seanparsons has quit [Read error: Connection reset by peer]
<gchristensen>
infinisil: definitely me
<infinisil>
Prepare yourself, as I'm about to unleash the beast
slyfox has quit [Ping timeout: 246 seconds]
* gchristensen
shrinks back
<infinisil>
>r=with lib;with builtins;let p=g: v: if isAttrs v then foldl({g,r}: n: let a=p g v.${n};in{n=a.n;r=r//{${n}=a.r;};}){r={};g=g;}(attrNames v)else if isList v then foldl({g,r}: e: let a=p g e;in{g=a.g;r=r++[a.r];}){r=[];g=g;} v else if isFunction v then let r=v g; in{r=r;g=r;}else{g=g;r=v;}; in v:(p builtins.currentTime v).r;
<{^_^}>
error: syntax error, unexpected ';', at (string):8:329
<infinisil>
>r=with lib;with builtins;let p=g: v: if isAttrs v then foldl({g,r}: n: let a=p g v.${n};in{n=a.n;r=r//{${n}=a.r;};}){r={};g=g;}(attrNames v)else if isList v then foldl({g,r}: e: let a=p g e;in{g=a.g;r=r++[a.r];}){r=[];g=g;} v else if isFunction v then let r=v g; in{r=r;g=r;}else{g=g;r=v;}; in v:(p builtins.currentTime v).r
<{^_^}>
r defined
deepfire has quit [Ping timeout: 260 seconds]
<gchristensen>
is that the gore? I can't tell
<infinisil>
Now what might this function do you ask yourself
<infinisil>
>r=with lib;with builtins;let p=g: v: if isAttrs v then foldl({g,r}: n: let a=p g v.${n};in{g=a.g;r=r//{${n}=a.r;};}){r={};g=g;}(attrNames v)else if isList v then foldl({g,r}: e: let a=p g e;in{g=a.g;r=r++[a.r];}){r=[];g=g;} v else if isFunction v then let r=v g; in{r=r;g=r;}else{g=g;r=v;}; in v:(p builtins.currentTime v).r
Synthetica has quit [Quit: Connection closed for inactivity]
seanparsons has joined #nixos
phreedom has joined #nixos
oida has joined #nixos
shabius has joined #nixos
klntsky has joined #nixos
<achambe>
infinisil: I'm not sure I understand when randomness would be desired in a build process
slyfox_ has quit [Ping timeout: 256 seconds]
<infinisil>
achambe: No? Aw.. you could make it such that your nixos system services each get a new random different port each time you rebuild! Isn't that nice? /s
slyfox has joined #nixos
<achambe>
lol
<symphorien>
Assigning a hourly cron job to a random minute might be legitimate.
<edef>
take it from the hash
<tertle||eltret>
you shouldsa heard the thunder here last night
<judson>
clever - sorry for late response: there were 13 profiles. did a full gc clean, got back 1G.
<judson>
Trying deploy again now.
jrolfs_ has quit [Ping timeout: 256 seconds]
<elvishjerricco>
Is there a nix channel suitable for both nixos and macos?
<samueldr>
the nixos channels
<elvishjerricco>
samueldr: Is there no advantage on macos to be using the nixpkgs-18.03-darwin channel over the nixos-18.03 channel?
<infinisil>
Oh I thought you meant IRC channel
<samueldr>
hmm, I was about to reply and then I realized I don't know one detail
<elvishjerricco>
infinisil: My bad :P
<samueldr>
are there darwin-specific tests in the -darwin channels?
<samueldr>
hmmm, and the nixos-* channel could realistically advance while the darwin channels don't
<samueldr>
(I think)
<elvishjerricco>
Hm. At work we have to pin our nixpkgs down, and we have both linux and macos users. So we kind of need a checkout that's suitable for both.
freeman42x[nix] has quit [Ping timeout: 256 seconds]
<samueldr>
ah, if you pin, you could verify beforehand whether that commit ID works for all channels
<elvishjerricco>
Yea but that doesn't provide a very seamless upgrade path
<samueldr>
though, there may be more efficient ways to get the same info!
<elvishjerricco>
Well, a commit being on nixpkgs-18.03-darwin doesn't necessarily mean that was ever the HEAD on that channel. can only really trust commits that were HEAD at one point. Everything else might just be intermittent states from a merge commit.
<samueldr>
hm, you're right
<clever>
there is a json file somewhere that tracks the head of each channel
<krkini>
How do I one-time install a package from the unstable channel? It seems that adding nixpkgs-unstable to my channel list will cause `nix-env -u` or `nixos-rebuild switch` to grab packages from nixpkgs-unstable as well. Do I have to do `nix-env -i -f https://nixos.org/channels/nixpkgs-unstable/nixexprs.tar.xz packageName` every time?
aisiszane has joined #nixos
<clever>
krkini: what does `sudo nix-channel --list` say?
aisiszane has quit [Read error: Connection reset by peer]
<clever>
krkini: if you `nix-channel --add https://nixos.org/channels/nixpkgs-unstable unstable` then you can just `nix-env -iA unstable.hello` to install the unstable version of hello
<clever>
and `nix-env -iA nixos.hello` to install the stable version
freeman42x]NixOS has quit [Ping timeout: 250 seconds]
<clever>
nixos will always use the channel called nixos
<krkini>
ahh, didn't know that part
<krkini>
thanks!
<clever>
and its best if you always manage the channels as root
<krkini>
what about `nix-env -i packageName` without -A ? Does it just check all channels and find the latest version based on that?
<clever>
it will check all channels, but i dont know what the priority is
<clever>
that command is also a lot slower, because it checks every single package, in every channel
<krkini>
Yes, and someone in here was telling me a while ago that it's best to install software long-term by adding it in /etc/nixos/configuration.nix anyway.
<clever>
another option is to create an attrset or buildEnv in config.nix, and then install that
Havvy has quit [Read error: Connection reset by peer]
nuncanada2 has quit [Quit: Leaving]
<elvishjerricco>
Anyone use bluetooth headsets on nixos? My adapter is kinda bad, and linux seems to be compensating by letting the audio back up when the connection becomes too weak. I'd rather it skip than become delayed. Is this configurable?
<petersjt014[m]>
palo: If nobody else can tell why, you may just want to try it on Impala just to see. It might not make a diff since nixos isn't as susceptible to 'works now because magic lol', but still.
mkoenig has joined #nixos
<jD91mZM2>
mount the ISO*
<petersjt014[m]>
I can try--one sec while I give virtbox usb ports to play with
<palo>
petersjt014[m]: I'm not 100% sure what you say. What is Impala.
Forkk_ has quit [Remote host closed the connection]
FunkyBob has quit [Remote host closed the connection]
<palo>
hmm was just a warning
<palo>
now it starts doing things
<palo>
but it looks disturbing
<palo>
but I've seen it today at somebodys computer as well
<palo>
so I'm not scared :D
rosa has quit [Quit: rosa]
<mog>
i have nixos stable installed and im trying to install the 4.16 kernel from unstable i have the let unstable import in and tried telling boot.kernelPackages to be 4_16 but rebuild made no change
rosa has joined #nixos
<jbboehr>
Does anyone know where the carnix source is? The link on crates.io is a 404.
<jD91mZM2>
srhb: I need to GC less if I want to win that prize you mentioned yesterday. I most likely GC'd yesterday as well, and because of that I only freed 3154.41MiB.
rosa has quit [Quit: rosa]
jperras has quit [Ping timeout: 246 seconds]
schoppenhauer has quit [Ping timeout: 240 seconds]
jb55 has joined #nixos
rosa has joined #nixos
schoppenhauer has joined #nixos
jrolfs_ has joined #nixos
rosa has quit [Client Quit]
<fearlessKim[m]>
after 10mn of compilation I get,`error: cannot create symlink '/home/teto/wireshark/result'; already exists`. Now it starts teh compilation again. Isn't it possible to check for "result" before the compilation ?
rosa has joined #nixos
<jD91mZM2>
fearlessKim[m]: What if the symlink is created during compilation?
<fearlessKim[m]>
jD91mZM2 it could check both or maybe I should alias it to check
<fearlessKim[m]>
beforehand
jrolfs_ has quit [Ping timeout: 240 seconds]
<jD91mZM2>
fearlessKim[m]: It could
<Cale>
rizary: reflex-platform has a bunch of android stuff in it
klntsky has quit [Remote host closed the connection]
<Cale>
(as well as iOS)
klntsky has joined #nixos
<rizary>
Cale: thanks, I thought it is only if we use reflex / reflex-dom.
<rizary>
not in general
<rizary>
thanks anyway, will read that.
<Cale>
Well, you can use it even if you're not using those
<Cale>
There's a bunch of nix stuff in there which solves a lot of the build issues with doing cross compilation
drakonis has quit [Remote host closed the connection]
<Cale>
rizary: Oh, I didn't realise which channel I was in until just now. ;)
<Cale>
Is this for a Haskell project at least? :)
<Cale>
It might be useful regardless.
<Cale>
Ask Sonarpulse about it whenever he's next around :)
andreabedini has quit [Ping timeout: 260 seconds]
<mog>
is anyone running an unstable kernel on a stable nixos? i cant get the setting to take and i am not sure why
andreabedini has joined #nixos
<rizary>
Cale: heh, not it is for java-ish Android project. Im considering using Haskell but I think right now that decision is still not possible. Thanks again Cale.
timclassic has left #nixos ["User left"]
nallar has joined #nixos
Ross has quit [Ping timeout: 255 seconds]
nallar is now known as Ross
krey has quit [Ping timeout: 256 seconds]
iqubic has joined #nixos
timclassic has joined #nixos
rosa has quit [Quit: rosa]
<petersjt014[m]>
If I'm using a motd and a window manager, will the motd be visible?
<jD91mZM2>
Does a block-based full disk encryption have any implications in terms of resistance? What happens if the computer loses power during a write?
<jD91mZM2>
iqubic: Message of the day. It's what shows the version number in a TTY.
<petersjt014[m]>
I thought that was the /etc/issue file?
<jD91mZM2>
Oh, whoops
<iqubic>
petersjt014[m]: What behavior are you looking for?
<jD91mZM2>
Oh yeah MOTD is when you log in with SSH, isn't it?
<iqubic>
And it will be visible even if using a WM. I see it when I drop into a TTY.
<jD91mZM2>
petersjt014[m]: Anyway, a MOTD is always in the shell, right? So if you're using X at all it shouldn't show(?)
<jD91mZM2>
Sorry if I'm stupid, I haven't messed around with motd
<petersjt014[m]>
I suppose so, yeah. I think I'll need to make a window with wmutil or something
<petersjt014[m]>
and open a txt file or something
<sphalerite>
it's definitely shown on SSH logins, I'm not sure about any other places
<sphalerite>
jD91mZM2: what do you mean by "resistance"?
<jD91mZM2>
sphalerite: Ignore my message, I didn't read enough about blocks
<jD91mZM2>
I was thinking "oh noEs WhaT iF wrITE FaiL Bc pOWer FailurE", without realizing that it could still read the filesystem journal from another block
tmaekawa has joined #nixos
iqubic` has joined #nixos
iqubic` has left #nixos [#nixos]
rosa has quit [Quit: rosa]
<petersjt014[m]>
make everything atomic and ur good 2 go
rosa has joined #nixos
<petersjt014[m]>
right down to the movment of the electrons
tmaekawa has quit [Client Quit]
iqubic has quit [Ping timeout: 246 seconds]
<petersjt014[m]>
we can do that right??
tmaekawa has joined #nixos
andreabedini has quit [Ping timeout: 264 seconds]
<petersjt014[m]>
I kid, I kid
<{^_^}>
[nixpkgs] @mmlb opened pull request #40202 → Netperf love → https://git.io/vpiQ8
orivej_ has quit [Ping timeout: 250 seconds]
<jD91mZM2>
Can I label filesystems by just changing fileSystms."/".label?
<clever>
jD91mZM2: it also auto-configures to mount via /dev/disk/by-label/${label}
<clever>
and i dont see any signs of it being used in autoFormat
<jD91mZM2>
clever: Oh okay. So I either use `device` or `label`?
<clever>
yeah
<clever>
and autoFormat likely wont find it if its not yet formatted with the given label
<jD91mZM2>
So if I just want my labels to show up nicely in gparted or whatever I can just use e2label?
<achambe>
yes
<achambe>
I usually just manually do my labels
<achambe>
but use the nix config to configure how they are mounted
<jD91mZM2>
Alright, thanks!
<jD91mZM2>
Does e2label have to be used on an unmounted system?
<achambe>
just try it, don't be a wimp
<jD91mZM2>
I mean yeah it should probably print an error if it fails
<jD91mZM2>
Seems to have worked... yay! Thanks!
jb55 has quit [Ping timeout: 246 seconds]
jgertm has joined #nixos
spear2 has quit [Remote host closed the connection]
drag has quit [Quit: Connection closed for inactivity]
MercurialAlchemi has joined #nixos
<jD91mZM2>
Is it recommended to use labels over UUIDs? The installation wiki says "this makes the file system configuration independent from device changes"
<jD91mZM2>
I mean it says right there that it's recommend. But... but...
jperras has joined #nixos
<Yaniel>
depends on your use case
robstr has joined #nixos
c0ffee152 has quit [Quit: Lost terminal]
<jD91mZM2>
Yaniel: I mean for usage inside hardware-configuration.nix
<jD91mZM2>
But then again I shouldn't edit that
Guanin has quit [Ping timeout: 246 seconds]
<Yaniel>
yeah well are you planning to swap drives or partition layouts often?
<jD91mZM2>
not really lol, I just wanted labels to show up in gparted
<Yaniel>
and yeah, if you want to mount a disk by label you should probably put it in configuration.nix and make sure it's absent from hardware-configuration.nix
<jD91mZM2>
Seems like nixos-generate-config doesn't use the labels by default. Interesting, since the wiki says it's recommended to set them
<Yaniel>
you set them when partitioning the disk, not in nixos config
jperras has quit [Ping timeout: 260 seconds]
<jD91mZM2>
Yaniel: Yeah but the wiki makes it sound like NixOS would use the labels as identifiers
<Yaniel>
and yes it is useful to set them because for most people it's a lot faster to recognize which disk "Carl" is than "AEB634BD"
<jD91mZM2>
btw when running `nixos-generate-config` from inside the system, it removes "usb_storage" from availableKernelModules
tmaekawa has quit [Quit: tmaekawa]
mahalel_ has quit [Ping timeout: 246 seconds]
endformationage has quit [Quit: WeeChat 1.9.1]
Guanin has joined #nixos
smichel has joined #nixos
rosa has quit [Quit: rosa]
tmaekawa has joined #nixos
rosa has joined #nixos
andreabedini has joined #nixos
xy2_ has joined #nixos
jD91mZM2 has quit [Quit: Rebooting]
<ghostyy>
hi i am trying to configure strongswan on nixos and when i try to define a key %default it complains about error: syntax error, unexpected $undefined,
<elvishjerricco>
O.o why do you have to build perl
<angerman>
elvishjerricco: part of the downstream packages
jeeves is now known as jeaye
<angerman>
it's building perl and ghc in parallel. perl just fails faster
<elvishjerricco>
Oh hm.
<angerman>
elvishjerricco: the `no-sys-dirs.patch` does not apply cleanly onto perl-5.24.3, locally... now that makes absolutely no sense. But for some reason the unpacked perl-5.24.3, ends up having the Configure file from a more recent perl-5.26 or so.
tmaekawa has quit [Quit: tmaekawa]
<elvishjerricco>
angerman: That's confusing. So `nix-build -A whaterver.perl` patches fine, but `nix-shell -A whatever.perl --run 'unpackPhase && cd perl* && patchPhase'` does not patch cleanly?
Ariakenom has joined #nixos
<angerman>
maybe I should try to see what the unpack phase says.
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vpiFB
<{^_^}>
→ b45fa92d by @rkoe: ext4magic: init at 0.3.2
<{^_^}>
→ 04814672 by @matthewbauer: Merge pull request #40148 from rkoe/add/ext4magic
<angerman>
sir_guy_carleton: use the --out-link ?
<clever>
angerman: the desktop puts out a decent amount of heat, and i keep the thermostat a tad high
<angerman>
sir_guy_carleton: or in a nixexpr "${pkg}/..."?
<sir_guy_carleton>
with nix-store?
<teto1>
When in a systemd service, I put Restart="always", the service should restart on rebuild right ? Looks like it doesn't for my telnet service so I can't test why CapabilityBoundingSet = "CAP_NET_BIND_SERVICE=+eip"; seems uneffective :
marusich has joined #nixos
<clever>
copying path '/nix/store/9hmhxgj4jk6jmxihgavj6gm0p759misc-cmake-3.10.2' to 'ssh://builder@amd.localnet'...
<clever>
angerman: it is now running cmake on rocksdb
<clever>
x86_64-pc-mingw32-gcc: error: unrecognized command line option '-rdynamic'
Guanin_ has joined #nixos
<sphalerite>
sir_guy_carleton: typically you'd use it through a profile, so something like /run/current-system/sw/bin/foo or ~/.nix-profile/bin/foo. Although those are on your PATH by default so it shouldn't be necessary to use the full path
Guanin has quit [Read error: Connection reset by peer]
<sir_guy_carleton>
i trying to write the path to another program in a configuration file in my home directory
alex`` has joined #nixos
<angerman>
clever: cross compiling is so much fun >D
<sphalerite>
sir_guy_carleton: yeah then use one of the paths I mentioned above. Or generate the config file using nix as well
<teto1>
clever: regarding my keyboard not working in boot.debug1 yesterday, sphalerite suggested to run boot.debug1devices instead and that fixed it. Thanks for the help once again
<sphalerite>
\o/
<clever>
teto1: oh, debug1 stops too soon, and you lack drivers
Arcaelyx has joined #nixos
hyper_ch2 has joined #nixos
toby1851 has joined #nixos
doyougnu has quit [Ping timeout: 255 seconds]
<teto1>
clever: I had the drivers in intramfs though
Arcaelyx_ has quit [Ping timeout: 246 seconds]
<jD91mZM2>
sir_guy_carleton: `nix-build '<nixpkgs>' -A hello --no-out-link` could work in the worst case scenario
<jD91mZM2>
This prints the nix store path to hello
<sphalerite>
teto1: they don't actually get loaded immediately though, debug1devices makes it stop only once they have bene loaded
<sphalerite>
sir_guy_carleton: jD91mZM2: this isn't a good path to use though because it might get GC'd away since nix doesn't know that something is depending on it
<pacmeister>
any advice on getting appimages to work?
simukis has joined #nixos
<sphalerite>
pacmeister: FHS user envs are useful for that sort of thing
<jD91mZM2>
sphalerite: I mean he said "path to a program" so it's not gonna get GC'd (assuming explicitly installed). Still not a good path to use, but that's why I said "worst case scenario".
<sphalerite>
hm, so root's NIX_PATH is the default nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels and /nix/var/nix/profiles/per-user/root/channels does contain unstable. But nix build unstable.procps says error: attribute 'unstable' in selection path 'unstable.procps' not found
<sphalerite>
any ideas?
<angerman>
clever: thanks! "make rocksdb use cmake (doesnt work)" :D
<clever>
angerman: all it does is switch to cmake, g++ then fails due to args cmake gave it
<sphalerite>
jD91mZM2: even if it's installed, it will change as soon as hello (or one of its deps) is upgraded, and once the old generation has been deleted it's fair game for the gc
<clever>
angerman: it also includes an lz4 fix
<angerman>
clever: saw that, thanks, will integrate, and merge with my stuff.
johann__ has joined #nixos
<jD91mZM2>
sphalerite: sir_guy_carleton: If it's installed system-wide it shouldn't upgrade without also updating nix-build tho?
<sphalerite>
jD91mZM2: yes, but the config file will still be referrign to the old path
<sphalerite>
so you need to either build it with nix, or point it to the program indirectly through a profile which is a GC root and will be kept up to date
<jD91mZM2>
sphalerite: Oh, I didn't realize he was writing to a config
<sphalerite>
sir_guy_carleton | i trying to write the path to another program in a configuration file in my home directory
<angerman>
elvishjerricco: now you go and fix it, so I can use it when I come back :p
<elvishjerricco>
angerman: Testing now :P
<elvishjerricco>
Seems to have fixed the issue! Now to see if there are further issues with android...
jperras has quit [Ping timeout: 240 seconds]
<angerman>
elvishjerricco: ha, I might possibly able to help with some of them...
tilpner__ has quit [Ping timeout: 240 seconds]
<angerman>
elvishjerricco: tritlo discovered that calling tzset from haskell on RPi triggers a segfault.
<angerman>
elvishjerricco: as such you can't use network ,... unless you provide TZ="UTC0" or something :D
<elvishjerricco>
eek. Is that only on RPi?
<angerman>
elvishjerricco: you may really be interested in using my stackage/hackage/haskell expressions for cross compilation thou.
<elvishjerricco>
angerman: I absolutely am. That's why I want that infra in nixpkgs! I'm tired of having to tell people to check out some *other* Nix repo (reflex-platform) to do cool stuff :P
allana has joined #nixos
<jD91mZM2>
Are there any permanent problems with changing /etc/resolv.conf? I know that it's not the nix way, but I need something quick and temporary until the nix way supports all the things I need it to support.
<jD91mZM2>
Like, if I change it now and then run nixos-rebuild, it'll still be overriden right? (which is good, so no changes outside of the proper nix ones are permanent)
<elvishjerricco>
angerman: I'm even interested in it just for GHCJS. Can't tell you how many times builds break because people don't realize `cabal2nix` doesn't translate `if impl(ghcjs)`...
<angerman>
elvishjerricco: I might turn my llvmng stuff into a separate patch for ghc8.4 as well.
<allana>
Hi, I am struggling through creating a custom build script for a derivation that requires some inputs that I am having a hard time finding. For example, my build script requires the "hostname" command-line tool but I don't know which nixpkg to declare as an input for that tool. Can anyone help me?
<elvishjerricco>
angerman: That would be nice. Improvements would be in compile time and LLVM compatibility, right?
<clever>
allana: nix sets the hostname to localhost inside builds, so even if you had the tool, it wont give the expected values
<clever>
allana: you probably want builtins.getEnv "HOSTNAME" which reads the env var at eval-time
butchery has joined #nixos
<elvishjerricco>
clever: I think allana is just saying that the build script for some existing package expects the `hostname` cli tool. Don't think they actually care to provide the build machine's real hostname
sir_guy_carleton has quit [Quit: Lost terminal]
<clever>
allana: its in the package called net-tools, add that to buildInputs
<elvishjerricco>
nativeBuildInputs, right? :P
<allana>
clever: and elvishjerricco: Thanks!
<clever>
thats better
<butchery>
is there an entirely unattended nixos installer, or a way to enable ssh on the installer with no input? I have a device that has a single USB port and no other inputs, and I have no USB hub on hand
<angerman>
elvishjerricco: yea, compiling will take a lot longer ;)
<elvishjerricco>
angerman: Oh? I thought llvmng was supposed to be faster?
xAFFE has joined #nixos
<clever>
butchery: using a second machine (even a fully working nixos), you can run nixos-install against a usb stick, and use whatever config you want, and setup sshd and passwords
reinzelmann has quit [Quit: Leaving]
<clever>
butchery: then you will have a custom USB that boots into that config on any machine
<angerman>
It generates better code but suffers from a space leak
<elvishjerricco>
angerman: Ah. Fixable, I hope!
reinzelmann has joined #nixos
<butchery>
ah, I see, is anything else needed other than --root and --system?
<angerman>
elvishjerricco: yea. It uses some neat mfix
<elvishjerricco>
angerman: Always love a good mfix :)
datakurre has joined #nixos
<clever>
butchery: should be as simple as formatting the USB stick as normal, mounting it to /mnt, and then doing nixos-generate-config --root /mnt and nixos-install --root /mnt/
Xal has quit [Ping timeout: 256 seconds]
<angerman>
elvishjerricco: it allows to compute some stupid sizes into the stream in a principled fashion. But that leads to really bad retention of thunks until the size has been computed.
Tucky has joined #nixos
<angerman>
elvishjerricco: I think that just inserting placeholder into the stream and filling them after the fact will perform better.
<angerman>
elvishjerricco: I just haven’t gotten around to doing that yet.
<{^_^}>
→ 6e7f4f99 by @nh2: manual: Document `types.attrs`
<{^_^}>
→ 85854ab2 by @magnetophon: vamp-sdk: 2.5 -> 2.7.1
<{^_^}>
→ 5855459f by @flokli: modules/system/boot/luksroot: remove comment about input_leds for caps lock
<angerman>
elvishjerricco: ha ;)
<steveeJ>
rycee: I've switch from nixos managed to home-managed zsh and now my zsh lacks the functions "git_prompt_info" and "prompt_char", which are in my PROMPT. do you know where to get them?
<elvishjerricco>
angerman: I'd also like to fix the enableShared thing but have no idea how
<angerman>
elvishjerricco: I don’t follow.
<elvishjerricco>
angerman: in that PR, I have `enableShared` disabled only because of a forbidden RPATH if you enable it. Not sure if it's android specific, and not sure why it's happening
<angerman>
elvishjerricco: I have some modifications in my ghc-8.4.2 branch to make he expr Match head closer.
<angerman>
elvishjerricco: well I built static almost always for cross compiler, I can hardly see any benefit of building dynamic.
<NinjaTrappeur>
Hey, I have a really strange nix-shell behaviour: regardless where I am trying to enter a nix-shell env, I have this error message: "error: getting status of '/home/minoulefou/Code/perso/haskell/nixpkgs-stackage/default.nix': No such file or directory". Thing is, this folder no longer exists, I don't understand why nix-shell is trying to access this file
<NinjaTrappeur>
Specifying the correct shell.nix file in argv[1] does not change anything. Does anybody already experienced this? This is really confusing
<NinjaTrappeur>
Godammit, it was an overlay... Sorry for the noise
orivej has joined #nixos
andreabedini has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mipmop has quit [Quit: Konversation terminated!]
LysergicDreams has quit [Ping timeout: 264 seconds]
jrolfs has joined #nixos
spear2 has quit [Quit: Leaving]
humanoyd has joined #nixos
<fearlessKim[m]>
overlays can mess up things in subtle (not here) ways (especially as one tends to forget about them)
asuryawanshi has joined #nixos
LysergicDreams has joined #nixos
asuryawanshi has quit [Remote host closed the connection]
<cheater>
ben: yeah, TMPDIR. but why is this set to /run/user/1000 (my uid)? is this normal for nix shell, or is this something weird in this project?
<silver_hook>
My current experiments with the /etc/nixos/sane-extra-config.nix were unsuccessful.
Dezgeg has joined #nixos
fendor_ has joined #nixos
<jD91mZM2>
cheater: That's weird, everything works like expected on my system. What version of NixOS are you running?
<gchristensen>
cheater: that is normal on NixOS/systemd I think
<symphorien>
silver_hook: there is a section about sane and networking on the wiki, maybe it can help
<ben>
Yeah I think it's systemd or something
<ben>
I kinda thought /tmp was tmpfs too tho
<ben>
probably easier for user quotas or something
<ben>
or user privacy :V
Maxdamantus has quit [Ping timeout: 256 seconds]
<cheater>
jD91mZM2: recent.
Maxdamantus has joined #nixos
<cheater>
jD91mZM2: 2.0
<cheater>
i have no idea who came up with this, but
<cheater>
it's pretty bad
<silver_hook>
symphorien: That’s what I did before. I’ll see if I can get it working this time :)
<cheater>
it makes stuff fail that shouldn't fail, for no good reason that i can immediately see. if there is a good reason for making mktemp return a place in /run, i'd love to hear about it
<cheater>
in my case: allocating a small (several MB) file in a /run that has 350 MB free fails with "out of memory" due to some stupid stuff related to /run
bbarker has quit [Remote host closed the connection]
allana` has joined #nixos
allana has quit [Ping timeout: 240 seconds]
ent has joined #nixos
rosa has quit [Quit: rosa]
rosa has joined #nixos
<Myrl-saki>
How do I callPackage with clang?
<silver_hook>
symphorien: So far no luck, with or without the sane-extra-configuration.nix. The only device I find (over USB or WiFi) is the built-in laptop camera.
<clever>
Myrl-saki: clangStdenv.mkDerivation ?
<Myrl-saki>
clever: Okay, so in my case, `... callPackage ./. {stdenv = ...}`?
<Myrl-saki>
Rather
<Myrl-saki>
{ stdenv = clangStdenv; }`
<clever>
Myrl-saki: if you want to force something to switch to clang, yeah, that should just work
<clever>
or .override as well if its already been callPackage'd
<silver_hook>
symphorien: On my work KUbuntu laptop I just got it to work, but the SANE driver is called “epson2”, which doesn’t seem like it wants to be passed to NixOS.
<Myrl-saki>
clever: Oh yeah. That's what I was gonna talk about.
<Myrl-saki>
clever: callPackage and .override seems to have different semantics.
<silver_hook>
At least: `hardware.sane.extraBackends = [ pkgs.epson2 ];` throws an error
<Myrl-saki>
clever: .override seems to only override the package itself, and callPackage seems to be a propagating "override"
<silver_hook>
makefu: That’s what I’m looking at, yes.
<makefu>
ah
<clever>
Myrl-saki: .override can only change the arguments passed into the file, the same ones callPackage overrides
<makefu>
this was working for me at least at the time i wrote it
<clever>
Myrl-saki: .overrideAttrs will override the attrset passed to mkDerivation
<clever>
Myrl-saki: and .overrideDerivation overrides it deeper, after mkDerivation has re-arranged some things, when its about to call something internal
<Myrl-saki>
clever: Oh okay.
<Myrl-saki>
Thanks, that's clearer now.
winem_ has joined #nixos
LysergicDreams has quit [Remote host closed the connection]
<silver_hook>
`imagescan -L` on KUbuntu shows this:
<silver_hook>
device `epson2:net:192.168.0.188' is a Epson PID 111F flatbed scanner
<silver_hook>
device `imagescan:esci:usb:/sys/devices/pci0000:00/0000:00:14.0/usb1/1-3/1-3:1.0' is a EPSON L386_Series
<silver_hook>
What works (i.e. builds) are “pkgs.epkowa“ and “pkgs.epson-escp”. But even then `sudo scanimage -L` doesn’t find any scanner.
<goibhniu>
silver_hook: silly question ... did you try disabling your firewall?
<silver_hook>
goibhniu: On router or in NixOS?
<goibhniu>
on NixOS
<silver_hook>
goibhniu: No, I didn’t know how to.
<clever>
silver_hook: checking the source of mkSaneConfig, it looks like all valid entries in extraBackends must contain either a lib/sane or an etc/sane.d directory
<jD91mZM2>
CrazedProgrammer: I'm not sure, but I think you want to base it on the master branch
<silver_hook>
goibhniu: But even with the firewall up, it should detect it through USB. (I’d still like to learn more about the firewall settings, since I want to get KDE Connect working).
<clever>
silver_hook: epson2 isnt even a valid package
<clever>
silver_hook: same for imagescan
<sphalerite>
CrazedProgrammer: I'd go with the latest release, and if they make a point release for the fix update to that
<silver_hook>
clever: It’s not a valid package, but that’s what the backend is actually callde.
<sphalerite>
jD91mZM2: I think CrazedProgrammer means upstream versions, not nixpkgs versions
<clever>
silver_hook: the backend may need to be packaged first
<goibhniu>
silver_hook: you might need firmware too
<silver_hook>
clever: `locate epson2` does show up quite a few hits.
<jD91mZM2>
sphalerite: oh
ericsagnes has joined #nixos
joehh has joined #nixos
<silver_hook>
clever, goibhniu: epson2 drivers are in sane-backends and present on disk
<steveeJ>
any home-manager users here that could tell me how to let it configure the default shell for a user?
<jD91mZM2>
steveeJ: You set default shell from the normal configuration
<steveeJ>
jD91mZM2: that's what I'm trying to avoid
<steveeJ>
I even manage the X session through home-manager
<jD91mZM2>
Oh. Well chsh requires sudo. For a workaround, look at whatever terminal emulator you're using, or maybe possibly `exec <shell>` from your .bashrc
<etu>
jD91mZM2: chsh requires you to enter your password :)
<jD91mZM2>
etu: Ok maybe not but it does change it system-wide
<jD91mZM2>
or on a system level, rather
<etu>
Yes, but only for one user
<etu>
So not system-wide
<steveeJ>
etu: I would guess so; I should've asked, how does it do that
<jD91mZM2>
I'm not exactly sure how it does that behind the scenes, but it's probably extending it's own privileges somehow
<jD91mZM2>
its*
jperras has joined #nixos
<{^_^}>
[nixpkgs] @CrazedProgrammer opened pull request #40217 → compsize: init at 2018-04-07 → https://git.io/vpPnF
smichel has quit [Quit: smichel]
xcmw has joined #nixos
<etu>
jD91mZM2: yeah, probably some capabilities like sudo, but you don't put sudo in front like you said :p
smichel has joined #nixos
<silver_hook>
Woohoo! Works now :D
<silver_hook>
It was the firewall, heh.
rosa has quit [Quit: rosa]
<makefu>
neat!
<jD91mZM2>
By the way, I think NixOS needs some kind of layered trust system. Something like having a certain amount of semi-trusted people vouch for a PR to have it automatically merged
<jD91mZM2>
Because there are currently 647 open PRs and 2 431 issues and I think that's a little insane
<silver_hook>
Now to only figure out how to properly tweak the firewall, so it doesn’t have to be off for cool stuff to work.
<CrazedProgrammer>
spheralite: sorry, i saw your message after writing the PR. i do think the fix is imporant for sshfs users since it would error if it encountered an sshfs mount, which is often in a users' home directory.
<silver_hook>
clever: Thanks :)
<sphalerite>
jD91mZM2: etu: usually chsh is setuid root
<sphalerite>
jD91mZM2: as for auto-merging, it's on the roadmap for borg
<jluttine>
goibhniu: just fyi, in case you're interested, fixed the xcb issue by adding xlibs.libX11 and xlibs.libXi to the environment
tertle||eltret has quit [Quit: Connection closed for inactivity]
rosa has joined #nixos
<allana`>
I am trying to create a derivation for a 3rd-party propietary software suite. The source is fetched using fetchurl and there are a series of scripts that need to be run that have a #!/bin/bash shebang which cause problems during the build phase. I'm curious to know if anyone else has had this problem and how they chose to address the issue. Is there a way to "ignore" these rigid shebangs? I would like to avoid modifying the underlying
<allana`>
source files to suit the needs of the derivation build process.
<clever>
sphalerite: my dhcp config detects that ipxe is working, and switches it over to an http based script
rosa has quit [Quit: rosa]
<sphalerite>
clever: yeah I've got that working, the problem is the kernel panic
<clever>
sphalerite: and line 46 gives it a relative path, so ipxe continues to fetch those files over http
<clever>
ah, *looks closer*
ericsagnes has quit [Ping timeout: 250 seconds]
<goibhniu>
silver_hook: nice! I think it's enough to do firewall.allowedTCPPortRanges = [{ from = 1714; to = 1764; }]; for KDE connect
<silver_hook>
goibhniu: Cool, I’ll try that. Any idea which ports need to be opened for SANE?
<goibhniu>
no idea, sorry
<sphalerite>
clever: and booting these same files with the same kernel parameters from grub works fine :/
<silver_hook>
goibhniu: I’ll look it up. Thanks for the help :D
knupfer has joined #nixos
<clever>
sphalerite: the ipxe embeded script says to run boot.ipxe, which doesnt exist in your tftproot derivation, and you also have grub config mixed in
<sphalerite>
clever: the actual tftp root is a stateful forest of symlinks so I can switch between ipxe and grub more easily
<clever>
yorick: i believe the ttl is only an hour
<yorick>
clever: okay, but if I update a channel and then do nix-channel --update, I kinda expect the channel to be updated
<sphalerite>
clever: the boot.ipxe I use just provides a menu, much like the grub one, which chainloads boot-x86_64.ipxe or boot-i686.ipxe as appropriate
<clever>
yorick: update the channel url, or change a channel you are hosting?
<yorick>
clever: change a channel I'm hosting
<sphalerite>
clever: the grub config isn't mixed in as such, it's in a completely separate file which GRUB loads.
<clever>
yorick: ah, if you blow away ~/.cache/nix/tarballs then nix will forget about all cached downloads
<clever>
sphalerite: id try adding boot.debug1mounts to the kernel params, and then confirm that $stage2Init exists within /mnt-root/
joehh has quit [Ping timeout: 250 seconds]
<yorick>
clever: I've settled for --tarball-ttl 0
<sphalerite>
ok, I'll try that next time I'm there. Thanks!
<yorick>
clever: but I'm using that cache in other places
<sphalerite>
clever: I also think it's odd that it doesn't finish printing the error message before exiting and the kernel dumps its stacktrace
<clever>
sphalerite: yeah, that is a bit weird
<clever>
sphalerite: how did you capture the console output?
<clever>
sphalerite: i think this can greatly slow down the backtrace as it panics
<Dezgeg>
maybe your initrd or squashfs filesystem is bad?
<sphalerite>
Dezgeg: nope, it boots fine with GRUB
<Dezgeg>
I'd add a sleep to the "can't execute ..." line
<Dezgeg>
to see the error
<sphalerite>
clever: I'll try that too, although I'm not sure that'll affect it?
<Dezgeg>
maybe boost loglevel up as well
<clever>
Dezgeg: yeah, i was also thinking of a sleep in stage-1-init.sh
<Myrl-saki>
I need help. GCC and Clang seems to have different `ld` mechanics. I get `undefined reference to `__libc_csu_fini'` from `clangStdenv`, but not from `stdenv`.
<jD91mZM2>
sphalerite: It's usually up to either the .zprofile/.bash_profile or in my case Display Manager to source the .profile
<jD91mZM2>
The profile is sourced. The problem is that everything is then overridden because /etc/zshenv gets executed every time you start a new shell
<jD91mZM2>
which is completely stupid
<steveeJ>
jD91mZM2: you're right that it's a lot slower when the option is enabled
<sphalerite>
jD91mZM2: ah yes, that stuff shouldn't be in zshenv then… or at least it should be set like : ${EDITOR:=nano} or something
<LnL>
jD91mZM2: yeah I don't think it looks at .profile, zshenv -> zprofile -> zshrc -> zlogin
Neo--- has joined #nixos
<jD91mZM2>
sphalerite: Eeeeeeexactly
<sphalerite>
jD91mZM2: PRs welcome ;)
<jD91mZM2>
Related: nixpkgs#33219, home-manager#183, probably more
LysergicDreams has quit [Ping timeout: 240 seconds]
<sphalerite>
infinisil: {^_^} should totally make links out of those :p
LysergicDreams has joined #nixos
<betaboon>
hello #nixos, i have a question regarding single-user-mode nix installations (eg on ubuntu) how do i configure remote-builds if there is no nix-daemon ?
<infinisil>
Hmm..
justan0theruser has quit [Ping timeout: 256 seconds]
<steveeJ>
jD91mZM2: without zsh enable in the global config I still get these environment variables, that's interesting
<infinisil>
Nah, I think it failed because the home-manager one wasn't valid
<jD91mZM2>
steveeJ: Seems like it didn't get updated
<jD91mZM2>
infinisil: ah
<steveeJ>
jD91mZM2: old terminal, whops. a new terminal has it unset
<jD91mZM2>
Ok so /etc/environment is where Arch and Ubuntu and other distros put their stuff. So globally using PAM instead of globally using the respective shells.
<steveeJ>
jD91mZM2: but e.g. PYTHONDONTWRITEBYTECODE is 1
<jD91mZM2>
Would it be a good idea to make NixOS also use /etc/environment?
<betaboon>
sphalerite: thanks for confirming this, i just figured it out with running strace nix-build, i was just confused that there was no /etc/nix/nix.conf on the system
<sphalerite>
betaboon: yeah it doesn't create one automatically, but it does use it. You may want to use ~/.config/nix/nix.conf instead though if you're not running as root
<sphalerite>
infinisil: :D
<{^_^}>
[nixpkgs] @limeytexan opened pull request #40220 → perlPackages.LinuxACL: init at 0.05 → https://git.io/vpP4x
<betaboon>
sphalerite: can i somehow enable remote-builders via env-vars aka on the fly (on multi-user and single-user) ?
<betaboon>
sphalerite: thats a nix2.0 option right ?
<sphalerite>
betaboon: yes
<sphalerite>
infinisil: but it doesn't print the second one!
<infinisil>
Hmm, should i allow it to work for all github projects Pr's?
MercurialAlchemi has quit [Ping timeout: 246 seconds]
<infinisil>
Ah maybe allowing a couple ones like hm, and other nix related ones
<sphalerite>
yeah I don't think we need all projects, but nix-related ones definitely
<sphalerite>
maybe use the same list as gchristensen's hound search
<betaboon>
sphalerite: using the mentioned command doesn't seem to result in building on the remote machine :(
acarrico has quit [Ping timeout: 246 seconds]
<sphalerite>
betaboon: does it work if you set it in ~/.config/nix/nix.conf? Or is it maybe using nix 1.11.x?
<betaboon>
sphalerite: i have nixos18.03 installed on my machine, but I'm running stuff with pinned nixpkgsChannel, which might result in it using nix 1.11.x, gonna have to check later
<sphalerite>
I want to delete some old generations and still be able to roll back to them. I've copied them to a binary cache, so I can get them back using nix-store -r /nix/store/hash-nixos-system-…, but how can I get the system to that generation after getting it back in the store?
<sphalerite>
Would just `nix-env -p /nix/var/nix/profiles/system --set <path> && /nix/var/nix/profiles/system/switch-to-configuration switch` be the right thing?
<clever>
sphalerite: that sounds like it would work
<clever>
sphalerite: i think nixops does something very similar
seanparsons has quit [Read error: Connection reset by peer]
<clever>
ah, thats the local rollbacks though
<sphalerite>
cool, thanks
<clever>
i think line 696 is when it does it to a remote system
<clever>
still pretty much identical to what you typed
<{^_^}>
[nix] @yorickvP opened pull request #2145 → make sure not to use cached channels for nix-channel --update → https://git.io/vpPu7
<sphalerite>
I'm guessing the only thing that will be different from a normal rollback where it hasn't been GCd is that the generation number will increase rather than decreasing?
<clever>
yep
<sphalerite>
I have `while sleep 60m ; do nix copy --from ssh-ng://bob@192.168.99.13 --to file://$HOME/www/cache --all; done` running on a server that's got plenty of space (192.168.99.13 is my chromebook which does not have a lot of space), it's pretty convenient
<sphalerite>
so I can gc without losing valuable build products
<sphalerite>
I should probably make it into a systemd timer or something
asuryawanshi has quit [Ping timeout: 255 seconds]
rosa has quit [Quit: rosa]
<Dezgeg>
huh, there's nix-copy --all... I should switch to that
rosa has joined #nixos
seanparsons has joined #nixos
<sphalerite>
Dezgeg: yeah, also works for at least path-info and sign-paths
<Dezgeg>
sounds much nicer than the current ssh jetson 'comm -23 <(echo /nix/store/* | tr " " "\n" | grep -v '\.lock$') <(nix-store --check-validity --print-invalid /nix/store/*)' I have x)
<sphalerite>
:D
<sphalerite>
it doesn't copy .drvs though
<Dezgeg>
I don't think that's a problem
rosa has quit [Client Quit]
shabius has quit [Quit: Leaving]
MercurialAlchemi has quit [Ping timeout: 260 seconds]
shabius has joined #nixos
cheshircat_ has quit [Ping timeout: 240 seconds]
cheshircat has joined #nixos
shabius has quit [Remote host closed the connection]
<{^_^}>
[nix] @edolstra pushed to master « In restricted eval mode, allow access to the closure of store paths »: https://git.io/vpPam
<betaboon>
sphalerite: on the topic of remote-builds and nixops, i now tested `nixops deploy --option builders "ssh-ng://some-host x86_64-linux" --option cores 0 --options max-jobs 0" which results in "error: unable to start any build; either increase '--max-jobs' or enable remote builds". which makes me believe that just setting "builders" isnt sufficient oO
<sphalerite>
betaboon: does it work if you set it in ~/.config/nix/nix.conf?
acarrico has joined #nixos
<GlennS>
Anyone have any experience with what you have to do to get a jupyter ipython kernel set up on Nix? I've installed jupyter, jupyter_core, jupyter_client, ipykernel and ipython, but when I run `jupyter kernelspec list` it can't find any kernels.
<betaboon>
sphalerite: when setting `nix.distributedBuilds = true;` and `nix.buildMachines ...` it tries to connect to the remote builder.
stumble has joined #nixos
<sphalerite>
betaboon: I thought you were using a single-user setup on non-nixos!?
<betaboon>
sphalerite: I'm using nixos18.03 on my machine, my colleague is using nix on ubuntu ;)
<betaboon>
sorry for being unclear about that
<betaboon>
I'm trying with ~/.config/nix/nix.conf aswell now tho
<sphalerite>
GlennS: not sure how to set it up with installing into a user env, but it works in nix-shell
<GlennS>
mmm
<GlennS>
I'm trying to get it working with org-babel
stumble has quit [Client Quit]
erasmas has joined #nixos
<betaboon>
sphalerite: with setting `builders = "ssh-ng://some-hoste x86-64"` in ~/.config/nix/nix.conf it doesn't work either
<sphalerite>
betaboon: the system string should be x86_64-linux
<GlennS>
may just have fixed it actually...
<sphalerite>
GlennS: installing python.withPackages (ps: with ps; [ipython jupyter]) might work too
<GlennS>
yeah I think it's either that or ipkernel instead of ipython
<GlennS>
just testing now
<sphalerite>
actually even just plain jupyter finds the python3 kernel
<bebarker>
Interesting, when I do nix-env --set foo.drv (generated from nix-instantiate foo.nix), the nix-env commands are no longer on my path
freeman42x]NixOS has joined #nixos
<betaboon>
sphalerite: ah yeah i used x86_64-linux, sorry for the typo
<clever>
betaboon: was nix in the storepath that foo.nix builds?
<sphalerite>
bebarker: it sets your profile to the result of building foo.drv, so *only* that will be in the resulting profile
<sphalerite>
bebarker: it's like "remove everything and replace it with this"
<betaboon>
clever: i don't understand your question :(
<clever>
betaboon: what sphalerite said
thefloweringash has quit [Quit: WeeChat 2.0]
<bebarker>
sphalerite, somehow I had assumed the nix binaries were implicilty included :)
<sphalerite>
betaboon: I think clever meant bebarker
<GlennS>
you're quite right
<GlennS>
thanks sphalerite
justan0theruser has quit [Ping timeout: 268 seconds]
<sphalerite>
bebarker: they're included in the system profile on nixos, on other OSes you only have the user profile, and nix manages itself directly
<bebarker>
I was getting a conflict with some file in a different version of the same profile, asking me to change priorities when just using nix-env -if, so I tried nix-env --set
<sphalerite>
bebarker: you can roll back using /nix/var/nix/profiles/per-user/(your username)-(previous generation number)-link/bin/nix-env --rollback
<sphalerite>
bebarker: if you want to make your profile declarative and write the expression yourself rather than having nix do it for you, you can use nix-env --set -f foo.nix, just make sure to include nix itself in the env :D
grp has joined #nixos
<bebarker>
sphalerite, ok, will try both of those suggestions - thanks!
justanotheruser has joined #nixos
obadz has quit [Ping timeout: 276 seconds]
smichel_ has joined #nixos
zopsi has quit [Ping timeout: 255 seconds]
smichel has quit [Read error: Connection reset by peer]
smichel_ is now known as smichel
xAFFE has left #nixos [#nixos]
jrolfs_ has joined #nixos
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
Neo--- has quit [Ping timeout: 250 seconds]
RyanGlScott has joined #nixos
knupfer has quit [Remote host closed the connection]
joehh has quit [Ping timeout: 246 seconds]
obadz has joined #nixos
troydm has joined #nixos
doyougnu has joined #nixos
<nliadm>
anyone here use kakoune from nixpkgs (18.03) ?
drag has quit [Quit: Connection closed for inactivity]
johann__ has joined #nixos
jb55 has joined #nixos
waleee has joined #nixos
<{^_^}>
[nixpkgs] @brian-dawn opened pull request #40225 → librealsense: init at 2.11.0 → https://git.io/vpPi5
<nliadm>
I can't get a non-manual reproducer, but if someone could check this: fire up `kak -n`, type `:%` (as to run a command in a subshell) and it should immediately trip an assert
<{^_^}>
[nixpkgs] @FRidh pushed commit from @vcunat to staging « python: 2.7.14 -> 2.7.15 (bugfix + security) »: https://git.io/vpPPW
simukis has joined #nixos
johann__ has quit [Ping timeout: 260 seconds]
johann__ has joined #nixos
<betaboon>
upgrading from nix 1.11.x to nix2.0 on a single-user system should simple be removing /nix and rerunning the installer from nixos.org/nix/install right ?
obadz has quit [Ping timeout: 246 seconds]
<goibhniu>
betaboon: I don't think there's any need for that. Can't you update nix with nix?
obadz has joined #nixos
<clever>
betaboon: nix-env -iA nixpkgs.nix is all you need to do on a single-user setup
<{^_^}>
[nixpkgs] @FRidh pushed 64 commits to python-unstable: https://git.io/vpP15
<{^_^}>
→ 15a49a2f by @FRidh: python: Cython: 0.28.1 -> 0.28.2
<{^_^}>
→ 48cab266 by @FRidh: python: Nikola: 7.8.14 -> 7.8.15
<{^_^}>
→ 56672d0e by @FRidh: python: XlsxWriter: 1.0.2 -> 1.0.4
jrolfs_ has quit [Ping timeout: 260 seconds]
<betaboon>
clever: thanks :)
mizu_no__ has joined #nixos
ryantrinkle has joined #nixos
<betaboon>
clever: that's assuming I'm on nixpkgs-unstable, right ?
<clever>
betaboon: yeah
griff_ has quit [Quit: griff_]
<{^_^}>
[nixpkgs] @limeytexan opened pull request #40226 → perlPackages.NetZooKeeper: init at 0.41 → https://git.io/vpPMx
Ariakenom__ has joined #nixos
Ariakenom__ is now known as Ariakenom
Ariakenom_ has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @Synthetica9 opened pull request #40227 → libomxil-bellagio: fix build on gcc8 → https://git.io/vpPDt
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite>
clever: Dezgeg: thanks for the pointers. Managed to work out that it was because of my stateful setup that it didn't boot — the path to the system profile was wrong >_>
<sphalerite>
Remaining TODO: make it *actually* faster than loading the images in GRUB via TFTP.
<sphalerite>
The DHCP is really slow for some reason
<allana`>
Hi, I'm working on a derivation for a piece of software that has a script that is run during the build phase and hard coded to look for /usr/bin/tr (which will not be a valid path in NixOS). Assuming that I cannot modify this script, does anyone know of a way that I can work around this problem?
<gchristensen>
surely you're not using tftp, sphalerite
<Dezgeg>
there should be a way to have the network configuration passed from one stage to next
<sphalerite>
gchristensen: I was with GRUB, now I'm using HTTP with ipxe
<sphalerite>
but it's still just as slow as GRUB with TFTP because DHCP takes so long
<clever>
sphalerite: the netboot expression in nixos auto-generates an ipxe script you can chainload
<clever>
sphalerite: that ensures the init= is always valid
<Dezgeg>
maybe in the long term it would be best to skip this iPXE totally and just load the root filesystem from Linux with curl or something
johann__ has quit [Read error: Connection reset by peer]
<goibhniu>
hi allana`, you could fix that in the patchPhase e.g. with `/usr/bin/env tr`
<sphalerite>
clever: I know, I'm generating my own script that has a valid init=. But I made a temporary one outside nix for testing purposes
<clever>
Dezgeg: i did similar in my rpi, one min
<clever>
sphalerite: ahh
<sphalerite>
Dezgeg: ooh, yes, that would be nice. Although I'd need to build the network drivers into the kernel for that, and building kernels on old hardware isn't fun
<Dezgeg>
you can always include the modules on the initrd
<sphalerite>
Dezgeg: oh right so make a smaller initrd that doesn't include the rootfs as a squashfs
<Dezgeg>
yes
<clever>
sphalerite: check the 2 links i just pasted
<sphalerite>
yeah I am
<clever>
and line 142 could be replaced with curl
<Dezgeg>
or you could be brave and switch to NFS x)
<Dezgeg>
s/brave/maniac/
<clever>
nfs left a bad taste in my mouth after 32bit failed so badly with inodes
<sphalerite>
Dezgeg: or NBD! :D
<clever>
if you have a 64bit nfs server, and inodes go over 4 billion, then a 32bit nfs client fails with EOVERFLOW when you readdir()
<Dezgeg>
well the exact same thing can happen with a local filesystem
<clever>
and half of nixos treats EOVERFLOW as an empty dir
<sphalerite>
I don't have a real end goal with this, it's just a playground. But what would also be nice is if I could use the hard disks in the machines as a cache while having the actual data come from the network
<Dezgeg>
add the hard drives as swap?
<sphalerite>
will that make it cache network stuff on them? Or is there some additional config to do for that?
<Dezgeg>
yes, nfs should use the page cache on the client as well
<Dezgeg>
hmm, maybe it doesn't get swapped though
<sphalerite>
yeah that's what I'd think
ryantrinkle has joined #nixos
<Dezgeg>
maybe you can put the filesystem on a loop device backed by a file on nfs x)
<sphalerite>
What I don't get is why the DHCP is so slow… It seems to send about 5 DHCPREQUESTS before any arrive at the server
freusque has quit [Quit: WeeChat 2.0]
johann__ has joined #nixos
<sphalerite>
although if I connect them directly they all arrive. I don't think the switch is faulty though
<sphalerite>
that is, I've tried with another switch and had the same issue
<nyanloutre[m]>
clever: Yes I could try this thanks
jb55 has quit [Ping timeout: 255 seconds]
<clever>
nyanloutre[m]: i suspect nixos will still insist that you also listen on a tcp port, so just set it to some dummy value and dont open that on the firewall
<nyanloutre[m]>
clever: thanks
<sphalerite>
clever: so I plugged a cable into the switch's "CONSOLE" port and now I've got a prompt saying Switch>. I'm guessing now I look up the switch's manual?
<sphalerite>
clever: is there a way to make DHCP work more quickly without allowing loops?
<clever>
sphalerite: is it a cisco product?
<sphalerite>
clever: yeah, catalyst 3750
<clever>
sphalerite: if its cisco, the keywords you want are: spanning-tree portfast
<Dezgeg>
I doubt your home network topology will become complicated enough to contain switching loops :P
<clever>
sphalerite: and "enable" is cisco for "sudo"
<clever>
and "disable" undoes it
<clever>
sphalerite: another handy thing, you can set an IP on the switch and enable telnet (and probably ssh on a newer model then mine), then get console over that
<sphalerite>
Dezgeg: it's the network that connects all this crap ^
<sphalerite>
and I'm likely to make really stupid mistakes that would break stuff in really confusing ways if I don't leave the safeguards against it on :p
<clever>
ok, thats a bigger network then mine :P
<serentty[m]>
clever: I thought you meant the Nintendo Switch, and I was excited because I thought you had found some hidden debugging mode.
<maurer>
serentty[m]: Google "reswitch" if you need to root a switch
<clever>
serentty[m]: that hidden debug mode has already been found, short out 2 pins on the right connector, and it will run unsigned code shoved over the USB port
<Dezgeg>
well, that's not quite true; rather you run an exploit that allows running unsigned code
<clever>
serentty[m]: but the boot rom and signing hasnt been fully defeated, and it can only boot unsigned code over usb, and once rebooted, the stock OS is back
<clever>
samueldr: nice
<samueldr>
I'm waiting for progress upstream on one or two issues to continue playing around with the nixos aarch64 port
<serentty[m]>
clever: Yeah, I knew about that. I was just misimagining what you said.
<samueldr>
but all that tethered nonsense works with the overlay
<clever>
samueldr: have you considered using my qemu-user stuff to build the kernel and system image?, and is that running nixos on the switch?
<samueldr>
clever: I already have the fail0vervlow kernel building using the nixos infra :)
* Synthetica
stumbles into this conversation
<Synthetica>
NixOS on a switch?
<samueldr>
but the issues with display init (fails randomly) and me not having 1.8b UART yet are making me wait
<samueldr>
1.8v*
<samueldr>
well, I just received yesterday the 1.8v USB serial, I'm left to receive a rail kit so I don't have to sacrifice a joycon
<clever>
samueldr: 1.8v??, thats fairly low
<Dezgeg>
it's the norm these days
<samueldr>
that's what the joycon's serial runs at
<sphalerite>
clever: I've made some progress, but I'm still not sure which interfaces to set portfast on. I'd like to do it on the physical ports numbered 33-48 but I'm not sure what that translates to for the purpose of the command
<clever>
sphalerite: one sec
<clever>
let me plug my catalyst back in...
<sphalerite>
clever: oh, show interfaces summary seems to give me what I want
<clever>
yeah, thats what i was going to find the name of
<clever>
it should also show the link status of each
<sphalerite>
yep
<clever>
unplug something and diff
<clever>
oh, also turn on snmp
<sphalerite>
it actually helped me discover that something that was supposed to be plugged in wasn't plugged in :D
<clever>
cacti and other tools can graph the bytes/sec on every physical port
<sphalerite>
now how do I set this on 16 ports at once?
<clever>
write a bash for loop that echo's to one terminal, then paste paste paste?
<sphalerite>
xD
<Dezgeg>
you can pipe to /dev/ttyUSB* directly
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vpPb9
<{^_^}>
→ bf02618e by @brainrape: idrisPackages.vdom: init at 0.6.0
<{^_^}>
→ b9054ec5 by @matthewbauer: Merge pull request #40168 from brainrape/init-idrisPackages-vdom
<clever>
sphalerite: there are also some minor security issues ive noticed with the IPFS protocol
jgertm has joined #nixos
<clever>
sphalerite: first, the node generates a keypair on first startup, and posts a pubkey=$MYIP to the DHT on every startup, so you can now be tracked at all times
<clever>
sphalerite: second, it also publishes the hashes of every file and block you are sharing
<clever>
anybody with knowledge of the protocol can probably iterate those, and download all of your shared files
<sphalerite>
clever: but it doesn't need to be connected to the internet, right? I can just use it on the LAN
<clever>
probably
<clever>
but i think the p2p bootstrap uses dns to find a stable node, and then the rest of the world
<sphalerite>
surely it would use local discovery too
<clever>
yeah, i would expect it to also have that
<Synthetica>
Hmm, I've recently sold a couple of those boxes 🤔
<{^_^}>
[nixpkgs] @emmanuelrosa opened pull request #40236 → gnucash: add upgrade instructions to NixOS release notes → https://git.io/vpXJD
<Synthetica>
sphalerit: what's it for?
<sphalerite>
Synthetica: mostly playing around with
<sphalerite>
I've got them all netbooting nixos, next step is to test the RAM and HDDs in all of them
<Synthetica>
Ah, cool :)
rauno has joined #nixos
Arcaelyx has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
knupfer has quit [Remote host closed the connection]
<ghostyy>
in nixos is there a way to include a file as a string
<ghostyy>
er, nix
<ghostyy>
specifically i dont want my wireguard private keys in my configuration.nix
rauno has quit [Ping timeout: 256 seconds]
<ghostyy>
i want them in a separate file which has different permissions
<electrocat>
ghostyy: ofcourse, the import statement
<ghostyy>
ah great
<ghostyy>
thanks!
<sphalerite>
ghostyy: you'll want to use the privateKeyFile option
<sphalerite>
ghostyy: otherwise it still gets imported into the nix store as a world-readable(!) file
<ghostyy>
oh! perfect
<electrocat>
sphalerite: right, good point
<ghostyy>
do i just copy my keyfiles into /run/keys ?
<ghostyy>
or is there a better place for them
jgertm has quit [Ping timeout: 240 seconds]
humanoyd has quit [Quit: WeeChat 2.1]
<steveeJ>
Profpatsch: zimbatm pointed me towards the container page you've set up on the wiki as I approached him about forming a SIG for Kubernetes. are you interested in helping to set that up?
TheAppleMan has quit [Ping timeout: 256 seconds]
<sphalerite>
ghostyy: /run/keys is a nixops-specific thing really. You can use it, but note that anything under /run doesn't persist across boots!
<ghostyy>
ahh i see!
<Profpatsch>
steveeJ: I don’t use kubernetes, but please don’t hesitate to add yourself to the container workgroup
<ghostyy>
do things in /etc persist?
<sphalerite>
ghostyy: I think the usual location for that sort of stuff is either /var or /etc, or you could put it in root to be surer of only root being able to read them
<{^_^}>
→ 33bc4a7b by @endgame: haskellPackages.blank-canvas_0_6_2: dontCheck
<{^_^}>
→ 85fc8338 by @peti: Merge pull request #40198 from endgame/dontCheck-blank-canvas_0_6_2
jgertm has joined #nixos
<ghostyy>
ill put it in /var then, thanks!
<steveeJ>
Profpatsch: what do you use for scheduling containers?
Arcaelyx has joined #nixos
Arcaelyx has quit [Client Quit]
<Profpatsch>
steveeJ: Scheduling? Right now only a pretty static docker-compose
<Profpatsch>
kubernetes feels like overkill for this use-case.
<Profpatsch>
Another few 100k lines of code to schedule some stuff. Hrm
<steveeJ>
Profpatsch: gotcha, is it in a single node environment?
<Profpatsch>
Plus made by a giant corporation for their thousands of production servers, other scaling requirements
<Profpatsch>
Yeah, right now it is.
betaboon has quit [Quit: WeeChat 2.0]
asuryawanshi has joined #nixos
<Profpatsch>
And postgres & rabbitmq we use do scale fairly easily without a lot of management.
<judson>
Is there a way for nixops to run... distributedly? It seems like you've got to `nixops deploy` from exactly one place.
rauno has joined #nixos
Sonarpulse has joined #nixos
<zimbatm>
judson: it depends :D
<zimbatm>
the nixops state file contains a few things, some of them can be discarded
<simpson>
steveeJ: Are you looking for people deploying Nix-built stuff on k8s, or people using Nix to administer and deploy k8s itself?
<Profpatsch>
judson: iirc that’s what disnix is trying to achieve
<zimbatm>
judson: 1. SSH key to ssh into the box. If you provision all your machines with your admin public keys then they can be discarded
<Profpatsch>
simpson: wat, it’s already customary to number-cut it like i11n?
* Profpatsch
shakes head
asuryawanshi has quit [Remote host closed the connection]
<zimbatm>
judson: 2. only use network "none"
<steveeJ>
simpson: primarily for the former, the latter would of course go very well with it
<Synthetica>
Well, i18n is an exception I think
<simpson>
steveeJ: Well, then yeah, I do that. I haven't really felt any need for a SIG or other coordinated effort, but I'm interested in the discussion.
asuryawanshi has joined #nixos
<Profpatsch>
Synthetica: Ah, it was 18 :)
<Synthetica>
Because it's so damn long to type
<Synthetica>
We're speaking English, not German :P
<sphalerite>
Profpatsch: and l10n
<steveeJ>
simpson: the goal would be keep up with upstream releases and always have the tests working
Sonarpulse has quit [Remote host closed the connection]
<samueldr>
a11y uses 11 (accessibility)
<simpson>
Which upstreams and which tests? Are we behind in some way?
<sphalerite>
samueldr: oh yeah that one!
<Synthetica>
I've never heard of l10n and a11y being used before now
<mightybyte>
When I do a nix-build I'm getting an error "file 'nar/xyz.nar' missing from binary cache"
<mightybyte>
I can't for the life of me figure out how to avoid this.
zincy has joined #nixos
<steveeJ>
simpson: the official releases of kubernetes, and the tests currently available in the nixpkgs tree, and eventually the e2e testsuite
<samueldr>
l10n is localization, closely related to i18n, internationalization
olto has joined #nixos
<samueldr>
(and I hate numeronyms too, but it's useful to know them)
<simpson>
steveeJ: Ah. Yeah, that's stuff I don't really care about, sorry. But it's probably good to get done.
coot has quit [Quit: coot]
coot has joined #nixos
raynold has joined #nixos
<steveeJ>
simpson: if you wanted to host a production cluster you'd probably care, right? :-D
<simpson>
I like this PR though.
<simpson>
Yeah. My position is that administering k8s directly is quite hard, and letting the public cloud vendor do it has been reasonable for me for now.
<simpson>
Also I think that it's unfortunate that there's no way to have a maintainer for our system service modules.
<elvishjerricco>
samueldr: Does your Bluetooth headset connect to your Linux machine automatically? Mine does, but not quite successfully, leaving it in a state where I have to manually disconnect it and reconnect it for it to work.
<steveeJ>
simpson: I think with the SIG we'd have plenty of maintainer responsibility
<samueldr>
elvishjerricco: no, but I think the headset is weirdly broken, it doesn't connect to anything automatically :/
<elvishjerricco>
samueldr: I have a bunch of devices that I switch it between, so I actually wish mine didn't autoconnect :P
<samueldr>
hah, not a bug, a feature!
<samueldr>
though yeah, this isn't helpful since I can't help with automatic reconnects
olto has joined #nixos
coot has quit [Quit: coot]
zincy has quit [Ping timeout: 260 seconds]
coot has joined #nixos
deepfire has quit [Ping timeout: 250 seconds]
<judson>
zimbatm: Having done those two things, the state file doesn't matter? Huh.
<steveeJ>
simpson: unfortunately it's too hard for me alone to pick up the current state of the module, there are too many errors in the log
asuryawanshi has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @megheaiulian opened pull request #40240 → distrobuilder: init at HEAD → https://git.io/vpXYd
coot has quit [Remote host closed the connection]
jrolfs has quit [Ping timeout: 256 seconds]
coot has joined #nixos
MichaelRaskin has joined #nixos
zincy has joined #nixos
Neo-- has quit [Ping timeout: 246 seconds]
<zincy>
Say I have two private haskell libraries in different directories. Library A imports B. How do I import B using A in nix
<zincy>
*within
jbboehr has quit [Quit: Quit]
<Synthetica>
Is it possible to get an email when a Hydra job finishes?
FRidh has quit [Quit: Konversation terminated!]
vaninwagen has quit [Ping timeout: 255 seconds]
GiGa has joined #nixos
GiGa has quit [Changing host]
GiGa has joined #nixos
<GiGa>
Evening all
vaninwagen has joined #nixos
<GiGa>
Sorry to keep asking, but anyone used a Yubikey with Nixos?
<electrocat>
GiGa: yes
<infinisil>
,ask GiGa
<{^_^}>
GiGa: Don't ask if anyone uses software xyz, just ask your specific question. This helps solve your problem quicker, and allows people who haven't used xyz but may still be able to help you to do so.
<{^_^}>
→ 32d8a7d2 by @va1entin: keepassxc: 2.3.2 -> 2.3.3
<{^_^}>
→ 59e94000 by @xeji: Merge pull request #40238 from va1entin/master
<GiGa>
electrocat: I would agree but having asked recently the question I figured I'd go with finding out if anyone had actually used the device in question
<GiGa>
rather than wasting everyone's time
<GiGa>
I have a Yubikey, I'm trying to get it working in Nixos
<GiGa>
not having any luck
<electrocat>
what doesn't work?
<electrocat>
anything?
<GiGa>
getting to that, needed to turn a light on
<GiGa>
lsusb shows the device is present (Yubico.com Yubikey Touch U2F Security Key)
<GiGa>
yet the yubico tools suggest there's no yubikey present
<electrocat>
ok, pressing the key generates OTP's?
<GiGa>
I'm assuming, quite possibly wrongly, that the problem relates to udev
<electrocat>
it's usually udev :p
<GiGa>
This is a fido2 U2f key so pressing the button won't do anything unless prompted
ftzm has left #nixos ["WeeChat 2.2-dev"]
<electrocat>
hm
<electrocat>
this is embarassing
<electrocat>
don't have my nixos machine on me right now
<GiGa>
what, me?
<GiGa>
ah, that could be a problem
<{^_^}>
[nixpkgs] @xeji merged pull request #40204 → [RDY] wireshark: run binary from nix-shell → https://git.io/vpi5x
<{^_^}>
→ ab613b93 by @teto: wireshark: run binary from nix-shell
<{^_^}>
→ c85d0bec by @xeji: Merge pull request #40204 from teto/wireshark_hook
<electrocat>
i also never use the yubico tools
<electrocat>
i just use it as gpg smartcard
<GiGa>
Using chromium (not chrome, which may also be an issue) plugging in the yubikey when prompted just says "something went wrong"
<zincy>
How do you import another derivation as an input into a given derivation? Essentially composing nix configs?
<GiGa>
which yubikey do you have electrocat ?
robstr has quit [Quit: WeeChat 1.9.1]
<GiGa>
zincy: like wanting to perform an include?
<electrocat>
yubikey 4
jrolfs has joined #nixos
<zincy>
Not sure what an include is?
<electrocat>
zincy: simply pass it as an argument?
<zincy>
Yes
<GiGa>
zincy: are you wanting to have one file with some common config (that perhaps you share across a few computers) and then just say in the main config "also include the config in that file over there?"
<zincy>
Get error: syntax error, unexpected '=', expecting $end, at /Users/tom/client/app.nix:1:9
<GiGa>
electrocat: Seems the security key doesn't work with google-chrome-stable either. Is there a way to tell if the key is being detected correctly? I can't use the security key as a smartcard with PGP
<GiGa>
zincy: are you able to paste your config at pastebin? Sounds like a syntax error to me
<electrocat>
GiGa: it's hard to help if i cant check why it works on my end :S
Sonarpulse has joined #nixos
<infinisil>
zincy: what does your shared/default.nix file contain?
Sonarpulse has quit [Client Quit]
<GiGa>
electrocat: Quite agree
nur0n0 has joined #nixos
Sonarpulse has joined #nixos
<infinisil>
zincy: if it's something like `{ shared = ...; }` then you need to replace the first line in your gist with `with import ../shared/default.nix;`
<{^_^}>
→ 882ed257 by @bfortz: owncloud-client: 2.3.4 -> 2.4.1
<zincy>
^ That is the default.nix I am importing
<{^_^}>
→ 96c2cad2 by @xeji: Merge pull request #39566 from bfortz/owncloud-2.4.1
<johnw>
how do I determine which Git version of nixpkgs a system is using, even if it comes from a channel?
<infinisil>
zincy: Ah, then you want to use `{ pkgs ? import <nixpkgs> {} }: let shared = pkgs.haskellPackages.callPackage ../shared/default.nix {}; in`
<infinisil>
Ah wait, it could be nicer
<zincy>
Oh then I get this error: anonymous function at /Users/tom/client/app.nix:1:1 called with unexpected argument 'miso', at /nix/store/l713q53x0h6lngb58g
<{^_^}>
→ 9410c39b by @limeytexan: perlPackages.LinuxACL: init at 0.05
<{^_^}>
→ df37647a by @limeytexan: Fix alphabetization of perl modules
knupfer has joined #nixos
LnL has quit [Ping timeout: 248 seconds]
tnks has joined #nixos
olto has quit [Read error: Connection reset by peer]
<tnks>
I tried to simplify this invocation with the -f switch, but got stuck: nix run '(with import <nixpkgs> {}; python36Packages.python.withPackages (p: with p; [ docker ipython ]))' -c ipython
zincy has quit [Ping timeout: 260 seconds]
<tnks>
the "with import <nixpkgs> {}" part seems verbose.
<jtojnar>
tnks: you can also do nix run -f '<nixpkgs>' python36 docker python36.pkgs.ipython -c ipython
justanotheruser has joined #nixos
<electrocat>
GiGa: on my current machine, pcscd is a dependency of gnupg, pcscd provides udev rules for the smartcards
<lejonet>
electrocat: pcscd iirc pcscd is the interface towards the smart card hardware and the OS and the scdaemon interfaces with pcsc on-behalf of applications to ease the use of smartcards
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tnks>
jtojnar: I think that's not quite what I want, because that "docker" is the docker client/server, and what I really want is the "docker" Python libary integrated into the interpretter's PYTHONPATH.
<sphalerite>
Sonarpulse: setfont <fontname>
<Sonarpulse>
sphalerite: yeah but remember the big font? :)
<sphalerite>
Sonarpulse: setfont --help tells you where it looks for them so you can list them
<antalsz>
so that has 18.09pre139198.9d0b6b9dfc9
<Sonarpulse>
ah ok
<sphalerite>
Sonarpulse: latarcyrheb-sun32
<jtojnar>
tnks: then nix run -f '<nixpkgs>' python36 python36.pkgs.docker python36.pkgs.ipython -c ipython
<samueldr>
installer image Sonarpulse?
<tnks>
infinisil: yeah, it's kind of unfortunate that it's more concise with "nix-shell -p" because I prefer the "-p" magic less architecturally.
<tnks>
infinisil: I think nix-shell is a lot more than a simple nix run... far more environment variables set, to start. And I think copumpkin makes good arguments too: https://github.com/NixOS/nix/issues/726
<antalsz>
any tips on how I might have figured that out?
<samueldr>
in the PR I linked I have added boot options which forces the VT to resolutions to help with HiDPI
<elex>
Hi, to short questions. Is there a full parsed list off all available parameters in configuration.nix? And second is there anything like a code completion for any editor?
<elex>
to= two
<infinisil>
antalsz: the version of when you do `import <nixpkgs> {}` comes from NIX_PATH, you have `/nix/var/nix/profiles/per-user/root/channels/nixpkgs` on there, which means it's gonna use the root's channels
<antalsz>
infinisil: aha. So _should_ that be root? Or should it be me?
<sphalerite>
Sonarpulse: for long-term use I usually prefer grabbing a font from terminus (i18n.consoleFont = "${pkgs.terminus_font}/share/consolefonts/ter-u32b.psf.gz";) because let's face it, latarcyrheb-sun32 is fugly.
<GiGa>
If building the system fails because one dependency couldn't be built do any of the changes get implemented or does NixOS back out the whole lot?
<Sonarpulse>
samueldr: that does look nice!
<sphalerite>
GiGa: none of it happens
olto has quit [Client Quit]
<samueldr>
while it brings no actual changes to the live system, it changes a bunch *right before*, the more hardware it's tested on, the better it would be
<sphalerite>
GiGa: as far as building is concerned, it's all-or-nothing. Switching can be partial though.
<lejonet>
okay, different question, if I want to fool around with different kernel options, what is the easiest way to do that? (I have an early Zen CPU that could benefit from having RCU_NOCB_CPU enabled, but have found that just changing extraConfig of the relevant kernel package with packageOverrides isn't enough to trigger a rebuild)
antalsz has quit [Quit: Page closed]
<{^_^}>
[nixpkgs] @Synthetica9 opened pull request #40246 → systemd: suppress certain warnings to comile under gcc8 → https://git.io/vpX8M
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vpX87
<{^_^}>
→ 2cac5d73 by @nicknovitski: cloud-sql-proxy: init at 1.11
<{^_^}>
→ 37cde003 by @matthewbauer: Merge pull request #36358 from nicknovitski/init-cloud-sql-proxy
olto has joined #nixos
<lejonet>
symphorien: good to know, will try that at once, I just couldn't for my life figure out why it wouldn't rebuild it when I just fiddled with the extraConfig in override
<ghostyy>
have any of you guys gotten wireguard working in nixos
<sphalerite>
!doesanyoneuse
<{^_^}>
Don't ask if anyone uses software xyz, just ask your specific question. This helps solve your problem quicker, and allows people who haven't used xyz but may still be able to help you to do so.
<sphalerite>
ghostyy: ^
<{^_^}>
[nixpkgs] @matthewbauer pushed to master « lib.makeSearchPath: allow null in search path »: https://git.io/vpX4v
<ghostyy>
so i was hoping someone would have a working config file lying around to show me
<symphorien>
you put this in boot.kernelPackages = ...
olto has quit [Client Quit]
qualiaqq has joined #nixos
<lejonet>
symphorien: ah, so you do the override already in kernelPackages instead of in nixpkgs.config.packageOverrides?
<sphalerite>
ghostyy: being more specific about how it doesn't work would be helpful
<infinisil>
,hint = Commands prefixed with , such as ,ask take a nickname as an argument, which you can use to direct the resulting message to that specific user, e.g. ,ask paul
<{^_^}>
hint defined
<sphalerite>
thanks infinisil :D
<infinisil>
,hint sphalerite
<{^_^}>
sphalerite: Commands prefixed with , such as ,ask take a nickname as an argument, which you can use to direct the resulting message to that specific user, e.g. ,ask paul
<ghostyy>
sphalerite, im just making a pastebin now
<infinisil>
Need to promote my new bot things!
olto has joined #nixos
<lejonet>
infinisil: does it have !botsnack? :P
<symphorien>
lejonet: the kernel is special because it is a scope, it may interfere
freeman42x[nix] has quit [Ping timeout: 250 seconds]
<infinisil>
,help lejonet
<{^_^}>
lejonet: Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<lejonet>
symphorien: so far I've figured :) just that the manual shows using packageOverrides as the way to enable config
<infinisil>
What does botsnack do lejonet?
<symphorien>
and since setting boot.kernelPackage changes the used kernel everywhere, I don't need packageOverride
<lejonet>
infinisil: gives the bot a snack :)
<ghostyy>
what id really like is some kind of error message about why it doesnt work...
<sphalerite>
lejonet: I've overridden the kernel using an overlay like self: super: { linuxPackages = self.linuxPackagesFor (super.linux.override[Attrs] …); }
<symphorien>
lejonet: ah sorry no idea then
<ghostyy>
but i dont see anything in dmesg, journalctl et al
<ghostyy>
oh, i forgot the part about it not working
olto has quit [Client Quit]
<lejonet>
I
<lejonet>
I
<lejonet>
... stupid enter!
<infinisil>
Hah
<ghostyy>
which is that i cant ping either address from each other
<ghostyy>
eg i can ping 10.8.0.1 from 10.8.0.1 and same for .2, but i cant ping .1 from .2 or vice versa
<sphalerite>
ghostyy: do the interfaces have the addresses assigned correctly (ip a)?
<lejonet>
I'll try the kernelPatches way and see if that works for me, else I'll probably have to do the overlay way
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #40248 → generic-stack-builder: attempt to add propagation → https://git.io/vpX4a
drakonis has joined #nixos
<sphalerite>
lejonet: kernelPatches does look more convenient, I didn't know patches could specify config options
<ghostyy>
.2 has inet 10.8.0.2/32 scope global wg0 and .1 has inet 10.8.0.1/24 scope global wg0
<sphalerite>
would probably have saved me considerable amounts of time >.>
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #27451 → simpleitk: init at 1.1.0 (note this only builds the C++ libs, not the Python/Lua/... bindings) → https://git.io/vQN2q
<{^_^}>
[nixpkgs] @matthewbauer pushed commit from @bcdarwin to master « simpleitk: init at 1.1.0 (note this only builds the C++ libs, not the Python/Lua/... bindings) (#27451) »: https://git.io/vpX4M
<ghostyy>
not sure if the different subnets change anything
<sphalerite>
ghostyy: the .2 one is wrong, that means it's the only thing in that subnet and can't actually talk to anything through it
<lejonet>
sphalerite: me neither until symphorien showed me it just now :P
<sphalerite>
ghostyy: they should both be /24
<ghostyy>
ooh
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<symphorien>
I had to read the source code to find this
<symphorien>
setting patch to null is rather a hack
olto has joined #nixos
<ghostyy>
May 09 16:51:14 computer2 ip[2850]: RTNETLINK answers: Invalid argument
<ghostyy>
AAAAAAAAAAA
<infinisil>
> botsnack = select [ "Om nom nom!" "That's very nice of you!" "Oh thx, have a cookie yourself!" "Thank you very much." "Thanks for the treat!" "{^_^} has been given a cookie, {^_^} is free!" ]
<{^_^}>
botsnack defined
<Drakonis[m]>
bcachefs is being prepared for the upstream gauntlet
<infinisil>
> botsnack /* like this lejonet? */
<{^_^}>
"{^_^} has been given a cookie, {^_^} is free!"
<sphalerite>
hm, I wonder…
<ghostyy>
ahhhh now i see
<lejonet>
infinisil: haha, yes :P
<sphalerite>
> builtins.trace "hello" "foo"
<{^_^}>
"foo"
<ghostyy>
it was .2 when it should have been .0
<lejonet>
infinisil: we had that for willikins, the gentoo channel bot :P just as a ploy thing
<sphalerite>
ghostyy: which one should have been .0?
<Ralith>
I have a private s3 bucket configured as a binary cache; `nix build` gives `AWS error checking bucket 'my-bucket': Access Denied`, but `nix copy` works fine. Anyone know what's going on?
<ghostyy>
allowedIPs
jrolfs has joined #nixos
<ghostyy>
on computer2
ertes-w has quit [Quit: Bye!]
<ghostyy>
now computer2 has inet 10.8.0.2/24 scope global wg0
<MichaelRaskin>
infinisil: if after receiving a snack the bot becomes free, does it mean it will stop answering requests?
<sphalerite>
can we please stop recommending packageOverrides :(
<lejonet>
MichaelRaskin: or answer them in any way it seems fit? :P
<infinisil>
,nixeval = I can evaluate nix if you prefix your expression with "> ". Use it here only for short demos or helping others with the syntax. If you wanna spam me, do it in #bottest!
<{^_^}>
nixeval defined
<samueldr>
symphorien: funny, this page is only a few hours old :)
<infinisil>
MichaelRaskin: Adding that to my todo list :P
<symphorien>
I just saw it because of the PR bot
<MichaelRaskin>
lejonet: as it sees fit, and not in the manner strictly bound by deterministic instructions?
<lejonet>
MichaelRaskin: exactly
<MichaelRaskin>
infinisil: the only risk is that people start calling it «poison cookie»
<lejonet>
for all we know, it could start cating /dev/urandom for answers
<Drakonis[m]>
sphalerite: what's better than overriding?
<MichaelRaskin>
lejonet: and in this case, maybe «hash cookie»
<infinisil>
MichaelRaskin: Or "russian cookie" hah
Arcaelyx has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite>
Drakonis[m]: overlays are better than packageOverrides (specifically packageOverrides, the .override and .overrideAttrs on packages are only tangentially related)
<MichaelRaskin>
infinisil: no, I think Russian cookies are pryanik cookies
<infinisil>
MichaelRaskin: But did you really call nix deterministic?
qualiaqq has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @basvandijk opened pull request #40250 → haskell: fix build of base-compat-batteries → https://git.io/vpXRo
winem_ has quit [Ping timeout: 250 seconds]
simukis has quit [Ping timeout: 250 seconds]
jrolfs has quit [Ping timeout: 256 seconds]
<infinisil>
MichaelRaskin: You can view variable bindings with `> :v rand` :)
<MichaelRaskin>
Ah, it uses currentTime
<sphalerite>
huh, builtins.currentTime doesn't work right in `nix repl`
<infinisil>
sphalerite: does for me?
<sphalerite>
infinisil: try evaluating it again
<infinisil>
yeah it's the same :)
<infinisil>
I think it's because nix-repl caches it
judson has quit [Read error: Connection reset by peer]
<lejonet>
symphorien: kernelPatches works awesomely sofar, thanks :D
<lejonet>
and now that I've added buildCores to the correct value, it probably won't take 30 mins to compile it all xD
Ariakenom has quit [Quit: Leaving]
<{^_^}>
[nixpkgs] @ariutta opened pull request #40252 → pgmanage: update to current name (minor). → https://git.io/vpX0Z
<sphalerite>
so I have 9 machines which are all running nixos and I can SSH into. What's a nice way of running a script on all 9 of them and collecting the results?
hamishmack has joined #nixos
grp has quit [Quit: box shutting down...]
Ariakenom has joined #nixos
rindolf-nixvm has joined #nixos
Arcaelyx has joined #nixos
<rindolf-nixvm>
hi all
<Synthetica>
sphalerite: A ssh-ing them the command in a for loop works in a pinch
<{^_^}>
→ 527abe59 by @matthewbauer: generic-stack-builder: attempt to add propagation
<{^_^}>
→ 12ee278a by @Ericson2314: Merge pull request #40248 from matthewbauer/fix-generic-stack-builder
<Sonarpulse>
Dezgeg: ping
<Sonarpulse>
wondering about -gnu
hiratara has joined #nixos
knupfer has quit [Remote host closed the connection]
coot has quit [Quit: coot]
alexteves has quit [Ping timeout: 264 seconds]
coot has joined #nixos
coot has quit [Client Quit]
coot has joined #nixos
alexteves has joined #nixos
hph^ has quit [Ping timeout: 240 seconds]
obadz has quit [Ping timeout: 240 seconds]
<ghostyy>
i just want to point out that the iptables command in https://nixos.wiki/wiki/Wireguard actually literally prevents the entire setup from working, i just spent two hours figuring this out
obadz has joined #nixos
saltsa has joined #nixos
Ariakenom has quit [Quit: Leaving]
<Ralith>
is there a way to run nix-shell with a package identified by store path?
<Ralith>
and why would `nix build` be unable to use a binary cache that `nix copy` talks to fine?
<infinisil>
Ralith: nix-shell -p $(nix-build --no-out-link '<nixpkgs>' -A hello)
<halfbit>
I mean, how do you get flycheck and all that working then
<halfbit>
that requires all the devtools for rust
<halfbit>
cargo, rustc, racer, etc
<Ralith>
halfbit: you use something like the tool I linked you
<{^_^}>
→ ca060419 by @teto: qemu-guest-agent: init module
<{^_^}>
→ 3d6f752c by @xeji: Merge pull request #39099 from teto/qemu_agent
<foldingcookie>
clever: what part of that would I change to use a toolchain for an architecture that isn't the RPi, and what part of that creates the /bin/pwd expected by the kernel's build system?
<clever>
foldingcookie: nothing will create a suitable /bin
<clever>
foldingcookie: everything in nix works by configuring $PATH
filwisher has joined #nixos
<foldingcookie>
so what do Linux kernel dev workflows look like on NixOS? the kernel's build system hard-codes paths like /bin/pwd
<clever>
foldingcookie: the nix expressions automatically patch the kernel source
<foldingcookie>
clever: how do I point the nix expressions to my kernel tree? do the nix expressions also add gitignore entries for the patches they apply?
<clever>
foldingcookie: it applies the patches in a copy of the source
<clever>
foldingcookie: what cpu are you trying to target?
<foldingcookie>
arm64, but I have a very rough linux tree (not my fault) that doesn't compile properly with a modern toolchain
<clever>
foldingcookie: aarch64-android-prebuilt, aarch64-multiplatform or aarch64-multiplatform-musl can be used for arm, in-place of raspberryPi
<clever>
for the kernel itself, i think you can run nix-shell against one of the kernels, like my original example, and then run patchPhase in there to edit the copy of the source in the working directory
<{^_^}>
infinisil: Don't ask if anyone uses software xyz, just ask your specific question. This helps solve your problem quicker, and allows people who haven't used xyz but may still be able to help you to do so.
<infinisil>
Thanks
<infinisil>
So I'm using the config from the linked site in the hopes of allowing me to use radicale with Let's Encrypt
<infinisil>
But currently LE can't get a valid certificate
<infinisil>
The error i get is "Not enough PEM encoded messages were found in fullchain.pem; at least 2 were expected, found 1."
<infinisil>
I more or less know what that means, but I have no idea why that's the case, because it works for all my other domains no problem