civodul has quit [Quit: ERC (IRC client for Emacs 27.1)]
lsix1 has quit [Ping timeout: 246 seconds]
pjt_tmp has joined #nixos
Dotz0cat has quit [*.net *.split]
ziguana[m] has quit [*.net *.split]
rhyfore[m] has quit [*.net *.split]
kateskips has quit [*.net *.split]
Kimat[m] has quit [*.net *.split]
Notkea[m] has quit [*.net *.split]
tcakoi has quit [*.net *.split]
Lorenzoas3ii[m] has quit [*.net *.split]
codebam has quit [*.net *.split]
davidak[m] has quit [*.net *.split]
leonardp has quit [*.net *.split]
Dotz0cat has joined #nixos
Notkea[m] has joined #nixos
leonardp has joined #nixos
kateskips has joined #nixos
davidak[m] has joined #nixos
tcakoi has joined #nixos
Kimat[m] has joined #nixos
ziguana[m] has joined #nixos
codebam has joined #nixos
Lorenzoas3ii[m] has joined #nixos
rhyfore[m] has joined #nixos
rb2k has joined #nixos
Rusty1 has quit [Quit: WeeChat 2.3]
EmoSpice has joined #nixos
fgz has joined #nixos
fgz has left #nixos ["Closing Window"]
madmaurice has quit [Quit: Leaving]
ZaraChimera has joined #nixos
acarrico has joined #nixos
fresheyeball has quit [Quit: WeeChat 2.9]
OCCASVS has quit [Quit: Connection closed for inactivity]
rb2k has quit [Read error: Connection reset by peer]
rb2k has joined #nixos
mbrgm_ has joined #nixos
mbrgm has quit [Ping timeout: 258 seconds]
mbrgm_ is now known as mbrgm
Rusty1 has joined #nixos
<attila_lendvai>
is there an easy way to read a non-existent set element as false instead of an error?
<energizer>
s.foo or false
<energizer>
i dont really understand how the parsing works on that tbh
<energizer>
or maybe `or` contains a try/catch?
<attila_lendvai>
energizer, heh, yeah, that's also surprising to a newcomer like me. thanks!
ManiacOfMadness has joined #nixos
jonringer has quit [Remote host closed the connection]
supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos
nf is now known as optimism
optimism is now known as nf
rb2k has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rb2k has joined #nixos
<attila_lendvai>
is there something like dpkg's postinst script? i need to run a script when a service is "installed" (enabled for the first time). i know about preStart, but it's not ideal. this is about generating keyfiles for crypto stuff upon first install.
<KarlJoad>
cole-h: I'm trying to gather feedback on my PR before my school starts again. You had some before, and I've made some pretty significant changes since then. Do you have time to look at #108562 again?
<{^_^}>
[nixpkgs] @erictapen pushed commit from @johanot to release-20.09 « ceph: 15.2.7 -> 15.2.8 »: https://git.io/JtTVi
ericsagnes has joined #nixos
ifthenelse4 has joined #nixos
mvnetbiz_999 has joined #nixos
ifthenelse has quit [Read error: Connection reset by peer]
berber has quit [Write error: Connection reset by peer]
buffet has quit [Write error: Connection reset by peer]
ifthenelse4 is now known as ifthenelse
h4ck3rm4n has quit [Write error: Connection reset by peer]
mvnetbiz_99 has quit [Read error: Connection reset by peer]
ece3 has quit [Read error: Connection reset by peer]
mvnetbiz_999 is now known as mvnetbiz_99
Lord_of_Life has quit [Excess Flood]
mauli_ has quit [Read error: Connection reset by peer]
ece39 has joined #nixos
mauli has joined #nixos
h4ck3rm4n has joined #nixos
berber has joined #nixos
buffet has joined #nixos
Lord_of_Life has joined #nixos
jophish has quit [Ping timeout: 256 seconds]
ddima has quit [Ping timeout: 256 seconds]
ddima has joined #nixos
jophish has joined #nixos
superherointj has joined #nixos
<superherointj>
I've used nixfmt but it made the code worse IMO, replacing "\n\n" with ''(multiple lines)'' was quite bad. Is there a way to disable this? This does not seem right.
zupo has joined #nixos
<superherointj>
It has made an explicit line break in a string to be implicit in a horrible way.
xqsl_ has joined #nixos
xqsl has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @xaverdh opened pull request #109518 → nixos-rebuild: put some important dependencies in PATH → https://git.io/JtTr6
grace has joined #nixos
grace is now known as Guest65753
<bennofs>
is there no nix option to enable ip forwarding? docker warns that networking will not work since net.ipv4.ip_forward is not enabled
<bennofs>
s/nix/nixos
<Guest65753>
hey all, I'm using node2nix on a bigger project and I'm getting `unexpected end-of-file` when I try to build it. the syntax is correct but one of the generated nix files has ~20.000 lines. Any idea how I
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #109520 → ocamlPackages.awa: init at 0.0.1 → https://git.io/JtTrQ
davidv7 has joined #nixos
xelxebar has quit [Ping timeout: 240 seconds]
<bigvalen>
I'm having problems doing a nixos install from Debian. Problem is that Debian sets up namespaces, and then nix tries to set more up later, when doing the build. I added 'sandbox = false' to nix.conf, and it helped until it got to 'nix-build --out-link /mnt/tmp.4i1ZUGoKyP/system --store /mnt --extra-substituters "auto?trusted=1" "<nixpkgs/nixos>" "-A" "system" "-I" nixos-config=/mnt/etc/nixos/configuration.nix'. That also seems to modify
<bigvalen>
/proc/$PID/uid_map. Any idea how .. to either disable the debian name space, so that succeeds, or to tell nix-build not to use namespaces for this ?
<b42>
bennofs: you can enable it using boot.kernel.sysctl."net.ipv4.ip_forward" = true;
<kandinski>
I thought putting pcre2 in the buildInputs of a shell.nix would put the include directory in the environment's $PATH, but it doesn't: https://termbin.com/lqto
<kandinski>
is this something I'm doing wrong?
<bennofs>
b42: thanks. I was just surprised to find that there was not special option for it (like networking.enableForward or something similar)
ambro718 has joined #nixos
xelxebar has joined #nixos
turlando_ has joined #nixos
Guest65753 is now known as ziprando
ziprando is now known as ziprandom
<kandinski>
this is using nix-shell from Nix 2.3.7 (and lorri, but same issue with invoking nix-shell by hand)
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 32 commits to staging-next: https://git.io/JtToO
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 33 commits to staging: https://git.io/JtTo3
turlando has quit [Ping timeout: 246 seconds]
attila_lendvai has quit [Ping timeout: 240 seconds]
<djanatyn>
you said the include directory, which i don't think gets added. you can see that only the `/bin` directory is on the path. i included the source of the derivation i'm using (pkg.pcre2) in my paste
<supersandro2000>
if you need to pass -E to sed because otherwise nix tries to interpret your regex as a var but you can't escape it because sed does not like that in the regex
<supersandro2000>
🤯
ManiacOfMadness has joined #nixos
berberman has joined #nixos
<supersandro2000>
and with -E I need to scape ${} but with -e not
<supersandro2000>
but for nix I need escape $\{ which sed does not like
<Duponin>
`nixos-install` refuses to use to use the `nixos-config=/foo/bar/configuration.nix` I set and have no idea why, it forces to use the one in `/mnt/etc/nixos/configuration.nix`
<Duponin>
someone experienced something similar?
hiro99 is now known as hiro98
<gchristensen>
at some point it does a chroot to change what `/` means, making /foo/bar/... stop making sense
<Duponin>
sure but it worked until today :|
<{^_^}>
[nixos-search] @turboMaCk pushed to turboMack/faceted-search-trap-click « Few small improvements »: https://git.io/JtT1H
<Reventlov>
hey
<gchristensen>
weird ... not sure
<Reventlov>
can I just add myself to some package maintainers if I want to update it ?
<{^_^}>
[nixos-search] @turboMaCk pushed 0 commits to faceted-search: https://git.io/JtT15
<gchristensen>
you can update it without that
<Reventlov>
yeah, but what if I want to
<gchristensen>
sure
ManiacOfMadness has quit [Ping timeout: 240 seconds]
<pbb>
Duponin: You could actually specify your config with the NIXOS_CONFIG environment variable and then it would check its existence correctly
rb2k has quit [Ping timeout: 264 seconds]
<cole-h>
FWIW: to get the value of a search path, it is as easy as `nix-instantiate -E --eval '<nixos-config>'`, which will give you the underlying path
magnetophon has joined #nixos
<Duponin>
pbb: oh, didn't knew about this ENV_VAR, interesting
rb2k has joined #nixos
fuzzypixelz has joined #nixos
rb2k has quit [Client Quit]
<pbb>
It's documented in the nixos-rebuild man page, but I didn't know about it for a long time either. Learned about it by reading nixos/default.nix.
<fuzzypixelz>
hello. I noticed that rustup wasn't packaged until recently in nix, why is that? I'd like to use rustup because my editor requires it to work with the RLS. Are there any deal-breaking issues with using rustup with nix? The nixpkgs manual says to install cargo and rustc
<pbb>
fuzzypixelz: rustup is packaged since 2017
<pbb>
fuzzypixelz: it's a bit of a hack since the downloaded binaries need to be patched to work on NixOS, but the rustup package takes care of that
roconnor_ has quit [Quit: Konversation terminated!]
<ajs124>
Hm, I'm trying to add lua support to dovecot right now and it fails to build with "*** No rule to make target '-L/nix/store/qp7s8wsq919p03alrchf5i9lpa2h3fn2-lua-5.4.2/lib', needed by 'libdovecot-lua.la'. Stop."
superherointj has quit [Remote host closed the connection]
superherointj has joined #nixos
magnetophon has quit [Ping timeout: 264 seconds]
<ajs124>
Duponin: lua is already compiled. it tries to compile that target, but that should just be a flag of where to find the libraries to link against, as far as I understand.
<ajs124>
I tried modifiying the lua pkg-config .pc file to not output these parameters, but then it obviously just fails to link. maybe there's something wrong with dovecots build system
<{^_^}>
[nixpkgs] @sternenseemann opened pull request #109550 → pythonPackages.ufoprocessor: fix test execution → https://git.io/JtT7E
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @dotlambda to master « python3Packages.zulip: 0.7.0 -> 0.7.1 »: https://git.io/JtTFu
<teto>
Reventlov: worst case naybe you can ignore it. the log doesn't show much. I haven't used ns-3 in a while maybe it can be made more verbose. I remember about NS_DEBUG but that may be too verbose. it's just lr-wpan-ack :(
KarlJoad has joined #nixos
<KarlJoad>
Is there any nix built-in to modify pkg-config's search path? Or do I _have_ to mess with $PKG_CONFIG_PATH?
<{^_^}>
[nixpkgs] @tesq0 opened pull request #109552 → lepton-eda: init at 1.9.13-20201211 → https://git.io/JtTFP
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
<superherointj>
I want to add hexchat themes packages to Nixpkgs (so it is easy to pick and switch themes in HexChat config). I'm wondering which path is the right one for this. Thoughts?
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 59 commits to staging-next: https://git.io/JtTxF
<{^_^}>
[nixpkgs] @github-actions[bot] pushed 60 commits to staging: https://git.io/JtTxb
vidbina_ has joined #nixos
<superherointj>
"<nixpkgs>/pkgs/misc/hexchat-themes" Is it a good idea?
<Ke>
yop commented out the check, zfs does not get built
<clever>
superherointj: <nixpkgs/pkgs/misc/hexchat-themes> would be better, then it picks a dir that actually has that path
<Ke>
so check gets evaluated, but the package is not installed
lsix has quit [Ping timeout: 246 seconds]
<Ke>
--show-trace is pretty much useless as always
<superherointj>
clever, I'm sorry, I did not understand.
<{^_^}>
[nixpkgs] @jonringer merged pull request #109539 → pythonPackages.check-manifest,pythonPackages.swaglyrics: Fix dependencies; botan: fix compilation on darwin → https://git.io/JtTXh
<{^_^}>
[nixpkgs] @AndersonTorres pushed 2 commits to master: https://git.io/JtThH
<KarlJoad>
When running a `nix search dolfin`, I end up with only one CPP header, `nixos.fenics`. But I have many python bindings. Does fenics not exist in nixpkgs?
<{^_^}>
[nixpkgs] @primeos pushed to master « google-chrome-dev: Add the new libxshmfence dependency »: https://git.io/JtTjr
ericsagnes has quit [Ping timeout: 246 seconds]
<bqv>
Nevermind, /nix/store/.links only goes up to 1z***...
<bqv>
For a minute I thought it might max out at zz***...
superherointj has quit [Quit: Leaving]
<KarlJoad>
I can't seem to find a way to get *just* the fenics package available for use. With the way it is now, is it possible to only use it through python? Because I need to compile some packages with the fenics library.
acarrico has joined #nixos
davidv7 has quit [Quit: Konversation terminated!]
ericsagnes has joined #nixos
kini has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @leungbk opened pull request #109562 → graph-cli: init at 0.1.7 → https://git.io/Jtkek
<remexre>
does nix guarantee that I can use `#!/usr/bin/env foo' in build scripts? I noticed that `#!foo' also works, which I guess doesn't depend on /usr/bin existing?
<inquisitiv3>
Can `inherit` take multiple arguments? So `inherit a b;` is the same as `inherit a; inherit b;`?
<clever>
remexre: /usr/bin/env wont exist in a build sandbox, but the patchShebangs function will replace that with absolute paths to the real binary
<clever>
inquisitiv3: yes
<inquisitiv3>
clever: Thanks!
superherointj has joined #nixos
isHavvy has quit [Read error: Connection reset by peer]
<energizer>
tried services.xserver.synaptics.scrollDelta = 20; but i dont think that did anything
<bqv>
Ok cool, at least this way I get a decent ETA too. Half an hour from now
<remexre>
clever: so in my builder, I would just write `patchShebangs $src' ?
<bqv>
Ty gchristensen
<clever>
remexre: and $src is read-only, you have to patch the copy in . after the unpackPhase
<remexre>
clever: ok, so would it be most idiomatic to do it in the patchPhase?
<{^_^}>
[nixpkgs] @HolgerPeters opened pull request #109565 → sphinx-autobuild: init at 2020.9.1 → https://git.io/Jtkv9
EmoSpice has quit [Ping timeout: 265 seconds]
<inquisitiv3>
I'm trying to prepare a PR for updating Pijul in NIxpkgs from alpha version 31 to alpha version 34 (latest). Pijul is a version control system written in Rust. The expression is located in https://github.com/NixOS/nixpkgs/blob/master/pkgs/applications/version-management/pijul/default.nix. Going prior version bump commits I only need to change the `version` attribute and two hashes. My problem is finding the hashes.
vidbina_ has quit [Ping timeout: 246 seconds]
o1lo01ol1o has quit [Remote host closed the connection]
<energizer>
inquisitiv3: replace them with e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 and try to build, it will tell you the right hashes
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/Jtkfc
<adisbladis>
superherointj: Neither is blessed
<adisbladis>
So use either, or none
<superherointj>
What is blessed?
<superherointj>
Allright. THanks.
<adisbladis>
superherointj: I mean there isn't an official recommendation
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<superherointj>
Got it. I'm about to do a PR. :)
<Reventlov>
teto: seems to be a bug upstream
slack1256 has quit [Ping timeout: 264 seconds]
<superherointj>
I remember there was an issue about removing stdenv. Does it affect `stdenv.mkDerivation`?
<qyliss>
there is no issue about removing stdenv
<adisbladis>
Huh
<adisbladis>
We're not removing stdenv
<superherointj>
I meant, references to it.
<qyliss>
the only thing that's changing is that top-level "lib" is now preferred over "stdenv.lib "
<superherointj>
I see.
riksteri has quit [Quit: riksteri]
<inquisitiv3>
I get this error when I try to build pijul using lib.fakeSha256:
<inquisitiv3>
error: hash 'lib.fakeSha256' has wrong length for hash type 'sha256'
<inquisitiv3>
(use '--show-trace' to show detailed location information)
<superherointj>
Is it acceptable to add myself as a CODEOWNER for a file that is included in a PR? I just want to be notified when that file is changed.
magnetophon has quit [Ping timeout: 272 seconds]
philr_ has joined #nixos
<{^_^}>
[nixpkgs] @dali99 opened pull request #109567 → metapixel: init at 1.0.2 → https://git.io/JtkJv
civodul` is now known as civoudl
ambro718 has joined #nixos
civoudl is now known as civodul
<bqv>
Oh wait, I literally can't unmount /nix, can i
<inquisitiv3>
energizer: I changed both hashes to `lib.fakeSha256` and tried to build Pijul.
<inquisitiv3>
When building I got a "hash mismatch in fixed output derivation '/nix/store/[...]" message with a hash, but I'm a little confused which of the two hashes (if even one of them) I got?
arjen-jonathan has quit [Ping timeout: 246 seconds]
<inquisitiv3>
The hashes that was used as an argument for `fetchCrate` is different from the `cargoSha265` hash. Now when I try to replace `lib.fakeSha256` with `"sha256-[hash]"` I get an error message saying "invalid SRI hash".
<energizer>
there are only two options so if you're not sure you can try each
<bqv>
etu: DigitalKiwi: any ideas how to get the kernel to let me at the goddamn block device?
<energizer>
but the first one should go in the src block
<energizer>
i think
<etu>
bqv: kexec another system from ram
<bqv>
Lemme guess, the issue is it's still mounted in some hidden namespace
cantstanya has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @petabyteboy pushed commit from @euank to master « libvirt: correct localstatedir (#109415) »: https://git.io/JtkUz
<DigitalKiwi>
bqv: just don't do what i did and reboot and not be able to login again and then even though you were smart and have a live usb img always available for these occassions zfs format updated since it was created so it can't mount r/w... no big deal you can make a new usb drive from the live disk... and then when you connect the sd reader it kernel panics and all of that time is wasted. finally you get a working one and delete a few snapshots to fre
<DigitalKiwi>
e up space. which would have been a lot easier to do before rebooting
<DigitalKiwi>
nix-build iso ~5 minutes, trying to figure out how ~250GB of free space went to 0 in <1 day... like 10 hours :(
trubi has quit [Ping timeout: 240 seconds]
cantstanya has quit [Remote host closed the connection]
cantstanya has joined #nixos
<DigitalKiwi>
oh and don't panic and forget that zfs is copy on write so deleting files requires space and since snapshotting is enabled it'll make it even worse :|
pbb has quit [Remote host closed the connection]
<gchristensen>
luckily zfs reserves some space to allow for deletes
<gchristensen>
does, uh, gnome / wayland use video drivers?
<DigitalKiwi>
would probably help to have killed the auto snapshot service and truncate files :(
<DigitalKiwi>
why wouldn't it? it even uses xorg :D
<gchristensen>
it doesn't look like I'm running X, I do have Xwayland though
<matthewcroughan>
Anyone here use Bitbake and Yocto?
<matthewcroughan>
I've been thinking that it would be possible to make a declarative Bitbake file, using Nix. Similar to how one could do so for a Dockerfile.
<pbb>
matthewcroughan: building yocto under NixOS is a huge pain
<matthewcroughan>
gchristensen: I'm sure you would know a lot about this, based on your docker layer knowledge. Since it is basically specced identically to Yocto, on purpose.
<{^_^}>
#106471 (by FRidh, 5 weeks ago, open): License of package versus license of source/project and propagation of licenses
<matthewcroughan>
This is for example, one of the things that GuiX does far superior to Nix, package license metadata.
<pbb>
I see
<matthewcroughan>
There's even Fosdem talks about replacing Yocto with GuiX. That can't happen with Nix quite yet because of this superfluous legality stuff.
<bqv>
This was such a mistake...
<bqv>
Imma reboot
<bqv>
Everything woulda been fine if I didn't umount -l >_>
<DigitalKiwi>
fine, maybe, yes, but not as entertaining
ManiacOfMadness has quit [Ping timeout: 240 seconds]
<matthewcroughan>
bqv: are you talking about the way the damned drive sticks around and the kernel doesn't know what to do?
<matthewcroughan>
There's a fix for that, have you rebooted yet?
<matthewcroughan>
pls say no
<matthewcroughan>
FreeBSD doesn't have that problem, funny enough ;D
<DigitalKiwi>
does nix work on freebsd again/yet
<pbb>
you can run it but there are basically no packages
<pbb>
because *BSD needs a lot of patches for upstream source
<{^_^}>
[nixpkgs] @flokli opened pull request #109573 → grafana-agent: init at 0.10.0 → https://git.io/JtkTy
<matthewcroughan>
pbb: is that real?
<pbb>
no xD
<matthewcroughan>
lmfao
<matthewcroughan>
I want it to be though.
<pbb>
same
<matthewcroughan>
NTPackages
gustavderdrache has joined #nixos
<matthewcroughan>
the kernel no longer matters
<pbb>
that's only for neurotypical people
<matthewcroughan>
hell, have the kernel be Minix 3.
<pbb>
you know, because it's the NT kernel
<matthewcroughan>
HAAAAAA
<pbb>
ehm, matthewcroughan, I think it's very possible to assert that all buildtime and runtime software is licensed in some way
simba1 has quit [Ping timeout: 246 seconds]
<DigitalKiwi>
surely everyone who uses nix willingly is neurodivergent
<matthewcroughan>
pbb: Yes, but you need to therefore maintain a list of those overrides of pkgs.<program>.licenses
<pbb>
overrides?
<matthewcroughan>
the licenses in your proposed scenario are not written by others.
<matthewcroughan>
You could similarly maintain your own `./pkgs/package_name` in your own repo on Github, but that's not very helpful is it/
<pbb>
I don't get it
<matthewcroughan>
How many Nixpkgs are there?
<pbb>
do you mean how many packages there are in nixpkgs?
<matthewcroughan>
We would have to change their .license attribute, all of them.
simba1 has joined #nixos
<pbb>
why?
<matthewcroughan>
because they're all wrong.
<pbb>
why are all of them wrong?
<matthewcroughan>
They're using the project license, not the propograted license.
lsix has quit [Ping timeout: 272 seconds]
<pbb>
Yes, but if you assert that none of your build dependencies use a restrictive license, you also know that none of your results use a restrictive license
<matthewcroughan>
Propogated license means like NetworkManager is GPL, but it needs to be GPLv3, because of the fact that it uses Readline.
<matthewcroughan>
NixPkgs says NetworkManager is GPLv2. Yes, that's what the NetworkManager code is.
<matthewcroughan>
but NetworkManager actually links to Readline, which is GPLv3. This is not reflected in the `.license` attribute of the nixos.networkmanager project.
<matthewcroughan>
-project, didn't mean to say that last word :D
<pbb>
Yes I understand your problem. But it's easy to check that all your build dependencies (including Readline) are _not_ using GPLv3
<matthewcroughan>
pbb: and now could we say to Nix, "Please swap readline out for a BSD alternative"?
<bennofs>
you can computed propagated license with nix probably. I mean the info is all there
<matthewcroughan>
currently NetworkManager.license = GPLv2
__monty__ has quit [Quit: leaving]
<matthewcroughan>
it'd be a recursive function bennofs
<pbb>
no
<bennofs>
nix can do recursive functions
<pbb>
it wouldn't
<pbb>
I mean recursive yes, but not infinite recursion
<matthewcroughan>
yes but it'd be terribly slow
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pbb>
no
<matthewcroughan>
I've not been shown a way to do it, so most of what I'm saying is pontification. I need to understand the language better.
<bennofs>
but I still don't see how that solves the issue of having a list of replacements. At the very least you'd need a list of interchangeable packages, and even then, you probably need different patches for each set that you want to build.
<bennofs>
not seeing how you'd be able to improve a lot on list of overrides + error if forbidden license is used
czer00 has quit [Remote host closed the connection]
<matthewcroughan>
-P | Force the kernel to scan the partition table on a newly created loop device. Note that the partition table parsing depends on sector sizes. The default is sector size is 512 bytes, otherwise you need to use the option --sector-size together with --partscan.
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
<Jonathan89>
Uh, this is not cool. I updated my nix rig and now I cannot switch to a TTY, when I try it switches back to my display manager.
<Jonathan89>
Any idea what is causing that?
fendor has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @lovesegfault merged pull request #106617 → firefox: fix running wayland firefox built with LTO and some miscellaneous improvements → https://git.io/JIwdu
<{^_^}>
[nixpkgs] @lovesegfault pushed 8 commits to master: https://git.io/JtkLd
<{^_^}>
[nixpkgs] @NomisIV opened pull request #109585 → Add package gruvbox dark gtk theme and icons → https://git.io/JtkLA
detran has joined #nixos
<Jonathan89>
Would that be SDDM, the display manager I am currently using, or is bash somehow missing?
troydm has joined #nixos
EmoSpice has joined #nixos
EmoSpice has quit [Ping timeout: 260 seconds]
<pbb>
Jonathan89: no idea, maybe a) try disabling SDDM and b) open an issue and tag the maintainers of SDDM, if it only happens with SDDM enabled
<KarlJoad>
I can't find a fenics package that includes the `dolfin.h` header I can use for compiling. I know it exists (it is in my nix store) because it is used to mkae a python bindings package for fenics. But, fenics by itself is not available. Anyone have any ideas?
<Jonathan89>
Yeah, I don't see a setting for it, like the did it intentionally.
<{^_^}>
[nixpkgs] @fabaff opened pull request #109586 → python3Packages.awsiotpythonsdk: init at 1.4.9 → https://git.io/Jtkta
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
ManiacOfMadness has quit [Ping timeout: 246 seconds]
<pbb>
matthewcroughan: so I built a networkmanager with linenoise instead of readline, but there are more GPL-3 dependencies. to debug this, I created an expression that can tell us what packages are still in the dependency tree and have this license.
<inquisitiv3>
madjar: How do you find the hashes for Pijul? Saw that you created the last update PR. I tested to generate them by replacing the two hashes with `lib.fakeSha256` and went through the resulting derivation to find where the hashes is located. But the hashes in the files isn't anywhere in the derivation file? :/
<Jonathan35>
I removed SDDM, and whichever one took it's place does the same thing.
<Jonathan35>
I stopped "display-manager" with systemctl and I got back to my tty.
<inquisitiv3>
cole-h: Thanks!
<colemickens>
how do I reference the out path of a devShell attribute?
o1lo01ol1o has quit [Remote host closed the connection]
<cole-h>
colemickens: `nix eval --json .#devShell.x86_64-linux.outPath` works for me
<colemickens>
cole-h: I wind up getting this though when it's built This derivation is not meant to be built, aborting
Rusty1 has joined #nixos
<colemickens>
I'm trying to build + pre-cachix it
<cole-h>
Don't mkShell it then ;)
Raimu[m] has joined #nixos
lsix has quit [Ping timeout: 246 seconds]
<cole-h>
mkShell has only one phase, the `nobuildPhase`, which prints that message :P
pisquo[m] has joined #nixos
* colemickens
has done this before and had to re-mkShell it for some reason.
<colemickens>
hm maybe im making that up
<colemickens>
I'll try
Jonathan35 has quit [Quit: Connection closed]
<cole-h>
mkShell is just a wrapper around mkDerivation.
<aanderse>
if i have a submodule written in a separate file from my module (think nginx calling vhost-options.nix from default.nix) what is the "best" way to pass config into the submodule? the submodule file (vhost-options.nix should have a { config, mainConfigFromDefaultDotNix, lib, ... } in it)?
<aanderse>
infinisil because <3
<aanderse>
ping
EmoSpice has joined #nixos
<infinisil>
aanderse: Unfortunately I don't think there's a best way. All of them are a bit clumsy. Main reason is probably just that this kind of breaks the encapsulation of submodules
<infinisil>
Because as soon as they depend on the parent submodule, they're not self-contained anymore
<colemickens>
cole-h: stdenv.mkDerivation? I get an error if I switch to that too :/
<infinisil>
s/parent submodule/parent module
<inquisitiv3>
cole-h: This is the output I get when I try to build Pijul. No hash that I get correspond to one of the ones you get in your diff.
<infinisil>
aanderse: I guess the nice solution would be to write submodule options for the inputs that you require from the parent module
<cole-h>
colemickens: Did the error change?
* cole-h
missed the gist
<colemickens>
cole-h: yeah
<aanderse>
ok, so in the case of `nixos/modules/services/web-servers/apache-httpd/vhost-options.nix` there is a `config` in the arguments at the top of the file
<aanderse>
this `config` represents the submodule, not global `config` - can i reference the global `config` from a submodule like this?
<infinisil>
aanderse: Then the parent module sets those options, then the submodule is self-contained
<colemickens>
it wants a $src
<cole-h>
colemickens: dontUnpack = true;
<cole-h>
probably
<colemickens>
cole-h: I guess I should probably just peek at mkShell source :|
<infinisil>
aanderse: Nope, again that breaks the submodule abstraction. submodules are entirely separate evaluations that don't know about the context in which they run
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtkqE
<inquisitiv3>
cole-h: But I'm also on Nix unstable? Running Nix on a Fedora machine.
<infinisil>
aanderse: You could also use `{ parentConfig, config, lib, ... }:` in the submodule, then import it with `types.submodule [ ./path/to/submodule { _module.args.parentConfig = config; } ]`