<excelsiora>
gchristensen: I read up on the evolution of mypy's treatment of `def __init__(self):` (looks like there's 3 cases in your branch) and looks like their resolution is that it's not considered covered if there are no declared types - do you want to add `-> None:` in that case, all cases, or leave it off (expecting to declare it covered when it's so configurable)?
<excelsiora>
(if/when)
<gchristensen>
excelsiora: what do you think is the best choice to make the code safest?
Arahael has quit [Ping timeout: 260 seconds]
<excelsiora>
In this context, I think safest is no noise when running checks. Adding it explicitly where not required... future churn/discussion possibilities regardless. I suppose I wouldn't otherwise add it explicitly to reduce code churn today.
<excelsiora>
Therefore I'd only add it in those three cases
<gchristensen>
which three cases are you thinking?
<excelsiora>
Python checks __init__() at run time to ensure it returns None... mypy check would catch it sooner though...
<gchristensen>
okay, sure, let's go ahead and add it
<gchristensen>
I think adding it to the others would be fine too, I don't really consider it churn but an investment in the code :)
aborsu has quit [Quit: aborsu]
<excelsiora>
ok, will add it to other cases opportunistically as well, mypy would catch issues before runtime but it does consider `def __init__(self, arg: type):` covered for -> None...
<MichaelRaskin>
It's not just unfree, it involves Nvidia and drivers…
<OmnipotentEntity>
I literally just grabbed colmap out of a bag at random because it was a cuda enabled package because I wanted to check if my problem was my configuration or my shell.nix for this one thing.
<OmnipotentEntity>
and it didn't work lol
mbrgm_ has joined #nixos
<OmnipotentEntity>
looks like the package broken between Nov 15 and Nov 20.
<OmnipotentEntity>
I can fix it if it's just an issue with cmake. But I'm thinking my CUDA setup is garbage, so I wouldn't be able to test it.
mbrgm has quit [Ping timeout: 240 seconds]
mbrgm_ is now known as mbrgm
<OmnipotentEntity>
MichaelRaskin, yeah, :/ But I do need it.
<excelsiora>
comment same line? `return True # TODO this should not unconditionally return True`
<gchristensen>
maybe something more clear about the types being incompatible between resources and machines
siiky has quit [Ping timeout: 256 seconds]
<excelsiora>
gchristensen: we could just say the parent class's method returns -> Any since it's just abstract anyways...
<gchristensen>
I think you found a very important question which needs answering properly
<excelsiora>
(it is abstract?)
<gchristensen>
don't add a type for now
<excelsiora>
ok
<tA_>
howdy again, ive been reading over the full disk encryption page on the wiki, and linked articles, and it really hasnt clicked for me, would someone who has an encrypted setup mind running me through how theyve got it set up?
<excelsiora>
how about: `def check(self): # TODO this return type is inconsistent with child class MachineState`
<gchristensen>
sgtm
<excelsiora>
perfect
werner291 has quit [Quit: werner291]
Guest29 has joined #nixos
work_ has quit [Quit: Connection closed for inactivity]
<cole-h>
Does anybody have an overlay prepped for nixfmt 0.4.0? It was marked broken as a part of the recent LTS 15.1 bump and it would be nice to get it back
tilcreator has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @stigtsp opened pull request #81480 → perlPackages.CPAN: fix build, patch url content has changed → https://git.io/Jv2Zj
tilcreator has joined #nixos
torque has quit [Quit: i am dead]
xelxebar has quit [Ping timeout: 240 seconds]
jluttine has quit [Ping timeout: 260 seconds]
drakonis has joined #nixos
jluttine has joined #nixos
xelxebar has joined #nixos
torque has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jv2n3
Guest29 has quit [Ping timeout: 258 seconds]
bhipple has quit [Ping timeout: 258 seconds]
bhipple has joined #nixos
<{^_^}>
[nixpkgs] @obadz merged pull request #77405 → nixos/zerotierone: prevent systemd from changing MAC address → https://git.io/JvvEO
<{^_^}>
[nixpkgs] @rnhmjoj pushed commit from @pacien to release-20.03 « riot-web: mention incompatible config change in release notes »: https://git.io/Jv2nj
<{^_^}>
[nixpkgs] @alyssais merged pull request #81465 → services.mailman: use RemainAfterExit so settings take effect properly → https://git.io/Jv2Yi
<{^_^}>
[nixpkgs] @alyssais pushed commit from @benley to master « services.mailman: RemainAfterExit so settings take effect properly »: https://git.io/Jv2cJ
<joehh1>
hello, how does one enter a negative number into a nix file?
<joehh1>
I'm using nix to generate some json config files and need to put a negative number in
<cole-h>
`(-1)`?
<joehh1>
great thanks cole-h - that seems to have done the trick
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
<cole-h>
:)
<tA_>
anyone here running an encrypted disk?
lovesegfault has joined #nixos
WhatisRT has quit [Ping timeout: 255 seconds]
<{^_^}>
[nixpkgs] @lovesegfault opened pull request #81483 → Revert "Merge pull request #79750 from doronbehar/update-luarocks" → https://git.io/Jv2cs
<cole-h>
lovesegfault++
<{^_^}>
lovesegfault's karma got increased to 15
<lovesegfault>
:)
<jackdk>
tA_: don't ask to ask, just ask
<cole-h>
lovesegfault: Can't wait to not have to `home-manager switch -I nixpkgs=$HOME/workspace/vcs/nixpkgs` lol (have it reverted locally as well)
<cole-h>
Wow man your GH skills suck imagine having that many failing CI builds
<cole-h>
(/s)
veske2 has joined #nixos
veske has joined #nixos
<lovesegfault>
:P
<tA_>
im either wanting to have multiple disks encrypted such that i get asked for a single decrption password at boot, or another method of encryption that is better that im not aware of
<jackdk>
physical or partitions?
<samueldr>
tA_: when a passphrase is valid for the first disk, it is tried for the next disk, and going, in our stage-1
<samueldr>
so if you have e.g. two encrypted drives with the same passphrase, it will unlock both
doublex__ has joined #nixos
<tA_>
and how do i refer to them in hardware.nix, will the installer pick it up properly?
<samueldr>
tA_: when you nixos-generate-config IIRC it should realize they are LUKS-backed partitions (assuming here) and output the right configuration
h0m1 has quit [Ping timeout: 240 seconds]
<samueldr>
it's been a good while since I last installed so I'm not trusting my flawed memory
<clever>
samueldr: it only detects that if your doing something simple like ext4 on luks, and i think ext4 on lvm on luks
<clever>
samueldr: but zfs on lvm on luks confuses it
NiceSun has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @bignaux opened pull request #81484 → sgt-puzzles: fix desktop files and add desktop sgt-puzzles.menu → https://git.io/Jv2c2
Flambe has quit [Remote host closed the connection]
Flambe has joined #nixos
felixfoertsch23 has joined #nixos
butterthebuddha has joined #nixos
<evanjs>
do you really even need LVM if you’re using ZFS, anyway?
butterthebuddha has quit [Client Quit]
<emily>
don't need luks either
<emily>
but i guess swap partitions ands tuff
felixfoertsch has quit [Ping timeout: 256 seconds]
butterthebuddha has joined #nixos
butterthebuddha has quit [Client Quit]
<clever>
evanjs: previously, you needed swap+zfs on lvm on luks, to not have to enter 2 passwords at login
<clever>
evanjs: but nixos now supports reusing the luks pw on multiple devices, so you can do zfs on luks + swap on luks
butterthebuddha has joined #nixos
mjrosenb has joined #nixos
mexisme_ has quit [Ping timeout: 260 seconds]
waleee-cl has quit [Quit: Connection closed for inactivity]
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
butterthebuddha has joined #nixos
orivej has quit [Ping timeout: 260 seconds]
<tA_>
i tried the generated config and on boot it says "waiting for /disk/byuuid/somelonguuid" then fails, have I formatted them wrong? each partition was luksFormatted, then pvcreate, vgcreate, and lvcreate to fill it, then mounted into the correct spot before generating the config
mexisme_ has joined #nixos
butterthebuddha has quit [Client Quit]
fusion809 has joined #nixos
butterthebuddha has joined #nixos
bhipple has joined #nixos
butterthebuddha has quit [Client Quit]
mexisme_ has quit [Ping timeout: 256 seconds]
<clever>
tA_: is the luks properly configured in hardware-configuration.nix ?
gustavderdrache has quit [Quit: Leaving.]
butterthebuddha has joined #nixos
<tA_>
i just went wit what the installer generated, it doesnt say anything about luks in it now that ive checked
<clever>
tA_: if you boot the installer, mount everything back to the right place, and edit the config, you can re-run nixos-install to update the install
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
lovesegfault has quit [Quit: WeeChat 2.7.1]
steshaw has joined #nixos
<evils>
samueldr: with3d is true for kicad and kicad-unstable, (the two failing), packages3d just installs the 3d models from a download (no building) and is quite big (hence the output limit being exceeded, it outputs like 1-2 lines per model (it builts fine for me, but does exceed my 65k line terminal scrollback...)), which is why the previous packager added the hydraPlatforms = []; line (i don't think there's a
<evils>
good reason for hydra to cache this)
<samueldr>
the issue here is that the line I copied is used for the wrapped, that wrapper depends on packages3d, packages3d needs to be built to be a dependency
<samueldr>
I think
<samueldr>
nix will not allow an unbuilt package to be a dependency of another
Rusty1 has quit [Quit: Konversation terminated!]
<evils>
so no getting around building it on hydra? does that mean the hydraPlatforms line does nothing? any way to solve the output limit issue?
<samueldr>
the hydraPlatforms, I'm pretty sure only makes it so no job is created when jobs are prepared for hydra...
<samueldr>
the size limit is 2.2GiB on hydra, I don't know enough about kicad to say how this could be solved in this particular case
<evils>
"Output limit exceeded" isn't about the number of lines to stdout but about the package size?
klntsky has quit [Remote host closed the connection]
<samueldr>
"Log limit exceeded" is when the log is too large
<evils>
huh, last time i checked it wouldn't show me the log, now i'm seeing "unpacking source archive", that would imply it's indeed the package
koleesch__ has quit [Remote host closed the connection]
bhipple has quit [Remote host closed the connection]
<evils>
huh, it's possible it previously failed with log limit exceeded and i didn't spot the difference
klntsky has joined #nixos
<samueldr>
that's an option
<samueldr>
you could also be looking at two distinct problems
<evils>
surely there's a way to set the limit higher? (though i'd prefer not to store a copy of what amounts to a 6GB download in hydra if it's not needed)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
orivej has joined #nixos
<{^_^}>
[nixpkgs] @jonringer merged pull request #81458 → vim-easytags: Fix to work with universal-ctags (which is in nixpkgs) → https://git.io/Jv2qf
<{^_^}>
[nixpkgs] @jonringer pushed commit from @KoviRobi to master « vim-easytags: Fix to work with universal-ctags (which is in nixpkgs) »: https://git.io/Jv24W
<{^_^}>
[nixpkgs] @vcunat pushed to master « nixpkgs manual: fix build after pandoc(?) update »: https://git.io/Jv24i
zupo has joined #nixos
<{^_^}>
[nixpkgs] @bb2020 opened pull request #81490 → nixos/sshd: disable openFirewall by default → https://git.io/Jv241
malSet has quit [Ping timeout: 240 seconds]
coco has joined #nixos
<domi-nix>
hello, I try to get my nixos running as a hyper-v guest-system on a win10. My problem is: I can't get any xserver started, because no screen get's detected. `virtualisation.hyperGuest.enable` is set, but does not help.
<Dandellion>
* https://github.com/NixOS/nixpkgs/pull/76723 has sat there for almost a month now (originally from new years), and it's not really clear to me how it's supposed to move forward
<emily>
dandellion: might be good to ask in #nixos-dev
<emily>
generally things are just stuck because there's a ton of PRs open at any given time, not because of any given problem, and a poke will help prioritize
knupfer has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #81485 → gtk3: fix build on darwin → https://git.io/Jv2C0
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/Jv2ud
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
fendor has joined #nixos
orivej has joined #nixos
coco has quit [Ping timeout: 256 seconds]
v88m has quit [Ping timeout: 256 seconds]
coco has joined #nixos
justanotheruser has quit [Ping timeout: 260 seconds]
kim0 has joined #nixos
<arianvp>
how do I load gnome user extensions? I added gnomeExtensions.dash-to-panel to my systemPackages, but the extension is not showing up in gnome-tweaks
<eon`>
arianvp: you can also reload gnome-shell with Alt+F2 and typing 'r'
<eon`>
used to work well back in the days
<stevenroose>
samueldr, simpson: So I tried the RPi1 armv6 installer and it seemed to work good. I only installed a minimal of new packages (1 IIRC), and tried the rebuild. The suggested binary mirror didn't work, so I removed it from the command. Then I left it overnight building.
<domi-nix>
may I reask my question: I try to get my nixos running as a hyper-v guest-system on a win10. My problem is: I can't get any xserver started, because no screen get's detected. `virtualisation.hypervGuest.enable` is set, but does not help, though :/
<stevenroose>
Now I'm getting this error that the bootstrap-tools resulted in an unexpected hash, making the build fail.
<raboof>
I'd like to access a keybase team folder. I've set 'services.keybase.enable' and 'servicds.kbfs.enable' to 'true' and logged into keybase from the commandline.
<raboof>
'keybase status' shows me as logged in, but I don't see a /keybase and no keybase services in 'systemctl --type=service'. What could I be missing?
<raboof>
oh, it's in 'systsemctl', it's just not a systemd service and placed under /home/me/keybase. sorry for the noise :)
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @Ma27 opened pull request #81518 → dmenu-wayland: init at 2020-02-28 → https://git.io/Jv2we
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
Synthetica has joined #nixos
oida has quit [Remote host closed the connection]
oida has joined #nixos
<Synthetica>
Is there a builder to use mkWrapper on a "Nix level", such that I could do something like `mkWrapper "${bash}/bin/bash" {PATH = [ hello ]}` and get a wrapped variant of bash with `hello` available?
<fendor>
can I create a shell.nix to setup a dev environment, comparable to docker containers? Assuming, can I shutdown the dev environment again automatically? e.g. open postgres database via shell.nix and shut it down again after leaving the shell?
<hyperfekt>
fendor: look into the trap command for teardown i guess
chiefgoat has quit [Remote host closed the connection]
<fendor>
maybe starting an interactive shell does not make much sense in the nix context? I would also be fine with a command like `nix run` or something like that?
<fendor>
simpson, that is really cool and yeah, that is the intention, too.
<__monty__>
simpson: That doesn't take care of teardown either, right?
WhatisRT has joined #nixos
<simpson>
__monty__: I guess not? It's not my recipe.
Acilim has joined #nixos
Acilim has joined #nixos
Acilim has quit [Changing host]
<fendor>
I think, it would be good enough if entering the shell gives you the possibility to run all the dev stuff in a dedicated terminal. Then teardown can be done by the commands itself
<fendor>
for every dev in my team, the environment would be still unified, they can see all the logs, etc...
<__monty__>
You expect the setup postgres instance to be shared?
<fendor>
shared between what? Within the same project, yes
rail_ is now known as rail
waleee-cl has quit [Quit: Connection closed for inactivity]
<ottidmes>
The manual says: "This kind of string literal intelligently strips indentation from the start of each line." And I know this was always the case. But just now I made a bit of config and both the script and service configuration have the indentation present
dm9 has quit [Quit: WeeChat 2.7.1]
<__monty__>
ottidmes: Are you mixing tabs and spaces maybe?
sigmundv__ has joined #nixos
jonten has joined #nixos
<ottidmes>
__monty__: I checked to be sure, made sure to convert them in the editor and checked manually afterwards, yet it still behaves incorrectly
<Taneb>
Is it because the closing '' is less indented?
domi-nix has quit [Remote host closed the connection]
<Taneb>
No, that wouldn't do this
veske2 has quit [Quit: This computer has gone to sleep]
veske has quit [Quit: This computer has gone to sleep]
<ottidmes>
Taneb: Tried it anyway, but no luck
<disasm>
CRTified: I think you can set `StrictModes off` in the ssh config, although that has other potential security concerns.
<__monty__>
No, only way I can kinda get this result is by having a line with less indentation preceding. But I'm on nix 2.2.2, maybe this is a regression?
<ottidmes>
If I do "/home/${name}/.vscode-server/bin/*/node IN_CLOSE_WRITE,loopable=true ${cronScript} $@/$#\n" I get exactly what I want, which I expected from the ''..'' string as well
<ottidmes>
__monty__: Could be, I am on 2 day old nixos-unstable
<__monty__>
ottidmes: And you're talking about indentation of the line, right? Not Trailing newlines?
<gchristensen>
excelsiora: are you doshitan?
erasmas has joined #nixos
<CRTified>
disasm: did my irc transport mix something up, or did you answer my old question regarding goklp?
<ottidmes>
__monty__: I always remove trailing newlines for nix files on save
<WhatisRT>
hi! is there some way to use a fetcher for a local git repository? I have some small changes I want to make to the original repository but I don't want to change the corresponding nix expression
<ottidmes>
__monty__: I will test the file on my other machine, if it succeeds there it should be Nix. Once I am home I will be downgrading to 20.03 and see if that helps
malSet has quit [Ping timeout: 256 seconds]
justanotheruser has quit [Ping timeout: 240 seconds]
<adisbladis>
whatisRT: You don't need to use a fetcher for that.
<{^_^}>
[nix] @edolstra pushed 2 commits to eval-optimisations: https://git.io/Jv2X1
<__monty__>
ottidmes: I don't mean in the nix file, I mean in the file produced with that string. Having the closing `''` on a new line appends a newline to the string.
malSet has joined #nixos
<ottidmes>
__monty__: In the produced file I see 2 newlines, while I expected 1 (the one produced by having '' on a newline)
karantan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ottidmes>
__monty__: Still seeing the issue on my other machine, which runs Nix 2.3 (my unstable machine runs 2.3.3)
<WhatisRT>
adisbladis: ah, thanks!
karantan has joined #nixos
<mikky>
hi, what's the current way to set io scheduler? (i.e. the former `elevator` kernel parameter)?
tjg1 has quit [Quit: DON'T KEEP CALM AND FUCK THIS. FUCK THAT. FUCK YOU. FUCK OFF.]
tjg1 has joined #nixos
<{^_^}>
[nixpkgs] @Ma27 pushed to master « nixos/release-notes: mention fix for predictable network-interfaces in initrd »: https://git.io/Jv218
karantan has joined #nixos
<ottidmes>
Build Nix 2.2 on my machine and get consistent behavior, so must be something on my end, nix eval "$(echo -e "(''\n test\n '')")", works as expected in both Nix 2.2.2 and 2.3.3, not sure what goes wrong in my file to make it misbehave, guess I will continue debugging it
<__monty__>
ottidmes: Oh, maybe you want echo -n? That adds a newline.
karantan has quit [Client Quit]
<ottidmes>
Found it! Stupid Windows! I was editing the file remotely on my Windows machine, and not being used to it, I have yet to set the line endings to be \n always...
<ottidmes>
Changing the editor to use LF rather than CRLF fixed it for me
<__monty__>
Oh.
<ottidmes>
Still kind of unexpected to me, I would have expected it to be able to deal with \r\n newlines
<fgaz>
I just created an imperative nixos container but I can't connect to it from the outside... do I need to enable networking somehow? I can't find info about it
<fgaz>
By connect I mean tcp
<hexa->
fgaz: did you start the container after creating it?
<{^_^}>
[nixpkgs] @teto pushed commit from @davidtwco to master « starship: 0.36.1 -> 0.37.0 »: https://git.io/Jv2HA
tokudan has joined #nixos
Mark_ has joined #nixos
<jared-w>
fresheyeball: I've done this before. It was a pain in the arse and required using pkgs.fetchgit, not builtins.fetchGit
tokudan has quit [Remote host closed the connection]
<jared-w>
pkgs.fetchgit { url = ...; leaveDotGit = true; }; and then in a mkDerivation you need a preBuild step where you call `git reset` or the .git directory will be there but the git index will be removed
veske2 has joined #nixos
veske has joined #nixos
<fresheyeball>
jared-w: I think I got it working with builtins.fetchGit
<fresheyeball>
I just set the url to be ssh://git@github ...
<evanjs>
fresheyeball: hopefully you don't need submodules, though :D
<fresheyeball>
seriously
<jared-w>
yeah fetchgit supports submodules, builtins.fetchGit doesn't. It's a royal pain in the arse because git is inherently super non-reproducible in all of its main usecases.
tokudan has joined #nixos
<evanjs>
I think there's a PR open but can't --- OH it's in nix, not nixpkgs lol
* jared-w
"same command => different output every time." "#yoloWeModifiedStateJustByCheckingForState". Etc...
LysergicDreams has quit [Ping timeout: 240 seconds]
veske2 has quit [Client Quit]
veske has quit [Client Quit]
<jared-w>
looks like this is gonna be yet another thing ye magical olde flakes fixes :p
<jared-w>
The year of the NixOS desktop is nearly upon us
<evanjs>
Idk, all I know is we can move a ton of stuff to proper derivations (re, not src = ./. ...) once something like this is merged
werner291 has quit [Ping timeout: 240 seconds]
<evanjs>
Oh, dang. The PR was updated 6 hours ago!?
<jared-w>
we as in nixpkgs or we as in internal code? My company has a decent amount of improper nix flying around due to required shennanigans with git so I'm also looking forward to it :p
<evanjs>
"I've incorporated all fixes. I just have to investigate the test failures."
<jared-w>
evanjs: "ready to merge" is code for "pls merge sometime before 2025 maybe"
<evanjs>
Bc overlays and etc
<evanjs>
jared-w: omg stop it ;_;
<jared-w>
evanjs: aw yiss, public code is best code
<jared-w>
evanjs: truth hurtz tho but where's the lie? :p
* jared-w
in search of fast way to check if nixpkgs channel has MR that I want
LysergicDreams has joined #nixos
<jared-w>
example: https://github.com/NixOS/nixpkgs/pull/81483. Pretty likely that this isn't in nixos-unstable yet given that it's 11 hrs old and nixos-unstable is older than that, but aside from doing some random screwing around in github's UI, I don't know of a way to check that
<evanjs>
All I'm doing is trying to make a minimal image for something we'd been creating by hand for a while, and figured a nix expression would be a good way to do so :D
<evanjs>
19M is my final image size (XZ compressed kernel) atm
phreedom has joined #nixos
<Shouou>
Do I need to do anything special to add a thumbnailer to GNOME/Nautilus other than make sure $out/share/thumbnailers/foo.thumbnailer exists, is formatted correctly, and said package is in my systemPackages?
<__monty__>
fresheyeball: You'll have to include a less indented line and then strip it away if you have to. What's your usecase?
xelxebar has quit [Ping timeout: 240 seconds]
<symphorien>
> " " + '' space''
<{^_^}>
" space"
zupo has joined #nixos
<fresheyeball>
__monty__: I want to dynamically create a git patch
<__monty__>
fresheyeball: Why do you need the indentation though? Maybe you can include an unindented comment? Usually patches have headers anyway, right?
<__monty__>
symphorien: That wouldn't work for a ''multiline'' string, which is probably the reason for using '' ''.
<fresheyeball>
__monty__: I am not totally sure yet
<fresheyeball>
but I generated the initial diff with git
<fresheyeball>
and it includes a line like this
<fresheyeball>
{
<fresheyeball>
just a curly boi all alone, but indented
<fresheyeball>
the indented on applies fine with git manually
<fresheyeball>
but when I have it was nix, it says that line is an error
teto has quit [Quit: WeeChat 2.7.1]
<fresheyeball>
I tried something like this
<fresheyeball>
''
<fresheyeball>
${ "
<fresheyeball>
ugh
<fresheyeball>
I put this in the multipline
<fresheyeball>
${" {"}
<__monty__>
Why do you want a patch as a multiline string? The git patches I've seen in nixpkgs are just in patch files and I think the patchPhase finds most of them automatically.
NoctisLa1 has quit [Read error: Connection reset by peer]
ixxie has quit [Ping timeout: 256 seconds]
mexisme_ has quit [Ping timeout: 258 seconds]
energizer` has joined #nixos
<HedgeMage>
How do I find out which of the things I'm installing is trying to pull in an outdated version of openssl? I'm not asking for it explicitly anywhere.
<HedgeMage>
(I'm doing a system rebuild, and got a warning)
<fresheyeball>
adisbladis++
<{^_^}>
adisbladis's karma got increased to 32
<fresheyeball>
that worked amazing
karantan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
LysergicDreams has quit [Ping timeout: 240 seconds]
<energizer`>
does cachix let me use other people's cached binaries, or only mine?
LysergicDreams has joined #nixos
<adisbladis>
fresheyeball: Nice :)
matthiaskrgr has quit [Read error: Connection reset by peer]
matthiaskrgr has joined #nixos
<cole-h>
energizer`: If you `cachix use` their repo, then yes
mexisme_ has joined #nixos
<Ericson2314>
any fish user noticed that command-not-found seems recently to be called all the time?
<Ericson2314>
whenver I control-C
<Ericson2314>
which is infuriating
<energizer`>
i don't really see why it'd be necessary to tell cachix where to look, i'm asking for "this thing", not "this thing, from this place"
<samueldr>
do you really trust random binaries someone uploaded saying it's "the thing"?
<fresheyeball>
adisbladis: if I run jq with one file
<fresheyeball>
jq "stuff" < foo.json > foo.json
<fresheyeball>
I get an empty foo.json
<adisbladis>
fresheyeball: That wont work, and that has nothing to do with jq
<fresheyeball>
do I need to create a tmp file?
<adisbladis>
Yeah
<samueldr>
`cachix use` probably (unverified) also adds the signing key for *that* cache as trusted
<adisbladis>
Otherwise you're overwriting the file while you are reading it :)
<cole-h>
samueldr: It does
<energizer`>
samueldr: i have a lockfile. nix can't use the hash from the lockfile?
<adisbladis>
energizer`: Sure, that could work for sources.
<adisbladis>
But not for build outputs, they are not necessarily reproducible
<energizer`>
given a package version with hash, how do i find out if nixpkgs has ever built it and tell nix that i want to use that version of nixpkgs for that package?
nerdmaxx has joined #nixos
linarcx has quit [Ping timeout: 258 seconds]
<Ericson2314>
huh maybe I was in too many nested nix shells
<energizer`>
maybe my questions aren't making sense because i'm doing something wrong. i'm using one of the lockfile to nix tools, but that means i have to build stuff manually. these are popular packages, with versions in nixpkgs, so i'm surprised if they've never been built by reputable sources before
<emily>
that only has a chance of being true if they use the exact same build system in nixpkgs
<emily>
e.g. depending on the language there's often multiple different builders with different properties
<emily>
in general hoping that a random derivation you have might match something hydra built just because the underlying source archive is the same is optimistic
<emily>
every single byte of every little shell script and every single dependency pulled in has to match
<adisbladis>
energizer`: Any foo2nix tool is _extremely_ likely to end up having different hashes than something manually packaged in nixpkgs.
<emily>
the "derivation -> output" mapping isn't cached at the abstract "build some module" level, but the "here's all the dependencies and files and a pile of shell, go execute it" level
hmpffff has joined #nixos
<jared-w>
energizer`: cachix behaves like a binary cache, so it's essentially a CLI tool that automates setting binary cache urls and adding public keys to the appropriate nix config settings
<jared-w>
so `cachix use` is a code-gen tool that generates the appropriate nix files.
<jared-w>
indeed `cachix use foo` is just gonna create a `/etc/cachix/foo.nix` file with the contents `{ nix.binaryCaches = [ "https://foo.cachix.org" ]; nix.binaryCachePublicKeys = [ "thekey" ]; }; }`
<jared-w>
(well, sudo cachix use foo will create it in /etc/cachix/foo.nix. Non sudo will create it somewhere in ~/.config iirc)
<energizer`>
so foo2nix will build a package a certain way, and there could be a cache of all the foo packages, that everybody could use?
chloekek has quit [Ping timeout: 260 seconds]
dsx has joined #nixos
aveltras has quit [Quit: Connection closed for inactivity]
excelsiora has joined #nixos
<energizer`>
cachix says "100GB of storage (once reached oldest entries will be garbage collected)". what does it mean by "oldest"? LRU?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
opthomasprime has left #nixos [#nixos]
<cole-h>
Probably means oldest as in the date it was entered into its cache
kristian_ has joined #nixos
<cole-h>
e.g. you add a package `foo` today, and 100+GB from now you add a package `baz`, `foo` will be garbage collected
knupfer has quit [Remote host closed the connection]
<excelsiora>
gchristensen sorry I'm not that person. I'm dealing with some personal issues that may prevent me from giving you a PR. so I should probably not hold up the work on that. I'll keep my eye on it and comment where apropos. My apologies for being out of it.
Flambe has quit [Ping timeout: 240 seconds]
esclear has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @flokli merged pull request #80096 → nixosTests.rsyslogd: Port to python → https://git.io/JvlqN
energizer` has quit [Remote host closed the connection]
gustavderdrache has quit [Quit: Leaving.]
justanotheruser has quit [Ping timeout: 256 seconds]
<ben>
hi, i'm on ubuntu, using nix-env to install rxvt-unicode and vim_configurable. I don't get colors in vim. TERM=rxvt-unicode=256color; if I manually set TERM=xterm I do get colors in vim. Is there a correct way to fix this?
kristian_ has quit [Quit: Konversation terminated!]
<ben>
ln -s ~/.nix-profile/share/terminfo ~/.terminfo seems like a workaround that may cause future problems
<fresheyeball>
how can I add an ssh key to the nix builder?
<fresheyeball>
I got the private repo thing working fine on my local, but I need it to work when ci evals this nix code
realrokka has quit [Ping timeout: 255 seconds]
<fresheyeball>
which is done with gitlab-runner service on nixos
lsix has quit [Quit: WeeChat 2.6]
bukkitgerman8 has quit [Quit: Ping timeout (120 seconds)]
bukkitgerman8 has joined #nixos
Rusty1 has quit [Quit: No Ping reply in 180 seconds.]
<fresheyeball>
actually, is there even a way to do that with nix code?
<fresheyeball>
can I give a user an ssh key? It seems like storing the private key might be a bad idea
<fresheyeball>
and I should do this imperatively
Rusty1 has joined #nixos
<{^_^}>
[nixpkgs] @jtojnar pushed to fetchpatch-base64 « fetchpatch: Allow using base64-encoded hash »: https://git.io/Jv2Ft
bukkitgerman8 has quit [Client Quit]
butterthebuddha has joined #nixos
nuncanada has joined #nixos
<{^_^}>
[nixpkgs] @jtojnar opened pull request #81564 → fetchpatch: Allow using base64-encoded hash → https://git.io/Jv2Fm
<{^_^}>
[nixpkgs] @jtojnar pushed to fetchpatch-base64 « fetchpatch: Allow using base64-encoded hash »: https://git.io/Jv2FY
rardiol has joined #nixos
maddo has quit [Quit: See ya]
esclear has joined #nixos
bukkitgerman8 has joined #nixos
philr has quit [Ping timeout: 268 seconds]
realrokka has joined #nixos
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bukkitgerman8 has quit [Quit: Ping timeout (120 seconds)]
butterthebuddha has joined #nixos
kleisli has joined #nixos
hmpffff has quit [Quit: nchrrrr…]
bukkitgerman8 has joined #nixos
realrokka has quit [Ping timeout: 256 seconds]
realrokka has joined #nixos
hmpffff has joined #nixos
butterthebuddha has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<lovesegfault>
Mic92: Sweet! Hi :) How did you build the git worktree functionality in nixpkgs-review where once you exit the shell the worktree gets nuked automagically?
<worldofpeace>
lovesegfault: I wanted Jan Tojnar review. Lately, well... we'll just say I'm prone to mistakes due to load
<worldofpeace>
lovesegfault: did you test it with lollypop?
<lovesegfault>
Mic92: Oh! Nice!
<lovesegfault>
worldofpeace: Nope, but I can, gimme one moment
<worldofpeace>
It's a mass rebuild, but if you copy the setup-hook and add another in all-packages.nix, wrapGAppsHook2, and use it in the package it can be avoided
malSet has joined #nixos
<jtojnar>
I am having a déjà vu regarding the worktree question