<{^_^}>
[nixpkgs] @thoughtpolice merged pull request #51409 → nixos/cockroachdb: supply defaultText for the package option → https://git.io/fpKV5
<{^_^}>
[nixpkgs] @thoughtpolice pushed commit from @pbogdan to master « nixos/cockroachdb: supply defaultText for the package option »: https://git.io/fpKDp
tejing has joined #nixos
jrolfs has quit [Ping timeout: 250 seconds]
Supersonic has quit [Ping timeout: 268 seconds]
Supersonic has joined #nixos
epicmetal has joined #nixos
acarrico has quit [Ping timeout: 250 seconds]
j_r0m has quit [Quit: Page closed]
<tejing>
I just tried switching channels, but when I try to 'nixos-rebuild --upgrade switch', it seems to hang after the downloads (specifically after 'building the system configuration...') I also get some warnings about about an unknown setting 'signed-binary-caches'. Am I doing somethign wrong or is something broken?
cyounkins has joined #nixos
<clever>
tejing: thats when its parsing all of the nix expressions
<clever>
tejing: just keep waiting
<tejing>
clever: but no cpu time is being used
<clever>
then its probably done that step, and checking the binary caches, which involves 100's of small http requests
fusion809 has quit [Remote host closed the connection]
<tejing>
ok, that took quite a while but it's finally doing something again. guess I jumped the gun a little bit.
drp has joined #nixos
<jackdk>
is there a nix 2.0-style equivalent to nixos-rebuild? `nix build` will show it checking caches and that might be nice for nixos too
lassulus has quit [Quit: WeeChat 2.2]
lassulus has joined #nixos
<clever>
jackdk: if you dont mind it being a 3 step process, do `nixos-rebuild dry-run`, find its .drv file, `nix build /nix/store/foo.drv` and when its done, go back to normal `nixos-rebuild`
<jackdk>
probably want a `--no-link` on that second command?
<{^_^}>
[nixpkgs] @veprbl pushed commit from @scalavision to master « mill-bin: init at 0.3.5 »: https://git.io/fpK9c
IRCsome has quit [Ping timeout: 246 seconds]
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tdeo>
i'm trying to compile a package with https://pastebin.com/jNBUZp6g but i get this error: src/password_generator_widget.cpp:47:59: error: 'log2' was not declared in this scopesrc/password_generator_widget.cpp:47:59: error: 'log2' was not declared in this scopesrc/password_generator_widget.cpp:47:59: error: 'log2' was not declared in this scope on compile
<tdeo>
oh
<tdeo>
that pasted 3 times
<tdeo>
i didn't mean to
<tdeo>
but i think it might have something to do with the package and not the code
cyounkins has quit [Remote host closed the connection]
mmlb has quit [Read error: Connection reset by peer]
mmlb has joined #nixos
vk3wtf has quit [Ping timeout: 268 seconds]
endformationage has quit [Quit: WeeChat 2.3]
gerschtli has quit [Quit: WeeChat 2.2]
gerschtli has joined #nixos
hellrazo1 has quit [Ping timeout: 250 seconds]
<j_r0m>
teto: thank you teto . at this point i've only ever used ns3 using (./waf --run NAMEOFSCRIPT) in non nixos environments . so it's a bit weird changing habits and workflow . i'm doing some things in postInstall at the end of the ns-3 derivation ... we'll see if that works
rprije has quit [Ping timeout: 240 seconds]
<teto>
j_r0m: I've had a patch for ns3.29 for quite some time in ase you are interested but I was waiting for the waf hook to land nixos-unstable before pushing my update
palo1 has joined #nixos
<j_r0m>
teto: rad . i'm still using 3.28 ; but i will certainly want to get with the program soon i reckon :)
<teto>
j_r0m: just curious but do you intend/have you ever used DCE ?
<teto>
(Direct Code Execution, a ns3 extension)
palo has quit [Ping timeout: 272 seconds]
orivej has quit [Ping timeout: 246 seconds]
monokrome has joined #nixos
<monokrome>
yooo
<monokrome>
Weird thing, but I was in here having issues w/ a laptop that has a GTX 1070 Q-Max and I'm having the same issues on a new PC that just arrived
<monokrome>
O_o
orivej has joined #nixos
<monokrome>
(where nvidia driver doesn't work and I am forced to use the intel driver even though I have GPUs in the machine)
<monokrome>
yeah that's wrong
<monokrome>
oops wrong widnow
jackdk has quit [Ping timeout: 244 seconds]
<monokrome>
(other GPU is a dedicated RTX 2080 Ti)
<oldandwise>
I did `# nixos-rebuild switch --upgrade` after `nix-channel --add https://nixos.org/channels/nixos-18.09 nixos` and my x64 machine has upgraded (/etc/os-release,login gui). But I saw that /etc/nixos/configuration.nix system.stateVersion = "18.03";
<{^_^}>
[nixpkgs] @vcunat pushed 3 commits to release-18.09: https://git.io/fp6ek
hjulle has joined #nixos
<v0|d>
exarkun1: move glib to natbuilinpt
<exarkun1>
that was my thought, indeed
<exarkun1>
I wasn't able to discern an effect
<hjulle>
I can't upgrade nix(os) because of the bug https://github.com/NixOS/nix/issues/2393. Is it any way to force it to install despite reporting an incorrect hash?
<{^_^}>
nix#2393 (by lopsided98, 13 weeks ago, closed): Hash mismatch with builtins.fetchUrl and .xz file
<sphalerite>
exarkun1: it looks like it's using the wrong C compiler, glib should be in buildInputs not nativeBuildInputs
<sphalerite>
exarkun1: not sure how sshfs's build works, but if you can get it to use $CC instead of cc it ought to work
<exarkun1>
it seems to use ninja
<exarkun1>
which seems to have its own ideas about what cross compilation should be
<exarkun1>
never actually used ninja myself, just browsed the docs briefly
<exarkun1>
sounds like maybe the answer is "it's work"
<sphalerite>
ninja doesn't care about that sort of stuff generally
<sphalerite>
AFAIU
<sphalerite>
ninja just executes the build, but the ninja files are generated by something else like meson or cmake
<exarkun1>
okay, maybe it's meson then
<exarkun1>
(meson is definitely involved)
Dedalo has joined #nixos
Dedalo has quit [Client Quit]
<exarkun1>
turns out it is easier to just do a native compile on rpi :/ I guess sshfs is tiny.
rprije has joined #nixos
jrolfs has joined #nixos
jrolfs has quit [Ping timeout: 246 seconds]
<tobiasBora>
anothertest: I know you are not connect anymore, but if by any chance you are interested by the answer to your question about iso mounting, I think to remember I managed to do something like that with the help of clever. If my memory is good, I ended up building a new iso with everything included directly in the loaded kernel/initramfs (it's only a problem when you have not enough RAM to load it in once). If
<tobiasBora>
you are still interested, I still have a memo at home with the .nix needed for that. And I was not able to find a solution with an external iso not fully mounted into ram, but I'm thinking that it may be possible to use the tool used by ubuntu with nix, as I think to remember that I saw some code linked with that in the git repo of nixpkgs...
<fendor>
yesterday, I tried to install NixOs on a Raspberry Pi 3, with the ARM Guide: https://nixos.wiki/wiki/NixOS_on_ARM#Installation . However, after following the instructions, I think the raspberry pi started to build the latest linux kernel from source. Naturally, after the night, it still hasnt finished compiling. Question: is that normal?
<ckauhaus>
shouldn't
<ckauhaus>
I did that quite a while ago and had zero local compiles
<sphalerite>
fendor: if you're using the armv7 image it is normal
<sphalerite>
I think
<fendor>
sphalerite, I used the ARM64 image
<fendor>
ckauhaus, damn it
<sphalerite>
right, then it is weird
<fendor>
welp, maybe I fucked something up. Has someone experience on how long the linux kernel takes to compile on a raspberry pi?
<fendor>
also, if the raspberry pi is not done in the evening, should I rather abort it and try to find the issue?
<fendor>
tilpner, yes and no, I didnt dare to change anything in the configuration.nix, moreover, since it should become a backup server using samba, I did not want to expose it to the internet
Ariakenom has quit [Ping timeout: 250 seconds]
vk3wtf has joined #nixos
<tilpner>
That's what you can use tinc (or similar) for, so it's not public but still accessible
drp has joined #nixos
simukis has joined #nixos
<fendor>
tilpner, didnt know about tinc, I will take a look at it and try to use it :)
<tilpner>
(tinc only if you have a public server, but I hear toxvpn works well if you don't have one)
sigmundv_ has joined #nixos
<fendor>
tilpner, I dont have a public server, the backup service is fine if it is only executed when you are in the local network
jrolfs has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @dotlambda pushed 2 commits to python-unstable: https://git.io/fp6LH
<{^_^}>
[nixpkgs] @FRidh pushed 496 commits to staging-next: https://git.io/fp6Ld
<{^_^}>
[nixpkgs] @vcunat pushed commit from @delroth to release-18.09 « nixos.tests.keymap: use new name of the colemak/en-latin9 keymap »: https://git.io/fp6Oj
wink has joined #nixos
<wink>
hi, is github.com/bendlas lurking here by chance? :)
<eyjhb>
How can I get nixos my ssh client to try all keys in '.ssh'? Currently it only uses 'id_rsa'
<hyper_ch>
what keys?
<wink>
I think that's an ssh-agent/keyring issue and should be distro-independent
ng0_ has quit [Quit: Alexa, when is the end of world?]
<eyjhb>
hyper_ch: general ssh keys for connecting. I have to specify 'ssh -i .ssh/id_rsa_tmp user@ip'...
<eyjhb>
wink: Running NixOS, not just the package manager. Any 'easy way' of doing it?
ng0 has joined #nixos
<wink>
eyjhb: I mean: I know of no "configure" option. ssh itself will only use the default names and agent/keyrign might scan ~/.ssh - but I could be totally wrong
dcol has joined #nixos
<wink>
but don't ask me about nix specifice, so I'd rather shut up ;)
<tilpner>
Try programs.ssh.startAgent
<dcol>
I'm trying to cross compile for an ARM device on NixOS. It's going pretty well using crossSystem and so on, however I realized that my ARM device has glibc 2.23 but my toolchain gets compiled with 2.27. I've been trying to figure out how to write a nixexpression to use 2.23 glibc when compiling, but it's not going too well. Anyone who could point me in the right direction?
<tilpner>
I suppose you could have an overlay of the older glibc
<tilpner>
But that would rebuild everything :/
<dcol>
hm I guess. but if I have a --pure shell that shouldn't be too bad?
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #51317 → audit-tmpdir hook: use abspath matching on RPATH entries → https://git.io/fpoK0
<{^_^}>
[nixpkgs] @Ericson2314 pushed 3 commits to staging: https://git.io/fp6cL
<symphorien>
dcol: is it possible to ship the nix glibc with your program ?
<{^_^}>
[nixpkgs] @Ericson2314 pushed 3 commits to staging: https://git.io/fp6cR
<dcol>
symphorien: possibly, but I'm trying to avoid that if possible. the device isn't that well documented, and I'm just hacking along here. I'm fine with rebuilding my cross.c environment should that be needed
<eyjhb>
sphalerite: but I do not want to do that, I am seeking ubuntu like behaviour, where it will try "all" files in my `.ssh`
Anton-Latukha has joined #nixos
<{^_^}>
[nixpkgs] @NeQuissimus pushed to master « slack: Add dark-mode version »: https://git.io/fp6CT
<tilpner>
Try programs.ssh.startAgent
jrolfs has quit [Ping timeout: 268 seconds]
<schopp0r>
hi. time to try again to compile nix from source with a different nix-store and state directory. can anyone suggest a good way? I was trying it manually from source, and I also tried to compile it on nixos and copy it to another machine.
<schopp0r>
the machine is a debian-i686 and I have no root access and no access to /nix
dcol has left #nixos ["ERC (IRC client for Emacs 26.1)"]
dcol has joined #nixos
<schopp0r>
proot also doesn't work.
<schopp0r>
but I have access to a virtual machine with nixos installed, on which I, in theory, could compile it.
<symphorien>
how far did you go last time ?
<eyjhb>
tilpner: will do and how that resolves it! - How well does switch work in that regard?
<schopp0r>
symphorien: I could manage (but don't know how exactly ...) to run nix on this machine, in /tmp/..., but when trying to compile nix from inside this /tmp/... with a store-directory in my home-dir, it failed when testing some aws-library. and then noone could tell me how to disable tests, and I could not figure it out...
<schopp0r>
symphorien: and it also wasn't really "clean". I just copied all environment-variables to this machine to run it.
<tilpner>
schopp0r - Does the target device allow unprivileged user namespaces?
<schopp0r>
symphorien: I think I did. hm. ok, then something else failed, I cannot remember :\ (should probably have written it down)
<symphorien>
the right way to remove aws-sdk-cpp would be to clone nixpkgs, modify the nix expression inside (should be easy) and then build with -I nixpkgs=/path/to/clone
<schopp0r>
symphorien: ah, but that brings me to the question: how do I "cleanly" copy the stuff from /tmp/blargh3 to my other machine?
<symphorien>
you don't need it
<symphorien>
nix can perfectly start from an empty store
<schopp0r>
symphorien: but how do I transfer the compiled nix to the other machine?
<symphorien>
tar the store where it lives
<schopp0r>
symphorien: I don't understand
<symphorien>
please give more details
<eyjhb>
This might be a stupid question, but isn't it good practice to start bash script with `#!/bin/bash`? But when doing that on nixos, it will fail to run, because bash does not reside in /bin/bash
<schopp0r>
symphorien: so what I would try is use this nix-expression on a nixos-VM.
<symphorien>
start with #!/usr/bin/env bash
<schopp0r>
eyjhb: you start it with #! /usr/bin/env bash
<schopp0r>
symphorien: and then somehow transfer the built nix package manager to the host machine
<schopp0r>
symphorien: and that is the problem: how do I transfer it.
<eyjhb>
I guess I have just been lucky all my life till now then ;) Thanks
<symphorien>
eyjhb: only /bin/sh is fine, because it is required by POSIX
arjen-jonathan has joined #nixos
<schopp0r>
symphorien: because when I build the nix-expression you linked, the then-built nix will still be in /nix/store somewhere
<symphorien>
yes
<schopp0r>
symphorien: so I would have to use *this* nix again to build a nix that is in /tmp/blargh
<symphorien>
because the nixos nix defaults to /nix
<symphorien>
BUT this is only a default
<symphorien>
with some environment variable, you can override on any nix the location of the store
<Graf_Blutwurst>
Hi I'm trying to get the clisp written nEXT browser to work on nixos and writing a package. but I haven't figured out how to handle sbcl in nix correctly. in the normal build there's a makefile which pulls quicklisp from some url and proceeds. when executing "make next-gtk" this obvs. explodes because the ${src} is read only
<nikivi>
the reason I ask is because this error seems to be due to me installing Go with Nix
<nikivi>
and Nix is immutable and now it can't install packages or something
<{^_^}>
[nixpkgs] @avnik opened pull request #51454 → mythes: fix link error on missed `tinfo` library (ncurses) → https://git.io/fp6BR
<ottidmes>
nikivi: the only workaround I got it to copy over/link stuff from the nix store to a writable location, that is what I used to make vsliveshare work in NixOS
<simpson>
Graf_Blutwurst: You're working in pkgs/development/lisp-modules/quicklisp-to-nix.nix, right? I don't really know this area of nixpkgs. It sounds like, if the Makefile has the bogus assumption that it can write to the source directory, you're going to have to either patch the Makefile or copy the source to a temporary directory first.
ng0 has quit [Quit: Alexa, when is the end of world?]
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
<__red__>
Greetings - I need to set up a wiki for a project. Any recommendations for the more minimal-fuss wiki implementtaion to run on a nixos webserver?
<Graf_Blutwurst>
@simpson no I'm not working on a existing package. rather I'm trying to get something to run that requires quicklisp etc
alex`` has joined #nixos
<Graf_Blutwurst>
simpson: ahh so that's user tagging here. I always forget. sorry
<{^_^}>
atlas-engineer/next#146 (by GrafBlutwurst, 3 days ago, open): Getting next to run under NixOS
m0rphism has quit [Ping timeout: 240 seconds]
fendor has quit [Ping timeout: 246 seconds]
samrose has joined #nixos
<symphorien>
Graf_Blutwurst: maybe you need wrapGappsHook
<Graf_Blutwurst>
symphorien: I'll need that for sure. But first I have to get to the point that GTK stuff fails ^^' but thanks for the pointer
<Graf_Blutwurst>
First I have to figure out how to move all that stuff to a temp dir. the make script curls a quick lisp file and right now that tries to write to ${src}
<symphorien>
using the network during a build will be prevented by the sandbox
<{^_^}>
[nixpkgs] @matthewbauer merged pull request #50936 → dhcpcd: Use runtimeShell in dhcpcd-run-hooks → https://git.io/fpRsU
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/fp6rb
<symphorien>
you must pull the file with feturl or something like that and then patch the makefile to use this instead of downloading
<{^_^}>
[nixpkgs] @grahamc merged pull request #50452 → Rust build-support: fixing a compilation error in some crates → https://git.io/fpZ2k
drp has quit [Ping timeout: 250 seconds]
rtjure has quit [Ping timeout: 250 seconds]
Twey has quit [Ping timeout: 250 seconds]
ottidmes has quit [Ping timeout: 246 seconds]
ottidmes has joined #nixos
disasm has quit [Ping timeout: 250 seconds]
jrolfs has joined #nixos
disasm has joined #nixos
philippD has quit [Ping timeout: 250 seconds]
travelion has quit [Ping timeout: 250 seconds]
alex`` has quit [Ping timeout: 244 seconds]
acarrico has joined #nixos
Twey has joined #nixos
travelion has joined #nixos
mmlb has quit [Read error: Connection reset by peer]
halfbit has joined #nixos
mmlb has joined #nixos
<catern>
hey #nixos, is there a way to override part of a recursive attribute set? (like override version in rec { name = "foo-{version}"; version = "1.0"; })
<gchristensen>
no
<gchristensen>
I mean, yes, you can override `version`, but it won't change `name`
<catern>
ok just checking
<catern>
the helpers that build all-packages and python-packages etc are essentially for this purpose, right? they make an overridable recursive attribute set
<gchristensen>
no, if you 'override' version on a package, it still doesn't update the name
<sphalerite>
catern: it's what the overlay pattern is there for. The above is an idea of what it could look like when it's integrated into the nix language directly.
jrolfs has quit [Ping timeout: 246 seconds]
Ariakenom has joined #nixos
<sphalerite>
catern: I think the overlay pattern (with the self: super: stuff) is what you're talking about when you're saying "the helpers that build all-packages and python-packages"?
dbmikus has joined #nixos
<{^_^}>
[nixpkgs] @bjornfor pushed to master « jpeginfo: init at 1.6.1 »: https://git.io/fp6K6
arjen-jonathan has quit [Ping timeout: 250 seconds]
<azazel>
hey guys I've enabled pcscd from configuration options (which in turn installs the pcsclite package) and them as a user I've installed opensc suite of applications that depends on pcsclite too, but looking into the hashes of the running software in nix store, pcscd and opensc depend on different releases of pcsclite and that's is probably causing the fact that I'm not able to see my smartcard, how that's possible? How can I fix it?
LogicAside has quit [Ping timeout: 250 seconds]
jperras has quit [Quit: WeeChat 2.2]
periklis has quit [Ping timeout: 244 seconds]
<clever>
azazel: if you install those extra things via environment.systemPackages, it will be more likely to come from the same nixpkgs set and not have issues
<clever>
azazel: and its likely channels that are to blame, what does `nix-channel --list` show as both root and non-root
LogicAside has joined #nixos
<ottidmes>
clever: maybe you know, do you know how to add a loopback device to NixOS. For now I have worked around it by using the dummy kernel module, which allows you to create dummy network devices for testing purposes, which comes down to basically the same.
<azazel>
clever: strange, a "nix-env -u opensc" did not upgrade opensc to a working release but a "nix-env -e opensc && nix-env -iA nixos.opensc" did install a version whose dependency is the same as the system pcscd
<ottidmes>
azazel: its just the different outputs, as you can see the -bin postfix in pcscd
pkx3 has joined #nixos
sir_guy_carleton has joined #nixos
<fendor>
to those that can remember my problem with raspberry pi building the linux kernel from source: the channel was unstable, switching to stable fixed the issue
<azazel>
ottidmes: it doesn't seem to me like a valid reason for the discrepancy
<ottidmes>
azazel: ah, I see, the 23 vs 24 version... hmm
<azazel>
ottidmes: yes ;-)
Mateon1 has quit [Ping timeout: 240 seconds]
Mateon3 is now known as Mateon1
jtojnar has joined #nixos
ixxie has joined #nixos
<ottidmes>
azazel: not sure if that only works well with having your own nixpkgs checkout or whether its useful when using nix-channel, but maybe nix-env --file <nixpkgs> -u opensc would have worked? (I don't use nix-env, just systemPackages and nix-shell for me)
<clever>
azazel: `nix-env -i opensc` will search for a package with `opensc` in .name
__red__ has quit [Remote host closed the connection]
<clever>
shouldnt be any others though...
dsx has quit [Ping timeout: 252 seconds]
jperras has joined #nixos
<sphalerite>
,imperative azazel
<{^_^}>
azazel: nix-env has multiple drawbacks as an imperative package manager. nix-env -u will sometimes upgrade to the wrong thing; the outputs to install are very finicky to override; and packages that have been removed or are otherwise unavailable will remain in your profile without any warnings. Consider using a ,declarative setup instead.
<gchristensen>
dmj`: probably should wait at least more than 10min :P
<gchristensen>
there are three people who'd like to review it, indicatedby their listing themselves as codeowners. if they don't respond soon (but they probably will!) then someone else could merge I suppose
<cyris212_>
Could someone point me in the right direction on what this error means?
<cyris212_>
builder for '/nix/store/h2cd26qdmv8m5h2gxc8amwq58ga2yasb-dirvish-1.2.1.drv' failed to produce output path '/nix/store/9agq7jk70xvdrbk7mb9fj9aw219ib4l8-dirvish-1.2.1'
<dmj`>
gchristensen: alright :)
pie___ has joined #nixos
<dmj`>
gchristensen: it's just some utility thing, thought it could be a quickie
<dmj`>
gchristensen: not like a derivation or anything
<symphorien>
usually it means that the build didn't create $out
<symphorien>
but it does not completely fit the files you linked
dsx has joined #nixos
<symphorien>
this package does not need a dedicated builder
<ottidmes>
cyris212_: installPhase is not called automatically when using a builder
pie__ has quit [Ping timeout: 260 seconds]
<ottidmes>
cyris212_: just use installPhase attribute rather than a dedicated builder like symphorien suggests and it should work
<cyris212_>
symphorien: ottidmes: thanks a lot. I'm going to to try it out.
<enleth>
,declarative
<{^_^}>
There are multiple ways of managing declarative profiles. 1) Attrset, compatible with imperative use of nix-env https://git.io/fAQHW ; 2) buildEnv, providing more control over the paths that are linked into the profile https://git.io/fp0aU ; 3) home-manager, providing nixos-like config for your ~ https://github.com/rycee/home-manager
<{^_^}>
[nixpkgs] @domenkozar pushed to master « chrome-token-signing: fix path to opensc-pkcs11 »: https://git.io/fp6yC
<{^_^}>
[nixpkgs] @domenkozar pushed to release-18.09 « chrome-token-signing: fix path to opensc-pkcs11 »: https://git.io/fp6y4
<dmj`>
clever: it can't be relative to the directory the command is executed, since users have the repo on different locations
<dmj`>
s/on/at
<gchristensen>
./foo.txt is relative to where the .nix file which contains ./foo.txt, not by where the program is executed
<gchristensen>
if you have foo.bar and foo.nix in the same directory, any reference to ./foo.bar inside foo.nix will always correctly refer to foo.bar
<dmj`>
gchristensen: so foo.txt will always be in a directory on the file system that is relative to the users home path, the .nix scripts could be anywhere
<gchristensen>
yeah, check out that impure.nix link ^
<tdeo>
i actually figured it out, i was able to run overrideDerivation on wineStaging which should be enough... remaking patches now and going to build
<sphalerite>
ixxie: does ~/.nix-profile exist (even if its destination doesn't)?
<tdeo>
aaaaand i accidentally made the patches for wine 3.14 instead of 3.18
fusion809 has quit [Remote host closed the connection]
worldofpeace has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/fp6dE
fusion809 has joined #nixos
<sphalerite>
anybody know tools suitable for manipulating PDFs more or less losslessly, e.g. splitting into pages and rejoining?
<sphalerite>
hyper_ch: I know you do PDF things :p
reinhardt has joined #nixos
acarrico has quit [Remote host closed the connection]
<c0bw3b_>
sphalerite: the "standard" tool is pdfsam but I believe it is not part of nixpkgs
<avn>
sphalerite: alhtough would be nice to replace it with pure java version, due gcj deprecation in 6.x epoch
<sphalerite>
__monty__: rendering doesn't seem relevant to what I'm trying to do…
<tdeo>
hmm, if i override wineStaging it adds my patch before the rest of the staging patches
<tdeo>
i need it to be after the staging patches
<avn>
tdeo: preConfigure probably then, or postPatch (can't remember on what stage staging patches applying)
jrolfs has quit [Ping timeout: 268 seconds]
<sphalerite>
hm I don't see anything that splits a PDF into pages in mutool
<avn>
tdeo: overrideDerivation staging (self: { postPatch = self.postPatch + yourcode; }) should work
<sphalerite>
avn: pdftk does seem to have something though, thanks!
<tdeo>
thanks, i'll try that
<avn>
sphalerite: only one disadvantage -- it pull gcc/gcj 6.x, which builds solely for it. Is a pure jave fork exists, but I am not too familiar with java stuff.
<sphalerite>
avn: eh, it's a one-off and I'm not short on space
acarrico has joined #nixos
<avn>
sphalerite: my weekly updates sometimes going too long ;)
<__monty__>
sphalerite: Unless you want to convert the pdfs to jpg or read them, sure. Having great rendering won't be an obstacle though.
jrolfs has joined #nixos
<sphalerite>
__monty__: not part of my use case :p I do already use mupdf for actually viewing pdfs, or with a tiny patch for presentation fun (view with the current slide on the projector and the current and next slides on the laptop monitor, + remote control from my phone) :D
<neonfuz>
Hello
<__monty__>
Ooo, nice.
<neonfuz>
the latest anki on nixos-unstable doesn't compile for me
<neonfuz>
I was wondering, just so I can finish this nix-env -u if there's an easy way to pin packages
<neonfuz>
I thought there was
<neonfuz>
I looked up pinning packages, but I just want to temporarily prevent nix-env from updating it
<neonfuz>
I don't want to make it permanant
<ottidmes>
neonfuz: so you are normally on nixos-unstable, but you want to use nixos-18.09's version for anki, I would use the following which also works the other way around (adding stable to otherwise unstable based system)
<sphalerite>
__monty__: https://github.com/lheckemann/presremote it's an awfully hacky HTTPish server written with bash and netcat but hey it works :p the viewing the next slide stuff is just by running multiple mupdf instances and shuffling them around with i3 and teeing the commands into all 3 instances
<neonfuz>
that's not what I want
<neonfuz>
I want to temporarily keep my installed version
<neonfuz>
I know there's a way to do this because I've done it before
<ottidmes>
neonfuz: then just add an root to it?
<neonfuz>
I just don't remember which nix command it's buried under
<neonfuz>
is that how you do it?
<sphalerite>
neonfuz: use nix-env -iA <packages> where <packages> are the ones that aren't anki
<neonfuz>
yeah ik
<neonfuz>
I'm sick of doing that though
<neonfuz>
every time a package beraks on me I manually do that
<neonfuz>
and it takes forever
<ottidmes>
neonfuz: adding a root for it would prevent it from garbage collecting, but if its still in nix-env, the newer version would be installed, its just that the old version will be kept in the store
Laalf has joined #nixos
<sphalerite>
neonfuz: actually maybe using --prebuilt-only with -u will do, as a hack
<ottidmes>
neonfuz: but if its just the bin folder you need, you could just add it to your $PATH
<sphalerite>
but -u is kind of awful
<sphalerite>
,imperative
<{^_^}>
nix-env has multiple drawbacks as an imperative package manager. nix-env -u will sometimes upgrade to the wrong thing; the outputs to install are very finicky to override; and packages that have been removed or are otherwise unavailable will remain in your profile without any warnings. Consider using a ,declarative setup instead.
<__monty__>
,declarative
<{^_^}>
There are multiple ways of managing declarative profiles. 1) Attrset, compatible with imperative use of nix-env https://git.io/fAQHW ; 2) buildEnv, providing more control over the paths that are linked into the profile https://git.io/fp0aU ; 3) home-manager, providing nixos-like config for your ~ https://github.com/rycee/home-manager
<neonfuz>
oh I found it
<neonfuz>
nix-env --set-flag keep true <package>
<ottidmes>
neonfuz: export PATH=/nix/store/...-anki-.../bin:$PATH, in environment.shellInit probably (or systemd service, or activation scripts (be careful with those))
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/fp6NR
<neonfuz>
ottidmes: I found the solution I was looking for ^
<Laalf>
hello! does someone run optimus on a laptop? i have a thinkpad w530 and i am not able to set it up. i have the external outputs wired to the nvidia chip. so i need the nvidia card
dsx has quit [Quit: dsx]
<{^_^}>
[nixpkgs] @peti pushed 4 commits to haskell-updates: https://git.io/fp6Nz
<neonfuz>
man it's sorta crazy that you guys didn't know about '--set-flag keep true', nix has too many features I swear lol
<ottidmes>
neonfuz: I dont use nix-env, pretty sure a lot of people dont
<neonfuz>
ah yeah
gspia has quit [Quit: Leaving]
m0rphism has joined #nixos
<Laalf>
i was able to get optimus to run with glxgears, that is no problem. what is a problem are my external outputs
<jw358>
is there a link i should read about locale problems on travis?
<fusion809>
Hi folks, I'm writing a nix file for building the Unplugged RA mod of OpenRA. The problem is that it needs its game engine (OpenRA itself) to be fetched using a script, fetch-engine.sh, that calls Python. As this is needed before the build, so that the engine files can be patched to work on NixOS, I have placed the fetch-engine.sh command in "extraPostFetch". The problem is that Python isn't being detected in extraPostFetch.
<fusion809>
How do I give it Python? What adjustments to PATH are needed?
jrolfs has quit [Ping timeout: 245 seconds]
<fusion809>
The dir /run/current-system/sw/bin doesn't have it, although $HOME/.nix-profile/bin does. But I'm not sure if I should point it to something that is only on my local system.
<vaibhavsagar>
fusion809, can you replace your invocation of python with "${pkgs.python3}/bin/python"
<fusion809>
That sounds excellent :). Thanks.
mehak has left #nixos ["ERC (IRC client for Emacs 26.1)"]
<fusion809>
vaibhavsagar: now I'm getting this error: https://pastebin.com/B3mLcjFT (Python 3 isn't used by OpenRA, Python 2 is, IIRC).
<avn>
fusion809: build a OpenRA as derivation, build mod using existing OpenRA as library?
<fusion809>
Must admit I'm a Nix beginner as what you're suggesting sounds beyond my abilities. I'm trying to build it the same way I build my RPM and Arch packages for this mod.
<avn>
fusion809: or if you want build all-at-once, just replace fetch-engine.sh to copy pre-fetched engine sources from store (and write fetch derivation)
<avn>
fusion809: network activity is prohibited during build
<fusion809>
OK, so you're suggesting adding an extra source to the src field in the Nix file?
<timokau[m]>
fusion809: Basically you can only access the network when you provide a hash of the expected result
<avn>
I can't say, until look on sources and/or your expression
<timokau[m]>
So you won't be able to download something in postFetch
<avn>
I'd suggest to patch out built-in fetching, and unpack OpenRA sources in prePatch
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #51469 → gimp: fix on darwin → https://git.io/fpiv0
<fusion809>
Here's my default.nix file, https://gist.github.com/8b621e5ee921960b16e1b8bee9a6d764. Must admit, most of what you's have said is beyond me. I have no idea what much of what you're saying means. As I said, I'm a beginner.
<fusion809>
I know how to add additional sources to RPM and Arch package spec and PKGBUILDs, respectively, but how to do so with a Nix file is beyond me.
<{^_^}>
error: syntax error, unexpected ')', expecting ID or OR_KW or DOLLAR_CURLY or '"', at (string):211:1
<dmj`>
boxscapeR: use that nix repl
<boxscapeR>
thank you
eliasp has joined #nixos
<vaibhavsagar>
boxscapeR: `nix-instantiate --eval '<nixpkgs>' -A "pythonPackages"`
<Phillemann>
Hm. anki fails with the latest nixos-unstable.
<Phillemann>
Shouldn't all software be built by hydra?
<timokau[m]>
fusion809: Is the `fetch-engine` script actually necessary? i.e. is it doing anything complicated? Otherwise you could just fetch the engine source directly using the nix fetchers
<c0bw3b_>
I'm pretty sure I still don't understand what goes into staging-next
<timokau[m]>
c0bw3b_: It is forked from staging, so it will be mass rebuilds. After the fork, it will only get fixes
<fusion809>
Not sure if it is necessary. They actually use their own fork of the OpenRA engine. How do I add additional fetchers? Like I've tried rm the ; at line 32 and ++ fetchzip { at line 33 (with under it the details of where the engine's URL is).
<timokau[m]>
fusion809: You could just do something like `engine = fetchFromGitHub ...`. Then `${engine}` will be a variable that points to the downloaded engine
<fusion809>
OK, and it'll automatically download it? Or do I have to call it somewhere>?
crmlt has joined #nixos
graphene has quit [Remote host closed the connection]
<c0bw3b_>
timokau[m] ok it gets clearer, but still find it strange that commits from 12hours ago are already in staging-next
<timokau[m]>
c0bw3b_: Wasn't staging-next just very recently forked from staging?
graphene has joined #nixos
<c0bw3b_>
timokau[m] no idea
jrolfs has joined #nixos
<timokau[m]>
c0bw3b_: Yes, looks like that was just 9 hours ago. So 9 hours ago we started a new cycle by taking everything in "staging" and putting it into "staging-next". New stuff will now go to "staging", while "staging-next" will stabilize (being built on hydra, bugs fixed). Then "staging-next" will be merged into master and "staging" becomes the next staging-next
<timokau[m]>
fusion809: It works just as it does with `src`. It will automatically download it (as long as it is accessed somewhere, its lazy)
<timokau[m]>
fusion809: Okay `src` is a bad example because it is used in the `unpackPhase` and you don't have to access it directly, but in principle you could access the `${src}` variable
ng0 has joined #nixos
<timokau[m]>
So you could do `echo ${engine}` which would print the path the engine was downloaded to. Or `cp ${engine} ./engine`. Note that `${engine}` is *not* a shell variable but a nix variable.
crmlt has quit [Quit: leaving]
<c0bw3b_>
timokau[m]: then I guess any PR toward staging that is ready should be merged sooner then
<{^_^}>
#51454 (by avnik, 6 hours ago, open): mythes: fix link error on missed `tinfo` library (ncurses)
<c0bw3b_>
yes i've seen it quickly, not familiar with libreoffice and such though
<c0bw3b_>
issue happened after a bump of hunspell?
<avn>
c0bw3b_: just broken transitive dependency, not LO itself. Probably yes, broken by hunspell. Because it was built week ago.
ixxie has quit [Ping timeout: 246 seconds]
<fusion809>
I'm looking at openttd and how it works with extra sources. So far I have: https://gist.github.com/e7e91345c9eb2f101f281edcba08cf6c. It doesn't seem to be fetching engine, and decompressing it as an engine subdirectory isn't being detected.
palo1 is now known as palo
<fusion809>
I'm hoping the engine is fetched, decompressed and placed within the uRA directory and named 'engine'. But, if I'm following how this works, it doesn't seem that directory is being created. I do not see any sign of the engine bing fetched.
<fusion809>
being^
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<hyper_ch>
sphalerite:pdf things?
<c0bw3b_>
avn I wonder if the real fix shouldn't rather be to declare ncurses as a propagatedBuildInputs in hunspell
<c0bw3b_>
will comment on the issue
civodul has joined #nixos
sigmundv_ has joined #nixos
<Phillemann>
Hm, steam also broken in nixos-unstable.
<avn>
c0bw3b_: you are right. Although I thrilling due upcoming ~12 hour rebuild of everything :/
<ottidmes>
Phillemann: not to say that its not a problem (i.e. it should not be broken on unstable), but if you want a more stable system in general, I would recommend just using the stable nixos-18.09 channel and combine it with an unstable channel for those probably few packages that you really need have the bleeding edge versions of
<c0bw3b_>
avn: ok I'll cook an express PR
<Phillemann>
ottidmes: Somebody told me once that every package gets built by hydra, and I'm wondering how these failures can even happen.
<Phillemann>
I mean, it's a 404 with steam.
<avn>
c0bw3b_: np, I I'll doing now myself.
<c0bw3b_>
avn: ok
<avn>
I believe no need to rebuild LO and chromium to proof that all works
<Phillemann>
ottidmes: It's actually ok for me to be unstable. I like to contribute issues and fixes. :)
<__red__>
if there's a way to tag a PR for security reasons - I don't know how
<eliasp>
what's the correct way to extend a list in Nix? mylist = mylist + ["foo"]? or is there some in-place operation like mylist.append("foo")?
freeman42x has quit [Quit: Leaving]
<c0bw3b_>
__red__: done
<__red__>
thanks c0bw3b_
lnikkila has quit [Ping timeout: 244 seconds]
<simpson>
eliasp: `mylist ++ [ "foo" ]`
<eliasp>
simpson: thanks
<simpson>
There are not any in-place mutable operations on Nix values that I know of, although I'm surely about to be proven wrong.
<__red__>
I didn't know if it was appropriate to add my name to the list of maintainers for the package
<__red__>
so I didn't
<boxscapeR>
ugh, I feel stupid. I just spend half an hour debugging because I used `<path>` instead of `import <path> {}`
<boxscapeR>
I feel like types would have helped with that...
<boxscapeR>
s/spend/spent
<ottidmes>
Phillemann: In that case, please keep doing that! :) But its not true that every package is succesfully build by Hydra before its moves the channel to a new commit. It is always the aim though to get as close to zero failures as possible at every stable release
<avn>
c0bw3b_: updated, at least pkgs.mythes builds with propagated (I push'd to same PR branch)
<c0bw3b_>
__red__ well adding yourself to the list without forcibly replacing any other name should always be accepted
<__red__>
RGR, thanks.
<__red__>
although strictly speaking I'm not sure it's a package "per se", so I just didn't want to futz with it
<__red__>
I just wanted the fix in
<Phillemann>
ottidmes: Ah, I see.
<c0bw3b_>
__red__: yheah exactly I'm writing a comment on your PR directly... fetching from a service module is weird.. it should be a package
<__red__>
Honestly - it's over my pay-grade to know how to fix that.
<__red__>
I figured it would be better for the fix to be in than for me to ignore it
<__red__>
<--- Still very much at the "Cargo Cult"-stage of NixOS contributions :-/
<ottidmes>
Phillemann: it just needs to pass certain tests, which is why you should use nixos-unstable rather than nixpkgs-unstable, because nixos-unstable has tests that make sure nixos should generally still work properly, but it will have a lot more issues and uncached packages than nixos-18.09 has
<c0bw3b_>
__red__: NP. it still a good move to just update the service module as-is
<__red__>
maybe it's something I can have a go at after this one gets committed
<__red__>
I'm happy to maintain it
<fusion809>
When I try to edit files in $engine, like with sed, in postPatch and preConfigure I get permission denied errors. How do I delete the fetch lines in it? This is what I have atm https://gist.github.com/a8b5066b02178d9dc486c12ea51dfa52. I'm trying to debug this, but this is turning into a time-consuming nightmare.
kisik21 has joined #nixos
<__red__>
c0bw3b_: I'm also going to add some configuration docs to the wiki as right now there isn't much (any?) in the way of documentation on how to configure it
<c0bw3b_>
avn: ah good :) you have a PR incoming?
<ottidmes>
fusion809: you are trying to edit things in the nix store, since ${engine} will result into the path to the nix store where the engine is stored, and nix store is readonly
<c0bw3b_>
__red__: karma points for you if you do. moar docs (anywhere) is always welcome yes ^^
hamishmack has joined #nixos
<ottidmes>
fusion809: and since you copy with attributes in places, you keep things readonly
<fusion809>
OK, so how do I edit what I need to, when I'm trying to build uRA? I need an engine directory in uRA, that contains those files, so I'm guessing this attributes talks means I'm going to have to find a cp option to remove these attributes?
vk3wtf has quit [Ping timeout: 268 seconds]
<fusion809>
Well all I find attributes-wise in the cp man pages pertains to the --attributes-only option, which sounds like the opposite to what I want.
<ottidmes>
fusion809: the proper solutions is probably srcs, but if you want to do it like you do know, you can do: cp --no-preserve=mode
<avn>
fusion809: `cp -v -r --no-preserve=all ${engine} engine` (remove -v when you see that all works)
m0rphism has quit [Quit: WeeChat 2.2]
m0rphism has joined #nixos
jackjennings has joined #nixos
<fusion809>
avn: thanks for trying to help, I do appreciate it, same thanks goes to ottidmes. Unfortunately the cp -v -r --no-preserve=all... copies it fine, but it seems still to be read-only as line 52 returns: "/nix/store/r88k69yw8zikrpsj1pkk3xgg1vi3cx0g-stdenv-linux/setup: line 101: engine/thirdparty/fetch-thirdparty-deps.sh: Permission denied"
<fusion809>
Oops, that might be line 53.
<fusion809>
I'll see what it's doing wrong.
<simpson>
fusion809: It's likely no longer executable; you may have to run sh on it directly, I think.
<fusion809>
Ah, so add `sh` to the start of line 53?
<avn>
fusion809: you need to do something with fetch-thirdparty deps as well ;)
<fusion809>
Ah, well I found what seems to have fixed things, chmod +x engine/thirdparty/*.sh
doyougnu has quit [Ping timeout: 268 seconds]
<gchristensen>
ah... that is a reason to not extract in the fetcher, infinisil :)
<tejing>
I'm trying to move to the 18.09 channel, and I'm finding that I can't install. It says it 'cannot download libacl1.deb from any mirror'. Tracing the dependencies, the problem seems to be coming from steam. Any ideas how to get around this easily?
<fusion809>
The problem is that now I'm getting the error: unzip: cannot find or open StyleCopPlus.MSBuild.zip, StyleCopPlus.MSBuild.zip.zip or StyleCopPlus.MSBuild.zip.ZIP. I copied the buildInputs from OpenRA, hoping that it would provide the files needed by engine/thirdparty/* scripts, as it uses the same inputs as OpenRA did.
<{^_^}>
#49048 (by xurei, 5 weeks ago, closed): Steam won't install - cannot download libvulkan1.deb from any mirror
<tejing>
c0bw3b: yeah, found that issue, and it does seem similar. still doesn't tell me what to do about it though
doyougnu has joined #nixos
vidbina has joined #nixos
<__red__>
I have steam installed on my local machine which means I probably have the source files in my nix/store if we need to upload them somewhere
<avn>
ottidmes: this approach works, if source designed to be unpacked side-by-side, nut by internal fetch-some-crap script
<avn>
fusion809: check from where this archive should come, or why it unpacked. Sometimes it even not need to build
<avn>
(although I feel pain each time when I see any dotnet code. Sometimes I though, that it not desogned for clean from-source build, only to co-existing with some artifact storage)
<c0bw3b_>
tejing: actually it may well be a new that could require running the steam update script again.. could you open a new issue please?
<ottidmes>
avn: sure it works, like I linked, I am using it myself too, but only because I had no other choice, but the srcs example I linked should end up with the a similar situation, maybe needing a simple move rather a cp from /nix/store
<c0bw3b_>
tejing: and mention Synthetica9 and Infinisil who tackled the previous issue
<ottidmes>
avn: tell me about it, vsliveshare was a pain to get working properly (half nodejs half dotnet)
sanscoeu_ has joined #nixos
mmlb has quit [Read error: Connection reset by peer]
mmlb has joined #nixos
<symphorien>
can I declare a remote builder to have several architectures ? I am building i686 derivation but they are not sent to my x86_64 builders
<gchristensen>
comma separated I think
sanscoeu_ has quit [Remote host closed the connection]
sanscoeur has quit [Ping timeout: 268 seconds]
freeman42x has joined #nixos
graphene has quit [Remote host closed the connection]
ng0 has quit [Quit: Alexa, when is the end of world?]
<attente[m]>
i'm struggling to understand what the proper approach to development is under nixos. how are things like pkg-config, CFLAGS, and LD_LIBRARY_PATH, etc handled properly in this environment?
<clever>
attente[m]: nix-shell handles all of those vars
<symphorien>
i686 also works for i686 bsd and so on, right ? I would like i686-linux
<clever>
attente[m]: `nix-shell -p pkgconfig`
<attente[m]>
clever: thanks, this looks useful :)
<{^_^}>
[nixpkgs] @bhipple opened pull request #51480 → fetchurl: fix and add extra KDE mirrors → https://git.io/fpiYE
<attente[m]>
clever: for some reason when i try `pkg-config --list-all` in a nix-shell, i get no results though :(
<ottidmes>
in Nix to get proper set like behavior, the only two options I got are lib.unique and attrNames (listToAttrs (map (value: nameValuePair value true) ...))?
<symphorien>
ah apparently meta.platform can be a list of strings
<clever>
attente[m]: you need to put other packages in the -p list, to see them
<ottidmes>
avn: it also does not like being in a readonly directory, or it generates binaries after the initial run, that do need to be patched in order to work, fun!
__monty__ has quit [Quit: leaving]
<symphorien>
gchristensen: now that ofborg counts as checks, if I force push my PR, the history of ofborg builds seems to be lost. It may make it hard to follow discussions.
epicmetal has joined #nixos
<avn>
ottidmes: not sure. but if you know how to force nuget work with chain on repos from buildInputs, you win ;) I am not big expert in dotnet, just updated mono couple times.
johanot has joined #nixos
<{^_^}>
[nixpkgs] @Mic92 merged pull request #51478 → libffi: swap src URL from FTP to HTTPS → https://git.io/fpiml
<ottidmes>
avn: I have done like 2 projects in .NET and packages 1 on NixOS, so don't know much about it either, but what you asks seems doable to me. Just create a global packages folder with symlinks to the right store paths and provide the path with: NUGET_PACKAGES
<{^_^}>
[nixpkgs] @polynomial opened pull request #51485 → Add mock to the list of whisper's test inputs. → https://git.io/fpiOP
<gchristensen>
symphorien: aye, what do you think we could do about that?
pkx3 has quit [Read error: Connection reset by peer]
nD5Xjz has quit [Ping timeout: 246 seconds]
pkx3 has joined #nixos
<symphorien>
Nothing comes to my mind right now
<xantoz>
I have a system that fails on systems-fsck@<dev-disk-by-uuid-to-the-EFI-partition>. It boots fine on the first generation, but any of my later generations fail for some reason.
<c0bw3b_>
what is the workaround to have ofBorg running a test?
<ottidmes>
avn: it would be cool if it could be something like: nugetWithPackages [ dep1 dep2 ]
<avn>
yup
doyougnu has quit [Ping timeout: 250 seconds]
<ottidmes>
which could just be a wrapper with NUGET_PACKAGES set
<eliasp>
when developing a module, how do I properly test it using my current environment? for a new one, I simply included it in my configuration.nix (see: https://github.com/NixOS/nixpkgs/pull/51474), but what's the right approach for one which is already part of my regular channel? just ending up with "The option … is already declared in …" this way for those changes:
<ramirez7>
Hello, I'm having trouble adding keys to my NixOps deployment: `users.extraUsers.root.openssh.authorizedKeys.keys` is set to a list of 3 keys. I successfully deploy, and I see that `/etc/ssh/authorized_keys.d/root` contains all 3 keys. But `/root/.ssh/authorized_keys` does not seem to have been updated.
<ramirez7>
And my new keys are rejected
<ottidmes>
avn: exactly, it would create a dir with those links set up properly and in /bin/nuget it would have a nuget wrapper with NUGET_PACKAGES pointing to that dir
<gchristensen>
symphorien: I think the new mechanism is a big improvement over the comments, even if it isn't quite as nice in the force push case
<ottidmes>
ramirez7: nixops add them on the fly, you could add them explicitly the way you want by just exporting them and adding them to your config yourself
<ramirez7>
ottidmes: hmm I'm not following. I have the public keys (from me + my team members) pasted into my nixops source code for the machine & would like `nixops deploy` to authorize all of them. fwiw, the machine itself wasn't spawned or managed by NixOps (I provisioned it on DO & put the IP in my Nix source)
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #51401 → all-packages.nix: Alias self to res, deprecating self → https://git.io/fpKEz
<avn>
ottidmes: pkgs/games/ckan/default.nix maybe? (but newer release, which is completely nuget based)
mkoenig has quit [Ping timeout: 246 seconds]
<ottidmes>
ramirez7: ah, I thought you had a case of, using nixops, and then rebuilding directly yourself, because then nixops can no longer auth in it, because the ad-hoc added public key used by nixops would then be removed by the rebuild
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
mkoenig has joined #nixos
<ramirez7>
I believe my `/root/.ssh/authorized_keys` file was not created my Nix (and instead created by DO upon provisioning when I pasted my public key into their UI)
<ottidmes>
avn: honestly I have almost no experience with mono, I am looking at ckan, but where would that nugetWithPackages come into play?
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<avn>
ottidmes: newer version hae all crap like ./build shellscript, which try downloading nuget.exe, *.csproj which call nuget restore from under guts, etc.
<avn>
honestly most (if not all) our mono/.net related stuff need to be revamped
<ottidmes>
avn: ah like so, yeah that sounds problematic, because that would mean things would get stale, right (if newer versions would no longer work with current setup)
<{^_^}>
[nixpkgs] @c0bw3b pushed 2 commits to release-18.09: https://git.io/fpisb
<__red__>
c0bw3b_: am I out of line kicking the can down the road on the operational things too?
doyougnu has joined #nixos
<__red__>
<-- still trying to understand NixOS culture
<avn>
ottidmes: well, I'd like to do something with (and for) opensource stuff.
<c0bw3b_>
__red__: no PR are here for this (review, feedbacks, taking another approach or completing stuff)
<ottidmes>
avn: of course! I am packaging something right now, after I am done with that, I will see if I can update ckan and see what kind of hell will break lose and what we can do about it :P
<c0bw3b_>
__red__: this PR is going to update on master only, we would then need to backport the change on release-18.09 for stable users
dbmikus has quit [Quit: WeeChat 2.3]
<c0bw3b_>
(precising because you mention exisitng installations)
<avn>
ottidmes: would be cool. Tell me if I can help with review/testing/whatever else.