spudly1 has quit [Read error: Connection reset by peer]
<alienpirate5>
ajs124:
<alienpirate5>
so the iso it generated is getting copied... somewhere?
justanotheruser has quit [Ping timeout: 264 seconds]
<ajs124>
the job was built on 48d0427b.packethost.net. now it's being copied from there, presumably over the hydra host into the binary cache. not sure how exactly the setup works.
<kalbasit>
gchristensen: I'm trying to build a json of all packages in our closure recursively. Will I get access to the meta.licenses in the closure info?
<pumpy>
gchristensen happen to know my ^ problem with bash prompt? :/
cognemo has quit [Ping timeout: 258 seconds]
leungbk has joined #nixos
<kalbasit>
gchristensen: it doesn't :( So I guess I need something that works on eval time
<fzakaria>
I am chasing a linking error in Java and LD_DEBUG shows:
<fzakaria>
```
<fzakaria>
2861762:binding file /lib/x86_64-linux-gnu/libc.so.6 [0] to /nix/store/bdf8iipzya03h2amgfncqpclf6bmy3a1-glibc-2.32/lib/libc.so.6 [0]: normal symbol `realloc' [GLIBC_2.2.5]
<fzakaria>
```
<fzakaria>
I'm not that well versed with `ld` can someone throw me some pointers on what to investigate O.o
<{^_^}>
[nixpkgs] @kalbasit opened pull request #103396 → terraform: withPlugins should inherit meta from wrapped derivation → https://git.io/JkIYk
<{^_^}>
[nixpkgs] @jonringer merged pull request #97261 → python3Packages.tensorboardx: fix build after upgrade to PyTorch 1.6 → https://git.io/JUGTv
<{^_^}>
[nixpkgs] @jonringer pushed commit from @akamaus to master « python3Packages.tensorboardx: fix build after upgrade to PyTorch 1.6 »: https://git.io/JkIcG
<Ke>
I think room state is incrementally computed and there are no state checkpoints
<jdnixx[m]>
I have pretty much no idea how it works technically
<jdnixx[m]>
or the differences to IRC
<Ke>
jdnixx: as I understand, matrix never forgets anything that has happened, which gets heavier over time
<jdnixx[m]>
ooooh
<jdnixx[m]>
IRC forgets everything by default I think right
<Ke>
O(N) heavier, or O(N log N) probably
<Ke>
yes
kamid has quit [Quit: WeeChat 2.9]
sneak has quit [Remote host closed the connection]
<jdnixx[m]>
no easy way to just prune old shit?
<bbigras>
matrix is nice for multiple devices and encryption. I'm looking forward for the p2p support.
<jdnixx[m]>
<bbigras "matrix is nice for multiple devi"> that's cool. didn't even know that was planned
<jdnixx[m]>
and yeah I tried using a few IRC clients on android and it was a fkn joke
<Ke>
p2p is just a prototyping project and I would not expect it to happen tomorrow
<bbigras>
I think they want an hybrid p2p and servers to make the network lighter and more resilient maybe.
orivej has joined #nixos
<bbigras>
yeah but they seems to say it's matrix's future.
<Ke>
the chat projects that started p2p really were not competitive, getting there is a research project
* colemickens
still doesn't understand how e2ee works with p2p clients
<Ke>
and people really do dissertations on this stuff
<Ke>
colemickens: how would p2p make e2e harder?
<bbigras>
I'm expecting e2e to be the same as the current model
<bbigras>
you cross sign your new key when you get a new device and you use the same key for a while on each devices.
<colemickens>
Emil Karlson: currently new clients get encryption keys for old conversations *from* existing clients, to my understanding. If my clients are ephemeral, how is a new "peer" client going to get my keys for old convos?
<colemickens>
maybe I'm thinking about it wrong though, I've done very little research into it
<notgne2>
aaaaa multi-device distributed e2e stuff is hard
<notgne2>
giving me flashbacks from when I used to work with the Tox devs
<bbigras>
I think clients are not ephemeral
<Ke>
colemickens: ok, I don't know how to resolve that indeed
<bbigras>
so for the "demo" they made. you just run a matrix server locally (with web assembly), but you still use the same account.
<bbigras>
your local server is federating or whatever with the others
<jdnixx[m]>
<notgne2 "aaaaa multi-device distributed e"> I've always thought this stuff might work well in some type of blockchain system
<notgne2>
uh oh
<bbigras>
isn't identity the hard part of p2p?
<Ke>
bbigras: but the point is, that occasionally you don't have the homeservre on
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « qmmp: 1.4.1 -> 1.4.2 »: https://git.io/JkI8k
<colemickens>
its not just identity though, you actually need the keys for old convos
<notgne2>
I will keep to my guns with insisting that 9/11 times you think you want a blockchain, you want a DHT
<Ke>
tox doesn't even do hard things like offline message delivery
<bbigras>
oh for like public rooms with old messages
<jdnixx[m]>
<notgne2 "I will keep to my guns with insi"> well yeah for sure always gotta keep that in mind
<jdnixx[m]>
I was sorta thinking to address the cost problems or whatever of hosting
<Ke>
retroshare does everything and the kitchen sink, but does not really work
<colemickens>
bbigras: no, like on a new device, when I first login, if I don't verify, I can't read _any_ existing conversations because the new client doesn't have the keys for those convos
<jdnixx[m]>
just have everyone pay microtransactions or something when going between servers
<notgne2>
Ke (chat.freenode.net): there's much that can be improved with it, it barely has groupchat support, and the new improved groupchats dropped A/V support
<colemickens>
my understanding is that when I verify a new client with an old client, it's able to "get" the conversation keys from the old client (again though I should go look this up before I keep guessing outloud)
malook has joined #nixos
<notgne2>
also while userfiles are nice and all, I would like it if Tox could use deterministic scattering to store an encrypted copy of your userfile on the DHT too, this would allow for faux username+password logins
<colemickens>
("any existing conversations " = one on one conversations that had encryption enabled)
<notgne2>
it's all technically possible, just nobody has done it yet
<Ke>
sure improvements are possible, but because the usability must come now, I believe matrix is currently the best option
<notgne2>
oh I agree, I'm using matrix right now to talk on this channel
<jdnixx[m]>
agreed
<colemickens>
Emil Karlson: the improvements seem to be numerous and steady, I'm excited.
<bbigras>
<colemickens "my understanding is that when I "> yes but you can also save the key. on your computers or on the server.
<jdnixx[m]>
<notgne2 "oh I agree, I'm using matrix rig"> why not join the actual matrix room then lol
<notgne2>
there's a reason I'm not involved with Tox anymore lol
<notgne2>
jdnixx[m]: I didn't know there was one
<jdnixx[m]>
well admittedly it was hard for me to join
<jdnixx[m]>
forget why but whatever link they had on nixos.org didn't let me join at first
<jdnixx[m]>
and element didn't list the room for some reason
<Ke>
I am not sure I believe it is possible to have fat p2p client on mobile
<jdnixx[m]>
or something like that
<Ke>
as in feasible
<colemickens>
bbigras: okay, but what happens when I start a convo with friend A from my phone and friend B from my laptop and want to continue the conversation with either on the other device? I have to manually store and sync convo keys out-of-band?
<Ke>
just too heavy
mallox has joined #nixos
<Ke>
too little sleep for the poor thing
<colemickens>
maybe there's an elegant solution, or I am missing a piece, I just haven't seen it spelled out in a way that's conceptually clicked for me yet
notgne2[m] has joined #nixos
<colemickens>
and there's no "server" if it's p2p...?
<bbigras>
if you are logued first on your computer you can verify the mobile with a QR code or emojis. it's pretty simple
<notgne2>
Ke (chat.freenode.net): you could support phones by having your network support active super nodes
<colemickens>
yeah, I know :)
<colemickens>
I was just, for some reason, thinking that wouldn't work with p2p sessions, but now I can't really explain why.
<notgne2>
Tox does that a little with TCP supernodes, but it's not really enough
<bbigras>
I'm guessing it would be the same with p2p. you would trust your friends keys
<bbigras>
and when they add a new device they would sign the new key or whatever it does
<colemickens>
Hm, ok. If I lose my cookie on my only other active session... ?
alp has joined #nixos
<colemickens>
I guess that's offline backups?
<bbigras>
yeah or the backup on the server
<bbigras>
you can recover the key too with a passphrase. maybe that's the server backup one
tenniscp25 has joined #nixos
<colemickens>
but there's no server in the p2p model? the server is ... in the browser, and stored in local session data?
<colemickens>
that's where I'm stuck.
<notgne2>
colemickens (chat.freenode.net): deterministic encrypted key scattering on the DHT could fix this
<colemickens>
Maybe there's a way to replicate key backups betweens servers?
<notgne2>
it would just need you to enter your password
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Ke>
notgne2 sure there are probably solutions and to me having someone else offloading sounds just like the phone is not a full p2p node anymore, which is probably ok
<notgne2>
you could of course also just throw your encrypted key on the DHT, but then you would need a nice strong password
<bbigras>
oh yeah maybe you would need to save an offline backup. but as long as you still have 1 device with the keys you can add new devices.
<notgne2>
so throwing in deterministic scattering makes it non trivial to pwn users with shitty passwords
<notgne2>
Ke (chat.freenode.net): it's still a part of the "mesh", any PC user could opt-in to taking on extra loads in the network to help support phone users
<bbigras>
I should try running it again. the demo was a bit buggy the last time I tried. I think they were focusing on the dendrite beta.
<notgne2>
I don't think dendrite is usable as a synapse replacement yet is it?
<colemickens>
that's very neat
<bbigras>
it's in beta since a couple of weeks ago
<bbigras>
for people with IOS. there was an official demo too
<bbigras>
dendrite is beta since 1 month ago
<notgne2[m]>
test
<bbigras>
there's the #p2p:matrix.org room if you have question for the devs.
<notgne2>
I think I forgot how slow matrix.org is, the NixOS matrix room just became available but the messages are a few minutes in the past
<notgne2>
their default homeserver is probably turning a lot of people away due to its performance
Izorkin has joined #nixos
<bbigras>
is the perf still that bad? I'm pretty sure they improved it a lot. maybe not enough. I think they are the victim of their success.
<bbigras>
I'm waiting for dendrite or conduit to run my own tiny server.
<notgne2>
every time I send a message here (the IRC channel, via my self hosted IRC Matrix bridge) it takes 20-120 seconds to arrive in the Matrix room (the official IRC bridged room)
<colemickens>
it varies but matrix.org perf keeps me from recommending it to others. (that or element needs more help than I thought)
<notgne2[m]>
at least NixOS makes it easy to self host :^)
<bbigras>
element's UX is getting better I think. I think they got some UX people a while ago and started to make progress.
<bbigras>
yeah
<Ke>
so you guys are not seeing element android freezes
<Ke>
?
<colemickens>
Emil Karlson: I've given up on Element Android and use FluffyChat instead
<Ke>
I was not the only one that had that issue, when I asked
<Ke>
clearing app cache helps for a while, but it goes bad after a while
<notgne2[m]>
element android never freezes for me, just every time you use it the amount of time it takes to send a message doubles until you clear the cache again
<bbigras>
some guy (in the core team I think) had an experimental matrix client. https://github.com/vector-im/hydrogen-web I wonder if it will help improve Element.
kalbasit has quit [Ping timeout: 240 seconds]
<colemickens>
hydrogen is one of the few sorta-multi-modal clients too -_- somehow
<{^_^}>
[nixpkgs] @jonringer pushed commit from @toothbrush to release-20.09 « mercury: 20.06 -> 20.06.1 »: https://git.io/JkI4N
<bbigras>
an emacs one?
<notgne2[m]>
yep, I don't know the name but I know a lot of people seem very happy with it
<bbigras>
> weechat-matrix is stable and quite usable as a daily driver. It already supports large parts of the Matrix protocol, including end-to-end encryption (though some features like cross-signing and session unwedging are unimplemented).
<{^_^}>
error: syntax error, unexpected ',', expecting ')', at (string):345:116
<bbigras>
sorry bot
<bbigras>
they are rewriting the weechat-matrix plugin from python to rust. It might take a while to get features parity.
<notgne2[m]>
I really like weechat, but something feels weird about the state of it as a whole
<notgne2[m]>
it feels like they haven't decided if they want to be a generic chat client or an IRC client
<{^_^}>
[nixpkgs] @jonringer merged pull request #103383 → linuxPackages.wireguard: fix the build on linux 5.4.76 → https://git.io/JkITU
<{^_^}>
[nixpkgs] @jonringer pushed commit from @ivan to master « linuxPackages.wireguard: fix the build on linux 5.4.76 »: https://git.io/JkIBI
<notgne2[m]>
they have a great API for non-IRC platforms, but then the core feature set is limited by IRC still, you need a plugin for multiline inputs even
<notgne2[m]>
and really there's not many chat plugins for it either, so the only way to use it as your client for everything is by using Matrix (or I guess spectrum2 or bitlbee, but those are more limiting)
<bbigras>
uh. I didn't use it in a while.
<Ke>
tbh, I would be happy with irc level of service, if it just works
<Ke>
better try it out soon, I guess
<bbigras>
you mean just irc? I like multiple-devices support of modern chats client. and not having to run an irc bouncer
<Ke>
well mostly I mean, I don't need to support pictures, videos etc. on all clients
numkem has joined #nixos
Fox0919 has quit [Ping timeout: 240 seconds]
<bbigras>
I find pictures support useful for screenshots. but I guess it's nice too when using it with friends: not having to upload the images to a 3rd party.
<bbigras>
well, maybe I could DCC send or whatever it was from my phone to my friend's
<typetetris>
Sometimes I have the problem, that nix doesn't work because some substituter isn't reachable. Couldn't it just print an error message and contiue with the other substituters or even build itself?
<inquisitiv3>
I"m trying to write a nix expression in a `shell.nix` file that will do the same thing as `nix-shell -p postgresql jetbrains.datagrip && ./startdb.sh`.
<{^_^}>
[nixpkgs] @FRidh pushed 5 commits to python-unstable: https://git.io/JkIwL
<{^_^}>
[nixpkgs] @erictapen pushed commit from @ivan to release-20.09 « linuxPackages.wireguard: fix the build on linux 5.4.76 »: https://git.io/JkIwo
<coco>
I'm using i3 and just upgraded to NixOS 20.09. Now I don't see icons anymore (networkmanager applet, seafile applet etc.) Any hints regarding what could be the problem?
<coco>
The applets seem to be running, I just can't see the icons
<das_j>
coco: Maybe `i3-msg restart`?
<das_j>
This happens to my i3 sometimes as well (already did on 20.03)
<coco>
das_j: Nope, restaring i3 doesn't help
<das_j>
oof
<{^_^}>
[nixpkgs] @erictapen pushed commit from @ivan to release-20.03 « linuxPackages.wireguard: fix the build on linux 5.4.76 »: https://git.io/JkIwQ
<coco>
Ah, I got it. My display wasn't primary, that's why they weren't shown. Fixed :-)
turion has joined #nixos
<turion>
NixOps feels unintegrated in the NixOs website to me. There is no link easily reachable from nixos.org, the user guide doesn't have the CSS of the other manuals, https://nixos.org/nixops forwards to Github... was this forgotten during the overhaul, or is it an intentional omission? (If the latter, what's the intention?)
<dminuoso>
turion: Nixops is in a sad place, currently
<dminuoso>
It has a largely undocumented plugin architecture, with largely undocumented plugins, some of which haven't been working in a long while, with no easily found documentation...
<{^_^}>
[mobile-nixos] @eamsden opened pull request #238 → Make the package set a parameter instead of using `<nixpkgs>` everywhere → https://git.io/JkIQz
coco has joined #nixos
graf_blutwurst has quit [Remote host closed the connection]
<FreeVariable>
Hello friends, how about bridging this (or another IRC room) to Telegram (and/or Discord) so as to allow users to join from their favourite client?
<FreeVariable>
Not implying that the current situation is bad, but it's always nice to have some "community"-like platform, whether for user support or simply hanging out.
<FreeVariable>
Setting up a bridge bot is not very difficult -- courtesy of Matrix --, and keeping things orderly requires about 4-5 moderators.
coco has quit [Quit: WeeChat 2.9]
graf_blutwurst has quit [Remote host closed the connection]
Sanchayan has joined #nixos
<typetetris>
Should `docker` find `docker-credential-pass` through the PATH automatically in nixos or do I need to configure something ?
<neiluj>
okay, fixed my issue, there was a missing brace around the argument of import, among other things
<axx>
hmm, just rebooted my rpi4 and all my changes to NixOS were lost (users, packages) but my configuration.nix file is still there, with all my system definitions. I've been using `nixos-rebuild switch` to manage the system, according to the manual "switch" will both activate and make it default for next boot. Is this an issue with the rpi4 not using GRUB as its boot loader?
<clever>
axx: that can happen if you didnt configure the filesystems properly, and it cant find what you said to mount at /
Mic92 has joined #nixos
<clever>
axx: did you run nixos-generate-config?
<axx>
ah, that would make sense given i've told NixOS to mount a BTRFS subvol from a hard drive that gets decrypted and mounted manually
<axx>
i honestly can't remember if i did clever, sorry
<axx>
i basically followed the instructions in the wiki
<clever>
axx: youll need to look at your config and whats under fileSystems., and confirm if those devices are still present and setup right
<axx>
looking at it again, i think i did and then edited the resulting config as suggested by the wiki
mananamenos has joined #nixos
<{^_^}>
[nixpkgs] @GovanifY opened pull request #103461 → pcsx2: update deps needed by upstream → https://git.io/JkLTI
malook has quit [Read error: Connection reset by peer]
<tom39291>
I want to make a patch for a project that is packaged in nixpkgs. It occurs to me that to get build deps and actually build the project, an easy option would be to use the nixpkgs definition to do that. Is there a nix-shell (or similar) way to drop me in an env with a git clone of upstream, suitable for building via nix?
<tom39291>
I could follow the project's build instructions... but why bother when nixpkgs already does this.
wavirc22 has joined #nixos
wavirc22_ has quit [Ping timeout: 240 seconds]
<__red__>
Yes there is, but what do you want to do at the end of your process?
<__red__>
what are you trying to achieve?
<tom39291>
I want to make a change to upstream, test that change (by building a new binary), and send that change to upstream.
<qyliss>
tom39291: nix-shell -A hello '<nixpkgs>'
<qyliss>
in that shell, you have access to a hello build environment
waleee-cl has joined #nixos
<qyliss>
you can run unpackPhase to get the source
<qyliss>
but it won't do stuff like git clone for you, because Nix packages generally use release tarballs
<{^_^}>
[nixpkgs] @maralorn pushed to haskell-updates « haskell-language-server: Fix hydra config for deps »: https://git.io/JkLke
nf has quit [Quit: Fairfarren.]
<tom39291>
qyliss: Perfect, thanks.
<qyliss>
In general, it's often better to write a shell.nix for the project, because e.g. you might want autoconf, which tarballs don't need
<qyliss>
But sometimes -A is fine. Just don't rely on it.
<{^_^}>
[nixpkgs] @FRidh pushed commit from @r-ryantm to staging « cmake: 3.18.2 -> 3.18.4 »: https://git.io/JkLIW
<neiluj>
why should i avoid function composition?
<axx>
<clever "axx: youll need to look at your "> i think i misunderstood ∿ und̷e̷l̷ě̷t̷e̷d̷'s advice yesterday, when they said they used udiskie to unlock and mount an external hard drive in a similar setup to mine: it has to be done as part of the conf file, not manually, right? (the unlocking + mounting)
<dminuoso>
You should not, nix has just poor ergonomics for it
<dminuoso>
in say haskell you'd write `(f . g . h . i) x`
<dminuoso>
but we cant define operators
<dminuoso>
in nix
<dminuoso>
neiluj: let comp = f: g: x: f (g x); in ...
<neiluj>
yes
<dminuoso>
lets you then write
Denommus has quit [Ping timeout: 246 seconds]
<dminuoso>
compL = builtins.foldl' comp id
<dminuoso>
Then you can just do `let f = compL [f, g, h]; in f x`
<dminuoso>
Or, `builtins.pipe x [h, g, f]`
<{^_^}>
[nixpkgs] @FRidh pushed 230 commits to staging-next: https://git.io/JkLIa
<dminuoso>
composition is not nesting parens, composition is the act of composing functions together as a first class action
<{^_^}>
[nixpkgs] @FRidh pushed 275 commits to staging-next: https://git.io/JkLIP
<dminuoso>
so its pkgs.lib.pipe I guess
JoKeRoNe has joined #nixos
<{^_^}>
[nixpkgs] @FRidh opened pull request #103462 → Staging next → https://git.io/JkLI1
<neiluj>
why is this preferrable over, making a simple composition
<dminuoso>
what do you mean by "making a simple composition"
<neiluj>
iterating over lists is faster?
<neiluj>
a function composition
<dminuoso>
what does "function composition" mean to you exactly?
<dminuoso>
You seem to have a different notion than I
<Henson>
say I've got a NodeJS project that I want to make a default.nix build file for. I can't run "npm install" in the buildCommand because it won't let me access the network. I also can't call "node2nix" in the buildCommand for the same reason. Do I have to run node2nix manually and then run nix-build on the resulting default.nix file, or can I somehow get nix-build to do all of those steps?
<neiluj>
hum, so here we're computing f(g(h(x)))
<neiluj>
right?
<dminuoso>
Right
<neiluj>
is the foldl form better?
<Henson>
oops, sorry to interrupt
<neiluj>
instead of just wrinting f(g(h(x))
<dminuoso>
neiluj: sure, or pipe
<dminuoso>
part of the reason is that it reveals your structure better, and it lets you move the composition into a separate binding
<neiluj>
okay thanks! i was doing function specialization ( like i defined a second order function and applied it with an other function)
<neiluj>
as an argument
<neiluj>
not sure if this differs from your point
<neiluj>
dminuoso: right, and it's more general
<dminuoso>
Also, for highly nested calls its much more readable
<dminuoso>
(f ∘ g ∘ h ∘ j ∘ k ∘ l) x
<dminuoso>
For example reveals that you have a sort of pipeline, and how it works.
<dminuoso>
You dont have to mentally keep track of parens.
<neiluj>
do you think this would benefit fronm using pipe?
<dminuoso>
I dont think so
<dminuoso>
You just need more bindings.
<yurb>
axx: udiskie is an applet for mounting / unmounting removable media
<neiluj>
thanks, this clarifies things!
<fling>
Is there a nixos vm image suitable for use with lxd?
<bk1603[m]>
I just can't get virt-manager to work, I added the kernel modules `kvm-amd` and `kvm-intel` enabled `libvirtd` in configuration.nix via `virtualisation.libvirtd.enable=true` and I added dconf via `programs.dconf.enable = true`. But everytime I open `virt-manager` I get the same warning message of not having any hypervisors installed. What should I do?
<dminuoso>
neiluj: Btw, there's an exotic and little known syntax where you can say...
<bk1603[m]>
(I added myself to the `libvirtd` group too)
<neiluj>
it's impressive that you can generate configuration for software like this
<JoKeRoNe>
hi, this might be a noob question, but i managed to break my R package versions it seems. I was using the 20.03 version and updated to unstable, but now the r packages are not working anymore. how do i re-install packages with rWrapper and configuration.nix?
heywoodlh has joined #nixos
<dminuoso>
neiluj: let { body = ...; ... };
<dminuoso>
neiluj: Im quite a big fan of that, because it's closer to haskell where style bindings
<neiluj>
oh if i got filterRule=filterRule; can i simplify this?
<dminuoso>
(I often find it easier to first define an expression, and then specify related bindings)
<neiluj>
in ocaml you can do filterRule;
<dminuoso>
so you could say
<dminuoso>
modules with inputs.nixpkgs; let { body = ...; rule = name: _: (lib.hasSuffix ".nix" name) && !(lib.hasPrefix "sysinfo" name)); };
<Henson>
JoKeRoNe: can you just roll back to the 20.03 version?
<dminuoso>
err
<dminuoso>
modules = with inputs.nixpkgs; let { body = ...; rule = name: _: (lib.hasSuffix ".nix" name) && !(lib.hasPrefix "sysinfo" name)); };
<neiluj>
dminuoso: oh so what does body represent?
<dminuoso>
neiluj: the thing the entire let expression produces
<dminuoso>
also, it turns on rec
<neiluj>
hmm
* dminuoso
hates let with a passion
<dminuoso>
but that makes it bearable
<dminuoso>
But Im Haskell damaged, so..
<neiluj>
i'm confused
<neiluj>
body is the entire expression of what?
<dminuoso>
> { x = let { body = a + b; a = 5; b = 10} };
<{^_^}>
error: syntax error, unexpected '}', expecting ';', at (string):345:40
<dminuoso>
> { x = let { body = a + b; a = 5; b = 10; }; };
<{^_^}>
error: syntax error, unexpected ';', expecting ')', at (string):345:46
<dminuoso>
uh Im bad at this
<gchristensen>
the "body" part of let is a detail of how nix implements the let in syntax and I'm not really sure it is important to know about or considered part of the language "spec" exactly
<neiluj>
let is nice, but i'm ocaml damaged :p
<Henson>
dminuoso: haha, Haskell damaged. I know the feeling.
<neiluj>
oh
<dminuoso>
> ({ x = let { body = a + b; a = 5; b = 10; }; }).x
<{^_^}>
15
<neiluj>
ok!
<dminuoso>
So this lets you visually first describe what x is bound to, and then declare a bunch of related and potentially (mutually) recursive bindings
Denommus` has quit [Remote host closed the connection]
<dminuoso>
Well.. "Quirks"
<dminuoso>
:>
JoKeRoNe has quit [Remote host closed the connection]
<gchristensen>
that ins't really officially documented
<dminuoso>
If you ever take that away, I demand where.
<gchristensen>
lol
<infinisil>
I think the `let { ... }` syntax is just a relic from the past and should probably be removed
<dminuoso>
Then I want my money back.
<gchristensen>
you got it!
sangoma has joined #nixos
<dminuoso>
No seriously, I think where bindings should be a thing in nix.
<dminuoso>
It makes for code that is easier to follow
knupfer1 has joined #nixos
knupfer1 has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @primeos pushed 2 commits to release-20.03: https://git.io/JkLty
<infinisil>
Aren't where bindings just inverse let bindings?
<dminuoso>
Sure, but often its easier to *start* with the expression and then the related bindings
orivej has joined #nixos
<dminuoso>
So you can chase the bindings if you care for it, as deeply as you want
<infinisil>
Hm true
<dminuoso>
With let you're forced to scan and find where the matching expression is. For trivial expressions thats not an issue, but anything non-trivial..
<{^_^}>
[darwin-stubs] @thefloweringash pushed 2 commits to main: https://git.io/JkLt5
<infinisil>
With NixOS modules, there's often a big `let in` block at the start, which can be very distracting from the main code
<dminuoso>
Indeed
bitmapper has joined #nixos
sangoma has quit [Read error: Connection reset by peer]
<dminuoso>
I mean you're a haskell developer. Do you find yourself using where or let more? :)
<dminuoso>
To me, it's often nice to see `f = handleOutput =<< generateOutput where ...` where I instantly have a rough idea what this does. If I care more, I can scroll down and look further
<roberth>
where will behave like an infix operator because of the lack of indentation-based syntax. That's not ideal but can be fixed with more keywords I suppose :/
<gchristensen>
rub more keywords in it
xlei has joined #nixos
ransom has joined #nixos
jonatanb has quit [Ping timeout: 260 seconds]
<Henson>
in Haskell I prefer using where, and only use let when I'm forced to with monads, or if I want to define some temporary things within an expression.
<Henson>
is there a reason why nix repl doesn't like semicolons at the end of lines? It makes copying and pasting from a derivation difficult
<pumpy>
huh? i know it's a whole block i gotta edit but i dont know how and that's part of the problem. i should have a simple nix config to turn off coloring
sangoma has joined #nixos
graf_blutwurst has quit [Remote host closed the connection]
<pumpy>
it's not even a universal thing to color prompt so defaulting on is really bad imo
<dminuoso>
Henson: "at the end of lines"?
<neiluj>
trying to factor code by writing the recurrent functions in a nix file
<dminuoso>
Henson: So roughly, nix doesn't have line based statements? The semicolons are artifacts of let bindings or attributes
<neiluj>
is there a way to import them in a module and pass it to the other modules?
<neiluj>
lieke defining a utils modules from a utils.nix, and call functions: utils.myfunc ...
<{^_^}>
[nixpkgs] @lzhang10 closed pull request #103374 → zfs: Force import non-root ZFS data pool → https://git.io/JkkjV
<dminuoso>
neiluj: the clean way is to use an option
<dminuoso>
that is, stuff the function into an option
<dminuoso>
Cant remember the details actually, but Ive had some discussions in here about this
<dminuoso>
`import ./foo` is generally bad for nixos
<dminuoso>
Part of the problem is that it can lead to double imports if you import by path.
captn3m0 has joined #nixos
<dminuoso>
oh no, wait, that was about something different
veleiro has quit [Ping timeout: 265 seconds]
<dminuoso>
that was about function values in imports
<Unode>
hi everyone, I'm n
<Unode>
oops
<Unode>
I'm currently experiencing some wifi issues and DHCP is defaulting to assigning a 169. IP to my connection. Does anyone know if there's an option to disable this behavior?
<axx>
<yurb "axx: udiskie is an applet for mo"> thanks, that i had understood :) I thought you were calling manually, after NixOS boot, to unlock and mount your ext. hard drive, and that the `fileSystems` part of the conf was capable of dealing with a disk being absent on boot
eahlberg has joined #nixos
graf_blutwurst has joined #nixos
<DavHau[m]>
If I reference a flake in cmdline via git url like for example `git+ssh://git@gitlab.com/company/project#package1`,
<DavHau[m]>
Is it possible to define the git branch or revision to use?
<yurb>
axx: that is correct
<yurb>
I only connect the drive occasionally so I mount it manually with udiskie
<FRidh>
DavHau[m]: yes, use ref? at the end of the url
graf_blutwurst has quit [Remote host closed the connection]
NaN-irc_ has joined #nixos
<FRidh>
ref?branch
<yurb>
But it adheres to that I have in `fileSystems`
<DavHau[m]>
<FRidh "$ nix shell git+https://github.c"> Thanks. Sadly for ssh this doesn't seem to work
<pumpy>
why doesn't programs.bash.enableLsColors = false; not turn off the blue and green file/dir colors pls?
heywoodlh has quit [Ping timeout: 260 seconds]
whatisRT has joined #nixos
heywoodlh has joined #nixos
sss2 has joined #nixos
<Valodim[m]>
does anyone know how flake-built systems can interact with the nixpath?
<Valodim[m]>
concretely: for a system built from a flake, can I pin the `<nixpkgs>` channel that's resolved from NIX_PATH (e.g. in non-flake `shell.nix` files) to the system nixpkgs?
<pumpy>
got more info, is this a nixos bug? why doesn't programs.bash.enableLsColors = false; not turn off the yellow, blue and green file/dir colors pls? `type -a ls` outputs alias ls='ls --color=auto' so i guess i can alias that back but shouldn't enableLsColors = false be doing that?
<lordcirth>
pumpy, did you open a new shell since switching?
<pumpy>
ya even rebooted
<axx>
ok, so the wiki says to define the filesystems for the rpi4 in configuration.nix, after having run `nixos-generate-config`, however, that generates a hardware.nix file that 1/ duplicates a definition for `/ and 2/ explicitly says not to edit the file, as it may be overwritten in the future, and to make changes to configuration.nix instead. What's the recommendation here, just skip defining the root fs in configuration.nix and
<axx>
stick to how it was automatically defined in hardware.nix?
<DavHau[m]>
<DavHau[m] "Thanks. Sadly for ssh this doesn"> I was wrong. It also works with ssh. Thanks again!
<axx>
* ok, so the wiki says to define the filesystems for the rpi4 in configuration.nix, after having run `nixos-generate-config`, however, that generates a hardware.nix file that 1/ duplicates a definition for `/` and 2/ explicitly says not to edit the file, as it may be overwritten in the future, and to make changes to configuration.nix instead. What's the recommendation here, just skip defining the root fs in configuration.nix and
<axx>
stick to how it was automatically defined in hardware.nix?
eeva has quit [Ping timeout: 265 seconds]
<lordcirth>
pumpy, do you have home-manager enabled?
<pumpy>
ya
<gchristensen>
pumpy: the NixOS option enableLsColors doesn't cause "alias ls='ls --color=auto'" to be setup
<pumpy>
i dont have a .bashrc created fwiw
<pumpy>
oh
<gchristensen>
all it does is export LS_COLORS=''
<pumpy>
how can i make the alias not be set?
<gchristensen>
dunno, step one is finding out what is setting it :)
<srhb>
pumpy: grepping through nixpkgs is usually helpful
<pumpy>
gchristensen btw fwiw, i am really liking nixos but some of these defaults are really bad decisions. like having shell colors by default. if someone wants color they won't get turned off by having to enable it. but if you default to color, now you gotta pick palette, and you'll turn off a lot more ppl. especially since it's so hard (impossible?) to
<lordcirth>
pumpy, my system has "ls is aliased to `ls --color=tty'"
<pumpy>
disable it with nix
<srhb>
pumpy: You should quickly find this.
<pumpy>
how do i grep through nixpkgs?
<lordcirth>
I used git grep and found where --color=tty is set
<lordcirth>
But I don't see where you are getting color=auto
<srhb>
pumpy: check out a clone of nixpkgs and grep -r "color=tty"
leungbk has quit [Ping timeout: 260 seconds]
<srhb>
oh, auto
<pumpy>
i dont have the repo local i don't think?
<lordcirth>
git grep is better
<gchristensen>
pumpy: yeah, I don't think this is a default since Nixpkgs doesn't do set an alias to --color=auto
<srhb>
Yes, that's unlikely to be nixpkgs
<pumpy>
gchristensen but nothing in my .nix configs enable coloring so it has to be a default
<srhb>
lordcirth: fyte me. grep is fine.
<gchristensen>
well, it isn't
<AmandaC>
You said you're using home-manager as well? It might be what's doing it. Home-manager isn't nixos, it's community-managed
<pumpy>
ill ask them
<srhb>
pumpy: I recommend having local clones of both home-manager and nixpkgs
<pumpy>
i still dont know what im supposed to grep
<pumpy>
why can't i just search github?
<srhb>
pumpy: You're supposed to find the file in which the alias is defined
<AmandaC>
github search is a mess
<srhb>
pumpy: Because _that_ search is pretty bad.
<lordcirth>
pumpy, it's not in nixpkgs anyway. But it's a good thing to try first in general.
<AmandaC>
IT'll return results that havn't been in the main branch for months, or more
<pumpy>
lordcirth where'd you find it?
<lordcirth>
pumpy, find what? I git grepped nixpkgs for '--color'
<lordcirth>
Nothing sets --color=auto globally
<pumpy>
pumpy, my system has "ls is aliased to `ls --color=tty'"
<pumpy>
that
<gchristensen>
before you said: `type -a ls` outputs alias ls='ls --color=auto'
<FRidh>
DavHau[m]: ok got it orking now $ nix shell git+ssh://git@github.com/FRidh/make-binary-wrapper.git?ref=macos#make-binary-wrapper
<pumpy>
gchristensen ya and?
<FRidh>
and $ nix flake info git+ssh://git@github.com/FRidh/make-binary-wrapper.git?ref=macos
<gchristensen>
well now it is --color=tty?
<lordcirth>
pumpy, I just did "type -a ls" like you did
<FRidh>
DavHau[m]: ah saw now you confirmed it works, good
<pumpy>
what im trying to figure out is how to turn that aliasing off
<pumpy>
i dont know where it's happening
ransom has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pumpy>
i just want ls to be ls ffs
<srhb>
pumpy: We realize you're frustrated, but people are trying to teach you to fish here. :)
mike[m]1 has left #nixos ["User left"]
<gchristensen>
well, it isn't `ls --color=tty` like you said a few minutes ago, or `ls --color=auto` like you said 15 minutes ago?
<pumpy>
it's --color=tty
<gchristensen>
:/
<lordcirth>
gchristensen, "tty" was him quoting me
<lordcirth>
Without quotes
<lordcirth>
Thus the highlight of himself
eahlberg has quit [Quit: Lost terminal]
<dutchie>
environment.shellAliases seems to be the option you need to change
simonpe^^ has quit [Remote host closed the connection]
<vika_nezrimaya>
How to get a native stdenv (declared in `nixpkgs/pkgs/stdenv/native/`) for experimentation purposes? I want to poke around the stdenv bootstrapping system a little bit :3
heyitsrama has quit [Ping timeout: 272 seconds]
zupo has joined #nixos
AmandaC has joined #nixos
seku has joined #nixos
Sanchayan has joined #nixos
Sanchayan has quit [Client Quit]
<Henson>
vika_nezrimaya: you can download your own nixpkgs, make changes to stdenv, and then import your custom nixpkgs for experimentation.
<vika_nezrimaya>
Nah, I want to explore how stdenvs get made
<vika_nezrimaya>
Not use that stdenv for something
<vika_nezrimaya>
Also I tried to replicate the bootstrap-tools tarball but it doesn't seem to be bit-by-bit reproducible, probably because of the version drift?
<vika_nezrimaya>
Can we even reproduce our bootstrap seeds for *-linux platforms?
alp has quit [Ping timeout: 260 seconds]
<vika_nezrimaya>
wait, wouldn't a binary-reproducible tarball be an overengineered quine with extra steps?
<vika_nezrimaya>
oh no
veleiro has joined #nixos
wnklmnn has joined #nixos
urkk has joined #nixos
vandenoever has quit [Read error: Connection reset by peer]
<pumpy>
dutchie i tried setting shellAliases = null but that was error. i didn't try = mkForce {} i just set null on the default aliases
<pumpy>
dutchie++
<{^_^}>
dutchie's karma got increased to 4
__monty__ has joined #nixos
<neiluj>
got some extensions for vscode to update
<pumpy>
in my configuration.nix i null the 3 default environment.shellAliases and that seems to work. but then in home manager home.nix i set my own programs.bash.shellAliases but it's coming up not found. what i doing wrong pls?
<neiluj>
is there a way to programmatically update them?
malook has quit [Read error: Connection reset by peer]
<neiluj>
the extensions are specified with vscode-utils.extensionsFromVscodeMarketplace
<{^_^}>
[nixpkgs] @roberth pushed to staging-20.09 « ghc: Force a rebuild on darwin for NixOS/hydra#830 »: https://git.io/JkLBv
<Henson>
is it possible to call nix-build inside of a mkDerivation's buildCommand block? I'm trying to build a NodeJS React program. I've figured out that I can call node2nix with the lock file such that it doesn't require network access. This generates a default.nix file. How can I use this default.nix file inside of another derivation, which is responsible for generating the default.nix file?
veleiro has joined #nixos
boxscape has quit [Ping timeout: 260 seconds]
<Henson>
I'm trying to avoid intermediate steps or a script that builds the derivation in steps, and would like to have a derivation check out the original source code and produce an output at the end of the day
<{^_^}>
[nixpkgs] @roberth pushed to revert-96767 « Revert "firefox: set build timeout at eight hours" »: https://git.io/JkLBA
nDuff has quit [Quit: zzz]
<{^_^}>
[nixpkgs] @roberth opened pull request #103474 → Revert "firefox: set build timeout at eight hours" → https://git.io/JkLRU
rajivr has quit [Quit: Connection closed for inactivity]
hnOsmium0001 has joined #nixos
malook has joined #nixos
vika_nezrimaya has quit [Ping timeout: 256 seconds]
veleiro` has joined #nixos
<roberth>
Henson: maybe with recursive nix, but I don't think it's worth the cost of recursive nix when default.nix is just a boilerplate that you could provide as part of your (generic?) Nix code
veleiro has quit [Ping timeout: 260 seconds]
<gchristensen>
I would love for more people to use recursive nix, but it gives me so many problems
dbmikus__ has joined #nixos
jonatanb has joined #nixos
dmedinag has joined #nixos
coco has joined #nixos
<Henson>
roberth: should I define a sub-derivation as an expression in a let block in the main default.nix file, and then use it in the buildCommand?
<Henson>
roberth: I suppose the equivalent would be creating sub-derivation .nix files and just callPackaging them from the main derivation.
jonatanb has quit [Remote host closed the connection]
<Henson>
roberth: the main problem is that the default.nix file contains a stdenv.mkDerivation whose buildCommand calls a function that generates a build.nix file, and it is this file that I want to then call nix-build on. That means either calling nix-build from within a derivation, or some other way of building this generated derivation that I don't know of
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
h0m1 has quit [Quit: WeeChat 2.9]
eahlberg has joined #nixos
h0m1 has joined #nixos
<clever>
Henson: import from derivation
astronavt has quit [Ping timeout: 256 seconds]
<clever>
Henson: let foo = callPackage ./default.nix; in callPackage "${foo}/build.nix" {}
* fzakaria
waves
* Henson
waves back
<roberth>
Henson: if that `build.nix` file is predictable, you can inline it / import it / callPackage it. Otherwise, you can IFD. Recursive Nix is last resort, but shouldn't be necessary
<Henson>
roberth: ok. I've never heard of import from derivation before, but I see there's a section on the wiki about it. Conceptually that's one avenue I thought would work, but wasn't sure if that was just a hack or the right way to do it, as you have an intermediate derivation that doesn't do anything other than encapsulate the code from another derivation.
zupo has joined #nixos
<Henson>
code from another -> code for another
<roberth>
IFD is subtle in the sense that you should avoid it if possible, but it can be the only solution in some cases
<maralorn>
jophish: Thank you for asking. But I am already finished. ;-)
<maralorn>
Or rather I was as of yesterday evening.
<evanjs>
anybody else failing to compile random rust crates right now? I stumbled across https://github.com/rust-lang/futures-rs/issues/2242, and that _does_ seem to be resolved with e.g. futures-0.3.8, but now crossbeam_utils is failing to compile, complaining about some GLIBC weirdness...
<{^_^}>
rust-lang/futures-rs#2242 (by jbaublitz, 2 weeks ago, open): futures-macro (pulled in by futures-util) appears not to compile on musl
<evanjs>
: version `GLIBC_2.32' not found (required by /home/evanjs/src/rjg/leapline/target/debug/deps/libconst_fn-60e096b97088955b.so) --> /home/evanjs/.cargo/registry/src/github.com-1ecc6299db9ec823/crossbeam-utils-0.8.0/src/atomic/atomic_cell.rs:15:5
boxscape has joined #nixos
<evanjs>
forgot crossbeam has a Discord. I suppose I could check there first
<zarkone>
hi everyone! i jsut installed nixos 20.09 on my new machine and copied home page from the previous one: it had ~/.config/nixpkgs/overlays which used to work perfectly for me. But now when I try to install it with `nix-env -iA`, it complains that attribute is not found. What am I missing and any tips on what should I check? thanks!
<{^_^}>
[nixpkgs] @kalbasit merged pull request #103396 → terraform: withPlugins should inherit meta from wrapped derivation → https://git.io/JkIYk
<{^_^}>
[nixpkgs] @kalbasit pushed to master « terraform: withPlugins should inherit meta from wrapped derivation (#103396) »: https://git.io/JkLri
<{^_^}>
[nixpkgs] @flokli pushed to remove-bluespec « bluespec: remove, as it's unmaintained in nixpkgs »: https://git.io/JkLrP
<deni>
anyone got and idea why a program work when ran from a shell but complains when executed from a systemd service? I'm getting the error: Error fetching client: failed to get token helper: error expanding config path "": exec: "getent": executable file not found in $PATH
<deni>
so apparently getent somehow isn't on the path...although it obviously is in the shell when I run the ExecStart line manually
<fzakaria>
roberth: I was investigating a problem with JNR (Java Native) using dlopen to load libc. It ended up pulling in the wrong libc and causing SIGSEGV
fendor has joined #nixos
<fzakaria>
I solved it by doing `-Djava.library.path=/nix/store/bdf8iipzya03h2amgfncqpclf6bmy3a1-glibc-2.32/lib` but that's kind of annoying given that the ELF binary for Java already is pulling that in.
<fzakaria>
Maybe this is just problematic of dlopen itself
<fzakaria>
I think this should be set in the JVM derivation OVERRIDE_LIBPATH
Supersonic has joined #nixos
<aminechikhaoui>
fzakaria I think java.library.path is for the directories where to lookup the JNI libs to load but I guess if after the dlopen is done your JNI lib is doing calls to the libc from the system path then maybe the JNI library (the .so file) has the wrong rpath ?
aasg has quit [Quit: killed]
veleiro` has joined #nixos
<fzakaria>
Yea and it hydrates that list of paths to system directories (/lib) by default
<fzakaria>
during compilation of the Nix JVM we should override that to be the /nix/store entries
<aminechikhaoui>
but I suspect that shouldn't impact libc right ? only if you're loading libFoo.so as JNI lib it would be looked up from those paths by default
<aminechikhaoui>
fzakaria yeah now I understand what you're saying but it seems like a really unique use case to https://github.com/jnr/jnr-netdb, so maybe pkgs that use it should make sure to have the wrapper add -Djava.library.path. But I don't have any strong opinions about it :)
<flox>
can someone help me find the magic override that will allow me to build all pkgs with clang instead of gcc in my nix flake when using nixpkgs as an input? Can't seem to figure it out
<tpw_rules>
it does not appear that the nix command supports --add-root?
sss2 has joined #nixos
<axx>
any package i try to install using `nix-env -iA` leads to:
<{^_^}>
[nixpkgs] @chkno opened pull request #103500 → doc: 20.09 release notes: nixos-YY.MM branches no longer in nixos-channels repo → https://git.io/JkLDT
sss2 has quit [Remote host closed the connection]
<axx>
* any package i try to install using `nix-env -iA` leads to something like:
<axx>
> error: attribute 'nixpkgs' in selection path 'nixpkgs.duf' not found
<{^_^}>
error: syntax error, unexpected $undefined, expecting ')', at (string):345:18
sss2 has joined #nixos
<axx>
what's going on? i'm seeing something about how nixpkgs does not exist in the top-level namespace, still trying to wrap my head around that :)
oxalica1 has joined #nixos
oxalica has quit [Ping timeout: 260 seconds]
oxalica1 is now known as oxalica
eahlberg has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @maralorn pushed to haskell-updates « haskellPackages.hls-hlint-plugin: Fix build on hydra »: https://git.io/JkLD1
flox has quit [Remote host closed the connection]
la-jesystani has joined #nixos
<woffs>
axx, try the channel name instead of nixpkgs, and look at $NIX_PATH in your environment
<colemickens>
axx: what woffs said, and `nix-channel --list` will help give you an idea of what channels you have
<la-jesystani>
howdy, im using sxhkd with home-manager, however whenever i boot my config doesnt work until i kill and start it manually, upon which its fine. im enabling it with services.sxhkd in my home.nix. as a side note, is there a way to specify the flags a service starts with? i need sxhkd to have `-m 1` every time its run
<colemickens>
`nix-channel --update` if you haven't in a while too
<{^_^}>
[nixpkgs] @roberth opened pull request #103501 → setup.sh: Support XDG_DATA_DIRS → https://git.io/JkLys
neiluj has quit [*.net *.split]
pushqrdx has quit [*.net *.split]
kalbasit has quit [*.net *.split]
expelledboy has quit [*.net *.split]
stoile has quit [*.net *.split]
mbrgm has quit [*.net *.split]
gthm has quit [*.net *.split]
woffs has quit [*.net *.split]
noonien has quit [*.net *.split]
rtjure_ has quit [*.net *.split]
abathur has quit [*.net *.split]
enteee has quit [*.net *.split]
fling has quit [*.net *.split]
akaWolf has quit [*.net *.split]
plp has quit [*.net *.split]
tsrt^ has quit [*.net *.split]
jeaye has quit [*.net *.split]
rail has quit [*.net *.split]
erhandsome has quit [*.net *.split]
jbo has quit [*.net *.split]
Aleksejs has quit [*.net *.split]
chipb has quit [*.net *.split]
c_wraith has quit [*.net *.split]
hexagoxel has quit [*.net *.split]
Siyo has quit [*.net *.split]
lovesegfault has quit [*.net *.split]
terrorjack has quit [*.net *.split]
Philonous has quit [*.net *.split]
nisstyre has quit [*.net *.split]
nightmared has quit [*.net *.split]
otavio__ has quit [*.net *.split]
Streetwalrus has quit [*.net *.split]
agb has quit [*.net *.split]
shapr has quit [*.net *.split]
afics has quit [*.net *.split]
nbathum has quit [*.net *.split]
Guest60204 has quit [*.net *.split]
greymalkin has quit [*.net *.split]
zemm has quit [*.net *.split]
tomaw has quit [*.net *.split]
DerGuteMoritz has quit [*.net *.split]
xe4 has quit [*.net *.split]
fiddlerwoaroof has quit [*.net *.split]
heath has quit [*.net *.split]
jkarni has quit [*.net *.split]
Luker has quit [*.net *.split]
lordcirth_ has joined #nixos
pushqrdx has joined #nixos
rtjure_ has joined #nixos
plp has joined #nixos
fling has joined #nixos
akaWolf has joined #nixos
DerGuteMoritz has joined #nixos
lovesegfault has joined #nixos
fiddlerwoaroof has joined #nixos
terrorjack has joined #nixos
heath has joined #nixos
expelledboy has joined #nixos
xe4 has joined #nixos
afics has joined #nixos
greymalkin has joined #nixos
c_wraith has joined #nixos
mbrgm has joined #nixos
nightmared has joined #nixos
nisstyre has joined #nixos
gthm has joined #nixos
Philonous has joined #nixos
shapr has joined #nixos
tomaw has joined #nixos
Siyo has joined #nixos
noonien has joined #nixos
kalbasit has joined #nixos
Guest60204 has joined #nixos
Streetwalrus has joined #nixos
jbo has joined #nixos
jkarni has joined #nixos
rail has joined #nixos
Aleksejs has joined #nixos
Luker has joined #nixos
enteee has joined #nixos
zemm has joined #nixos
abathur has joined #nixos
erhandsome has joined #nixos
neiluj has joined #nixos
hexagoxel has joined #nixos
woffs has joined #nixos
otavio__ has joined #nixos
chipb has joined #nixos
agb has joined #nixos
jeaye has joined #nixos
nbathum has joined #nixos
stoile has joined #nixos
seku has quit [Quit: Connection closed]
<tpw_rules>
okay so i've sort of got it working by using buildEnv instead of mkShell. but then develop still does not create the GC root, i have to use build (which does not do anything) to do it. is there a better way?
afics has quit [Max SendQ exceeded]
nisstyre has quit [Max SendQ exceeded]
terrorjack has quit [Max SendQ exceeded]
lordcirth__ has quit [Ping timeout: 264 seconds]
tsrt^ has joined #nixos
neiluj has quit [Read error: Connection reset by peer]
nisstyre has joined #nixos
lordcirth__ has joined #nixos
terrorjack has joined #nixos
<tpw_rules>
not sure if this is the right approach for what i want
afics has joined #nixos
lordcirth has quit [Ping timeout: 264 seconds]
cole-h has quit [*.net *.split]
dbmikus__ has quit [*.net *.split]
leungbk has quit [*.net *.split]
acarrico has quit [*.net *.split]
numkem has quit [*.net *.split]
viric has quit [*.net *.split]
spudly has quit [*.net *.split]
vk3wtf has quit [*.net *.split]
grumble has quit [*.net *.split]
amfl has quit [*.net *.split]
CMCDragonkai has quit [*.net *.split]
night has quit [*.net *.split]
sulu has quit [*.net *.split]
icey_ has quit [*.net *.split]
hauleth has quit [*.net *.split]
rawtaz has quit [*.net *.split]
kwork has quit [*.net *.split]
chiiba has quit [*.net *.split]
__red__ has quit [*.net *.split]
_deepfire has quit [*.net *.split]
konobi has quit [*.net *.split]
suchar has quit [*.net *.split]
NieDzejkob has quit [*.net *.split]
marcusr has quit [*.net *.split]
notgne2 has quit [*.net *.split]
puck has quit [*.net *.split]
rodgort has quit [*.net *.split]
HackerFoo has quit [*.net *.split]
mzan has quit [*.net *.split]
maljub01 has quit [*.net *.split]
ajp has quit [*.net *.split]
hspak has quit [*.net *.split]
bgamari has quit [*.net *.split]
awaz has quit [*.net *.split]
utsl has quit [*.net *.split]
tldr32 has quit [*.net *.split]
nek0 has quit [*.net *.split]
buffet has quit [*.net *.split]
mxco86 has quit [*.net *.split]
stiell has quit [*.net *.split]
wolfshappen has quit [*.net *.split]
siers has quit [*.net *.split]
pjan has quit [*.net *.split]
aminechikhaoui has quit [*.net *.split]
evhan has quit [*.net *.split]
outerpassage has quit [*.net *.split]
evils has quit [*.net *.split]
mkaito- has quit [*.net *.split]
gambpang_ has quit [*.net *.split]
pbogdan has quit [*.net *.split]
Mirrexagon has quit [*.net *.split]
luigy has quit [*.net *.split]
clever has quit [*.net *.split]
spudly- is now known as spudly
justanotheruser has quit [*.net *.split]
civodul has quit [*.net *.split]
yeongsheng_tan[m has quit [*.net *.split]
GuillaumeChrel[m has quit [*.net *.split]
manveru[m] has quit [*.net *.split]
blindidiotgod has quit [*.net *.split]
lasadd[m] has quit [*.net *.split]
daddy_james[m] has quit [*.net *.split]
miangraham[m] has quit [*.net *.split]
felschr[m] has quit [*.net *.split]
JaakkoLuttinen[m has quit [*.net *.split]
ugubok[m] has quit [*.net *.split]
zane has quit [*.net *.split]
IslandUsurper[m] has quit [*.net *.split]
txt-file has quit [*.net *.split]
goodnight_noom[m has quit [*.net *.split]
ilya-fedin has quit [*.net *.split]
Hirmes[m] has quit [*.net *.split]
TheSirC[m] has quit [*.net *.split]
ma27[m] has quit [*.net *.split]
cyberwolf[m] has quit [*.net *.split]
Skallwar has quit [*.net *.split]
peterhoeg has quit [*.net *.split]
danielrf[m] has quit [*.net *.split]
alexarice[m] has quit [*.net *.split]
mt[m] has quit [*.net *.split]
regnat has quit [*.net *.split]
siraben has quit [*.net *.split]
Ke has quit [*.net *.split]
yangm has quit [*.net *.split]
nilsirl[m] has quit [*.net *.split]
bohan[m] has quit [*.net *.split]
fgaz has quit [*.net *.split]
ptotter[m] has quit [*.net *.split]
mla has quit [*.net *.split]
solene has quit [*.net *.split]
pinkieval has quit [*.net *.split]
develCuy has quit [*.net *.split]
S0rin has quit [*.net *.split]
MinceR has quit [*.net *.split]
komasa has quit [*.net *.split]
IRCsum has quit [*.net *.split]
evax has quit [*.net *.split]
garbas has quit [*.net *.split]
asymmetric has quit [*.net *.split]
liori has quit [*.net *.split]
doublej41 has quit [*.net *.split]
eta has quit [*.net *.split]
<{^_^}>
[nixpkgs] @mmahut merged pull request #99212 → openethereum: linux -> unix, remove systemd from buildInputs → https://git.io/JUyP0
Darkmatter66_ has quit [Quit: ZNC 1.7.5 - https://znc.in]
r0bby has joined #nixos
omnigoat has joined #nixos
diamondman has joined #nixos
raoul has joined #nixos
Cynthia has joined #nixos
jb55 has joined #nixos
feepo has joined #nixos
Darkmatter66 has joined #nixos
betawaffle has joined #nixos
<zebrag>
On my debian box, I've got 6 different package managers: pip, npm, melpa, stack (haskell), opam, debian. I want to move it on a nixos box. last time I tried to do it, I had issues with "npm", even to the point, some packages/executables weren't finding the `so`-libs they wanted.
<zebrag>
Can I have the transfer automated? ;)
awwayaya24hours[ has joined #nixos
ilya-fedin has joined #nixos
goibhniu has joined #nixos
Notkea[m] has joined #nixos
GuillaumeChrel[m has joined #nixos
zane has joined #nixos
Skallwar has joined #nixos
goodnight_noom[m has joined #nixos
jdks[m] has joined #nixos
<{^_^}>
[nixpkgs] @bennyandresen opened pull request #103505 → release-20.09: cherrypicked graalvm-ce, clj-kondo + babashka from master → https://git.io/JkL9W
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to staging: https://git.io/JkL94
<{^_^}>
[nixpkgs] @flokli closed pull request #103487 → bluespec: remove, as it's unmaintained in nixpkgs → https://git.io/JkLrQ
<typetetris>
I have this in some expression to work around a non deterministic build of `texlive.combine`. Shouldn't `tex` never be substituted now? Somehow my ci server still try to substitute the resulting derivation. https://www.irccloud.com/pastebin/nSG4xb87/
<zebrag>
la-jesystani: sort of; it was 6 month ago though; also I was told not to procrastinate, to make it short.
ncm[m] has joined #nixos
dmj` has joined #nixos
M0[m] has joined #nixos
jojosch[m] has joined #nixos
fatjedi[m] has joined #nixos
betux[m] has joined #nixos
sarcasticdream[m has joined #nixos
Rian[m] has joined #nixos
eadwu[m] has joined #nixos
Jezhehim[m] has joined #nixos
pheoxy has joined #nixos
cirno[m] has joined #nixos
rednaZ[m] has joined #nixos
shu9 has joined #nixos
tilcreator has joined #nixos
DanP[m] has joined #nixos
humancalico[m] has joined #nixos
us3r[m] has joined #nixos
domenkozar[m] has joined #nixos
yoctocell[m] has joined #nixos
GerdFlaig[m] has joined #nixos
<la-jesystani>
pumpy: i had a similar problem, try checking the home-configuration.nix man page and seeing if its listed there? i think a version didnt have a bunch of the xsession options
Valodim[m] has joined #nixos
alexfmpe has joined #nixos
ryantm has joined #nixos
artturin has joined #nixos
symphorien[m] has joined #nixos
bielgeets[m] has joined #nixos
<pumpy>
recently?
ylsdfh27[m] has joined #nixos
Hm7000 has joined #nixos
<pumpy>
like 20.09
qbit[m] has joined #nixos
raoul has quit [Read error: Connection reset by peer]
<la-jesystani>
im on revision 7339784e07217ed0232e08d1ea33b610c94657d8 of home-manager and the options are working again
manveru[m] has joined #nixos
JaakkoLuttinen[m has joined #nixos
felschr[m] has joined #nixos
blindidiotgod has joined #nixos
daddy_james[m] has joined #nixos
regnat has joined #nixos
miangraham[m] has joined #nixos
IslandUsurper[m] has joined #nixos
peterhoeg has joined #nixos
ugubok[m] has joined #nixos
bohan[m] has joined #nixos
danielrf[m] has joined #nixos
wavirc22 has quit [Ping timeout: 272 seconds]
alexarice[m] has joined #nixos
lasadd[m] has joined #nixos
Ke has joined #nixos
ptotter[m] has joined #nixos
nilsirl[m] has joined #nixos
abstract300 has joined #nixos
jonathan[m]3 has joined #nixos
barn0 has joined #nixos
georgyo[m] has joined #nixos
mt[m] has joined #nixos
Ericson2314 has joined #nixos
nikivi[m] has joined #nixos
siraben has joined #nixos
hsiktas[m] has joined #nixos
<tpw_rules>
using the --profile switch does not work for nix develop because it doesn't mark some stuff. so if i collect-garbage i can't use nix develop without an internet connection
colemickens has joined #nixos
wavirc22 has joined #nixos
alexherbo2 has joined #nixos
ThaEwat has joined #nixos
SpaghettiCthulhu has joined #nixos
TheSirC[m] has joined #nixos
ndarilek has joined #nixos
omasanori[m] has joined #nixos
Synth[m] has joined #nixos
<colemickens>
axx: yeah, you don't want to do that
<colemickens>
axx: instead, register a channel for your regular user and reference it
<colemickens>
axx: it's a bit weird too, `-f <nixos>` is just incorrect, `-f <nixpkgs>` would be telling nix-env to use <nixpkgs> from NIX_PATH, and then would look for a derivation names "nixpkgs.duf" rather than "duf".
<pumpy>
la-jesystani i dont get syntax error anymore but i don't see the i3.extraConfig in the generated config
sum1[m] has joined #nixos
michaelpj has joined #nixos
axx has joined #nixos
porphyrogenetos has joined #nixos
<pumpy>
know anything about that?
<pumpy>
la-jesystani++
<{^_^}>
la-jesystani's karma got increased to 1
danderson[m] has joined #nixos
leonardp has joined #nixos
Sabbat[m] has joined #nixos
notgne2[m] has joined #nixos
Fafnir[m] has joined #nixos
alan[m]1 has joined #nixos
tristan[m] has joined #nixos
tyrion-mx has joined #nixos
ThornAvery[m] has joined #nixos
sshow[m] has joined #nixos
witchof0x20[m] has joined #nixos
kraem has joined #nixos
jlv[m] has joined #nixos
tet[m] has joined #nixos
risson[m] has joined #nixos
mindtree[m] has joined #nixos
timclassic has joined #nixos
timokau[m] has joined #nixos
puzzlewolf has joined #nixos
NobbZ[m] has joined #nixos
davidak[m] has joined #nixos
<axx>
thanks a lot for your explanation
aanderse has joined #nixos
nh2[m] has joined #nixos
jojosch|Johannes has joined #nixos
esclear_m has joined #nixos
fzakaria1 has joined #nixos
betrion[m] has joined #nixos
<colemickens>
hopefully the command I gave (works and) makes sense :)
<{^_^}>
[nixpkgs] @jonringer pushed commit from @r-ryantm to master « k9s: 0.23.3 -> 0.23.10 »: https://git.io/JkLQ0
<axx>
am i to understand that when defining a user in configuration.nix, you still have ta manually add a channel for that user? (and then update it 😉)
<pushqrdx>
so there're only 2 remaining mini pain points that are standing between me and a fully migrating to linux, i know it's not nixos specific but i would love someone's help/hints
<pushqrdx>
the bigger and more important one is being able to route and apply vst/audio effects at low latency on the audio output
<pumpy>
im using i3 and i see some stuff in the generated configi dont understand. it's "client.x y" where x is a window state like focused or urgent and y is 1+ hex colors. what is that?
<pumpy>
nvm
fzakaria1 has left #nixos ["User left"]
bridge[evilred] has quit [Remote host closed the connection]