<simpson>
hyperfekt: Can the locations of the scripts be patched inside transcrypt's source before it's even built?
__monty__ has quit [Quit: leaving]
<hyperfekt>
simpson: No, since that wouldn't work for anyone else using the repository, and no security updates would ever be applied to the once written scripts.
<hodapp>
hmm, "just don't use dropbox" is looking like a better and better option
<hodapp>
seeing how it insists on downloading and managing its own binary
<simpson>
hyperfekt: Oh ew, that's terrible. It does that on every platform?
<hodapp>
because obviously my package needs its own package manager
jluttine has joined #nixos
<ottidmes>
does anyone know if its possible in bash/zsh to get the arguments passed to a command while still triggering completions? like cmdfoo "$(cmdbar --option param)" and have cmdfoo get ( cmdbar --option param ), right now I workaround it by copying over the completion files and making them work for my alias that returns the arguments, it works, but I would love to just be able to use the original name
adamt has quit [Ping timeout: 255 seconds]
<hyperfekt>
simpson: It appears to be the way filters are added to a git repository in order to not require the tool itself, but only openssl.
<hyperfekt>
Hmm, maybe we could patch the scripts so they try to call nix-shell with the required attributes
jluttine has quit [Ping timeout: 255 seconds]
fusion809 has joined #nixos
eadwu has joined #nixos
endformationage has joined #nixos
<hyperfekt>
That would have the advantage of automagically installing the necessary dependencies on other nix systems. I'm not quite sure how to do it, though.
illegalprime has quit [Remote host closed the connection]
hyperfekt has quit [Ping timeout: 256 seconds]
thc202 has quit [Ping timeout: 246 seconds]
infinii has joined #nixos
infinii has quit [Ping timeout: 246 seconds]
matijja has quit [Ping timeout: 268 seconds]
fusion809 has quit [Ping timeout: 252 seconds]
yl has quit [Ping timeout: 245 seconds]
doyougnu has joined #nixos
jtcs has quit [Ping timeout: 268 seconds]
jluttine has quit [Ping timeout: 246 seconds]
Makaveli7 has quit [Ping timeout: 245 seconds]
jluttine has joined #nixos
iqubic has left #nixos ["ERC (IRC client for Emacs 26.1)"]
mounty has quit [Remote host closed the connection]
mounty has joined #nixos
infinii has joined #nixos
schjetne has joined #nixos
schjetne has quit [Ping timeout: 255 seconds]
andersk has quit [Remote host closed the connection]
doyougnu has quit [Ping timeout: 246 seconds]
andersk has joined #nixos
cthachuk has quit [Remote host closed the connection]
hiroshi has joined #nixos
aanderse-liveusb has joined #nixos
<aanderse-liveusb>
my hard drive half died. i was running 18.09 but only had a 18.03 live usb. i've copied / over from the old hard drive to another hard drive. trying to recover system
<aanderse-liveusb>
so my system would build against 18.09 instead of 18.03 (because my configuration.nix has 18.09 specific stuff in it)
<aanderse-liveusb>
nixos-install is complaining that some options don't exist, so it appears that nixos-install is still building against 18.03
<aanderse-liveusb>
any help?
jluttine has quit [Ping timeout: 255 seconds]
fusion809 has joined #nixos
<clever>
aanderse-liveusb: did you run `nix-channel --update` as well?
<aanderse-liveusb>
>_<
<aanderse-liveusb>
so... twice in a day
<aanderse-liveusb>
where do i send the obligatory beer to?
<clever>
i dont drink :P
<ottidmes>
clever++
<{^_^}>
clever's karma got increased to 99
<ottidmes>
aanderse-liveusb: now you, then you get the honor of getting him to 100 :)
<aanderse-liveusb>
clever++
<{^_^}>
clever's karma got increased to 100
<aanderse-liveusb>
:)
<ottidmes>
whohhooh! gratz clever :)
<clever>
:D
<aanderse-liveusb>
well chosen nickname as well i'll say
<aanderse-liveusb>
ha ha
furrycatherder has quit [Remote host closed the connection]
<aanderse-liveusb>
i wanted to hack on a module tonight... not restore my os from a faulty hard drive :(
<ottidmes>
but NixOS should help with those painful moments
<aanderse-liveusb>
sure does
<aanderse-liveusb>
seems like i was able to recover the mutable parts of /etc and all of /var
<aanderse-liveusb>
so in theory i should be able to boot into my old system with no difference except the hard drive it is running on
slyfox has quit [Ping timeout: 272 seconds]
<aanderse-liveusb>
i love the combination of zfs for mutable data + nixos for system
rcshm_ has quit [Remote host closed the connection]
<aanderse-liveusb>
especially at times like these
<aanderse-liveusb>
:)
<ottidmes>
yeah, I will convert all my systems to that as well at some point
<ottidmes>
right now only my backups are on ZFS
rcshm has joined #nixos
<clever>
i need to automate `zfs send` on my end
<aanderse-liveusb>
yeah when i upgraded to18.09 i decided it was time to switch everything over to zfs on root... because nixos makes it so easy
<aanderse-liveusb>
yeah i still use rsync for backups... mostly because i haven't had a chance to experiment with zfs send recv enough yet
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
<ottidmes>
I still plan to make it so that eventually I can just take a new harddrive, boot a custom installer via usb/kexec and just have it reproduce my system without manual intervention, but that takes a lot of work
jluttine has joined #nixos
<aanderse-liveusb>
sounds great
<clever>
ottidmes: that reminds me, let me find it...
dramforever has joined #nixos
<aanderse-liveusb>
ok rebooting... wish me luck
<ottidmes>
good luck!
aanderse-liveusb has quit [Remote host closed the connection]
<clever>
ottidmes: having trouble searching up the gist...
<acowley>
I'm still having trouble applying a specific config file for a kernel build. Does anyone have an example of calling `make my_config` with a `my_config` found in the kernel's source tree as part of the build?
dramforever has quit [Remote host closed the connection]
silver has quit [Read error: Connection reset by peer]
<ottidmes>
clever: anything special? most effort is going to be all those small things that add up, getting the files in the right place, correct formatting, etc.
<acowley>
An annoyance with this challenge is that the build tends to keep going past errors only to fail much later rather than quit at the first sign of something being wrong.
<clever>
ottidmes: mostly, to skip the 1st generation nixos-install makes, and let you deploy to a 100% empty hdd
<clever>
ottidmes: basically, you can boot a nixos livecd, mount the rootfs to /mnt/, then run this to copy a pre-made nixos, directly into /mnt/, over ssh
<clever>
ottidmes: then you just chroot in (nixos-enter), and run `/nix/store/hash-nixos/bin/switch-to-configuration boot --install-bootloader`
<clever>
boom, your installed
<ottidmes>
clever: thats nice :)
infinii has quit [Ping timeout: 255 seconds]
aanderse has joined #nixos
<ottidmes>
clever: thanks for the tip, made a note for it :)
<aanderse>
:)
<clever>
ottidmes: with some minor tweaks, you could teach nixops about this, and then it can install to a blank hdd
<aanderse>
well that was the smoothest hard drive crash i've ever had
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
<clever>
NIX_REMOTE=local?root=/mnt/img/ nix-env --profile /mnt/img/nix/var/nix/profiles/system --set -f '<nixpkgs/nixos>' -A system -I nixos-config=/mnt/img/etc/nixos/configuration.nix --option system aarch64-linux
<clever>
Raw
<clever>
ottidmes: this is another fun one
<clever>
ottidmes: this will setup the system profile, inside a chroot, and cross-compile
<clever>
but variations can be made for other uses
<clever>
ottidmes: this is the most complex one of them all
<clever>
that makes a special build of nix, that expects the store to live in /home/$X/nix/store
<clever>
but also, at the same time, it will chroot into /home/clever/rootfs/ while building that
<ottidmes>
that needs a lot of know how of how of the internals
<clever>
so it creates a /home/clever/rootfs/home/$X/nix/store
<clever>
you then move the whole /home/clever/rootfs/home/$X to /home/$X on another machine
<clever>
and nix will continue to work there
<clever>
but you wont have any help from the binary cache
<ottidmes>
right, was about to say that
goibhniu has quit [Ping timeout: 255 seconds]
<clever>
the neat thing, is that you can use nix on one box (without root) to build it, in a pure manner, and then copy it to another box (also without root)
<ottidmes>
yeah, I really ought to play more with that, nix copy and such
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
vidbina has joined #nixos
<Ashy>
hmm, anyone here using sway-beta? swaygrab does not seem to be part of sway-beta and the version from the sway package fails to run with a json assert abort
ayerhart has quit [Read error: Connection reset by peer]
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
ayerhart has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
ayerhart has quit [Read error: Connection reset by peer]
ayerhart has joined #nixos
vidbina has quit [Ping timeout: 255 seconds]
ym555_ has joined #nixos
ym555 has quit [Ping timeout: 255 seconds]
ym555_ is now known as ym555
vidbina has joined #nixos
freeman42y has quit [Ping timeout: 258 seconds]
kvda has joined #nixos
ym555 has quit [Ping timeout: 245 seconds]
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
ym555 has joined #nixos
adamantium has joined #nixos
ddellacosta has quit [Ping timeout: 255 seconds]
ddellacosta has joined #nixos
orivej has joined #nixos
rcshm has quit [Remote host closed the connection]
i1nfusion has quit [Remote host closed the connection]
schjetne has joined #nixos
ericsagnes has quit [Ping timeout: 240 seconds]
i1nfusion has joined #nixos
<gchristensen>
anyone familiar with rust + darwin + llvm + doctests? when I run `cargo test` with a rust project which uses doctests, I get dyld: Symbol not found __ZTIN4llvm2cl18GenericOptionValueE at compile time
schjetne has quit [Ping timeout: 245 seconds]
eadwu has quit [Quit: WeeChat 2.4]
eadwu has joined #nixos
eadwu has quit [Client Quit]
softinio has quit [Quit: Connection closed for inactivity]
<jasom>
is user namespaces support required for nixpkgs on linux?
<clever>
jasom: its used by the sandboxing, which could be disabled
<clever>
but i prefer to have the sandbox on 24/7
* jasom
grumbles and considers whether or not to disable sandboxing or to recompile his kernel
ottidmes has quit [Ping timeout: 255 seconds]
<clever>
jasom: gentoo?
<jasom>
yup
<clever>
jasom: should be fairly fast to build the kernel on a modern system
<clever>
not a 10min job anymore
<gchristensen>
interesting, do they turn it off by default?
<jasom>
the reboot will take longer than the compile
<clever>
gchristensen: gentoo lets you do crazy things, like not even have a libcurl, anywhre
<clever>
jasom: heh, i know what you mean
<gchristensen>
yeah, I know, but is it off by default?
<jasom>
gchristensen: they do: "run make menuconfig and configure it yourself" by default
<clever>
gchristensen: that lack of curl, is how i discovered a bug in the adobe flash player
<jasom>
regardless the feature probably didn't exist when I first installed gentoo and I've been bringing my kernel config with "make oldconfig" as I go, so "make oldconfig" clearly doesn't enable it by default.
<clever>
jasom: i did the same thing when i ran gentoo
slyfox has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
Glider_IRC__ has quit [Remote host closed the connection]
jasom has quit [Quit: WeeChat 2.3]
Glider_IRC__ has joined #nixos
jasom has joined #nixos
<jasom>
and it's installed now; thanks clever!
<clever>
jasom: your welcome :)
<gchristensen>
clever++
<{^_^}>
clever's karma got increased to 101
<clever>
gchristensen: now my karma is binary, lol
<gchristensen>
perfect, up to 5 karma
<ar1a>
considering giving someone karma takes away from yours... is the sum of karma 0?
<ar1a>
or was some given out
ericsagnes has joined #nixos
<gchristensen>
heh
<aleph->
Sigh
<clever>
gchristensen: oh, random idea about packet.net and nixops
<aleph->
So salt stack has horrifying crypto defaults.
<clever>
gchristensen: why install nixos (with a random nixpkgs rev) when nixops is going to deploy a closure that likely is a mass-rebuild?
<clever>
then deploy the final build, directly to the EBS
vidbina has joined #nixos
<samueldr>
I just used the rescue_boot one merely an hour ago
<clever>
no wasted space from the old generation in the AMI
<samueldr>
(had misconfigured the added drive and didn't want to deal with booting a previous generation which would dirty up something I was getting rid of)
<clever>
that minor tweak, would add aws to the list above, and allow any FS on aws
<clever>
samueldr: nice
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
the only major fault i can see with the design of rescue-boot, is that it rebuilds on every nixos-rebuild
<clever>
so if your kernel becomes faulty, the rescue-boot kernel is also faulty!
<samueldr>
though it's not taking space on the boot partition since only one is copied over
<samueldr>
ah, yeah :)
<clever>
but you could always change the pkgs.path to be a pinned one
<samueldr>
though it also wouldn't necessarily be the same kernel, the rescue-boot one would be the default one, since it doesn't look at your system config
<samueldr>
so I can screw up my system's kernel without much fear
<clever>
yeah
<clever>
more a matter of the nixpkgs rev breaking core things
<samueldr>
though I think it'd be affected by overlays
<clever>
Arahael: and android-studio still exists there, did you enable unfree?
<Arahael>
I think I did, yes. android-studio is at version 3.1, unless I use the canary version.
rprije has joined #nixos
<clever>
yeah, 18.09 has 3.1.4.0
<clever>
nixos-unstable has 3.3.0.20
<Arahael>
Ah, I'll make do with 3.1, then, for now. I'm only poking around, anyway.
<Arahael>
Doing the nixos-rebuild switch now, incidentially, I have another issue there: It seems that a masked systemd unit for some reason: Failed to start local-fs.target: Unit -.mount is masked.
<Arahael>
I asked in nixos-chat but that channel is quiet at the moment.
<clever>
not sure what causes that
<clever>
but mask/unmask/enable/disable just wont work on nixos
<Arahael>
So what's the solution?
<clever>
unknown
<Arahael>
I tried nixos-rebuild switch --rollback, but it doesn't fix it.
<clever>
rebooting?
<Arahael>
I haven't tried that.
<Arahael>
But if it's masked, why would a reboot fix that?
<clever>
something might be in a weird state
<Arahael>
The systemd state like that is that dynamic? :/
<clever>
some of it is
<Izorkin>
How to need to merge this PR #54903 #51902 #44343 ?
jtojnar has quit [Remote host closed the connection]
Ralith_ has quit [Remote host closed the connection]
akaWolf has quit [Remote host closed the connection]
MinceR has quit [Remote host closed the connection]
callahad has joined #nixos
mmlb has joined #nixos
gurmble is now known as grumble
DanC has joined #nixos
copumpkin has joined #nixos
pepesza has quit [Ping timeout: 249 seconds]
pepesza has joined #nixos
yokyolu has joined #nixos
<yokyolu>
i'm trying to add a package from a pinned nixpkgs to an overlay
<yokyolu>
more specifically a haskell package
<yokyolu>
but once in the shell cabal doesn't pick it up
hedning has joined #nixos
rgrau has quit [Ping timeout: 272 seconds]
Czen has quit [Ping timeout: 272 seconds]
rixed has quit [Ping timeout: 272 seconds]
MasseR has quit [Ping timeout: 272 seconds]
rixed has joined #nixos
Czen has joined #nixos
MasseR has joined #nixos
rgrau has joined #nixos
yokyolu has quit [Ping timeout: 264 seconds]
neminis has quit [Quit: WeeChat 2.2]
FRidh has joined #nixos
crmlt has joined #nixos
<bpye>
Hm, rclone does weird things, it updates it's own config file to update tokens, any idea how I can handle this in a service?
<bpye>
Should I just put the config file somewhere mutable? I don't like that...
crmlt has quit [Remote host closed the connection]
slack1256 has joined #nixos
palo has quit [Quit: WeeChat 2.2]
jacereda has joined #nixos
switchsox has quit [Remote host closed the connection]
Glider_IRC_ has joined #nixos
Glider_IRC__ has quit [Ping timeout: 245 seconds]
jasongrossman has joined #nixos
<bpye>
Is there a good way to handle tmp files in a systemd service? It seems ideally Id created the temp file in PreStart and delete it in PostStop, but I don't know if that's possible?
knupfer has joined #nixos
<bpye>
Since I need to pass the filename into my main script
jacereda has quit [Ping timeout: 245 seconds]
dwdr has joined #nixos
dwdr has quit [Remote host closed the connection]
dwdr has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
<Havvy>
bpye: Is it okay if the file gets deleted or messed with while the service is running?
Ariakenom has joined #nixos
<bpye>
Havvy: I need the service to be able to modify it (updates a token), but the token expires before it runs again so it doesnt matter that I don't keep it. the file shouldn't be modified under the service though
<bpye>
Whilst it's running, only the service should be modifying it
simukis has quit [Quit: simukis]
<Havvy>
Well, /tmp doesn't work then.
<bpye>
I suppose that's true, I can always use mktemp with a different directory though can't I?
<bpye>
Same general idea
<Havvy>
Yeah.
<Havvy>
I wouldn't worry about cleaning up the file until you uninstall the service personally.
<bpye>
Havvy: It's a timer task, for backups. But hm, I need to use su to run a command as some other users to ensure that I copy a consistent state, but if I'm not using /tmp I should probably create a user to run this task, but then they need to be able to run su
<bpye>
Which if I'm not being dumb is something only root can do
<kyren>
hey, I'm trying to get looking glass working on a vm, and I'm setting up a shared memory file for it vaguely following what the arch wiki is telling me:
<kyren>
my plan is to create a systemd job just like that guide is saying to create a file in /dev/shm/ of something like root:mycustomgroup, then add my user to mycustomgroup so that I can read / write to it
<Havvy>
I know not about `su`'s restrictions on who can use it.
<kyren>
I'm here to ask whether or not I'm overcomplicating things and if there's some easier way to do it that anybody knows of
dermetfan has joined #nixos
zupo has joined #nixos
<kyren>
I guess using a group like libvirtd is easier
<bpye>
Havvy: I was being silly, no reason it has to run as non root to use a different dir eh, can just create /var/lib/rclone, done 🙂
<bpye>
And now I can auto backup my git repos and my samba shares and everything is happy and I get to go to bed, good night
wfranzini has quit [Ping timeout: 246 seconds]
Guanin has joined #nixos
palo has joined #nixos
goibhniu has joined #nixos
leotaku has quit [Ping timeout: 245 seconds]
wfranzini has joined #nixos
pointfourone has joined #nixos
knupfer has quit [Remote host closed the connection]
knupfer has joined #nixos
cwhy has quit [Ping timeout: 256 seconds]
__monty__ has joined #nixos
srhb- is now known as srhb
<dminuoso>
If I set up a `nix-shell -p haskell.compiler.ghc863` is there a way to also include the previous profile/environment?
schjetne has joined #nixos
pointfourone has quit [Quit: Leaving]
<__monty__>
dminuoso: Without --pure nix-shell doesn't clear much from the environment afaik. Do you mean nested nix-shells or something?
<dminuoso>
__monty__: No I want the opposite of --pure. So lets say I have a vim package installed via home-manager. If I run that shell above, vim is no longer in my PATH.
<dminuoso>
Trying to figure out how to manage multiple GHC versions nicely with nix
<__monty__>
Ordinary nix-shell wouldn't do that afaik. So I suspect it's a home-manager quirk.
<dminuoso>
__monty__: Im so close to downloading the ghc bindists and manage the versions myself.
<dminuoso>
__monty__: actually wait, vim is installed via /etc/nixos/configuration.nix even
<dminuoso>
Oh hah wait!
<dminuoso>
This is embarrassing.
<dminuoso>
With nix-shell my .zshrc is no longer loaded, so my alias vi="vim" is no longer there.. that took me just about 30 minutes to figure out!
ericsagnes has joined #nixos
<__monty__>
Just symlink instead of aliasing : )
Makaveli7 has joined #nixos
<dminuoso>
__monty__: How would I even do that in nix?
<__monty__>
That's a good point. Maybe an overlay that adds the symlink to the vim package?
<srhb>
It'll probably be easier to create a package that is _just_ the symlink
<srhb>
Less recompiling.
<DigitalKiwi>
you could just type vim
<dminuoso>
DigitalKiwi: Do you know how much time I would lose over the duration of my life?
<DigitalKiwi>
less than 30 minutes
astrofog has joined #nixos
<__monty__>
I tend to avoid aliases/functions, usually unnecessary anyway and scripts are way more portable.
<dminuoso>
Also, muscle memory be damned!
<srhb>
As in, runCommand "vi" "mkdir $out/bin && ln -s ${pkgs.vim}/bin/vim $out/bin/vi"
<srhb>
uh, forgot the second argument, empty set should do..
<__monty__>
srhb: Hmm, I thought overlays prevented all the unnecessary recompilation that comes with overrides?
<srhb>
__monty__: No, not at all.
astrofog has quit [Remote host closed the connection]
<srhb>
If you modify the vim package in any meaningful way, you'll get a new hash and thus a rebuild, regardless of the method of the change.
<srhb>
That is an important feature. :)
<DigitalKiwi>
also isn't vi usually a symlink to vim already anyway?
<srhb>
DigitalKiwi: Not in the vim package at least.
<srhb>
Uh, yes it is..
<__monty__>
srhb: I remember you recommending switching from overrides to an overlay for my ghcWithHoogle setup though. You said something along the lines of "You must be doing a ton of rebuilding because of overrides."
<srhb>
__monty__: Then you were probably doing something that would recursively alter the hashes of a lot of things (through very base-of-the-tree dependencies, or maybe by modifying callPackage)
<srhb>
__monty__: Really, the only thing you can do here is be clever about what you alter to get the minimal amount of rebuilds. But anything you _do_ alter will need to be rebuilt.
<DigitalKiwi>
oh i have a question how come sometimes i get syntax highlighting and sometimes i don't :(
<srhb>
DigitalKiwi: Basically a flexible derivation for setting up a more customized Vim.. It doesn't come with the symlink
<DigitalKiwi>
oh
<srhb>
__monty__: At a guess, Vim probably looks at the name it was called as and decides compat from there.
<__monty__>
Is that what I need to get vim +perl +python?
sigmundv has joined #nixos
<srhb>
__monty__: It's one way.
<srhb>
DigitalKiwi: By "sometimes" do you mean for the same filetype?
schjetne has quit [Ping timeout: 245 seconds]
<DigitalKiwi>
srhb: yeah
<srhb>
DigitalKiwi: Hm, not sure.
<DigitalKiwi>
like in the same window same command same filename/extension one will have syntax one won't
<DigitalKiwi>
it's more than mildly irritating
<srhb>
DigitalKiwi: Can you try forcing the filetype next time? It might help debugging it.
hoijui has joined #nixos
<DigitalKiwi>
how do that? mostly i notice it with default.nix files
hoijui has quit [Remote host closed the connection]
schjetne has joined #nixos
talyz has quit [Ping timeout: 245 seconds]
thc202 has joined #nixos
<DigitalKiwi>
i feel like it's less common for vim to behave differently when it's called as vi instead of vim, which for me, is nice...never have i wanted to use vi instead of vim >.>
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<qyliss>
Arahael: I’ve seen that before too. A reboot makes it go away for me.
<Arahael>
qyliss: Awesome - thanks.
talyz has joined #nixos
dwdr has quit [Remote host closed the connection]
dramforever has joined #nixos
the-kenny has quit [Ping timeout: 250 seconds]
regulus_ has quit [Ping timeout: 246 seconds]
pointfourone has joined #nixos
pointfourone has quit [Client Quit]
srl295 has quit [Quit: Connection closed for inactivity]
slack1256 has quit [Remote host closed the connection]
cmacrae has joined #nixos
_kwstas has joined #nixos
<cmacrae>
o/ With 'mpv' in my systemPackages, how can I ensure 'libmpv.so' is also in the resulting output?
StardustX has joined #nixos
StardustX has left #nixos [#nixos]
regulus_ has joined #nixos
jacereda has joined #nixos
petrkr has joined #nixos
emily has quit [Remote host closed the connection]
xkapastel has joined #nixos
emily has joined #nixos
petrkr has quit [Ping timeout: 258 seconds]
<__monty__>
cmacrae: That doesn't sound like a nix-y thing to do. Usually you specify runtime dependencies in the package that depends on them. And buildtime dependencies in your default/shell.nix, generally you don't install libraries outright.
johanot has joined #nixos
<cmacrae>
__monty__: Sure, that's fair enough. You make a good point, this is something I want for a Python project I'm working on, so it'd make more sense to express it in the shell.nix
<cmacrae>
I don't suppose you have any advice for how I might achieve that? Perhaps how I might express it through buildInputs
silver has joined #nixos
_kwstas has quit [Quit: _kwstas]
<srhb>
cmacrae: Yeah, just having mpv in buildInputs should do it.
simukis has joined #nixos
<cmacrae>
srhb: Oh! Haha, alright
<cmacrae>
Thanks guys, I'll give it a go
petrkr has joined #nixos
jeaye has quit [Ping timeout: 250 seconds]
zupo has joined #nixos
jeaye has joined #nixos
rprije has quit [Ping timeout: 268 seconds]
regulus_ has quit [Ping timeout: 244 seconds]
<cmacrae>
Hmm, doesn't seem to have worked with just mpv in my buildInputs
<__monty__>
Maybe it needs to be propagatedBuildInputs or something? Considering you said python.
dermetfan has quit [Ping timeout: 250 seconds]
sigmundv has quit [Ping timeout: 246 seconds]
asymmetric has joined #nixos
<therealwaphire[m>
guys, have any of you tried using `/dev/sda9` created by ZFS as the ESP?
<Myrl-saki>
Hm. grub2_efi seems to be installed on my system but I don't know hwy.
<cmacrae>
__monty__: Tried using propagatedBuildInputs as well, but it still complains it can't find the shared object at the nix path: OSError: dlopen(/nix/store/h68692jk7bxynp1y8q2j2icadnzjz815-mpv-0.29.0/lib/libmpv.so, 6): image not found
dwdr has joined #nixos
<__monty__>
cmacrae: I'm out of my depth, I have no idea what an image is in this context. Did you check that path in the store?
<Myrl-saki>
Seems like it's because of dbus. RIP.
regulus_ has joined #nixos
<cmacrae>
__monty__: Sure, thank you for your help anyway :) I appreciate it. Yeah checked the path, and the file isn't there. Wondering if there's something I need to pass to the mpv expression to say I want the library included in the output
deltasquared has joined #nixos
<deltasquared>
hrm, so I go to install warp via a hosted nix's cabal, that blows up because it couldn't find zlib in it's search paths. I try installing zlib via nix, *that* blows up citing EPERM cloning a build process
<deltasquared>
something tells me nix is trying to do something that requires root permissions to isolate builds
<__monty__>
cmacrae: I don't think so. I think maybe your expression needs to explicitly say it needs the path like with `mkdir $out; mv libmpv $out/lib/libmpv` or something. But as I said I have no experience with this so I don't actually know the proper way.
<cmacrae>
Sure, I'll have a play around. Thanks again __monty__
<dramforever>
deltasquared: Are you running Nix in docker?
<deltasquared>
dramforever: negative
<deltasquared>
it's in it's own user however
<deltasquared>
for most nix things that tends to work...
<dramforever>
I think this too: 'something tells me nix is trying to do something that requires root permissions to isolate builds'
<dramforever>
Not exactly root, but some sort of permission regarding namespaces
<deltasquared>
faq, namespaces again
<deltasquared>
most distros disable user namespaces because they tend to cause problems
<deltasquared>
idk if that's something I'd have to recompile my kernel to include
<dramforever>
nix-daemon *is* running as root on my machine
<dramforever>
... which is the NixOS default
<deltasquared>
dramforever: "hosted nix"... I don't know if that's the right term.
<deltasquared>
I don't run nixos. I have some personal disagreements.
<symphorien>
dramforever: on debian there is a sysctl knob for example
<deltasquared>
just nix on arch linux...
<dramforever>
multi-user?
<deltasquared>
dramforever: no, just one user (also called "nix" because I'm not imaginative)
<__monty__>
deltasquared: Personal disagreements with the OS?
<dramforever>
No I mean the 'nix-daemon' thing is called multi-user
<deltasquared>
dramforever: I don't have a deamon running no.
<dramforever>
Multi-user setup vs. Single-user setup
<Myrl-saki>
I'M STUPID
<deltasquared>
so single user I guess.
<Myrl-saki>
I had grub2_efi in my configuration.nix
<Myrl-saki>
Crap.
<dramforever>
deltasquared ok then just disable sandboxing, I guess
<deltasquared>
__monty__: you could say that... I didn't have the time to invest to fix it as per OSS philosophy :>
<deltasquared>
dramforever: next stupid question, is there a config file for that
<dramforever>
~/.config/nix/nix.conf
<dramforever>
Put a line 'sandbox = false' in it
<__monty__>
deltasquared: What's broken about it though?
<dramforever>
Uh, just the home dir of whatever you're running Nix as, I suppose
<MichaelRaskin>
__monty__: I do have personal disagreements with NixOS although I agree with OSS stance taken by NixOS/Nixpkgs almost precisely. Feel free to ask in #nixos-chat for details
<deltasquared>
__monty__: many numerous things, I found it too difficult to shape to my liking and I wasn't interesting in sinking weeks on a learning curve just to get to an equivalent setup to the one I have now.
<deltasquared>
but yeah, maybe best kept to the offtopic
<simpson>
deltasquared: It could be extremely on-topic; it's hard for folks to support a custom Nix installation which has gone off the beaten path and what's described in the manuals.
<deltasquared>
simpson: you should see the crazy stunts I pull on arch. there's a lot of custom scripts and services.
ottidmes has joined #nixos
<deltasquared>
also, there's just the decreased size of software packaged by default (at least last time I tried)
<deltasquared>
I don't have the patience for source compilations. sure it's hand configured, but after that I expect to just be able to get on with things...
<simpson>
deltasquared: No worries. NixOS isn't for everybody.
<deltasquared>
simpson: naturally. don't get me wrong, the concept is awesomesauce, but I don't have the free time to invest into it
palo has quit [Quit: WeeChat 2.2]
<Twey>
Yeah, it's still very much contributor software :)
<Twey>
Some day!
<simpson>
deltasquared: Is that what you tell me or what you tell you? It sounds like you have time for lots of custom scripts and services, something I don't really have time for.
orivej has quit [Ping timeout: 268 seconds]
<dramforever>
lots of scripts and services accumulated over time and difficult to port, I support?
<Twey>
Well, most of the time I put in is into Nix (and nixpkgs), not NixOS
<Twey>
NixOS basically just works for me after a little while setting it up
<deltasquared>
simpson: I used to. these days I just maintain them and occasionally need a drink after looking at really old ones.
_kwstas has joined #nixos
<deltasquared>
dramforever: precisely
palo has joined #nixos
<simpson>
deltasquared: Ah! I can't drink, so I've proceeded down a path of being able to hack and replace pieces of infrastructure on an as-needed basis. I used to run (and contribute to?) Gentoo before NixOS.
<dramforever>
So, uh, back to unable to start builds
<dramforever>
I suppose you installed Nix in some other user, and switch to that user when you want to install stuff
<deltasquared>
dramforever: your line worked perfectly, btw
<deltasquared>
and yes.
eadwu has joined #nixos
<dramforever>
Oh
<dramforever>
Okay, no more unable to start builds at least for now :)
<deltasquared>
basically all non-native packaging systems are their own user, because nothing should need root
<deltasquared>
luarocks, cabal, nix, npm...
<deltasquared>
cabal was an interesting one, I need to source a script to make that work but otherwise it hold up well
<deltasquared>
cabal from nix, naturally, because arch's haskell packaging sucks
<Twey>
So I just learnt about `placeholder` and I don't really understand it :þ How does it work?
petrkr has quit [Remote host closed the connection]
<deltasquared>
ok so this is annoying, cabal from nix still can't zlib, but at this point I doubt it's a nix problem per se
<deltasquared>
*can't find
<dramforever>
Try something like nix-shell -p zlib
<deltasquared>
dramforever: then run cabal in that...?
<dramforever>
yeah
<celeron55>
i'm trying to get nixos grub menu to work from grub that was installed by fedora by adding a menuentry that does root, search, configfile '/boot/grub/grub.cfg' and it works fine otherwise, but the problem is that fedora's grub is patched to use the "linuxefi" and "initrdefi" commands instead of "linux" and "initrd" that work exactly the same except for the name... so i need to get nixos to generate menu.cfg with linuxefi and initrdefi ...
<celeron55>
... instead of linux and initrd
<celeron55>
how would i go about making that happen
<celeron55>
just having linuxefi and initrdefi there would make everything work, but as a newbie i don't really see how to make it happen in a manageable way
<celeron55>
like, is there a way to add a post-hook to the generation of grub.cfg, or a way to persistently automatically use an install-grub.pl supplied by myself, or a way to automatically patch install-grub.pl
<celeron55>
or really anything that would allow me to customize the automatic grub.cfg entries
<celeron55>
well i guess patching grub with the same patch fedora uses would also work
<deltasquared>
dramforever: to run the nix-shell command
<simpson>
Twey: Search and replace? Not sure. Never thought about it before.
<dramforever>
Ah
<deltasquared>
dramforever: so uh, no idea which way around to do it xD
<dramforever>
So you mean you need to be nix user to use nix, but also cabal user to use cabal?
<Twey>
simpson: It would have to register a mapping, right? But I can't see anything like that in the source. Might just be missing it :þ
<deltasquared>
dramforever: essentially... >_<
<Twey>
deltasquared: You've carefully separated the user who has `zlib` from the user who has `cabal`, and now you're trying to combine `zlib` and `cabal` :þ
<deltasquared>
I think I may try passing the store path directly.
<Twey>
deltasquared: You could use groups maybe?
<Twey>
Oh yeah, or you can do that if the cabal user has read access
<deltasquared>
Twey: oh yeah, the entire system can read it
<Twey>
But beware of transitive dependency conflicts…
<Twey>
If you e.g. pull in glibc somewhere else from cabal, and zlib expects the Nix glibc, and the two are sufficiently different versions, you're going to have a bad time
acarrico has quit [Ping timeout: 255 seconds]
ixxie has joined #nixos
<deltasquared>
Twey: I'll bear that in mind.
<deltasquared>
thankfully it seems nixos's zlib is statically linked only...
<deltasquared>
err nix's rather
<deltasquared>
ah, it's all building fine now
<deltasquared>
all four cores at 100% during compilation is a fine thing to see
jabranham has joined #nixos
ng0 has quit [Quit: Alexa, when is the end of world?]
<deltasquared>
oooh, nix-env upgrade got me ghc 8.6.3 finally!
<deltasquared>
...
<deltasquared>
Could not load module ‘Network.Wai’
<deltasquared>
It is a member of the package ‘wai-3.2.2-K52gsfZg0b72WhZKsLd0DA’
<deltasquared>
which is unusable due to missing dependencies:
<deltasquared>
>_<
<Twey>
deltasquared: Always ghcWithPackages
<deltasquared>
this is going swimmingly...
<Twey>
GHC doesn't have a means to pass in dependencies not in the registry :(
dermetfan has joined #nixos
<deltasquared>
Twey: ghcWithPackages, so where exactly do I specify that...
<Twey>
deltasquared: I think the way most people like to do it is to have a separate default.nix/shell.nix per project with the dependencies they want, and maybe an expression in ~/.nixpkgs/config.nix's packageOverrides
<Twey>
(for a commonly used interactive/non-project environment)
<deltasquared>
hrm, that rather complicates using nix's ghc(i) as a drop-in for arch's one for random users on the system...
<deltasquared>
evidently that setup was rather fragile
<Twey>
deltasquared: On NixOS you can use environment.systemPackages, but without NixOS I don't think there's a good way to install something for all users
<__monty__>
deltasquared: I still do the all of the packages any of my projects need in one big ghc env.
boogiewoogie has joined #nixos
<deltasquared>
Twey: well seeing as I'm just doing *cabal install*, I just did export GHC_PACKAGE_PATH="/usr/local/cabal/ghc/x86_64-linux-8.4.3/package.conf.d/:";
<Twey>
deltasquared: I guess you could explicitly add a build root somewhere under / using nix-store --add-root?
<celeron55>
can i just replace system.build.installBootLoader with my own code
<deltasquared>
I needed to update said script
<deltasquared>
it's pointing at old version
<celeron55>
in configuration.nix
<Twey>
deltasquared: Ahhhh that's a nice approach! :)
<deltasquared>
Twey: so yeah, nix's ghc and cabal, from there on it's locally compiled haskell packages
<deltasquared>
I'm told haskell favours the source approach anyhow.
<deltasquared>
(note to self, all my other source is probably broken until I re-fetch cabal packages...)
<deltasquared>
yas it works perfectly now
<deltasquared>
problem solved :>
<deltasquared>
(famous last words)
<deltasquared>
my janky setup gets to hang together for another day
<Twey>
deltasquared: GHC does, yeah
<Twey>
deltasquared: I'd say per-project dependencies are probably nicer in general though
elgoosy has joined #nixos
<Twey>
deltasquared: Stack also has Nix integration
<deltasquared>
Twey: I guess I'd need nix-daemon for that?
<deltasquared>
so you could "talk" to the nix store to request packages
<deltasquared>
Twey: also re: project specific deps, well, I guess I write small enough things in haskell that I haven't bothered writing proper project files yet.
<deltasquared>
back in a bit, laundry
<elgoosy>
hi, i've just installed nixos. im in login screen with the only user i defined in configuration.nix. However i still haven't done passwd to that user which is why i cannot login. If i hit ctrl alt f1 i get black screen and am unable to write
sondr3 has joined #nixos
Makaveli7 has quit [Remote host closed the connection]
<symphorien>
celeron55: doesn't boot.loader.grub.extraPrepareConfig solve your problem ?
<tilpner>
elgoosy: Login as root, then "passwd youruser"
Makaveli7 has joined #nixos
Makaveli7 has quit [Max SendQ exceeded]
<symphorien>
something like "sed -i s/linux/linuxefi/ /boot/grub/grub.cfg"
<tilpner>
elgoosy: If your login manager doesn't allow logging in as root, ctrl+alt+f2
<elgoosy>
tilpner, i cant. this 'by default' login screen of nixos only shows me one user of mine
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jacereda has quit [Ping timeout: 245 seconds]
<sondr3>
does anyone have any examples of using the Rust overlay in a default.nix file? All the ones I see are fetching the overlay directly but I want to use my local one
zupo has joined #nixos
<celeron55>
doesn't work, maybe the file doesn't exist at that point in the process
<dramforever>
elgoosy: nixos-enter
<dramforever>
Basically 'chroot into nixos mountpoint'.
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
<Myrl-saki>
Well crap. I just found something that works on qemu-system-x86_64 but not on qemu-kvm.
<elgoosy>
dramforever, that's nice, thanks!
<symphorien>
celeron55: I would patch grub with nixpkgs.overlay as I said above, then
asymmetric_ has joined #nixos
asymmetric has quit [Ping timeout: 246 seconds]
FRidh has joined #nixos
<emily>
Myrl-saki: now I'm curious how it behaves on bare metal
<Myrl-saki>
emily: It works under VirtualBox too.
<emily>
with software or hardware virtualisation?
<Myrl-saki>
Hm.
<emily>
I guess VB only does the latter for x86_64
<tilpner>
elgoosy: Sorry, was gone. Did you get it working?
knupfer1 has joined #nixos
<Myrl-saki>
With VT-x/AMD-V
knupfer has quit [Ping timeout: 250 seconds]
knupfer1 is now known as knupfer
xkapastel has quit [Quit: Connection closed for inactivity]
<Myrl-saki>
emily: OTOH, seems like x86-64 *requires* hardware virtualization under VirtualBox.
<deltasquared>
Twey, dramforever: so after all that I finally managed to make a hello world using wai and warp. https://ptpb.pw/Klj1.hs ... I think I'll take the rest of the day off
<emily>
Myrl-saki: right
xkapastel has joined #nixos
<emily>
what are you doing, ooc?
arjen-jonathan has quit [Ping timeout: 250 seconds]
acarrico has joined #nixos
jacereda has joined #nixos
<Myrl-saki>
emily: I needed a rewirtable NixOS image.
<Myrl-saki>
...
<Myrl-saki>
Crap.
<Myrl-saki>
Does make-disk-image.nix do that?
<celeron55>
symphorien: can i patch nixos itself if i want to? i mean, to patch install-grub.pl
<symphorien>
clone nixpkgs
<symphorien>
edit the file
<Myrl-saki>
AaaaAAaa even the source code of make-disk-image.nix is pretty similar.
<symphorien>
if you update nixpkgs and not the module you can have problems
<symphorien>
this is too invasive to be carried long
<symphorien>
this is why patching grub is in my opinion the cleanest and easiest way
<celeron55>
kind of lame, but i guess it's patching grub then
<symphorien>
hum
<symphorien>
now that I think of it
<symphorien>
it won't work
<symphorien>
grub is not used to generate the configuration, is it ?
<celeron55>
oh, you're right
<celeron55>
it's the module that generates it, and fedora's grub that reads it
<celeron55>
the module installs the perl script and a wrapper shell script, and nixos-rebuild uses the installed script to generate grub.cfg
<celeron55>
nixos-rebuild could have a post-grub.cfg-hook but i doubt that
petrkr has joined #nixos
<Myrl-saki>
Is it possible to pad the start of a file?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<celeron55>
the script calls $extraPrepareConfig just before it writes $conf to grub.cfg
<celeron55>
this is silly
<celeron55>
install-grub.pl that is
<symphorien>
it is meant to copy files
<symphorien>
apparently nobody had you use case before
aanderse has quit []
<DigitalKiwi>
so i was able to package ocrmypdf the other day and i had to package a few of its dependencies too and there was one that was already in nixpkgs but it is a lower version than was needed so i had a few questions about making it proper and maybe submit as PR if wanted?
<celeron55>
if grub had an alias system, i could alias linux to linuxefi in grub.cfg before the entries, but i don't think it does
<{^_^}>
nix#711 (by vizanto, 3 years ago, open): nix-shell command line option to print target derivation to stdout
<wavewave>
yup. thanks!
<softinio>
so I am following this guide to add more vim plugins to nix: https://nixos.wiki/wiki/Vim if you look at the add plugins section it mentions editing /vim-plugins/vim2nix/additional-nix-code this does not exist. Anyone know if this has changed to be something else or how I would use the last bullet point in that doc?
<wavewave>
hmm. `--print-derivation` was not implemented :-P
rcshm has joined #nixos
<wavewave>
ah.. now i got it. i can use `nix-instantiate` to create drv and `nix-copy-closure` with `--include-outputs`
<wavewave>
yes. it worked!
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
acarrico has quit [Ping timeout: 245 seconds]
Guest85472 has quit [Read error: Connection reset by peer]
vidbina has joined #nixos
agander_ is now known as agander
<Orbstheorem>
Hello o/ How can I see non-haskell dependencies of a haskell package?
<wavewave>
`nix-store -q --requisites (package in nix/store)` will show all the deps?
<WilliamHamilton[>
srhb: hi, I have a question on workflow on developing haskell packages: when developing package A, I generate the default.nix and shell.nix for my package with cabal2nix, but then manually patch A's default.nix to depend on a local package B instead of the hackage version. At some point I want to add some other dependencies in A's cabal file. Do I have to use cabal2nix again and patch default.nix manually again or is
<WilliamHamilton[>
there a better way?
<Orbstheorem>
wavewave: that gives me a very long list xD any way I could get only the top-level deps? x)
<wavewave>
Orbstheorem: eh. depends on the def of top-level deps. :-P
<WilliamHamilton[>
Orbstheorem: I'd find the derivation in nixpkgs
<wavewave>
yeah. probably that's more what you want.
<Orbstheorem>
Let's start over x) I want to use stack to build my xmonad configuration. afaik xmonad has no direct native requires, but its dependencies do
<WilliamHamilton[>
Orbstheorem: so you are not using nixos, correct? You are using nix in some other OS?
<Orbstheorem>
So when I run stack build, I get configure errors because of missing libraries
<Orbstheorem>
I am using nixos
<WilliamHamilton[>
Orbstheorem: so why are you trying to build xmonad with stack? Let me paste some configuration
<Orbstheorem>
I have xmonad in my configuration.nix
<Orbstheorem>
and I know I can add additional packages
<Orbstheorem>
but I want to change to stack because my current config does not build on non-nixos systems
<[Leary]>
Don't you just want to give stack a shell like `haskellPackages.shellFor { packages = p: [ p.xmonad p.xmonad-contrib ]; }`?
_d0t has joined #nixos
<wavewave>
oh. TIL `shellFor`!
<wavewave>
great.
<Orbstheorem>
Oh!
<Orbstheorem>
Yeah
<asymmetric>
does anyone know how to edit connections settings with iwctl?
<[Leary]>
I'm pretty sure that will work with cabal anyway, I don't think stack uses the haskell deps it's given though.
ddellacosta has quit [Ping timeout: 255 seconds]
<_d0t>
Could you guys help me with ZFS again? People on #zfs seem to be idling. I replace a disk with zpool replace, and now the pool fails to import.
ddellacosta has joined #nixos
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
<[Leary]>
Oh, you might need to manually inherit nativeBuildInputs from (xmonad/xmonad-contrib).env to make `shellFor` work properly... In that case I'd actually just try haskellPackages.xmonad.env first.
<Orbstheorem>
I get “I don't know how to install GHC on your system configuration, please install manually
<Orbstheorem>
when I run stack build
<Orbstheorem>
I told it to use nix to build the derivation and specified shellFor in a shell.nix
<Orbstheorem>
(I also added shell-file: shell.nix)
johanot has quit [Quit: WeeChat 2.2]
mmlb has quit [Ping timeout: 245 seconds]
<softinio>
just created my first PR for nixpkgs. Do I tag/assign anyone to review it or do they automatically get looked at?
mmlb has joined #nixos
<infinisil>
softinio: Mostly automatically, but this might take a while. It's a good idea to ping relevant people by looking at who is maintainer or who touched the files previously
<infinisil>
(by automatically I mean that people just stumble through PR's, there's nothing that actually automatically assigns people to PR's)
<asymmetric>
in order to disable mutableUsers, should i define a root user in configuration.nix ?
<infinisil>
Izorkin: Not sure what there is to check, delroth closed it, the change is very ugly, and I don't have the motivation to rework it, I don't even know if it could be done better. I know running things as root isn't good, but I'd rather do that than have ugly workarounds
<softinio>
infinisil: just saw the failure and your comment. How can I run the checks locally before pushing?
dwdr has quit [Ping timeout: 250 seconds]
<infinisil>
softinio: In your case you just need to build the vim packages you added (I'm surprised you didn't notice that failure earlier)
<_d0t>
is there a way to run a shell from nixos initrd?
<Izorkin>
infinisil: is there any alternative way?
<infinisil>
Izorkin: I just said "I don't even know if it could be done better"
<samueldr>
_d0t: those are the kernel commandline nixos' stage-1 understands
winem_ has joined #nixos
<infinisil>
Izorkin: How does it solve the privileged port problem? delroth mentioned: "The reason why it needs a wrapper is that for some reason the nginx config test checks if ports can be listened on, and this fails for privileged ports without the proper cap"
rcshm has quit [Read error: Connection reset by peer]
equivrel has joined #nixos
rcshm has joined #nixos
<_d0t>
samueldr: thank you!
<Izorkin>
infinisil: code is removed using security.wrappers.nginx and ${pkgs.su-exec}/bin/su-exec ${cfg.user}:${cfg.group} /run/wrappers/bin/nginx -c ${configFile} -p ${cfg.stateDir} -t
<Izorkin>
or option - completely remove code "nginx -t"
<infinisil>
Izorkin: What code is removed where exactly?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Izorkin>
infinisil: minus - no configuration check. But you can add it manually services.nginx.preStart
<infinisil>
Oh, didn't see that, that's not very good then actually
<infinisil>
Well, I guess the config check doesn't really help all that much anyways, because if it's a bad config, the server is going to fail at runtime anyways
<infinisil>
What would be better is to have a build time config check
<Izorkin>
infinisil: nginx -t runs as root, and changes permissions i files.
s2mitrov has joined #nixos
<Izorkin>
because of this, we tried to add security.wrappers.test-nginx
<s2mitrov>
what would be the easiest way to install chunkwm (macOS tiling WM) through nix, but without adding/setting up nix-darwin? I thought maybe I could get away with just adding the nix-darwin channel but `nix search chunkwm` doesn't give me any results
<__monty__>
s2mitrov: Maybe you can specify the channel with -I or -f '<darwin>'?
arjen-jonathan has quit [Ping timeout: 258 seconds]
johanot has joined #nixos
matijja has joined #nixos
<s2mitrov>
oh
<s2mitrov>
hmm
johanot has quit [Client Quit]
<s2mitrov>
i think it might just be the folder structure? nix search -f '<darwin>' does give me darwin.installer and darwin.uninstaller
kuznero has joined #nixos
<__monty__>
nix search -f '<darwin>' chunkwm, it's just to indicate what channel to search. I don't know if nix search searches all your channels, I'd be surprised if it did tbh.
<kuznero>
Hi All!
rcshm has quit [Read error: Connection reset by peer]
<s2mitrov>
there's no chunkwm unfortunately
johanot has joined #nixos
rcshm has joined #nixos
<s2mitrov>
ls ~/.nix-defexpr/channels/darwin has two main folders - pkgs (which contains darwin-installer/uninstaller) and modules which has things like chunkwm skhd etc
o1lo01ol1o has quit [Remote host closed the connection]
xkapastel has joined #nixos
<s2mitrov>
is there any reason chunkwm isn't in nixpkgs? do they not want OS specific stuff, or?
o1lo01ol1o has joined #nixos
<kuznero>
I scaffolded simple library+executable haskell project with cabal. Then cabal2nix generated nix file. I put simple default.nix and shell.nix after that. I opted to using integer-simple.ghc844 by default. Then running nix-build or nix-shell eats up all available memery and make my KDE hang... Any ideas? Here is the project reference: https://github.com/kuznero/toggler
<charukiewicz>
What is the "Wrong ELF type" warning that I see at the end of my Haskell build sometimes? How can I solve the underlying issue?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<_d0t>
is there a way to rebuild initrd only?
<_d0t>
after I changed hardware-configuration.nix for example
rcshm has quit [Read error: Connection reset by peer]
o1lo01ol1o has joined #nixos
rcshm has joined #nixos
zupo has joined #nixos
s2mitrov has quit [Quit: Lost terminal]
vidbina has quit [Read error: Connection reset by peer]
<kuznero>
I was simply following haskell section from nixpkgs manual
<kuznero>
;(
vidbina has joined #nixos
o1lo01ol1o has quit [Ping timeout: 255 seconds]
<kuznero>
perhaps it mixed up lib/exe into same /src in cabal file
<kuznero>
and recurses somehow
ted_ripper has joined #nixos
o1lo01ol1o has joined #nixos
ted_ripper has quit [Client Quit]
jfroche has joined #nixos
<equivrel>
Hi, I think there's something wrong with how the nixos installer (or possibly nixops) handles systemd's "predictable interface names"
<kuznero>
nope, checked with single exe cabal - same effect. nix-shell eats up all available memory - ~15GB
<equivrel>
my interfaces are eth0, eth1, eth2, etc. are install time, and so my configuration gets populated with stuff like "interfaces.eth2 = { ..."
<equivrel>
however systemd later renames my interfaces to things like enp98s0f1 and then later when trying to restart the interface service I get an error saying that the interface doesn't exist.
<equivrel>
And worst of all, this is happening with a remote dedicated server (provisioned with nixops), so that when I reboot the system I lose network access and can't get back in, forcing me to reprovision it and start over.
rcshm has quit [Read error: Connection reset by peer]
rcshm has joined #nixos
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<__monty__>
kuznero: I suspect it's because it needs to compile ghc.
<equivrel>
ottidmes: the problem is I can't quite figure out if this is an issue with the nixos installer or with the hetzner nixops backend...
<equivrel>
srk: do you happen to know if this is equivalent to `networking.usePredictableInterfaceNames = false;`
<srk>
sounds like it!
<kuznero>
__monty__: not sure I understand. Will that be different without integer-simple version of ghc?
<equivrel>
Unfortunately it seems to make matters worse: since I have three network interfaces on this machine, I think this leaves their naming up to chance, and when I tried it I wasn't able to connect to the machine again...
<srk>
equivrel: not exactly (according to nixpkgs)
<srk>
equivrel: you need to create udev rules
<ottidmes>
I indeed have boot.kernelParams = [ "net.ifnames=0" "biosdevname=0" ]; and networking.usePredictableInterfaceNames = false; and some udev rules that makes the interfaces names stable (explicit MAC addresses coupled to interface names)
<kuznero>
__monty__: it is stuck on [evaluating file '/nix/store/sxrlyz64dmvm7spzv141lqg65avg4w1w-nixos-18.09.2247.3276e04a39e/nixos/pkgs/desktops/gnome-2/default.nix] when nix-shell
<equivrel>
ottidmes: and in addition to that you need to set networking.usePredictableInterfaceNames = false;, right?
<srk>
looks like it, no experience with that backend tho
<__monty__>
kuznero: No idea tbh. I thought you overrided ghc844 to get integer-simple support.
<ottidmes>
equivrel: yep
<srk>
not quite .. udev rule contains ENV{net.ifnames}=="0", GOTO="net_name_slot_end"
<srk>
it probably respects kernel cmd line
<ottidmes>
equivrel: this way I can just put it in my generic code, and whenever I want networking.usePredictableInterfaceNames = false; for a host, I want to run that bit of config
<averell>
i have a related question, i have initrd networking, which breaks predictable names because the interface is in use. how can i fix that?
<equivrel>
ottidmes: and you are reading the MAC addresses and adding them to udev rules automatically right? That sounds better than having to write them in manually...
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ottidmes>
equivrel: yep, it reads it from your networking config as you can see, so its automatic
<ottidmes>
averell: you have to reset your interface, its possible, I have seen people do it, but I kinda like having eth0, eth1, etc. and by binding them to a mac address, they are as stable as the predicatable interface names that are being generated, so thats how I worked around it
johanot has quit [Ping timeout: 246 seconds]
Thra11 has joined #nixos
Akii has quit [Remote host closed the connection]
Akii has joined #nixos
<averell>
hm. maybe i'll do it like that und just use the predictable names.
ij has joined #nixos
rcshm has quit [Read error: Connection reset by peer]
rcshm_ has joined #nixos
reinhardt has joined #nixos
vidbina has quit [Ping timeout: 255 seconds]
<Izorkin>
infinisil: created PR #56255 Please check)
<equivrel>
ottidmes: feels like a silly question but how do you actually import that snippet? :>
<ottidmes>
equivrel: I have it in a config = mkMerge [ ... (...the snippet ...) ... ]; but you could also just put it in a seperate file and have it there under config = ... the snippet ...;
<equivrel>
ah, because of the additive nature of = in nix?
jtojnar has joined #nixos
boogiewoogie has quit [Remote host closed the connection]
rcshm_ has quit [Read error: Connection reset by peer]
<ottidmes>
equivrel: thats part of the types in the NixOS module system, so whether = merges depends on the type, but the config overall merges, yes
rcshm has joined #nixos
_kwstas has quit [Quit: _kwstas]
<immae>
Hey there! I have a derivation of the form https://paste.ee/p/yF8Zy that I don’t have control on, and I would like to obtain the foo-config (as a string is fine) from the foo variable only (because foo-config is not "exported"), is there a way to do that?
elgoosy has joined #nixos
elgoosy has quit [Remote host closed the connection]
<tilpner>
immae: Only if you're fine with doing string-parsing
lekstowsky has joined #nixos
<immae>
That’s how I’m doing it yes :D , but I wondered if it had the dependency somewhere hidden and accessible
<equivrel>
ottidmes: definitely another noob question but, where do I get lib.filterAttrs from? I don't have a lib anywhere in this function...
<__monty__>
equivrel: lib is part of nixpkgs. So if you import <nixpkgs> ... it should be in scope.
<__monty__>
*with import <nixpkgs>...
<immae>
tilpner: hmm by string-parsing, you mean pattern matching on the file content right? Or is there a better way?
<tilpner>
immae: With this new builtin you might be able to find foo-config
rcshm has quit [Read error: Connection reset by peer]
<tilpner>
But it's not in any released Nix version yet
<immae>
ok
<equivrel>
ottidmes: ah got it, thanks
<immae>
I’ll have a look at it thanks
rcshm has joined #nixos
<__monty__>
ottidmes: No idea, I often see with import <nixpkgs> {}; at the top of expressions.
<immae>
I’m fine with hackish solution but builtins.match is a bit too much :p
<tilpner>
immae: You can also do a bunch of splits and index by a constant. But that's not really better
<immae>
yes
<immae>
I could also just match /nix/store/*foo-config in my case, but I don’t like it at all (and specific to /nix/store prefix) :p
<tilpner>
You can get the current */nix/store* as builtins.storeDir
rcshm has quit [Read error: Connection reset by peer]
<immae>
oh
rcshm has joined #nixos
<immae>
That’s better than nothing!
<immae>
thanks
<tilpner>
But what is this expression you have no control over?
acarrico has joined #nixos
lekstowsky has quit [Remote host closed the connection]
<ottidmes>
immae: just saying, so you know your options, you could also use scopedImport or libredirect to replace the file to make it export the thing you want or whatever changes you want
rcshm has quit [Remote host closed the connection]
rcshm has joined #nixos
<immae>
ok, thanks both, I think I’ll find my way with that
<ottidmes>
with scopedImport, Nix becomes a whole lot more flexible, but also slower, since it does not do caching for scopedImport
<immae>
Yes, I saw your comment about that
freeman42x has joined #nixos
<stepcut>
I am trying to use the ssh-ng:// substituter. When it tries to download the packages it often dies with 'writing to file: Broken pipe'. But if I try it more times, it will eventually work :-(
<stepcut>
running nix-copy-closure by hand doesn't seem to fail though -- so I don't think it is a connectivity issue
Thra11 has quit [Ping timeout: 244 seconds]
Guest27871 has quit [Ping timeout: 246 seconds]
<equivrel>
ottidmes: it looks like it wasn't that simple. Reinstalling the system with your snippet and it loses connectivity after reboot :(
winem_ has quit [Ping timeout: 264 seconds]
DrLambda has joined #nixos
matijja has quit [Remote host closed the connection]
<ottidmes>
equivrel: and with networking.usePredictableInterfaceNames = false; and with mac address explicitly configured?
<{^_^}>
[nixpkgs] @Profpatsch opened pull request #56258 → skawarePackages.execline: wrap execlineb with tools → https://git.io/fhbte
ddellacosta has quit [Ping timeout: 250 seconds]
<equivrel>
ottidmes: I had usePredictableInterfaceNmaes = false; yes. are you suggesting I try it again but with explicit udev rules by mac address?
ddellacosta has joined #nixos
sakalli_ has joined #nixos
<ottidmes>
equivrel: no, all I am saying is that macInterfaces != {} check is probably failing for you. Did you set the macAddress on config.networking.interfaces
<equivrel>
ottidmes: oh I actually removed the check and just used the body.
<ottidmes>
same problem, the check is there for a reason, if its empty, it wont write any mac addresses, and it cannot determine them automatically, they have to be stated in the config somewhere, so I used the config you normally use to set it
<equivrel>
ah, I see
<equivrel>
So I do need to set the mac address of each interface manually after all?
<sakalli_>
hi, just started to get a bunch of these warnings while building. "/etc/cups directory contains user files. Symlinking may fail. at /nix/store/amhpx347b7n6m2bq24siw0qq27w3np8n-setup-etc.pl line 114." not exactly sure how i created the problem. any ideas?
<ottidmes>
equivrel: yeah sorry, when you mentioned automatic, I thought you mentioned do anything special outside setting that (that option to false and setting mac addresses), within Nix, without being impure, you simply cannot determine that kind of info
<{^_^}>
[nixpkgs] @dtzWill opened pull request #56259 → minetime: init at 1.4.10 → https://git.io/fhbtY
<ottidmes>
sakalli_: you, or some program, probably wrote some stuff to /etc/cups, while NixOS also wants to symlink to that path, probably something like that
<sakalli_>
so the folders are cups, dbus-1, terminfo, zoneinfo, fonts, kdb, systemd/system, systemd/user, sustemd/system-generators, udev/rules.d, dovecot/modules
rauno has joined #nixos
<sakalli_>
any suggestions how i could fix it.
ixxie has joined #nixos
<symphorien>
if you don't have configured your printers yet, you could just remove /etc/cups
orivej has quit [Ping timeout: 245 seconds]
<sakalli_>
two thins i did: removed packages from elpa folder to let use-package to reinstall them. and then did nixos-rebuild from userspace by mistake. perhaps the latter is the culprit?
<sakalli_>
just removing cups is ok, but there are those other folder as well.
<equivrel>
ottidmes: I see. As a matter of fact this is a bit confusing since it looks as if two of my interfaces were renamed from the same thing... So I'm not even sure how to name them now: https://gist.github.com/livnev/8fea5e78c663fb86bbd52314e6745820
orivej has joined #nixos
<ottidmes>
equivrel: just tested on my desktop, same command gives me no renames, so it much be something additionally being done in your config
<gchristensen>
goibhniu: ping?
<sakalli_>
how does this work, can one remove stuff from those folders and let nixos rebuild them?
<ottidmes>
equivrel: almost seems like your config did not run with that snippet you showed, because its still doing predictable naming
<ajs124>
I'm having a kind of stupid problem with the linux kernel right now. What is the best way to git bisect the kernel on a nixos system?
<equivrel>
ottidmes: sorry for the confusion, that output was from before I added the snippet. Since I added the snippet I haven't been able to connect to the system so I can't tell you what's in there.
<equivrel>
I was just investigating that old output since I wanted to know what to name my interfaces...
sakalli_ has quit [Remote host closed the connection]
<ottidmes>
I always have my main ethernet connection as eth0, as that is what you'd expect, and I tend to just disable the others if I don't use them anyway, can only cause confusion
<clever>
bsima: line 57, preStart gets ran as root
<bsima>
oh, i didn't know what that line did, thanks clever
ixxie has quit [Ping timeout: 244 seconds]
Glider_IRC_ has joined #nixos
reinhardt has quit [Quit: Leaving]
_kwstas has joined #nixos
Glider_IRC__ has quit [Ping timeout: 245 seconds]
<lunabo>
Is it normal that adding scipy to my global Python packages causes it to build scipy from source? Shouldn't it use a prebuilt version somehow?
juliendehos has joined #nixos
<lunabo>
Ah at least it doesn't rebuild every time I add a Python package, then it's something I can live with :)
<{^_^}>
[nixpkgs] @FRidh pushed 44 commits to staging-next: https://git.io/fhbqT
<{^_^}>
[nixpkgs] @FRidh closed pull request #55962 → buildPythonPackage: initial support for PEP 517 → https://git.io/fh5DB
<bsima>
lunabo: nix will only rebuild packages like scipy if it can't find a compiled version in the nix cache
<bsima>
cache.nixos.org i believe
eadwu has joined #nixos
wfranzini has quit [Remote host closed the connection]
<lunabo>
bsima: yes, I sort of expected scipy to be in the cache, but maybe it just wasn't there. I was also thinking it could be because it is a "subpackage" of the Python interpreter, but I don't understand things well enough to know if that is the issue
winem_ has quit [Remote host closed the connection]
xkapastel has quit [Quit: Connection closed for inactivity]
<equivrel>
ottidmes: hm so I tried assigning networking.interfaces.eth0.macAddress from eth0 to eth3 but still lose connection...
Cale has joined #nixos
<equivrel>
to be honest, I still don't understand why it's necessary to manually assign names to these interfaces
neminis has joined #nixos
zupo has joined #nixos
<ottidmes>
its not necessary, normally you dont have the installer generate network config for you, so you just used systemd predicatable names, in my use cases I have network needed in initrd, so then its easier to just use the old naming of eth0. I have no experience with your host, but I would try and keep things as simple as possible. You don't have a VNC connection to the server, all my VPS providers do, regardless
<ottidmes>
of internet connection of the host itself
wedens has quit [Quit: Connection closed for inactivity]
<averell>
is there a way to specify zfs pool id for the import job on boot? i'm getting "importing ZFS pool "data"...cannot import 'data': more than one matching pool"
<averell>
(there is only one pool named such)
otulp has quit [Remote host closed the connection]
matijja has joined #nixos
camsbury has joined #nixos
winem_ has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
camsbury has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @buffet opened pull request #56264 → creep: init at 0.31 → https://git.io/fhbmO
<tokudan[m]>
averell: if you enter "zfs import", then you should get a list of pools that can be imported and their status. you can import a specific one by specifying the ID that's shown
<{^_^}>
[nixpkgs] @aanderse opened pull request #56265 → replace deprecated usage of PermissionsStartOnly → https://git.io/fhbm3
<clever>
adamantium: if you create a /etc/NIXOS_LUSTRATE and reboot, then the initrd will move EVERYTHING in / to /old-root/
<clever>
so it will practically be a fresh install
<{^_^}>
[nixpkgs] @arianvp closed pull request #47177 → nixos-manual: Include all modules in documentation generation → https://git.io/fAd6v
<clever>
i think you can also `echo /home > /etc/NIXOS_LUSTRATE` to make it not move /home, and similiar for other paths you care about
<clever>
and since its only a mv, you can undo anything thats important
aw has quit [Remote host closed the connection]
spacefrogg_ has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @catern opened pull request #56267 → perlPackages.NetServerSSPrefork: init at 0.05 → https://git.io/fhbmB
agander_ has joined #nixos
aw has joined #nixos
ddellacosta has quit [Ping timeout: 250 seconds]
spacefrogg has joined #nixos
agander has quit [Ping timeout: 244 seconds]
matijja has quit [Ping timeout: 268 seconds]
<adamantium>
clever: ok clever, i will try echo /home > /etc/NIXOS_LUSTRATE
o1lo01ol1o has joined #nixos
ddellacosta has joined #nixos
johanot has quit [Quit: WeeChat 2.2]
_kwstas has quit [Quit: _kwstas]
ddellacosta has quit [Quit: WeeChat 2.2]
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<{^_^}>
[nixpkgs] @catern opened pull request #56268 → pythonPackages.pykerberos: take krb5 as a nativeBuildInput → https://git.io/fhbmP
zupo_ has joined #nixos
<adamantium>
clever: i did the lustrate, i rebooted, it did what it was supposed to do, preserved my /home /etc/nixos and /nix-config directories ... and created /old-root. But I still get the same warnings when running nixos-rebuild switch https://gist.github.com/f4987e80ca12594cfd9d164a8db06cde
zupo has quit [Ping timeout: 255 seconds]
<adamantium>
I mean, it totally worked, even my password was reset. The problem is still here though. So i'm confused
<ryantm>
worldofpeace: I'm curious how you manage your workflow where you re-write commits of people and force push them to GitHub. I notice it says the commits are authored by you and the other person. Do you have to add their fork as a remote in git?
<clever>
adamantium: probably a bug in nixos then
<adamantium>
nothing to worry about probably ? :D
<clever>
adamantium: yeah, just ignore it till somebody else fixes it
<adamantium>
kthx
<adamantium>
That was pretty amazing though. Thanks for that bit
<adamantium>
I thought it would take a long time, but it was fast.
<clever>
yeah, mv is atomic, so its basically instant
<NickSeagull>
Hey all, has somebody disabled auto screen blanking after 10 mins on NixOS? I've tried xserver.sessionCommands with `xset s off` and `xset s off` by itself and nothing :(
<tilpner>
NickSeagull: You could try setting services.xserver.serverFlagsSection
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « apt-dater: 1.0.3 -> 1.0.4 (#55901) »: https://git.io/fhbYO
<infinisil>
NickSeagull: When did it stop working by default?
sigmundv has joined #nixos
<NickSeagull>
What do you mean infinisil ?
<infinisil>
Ohh
xkapastel has joined #nixos
theseriousadult[ has joined #nixos
<infinisil>
I interpreted "has somebody disabled auto screen blanking after 10 mins on NixOS?" as you updated NixOS and it stopped working with that, and you want to turn it on again
ddellacosta has quit [Ping timeout: 250 seconds]
benkerry[m] has joined #nixos
<NickSeagull>
Oh no, I just didn't know how to do it properly ^_^"
juliendehos has quit [Quit: Leaving]
<infinisil>
Yeah got it, nvm then
<tilpner>
dtz: appimageTools was merged, and it's not the same commit as the one your branch is based on
<tilpner>
dtz: I don't know if that'll work. To be sure, you could reset to master, and cherry-pick your commits on top
NickSeagull has quit [Client Quit]
<clever>
tilpner: thats the same as `git rebase origin/master`
agander_ is now known as agander
<tilpner>
clever: I remember trying to rebase, but I ended up rewriting a few commits
<tilpner>
So I probably did something wrong
<clever>
it can have trouble if merge conflicts exist
<tilpner>
All the commits were still there, but they were suddenly mine
<clever>
ah, yeah
<clever>
it will change the headers in the commits, to say you where rewriting history
<clever>
and it will re-sign them with your gpg, because you lack the authors gpg
<ryantm>
ddellacosta: scanning it briefly, it looks like it should still work. What are you trying to do though?
<ddellacosta>
ryantm: oh basically exactly what is in that blog post, in the sense of create my own custom package repo that I can add scripts and whatnot to, and be able to refer to easily in my configuration.nix file and elsewhere
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « tvheadend: 4.2.7 -> 4.2.8 (#54038) »: https://git.io/fhbYw
<ddellacosta>
rather than a one-off package I add to the root profile or something uglier like hard-code paths into my configuration.nix file, I figured it would be cleanest to just keep my own private repo of custom scripts
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to staging « securefs: 0.8.2 -> 0.8.3 (#54671) »: https://git.io/fhbYi
<ddellacosta>
or would the suggested approach rather be something like (if I'm talking about just some scripts) using bash as the base and then overlaying custom scripts on top of that?
<ryantm>
ddellacosta: You can add packages. In 12.2 the first example adds a package called rr
<ddellacosta>
ah okay, thanks, will read it more deeply then
<ddellacosta>
I still feel like I haven't quite gotten the nix mindset...hopefully after fumbling around for a while with this stuff it'll click
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « stubby: 0.2.4 -> 0.2.5 (#54674) »: https://git.io/fhbYX
<ryantm>
ddellacosta: If you look at the nixpkgs repo, there is a file called pkgs/top-level/all-packages.nix , and when you use an overlay, it's basically like you are adding stuff to that giant attrset
o1lo01ol1o has joined #nixos
<ddellacosta>
I see
<ryantm>
or editing stuff in there :)
<dtz>
<freenode_til "dtz: appimageTools was merged, a"> good to know, I'll rebase accordingly. (and sort out whatever conflicts may be involved). Thanks for the heads up :)
<ddellacosta>
that's great, those kinds of tips are super helpful, thanks ryantm
<{^_^}>
[nixpkgs] @xeji pushed commit from @r-ryantm to master « cups-filters: 1.21.6 -> 1.22.0 (#54841) »: https://git.io/fhbYM