CMCDragonkai1 has quit [Remote host closed the connection]
<bigvalen>
Has anyone tried to play with iSCSI on nixos ? I was going to use the 'targetcli' tool - but it tries to write to /etc/target. What's a usual way to fix that ? Patch the source to write to /var/lib/target instead? Or make a symlink in /etc/ pointing to /var/lib/target?
<bigvalen>
clever, The iscsi-boot.nix - that's for making nixos RPi's that boot iSCSI ? For tgt_service - What is 'backingStore' ? Is that a directory ? A partition ? A flatfile ?
<c74d>
I have this issue: <https://github.com/NixOS/nixpkgs/issues/89599>. Chrome's sandbox helper program needs to be setuid root, which it can't be in the store, but security.wrappers doesn't work, because Chrome invokes the program by its absolute path. As a slightly desperate 'fix', how unreasonable would it be to copy the whole google-chrome store directory out of the store so that it can work?
<{^_^}>
#89599 (by 8573, 40 weeks ago, open): Google Chrome complains that its SUID sandbox isn't configured correctly
<c74d>
(E.g., rsyncing it from the store on activation)
<c74d>
s/so that it can work/so that the helper program can be setuid root)
gustavderdrache has joined #nixos
growpotkin has quit [Ping timeout: 276 seconds]
NeoCron has quit [Remote host closed the connection]
<matthewcroughan_>
looks like this does it, I don't understand it though
Negher_ has quit [Quit: Leaving ...]
Negher has joined #nixos
<simpson>
I want to say "what have you tried?" and "yeah, that's about how I would do it", but I don't think that those are helpful. What's currently blocking you?
gustavderdrache has quit [Quit: Leaving.]
<matthewcroughan_>
simpson: understanding what L71-76 of that file I linked is doing
<matthewcroughan_>
because 65 shows that default = {}, so what is 71-76 doing?
<matthewcroughan_>
"Function that converts the option value to something else."
<matthewcroughan_>
What exactly does this mean?
<DigitalKiwi>
one does not need to have a physics degree to swing a hammer
<simpson>
Er, not defaults, but forced overrides.
<clever>
matthewcroughan_: after the final merge is done, nixos will run the apply function, and the result of apply is what you read with services.activemq.javaProperties
<simpson>
matthewcroughan_: It's like a post-processing step.
<matthewcroughan_>
Yes, but default = {}
<matthewcroughan_>
so the app should by default never have any of those options set? So what is the point?
<clever>
matthewcroughan_: default = { foo = 1; }; would entirely be replaced by javaProperties = { bar = 2; };
<matthewcroughan_>
at what point are any of those options in 'apply =' used?
<clever>
so then you loose the foo=
<clever>
they are acting like defaults, because they are on the left of a //
<matthewcroughan_>
Okay. So, if I choose to fill javaProperties with *anything*, these values in the attrset of `apply=` will be forcibly concatenated?
<clever>
but they have higher priority, and wont magically vanish if you add anything
<matthewcroughan_>
but if I do not set javaProperties to anything, then indeed, nothing will be passed?
<clever>
but they also cant conflict with your own things
<clever>
matthewcroughan_: apply is putting those values on the left side of //, so they have a lower priority
<clever>
so if you ever set activemq.base, your version has priority
<matthewcroughan_>
clever: They should never be set, because default = {} though
h0m1 has quit [Ping timeout: 260 seconds]
<matthewcroughan_>
it should be null, in all circumstances
<clever>
the problem is that if you set default = { a=1; b=2; };
<clever>
then a/b are at the default priority level
<clever>
if you ever do javaProperties = { c=3; }
<clever>
then that is at the non-default priority, and so ALL of the default ones (a/b) vanish
<matthewcroughan_>
I do not understand when anything in apply will be set.
<matthewcroughan_>
If default = {}, then apply = { something } is ignored, no?
<clever>
it isnt ignored
john-appleseed[m has joined #nixos
<matthewcroughan_>
then when is it used?
<clever>
apply always gets ran, when you read javaProperties
<matthewcroughan_>
is default = {} false then? Clearly some values *are set*?
h0m1 has joined #nixos
<clever>
if you never set javaProperties, then the attrs on line 71 is the default of {}
<clever>
yeah, it is very funky
<matthewcroughan_>
the default being = to {} is a complete lie then?
<clever>
its acting like a default, by different rules
<matthewcroughan_>
so it's a complete misuse of the function?
<clever>
its a different way of providing defaults at a higher priority
growpotkin has joined #nixos
<matthewcroughan_>
sure, but if I went to search.nixos.org and looked for this option, I'd be misled in thinking that it was empty
<clever>
matthewcroughan_: the default of this option dynamically changes, based on other options
kaivai has quit [Ping timeout: 265 seconds]
<matthewcroughan_>
Yes, but that's unrelated.
<matthewcroughan_>
If the default = {}, I expect it to reflect the truth.
<matthewcroughan_>
But when apply = is used, it lies.
<clever>
yeah
<clever>
you could pass that same set to default, but anybody knowing the rules would still be confused more
<clever>
getting the auto-generated docs to explain these edge cases isnt easy
<simpson>
From the perspective of a user, it's not a lie; the user probably *doesn't know* about those flags being required.
<matthewcroughan_>
That hydra option being dependant on other options is really cool though. WOw.
<matthewcroughan_>
clever: " # Function that converts the option value to something else."
kaivai has joined #nixos
<matthewcroughan_>
What does this mean? It seems meaningless to me.
<matthewcroughan_>
"To something else".
<matthewcroughan_>
Windows 10: Something happened.
<simpson>
It's just open-ended. Like a polymorphic type which hasn't been specialized. The idea is that the post-processing can completely change the configured data's type, if that's easier/faster to work with.
<matthewcroughan_>
simpson: deep nix fu, I'll understand that when my beard is grey.
<clever>
matthewcroughan_: after nixos has finished merging every value assigned to that option (or picking the default), it will run apply on the merged result
<clever>
matthewcroughan_: the result of apply, is what you see at config.services.activemq.javaProperties
<clever>
matthewcroughan_: so line 126 is reading the result of the apply function, at all times
<clever>
matthewcroughan_: the other package, was using an attrset, so you could do { foo = "bar"; }, rather then [ "foo=bar" ];
<matthewcroughan_>
I want it to be foo=bar because this is literally how you interact with nodered on the cli
<clever>
matthewcroughan_: because [ "foo=bar"] and [ "foo=baz" ] dont conflict, and you wind up with [ "foo=bar" "foo=baz" ]
<clever>
do you want to silently allow duplicates?
<matthewcroughan_>
It doesn't seem to matter on the cli
<clever>
does it use both, or just one?
<matthewcroughan_>
just one, and I'm not sure on the precedence
<clever>
then you want nixos to error out when duplicates are given
<clever>
so the type should be attrs
<matthewcroughan_>
and the only way to do that is the way activemq has done it?
<matthewcroughan_>
ah, okay
<meatcar>
Hi all, can I get someone to take a look at and potentially help merge a PR? It's been stuck open for a couple weeks now. Thanks :) https://github.com/NixOS/nixpkgs/pull/113871
<remexre>
did something change how sh (the python module) interacts with python recently? A build that does pkgs.python39.withPackages(p: [ ... p.sh ...]) is now trying to build python from source...
<remexre>
(and doesn't when p.sh is commented out)
<remexre>
(git log of pkgs/development/python-modules/sh shows nothing incriminating; this change would've happened in the last month, I think)
orivej has joined #nixos
<jonringer>
remexre: the build is broken, therefore not cached
<hyper_ch>
I used this here: https://paste.simplylinux.ch/view/raw/cc5af7e2 but the problem was that there was no zstd compression available for zfs... but that was put into unstable channel half a year ago
<hyper_ch>
it uses unstable 20.09... but I need newer
mfernandez has quit [Ping timeout: 260 seconds]
mikeplus64 has joined #nixos
noudle has joined #nixos
<mikeplus64>
hi all, i have a haskell package that links to the c library libgit2 that i've updated to use ghc102 and now get a undefined symbol (for a libgit2 func) error when compiling the executable, which previously worked on an older nixpkgs/ghc combination, anyone have any ideas?
<mikeplus64>
ahaaaaaaaaaaa nevermind, this specific function has changed in the new version of libgit2
<pie_>
hyper_ch: Idk offhand how things get put into the iso, but for building the iso youd probably do something like -I nixpkgs=channel:nixos-unstable to set the <nixpkgs> value which is used by impure.nix, if I recall correctly, to set the pkgs thats in scope, and such
<mikeplus64>
oh, nevermind my nevermind, it's actually that the GIT_THREADS opt has been removed from libgit2 in nixpkgs...
<Miyu-saki>
Kinda! I think what your actual question is, "how is <nixpkgs> in general used." So what nixpkgs does is that when you `import` it, the resulting attrset is basically similar to `/pkgs/top-level.nix`.
<Miyu-saki>
This provides you with a function called `callPackage`.
<pie_>
no im asking the exactly converse question :p
<mikeplus64>
(fwiw, my problem was *actually* an incorrect version check. as it's been a while since I touched this codebase I forgot all about it and spent ages doing stuff unrelated to the actual issue. doh)
<pie_>
how does nix-build or eval.cc or (well this is part of figuring out the answer) whichever part of nix is responsiblefor this, figure out where that first default.nix that it should be calling, is
<pie_>
the obvious part is that it checks NIX_PATH
<pie_>
but i havent found the source yet
<Miyu-saki>
So in the "usual use-case", you'll have something like `with import <nixpkgs> {}; callPackage ./. {}`. And when you override that with `-I` it will instead use the nixpkgs from your /nix/store, with no regards to any other nixpkgs installed in your systme.
Lord_of_Life has joined #nixos
<Miyu-saki>
And oh
<Miyu-saki>
TL;DR, it is stored in like ~/.nix/cache or something
<Miyu-saki>
I think
<Miyu-saki>
Err, ~/.cache/nix
<pie_>
actually yeah i guess it makes more sense for it to not have a separate mechanis
<pie_>
im looking for whatever nix-build something.nix does
<Miyu-saki>
Okay, I think I'm a bit confused. Are you trying to figure out how `NIX_PATH` is resolved, or how channels are reified to /nix/store paths?
<Miyu-saki>
Oh
<pie_>
hm. ok i guess nix-build something.nix doesnt need to invoke the nixpkgs variable at all
<pie_>
if it does its because _youre_ using it in your nix code
<pie_>
so yeah i guess that makes sense.
<Miyu-saki>
Yep!
<pie_>
its almost as if i havent been using nix for years :'c
<Miyu-saki>
I mean, it could also be that, y'know? You've been using it for so long, and it's so automatic, that you don't really think about it kinda thing.
<pie_>
hyper_ch: fwiw, i didnt look into it, but when i did nixos-install -I nixpkgs=channel:nixos-unstable at some point, it seemed to set that as the default nixos channel in the new system, so it sounds like it should work
<pie_>
Miyu-saki: no i just never really understood how things work on a deeper level, just accumulating bits and pieces and forgetting things :P
<pie_>
anyway
est31 has quit [Remote host closed the connection]
<hyper_ch>
pie_: the problem is I need a current live iso where I have zfs available with zstd
<hyper_ch>
it's not working with building the iso
<hyper_ch>
it just uses 20.09
<pie_>
well i think that should work with what i said because it should build using the unstable channel
orivej has joined #nixos
dbmikus has quit [Ping timeout: 265 seconds]
zakame has quit [Ping timeout: 265 seconds]
<pie_>
hyper_ch: do you have enough cpu/ram to run a vm?
<hyper_ch>
pie_: yes, but how is that of any help?
ahmedelgabri has joined #nixos
<pie_>
hyper_ch: i highly recommend using vms to test this stuff, faster than rebooting and shuffling images around
<hyper_ch>
no idea what that's good for
zakame has joined #nixos
<hyper_ch>
how can I create an iso from the current nixos-unstable channel
<pie_>
the idea would be that for example, you can test if the image you built has zfs
<pie_>
but thats tangential
<hyper_ch>
pie_: the built images have zfs
<hyper_ch>
but it uses 20.09 and not current unstable
<hyper_ch>
so that zfs version does not have zstd
<pie_>
ok right
<pie_>
hyper_ch: did you try what i said
<hyper_ch>
and hence I can't import the pool on the server
<hyper_ch>
pie_: nix-build '<nixpkgs/nixos>' -A config.system.build.isoImage -I nixos-config=customIso.nix -I nixpkgs=channel:nixos-unstable --> it did build, but again with 20.09
<pie_>
ok thats odd
<hyper_ch>
how do I set in the customiso to use that channel?
<pie_>
well, what happens is that any references to <nixpkgs> now point to nixos-unstable
<pie_>
do you have any other channels, and are you referring to anyhting else with <>?
<pie_>
somethng you can try is set NIX_PATH= in the environment and see what breaks (though some misc. stuff might have to be fixed before we can see what we really want)
<hyper_ch>
pie_: not sure what you're refering to
<pie_>
run it like: NIX_PATH= nix-build ...
jonringer has quit [Ping timeout: 264 seconds]
<hyper_ch>
pie_: trying with an overlay in the customIso.nix now
<pie_>
hyper_ch: I don't know what exactly is going on, but it sounds like you have a different channel leaking in from somewhere, so we need to figure out why
luxemboye has quit [Remote host closed the connection]
philr_ has quit [Ping timeout: 245 seconds]
luxemboye has joined #nixos
<hyper_ch>
pie_: still no luck
<pie_>
hyper_ch: what version is the zfs
<pie_>
that the system is giving you
<hyper_ch>
pie_: some 0.8.x
<hyper_ch>
becuase 2.0 contains zstd
<pie_>
do you _know_ the problem is the channels? if you know the problem is the channels, stop trying random things and figure out whats wrong with your channels
orivej has quit [Ping timeout: 264 seconds]
<hyper_ch>
pie_: I have no idea what the problem ist
<hyper_ch>
pie_: as you can see, I have two -I -- that should work since different things are defined, right?
<Miyu-saki>
Just so you know, `<nixpkgs/nixos>` is not the same as `<nixpkgs> + /nixos` (but it *should* be equal, otherwise, you have a misconfiguration.)
<Miyu-saki>
hyper_ch: Also, just run that.
<hyper_ch>
run what?
<Miyu-saki>
IIRC, it does work, if not, then you have to use nixos-config=customIso.nix:nixpkgs=/tmp/nixos-master
<Miyu-saki>
(I'm not sure if -I uses the : delimiter, but NIX_PATH does)
<hyper_ch>
it did not complain about running it like that
<Miyu-saki>
Right, then there's no problem!
<hyper_ch>
we'll see
<Miyu-saki>
Most of the time, if Nix doesn't complain, you did it correctly.
<pie_>
error: file 'bar' was not found in the Nix search path (add it using $NIX_PATH or -I), at (string):1:19
<pie_>
Miyu-saki: should we expect : to work there? maybe needs an issue if it doesnt have one already
boolman has quit [Quit: leaving]
<Miyu-saki>
Nah, I think multiple `-I` is exactly how you're supposed to pass multiple files. The problem with accepting : as a delimiter is that URIs get a bit more weird.
teto has joined #nixos
<pie_>
ok so i did a quick test and the order of -I as expected, should not be an issue here
<pie_>
hyper_ch: the problem is that youre saying the channel isnt taking effect. trying a different channel wont help
<pie_>
anyway im still catching up on some threads
<pie_>
oh? <Miyu-saki> Just so you know, `<nixpkgs/nixos>` is not the same as `<nixpkgs> + /nixos` (but it *should* be equal, otherwise, you have a misconfiguration.)
<Miyu-saki>
Yeah
<Miyu-saki>
You can for example add nixpkgs/nixos to your NIX_PATH.
<pie_>
<Miyu-saki> (I'm not sure if -I uses the : delimiter, but NIX_PATH does)
<pie_>
<hyper_ch> it did not complain about running it like that
<pie_>
wait what? re:
Milanfallbackacc has joined #nixos
<pie_>
Miyu-saki: ok lol, i guessed as much but wasnt sure
<hyper_ch>
pie_: Miyu-saki: leonardp: it's a typical PEBKAC ..... on the server, there was another usb stick still on the back inserted with 20.09 which I always selected fromthe boot menu
dbmikus has quit [Ping timeout: 256 seconds]
hypokeimenon[m] has joined #nixos
fuzzypixelz has joined #nixos
<Miyu-saki>
lol
<Miyu-saki>
Pain
<pie_>
hyper_ch: haha yeah i was gonna ask something like are you running the right iso
<hyper_ch>
sorry for wasting everyone's time
<leonardp>
hyper_ch: oh wow
<pie_>
you didnt waste much this time ;P i spent it learning some stuff
<ldlework>
pie_: what are you telling me
<pie_>
but yeah thers a heuristic like "always check the things youd be most embarrassed about first" xD
<leonardp>
yeah i think everyone did a bit of learning here :9
<fuzzypixelz>
is managing vscode extentions through home-manager supported? it only seems to work with vscodium for me
<hyper_ch>
well, there were a ton of entries in the list to select which boot devices from efi, and I knew the usb stick just shows as usb generic.... and on the back I still had an identical one in there :(
<pie_>
All hosted environments that are still of interest for portable code provide all of the headers specified in ISO C90 (as amended in 1995).
<pie_>
This macro is obsolescent. Its sole effect is to make sure that all the headers that are included by AC_INCLUDES_DEFAULT (see Default Includes), but not part of ISO C90, have been checked for.
<jboy>
I'm trying to write a simple derivation to build a "./configure --prefix=... && make" type package. I set `makeFlags = [ "prefix=${placeholder "out"}" ];` but it still fails trying to write to `/etc`. Any ideas what might be going wrong?
<pie_>
i probably shouldnt be looking at a 2 year old repository
* pie_
yonks the url from the nix expresision
ahmedelgabri has joined #nixos
<pie_>
jboy: do you know why its trying to access /etc?
<raboof>
jboy: shouldn't that be `configureFlags` then?
<jboy>
pie_: no, I looked at the configure script and it seems to do everything right in terms of prefixing everything
<raboof>
oh, okay
<jboy>
oh, good point raboof, I'll try that
<pie_>
let me rephrase, i dont know your the stuff looks like but i mean you should figure out why its still trying to write to /etc then :P
<pie_>
idk if there's a more convenient way to do it but you can then also set that temporary directory writable with chown/chmod , drop into a nix-shell, and run the phases manually
<pie_>
oh yeah, just to check that the header isnt set by the compiler by default: gcc -E -dM -xc++ <(echo "int main(){}") | grep STDC
sangoma has joined #nixos
respawn_ has joined #nixos
m0rphism has joined #nixos
<pie_>
ok its actually defined in /nix/store/sr0ci8f8pgby77fj4mpcl9bcgxji3676-gcc-10.2.0/include/c++/10.2.0/x86_64-unknown-linux-gnu/bits/c++config.h:#define STDC_HEADERS 1
<pie_>
ldlework: for a while i figured its an include order issue
<pie_>
ldlework: adding #include "/nix/store/sr0ci8f8pgby77fj4mpcl9bcgxji3676-gcc-10.2.0/include/c++/10.2.0/x86_64-unknown-linux-gnu/bits/c++config.h" to sysdeps.h before the check seems to fix it, so the question is why is the include order the way it is
<ldlework>
You got it to compile?
turlando has joined #nixos
ptrcmd_ has joined #nixos
<pie_>
ldlework: no but i got the error to go away
evils_ has joined #nixos
diamondbond_ is now known as diamondbond
<pie_>
idk if thats how it works, but note that here sysdeps is included before like, all the other stuff
<pie_>
specifically, the check happens before /nix/store/f8ahd6cxx9ia4fzdiyczmhcv3c2qmd42-glibc-2.32-37-dev/include/sys/types.h \
<pie_>
and only config.h user_strings_unix.h are included between sysdeps.h and types.h
<pie_>
we probably want all the system headers to show up before the user ones
philr_ has joined #nixos
<pie_>
(mostly ive just been googling and lookig at SO posts for what the commands are to get things out of GCC, ive become more comfortable with this stuff ever since i had a good fight with some clang tooling a while back)
<pie_>
half of it is just figuring out wtf even exists and how to look at it
<pie_>
i really need to start that "how to debug whatever" website..
turlando has quit [Ping timeout: 260 seconds]
* pie_
tries to remember where g++ gets its include paths...
<pie_>
the assumption is nix isnt insane so theres probably a reason these are appanded afterwards and it _should_ work?
SomeoneSerge has joined #nixos
<pie_>
looking for -isystem shows a description of the lookup order
<pie_>
understandably, user directories are searched first...but actually i just realized....theres no reason to expect system includes will be arbitrarily included :P
<everythingTaken>
hi guys, rather basic question, why name of iso different in sha256 file? downloaded iso is latest-*, but in sha256 file starts with nixos*
orivej has joined #nixos
dbmikus has joined #nixos
thehottestpotato has joined #nixos
<everythingTaken>
how do I compare sha256sum latest-* with hash I copied from sha256 file
meh` has joined #nixos
Qwerky has quit [Ping timeout: 265 seconds]
Qwerky has joined #nixos
respawn_ has quit [Quit: Leaving]
mkaito has joined #nixos
mkaito has joined #nixos
mkaito has quit [Changing host]
<everythingTaken>
I had to rename the file name in sha256 file
<everythingTaken>
anyone gonna look into this? perhaps fix it?
<bb_>
it's not installing it in a "reproducible way"
<bb_>
does not seem very nix like
<jumper149>
bb_: There are different approaches to declarative userspace packages.
<etu>
bb_: That kind of interactive install just adds it to the users profile, but yeah, it doesn't add it to any config automatically
<bb_>
I understand
* etu
tends to stay away from such installs
<jumper149>
There is a wiki page, one sec
<bb_>
I have read somewhere
<bb_>
that I can start a temporary shell and install there
<srk>
,imperative
<{^_^}>
nix-env has multiple drawbacks as an imperative package manager. nix-env -u will sometimes upgrade to the wrong thing; the outputs to install are very finicky to override; and packages that have been removed or are otherwise unavailable will remain in your profile without any warnings. Consider using a ,declarative setup instead.
<bb_>
for testing puposes
<bb_>
but I forgot how to do that
<jumper149>
bb_: nix-shell -p firefox
<bb_>
makes sense, thank you
<{^_^}>
[nixpkgs] @primeos pushed to master « chromiumBeta,chromiumDev: Mark as broken »: https://git.io/JmY31
* devhell
agrees with etu
Chiliparrot has joined #nixos
<jumper149>
bb_: I can't seem to find the wiki page I was talking about. But the 2 main ideas are either you use home-manager, or you set up an overlay with your packages.
<bb_>
I use home-manager and shell.nix in my projects folders
<etu>
Or use a shell.nix to declare dependencies :)
<Aleksejs_Home>
hey, does libinput restart during switch?
s1341_ has joined #nixos
<sterni>
oreoking[m]: you don't really need to worry about flakes just yet, it' an experimental feature at the moment and you don't need to necessarily use them for anything currently
<{^_^}>
[cabal2nix] @peti pushed to master « Bump to version 2.17.0 for the upcoming release. »: https://git.io/Jm3EU
dev_mohe has joined #nixos
nek0 has joined #nixos
erasmas has joined #nixos
dbmikus has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<numkem>
I have build nodes for hydra running nixFlakes as their nix and they are now getting the error basic_string::_M_create when hydra is trying to send them a build to do
<nschoe>
Hey everyonem I've got a quick issue: I'm trying to boot on a nixOS installer just to resize my LVM partitions of my main NixOS installation. I've DLed and dd-ed the latest (non grpahical) ISO on a USB stick, but when I boot, it says "error occured on stage 1", I tried to simply "launch an interactive shell", but I don't have cryptsetup commands.
<fuzzypixelz[m]>
hello. I'm using `programs.vscode.extentions` in `home-manager` to ... manage vscode extentions. Except it doesn't seem to work with VSCode, only VSCodium for some reason. I want to do this because the MS C/C++ extensions throws a `no such file or directory) error when running `cpptools` (obviously the binary has linking errors) Any ideas? the `vscode.package` doesn't seem to have any requirements:
<fuzzypixelz[m]>
* hello. I'm using `programs.vscode.extentions` in `home-manager` to ... manage vscode extentions. Except it doesn't seem to work with VSCode, only VSCodium for some reason. I want to do this because the MS C/C++ extensions throws a `no such file or directory` error when running `cpptools`(obviously the binary has linking errors) Any ideas? the`vscode.package` doesn't seem to have any requirements:
<nschoe>
But is it new that the installer needs the USB to be labelled this way? I don't remember is to tbe the case when I last installed nixos, about a year ao
sangoma has joined #nixos
dmoerner has joined #nixos
<fuzzypixelz[m]>
good luck, I know nothing about how exactly these issues come up. I've had the same usb installer on the same machine bootup or not purely on a whim
<ToxicFrog>
I last installed it last week and it definitely didn't need any "labeling"
matsaman has joined #nixos
<ToxicFrog>
Download iso, dd to usb drive, boot from USB and away it goes.
<fuzzypixelz[m]>
<nschoe "But is it new that the installer"> no I don't think so either, I'm sure my last installer had "GNOME" in it
matsaman has quit [Quit: leaving]
<ToxicFrog>
That said, all the systems involved were fairly straightforward amd64 arch UEFI boxes made within the last five years
<nschoe>
ToxicFrog: that's literally what I've just done. It boots all right: I see the 3 or 4 menu entries, I chose the default "installer". But then it immediately fails at stage 1 o_O
<nschoe>
I'm trying Etcher and an old iso file I've just found laying in a folder. Will see :)
<nschoe>
Maybe it's my USB stick, dunnow.
<fuzzypixelz[m]>
wait I never asked, what the full error?
<ToxicFrog>
Oh, so it fails at nixos stage1, not grub stage1
<nschoe>
ToxicFrog: indeed.
<nschoe>
Oh so now it works: either it's Etcher that did the job or the fact that it was an old ISO.
<nschoe>
Dunnow :/
Qwerky has quit [Remote host closed the connection]
<simpson>
Marcelo2324: I don't see `csv`, `json`, or `sys` either. Those are typically part of the Python stdlib; they wouldn't be needed as extra libraries. You should confirm with the person who sent it to you that it works for them.
<simpson>
Sure. Your interviewer should be the one explaining all of this to you.
<Marcelo2324>
I don't wanna spend too much time on it tbh, just wanna make sure the other person can run the project
<Marcelo2324>
thx a lot :)
<Marcelo2324>
how do you know it's for an interview? lol
<simpson>
Because you left the string "interview" in the file. If you're administering the interview, then I'd hope that you know the tools you're asking folks to use.
<__monty__>
Hmm, when I build llvmPackages_7.clang-unwrapped I see it build cmake-root. I assume it's because cmake is a buildInput in the clang expression. However when I try to build cmake it tries to build a lot more than just cmake-root.
<yaymukund>
if I add variables to sessionVariables and variables and they are not accessible from the environment in my terminal emulator, does that mean I've set it up incorrectly?
<jumper149>
yaymukund: I'm not sure, but you might have to start a new session for environment variables.
endformationage has joined #nixos
<evanjs>
What's the best way to "patch" nixpkgs/grab an expression before it's merged? Just grab nixpkgs from the PR's source branch, import and use that package set?
<evanjs>
Trying to use #112886 right now and trying the approach I mentioned
<jumper149>
What you can do is either to just copy that default.nix and use it in your overlay or you could try to use pkgs.qemu.override or pkgs.qemu.overrideAttrs to just override what you want to change
<evanjs>
jumper149: yeah I figured as much
<jumper149>
So postPatch and configureFlags. I think you need overrideAttrs for that
emmanuel_erc has quit [Read error: Connection reset by peer]
emmanuel_erc has joined #nixos
<evanjs>
fuzzypixelz: I am pretty sure my GDM is dark right now... lemme check lol
dupon19 has joined #nixos
<evanjs>
ahhh gdm is grey by default so it just happens to work with breeze-dark on my system haha
<evanjs>
It looks like other DMs have theme options... I wonder how much work it would be to add such to gdm, or if there's an easy way to hook into that to begin with
<andoriyu>
I think it was dead before that. It wasn't a drop-in replacement in any sense. It was good around when it was first released. Ever since then it was lagging behind OpenSSL in every way. Today it offers zero benefit over OpenSSL with bigger cost.
<simpson>
supersandro2000: Merely pointing out that copying other distros is not helpful. Instead, we should copy their *reasons* for dropping certain packages.
<ambro718>
Hi. Did anyone try nvidia driver with kernel 5.10? I get complete breakage, the nvidia kernel module immediately complains that the card is not responding to commands. Works with no problems with kernel 5.4.
<srhb>
ambro718: Think I skipped 5.10 (so maybe this is no help) but 5.11 works here for whatever nvidia card is in my laptop
<andoriyu>
Maybe I don't understand something about linux or nixos, but: I'm tracking unstable nix channel and running linuxPackages_5_11. `uname -a` shows that I'm running 5.11.0, while master is 5.11.6? Why?
<srhb>
andoriyu: What does nixos-version say?
<andoriyu>
`21.05pre275232.29b0d4d0b60 (Okapi) `
<andoriyu>
(had to run use that because of intel xe)
<ambro718>
Ah, 5.11 is not in nixos-20.09 though.
ManiacOfMadness has quit [Ping timeout: 260 seconds]
<srhb>
andoriyu: That is indeed 5.11.5. ls -l /run/current-system/kernel /run/booted-system/kernel ?
<ambro718>
(I updated the nvidia to the latest version manuallt otherwise it doesnt compile with 5.10)
<srhb>
ambro718: I don't know if it makes a difference anyway on your system, just a data point :)
<ambro718>
Yeah thanks.
<andoriyu>
srhb first one goes to 5.11, second goes to 5.11.2
<srhb>
andoriyu: I don't think you have, maybe this is channel confusion? Might you have run nix-channel --update as your own user?
<andoriyu>
I think I did as my user like you said
emmanuel_erc has quit [Read error: Connection reset by peer]
<andoriyu>
oh well
<andoriyu>
thanks
<srhb>
Sure thing.
jumper149 has joined #nixos
emmanuel_erc has joined #nixos
<srhb>
(Getting rid of user channels might be helpful to avoid that pitfall)
<andoriyu>
idk why I'm updating. Still waiting on 5.12 because that's where my wifi supposed to start working.
<srhb>
Heh :P
xelxebar has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
drakonis has joined #nixos
<srhb>
ambro718: I hadn't even realized 5.10 got the LTS label. That saddens me, since a bunch of my stuff only works on 5.11. Here's hoping for backports I guess.
<ambro718>
This line is supposed to be something like: if (!priv->has_hw_rfkill_switch || read_ec_data(...))
<ambro718>
otherwise the bug causes my wifi to be RFkill-blocked (Lenovo Y720)
<ambro718>
Any idea how to get the fix upstream?
<ambro718>
This issue can be found by googling for something like "linux lenovo Y720 wifi not working" and seeing suggestions to blacklist ideapad-laptop module.
<q3k[m]>
thoughtpolice: did you attempt to package bluespec's bdw yet? i'm giving it a shot and i'm running into some incrtcl/tcl issues, wondering if i could pick your brain on this
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JmcyY
emmanuel_erc has quit [Read error: Connection reset by peer]
emmanuel_erc has joined #nixos
<jmercouris>
how do you guys make your shell scripts visible in konsole? I can't use .profile to append to my path
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/Jmcyy
<jmercouris>
my path is just "/run/wrappers/bin:/home/jmercouris/.nix-profile/bin:/etc/profiles/per-user/jmercouris/bin:/nix/var/nix/profiles/default/bin:/run/current-system/sw/bin"
<{^_^}>
[nixpkgs] @peti pushed 840 commits to haskell-updates: https://git.io/Jmcyx
<{^_^}>
[rfc39-record] @grahamc pushed commit from rfc39 to main « Automated team sync results. »: https://git.io/Jmcba
bakerboy has joined #nixos
fuiltilt has joined #nixos
<matthewcroughan_>
clever: How would you symlink to rename binaries in a derivation?
orivej has joined #nixos
<matthewcroughan_>
installPhase
<matthewcroughan_>
does anyone know if there's a way to get access to function definitions in the nix cli? E.g nix edit installPhase to see how installPhase works? I'm getting tired of using my web browser :D
ambro718 has quit [Quit: Konversation terminated!]
<extrowerk>
Hi, i am with the HaikuPorts team here, and just built nix on Haiku. So far success. But i don't know anything at all about nix, so i tried to follow the user guide, and met with some problems. Basically nix tries to write files into /nix/ folderstructure. But unlike unixes / on Haiku is not the root, it is a virtual folder, where every other drive gets mounted. So it is not allowed to create files here, only folders. How can i tel to
<extrowerk>
nix to use a different path?
<tpw_rules>
that's possible, but a bad idea because everything will have to be rebuilt for the new path.
<tpw_rules>
on macos where writing to / is impossible too, the installer creates a separate logical volume to hold everything under /nix . can you not do something similar in haiku?
<extrowerk>
i can create a diskimage and mount it as /nix
<extrowerk>
does it needs to be big?
<lassulus>
depends on how much stuff you want to install with nix
<tpw_rules>
yes. everything nix lives under that
<tpw_rules>
on macos it doesn't matter because the volume will expand as necessary and share space with the rest of the system
<tpw_rules>
you don't have to pre-allocate a fixed amount of space
<extrowerk>
i create a ramdisk for now to test the basic functionality.
<lassulus>
if you want to boot something which depends on nix (like nixos) it should be persistent
<tpw_rules>
you can use that compiler flag to change the directory too. but it can't be changed at runtime, and packages built for one directory couldn't be used in another. maybe the haiku community can come up with a common location that works on every machine, because i imagine haiku cannot run x86_64 linux binaries without modification anyway?
<matthewcroughan_>
extrowerk: NIX_PATH envvar or --store
<matthewcroughan_>
`./nix --store $HOME/.cache/nix/store run -f channel:nixos-20.09 magic-wormhole` for example
<matthewcroughan_>
tpw_rules: what about what I just gave isn't going to work?
emmanuel_erc has quit [Read error: Connection reset by peer]
<tpw_rules>
i don't know, i haven't tried it
emmanuel_erc has joined #nixos
pushqrdx has joined #nixos
<tpw_rules>
but now that i think about it you'll have to modify nix with a haiku target too
<extrowerk>
nix built with debug enabled here, all the object files are huge, also the generated nix binary is over 70Mb+. Can i somehow disable the debug mode? I don't see any obvious way.
<matthewcroughan_>
tpw_rules: it's kind of amazing
ahmedelgabri has quit [Ping timeout: 244 seconds]
<matthewcroughan_>
though I do know some things don't quite work for some reason.. maybe you could help with it, let me try and fetch firefox/chromium again for example.
pushqrdx has quit [Remote host closed the connection]
ManiacOfMadness has joined #nixos
<genevino>
ok sound in chromium is suddenly gone here. even when starting it from a terminal, i see no sink popping up in pavucontrol when i play a video on youtube. suddenly happened, definitely nothing changed. neither rm -rf .config/chromium, nor rebooting, worked. anyone a clue what's going on? firefox continues to work fine.
bitmapper has joined #nixos
<matthewcroughan_>
supersandro2000: what is the "file that collects all modules" ?
codygman__ has quit [Ping timeout: 260 seconds]
<matthewcroughan_>
There doesn't seem to be one in the PR that added cryptpad.nix for example.
ambro718 has quit [Quit: Konversation terminated!]
extrowerk has joined #nixos
<extrowerk>
Ah, yet something: please consider to not use dotfolders and dotfiles in the $HOME. Thats unacceptable on Haiku, even for ported programs. It is just bad practice.
<colemickens>
Nix does?
<extrowerk>
t made .nix-channels file in $HOME
<extrowerk>
also ~/.config/nix and ~/.nix-defexpr
<samueldr>
the .nix-channels and .nix-defexpr are old things that won't follow XDG guidelines, though .config/nix I think should
<samueldr>
I guess for a new port it would need to be fixed for the platform in defines maybe?
<{^_^}>
[nixos-homepage] @garbas pushed to fix-update-script « nix flake update is now nix flake lock »: https://git.io/JmCgn
<samueldr>
(because there is no backwards compat needed for .nix-channels and .nix-defexpr)
<extrowerk>
Haiku uses XDG env vars, so it should never happen
<extrowerk>
still nix made those files, so it is obviously not respecting the XDG envvars somewhere
<samueldr>
well, for .nix-channels and .nix-defexpr, it's handled separately... but I assumed for .config/nix
fendor has quit [Remote host closed the connection]
<samueldr>
the former .nix-channels and .nix-defexpr *I think* can't be handled through the XDG paths (and is quite an old feature that cannot be broken easily)
<samueldr>
(for existing platforms)
jboy has left #nixos ["WeeChat 3.0"]
cosimone has joined #nixos
fuzzypixelz has joined #nixos
<fuzzypixelz>
if you're familiar with https://github.com/divnix/devos, please explain to me how to use profiles, i.e how to import them into my hosts/users
emmanuel_erc has quit [Read error: Connection reset by peer]
<supersandro2000>
matthewcroughan_: if you read it you might know where it needs to be added
lobre has joined #nixos
<lobre>
Hi, I just switched to the unstable channel of NixOS and I am trying to build with NVIDIA drivers (services.xserver.videoDrivers = [ "nvidia" ];) but it fails with error "The required package gtk+-3.0 was not found on your system". I am not sure if the problem comes from my configuration or if it something wrong with unstable...
<unclechu>
hi. about Rust packages. is there some Nix function for Rust like `pkgs.haskellPackages.callCabal2nix` for Haskell?
<unclechu>
a function that would produce Rust package derivation from a given path?
<talyz>
aanderse: I am now :)
gustavderdrache has joined #nixos
justanotheruser has quit [Quit: WeeChat 2.9]
<unclechu>
or is there a function to read and parse *.toml file? in order to extract some information from it?
<talyz>
unclechu: there is a builtins.fromTOML that does that
<unclechu>
talyz: wow, didn’t expect to find it in `builtins`, thanks
<aanderse>
talyz: ah i had a snippet i use for a few web apps for inbound e-mail
<aanderse>
basically just push the contents of the e-mail to a url via curl
emmanuel_erc has quit [Read error: Connection reset by peer]
<aanderse>
i think discourse can work that way too
emmanuel_erc has joined #nixos
supercoven_ has quit [Ping timeout: 264 seconds]
Baughn has joined #nixos
<aanderse>
talyz: for reference, you should be able to tweak something like this to get inbound email working: `services.postfix.extraAliases = '' test: "|${pkgs.curl}/bin/curl --data-binary @- https://${cfg.hostName}/api/email" '';`
<talyz>
unclechu: no problem :)
justanotheruser has joined #nixos
<talyz>
aanderse: ah, i've actually packaged the upstream solution now and am on my way to try it out
<aanderse>
sounds good
orivej has quit [Ping timeout: 256 seconds]
<talyz>
aanderse: yeah, it should do some additional niceties like reject obvious spam, if it works
<aanderse>
awesome
<aanderse>
maybe one day we self host discourse for nixos ^^
<samueldr>
stable releases don't get major package upgrades
<viric>
matthewcroughan_: well I'm a poor maintainer
orivej has joined #nixos
shibboleth has joined #nixos
<viric>
matthewcroughan_: I used freecad in a period of time but I don't, anymore. I've no idea about that piece you mention. Do you mean it'd be a trivial change, or you tried and failed to?
crenfrow has quit [Remote host closed the connection]
emmanuel_erc has quit [Read error: Connection reset by peer]
emmanuel_erc has joined #nixos
<colemickens>
clever: the command does, or release.nix does?
<clever>
colemickens: the above cmd uses default.nix
<clever>
release.nix is more aimed at running default.nix multiple times, once for every os and cluster combination
<colemickens>
I'm doing the xy thing. Would it make sense for that readme to have recommended usage instructions for nixos users?
<colemickens>
And if so, what would that look like? Presumably I want sandboxing?
<clever>
colemickens: if you want the sandboxing and auto-update, just install it the same way as any other linux user
<clever>
ive made sure the sandbox stuff works on both nixos and normal linux
dbmikus has quit [Ping timeout: 256 seconds]
<clever>
colemickens: the main point of the sandbox, is to let things work the nix way, even if you lack write perms to /nix, and on a nixos system, if you are not a trusted user in nix.conf
ahmed_elgabri has joined #nixos
xelxebar has quit [Remote host closed the connection]
xelxebar has joined #nixos
svrana has quit [Ping timeout: 260 seconds]
orivej has quit [Ping timeout: 245 seconds]
<colemickens>
I was looking for something to put in my configuration.nix. I did run the bin just now and I see it did a bunch of stuff, but then seems to have just exited.
ahmedelgabri has quit [Ping timeout: 244 seconds]
couldshift has joined #nixos
svrana has joined #nixos
<clever>
colemickens: after that bin has finished, you will have a binary in ~/.local/bin and its also in the xfce apps menu
<colemickens>
oh jeez I guess I dont have that on path
Baughn has quit [Ping timeout: 260 seconds]
couldshift has quit [Client Quit]
<colemickens>
and embarrasingly it is in my launcher, ok thanks
<clever>
and then it will just magically update itself when updates are needed
jumper149 has quit [Quit: WeeChat 3.1]
venom17[m] is now known as KiraXD[m]
<colemickens>
I think get why, but the nixos-er in me is a bit uncomfortable
<colemickens>
for example, I want to build a nixos image with some of these types of tools pre-installed for a portable self-contained, replicable hard drive
<clever>
[clever@amd-nixos:~/iohk/daedalus-master]$ nix-build -A daedalus --argstr cluster mainnet
<colemickens>
clever: okay, as you said. understood.
<clever>
you can just do something like `daedalus = import sources.daedalus { cluster = "mainnet"; };` in an overlay, and toss it into systemPackages
<clever>
and then it wont be able to update itself, so you need to use `niv update` and rebuild the whole system
* colemickens
nods
<clever>
adding the iohk hydra to your nix will also speed up the build, directions are in the readme
emmanuel_erc has quit [Read error: Connection reset by peer]
<colemickens>
it imports iohk-nix without passing system, so it fails in strict eval
<colemickens>
unless im missing something again -_-
emmanuel_erc has joined #nixos
<colemickens>
i can hack at it though, maybe send a pr
<colemickens>
thank you for the help!
thc202 has quit [Ping timeout: 240 seconds]
crenfrow has joined #nixos
xwvvvvwx has quit [Ping timeout: 264 seconds]
Baughn has joined #nixos
sigmundv__ has quit [Remote host closed the connection]
Baughn has quit [Client Quit]
Synthetica has quit [Quit: Connection closed for inactivity]