Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v7oRA
<NixOS_GitHub>
nixpkgs/master 1ec7242 Tim Steinbach: linux-copperhead: 4.12.4.a -> 4.12.5.a
NixOS_GitHub has left #nixos []
rpifan has quit [(Read error: Connection reset by peer)]
stasku has joined #nixos
takle has joined #nixos
mizu_no_oto has joined #nixos
Mateon3 has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
Sonarpulse has joined #nixos
Mateon2 has quit [(Ping timeout: 260 seconds)]
Mateon3 is now known as Mateon2
rpifan has joined #nixos
mkoenig has quit [(Quit: Lost terminal)]
mkoenig has joined #nixos
mkoenig_ has joined #nixos
takle has joined #nixos
mkoenig_ has quit [(Client Quit)]
mkoenig has quit [(Client Quit)]
mkoenig has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
takle has joined #nixos
Piece_Maker has joined #nixos
Acou_Bass has quit [(Ping timeout: 260 seconds)]
Piece_Maker is now known as Acou_Bass
takle has quit [(Ping timeout: 260 seconds)]
Infinisil has quit [(Quit: leaving)]
jgertm has quit [(Ping timeout: 240 seconds)]
tommd has quit [(Ping timeout: 260 seconds)]
takle has joined #nixos
Wizek has quit [(Quit: Connection closed for inactivity)]
mkoenig has quit [(Ping timeout: 240 seconds)]
takle has quit [(Ping timeout: 260 seconds)]
phinxy has quit [(Read error: Connection reset by peer)]
takle has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
eschnett has joined #nixos
takle has joined #nixos
LinArcX has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] benley pushed 1 new commit to master: https://git.io/v7ouW
<NixOS_GitHub>
nixpkgs/master 989514f Benjamin Staffin: jsonnet: 0.8.9 -> 0.9.4 (#27989)
NixOS_GitHub has left #nixos []
Infinisil has joined #nixos
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
mkoenig has joined #nixos
takle has quit [(Ping timeout: 240 seconds)]
<adisbladis>
Having some issues with nixops.. Stuck on "waiting for ssh" on ec2
<adisbladis>
Don't know where to go from here.. The aws debugging experience is not great
<simpson>
adisbladis: On GCE, those sorts of situations usually happen because I've accidentally stopped a machine and forgotten to restart it and nixops also forgets to restart it. So maybe go into the AWS console and make sure that the machine's up and running?
<simpson>
And yes, the AWS debugging experience sucks.
<adisbladis>
simpson: It was just created by nixops, its in a security group where all traffic is allowed
<adisbladis>
I hate that I can't attach a console on aws :/
schoppenhauer has quit [(Ping timeout: 268 seconds)]
slyfox has quit [(Ping timeout: 240 seconds)]
schoppenhauer has joined #nixos
<adisbladis>
simpson: Ahhh Amazon in all their wisdom decided that letting through "all" traffic does not mean ssh for some reason
<adisbladis>
Explicitly adding port 22 to the security group works
<simpson>
adisbladis: AWS has wisdom? I thought that they just had two pizzas and zero outages~
<Infinisil>
unlmtd: I think there's an issue about wayland support
nix-gsc-io`bot has quit [(Client Quit)]
takle has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
sary has quit [(Ping timeout: 246 seconds)]
dev1 has joined #nixos
reinzelmann has joined #nixos
<unlmtd>
update: ran ```way-cooler``` from tty and POOF
<unlmtd>
but then I couldnt do anything, and couldnt switch tty nor switch back to way-cooler, but I saw the last few lines of the tty running it, and it looked really sane
<unlmtd>
so I guess it works
<unlmtd>
I just need a way to launch it. lightdm seems to be out
<dev1>
what does "..." mean in a .nix file?
<dev1>
such as { config, pkgs, ... }
<clever>
dev1: that lets it ignore any other arguments it was given
<dev1>
whats does ... actually do? can i use it elsewhere?
<clever>
it only works in the function arguments like that
<dev1>
so as a function parameter ... tells it to ignore all other parameters?
<clever>
yep
<dev1>
why do we want to ignore "config, pkgs"?
<dev1>
since its default nix configuration
<dev1>
whats the use case for it?
<clever>
it ignores all OTHERS
<clever>
config and pkgs are the only ones it will use
<dev1>
oh I see...
<dev1>
I understand
<clever>
nixos may pass it more, but only config and pkgs get used
<dev1>
so why do we want to make sure only config and pkgs get used?
<clever>
because your module may break in 6 months when anybody adds a new argument to all modules
<clever>
and it wont accept unknown argumentsd
<joehh>
what is the best way to delete a profile
<joehh>
?
<joehh>
or equivalently uninstall all packages from it?
<clever>
joehh: nix-env -r
<dev1>
thank you clever I understand now
<joehh>
great - so nix-env -p myProfile -r -i mypackage will ensure that only mypackage is installed
<joehh>
exactly what I need - thanks for that
_rvl has joined #nixos
<joehh>
hey clever - what timezone/where in the world are you? Whenever I am on, you seem to be there and active :)
<clever>
NB canada, atlantic standard time
<joehh>
wow - almost completely out of phase! -I'm in Melbourne, Australia +10 at the moment - thanks for all of your assistance, must all have been at crazy hours for you
<clever>
lol
<clever>
joehh: i'm more of a nightowl, so i can see how we might be on at the same times
<joehh>
:)
rpifan has quit [(Remote host closed the connection)]
rpifan has joined #nixos
alx741 has quit [(Quit: alx741)]
Ivanych has joined #nixos
LinArcX has joined #nixos
<Infinisil>
Same for me, stayed up all night again..
<Infinisil>
7am now
alx741 has joined #nixos
<LinArcX>
Hi. I install tor but I want to use polipo to convert socks to https. How do that?
<sphalerite>
Is it sensible at all to use fail2ban to ban IPv4 addresses that are scanning for phpmyadmin and stuff, or is CGNAT deployed widely enough that that's likely to be a problem for legitimate visitors?
takle has joined #nixos
takle has quit [(Ping timeout: 246 seconds)]
oida_ has joined #nixos
oida has quit [(Ping timeout: 240 seconds)]
jgertm has quit [(Ping timeout: 240 seconds)]
alx741 has quit [(Quit: alx741)]
<joehh>
sphalerite: I think you need to assess your likely visitors
<joehh>
Here in oz, my gut feel is that it is pretty widely deployed, especially on our mobile networks
<joehh>
but last time I was involved with running fail2ban in that environment, all of our scans were coming
<adisbladis>
sphalerite: cgnat is pretty widely deployed on mobile networks worldwide
<joehh>
from countries that were irrelevant to the local company
<joehh>
I ended up banning them all, and reviewing each day or so. Only once or twice decided to
<joehh>
remove a ban manually, not sure if it affected any visitors or not though
takle has joined #nixos
<adisbladis>
I find that most of the time even a short-lived ban is pretty effective
<adisbladis>
If requests time out they just give up
takle has quit [(Ping timeout: 255 seconds)]
Capprentice has quit [(Ping timeout: 240 seconds)]
LinArcX has quit [(Remote host closed the connection)]
reinzelmann has joined #nixos
revtintin has joined #nixos
periklis has joined #nixos
leat has quit [(Ping timeout: 276 seconds)]
takle has joined #nixos
MP2E has joined #nixos
joehh has quit [(Ping timeout: 260 seconds)]
dev1 has quit [(Quit: WeeChat 1.7.1)]
takle has quit [(Read error: No route to host)]
[0x4A6F] has quit [(Remote host closed the connection)]
joehh has joined #nixos
ertes has quit [(Ping timeout: 248 seconds)]
ebzzry has quit [(Quit: WeeChat 1.9)]
oahong has quit [(Ping timeout: 260 seconds)]
Phlogistique has joined #nixos
oahong has joined #nixos
oahong has quit [(Changing host)]
oahong has joined #nixos
Wizek has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gebner pushed 2 new commits to master: https://git.io/v7oKO
<NixOS_GitHub>
nixpkgs/master 5b996f9 Matthew Bauer: emscripten: 1.37.13 -> 1.37.16
<NixOS_GitHub>
nixpkgs/master 28ca51c Gabriel Ebner: Merge pull request #27899 from matthewbauer/emscripten-update...
NixOS_GitHub has left #nixos []
hydraz has quit [(Ping timeout: 255 seconds)]
rpifan has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 1 new commit to master: https://git.io/v7oKP
<NixOS_GitHub>
nixpkgs/master f152749 Peter Simons: hackage-packages.nix: automatic Haskell package set update...
NixOS_GitHub has left #nixos []
hydraz has joined #nixos
hydraz has quit [(Changing host)]
hydraz has joined #nixos
silver_hook has joined #nixos
silver_hook has quit [(Changing host)]
silver_hook has joined #nixos
sjourdois has quit [(Quit: ZNC - 1.6.0 - http://znc.in)]
sjourdois has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 6 new commits to staging: https://git.io/v7o6n
<NixOS_GitHub>
nixpkgs/staging 5ba3972 John Ericson: cc-wrapper: Stop preWrapping...
<NixOS_GitHub>
nixpkgs/staging 9f1e009 John Ericson: cc-wrapper: Unconditionally use @infixSalt@ accross the board...
<NixOS_GitHub>
nixpkgs/staging 42f3550 John Ericson: cc-wrapper: Make hygienic...
NixOS_GitHub has left #nixos []
sjourdois has quit [(Client Quit)]
sjourdois has joined #nixos
zeus__ has quit [(Remote host closed the connection)]
ertes-w has joined #nixos
takle has joined #nixos
FRidh has quit [(Quit: Konversation terminated!)]
linarcx_ has joined #nixos
LinArcX has joined #nixos
<linarcx_>
hi. i want to set zsh as default shell.how do that?
<tilpner>
users.users.linarcx.shell = pkgs.zsh;
<tilpner>
And probably programs.zsh.enable = true; somewhere
<linarcx_>
and what is it: users.defaultUserShell = pkgs.zsh ?
<linarcx_>
what is difference between two approaches?
<tilpner>
Oh, that will affect all users, instead of just you. I assumed you wanted to change the default shell for just your user
<linarcx_>
ok thanks. i will use your approache :)
<linarcx_>
thanks.but another question. i want to use prezto instead of oh-my-zsh. how do that? just copy .zpreztorc to home directory or exists an option for this?
<mg>
I'm trying to create a script to setup a tmux session a project. The first line of the script is "nix-shell", and it seems nothing after that gets executed. Why is that? Is there any way around it?
MoreTea has joined #nixos
joehh has quit [(Ping timeout: 240 seconds)]
thc202 has joined #nixos
<clever>
mg: nix-shell spawns a new shell, and only returns once the user exits from that
<clever>
mg: you may want to look into nix-shell --run
* clever
heads to bed
<mg>
clever: thanks, but it's a bit tedious to run nix-shell --run on every command.. Ideally I'd just like everything in the tmux session spawned later in the script to be run in the context of the nix-shell..
linarcx_ has quit [(Remote host closed the connection)]
<mg>
ah, I can just run nix-shell --run ./myscript.sh ofcourse. Thanks clever :)
linarcx_ has joined #nixos
faffolter has joined #nixos
faffolter has quit [(Changing host)]
faffolter has joined #nixos
moesasji has joined #nixos
dywedir has joined #nixos
takle has quit [(Remote host closed the connection)]
<moesasji>
Hi, I'm trying to get a patch ready for submission. Unfortunately stuck on "git rebase master". Anyone able to help?
<MoreTea>
moesasji, stuck how exactly?
<moesasji>
I get an error when running rebase: could not apply 69d75ce207
<moesasji>
This is completely unrelated to any file I touched
<MoreTea>
I recommend to learn more about git via 'git-from-the-bottom-up'
<moesasji>
Have read those pages in the past (and yesterday when hitting this). Not something that helps me out of this as I don't see why this happens.
<MoreTea>
Sorry, can't help you then.
MP2E has quit [(Quit: leaving)]
takle has joined #nixos
<moesasji>
OK....thanks for looking at it! Hopefully someone who has seen this will pop by.
_rvl has quit [(Ping timeout: 240 seconds)]
goibhniu has joined #nixos
takle has quit [(Ping timeout: 260 seconds)]
<LnL>
what I usually do when a rebase fails is to do the cherry-pick manually
<LnL>
so git reset --hard origin/master and git cherry-pick <commits...>
<linarcx_>
friends. i install nixos and i want to install python.there is a lot of pythonpks. how to properly install python?
<mg>
Is there a way to configure what nix-shell does to $PS1?
<moesasji>
LnL: would the cherry-pick happen with the commit that gives trouble and then do git rebase master?
<simpson>
linarcx_: Usually you won't want Python installed; you'll want either pythonPackages.virtualenv or a nix-shell invocation like $(nix-shell -p pythonPackages.{treq,twisted})
fre has joined #nixos
<LnL>
the reset drops all of the changes you want to rebase, so you need to cherry pick whatever you where working on
<LnL>
(the commits that are not part of upstream master)
takle has joined #nixos
<spacefrogg>
mg: If you set PS1 inside the derivation, it should not get modified by nix-shell.
<moesasji>
My patch is in a separate branch on 17.03. Just trying to get that to rebase on master (and lost how to do that)
Tucky has joined #nixos
<avn>
moesasji: cherry-pick can be better approach, branches diverged too much to smooth rebase
<avn>
Altough my WIP branch have commit which older than year
<moesasji>
So start on base and just cherry-pick the commits that are in my working branch?
<grantwu>
moesasji: What was the exact error?
<grantwu>
Is "could not apply" literally all it said
<moesasji>
Error: could not apply 69d75ce207... release-notes: add new services for 17.03 + misc additions
<LnL>
oh you're trying to rebase 17.03 onto master, that's going to be a pain
<grantwu>
do you get merge conflicts you can manually resolve?
<avn>
moesasji: git rebase --skip for it ;)
<grantwu>
nvm, it seems like other people know what to do
takle has quit [(Ping timeout: 276 seconds)]
Tucky has quit [(Remote host closed the connection)]
<LnL>
it's probably a lot easier just to cherry-pick your own commits onto master instead of doing a full rebase
<moesasji>
LnL: indeed as that is what the manual says to do when preparing patches.
LinArcX has quit [(Remote host closed the connection)]
<symphorien>
I usually do git rebase --onto master mybranch~ mybranch
<symphorien>
it just copies the relevant commits
Tucky has joined #nixos
takle has joined #nixos
<moesasji>
The git rebase --onto master mybranch indeed appears to complete. Thanks symphorien, avn, grantwu, LnL!
<LnL>
ah yes I think that does something similar to what I was suggesting but automatically :)
<cocreature>
ah I think following the symlinks works
<cocreature>
that leads me to /nix/store/p845zp2mv9a6q81mczplr3n0nqcvgipz-nixpkgs-17.09pre111388.00512470ec/nixpkgs and 00512470ec should be the commit
<goibhniu>
cocreature: right ... ~/.nix-defexpr/channels/<yourchannelname>
<avn>
adisbladis: nope, is other problem, I already have this commit
<globin>
adisbladis: that is in master already
<sellout->
Is anyone using Scion for Haskell? The version in haskellPackages seems to be pretty old and isn’t compiling for me (on darwin) because my version of base, etc. is too new.
Myrl-saki has quit [(Ping timeout: 240 seconds)]
Infinisil has joined #nixos
faffolter has quit [(Remote host closed the connection)]
<adisbladis>
Huh. I didn't know that.. I'm running a working kde setup from master so I musted have done something to fix qt apps..
<avn>
globin: also I feel you can push fix right to master, because it broken anyway ;)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] moesasji opened pull request #28004: xfce4-session: fix lock screen not working with light-locker (master...master) https://git.io/v7oQP
NixOS_GitHub has left #nixos []
wigust has quit [(Remote host closed the connection)]
mpcsh has quit [(Quit: THE NUMERICONS! THEY'RE ATTACKING!)]
ison111 has quit [(Ping timeout: 240 seconds)]
mpcsh has joined #nixos
<globin>
Sonarpulse: o.O kidletime is another example, but the build doesn't fail with NIX_DEBUG=1...
Tucky has quit [(Remote host closed the connection)]
adev has joined #nixos
seanparsons has quit [(Read error: Connection reset by peer)]
<qknight>
nixos submodules can get arguments magically, like: { config, ... }: which is not the global config but holds the attributes of the submodule as enable, and whaterver one has
<qknight>
how can i call a submodule with one additional parameter, a string?
<qknight>
when i try that here, it will screw up the submodule arguments and i have to pass in all of them, like the config, which is stupid as i don't know how to do that
seanparsons has joined #nixos
ison111 has joined #nixos
<sphalerite>
qknight: can't you just make the additional parameter a config option?
<sphalerite>
qknight: alternatively, have the module look like yourExtraParam: {config, pkgs, ...}: {...} and put (import path/to/module.nix extraParamValue) in the imports
spear2 has joined #nixos
mudri has joined #nixos
Tucky has joined #nixos
zraexy has joined #nixos
pie_ has quit [(Ping timeout: 240 seconds)]
gnuhurd has joined #nixos
Ivanych has quit [(Ping timeout: 260 seconds)]
simukis has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nonsequitur opened pull request #28006: ruby docs: fix text and example (master...patch-2) https://git.io/v7oAi
NixOS_GitHub has left #nixos []
<qknight>
sphalerite: will think about that, thanks!
saltsa has quit [(Ping timeout: 255 seconds)]
saltsa has joined #nixos
spear2 has quit [(Quit: ChatZilla 0.9.93 [Firefox 54.0/20170616104331])]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nonsequitur closed pull request #28006: ruby docs: fix text and example (master...patch-2) https://git.io/v7oAi
NixOS_GitHub has left #nixos []
hotfuzz has quit [(Quit: WeeChat 1.7.1)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nonsequitur reopened pull request #28006: ruby docs: fix text and example (master...patch-2) https://git.io/v7oAi
NixOS_GitHub has left #nixos []
cpennington has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 2 new commits to master: https://git.io/v7opP
<cocreature>
Hey, ghcjs is currently broken in unstable because haddock-api and haddock 2.18 don’t build with GHC 8.0. I know it’s fixed in master but to improve my nix foo I thought I could try to figure out how to overwrite just haddock and haddock-api. I came up with http://lpaste.net/357490 where haddock-api.nix and haddock.nix contain the derivations for the 2.17 releases but when I run "nix-build" it’s still
<cocreature>
trying to build haddock-api 2.18. what am I missing here?
adeohluwa has joined #nixos
<adeohluwa>
wondering if nix os supports sublime text & yarn package manager
cement has joined #nixos
erictapen has quit [(Ping timeout: 240 seconds)]
<MoreTea>
adeohluwa, for yarn, take a look at my yarn2nix project
<nwuensche>
Hi everybody! I'd like to install NixOS in a virtual machine. However, when I nixos-install, I always get the error message: "curl: (6) Couldn't resolve host 'cache.nixos.org'. What can I do?
erasmas has joined #nixos
<srhb>
nwuensche: can you resolve or ping anything in the vm? does it have internet at all?
<nwuensche>
Jup, I pinged 8.8.8.
<nwuensche>
Jup, I pinged 8.8.8.8
jbaum98 has joined #nixos
<srhb>
nwuensche: And how about lookups?
<srhb>
ie. host google.com
<srhb>
eg*
<clever>
nwuensche: and what is in /etc/resolv.conf
<clever>
nwuensche: and what command fails with an error?
sauyon has joined #nixos
Tucky has quit [(Remote host closed the connection)]
<nwuensche>
build of '/nix/store/4b...-nixos-system-17.03.1645....drv' failed
sauyon has quit [(Client Quit)]
Myrl-saki has quit [(Ping timeout: 255 seconds)]
cmcdragonkai1 has joined #nixos
sauyon has joined #nixos
cmcdragonkai has quit [(Ping timeout: 268 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v7KgJ
<NixOS_GitHub>
nixpkgs/master 4825e48 Tim Steinbach: linux: 4.9.40 -> 4.9.41
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/v7KgI
<NixOS_GitHub>
nixpkgs/release-17.03 92dfc19 Tim Steinbach: linux: 4.9.40 -> 4.9.41...
NixOS_GitHub has left #nixos []
joachifm has quit [(Ping timeout: 240 seconds)]
jonte_ has quit [(Ping timeout: 240 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v7Kg6
<NixOS_GitHub>
nixpkgs/master ea2a10e Tim Steinbach: linux: 4.4.79 -> 4.4.80
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to release-17.03: https://git.io/v7KgP
<NixOS_GitHub>
nixpkgs/release-17.03 4dc413a Tim Steinbach: linux: 4.4.79 -> 4.4.80...
NixOS_GitHub has left #nixos []
vinymeuh has joined #nixos
eschnett has quit [(Quit: eschnett)]
alx741 has joined #nixos
alx741 has quit [(Client Quit)]
Ivanych has joined #nixos
alx741 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v7K2O
<NixOS_GitHub>
nixpkgs/master 06af1df Tim Steinbach: linux: 4.13-rc3 -> 4.13-rc4
NixOS_GitHub has left #nixos []
ison111_ has joined #nixos
pie_ has joined #nixos
ison111_ has left #nixos []
<nwuensche>
Nobody knows what I can do?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NeQuissimus pushed 1 new commit to master: https://git.io/v7K25
<NixOS_GitHub>
nixpkgs/master 4edcfa3 Tim Steinbach: gradle: 4.0.2 -> 4.1
NixOS_GitHub has left #nixos []
tmaekawa has joined #nixos
<srhb>
nwuensche: It sounds to me like you're having an intermittent networking problem. First, you couldn't resolve cache.nixos.org, but as I understand it, now you can.
<srhb>
That's hard to help with.. :/
<nwuensche>
I can ping it, but curl still doesn't know what to do...
<srhb>
nwuensche: Even if you manually curl something?
<clever>
nwuensche: this is what fdisk should say about it when you run the i command on it
mkoenig has quit [(Ping timeout: 260 seconds)]
<nwuensche>
Sorry for the newbie question, but how can I decide if my system can use UEFI?
<clever>
look in the bios config to see if it mentions uefi or efi or secure boot
mkoenig has joined #nixos
<nwuensche>
Ok, so I will reboot my PC and find it out. Thank you so much so far!
nwuensche has quit [(Quit: Leaving)]
<nliadm>
hm, makeWrapper --prefix doesn't work with LUA_PATH, I'm assuming because it doesn't quote the '?'s in the string
jgertm has joined #nixos
nwuensche has joined #nixos
<nwuensche>
Hello. So the BIOS says: "UEFI/Legacy Boot - Both" + "UEFI/L Boot Priority - L first" Do I have to change this to make UEFI work or not?
<clever>
nwuensche: simplest thing is to just make the bios boot partition and boot via legacy
tmaekawa has quit [(Quit: tmaekawa)]
<nwuensche>
Ok. I have a 500 MiB partition that I mounted into /mnt/boot. Can I unmount that and use this one?
<clever>
umount it, delete it, then create 2, a bios boot (1mb) and a replacement /boot (499mb)
<moesasji>
I'm trying to get xfce user switching to work; on other linux systems I do this by creating /usr/bin/gdmflexiserver calling light-locker-command -l.
<clever>
and then mount the 499mb to /mnt/boot after formatting again
<clever>
nwuensche: and set the type right on the 1mb
<clever>
nwuensche: then nixos-generate-config --root /mnt and nixos-install
<clever>
it will reuse the existing rootfs, as long as you mount it to /mnt/
<moesasji>
NixOS doesn't have /usr/bin....any pointers how to create a bash-script that gets called ?
ison111 has quit [(Ping timeout: 240 seconds)]
<clever>
moesasji: #!/usr/bin/env bash
<moesasji>
that is for a bash script....I need to install a script in /usr/bin or equivalent that gets called. That is the bit I can't figure out.
<clever>
moesasji: this creates 3 bash scripts, and then i can just do something like environment.systemPackages = [ (pkgs.callPackage ./util.nix {}) ];
<dtzWill>
blargh I probably can't repro but just b/c I need some sanity: if a store path (derivation output produced with nix-build) contains 1)trivial textual references to store other store paths or especially 2)symlinks to said store paths then ... "nix-store -qR /nix/store/my-output" should definitely include things mentioned via 1 and 2... right??
<clever>
dtzWill: yes
<moesasji>
Thanks! Only puzzle that remains is how to let xfce-session find the script I create that way. This system really takes some getting used to.
<dtzWill>
usually just mentioning in a log or something is enough, which is "annoying" for various jobs I have that analyze paths (without doing something like nuke-references)
<clever>
dtzWill: but only if you got those paths via a derivation
<dtzWill>
clever: thank you
<clever>
dtzWill: if you hard-code a storepath as a string, it wont
<dtzWill>
clever: lol oh >.<
<clever>
dtzWill: every string in nix has some magic context on it, the dependencies of the string
<dtzWill>
clever: what if I manage to sneak a store path (let's say I put it in an encrypted zip for sake of discussion) through an intermediate derivation
<clever>
dtzWill: the builtins.derivation function creates a string (the output path) that depends on the derivation
<dtzWill>
clever: and then extract them all
<clever>
dtzWill: and if any string with context is pased to derivation, they become build-time deps
* dtzWill
looks and reads clever's suggestion carefully
<clever>
and only the build-time deps are looked for in the output
<dtzWill>
okay
<dtzWill>
you don't say!!
<dtzWill>
okay
<clever>
also, you should really build with sandboxing on
<dtzWill>
well then that's enough to go by, I think I can go from here. Guess I just never managed to produce store paths during a derivation without them being (transitively) build inputs somehow haha
<dtzWill>
I definitely do! haha
<clever>
that will enforce you to only access things that are properly defined as build inputs
<clever>
so if you cheat, the files just wont exist
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] wmertens pushed 2 new commits to master: https://git.io/v7KKt
<dtzWill>
I managed to obfuscate the store paths (and don't put them in nix-support/something-or-other), and then later derivations that analyze these paths are magically producing paths that the sandbox is like NOPE xD
<dtzWill>
so I created intermediate derivation that tries to tell Nix about the paths
<clever>
ah
<dtzWill>
hmm interesting. I suppose this means I'm doing something the wrong way, thank you very much
<clever>
yeah
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] wmertens pushed 2 new commits to master: https://git.io/v7KKz
<clever>
builtins.unsafeDiscardStringContext is also a fun function
<clever>
that strips all context from a string
<clever>
so i can make a config file that does "foo=${builtins.unsafeDiscardStringContext bar}", and that config wont depend on bar
<clever>
but it will have bar's absolute path
<dtzWill>
it's actually impressive I went so long (I rely heavily on references for various things) under the false impression Nix would always know if I managed to reference a path
<dtzWill>
ahh okay
<clever>
the last time i used that, i squashfs'd an entire nixos up, then jammed that into an initrd
<dtzWill>
xD
<clever>
but i needed to refer to a path that will exist inside the squashfs, on the kernel cmdline
<clever>
and nix wants to haul along the entire uncompressed copy of nixos
<clever>
dtzWill: there is also a related problem that cant easily be fixed
<nwuensche>
clever: Is the right Label ef02 BIOS boot partion?
<clever>
nwuensche: i think so
<clever>
nwuensche: double-check what the i command says about it
<nwuensche>
I installed it! Thank you all so much!
<srhb>
mpcsh: More relevantly, can you tell when it will update in the future? No, it will update when the tests pass.
<srhb>
nwuensche: Yay! :)
<clever>
srhb: right now, sddm is broken on master
<srhb>
Yeah, I noticed.
<srhb>
QT stuff?
<clever>
yeah
adeohluwa has quit [(Quit: Connection closed for inactivity)]
erictapen has joined #nixos
ambro718 has joined #nixos
goibhniu has quit [(Ping timeout: 240 seconds)]
MoreTea has quit [(Ping timeout: 246 seconds)]
ThatDocsLady has quit [(Ping timeout: 255 seconds)]
rauno has joined #nixos
tvon has joined #nixos
<rauno>
hey all, is there an alternative to python-parse in nix packages ?
<rauno>
coudln't find anything atm
<rauno>
can't install with pip inside nix-shell either
zraexy has joined #nixos
gnuhurd has quit [(Remote host closed the connection)]
_rvl has joined #nixos
<srhb>
rauno: Is it the package called parse on pypi?
<rauno>
ye
zarel has joined #nixos
<kragniz>
any nixos people at sha2017?
<srhb>
rauno: `pypi2nix -V 3.5 -e parse` generates a nix shell that appears to work
<srhb>
rauno: (That is, I can do from parse import * and the small example works)
<rauno>
oh
<simpson>
kragniz: You are not the first to ask, but I cannot remember the nicks of the others.
<rauno>
this pypi2nix can work with requirements.txt also ?
<mpcsh>
srhb: srhb is there a way to see test progress?
<srhb>
rauno: That's the default behaviour
<srhb>
mpcsh: Click the link :)
<rauno>
nice, i'll check it our right away :)
<srhb>
mpcsh: Leftmost column is the most recent run
kiloreux has quit [(Ping timeout: 255 seconds)]
<kragniz>
simpson: /me looks through logs
<mpcsh>
srhb: wait, which link?
<srhb>
mpcsh: On howoldis.herokuapp.com, the rightmost column "Hydra job for tests" - click the one corresponding to nixos-unstable
<srhb>
mpcsh: The leftmost column there (the most recent build of the test set) has to be all checkmarks for nixos-unstable to advance to master..
ThatDocsLady has joined #nixos
<mpcsh>
srhb: is that top axis git commits?
<mpcsh>
srhb: ah nvm, they're hydra build ids
<srhb>
mpcsh: In the hydra page? No, they are build numbers.
<srhb>
Yeah :)
<srhb>
You can click them and get details or reproduce a specific test by clicking the icons next to the specific test.
<srhb>
(Most if not all of them right now is: error while loading shared libraries: libQt5Xml.so.5: cannot open shared object file: No such file or directory)
<mpcsh>
srhb: ok yeah, this is the page I needed, thank you!
<srhb>
Oh, no, there's also a few others
<srhb>
mpcsh: Cool! :)
erictapen has quit [(Ping timeout: 260 seconds)]
thblt has quit [(Remote host closed the connection)]
ericsagnes has quit [(Ping timeout: 240 seconds)]
_rvl has quit [(Ping timeout: 240 seconds)]
kiloreux has joined #nixos
detran has joined #nixos
gnuhurd has joined #nixos
<copumpkin>
zimbatm: are you looking into the terraform 0.10 stuff? I might soon
frankpf has quit [(Ping timeout: 240 seconds)]
mbrgm has quit [(Ping timeout: 255 seconds)]
ericsagnes has joined #nixos
mbrgm has joined #nixos
mbrgm has quit [(Client Quit)]
<zimbatm>
copumpkin: not actively
takle has joined #nixos
<copumpkin>
ok :)
<copumpkin>
just didn't want to duplicate work if you were on it
<nwuensche>
So I have another question. I used the nixos .iso without the graphical stuff, but now I want to use i3(wm). Is it possible to install X11 with a package? I couldn't find one
arximboldi has quit [(Quit: arximboldi)]
<srhb>
nwuensche: In the installer or on your installed system?
<nwuensche>
On my installed system
<srhb>
nwuensche: In general you'll want to enable X via configuration.nix
<srhb>
nwuensche: And then nixos-rebuild switch
<srhb>
services.xserver.enable = true; will enable the xserver, but perhaps you also want a display manager and a de.
<nwuensche>
I had a problem doing this, but now I found a typo in my config... Thanks nevertheless
takle has quit [(Remote host closed the connection)]
takle has joined #nixos
<cocreature>
I’ve asked this a few hours ago but hadn’t received a response so I’ll try again. I hope that’s ok.
<cocreature>
Hey, ghcjs is currently broken in unstable because haddock-api and haddock 2.18 don’t build with GHC 8.0. I know it’s fixed in master but to improve my nix foo I thought I could try to figure out how to overwrite just haddock and haddock-api. I came up with http://lpaste.net/357490 where haddock-api.nix and haddock.nix contain the derivations for the 2.17 releases but when I run "nix-build" it’s still
<cocreature>
trying to build haddock-api 2.18. what am I missing here?
jensens has quit [(Ping timeout: 258 seconds)]
_rvl has joined #nixos
vinymeuh has joined #nixos
takle has quit [(Ping timeout: 255 seconds)]
<sphalerite>
Anyone know of a decent way to rsync between two machines neither of which has a public IP address?
<sphalerite>
somehow relaying via a third machine which does is an option, but I do want the connection to be secure
arximboldi has joined #nixos
<grantwu>
IPv6 tunnels?
<grantwu>
Wait, no that doesn't work
<rauno>
another question about nix-shell, wondering why i can't connect to qemu from there ?
<thblt>
both rsync in turn to third machine?
ThatDocsLady has quit [(Ping timeout: 255 seconds)]
<zarel>
sphalerite: I used tinc + a discovery node (which is part of the VPN) and it managed to direct connect tho machines behind NAT
athan has quit [(Remote host closed the connection)]
<srhb>
cocreature: iirc you'll want packages = super.haskell.packages // { ... as well
<srhb>
cocreature: And if that's your user config, you might want to try using an overlay
<srhb>
Oh, that's a local package config. Nevermind then :)
gnuhurd has quit [(Remote host closed the connection)]
gnuhurd has joined #nixos
<srhb>
cocreature: I'm actually a little unsure what's going on with the inherit there. Why not simply create myHaskell = super.haskell // { etc ...} and then use callPackage from there?
<srhb>
Or pass in the overridden derivations explicitly in callPackage
tommd has joined #nixos
simukis has quit [(Quit: simukis)]
<cocreature>
srhb: I just copied that from somewhere, I’m new to nix :)
simukis has joined #nixos
<srhb>
cocreature: Oh, ok!
<srhb>
cocreature: Let me see if I have a small example lying around
<thblt>
Apologies in advance, for I'm probably missing something obvious, but. I'm trying to change the default console font to Terminus28, so I'm adding terminus_font to environment.systemPackages and i18n.consoleFont = pkgs.lib.mkForce "Lat2-Terminus28"; nixos-rebuild switch fails because systemd-vconsole-setup.service fails to start.
<thblt>
Full code is here: https://pastebin.com/FwdgE8RZ This is /etc/nixos/class/hidpi.nix, which is imported from configuration.nix
<sphalerite>
zarel: that looks like a neat option, do you know anything about configuring tinc on nixos? It seems to be pretty simple but I don't understand how to set up the hosts part
<srhb>
cocreature: The top comment is funny :-P
<cocreature>
:)
<srhb>
cocreature: Anyway, when you have a callPackage like that, you can just override the arguments to it
<cocreature>
srhb: at least the overrides in packages do take effect. if I just replace ghcjsHEAD by some garbage, it actually fails
<cocreature>
but apparently I’m not overriding all references to haddock and haddock-api
_Vi has quit [(Ping timeout: 258 seconds)]
<srhb>
cocreature: Oh, really. Let me reread.
takle has joined #nixos
<cocreature>
the "compiler" stuff I came up with myself so I suspect that that’s broken :)
<cocreature>
although I guess the nix config could count as minimal
<srhb>
cocreature: haddock-api.cabal: This package requires at least Cabal version 2.0
<srhb>
?
<cocreature>
srhb: exactly
<srhb>
cocreature: OK
nwuensche has quit [(Quit: Leaving)]
eacameron has quit [(Quit: Leaving...)]
joachifm has joined #nixos
joachifm has quit [(Client Quit)]
<cocreature>
as I mentioned before it’s fixed in master. at this point I’m mostly trying to improve my nix-foo by figuring out how I could overwrite haddock to get this working again
joachifm has joined #nixos
<srhb>
cocreature: Yeah :)
<cocreature>
maybe I should have chosen something with shorter compile times to make it easier to experiment :P
leat has quit [(Quit: leat)]
gnuhurd has quit [(Remote host closed the connection)]
gnuhurd has joined #nixos
goibhniu has joined #nixos
leat has joined #nixos
ambro718 has joined #nixos
periklis has joined #nixos
vinymeuh has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
gnuhurd has quit [(Remote host closed the connection)]
gnuhurd has joined #nixos
gnuhurd has quit [(Remote host closed the connection)]
gnuhurd has joined #nixos
ertes has joined #nixos
m0rphism has quit [(Quit: WeeChat 1.9)]
<srhb>
cocreature: I didn't manage to clean it up much, really, but perhaps you can use some ideas..
<srhb>
Oops, I named packages.ghc802 inconsistently, it should have been myghc802
<srhb>
cocreature: Anyway, it's not really an improvement as much as it is a style change. q_q
Fare has joined #nixos
<Fare>
hi!
<srhb>
Fare: Hello.
<cocreature>
srhb: thanks so much your help! as I said before I’m new to nix so style changes are very helpful :)
<srhb>
cocreature: I hope you glean something from it at least. I learned something as well. :)
<srhb>
(bootPkgs)
<Fare>
I'm trying to setup a nix build server, and I'm not sure how to configure client and server to both use the same non-channel fork of nixpkgs
<srhb>
Fare: nix.nixPath ?
<srhb>
It can get a bit finicky since if you use a fetchfromGitHub there, it will only actually affect the nixpkgs used *after* a rebuild
<Fare>
srhb: I have no idea. Where would I set that up? Is it a matter of environment variables? How do I tell, e.g. the server (over ssh) to use (or initialize) the correct NIX_PATH (or whatever)
<srhb>
It's not really a clean solution imo, perhaps someone else has a better idea.
<Fare>
oh, so first use nixos-rebuild switch to build the server, with the server pointing to the right place?
<Fare>
then the server uses the configured nix.nixPath
<srhb>
Yes.
<Fare>
thanks
<srhb>
The ugly thing is that it now takes two rebuilds: one to switch the revision, another to build the system resulting from the altered revision
<srhb>
ymv
<srhb>
ymmv*
erictapen has quit [(Ping timeout: 240 seconds)]
<Fare>
meanwhile, when I install the server itself, it will install from its own channels.
<srhb>
Well, purity means that doesn't matter. :-P
<Fare>
usability means it's annoying, but oh well, there is worse.
ison111_ has joined #nixos
<srhb>
Yes... Manually keeping a checkout in sync between servers might be more annoying though
<srhb>
I know, network mount nixpkgs! What could possibly go wrong. :-)
<Fare>
:-(
tvon has quit [(Quit: Peace out, y'all!)]
erictapen has joined #nixos
<srhb>
Fare: Hang on, doesn't the build server actually get nixpkgs from the client?
<srhb>
I mean, for the purpose of being a build server.
Fare has quit [(Ping timeout: 260 seconds)]
MP2E has joined #nixos
Nobabs27 has joined #nixos
tnks has joined #nixos
<tnks>
I'm reading the code for makeOverridable, and as expected simultaneously vommitting and getting a headache.
_Vi has joined #nixos
<tnks>
I just forget some of the nuances.
<tnks>
But I was trying to figure out if there's really a difference between .override and .overrideAttrs.
<tnks>
does someone know the answer to that? Otherwise, I'll puzzle over this snippet more.
<LnL>
lol, it's not that bad is it?
<ndash_>
tnks: yes, .override replaces the args to a function returning a derivation
<ndash_>
.overrideAttrs replaces attributes in the result itself
<tnks>
ndash_: ah, that makes more sense now.
<tnks>
ndash_: thanks.
<LnL>
well it replaces the arguments passed to stdenv.mkDerivation I think
<srhb>
LnL: "it" ?
<LnL>
overrideAttrs
Fare has joined #nixos
<srhb>
LnL: Thanks :)
<tnks>
ndash_: however, this is what the docs say, "The function overrideAttrs allows overriding the attribute set passed to a stdenv.mkDerivation call"
<tnks>
ndash_: that seems to conflict what you just said, which is why I ended up in the source.
<LnL>
override is more generic, it allows overriding arguments of a function that returns an attrset of another function
<MP2E>
hmm, I pulled the latest nixpkgs master, but it seems the recent staging merge containing binutils 2.29 broke some things. When building a wine wow64 derivation, it fails on linking, it's seemingly unable to find the 32-bit libdl.so.2
<MP2E>
just a heads up
erictapen has quit [(Ping timeout: 240 seconds)]
<tnks>
LnL: override now has an "isFunction" check and seems to do both.
gnuhurd has quit [(Remote host closed the connection)]
<tnks>
I suppose I should look at the history of this function in Git too.
gnuhurd has joined #nixos
tokudan has quit [(Quit: Leaving)]
<cocreature>
srhb: after waiting way to long for ghcjs to build, I can confirm that it actually worked \o/ :)
<srhb>
cocreature: Hooray! I cannot confirm that mine does. ;-)
<cocreature>
srhb: it does work too and it’s even close enough to mine that it doesn’t rebuild everything again :)
<srhb>
Oh, magical!
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] MP2E pushed 1 new commit to master: https://git.io/v76JI
<tnks>
I think I might understand more now. I was thinking of it wrong. That's why I'm revisiting this in the first place... it's easy to forget the details.
ison111_ has left #nixos ["WeeChat 1.7.1"]
<tnks>
I lost it... and I can't tell the difference between .override and .overrideAttrs again.
silver_hook has joined #nixos
silver_hook has quit [(Changing host)]
silver_hook has joined #nixos
Nobabs27 has quit [(Quit: Leaving)]
<LnL>
they are kind of similar
Infinisil has joined #nixos
<sphalerite>
tnks: override is provided by callPackage and lets you replace the arguments to a package function
<sphalerite>
tnks: overrideAttrs is provided by mkDerivation and lets you override what's given to mkDerivation
Fare has quit [(Ping timeout: 260 seconds)]
<tnks>
sphalerite: got it.
<tnks>
that finally occurred to me.
<tnks>
it took a while to realize that "override" would get overridden by trying to make nested functions both overrideable.
dejanr has joined #nixos
<sphalerite>
well you can combine override and overrideDerivation
<tnks>
I thought that overrideDerivation is discouraged.
dejanr has quit [(Client Quit)]
<tnks>
and I thought you can combine override with overrideAttr also.
<sphalerite>
err I meant overrideAttrs
<tnks>
right.
<srhb>
Now I'm confused. Is callPackage a special case of override, but works on attributes instead of source files?
dejanr has joined #nixos
<srhb>
er, but override works on attributes instead of source files
DutchWolfie has quit [(Quit: Konversation terminated!)]
<tnks>
actually, I get a headache thinking about the doubly-nested .override calls.
dejanr has quit [(Client Quit)]
colabeer has joined #nixos
deltasquared has joined #nixos
<LnL>
I'm not sure if you can chain overrideAttrs
<sphalerite>
srhb: callPackage takes a function (or a path which it imports to get the function). To that function, it passes the attributes in nixpkgs that the function wants. It adds the attribute `override` to the result of evaluating that, and that attribute is a function that takes a set which will replace parts of what's passed into the package function
<sphalerite>
LnL: I believe you can. `(bash.overrideAttrs (o: {})).overrideAttrs (o: {})` evaluates just fine for me
gnuhurd has quit [(Remote host closed the connection)]
gnuhurd has joined #nixos
<tnks>
maybe there's just many ways to skin this cat, and I should stop worrying about it.
<LnL>
yes that works but if you actually override something you might loose the previous override
ison111 has quit [(Ping timeout: 260 seconds)]
<LnL>
(that's why makeOverridable looks so complicated)
<sphalerite>
well, yes. That's the nature of a destructive override function
<sphalerite>
but if they only add to super or touch different parts, then yes it will work
<LnL>
hmm, looks ok
<srhb>
sphalerite: Thank you :)
lucasOfBesaid has joined #nixos
ericsagnes has joined #nixos
<LnL>
oh yes, it returns the new mkDerivation call
zeus__ has quit [(Read error: Connection reset by peer)]
m0rphism has joined #nixos
zeus__ has joined #nixos
zarel has quit [(Quit: Leaving)]
<lucasOfBesaid>
Hi, all! I'm trying to build my package with a cross compiler, and I can't seem to get there. Building uclibc fails at the linking stage when it can't find some symbols. I don't actually need uclibc, but I can't seem to make a derivation that doesn't depend on it.
<sphalerite>
lucasOfBesaid: let me guess, ARM, and it's something to do with EABI/OABI?
<sphalerite>
lucasOfBesaid: the uclibc in nixpkgs is ancient and nixpkgs's cross-compilation support is still a bit flaky afaik.
<lucasOfBesaid>
sphalerite: No. i686, and the error is "undefined reference to `_dl_loaded_modules'"
<sphalerite>
oh ok
<sphalerite>
You might be able to make a derivation using the native mkDerivation with the foreign gccCrossStageStatic, if you don't need a libc
<sphalerite>
But if it's i686 then I think you should be using i686Packages
<colabeer>
Hi there. When I start wine with some app. It complains with a libGL error "unbale to load driver: i965_dri.so" I've managed to get rid of it by using a custom LD_LIBRARY_PATH. But this does not seem like good style in nixos. What is the best way to overcome that?
viric has joined #nixos
<viric>
hello
<sphalerite>
lucasOfBesaid: sorry, not i686Packages. pkgsi686Linux
<lucasOfBesaid>
sphalerite: What's i686Packages?
<viric>
what was the nix-shell thing to compile with -O0?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 2 new commits to master: https://git.io/v76ka
<sphalerite>
colabeer: get back the standard `LD_LIBRARY_PATH`, and customise it by adding to it rather than overwriting it.
<LnL>
hardeningDisable = [ "fortify" ];
<LnL>
viric: ^
<viric>
LnL: ah right. Where?
<LnL>
the derivation, or export it
<viric>
ok
<sphalerite>
colabeer: the specific entry in question would be /run/opengl-driver/lib
m0rphism has quit [(Quit: WeeChat 1.9)]
<viric>
LnL: hm it fails to work here...
<sphalerite>
lucasOfBesaid: e.g. nix-build '<nixpkgs>' -A pkgsi686Linux.hello will get you a 32-bit GNU Hello
<LnL>
viric: you still have -O2 in NIX_CFLAGS_COMPILE if you restart your shell?
<viric>
LnL: ah, export it BEFORE nix-shell'
<viric>
?
<viric>
LnL: I thought it was read at gcc call time. Now it works. Thank you!
<lucasOfBesaid>
sphalerite: pkgsi686Linux looks like a way to get multilib packages for Linux. That's not what I need. I'm trying to work on an OS kernel, and I need a cross-compiler that targets i686-elf
<LnL>
viric: might have changed recently, not sure
m0rphism has joined #nixos
<viric>
LnL: it's correct: export before call to nix-shell. NIX_DEBUG=1 shows it right
<sphalerite>
lucasOfBesaid: pkgsi686Linux.gcc should do that
<clever>
lucasOfBesaid: what about just directly using pkgsi686Linux.stdenv.mkDerivation ?
<clever>
lucasOfBesaid: then everything will be in 32bit mode
<clever>
lucasOfBesaid: you can also do import <nixpkgs> { system = "i686-linux"; }, then it wont even have a 64bit option
<clever>
it will just default to 32bit
<colabeer>
sphalerite: actually LD_LIBRARY_PATH is set to /run/opengldriver/lib
<avn>
btw folks, any reasons, why build under nix can be stalled?
<clever>
lucasOfBesaid: i686-linux is an internal code for nixos, that just says everything must be 32bit
bkchr has joined #nixos
<sphalerite>
lucasOfBesaid: most modern linuxes (including the one in nixos) use ELF. That compiler will generate ELF executables, all you need for OS dev should be to make sure it's not linking against libc dynamically and stuff
<sphalerite>
I think.
<viric>
LnL: thank you, btw!
cement has quit [(Ping timeout: 260 seconds)]
<lucasOfBesaid>
sphalerite: Unless you compile GCC with --target='i686-elf' instead of --target='i686-linux', it will make all kinds of assumptions about the OS it's running on, which in the case of an OS kernel will be wrong.
<LnL>
viric: np, I fixed a package with something like that very recently :)
<lucasOfBesaid>
clever: So, the question becomes, is pkgsi686Linux.gcc compiled with --target='i686-elf'?
Filystyn has joined #nixos
<sphalerite>
I don't see what that would be, all that would affect actual compilation/assembly (as opposed to linking) that I know of is the calling convention
<sphalerite>
but I'm not an expert.
<lucasOfBesaid>
sphalerite: That's why I threw that link there. It leads to a page written by experts. Anyway, my question stands. And if pkgsi686Linux.gcc is NOT compiled with --target='i686-elf', what's the most sensible way I can get one that is? I'm happy to assist in improving the state of Nix's cross-compilation machinery if necessary.
<lucasOfBesaid>
Ah. It seems, clever, that pkgsi686Linux.gcc is not an appropriate cross-compiler for my purposes. After nix-shell -p pkgsi686Linux.gcc and gcc -v, I see "Target: i686-pc-linux-gnu". Unfortunate.
ison111 has joined #nixos
<lucasOfBesaid>
Is there some way I can leverage the existing GCC derivation? Some combination of arguments I can give it, or some configuration I can provide that would prevent it from needing unnecessary dependencies and give it the right TARGET?
<clever>
lucasOfBesaid: line 1 sets the target arch that gets passed to gcc
Filystyn has quit [(Read error: Connection reset by peer)]
Filystyn has joined #nixos
Filystyn has quit [(Changing host)]
Filystyn has joined #nixos
mkoenig has quit [(Remote host closed the connection)]
<clever>
you can also nix-build arm-baremetal.nix --argstr arch i686-elf
<clever>
so you just need to find a string that makes gcc happy
<lucasOfBesaid>
clever: I tried this. I get the error "error: Unknown libc"
<clever>
ah, things may have changed in the newest master
<clever>
try it against a checkout of 16.09?
<lucasOfBesaid>
Oh dear. I've been using channels. And my skills with nix are still kinda nascent. Are there some instructions I can follow somewhere to walk through that?
<clever>
then in another dir, nix-build arm-baremetal.nix -I nixpkgs=/home/clever/nixpkgs-channels
<sphalerite>
clever: why 16.09 and not 17.03?
pie_ has quit [(Read error: Connection reset by peer)]
<clever>
going to a fairly old one to avoid stdenv improvements
<clever>
which arent compatible with the example
pie_ has joined #nixos
kiloreux has quit [(Ping timeout: 240 seconds)]
<sphalerite>
ah right
<lucasOfBesaid>
Cool. So far, this is definitely doing something. Let's see how far it gets.
<lucasOfBesaid>
Ah. I love having 24 cores. ^_^
pie_ has quit [(Remote host closed the connection)]
pie_ has joined #nixos
<deltasquared>
lucasOfBesaid: I missed context, what CPU is this
<deltasquared>
it can't be threadripper yet... right!?
deltasquared has left #nixos ["damnit, blackout time"]
<lucasOfBesaid>
model name: Intel(R) Xeon(R) CPU E5-2687W v4 @ 3.00GHz
<lucasOfBesaid>
Whoops, missed them.
<lucasOfBesaid>
clever: That appears to have worked. Thanks! And it doesn't seem to have pulled any useless dependencies in either. Why does this not work with the current version of Nixpkgs?
<clever>
lucasOfBesaid: something to do with the cross-compiler and libc was changed
<clever>
lucasOfBesaid: so that cheaty trick of setting libc to null no longer works
colabeer has quit [(Remote host closed the connection)]
dejanr has quit [()]
<lucasOfBesaid>
clever: I really appreciate your help. Of course I'd like to be able to write a derivation that will work with current and future versions of Nixpkgs. Is there someone who regularly works on the cross compiler that I should bring this issue up with? Should I file an issue on Github?
goibhniu has quit [(Ping timeout: 255 seconds)]
elurin has joined #nixos
<clever>
lucasOfBesaid: id have to re-read the code in all-packages.nix and see how it changed
<clever>
lucasOfBesaid: Sonarpulse may know more as well
<Sonarpulse>
hmm?
<Sonarpulse>
uh oh
<Sonarpulse>
more breakage???!?!
<lucasOfBesaid>
Sonarpulse: Pleasure to meet you.
colabeer has joined #nixos
<Sonarpulse>
lucasOfBesaid: what pain have I caused you :)
<Sonarpulse>
(and pleasure, likewise :))
arximboldi has joined #nixos
<lucasOfBesaid>
Sonarpulse: Cross-compiling for OS development is currently ... inconvenient. I can't get a cross-compiler without also pulling in a libc.
<Sonarpulse>
lucasOfBesaid: oh!
<lucasOfBesaid>
and Linux headers
<Sonarpulse>
did it work before?
<colabeer>
when i want to be dropped in a nixshell with a lib in the ld library path say libpng. How do I do that?
<sphalerite>
colabeer: nix-shell -p libpng
<lucasOfBesaid>
Sonarpulse: clever helped me confirm that it seems to have, yes. Used to be able to specify "libc = null" in a crossSystem.
<sphalerite>
colabeer: nativeBuildInputs is probably the most helpful one
<Sonarpulse>
I need to get that landed
<Sonarpulse>
then we will be in cross nirvana
<Sonarpulse>
(well, baring libc ugliness)
eacameron has joined #nixos
<colabeer>
sphalerite: when i run a binary inside nix-shell, how does it know where to search for the libraries when I did nix-shell -p ...
Fare has joined #nixos
<clever>
Sonarpulse: i was mainly setting libc to null to make sure any attempt at using libc would fail
<Sonarpulse>
clever: ah ok, that's fair
<clever>
Sonarpulse: so i would get a purely libc-less gcc
<sphalerite>
colabeer: oh wait you want to set `LD_LIBRARY_PATH`?
sellout- has joined #nixos
<colabeer>
sphalerite: Iam not sure, how do other nix packages make sure, their provided binaries can access shared libraries?
Turion has quit [(Quit: Leaving.)]
<sphalerite>
colabeer: by specifying them as build inputs, then nix checks which paths are referenced and makes sure that if A depends on B, B must be present for A to be present
Ivanych has quit [(Ping timeout: 260 seconds)]
Filystyn has quit [(Remote host closed the connection)]
<lucasOfBesaid>
Time for me to head out. Thanks all, for the help.
<avn>
colabeer: just by binary search, all what reference to /nix/store/$hash is a dependency
<colabeer>
sphalerite: the man nix-shell says, that it adds the lib locations to NIX_LDFLAGS, shall i append NIX_LDFLAGS to LD_LIBRARY_PATH then to make them accesable
<colabeer>
clever: thx for the hin. I've seen patchelf somehwere before, I think I should check out how it works :)
tnks has left #nixos []
<sphalerite>
colabeer: I used pkgs.runCommand there but you could use builtins.derivation instead to get an env with only the `LD_LIBRARY_PATH` added rather than all the stuff stdenv brings with it
vinymeuh has joined #nixos
vinymeuh has quit [(Client Quit)]
<sphalerite>
colabeer: which would look like nix-shell -E 'with import <nixpkgs> {}; builtins.derivation {name = "env"; LD_LIBRARY_PATH = pkgs.lib.makeLibraryPath (with pkgs; [libpng]); builder = "<this will not run>"; inherit system;}'
<sphalerite>
clever: why the gcc buildInput?
<colabeer>
hui then I'd have a very clean environment :)
<sphalerite>
clever: and wouldn't writeScript be sufficient?
<NixOS_GitHub>
[nixpkgs] volth opened pull request #28016: orackejdk: untie from gnome2 (master...patch-7) https://git.io/v76ZO
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 opened pull request #28017: ghcjs / nodejs: Add back older 6.x node version for GHCJS (release-17.03...ghcjs-node-version) https://git.io/v76Zs
NixOS_GitHub has left #nixos []
simukis has quit [(Ping timeout: 240 seconds)]
joehh has joined #nixos
Fare has quit [(Ping timeout: 260 seconds)]
cpennington has quit [(Ping timeout: 240 seconds)]
georgiy has quit [(Ping timeout: 260 seconds)]
zeus__ has quit [(Read error: Connection reset by peer)]
zeus__ has joined #nixos
justbeingglad has joined #nixos
justbeingglad has left #nixos []
phreedom has joined #nixos
silver_hook has quit [(Ping timeout: 240 seconds)]
<jophish>
Hmm, deepClone doesn't seem to be working for me
<Infinisil>
jophish: With fetchFromGitHub you also need to set fetchSubmodules = true; to get the .git directory
<mpcsh>
hey there, I have a script for locking my screen and suspending that hangs on nixos. however, if I run it with `bash -x suspend`, it works fine! anyone with better bash knowledge know what could be the cause?
<Infinisil>
mpcsh: When you check the contents of this script, does it have a shebang at the beginning?
<jophish>
Infinisil: ah, I've been using fetchgit
<Infinisil>
vim $(which suspend)
<jophish>
I'll try with fetchFromGitHub
<mpcsh>
Infinisil: #!/usr/bin/env bash
<Infinisil>
jophish: Ah, well no, fetchgit should work, fetchFromGithbu with fetchSubmodules = true just uses fetchgit underneath
<jophish>
Ah, I thought that might be the case :)
<Infinisil>
jophish: What parameters do you pass to it?
<Infinisil>
mpcsh: No idea what's wrong with that, where does `suspend` come from? Your DE?
<Infinisil>
mpcsh: No idea what's wrong with this script, maybe you could try bisection to find out where the problem lies
<mpcsh>
Infinisil: well there's nothing to bisect
<mpcsh>
I wrote these back on arch and haven't changed them in years
<mpcsh>
And now after moving to nixos the suspend script just hangs
<LnL>
jophish: I think/hope fetchgit won't include all the branches
<mpcsh>
Infinisil: ok after inserting some "echo"'s, I've determined that it doesn't make it past the shebang
<Infinisil>
mpcsh: You could move the lock script into the suspend one, remove pics, maybe remove the IFS
<mpcsh>
Infinisil: not even the `set -euo` executes
<mpcsh>
Infinisil: (without bash -x, that is)
<LnL>
jophish: that would break the hash if a branch is updated
<Infinisil>
mpcsh: Maybe just write a small bash script that just does echo hi
<jophish>
LnL: I was hoping it woulld keep the default one though
<mpcsh>
Infinisil: it works
<jophish>
but that does make a lot of sense now that I think about it
<mpcsh>
Infinisil: the lock script works fine
<jophish>
I'll just patch dronin not to do this silly thing
<mpcsh>
Infinisil: I'm an idiot
Sonarpulse has quit [(Ping timeout: 248 seconds)]
<mpcsh>
Infinisil: it's clobbering the suspend bash builtin
<Infinisil>
mpcsh: Oh lol
<mpcsh>
Infinisil: why wasn't it doing that on arch???
<Infinisil>
did you use bash on both systems?
<mpcsh>
Infinisil: yeah!!
odites has joined #nixos
<Infinisil>
Maybe different versions
<Infinisil>
Or different default settings, I don't know if nix modifies these in any way
<Infinisil>
s/nix/nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] expipiplus1 opened pull request #28018: dRonin: init at 26-06-2017 (master...dronin) https://git.io/v76lI
NixOS_GitHub has left #nixos []
Fare has joined #nixos
slyfox has quit [(Ping timeout: 240 seconds)]
Fare has quit [(Read error: Connection reset by peer)]
<jophish>
night all!
slyfox has joined #nixos
Fare has joined #nixos
odites has quit [(Quit: Konversation terminated!)]
silver_hook has joined #nixos
markus1189 has joined #nixos
<Infinisil>
jophish: Night!
Fare has quit [(Ping timeout: 240 seconds)]
markus1199 has quit [(Ping timeout: 240 seconds)]
Fare has joined #nixos
joehh has quit [(Ping timeout: 255 seconds)]
digitus has quit [(Quit: digitus)]
rpifan has joined #nixos
silver_hook has quit [(Ping timeout: 260 seconds)]
sellout- has quit [(Quit: Leaving.)]
Fare has quit [(Ping timeout: 260 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Infinisil opened pull request #28019: xmonad service: add defaultText to extraPackages to fix rendering in docs (master...fix-default-text-xmonad) https://git.io/v7643
NixOS_GitHub has left #nixos []
erictapen has joined #nixos
bennofs has joined #nixos
johnw_ has joined #nixos
alunduil has joined #nixos
johnw has quit [(Ping timeout: 255 seconds)]
erasmas has quit [(Quit: leaving)]
amen has joined #nixos
phinxy has joined #nixos
kiloreux has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] michalrus opened pull request #28020: geekbench: init at 4.1.1 (master...geekbench) https://git.io/v76RK