<nh2[m]>
when I use `fileSystems."/".device = "/dev/disk/by-label/root";`instead of `/dev/disk/by-uuid/...`, my server doesn't boot any more. In the stage-1 rescue shell, `/dev/disk` doesn't exist at all. But how can `/dev/disk/by-uuid/` work when `/dev/disk` doesn't exist?
supersandro2000 has quit [Disconnected by services]
<vikanezrimaya>
Say, if I need to launch a program from a user service that needs an environment maximally close to one I can get by opening a login shell, how would I approach writing it as a NixOS module?
supersandro2000 has joined #nixos
<vikanezrimaya>
Specifically I need code-server and I'm not sure how well it'll work when launched from outside a proper environment
<vikanezrimaya>
(btw code-server is amazing, if you love VSCode you'll love code-server - it's like VSCode but in your browser, saving memory)
<andoriyu>
I get` zsh: no such file or directory: ./sdm` (./sdm is present)
novafacing has quit [Remote host closed the connection]
lovesegfault has joined #nixos
vikanezrimaya has quit [Ping timeout: 240 seconds]
vikanezrimaya has joined #nixos
riksteri has quit [Quit: riksteri]
ericsagnes has quit [Ping timeout: 272 seconds]
<andoriyu>
ahh patch elf
<nh2[m]>
I solved my question above: Indeed `/dev/disk/by-uuid/` doesn't work either on my new system config, so it's something else that breaks it overall.
justanotheruser has quit [Ping timeout: 240 seconds]
monokrome has joined #nixos
<monokrome>
hey y'all - I ran nixos-infect and wondering if anyone has ideas why `nixos-rebuild switch` doesn't seem to read /etc/nixos/configuration.nix here?
<monokrome>
Probably something dumb but I'm a bit confused about it
cript0nauta has quit [Ping timeout: 245 seconds]
<clever>
monokrome: by default, it uses whatever <nixos-config> points to, which relies on $NIX_PATH
<monokrome>
clever: This seems right, but it doesn't seem to build to here? /root/.nix-defexpr/channels:nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixos:nixos-config=/etc/nixos/configuration.nix:/nix/var/nix/profiles/per-user/root/channels
<monokrome>
(for NIX_PATH)
<clever>
monokrome: that looks like the default NIX_PATH and it has the expected nixos-config=/etc/nixos/configuration.nix
<monokrome>
hmm, but if I rebuild then it doesn't seem to even install the packages from that one (http://sprunge.us/JRYk3g)
griff_ has quit [Quit: griff_]
<monokrome>
but if I add a syntax error then it shows the syntax error
<monokrome>
so the environment must just be pointing to the wrong place somehow
<clever>
monokrome: [clever@amd-nixos:~]$ vi ~/apps/nixpkgs/nixos/default.nix
<monokrome>
if I do `python3`, though, it's not found
<monokrome>
(weechat may not be the best example because when I reboot it seems to show up fron the nix-env -iA whatever happened)
<clever>
what if you `nix-build '<nixpkgs/nixos>' -A system && ls -l result` ?
<monokrome>
what does that one do
<clever>
it should be the same as `nixos-rebuild build`
<monokrome>
weird, that command just says "Killed"
<clever>
check dmesg
<monokrome>
omg
<monokrome>
OOM killer
* monokrome
falls over
<clever>
your nixos-rebuild was likely dying the same way
<monokrome>
yep
<clever>
so its never actually rebuilt
<monokrome>
rofl
<clever>
slap some swap on that sucker
<monokrome>
no wonder this was so weird at least
<clever>
how much ram does this poor thing have??
<monokrome>
not enough lol
<clever>
`free -m` ?
nicoo has joined #nixos
<monokrome>
but it seems nixos-infect doesn't generate an fstab w/ swap in it
<monokrome>
so I need to nixos-generate-config again maybe
<clever>
you can add your own too, one sec
<monokrome>
once I mounted the already preexisting swap, it built
<clever>
[root@amd-nixos:~]# man configuration.nix
<clever>
ah, that works too
<monokrome>
512MB RAM
<clever>
and nixos-generate-config will add the swap to hardware-configuration.nix
<monokrome>
omg python3 works now
<monokrome>
thank you
vidbina_ has quit [Ping timeout: 276 seconds]
<monokrome>
a lot >.<
<clever>
yep
<monokrome>
is there another way
<clever>
?
o1lo01ol1o has joined #nixos
<monokrome>
other than nixos-generate-config or manually updating fileSystems
<monokrome>
seems 512MB of RAM should have been enough regardless
<clever>
swapDevices.*.size
<clever>
you can set swapDevices = [ { device = "/path/to/file"; size = 1024; } ];
<clever>
and nixos will auto-create the file and swapon for you
<clever>
if you didnt have a dedicated partition
ahmed_elgabri has quit [Ping timeout: 264 seconds]
<monokrome>
should have probably done it that way
<monokrome>
I wonder if the VPS themselves forgot to enable the swap in their images lol
<clever>
a swap partition is better performance
<clever>
if you have control of the vm, you can create an entire virtual disk just for swap, and then you can freely throw it out and replace it with a bigger one
<monokrome>
yeah, they had a default swap partition (unless nixos-infect created that which I assume not)
<clever>
justdoit and my kexec tool is more powerful then nixos-infect
<clever>
it lets you repartition the disk, and do anything
<clever>
build a magic tarball on one machine, unpack it on the un-suspecting victim, and run /kexec_nixos, and boom, your now running nixos entirely from ram
<clever>
wipe the disk and install as usual, like you used a livecd
<clever>
this module is also installed, it makes the system reboot within an hour, if you fail to ssh into it and gain control
<clever>
once your in, if you wipe the disk, then it obviously cant boot, so if you fail to install it right, its bricked
<clever>
justdoit automates things enough that you cant really mess up
<clever>
the original intent, was to infect remote servers, without any console access to fix mistakes
<monokrome>
I think you underestimate me 😹
amfl_ has joined #nixos
sangoma has joined #nixos
<patagonicus>
Ok, the kexec for switching to NixOS is really nice. When I installed NixOS on my rented server, I spent quite a long time trying to figure out how to get nix installed in a usable way on one of the available distros, then installed a basic nix to an extra partition and then finally used that to install the real nixos on the full disk.
<clever>
patagonicus: the kexec stuff is a multi-stage trick, make-system-tarball.nix generate a tarball with a complete /nix/store containing the closure of anything
<clever>
in this case, its the closure of a shell script, containing kexec, a kernel, and an initrd
<clever>
so you dont even need a working kexec installed on the victim
<clever>
the kernel/initrd itself, comes from the netboot utils, which packages the whole rootfs up as a squashfs inside the initrd
est31 has quit [Ping timeout: 268 seconds]
<patagonicus>
Yeah, I skimmed the nix files. I should really play around with kexec more, never really used it before. But then I don't really plan on installing nixos on any system that I can't just easily boot a NixOS live CD on any time soon.
<monokrome>
so most things seem to work now but I think the system is still unstable
<monokrome>
`mkpasswd -m sha-512 >> /etc/nixos/configuration.nix` and putting the resulting string in hashedPassword doesn't generate a working password
<monokrome>
which seems... messed...
<clever>
monokrome: hashedPassword only works if you have set mutableUsers = false i believe
<clever>
monokrome: there is also initialHashedPassword, but it only works if the user hasnt been created yet
<monokrome>
oh O_o
<monokrome>
I've used hashedPassword w/out mutableUsers = false before
<monokrome>
maybe I am wrong though
queiw has quit [Remote host closed the connection]
<patagonicus>
Documentation says "If the option users.mutableUsers is true, the password defined in one of the three options will only be set when the user is created for the first time."
<viric>
clever: oh 500MB were gzipped. 750MB gunzipped...
skinnedcat[m] has joined #nixos
<CFS>
Hi... I'm (eternally) new to nix. I'm making a small change to a derivation. Specifically, I'm changing a reference to an absolute path (files being copied from a specific place in the HD) to a relative one ( ../..). I thought that would be as trivial as it sounds, but no...
<viric>
how would I find out why the initrd is so big
<CFS>
looks like relative path points to the nix store
mananamenos has joined #nixos
<CFS>
So going from this:
<CFS>
infraRepo = p.config."dev:infra".repo;
<CFS>
dir = "${infraRepo}/ansible/roles/prometheus/files";
<CFS>
to this:
<CFS>
dir = "${../../../dashboards}";
<CFS>
break stuff
<clever>
viric: and it needs 500+750 to perform that whole unpacking
<CFS>
(the dirs have the same contents exactly)
<CFS>
any hint? I'm rather frustrated at this point
<clever>
CFS: how exactly does it break?
<viric>
clever: AH the initrd includes the squashfs root in it
<clever>
viric: and that squashfs will remain compressed, so it shouldnt expand any further
<viric>
clever: but maybe the kernel won't be able to free the initrd
<viric>
because there is the squashfs
<clever>
viric: but you may want to built config.system.toplevel, and get the closure size down
<CFS>
@clever: error: getting attributes of path '/nix/dashboards': No such file or directory
<eliasp[m]>
CFS: do you need help in "how do I customize an expression and make it part of my system" (e.g. when using "nixos-rebuild") or do you need help with this specific issue regarding your expression?
<viric>
I see. I'll try, clever
<clever>
viric: yeah, you need the whole initrd to run, but you can shrink it
<viric>
clever:if I could use a network root
<clever>
CFS: that looks like something did /nix/store/hash-foo/../../dashboards
<CFS>
for reference this is using nix on Ubuntu, not nixOS, and nix is being invoked as part of a large process I don't control (much) so even adding an arg is not totally trival
<viric>
I'll try but I've to go < 500MB unpacked
<CFS>
@eliasp[m], I need help going through this specific hurdle (for now)
<clever>
viric: modify that nix expr, so the symlinkJoin is part of a set, so you access it via `nix-build foo.nix -A netboot`
<clever>
viric: then add a second attr to the set, to expose bootSystem.config.system.toplevel
<clever>
viric: then build toplevel, and run `du -hc --max=0 $(nix-store -qR ./result) | sort -h | tail` on it, and look for fat you can trim
<CFS>
@clever yes, that's clearly what's happening. Of course /nix/store doesn't have a dashboards directory, but it does have a yij69859qxzfb13lnynj8ad0jbldmmhz-dashboards
<CFS>
so the question is how do I get the "nix name"
<clever>
CFS: what config file did /nix/dashboards wind up in? can you pastebin that whole file?
<CFS>
You mean the .nix file?
<clever>
CFS: the config that nix generated
<clever>
CFS: that is being passed to grafana
simba1 has quit [Ping timeout: 260 seconds]
<CFS>
dashboards contain the .json files that were externally copied from somewhere else. But if you tell me how to get to that config file from the .nix file I'll be happy to paste it
mayhew has quit [Remote host closed the connection]
<CFS>
The .nix file looks like this:
<CFS>
dashboardJsonFiles =
<CFS>
dirEntries = builtins.readDir dir;
<CFS>
dir = "${../../../dashboards}";
<CFS>
let
<CFS>
jsonFiles = lib.filterAttrs
<CFS>
(path: type: type == "regular" && lib.hasSuffix ".json" path)
<clever>
viric: 20.09 is just using a stray var in a let block, so you cant force it off, but you can still use disabledModules to just omit zfs support from nixos entirely
<viric>
I'll mkForce the supported fs in boot to []
mkaito has quit [Quit: WeeChat 3.0.1]
virkony has quit [Ping timeout: 245 seconds]
gueorgui has quit [Ping timeout: 276 seconds]
MmeQuignon has joined #nixos
CFS has quit [Remote host closed the connection]
simba1 has quit [Ping timeout: 240 seconds]
simba1 has joined #nixos
CFS has joined #nixos
<viric>
clever: it booted further. "/initrd.image incomplete write" linux says
<clever>
viric: havent seen that before, how big is the initrd? put the gz back?
<CFS>
OK this is working. thanks a lot for your help @clever
<clever>
yep
<viric>
clever: the other day I had a question... how would we use the nix toolchain to produce binaries for a standard sysroot?
<viric>
nixpkgs gcc's I mean
<CFS>
Now I need to decide about my nix life :-) I guess I should just switch to nixOS and embrace it fully and I'm going to be miserable each time I need to change a small thing
ahmed_elgabri has quit [Ping timeout: 272 seconds]
<clever>
viric: the gcc that comes with nix has been heavily patched, to actively reject /usr/lib and to add rpath's into the store
<viric>
clever: exactly
<clever>
viric: you could just build it the nix way, and then use patchelf to un-nix it
mkaito has joined #nixos
mkaito has joined #nixos
mkaito has quit [Changing host]
<viric>
patchelf used to break some binaries, long ago...
<clever>
the FHS env used by things like steam, just symlinks them to look like a normal /usr/bin, but it still has the normal nix RPATH
m0rphism has quit [Ping timeout: 245 seconds]
<clever>
but its also mixing in some libs from debian
<viric>
ah
<clever>
to avoid compatability problems from using the wrong versions
<viric>
I thought all magic was in gcc-wrapper
<viric>
but if gcc is also patched...
<clever>
i think its a bit of both
<viric>
clever: linux didn't like my "xz -T0" image
stampirl has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
<clever>
viric: you would build a config using fileSystems."/" = { device = "UUID=whatever"; iscsi = { enable = true; host="1.2.3.4"; lun = "something"; }; };
<viric>
Ah I see
<viric>
why isn't that on nixpkgs? :)
<clever>
viric: but to create the rootfs itself, you would need to either build a disk image and share that, or connect it to a temporary box, and nixos-install to it
<clever>
viric: so i can expose a zvol to my windows box, and windows just says "sure" and treats it like another block dev, toss ntfs on, and your apps are all happy
<clever>
removing the firmware dropped mine from 1.4gig to 900mb
<viric>
sure, it's big
<isaacr24>
Hi. The test suites for time-compat-1.9.2.2 and GHC 8.6.5 are failing. For whatever reason, when I nix-build and use the `dontCheck` override, the tests go anyway.
<clever>
system.extraDependencies shaved another 100mb off
<viric>
clever: I'm at 770MB
<isaacr24>
I'm a nix novice, any tips would be much appreciated
dev_mohe has joined #nixos
stree has quit [Ping timeout: 245 seconds]
<viric>
isaacr24: what is "that" override?
dev_mohe has quit [Client Quit]
<viric>
clever: I think the problem is more pxe than ram, now
<isaacr24>
it's very odd, since this works for my other packages
<isaacr24>
builder for '/nix/store/g8hnl1zjiq8fsp9gnmgi4crpz1ymv9k9-time-compat-1.9.2.2.drv' failed with exit code 1 cannot build derivation '/nix/store/xvdp9dl6jcaisgr9dsckl2kahdr01k5d-aeson-1.4.4.0.drv': 1 dependencies couldn't be built
<isaacr24>
I think the issue is that the derivation for aeson is choosing `time-compat` on its own, and ignoring that I'm trying to `dontCheck` on it
<clever>
viric: profiles/base.nix looks like more fat to trim...
<isaacr24>
it's also weird that I can change the version of time-compat in default.nix, but it always builds 1.9.2.2
<viric>
clever:
<viric>
clever: I think I give up... every attempt is frutiless
<clever>
viric: sometimes, its simpler to just rip the hdd out, mount it elsewhere, nixos-install, and then boot that back in the target
<viric>
clever: I'll try to get a USB thing
<clever>
isaacr24: assuming default.nix is using the overrides on line 31 right, that should work
<viric>
clever: It looks like should be able to have a 4MB initrd image that just mounts a nixos from a signed http thing
<viric>
Doesn't that exist?
<q3k[m]>
does anyone know of a buildFHSUserEnv alternative (for nix-shelling) that would 'properly' build an ld cache instead of using LD_LIBRARY_PATH to point at target libraries?
<isaacr24>
Not working -- it always builds 1.9.2.2, even when I change it to 1.9.4, for example
<q3k[m]>
ie. that would have a function linker with a fully yeeted env
<clever>
viric: ive not seen any that mount over http(s)
<viric>
some kind of 'verity' signed from the net directly
<patagonicus>
IIRC either Firefox or Chrome or both at some point had to move their x86 builds to amd64 since they wouldn't build on 32 bit systems anymore. So Firefox is probably not tested for building on 32 bit, just that 32 bit Firefox built on 64 bit systems works.
<viric>
so they cross-build?
<clever>
yeah, cross-compile from 64->32 is one simpler option
<clever>
nixos doesnt give you that option right now
<clever>
rather then doing multi-lib or cross, nixos just makes a 100% pure 32->32 env
<viric>
sure
<viric>
I'll have a look
<viric>
Just in case they use a too pretentious linking
<avn>
Just curious -- what mean `Resolved derivation: '/nix/store/dsnhbm2a2sc2waf2x61h3fngw7dllyms-zoom_x86_64.pkg.tar.xz.drv' -> '/nix/store/dg2xvzvrxy8bh05psqagv2hb51rpwgdg-zoom_x86_64.pkg.tar.xz.drv'...`
<avn>
looks like something new
alexherbo2 has joined #nixos
<gchristensen>
huh is that something with CA derivations
<PotentialUser-93>
Hello. How should systemd DynamicUser work with TLS certs? I use NixOS to copy LetsEncrypt certs to /var/certs/matrix but naturally that doesn't work. I'm trying to run matrix-dendrite which is PR109561.
<PotentialUser-93>
Should have been clearer: the issue is permissions, due to DynamicUser I don't know which user accesses the certs so I can't set the permissions manually. When the certs are updated I don't know if dendrite is running so I can't copy the certs to /var/lib/matrix-dendrite (and even then the permissions may be wrong).
<dominikh>
is anyone here familiar with CUPS and foomatic? I'm trying to package https://github.com/philpem/printer-driver-ptouch but it only comes with XML files and I have no idea how to turn them into PPD files, or how to hook them into CUPS and automatic printer detection without turning them into PPD files. I couldn't find any other printer package in nixpkgs that has to do this, either. everyone else
<dominikh>
seems to ship PPD files in their tarballs…
civodul has joined #nixos
zebrag has quit [Quit: Konversation terminated!]
simba1 has joined #nixos
zebrag has joined #nixos
<dominikh>
I'm beginning to think that we don't package enough of foomatic for this
alexherbo2 has quit [Read error: Connection reset by peer]
alexherbo21 is now known as alexherbo2
<prsteele>
Hello everyone. I'm trying to figure out an idiomatic approach to the following problem. I have a valid shell.nix file that I would like to instantiate with one additional dependency. I'm happy to invoke this alternate shell in another file, such as linting.nix. Is there a common pattern for gluing two shell.nix files together?
Rusty1 has left #nixos ["Konversation terminated!"]
stree has joined #nixos
<prsteele>
I could imagine a solution that involves concatenating two imported buildInputs, but I'm not sure if there is anything else I'd need to propagate from each shell
alexherbo24 has joined #nixos
alexherbo2 has quit [Ping timeout: 256 seconds]
alexherbo24 is now known as alexherbo2
proofofkeags has joined #nixos
rajivr has quit [Quit: Connection closed for inactivity]
<ronthecookie>
im working on a rust project that links to libpulse yet ldd cant find libpulse: https://termbin.com/w1jz
<infinisil>
prsteele: That's exactly what the `inputsFrom` argument from the `pkgs.mkShell` function is for :)
<prsteele>
that definitely looks promising, thanks infinisil
<Yaniel>
viric: one of the non-graphical consoles yes
<prsteele>
in that example there are three packages mentioned --- do you know what the semantic difference is between using inputsFrom and buildInputs? Naively I would have expected all three to be in either buildInputs or inputsFrom
<infinisil>
prsteele: inputsFrom takes the buildInputs (and co.) from the listed packages, and inserts them into the buildInputs of the shellFor itself
<infinisil>
So inputsFrom doesn't actually need to build its inputs, it only needs the dependencies
<infinisil>
Which is in contrast to buildInputs, which do need to be built (as do all dependencies of all inputsFrom)
fendor_ has joined #nixos
<prsteele>
okay, I think I have a better mental model now. I think I actually want to shove something in buildInputs (my linter) and inputsFrom (my package derivation that I want to lint)
kenran has joined #nixos
<infinisil>
Hmm not sure
<prsteele>
(concretely I'm trying to get a shell with haskell-language-server built to introspect my package)
<infinisil>
Oh, I guess mkShell probably isn't what you want then actually
<infinisil>
prsteele: How about just adding `withHLS ? false` as an argument to your shell.nix, then using `buildInputs = [ ... ] ++ lib.optional withHLS haskell-language-server`
_bin has quit [Ping timeout: 256 seconds]
fendor has quit [Ping timeout: 265 seconds]
<infinisil>
Could then also have a separate file `hls-shell.nix` just consisting of `import ./shell.nix { withHLS = true; }`
<prsteele>
that also sounds reasonable
<prsteele>
(I can solve this trivially by just making my package depend on hls directly, but that isn't a reasonable thing to do if I want to publish it)
_bin has joined #nixos
<prsteele>
although mkShell appears to be working
higherorder__ has quit [Ping timeout: 240 seconds]
davetapley has quit [Ping timeout: 240 seconds]
<prsteele>
hm, perhaps not --- cabal build is missing some dependencies with the mkShell approach
<prsteele>
thanks infinisil. I'll report back when I find something that works... should probably go update some stackoverflow questions that ask this same quesetion without a satisfactory answer
davidv7 has quit [Remote host closed the connection]
<ambroiseur>
Okay so for future reference, it was because the plugins have been updated for the next version of jellyfin, and the current version does not check if a plugin is compatible with itself
<clever>
viric: ah, rust wasnt even started (i think), when i first had such linker problems
<prsteele>
essentially, swap out mkDerivation for one that injects new dependencies. This allows me to reuse whatever magic is in the Haskell mkDerivation
<clever>
back when i still ran gentoo, and had to compile firefox
<ronthecookie>
sigh, my nix-shell got gc'ed out of the store so when i tried to use a binary that links with something built in it it broke
<ronthecookie>
i have been debugging this for hours
griff_ has quit [Quit: griff_]
<clever>
ronthecookie: run ldd on the binary, then run `nix-store -r` on the relevant missing paths
bhoman has quit [Ping timeout: 246 seconds]
<clever>
if its in the cache, that will re-fetch it
<ronthecookie>
clever: it wouldn't be in the cache, and anyhow i just rebuilt it
<m1cr0man>
Hey folks! I'm trying to convert an openldap config that was using extraConfig options to use the new "settings" module (AKA slapd.conf -> OLC conversion). I was wondering if someone knew a good bit about doing this that I could bounce questions off?
<kalbasit[m]>
is there anything special about httpd's SSLCertificateKeyFile? I have it set to `/run/secrets/key` and even though the file is owned by wwwrun and its mode is set to 0444 httpd still fails saying the SSLCertificateKeyFile does not exist or is empty
<kalbasit[m]>
the secret file is setup by Mic92 's sops-nix
<kalbasit[m]>
hmm actually with mode 0444 it works so seems like the user of httpd can't access it even though it's the same user as far as I can tell
<m1cr0man>
What about /run/secrets itself? Also is selinux enabled on your system?
karantan has joined #nixos
<kalbasit[m]>
`/run/secrets` is world-readable. I don't have selinux enabled
mananamenos has quit [Ping timeout: 276 seconds]
<kalbasit[m]>
it's interesting I think I spoke too fast. It does not work with permissions `0666` either
<m1cr0man>
Yeah I would just check with ps what user your httpd is really running.
<kalbasit[m]>
I did, it's running wwwrun
simba1 has quit [Ping timeout: 258 seconds]
respawn_ has joined #nixos
<m1cr0man>
Do a systemctl cat httpd. I cant remember but if there's any sandboxing options enabled I belive that will make a unique /run appear for httpd
<kalbasit[m]>
There's `RuntimeDirectory` set to `/run/httpd` but that's not sandboxing is it? Because it's able to load files from the Nix store as evident by the loading of the httpd.conf it writes
mayhew has quit [Ping timeout: 264 seconds]
alexherbo2 has quit [Quit: Ping timeout (120 seconds)]
simba1 has joined #nixos
ahmedelgabri has quit [Ping timeout: 246 seconds]
alexherbo2 has joined #nixos
<ScottHDev5>
Hello, how coud I get the output of a derivation from it's .drv path example : from get /nix/store/...-hello-2.10 from /nix/store/...-hello-2.10.drv
<ScottHDev5>
?
<kalbasit[m]>
Dang I figured it out :facepalm: `/run/secrets` itself is a symlink to a directory that's not world-readable and the solution was to allow systemd unit to access that specific group.
<mindtree[m]>
Hi folks - is it possible to use nix without any channels whatsoever and instead declare them in nix exprs as necessary?
<mindtree[m]>
I'm considering trying this with my configuration.nix so that I can keep dependencies like `nixos`, `home-manager` and `nixos-hardware` declarative and more reproducible across machines and in my git history
mananamenos has joined #nixos
alexherbo2 has quit [Read error: Connection reset by peer]
<m1cr0man>
mindtree[m]: The best you could really do is add "-I nixpkgs=$NIXPKGS" where $NIXPKGS is the path to your own local version of the nixpkgs repo. That way you can keep the same version across machines
<m1cr0man>
add* to nixos-rebuild. if you're using nixops you can do a similar think afaik (I dont use it though)
<viric>
clever: interesting this thing of i686-linux without firefox... the processor can do 64-bit but if I use 64-bit, nix takes 2x ram as it takes on 32-bit
alexherbo26 has joined #nixos
alexherbo2 has quit [Ping timeout: 264 seconds]
alexherbo26 is now known as alexherbo2
<munksgaard>
mindtree: Have you considered looking at something like niv?
<mindtree[m]>
m1cr0man: ah interesting, so a `nixos` channel must exist for `nixos-rebuild` to function at all?
<kalbasit[m]>
mindtree: or flakes
<m1cr0man>
mindtree[m]: Well really an instance of nixpkgs must exist, which is what the channels provide
jmeredith has quit [Quit: Connection closed for inactivity]
<mindtree[m]>
<m1cr0man "mindtree: Well really an instanc"> But can't I just declare this nixpkgs source in my nix expression?
zupo has joined #nixos
<mindtree[m]>
<kalbasit[m] "mindtree: or flakes"> Yeah I'll likely switch to flakes once they eventually land in stable nix
<m1cr0man>
mindtree[m]: Ah yeah you could do that actually. You would be redeclaring the "pkgs" variable at the top of your configuration.nix I think
<kalbasit[m]>
<mindtree[m] "Yeah I'll likely switch to flake"> See github.com/SoxinOS/soxin for inspiration on that. If you still want the non-flake then my archived project can probably help you github.com/kalbasit/shabka
simba1 has quit [Ping timeout: 258 seconds]
blissdev has joined #nixos
simba1 has joined #nixos
<vulpine>
is there a way to put a package thats not in a channel into environment.systemPackages? putting a path to a nixfile seems to not work
<mindtree[m]>
<vulpine "is there a way to put a package "> I think this should work, as long as the file is an expression that evaluates to a derivation. That said, it seems to be more common to declare packages as a function that evaluates to a derivation, in which case you might need to use `callPackage /path/to/local/pkg.nix { }` to produce the actual derivation. At least I think it's something along these lines, I'm still new!
<davidak[m]>
Is someone interested in continuing the work on a Calamares GUI Installer for NixOS? Maybe C++ experience is needed to extend it. https://github.com/NixOS/nixpkgs/issues/100475
<ldlework>
error: Package ‘gst-plugins-base-0.10.36’ in /nix/store/21r9dxdkh1xxihjpc4axf6rcqq375h5m-nixos-20.09pre239318.c59ea8b8a0e/nixos/pkgs/development/libraries/gstreamer/legacy/gst-plugins-base/default.nix:55 is marked as broken, refusing to evaluate.
<ajs124>
ldlework: pgadmin3 apparently depends on gstreamer 0.10, which is old and broken
<ldlework>
shouldn't pgadmin3 then be marked as broken or some such?
<tpw_rules>
pgadmin3 is old and broken itself
<ajs124>
it's transitively marked as broken. apparently it also depends on an old openssl.
<tpw_rules>
unfortunate that 4 is not packaged
<ajs124>
there's #33103, but apparently nobody got around to it