worked for me, to be clear)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] aszlig pushed 1 new commit to master: https://git.io/vN5hZ
<ij>
Can I make buildEnv change the paths? i.e. give paths = ["foo" "bar"], but have "dummy-foo" and "dummy-bar" links in the output?
<clever>
the talk at 58 minutes in, he starts diving into obscure utilities that are installed on every mac
<clever>
yet nobody knows what the utility even does
<clever>
ij: there is a post hook you can set, which can just rename whatever it wants
<viric>
clever: interesting, but I don't see the link
<ij>
clever: Are you talking about bitsquating?
<viric>
clever: the link with our talk :)
aarvar has joined #nixos
<clever>
viric: it was next to the other one in my bookmarks, lol
<clever>
ij: yeah
<viric>
clever: ah :)
michaelpj_ has quit [Ping timeout: 260 seconds]
<viric>
very interesting, the bitsquating
<viric>
I wonder whether those failing systems would have memtest catching anything
<viric>
but there is nothing better than memtest, right?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 6518772 Yegor Timoshenko: linuxPackages.broadcom-sta: fix build with Linux 4.15...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] yegortimoshenko pushed 1 new commit to master: https://git.io/vNdvO
<ij>
Hm, renaming the links will be hard. I might just change the link target paths. How does buildGoPackage figure out what to call the bin/ executable?
<viric>
clever: so do you expect to be helped by zfs crcs?
<viric>
clever: I understand snapshots and other things, but the crcs?
<clever>
viric: zfs would prevent the issues on-disk, but in my case, problems could still happen in ram and over the nfs layers
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] catern opened pull request #1831: document ability to set NIX_REMOTE=unix://path/to/socket (master...master) https://git.io/vNdvC
NixOS_GitHub has left #nixos [#nixos]
<clever>
so bit-rot wont be an issue, but while moving, it may become corrupt
<viric>
I know
<viric>
but does bit-rot happen?
<clever>
unknown, i wasnt verifying the files often enough when they did become corrupt
<joepie91>
viric: I had some issues with my GPU writing junk to my RAM due to power management issues, and this actually seriously corrupted a bunch of stuff in my Nix store
<viric>
auhm
<clever>
joepie91: oh, that reminds me, for the longest time, ive found evidence of my GPU using tooltip textures before they where initialized
<viric>
interesting. One of the computers has weird broken graphics, to be honest
<joepie91>
viric: AMD?
<clever>
so my tooltips always rendered as garbage on the first drawing
<viric>
joepie91: not AMD gpu
<viric>
joepie91: AMD cpu, nvidia gpu
<clever>
joepie91: i have even seen my windows wallpaper, slightly corrupted, upon logging into nixos!!
<joepie91>
viric: disable DPM in BIOS, reboot, see if it persists
<joepie91>
er, UEFI*
<clever>
i have since switched to the amd-gpu driver, and most of those issues have gone away
<viric>
disable what?
<clever>
but it still locks X up solid every now and then
<joepie91>
viric: it'd be very interesting if it's the same issue on your end, since I haven't spoken to anybody who recognized the issues
<clever>
joepie91: and changing the power management options breaks the gpu entirely on my end
<joepie91>
viric: DPM, Dynamic Power Management, it may also be called APM or something like that
raynold has joined #nixos
<joepie91>
something ending in "Power Management" that suggests it changing
<viric>
joepie91: It's a very old nvidia; I thought drivers were not handling things ok
<viric>
I planned to change the gpu
<joepie91>
viric: well, this option is free, so it's worth a try :)
<viric>
:)
<viric>
ok
<joepie91>
viric: please do let me know whether it fixes it, I'm still looking for more cases of the issues I was having
<viric>
In fact the issues appeared not when I replaced the cpu/motherboard, but when I updated to last kde
<joepie91>
clever: eep
<viric>
joepie91: ok
<clever>
52 #"amdgpu.dpm=0"
<clever>
joepie91: i had to comment that out to even get graphics
<viric>
all is quite fragile, then
<joepie91>
viric: I specifically had KDE failing on my end due to this bug, but in other cases I had things like Chromium having rendering glitches
<joepie91>
viric: it particularly manifested when something was initialized on or uploaded to the GPU; a new OpenGL context, a new texture, etc.
<joepie91>
as far as I can tell
<clever>
joepie91: chromium is the only place i noticed that tooltip issue
<viric>
joepie91: ok
<joepie91>
viric: (KDE just outright did not boot for me)
<viric>
I see broken graphics in all desktop subicon text
<joepie91>
clever: oh, amdgpu has a dpm switch now?
<viric>
and only there.
<joepie91>
this is news to me
<joepie91>
but yeah, I have the same setting at 0, on radeon though
<joepie91>
if DPM is not enabled in radeon, all graphics hell breaks loose
<joepie91>
:p
<joepie91>
er
<clever>
joepie91: i cant remember who brought it up last time in #nixos, but they mentioned the switch
<joepie91>
not disabled*
<joepie91>
clever: I mean, I knew about radeon having it, but last I checked amdgpu was missing it, maybe it's been added since
<viric>
^Z.
<joepie91>
I think the summary of this all is: GPUs are a bloody mess :)
kakuzade has joined #nixos
<joepie91>
viric: btw, to describe my symptoms a bit more: under `fglrx` it caused random graphical glitches/corruption (because it wrote junk into the GPU part of the RAM, presumably), under every other driver (`radeon`, `amdgpu`, `amdgpu-pro`) it caused processes misbehaving and eventually a sudden and total system hang (because it wrote junk into userland/kernel RAM)
Lisanna has quit [Ping timeout: 268 seconds]
<joepie91>
my suspicion is that when dynamic/adaptive power management is enabled, it sometimes writes beyond the part of the RAM it should be writing in
<joepie91>
and that different drivers allocate it different parts of the RAM, with more or less risk of overlap with non-GPU stuff
<joepie91>
viric: also, I've had this issue with DPM enabled for *either* my GPU *or* my CPU
<joepie91>
hence why I suspect it may even apply with your nvidia card
<joepie91>
but the nature of the bug means that it can manifest in completely different ways on different systems
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 2 new commits to staging: https://git.io/vNdJ7
NixOS_GitHub has left #nixos [#nixos]
<clever>
joepie91: i'm also reminded of a cheap trick i used, back when i had 64mb of ram
<clever>
joepie91: the gpu also had 64mb of ram, and it was used as a server
<ij>
I'm opening up a nix-shell for terraform with nix-shell -I $HOME/nixpkgs-copy -E 'import and stuff…', but no terraform is in $PATH.
<clever>
joepie91: some quick math confirmed how big the framebuffer was, then i used MTD to convert a chunk of video ram into a block device, mkswap, double the ram!
<joepie91>
lol
<contrapumpkin>
ij: use -p, not -E
<clever>
ij: thats likely giving you a shell suitable for building terraform, not using it
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dotlambda opened pull request #34479: postfix: 3.2.3 -> 3.2.5 and correct license (release-17.09...postfix-17.09) https://git.io/vNdUf
NixOS_GitHub has left #nixos [#nixos]
<joepie91>
clever: that sounds like an incredibly bad idea, haha
<ij>
contrapumpkin: Hello mister, are you copumpkin?
vidbina has joined #nixos
<contrapumpkin>
ij: sometimes
<ij>
contrapumpkin: I've created an issue surrounding terraform's plugins and I've solved the issue with a postBuild in the buildGoPackage.
<contrapumpkin>
ah
erictape1 has joined #nixos
<ij>
I chased this bug for two hours, I might've squashed it.
<contrapumpkin>
nice!
Lisanna has joined #nixos
<contrapumpkin>
I can't look at any real code right now but if you ping me on the issue I'll take a look when I get a moment