<maurer>
emacsomancer: Also note that if you want to use loadable modules, the target system will probably want to have a /nix directory on it, though this could likely be worked around
<emacsomancer>
maurer: ty. the machine I have in mind has the nix standalone pkg manager on it and thus a /nix directory. I'll have to do some research on this it looks like (I momentarily entertained the idea that the prebuilt kernels might automagically work on foreign distros)
<maurer>
If you are lucky, they might - the thing is, foreign distros may have assumptions about what things are built into the kernel, available as modules, etc. based on their package manager state
<maurer>
(for example, if you have LVM-based partitioning, and you don't tell nix about it when building the kernel, you could end up with a kernel that doesn't boot)
<maurer>
(in the NixOS case, this is handled by the entire system being declarative, so it can see that there will be LVM partitions in the configuration.nix directly)
ddellacosta has quit [Ping timeout: 268 seconds]
<emacsomancer>
maurer: I was interested particularly in the zfs-kernels, but those surely add complexity to the whole thing.
<infinisil>
DigitalKiwi: I usually just use tab completion in a nixpkgs checkout, but you can also use something like `nix-env -qaPA nixos.nodePackages`
<DigitalKiwi>
i mean if i didn't know nodePackages was a thing
Guanin has quit [Remote host closed the connection]
<infinisil>
All node packages that are applications should be aliased in the top-level pkgs definition
wfranzini has quit [Remote host closed the connection]
<infinisil>
So if you search for packages like normal you'll find it
wfranzini has joined #nixos
<DigitalKiwi>
haskellpackages aren't
mexisme has quit [Ping timeout: 245 seconds]
n0qqe has joined #nixos
dansho has joined #nixos
drakonis1 is now known as drakonis
<DigitalKiwi>
i don't know if i'm not asking the right way
noqqe has quit [Ping timeout: 248 seconds]
n0qqe is now known as noqqe
<DigitalKiwi>
say i didn't know that haskellPackages was a thing i could query that didn't show up in normal searches, is there a way to (programmatically) discover it (and anything else like it) without digging through git repos or finding it on the manual
jasongrossman has quit [Ping timeout: 248 seconds]
<{^_^}>
[nixpkgs] @colemickens opened pull request #63273 → data/fonts/gelasio: init at v2018-08-12 → https://git.io/fjaBR
<colemickens>
I have a scenario where I'm building packages on a spot machien and want them always pushed to cachix, even if some of the current build fails.
<colemickens>
I'm realizing that I may already get the behavior that I want, as long as I don't have `-o pipefail` like I usually do
slack1256 has quit [Remote host closed the connection]
oborot has quit [Ping timeout: 248 seconds]
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
hmpffff has quit [Quit: nchrrrr…]
jasongrossman has joined #nixos
Rusty1 has quit [Quit: Konversation terminated!]
<jasongrossman>
How does one tell why a package's been marked as broken?
<jasongrossman>
I mean, obviously the answer is "because it's broken"! But marking it as broken is done by a human, right? So someone might have said something interesting about why.
wfranzini has quit [Remote host closed the connection]
<clacke_movim>
jasongrossman: You'd have to `git blame` that line, and read the commit message and associated PR to see the motivation for it.
mexisme_ has quit [Ping timeout: 248 seconds]
<clacke_movim>
If you do the git blame on github, you can just click the commit and it will have a link to the PR.
o1lo01ol1o has joined #nixos
<clacke_movim>
What's an example of a package marked broken, and what is its nix file?
<jasongrossman>
Huh. I knew about this problem but was hoping not to have to deal with it.
<clacke_movim>
So you'll have to do it offline.
<jasongrossman>
Anyway, I believe somebody has their eye on gitit. Thanks for answering my general question.
InFlames has joined #nixos
<clacke_movim>
In a checked out nixpkgs you do `git blame haskell-packages.nix` and each line will start with a header indicating which commit last changed that line.
<jasongrossman>
I see, good.
<clacke_movim>
So you look at what thatvline somewhere near line 91317 says, and follow the evidence from there.
<jasongrossman>
Hm. But gitit is not marked as broken in that file.
toppler has joined #nixos
<jasongrossman>
I'm happy now, though, because I just noticed that someone was saying sensible things about gitit in nixpkgs github issues recently.
<clacke_movim>
Oh, maybe a dependency. I wonder. .. I think Nix 2.0 added tooling to explain why a particular package is considered broken?
<clacke_movim>
+1
kvda has joined #nixos
<jasongrossman>
A dependency was my first thought, because it has many ... but the error message from nixos-rebuild suggests it's gitit itself. Oh well. Doesn't matter.
<jasongrossman>
(Just to clarify though: the error message say that gitit is marked as broken, and I still don't see where it's marked as broken!)
<jasongrossman>
(But maybe the error message is wrong, and it means one of its dependents is marked as broken.)
<clacke_movim>
Are you looking at the same commit of nixpkgs?
<clacke_movim>
Could be fixed in master but still marked broken in latest nixpkgs-unstable because master didn't pass the gate yet.
<jasongrossman>
Ah, right, true.
<jasongrossman>
I hadn't thought of that.
<jasongrossman>
But the history for the "ginormous blob" you linked to says it hasn't changed recently.
<jasongrossman>
... the version of that file in my Nix store DOES have gitit marked as broken.
<clacke_movim>
OK so time to `nix-channel --update` :-)
<jasongrossman>
I don't understand git very well, which is ironic given the name of the package I'm looking at (although, also ironically, that package doesn't actually depend on git).
<clacke_movim>
:-)
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jasongrossman>
I think nixos-rebuild --build does an update, doesn't it?
<jasongrossman>
Sorry, typo.
<clacke_movim>
No
jgt has joined #nixos
<jasongrossman>
I mean `nixos-rebuild build --upgrade` does an update.
<jasongrossman>
I think.
<clacke_movim>
It creates the system based on what your nixos channel state is
kvda has joined #nixos
<clacke_movim>
Ah, `--upgrade`, ik
<clacke_movim>
ok*
<jasongrossman>
Right.
day|flip has quit [Remote host closed the connection]
<jasongrossman>
The git history for that hackage-packages file shows why OTHER things are marked as broken!
adamantium has joined #nixos
<jasongrossman>
I'm happy to live with this small amount of mystery though.
jgt has quit [Ping timeout: 257 seconds]
<clacke_movim>
If you find out what commit your nixpkgs is from, you can `git blame loremipsumdeadcafeloremipsumdeadcafe1234 haskell-packages.nix`
<clacke_movim>
And that will show you blame from that point in time
<{^_^}>
[nixpkgs] @fpletz pushed commit from @mmahut to release-19.03 « bubblewrap: CVE-2019-12439 »: https://git.io/fjaB5
<li_matrix>
which packaging nix translator do you think produces the most sane nix code? im reviewing crate2nix and it needs to produce much more efficient nix; evaluation takes forever. Could it be because it uses a single recursive attribute set?
<clever>
li_matrix: the next trick, is that you can mix makeScope and newScope together with your `self: { stuff }` to make a set that contains an .overrideScope'
<slack1256>
etu: Thanks a lot!, will nixos-rebuild now
<clever>
li_matrix: and then people can take your stuff and setOfCrates.overrideScope' (self: super: { overlays })
<avitex_>
Heya folks, how do you become a maintainer for a package, does it require pre approval?
sigmundv has joined #nixos
<avitex_>
I have a bug fix patch I'm going to submit as a PR, if I can get added as a maintainer for that package if problems arise I'm happy to help
jgt1 has quit [Read error: Connection reset by peer]
gratto has quit [Ping timeout: 244 seconds]
gratto has joined #nixos
inquisitiv3 has quit [Quit: Leaving]
inquisitiv3 has joined #nixos
<johanot>
avitex_: I'd just add myself to the maintainer section of the package meta and PR that. Then, request approval/review from existing maintainers of the package.
<johanot>
in general, fastest way of becoming maint. of an existing package is to talk to the current maintainer(s). :)
<pbb>
Hi, I recently set up a server with an encrypted ZFS root. Regularly all disk IO gets stalled, I can't start new ssh sessions or applications that are not already in memory and nothing is logged anymore. This usually happens after a few hours. Any ideas how I could debug this?
<avitex_>
johanot: cheers mate, didn't want to force myself on the package on a PR, so I thought I'd check
hmpffff has joined #nixos
<Zer0xp>
andi-: So that command or directory doesn't exist for me
<johanot>
avitex_: you can always try to pre-ping existing maint. on IRC first, but tagging them in a PR should be sufficient.
<{^_^}>
[nixpkgs] @fpletz pushed to gcc-8 « linuxPackages.usbip: fix build with gcc8 »: https://git.io/fjauW
<Zer0xp>
andi-: Sorry I meant to say file or directory
<avitex_>
johanot: i think that's what i'll do then :)
<clever>
pbb: this causes mingetty to print an ansii escape code on startup, which disables power-saving in the gpu, so the monitor output is never disabled
inquisitiv3 has quit [Quit: Leaving]
<clever>
pbb: and because of that, any errors the kernel prints to the console, are visible, even after the kernel locks up solid
<clever>
so i can just turn the monitor on to see what blew up
<pbb>
ok. I don't have physical access to the server unfortunately. shouldn't it turn up in dmesg too?
<pbb>
I don't get new dmesg entreis when it happens
<clever>
pbb: ah, the machine is responding enough to run dmesg?
<pbb>
I left dmesg -w running
<clever>
ah
<pbb>
and I can also run it again after it locks up
<clever>
pbb: this lets you trigger sysrq without physical access, l t and w could be of use
<clever>
pbb: and b will just hard-reset the machine, without even syncing the disks, so you can get it back online quickly
hmpffff has joined #nixos
<clever>
b is basically identical to the reset button on the motherboard
<pbb>
I knew the reboot option
<pbb>
but not the others
<pbb>
that's very helpful, thanks
<pbb>
clever++
<{^_^}>
clever's karma got increased to 136
hmpffff has quit [Client Quit]
<immae>
I won’t try now because I don’t need it, but can we do `echo "reisub" > /proc/sysrq-trigger` (the "usual" sequence when a computer is stuck) and it will do the sequence as if it was done from keyboard?
gratto has quit [Read error: Connection reset by peer]
gratto has joined #nixos
<clever>
immae: could try with some of the read-only things, like ltw and see if it prints multiple debug things at once
<immae>
clever: Yes but usually when you do the "i" in reisub you already killed everything so it cannot continue, you’d have to adapt the sequence
cfricke has quit [Quit: WeeChat 2.5]
<clever>
immae: ack!
<immae>
anyway, it was just out of curiosity, I wasn’t aware of that facility and wondered how far it could go
<immae>
It’s ages since I last used it actually :D
<clever>
immae: related, if you send a break signal over the serial port, you can follow it with one of those magic letters
<immae>
(good point for GNU/linux I’d say)
<clever>
immae: so <break>o will just cut power to the whole system
<immae>
ok
<clever>
immae: and a break signal on a serial port, is i think 10 bits of low, when the normal pattern is 1 low for the start bit, 8 data bits, and 1 high for the stop bit
<immae>
you lost me there
<clever>
some serial port programs have a hotkey to send a break code over the serial port
<clever>
which will do that for you
Makaveli7 has quit [Quit: WeeChat 2.5]
orivej has joined #nixos
<immae>
in my case I wondered about my remote servers, which I usually access only by ssh. (I have no serial access as far as I know). In my desktop/laptop I didn’t need to use those magic keys for ages (won’t complain about that), and I have a keyboard anyway, so...
<clever>
tobiasBora: do you have the output path for something its trying to build?
<tobiasBora>
clever: I can read lines like /nix/store/i4r7xx7sj1bjgvj9p6dv59f1mb329ivw-coreutils-8.31/bin/install -c -m 644 scan/__init__.py scan/sane.py '/nix/store/36xbdg33gdlyzpjh3dmr07y0znxy0kkk-hplip-3.19.1/share/hplip/scan'
<clever>
/nix/store/36xbdg33gdlyzpjh3dmr07y0znxy0kkk-hplip-3.19.1 then
<tobiasBora>
so I guess /nix/store/36xbdg33gdlyzpjh3dmr07y0znxy0kkk-hplip-3.19.1/share/hplip/scan is a path that is being built
<tobiasBora>
I found /nix/store/36xbdg33gdlyzpjh3dmr07y0znxy0kkk-hplip-3.19.1/share/hplip/dat2drv ==> 64 bits (dat2drv: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked, interpreter /nix/store/g2p6fwjc995jrq3d8vph7k45l9zhdf8f-glibc-2.27/lib/ld-linux-x86-64.so.2, for GNU/Linux 2.6.32, with debug_info, not stripped
<clever>
tobiasBora: then its not a matter of 32bit being uncached
<clever>
tobiasBora: can you pastebin your configuration.nix file?
<clever>
tobiasBora: i would expect that to all work just fine
<clever>
tobiasBora: only thing i can complain about, is that your opening port 22 udp, but ssh is tcp only
<clever>
tobiasBora: and ssh opens 22 tcp on its own, so you dont need either
<tobiasBora>
clever: true
<tobiasBora>
thanks
<tobiasBora>
but does not solve the recompilation issue ^^
<clever>
tobiasBora: yeah, the next step would be to see what depends on this modified hplip
<immae>
tobiasBora: I suppose you don’t do anything funny in hardware-configuration.nix or radio.nix?
<clever>
tobiasBora: once the rebuild is done, use `nix why-depends` to see how your final nixos build depends on the above hplip
ThatDocsLady has joined #nixos
<tobiasBora>
immae: untouched hardware-conf, and in radio it's just a services.icecast.{...}, some allowed ports, a environment.systemPackages = with pkgs; [ darkice ];, and a environment.etc."darkice.cfg".text
<immae>
ok
<immae>
This seems all legit to me too, I wouldn’t expect a rebuild except if it’s missing from upstream
<cfricke>
Ahoy, can't find in the IRC logs, nor in the github issues any mention about the most recent unstable channel update causing timesyncd to malfunction. Anyone experiencing the error where timesyncd is looking for /var/lib?
psy3497 has quit [Ping timeout: 248 seconds]
inquisitiv3 has joined #nixos
ericsagnes has quit [Ping timeout: 252 seconds]
<inquisitiv3>
In my `/etc/nixos/configuration.nix` file I import a Nix file with `imports = [ ../../home/user/.home/modules/roles/fat-client/system.nix ];`, but when I try to rebuild the system configuration I get the error message "error: getting status of '/home/user/.home/modules/modules/roles/fat-client': No such file or directory".
<inquisitiv3>
Why does Nix add an extra "modules/" in the path?
<inquisitiv3>
Same problem when I use the absolute path.
[Leary] has joined #nixos
Lears has quit [Read error: Connection reset by peer]
kaliumxyz has joined #nixos
ericsagnes has joined #nixos
cjpbirkbeck has quit [Quit: Quitting now.]
cjpbirkbeck has joined #nixos
avn has joined #nixos
iyzsong has joined #nixos
<clever>
inquisitiv3: are there any symlinks along the path?
toppler has quit [Ping timeout: 248 seconds]
mexisme_ has quit [Ping timeout: 245 seconds]
<inquisitiv3>
clever: I fixed it. I had an import expression in the file I imported, the that path had also changed.
<clever>
inquisitiv3: ah
xelxebar has joined #nixos
cfricke has quit [Quit: WeeChat 2.5]
orivej has joined #nixos
mexisme_ has joined #nixos
toppler has joined #nixos
kaliumxyz has quit [Ping timeout: 246 seconds]
<Athas>
How do I write derivations that use OpenCL on macOS? Is there an existing one I could crib?
<Athas>
The problem is that ocl-icd does not build on macOS, so I need access to the system-level OpenCL (or a functional copy; not sure how Nix works on macOS).
<tobiasBora>
immae: ahah indeed, I copied the wrong line ^^
<tobiasBora>
clever: here is the last command /nix/store/9in24jb883n8ddzb6x05v50klgx7v71c-hplip-3.19.1.drv
<tobiasBora>
(output)
<tobiasBora>
What is it supposed to do?
<tommyangelo>
can anyone point me to an example of a nix package which downloads a deb package and installs it, where the deb package has a systemd service file in it?
<etu>
tommyangelo: Nix doesn't copy the systemd service files, it uses nix services to create those itselfs
clacke_movim has left #nixos [#nixos]
Lears has joined #nixos
<tommyangelo>
etu: ok thanks, I'll do that then
[Leary] has quit [Read error: No route to host]
<tobiasBora>
"Warning: the new NixOS configuration has an ‘init’ that is incompatible with the current configuration. The new configuration won't take effect until you reboot the system." ==> funny, never had this error before ^^
<{^_^}>
[nixos-homepage] @edolstra merged pull request #277 → Reviews the nixos.org main page → https://git.io/fjaaY
<gchristensen>
:o samueldr
Fzer0 has joined #nixos
Fzer00 has joined #nixos
Fzer0 has quit [Client Quit]
Fzer00 has quit [Client Quit]
dansho has quit [Ping timeout: 272 seconds]
dansho has joined #nixos
<tobiasBora>
Hum
<tobiasBora>
I'm trying to understand why sage keeps using an old version (8.3) despite my update this morning
<tobiasBora>
when I type "which sage", I get ~/.nix-profile/bin/sage. And it appears that this folder ~/.nix-profile/bin contains a single file, which is sage
<tobiasBora>
what is this folder for?
<tobiasBora>
Why is there only one file (sage) in this folder?
<gchristensen>
that sage was installed by using nix-env to install sage
<inquisitiv3>
I'm trying to understand the `xsession.windowManager.xmonad.config` option in Home manager. The function accepts either a null or a path, but the example shows with `pkgs.writeTest` and a string. T
<asymmetric>
is there a way to make `nix-shell --pure` the default?
<inquisitiv3>
The documentation also says it should be an absolute path, but when I look at the documentation for the type it's either an relative or absolute path.
Matthieu_ has joined #nixos
<andi->
asymmetric: a shell alias? Probably not what you were hoping for
<andi->
tobiasBora: welcome to the glorious new world of system v242 :) The protocol changed so that requires a reboot..
hoijui has quit [Ping timeout: 248 seconds]
<Shoubit>
alex```, "what -A is for in that context?, nix-env --install <package> vs nix-env --install --attr <package>?" -- in this case it allows you to select a different channel, in the example I posted, `nix-env -iA unstable.rustc`, it would get rustc from the unstable channel if you've aliased it to "unstable"
<Shoubit>
Late reply, but hope it answers your question!
toppler has quit [Ping timeout: 272 seconds]
<{^_^}>
[nix] @CSVdB opened pull request #2950 → Cleaned up HandleLockFile and the flags to steer it → https://git.io/fjaVY
<{^_^}>
[nixos-homepage] @davidak opened pull request #283 → Remove Nix link → https://git.io/fjaVg
toppler has joined #nixos
<jD91mZM2>
Is there a nix-shell alternative to `nix-build file.nix -A attribute`?
<jD91mZM2>
That adds the attribute's outputs to path, that is
<tommyangelo>
where is `callPackage` defined? I am importing a file that declares a systemd service, but it needs access to the package file, so I want to use `callPackage`. But I get 'error: attribute 'callPackage' missing'
<jD91mZM2>
tommyangelo: It's supposed to be there on the root
<{^_^}>
[nixos-homepage] @edolstra pushed 2 commits to master: https://git.io/fjaVy
<tommyangelo>
hmmm, I have { imports
cfricke_ has quit [Quit: WeeChat 2.5]
jmeredith has joined #nixos
aristid has quit [Ping timeout: 248 seconds]
<tommyangelo>
hmmm, I have `{ imports = [ ./myservice.nix ]; }` in my base config. Then in myservice.nix I have `{ }: let mypackage = callPackage ./mypackage; in { systemd.services.myservice = { enable = true; # other stuff }}`
<clever>
tommyangelo: ah, you want pkgs.callPackage and { pkgs, ... }: as your first line
<jD91mZM2>
tommyangelo: You need to 1. "import" it by adding it as a function argument like so `{ callPackage }: ` 2. you need to add an empty {} after you callPackage call
jgt1 has joined #nixos
<tobiasBora>
I wanted to know, what is the regular way to integrate python packages (numpy...) into a jupyter notebook? For now I installed jupyter with "environment.systemPackages = with pkgs; [jupyter];", but I'm not sure how to add custom python
jgt has quit [Read error: Connection reset by peer]
<tobiasBora>
hum pkgs.jupyter.override seems the solution
<gchristensen>
samueldr: hmm I'm sort of missing the call to action the prior homepage had
<JosW>
would be the thirst time I'll do this and am searching for some kind of howto
<etu>
JosW: Sometimes projects write in their docs what things they need
<etu>
It's very different for different projects
<JosW>
but i'll give it a try
<tobiasBora>
let's try with notebook instead
<JosW>
etu: spec files you mean?
mmercier has joined #nixos
<etu>
JosW: No, sometimes projects writes in their readme, wikis, install guides, whatever which deps they need
<JosW>
Ok, gonna start with it some time :) I'll go for it and see what happens
<tobiasBora>
ok, notebook seems to be the solution
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<tommyangelo>
I'm getting 'cannot coerce a set to a string' on this package now. Paste https://pastebin.com/yX8MGDXH What am I missing?
<clever>
tommyangelo: src is a set, when it should be a string, you want to run fetchurl there
<tommyangelo>
oh O_o it was definitely there before
<mmercier>
Hi all! I'd like to copy a list of package into container in a specific place /var/lib/something but I'm always facing a `cannot coerce a list to a string` error...
<mmercier>
It seems that the derivations on my list are not instanciated and do not have an `outPath`
<clever>
mmercier: can you pastebin your nix exprs?
<{^_^}>
cannot coerce a set to a string, at (string):255:2
<Miyu-saki>
TIL.
<{^_^}>
[nixpkgs] @veprbl merged pull request #63042 → waf: Added support for optional extra tools → https://git.io/fj2TR
<{^_^}>
[nixpkgs] @veprbl pushed commit from @xbreak to master « waf: Added support for optional extra tools (#63042) »: https://git.io/fjar9
<clever>
> "${ { outPath = "foobar"; } }"
<{^_^}>
"foobar"
mmlb has joined #nixos
<Miyu-saki>
Oh shit lmao
<Miyu-saki>
Way simpler.
<clever>
> let f = { __functor = self: arg: self.a + arg; a = 40; }; in f 2
<{^_^}>
42
<Miyu-saki>
mmercier: What do you mean by "nested string"?
<clever>
Miyu-saki: sets can also be functions!
<Miyu-saki>
clever: Yep, and there's also a __toString or something, right?
<clever>
> let x = { __toString = self: "a${self.b}"; b = "42"; }; in "${x}"
<{^_^}>
"a42"
<Miyu-saki>
There we go. :)
o1lo01ol1o has quit [Ping timeout: 244 seconds]
<Miyu-saki>
mmercier: Also, what are you trying to concatStringSep?The variables you're using are not in the pastebin.
<clever>
mmercier: what happens if you just use `lib.concatStringsSep " " [ ui-image api-image ]` ?
<mmercier>
Miyu-saki: I mean that the string output by this then evaluated in a script
<mmercier>
I don't understand, it works on simple example but not on my full implementation...
<mmercier>
I'll figured out thx guys!
<clever>
mmercier: can you pastebin more of the nix files, so i can see everything that is going on?
<mmercier>
It is a bit large
<clever>
mmercier: at least how this file is being called, and given ui-image and api-image
<mmercier>
with a callPackage
<clever>
mmercier: and how are you passing it images?
<mmercier>
humm ok
zacts has quit [Quit: WeeChat 1.9.1]
domogled has joined #nixos
<mmercier>
Ok I find out what was the problem: one of the elements in my list was in fact a set of derivation instead of one derivation... (trailling "s")
<mmercier>
sorry guys
<mmercier>
in the end the `pkgs.lib.concatStringsSep " " my-list` is the right way to do this
<clever>
mmercier: stdenv.lib is another way to get lib
<clever>
mmercier: i also try to avoid using pkgs when callPackage is at work, it makes .override harder
<mmercier>
oups no the right way is actually "echo ${pkgs.lib.concatMapStringsSep " "(x: "${x}") mylist}"
<clever>
mmercier: the map shouldnt be needed, since derivations turn to strings automatically
kaliumxyz has quit [Remote host closed the connection]
<Ankhers>
Okay, so I just uninstalled and am trying ot reinstall nix. Running `nix-shell -p nix-info --run "nix-info -m"` gives me a permission denied error trying to open /nix/var/nix/db/big-lock. I did the multi user install. Anyone know what I can do about this?
<nilsirl[m]>
How can I run firefox on wayland natively?
<nilsirl[m]>
(without xwayland)
<gchristensen>
MOZ_ENABLE_WAYLAND
<gchristensen>
=1
<nilsirl[m]>
@gch
<nilsirl[m]>
Is there a setting to put in configuration.nix?
dansho has quit [Ping timeout: 246 seconds]
<gchristensen>
yeah, there is a way to set environment variables
<asymmetric>
how do i see the difference in what's included between nixos-unstable and nixos-unstable-small?
<nilsirl[m]>
gchristensen: How are you supposed to know the configuration settings. In the manual, it says there is no way to query the configuration. But even my reading the .nix file for firefox I am unable to understand
<jD91mZM2>
I'm getting "Cannot write: No space left on device" from Nix trying to fetch a large binary into a store path, even though I have about 100G free after running some commands to clear garbage
<jtojnar>
clever: I have i7 laptop and downloading anything worked fine until recently. Do you really think that could be issue outside something like Pi?
<wucke13>
gchristensen: How are the keys for the signing of hydra/the binary cache stored and accessed?
domogled has quit [Quit: domogled]
<codedmart>
I am trying to figure out why with xmonad/hidpi my cursor is half the size with no apps open. Correct size with some apps such as gnome-terminal, chrome, but half the size with nautilus.
<Ankhers>
Are there any books that go into a deep dive of nix?
<codedmart>
I feel like in Arch I did something with dbus or merging similar configs that gnome does before launching xmonad, but that was the wrong solution.
<codedmart>
Anyone that can help me with my small cursor problem in xmonad?
<slabity>
Ankhers: The manuals goes pretty deep. Is there something specific you were looking for?
<slabity>
codedmart: Can you try setting the `Xft.dpi` variable in your Xresources? That would narrow it down which apps are seeing the correct value
<codedmart>
slabity: I have that set already to 192.
<codedmart>
Gnome seems to work fine all around. With xmonad it is only small when no apps are open or only with certain apps. Some apps once the mouse is over them the cursor is the correct size.
<filedesless>
says it can't find module XMonad on my side. anything I must configure?
<Ankhers>
slabity: Nothing specific. I just feel like I still don't fully grawk how everything works. I don't know how much of the manual I have properly gone thorough. I will start with that. But I am also trying to build up the BEAM ecosystem within nix and I can't really seem to find much of a "best practices" for that type of stuff.
<slabity>
nilsirl: Do you get any `Gtk-WARNING` errors? Do you only get the locale error?
<filedesless>
thanks a lot, I'll look into this
<nilsirl[m]>
it's Gtk-Warining
<nilsirl[m]>
It talks about something about using C as default
<nilsirl[m]>
because my locales aren't set
<slabity>
Locale not supported by C library?
<nilsirl[m]>
yeah
<nilsirl[m]>
smth like that
<nilsirl[m]>
And it's a Gtk warning
<slabity>
nilsirl: Variables LC_ALL, LANG, and C_CTYPE are all set to "en_US"?
<slabity>
Or whatever your locale is?
<nilsirl[m]>
LANG is set to en_UK
<slabity>
That's weird
<slabity>
Nothing that says "Error: no display specified"?
<nilsirl[m]>
Nothing talking about the fact I'm on wayland
<nilsirl[m]>
or about display
alexherbo2 has joined #nixos
<nilsirl[m]>
also the time it worked (with xwayland) I had the same locale warning so it's probably unrelated
Guanin has joined #nixos
<slabity>
Unfortunately the only thing I can think of now would be to run strace and probe information. But I'
<nilsirl[m]>
Is there anything you have to activate (like a systemd unit) for xwayland work? Because the only thing I did the time it worked I just add xwayland to systemPackages
<slabity>
nilsirl: I believe there's an xwayland service in the wayland-overlay I posted earlier that should take care of that
<nilsirl[m]>
slabity: I didn't use the wayland overlay
<colemickens>
However, the existing displaymanager options do more or less work for Wayland
<colemickens>
Gdm, for example, boots with Wayland by default, boots into a Wayland session. Sway just works out of the box, though we could introduce a service for it
<slabity>
colemickens: displayManager options are all exposed behind `services.xserver`.
<Yaniel>
and there doesn't seem to be such an option for kde
<colemickens>
<codedmart "slabity: Oh I think we are confu"> (The cursor size issue is covered on the away wiki and the provided workaround worked for me)
<colemickens>
Yeah, the option naming is weird right now. Yaniel, it's for "plasma" it should be in there.
<codedmart>
colemickens: The away wiki?
<Yaniel>
colemickens: do you mean that "plasma5.enableQt4Support" option
<colemickens>
Sway, sorry, on mobile
<Yaniel>
because that's the only one I see
<colemickens>
I guess maybe I don't know what option you were talking about... I thought you just meant to get a Plasma/Wayland session going.
<{^_^}>
[nixpkgs] @timokau pushed commit from @r-ryantm to master « jmol: 14.29.42 -> 14.29.46 (#63277) »: https://git.io/fjaXj
<filedesless>
devalot: I can build the project with cabal now, with my nix-shell loaded by direnv, but flycheck still complains
<nilsirl[m]>
slabity: going to restart into nixos to try and figure it out
<colemickens>
I guess I haven't booted my plasma config lately, but iirc it's the same as gnome. Just enable the xsever option, pick the Wayland session at the login manager
<marek>
any reason of the location of this package? lib/testing.nix: jquery = callPackage ./testing/jquery.nix { };
<codedmart>
colemickens: The cusor size works fine in most gtk apps, gnome-terminal, chrome, etc. It is super small in xmonad blank (no apps open) and also super small in nautilus.
o1lo01ol1o has quit [Remote host closed the connection]
<colemickens>
Yaniel ugh, it may be that plasma wayland support is still bad enough that no one has bothered wiring it :(
mexisme_ has joined #nixos
<colemickens>
codedmart that would probably be the default Xcursor size then, which is in a different place as I recall. But I'm guessing you've editted the relevant X file as well...
Anton-Latukha has quit [Ping timeout: 268 seconds]
<codedmart>
colemickens: I am not sure which files you are referring to. I have dpi and Xcursor.size set in ~/.Xresources.
<codedmart>
I have the correct cursor size set in theme settings.
<devalot>
filedesless: I use the flycheck checker that comes with the dante Emacs package instead of the stock ghc checker. To use the sock checker you need to make sure you have the full GHC package database in your shell.nix. Let me track down an example.
<devalot>
filedesless: That's the old way I did things. Now I just use dante and have it use my nix-hs tool to get a ghci repl. It then uses that repl to do flychecking.
<filedesless>
dante provides an interactive repl?
<slabity>
Can you try running `xrdb ~/.Xresources` and then run your programs?
<filedesless>
from my understanding it was merely just executing comments
<codedmart>
It doesn't seem like it should be this hard to change a cursor size in xmonad/xorg.
<slabity>
Does that do anything?
<codedmart>
slabity: Nope if I run that from command line then run nautilus the cursor is still really small.
gratto has joined #nixos
<slabity>
codedmart: Instead of `Xft.dpi`, can you try `*.dpi`? Then run the `xrdb` command to reload it and then run the apps
fendor has joined #nixos
<arianvp>
Huh
<arianvp>
Since when does gnome3 on NixOS run Wayland?
<devalot>
filedesless: I don't think dante provides an *interactive* repl, but it does start ghci so it can do flycheck and completion.
Matthieu_ has quit [Ping timeout: 272 seconds]
<codedmart>
slabity: Doesn't seem to have had any effect.
cantstanya has quit [Remote host closed the connection]
work_ has joined #nixos
work_ has quit [Client Quit]
cantstanya has joined #nixos
<filedesless>
oh I see
Soo_Slow has quit [Remote host closed the connection]
logand has quit [Remote host closed the connection]
<slabity>
codedmart: Do you happen to use an Nvidia card?
<filedesless>
followed your older example for the shell.nix, and now I can see xmonad when I type `ghc-pkg list`
<filedesless>
yet flycheck still seems unable to find XMonad
<matthewbauer>
it should be ready to merge soon, but I was hoping more people could try it out to find regressions
<fdsffgdd>
slabity: thank you for the help. I can now confirm that the time it worked it was because of xwayland which means that I still need to figure out how I can run firefox under wayland natively (sway). (I am NilsIRL)
Anton-Latukha has joined #nixos
Anton-Latukha has quit [Client Quit]
<devalot>
filedesless: And Emacs is using direnv to load the shell.nix environment?
<pie__>
i havent had time to clean it up more but i was trying to make a better user-facing interface for language stuff :V maybe provide a bit of a framework for packaging as well
<pie__>
so if youre feeling adventurous...
<pie__>
Ankhers, i meant to link the master branch
lemonpie has left #nixos ["ERC (IRC client for Emacs 27.0.50)"]
<Ankhers>
pie__: I will take a look at that. If you have any suggestions on how to go about getting better Erlang support into nix, I would love to discuss things.
phreedom has quit [Ping timeout: 256 seconds]
<pie__>
Ankhers, im just a semi-active low tier guy still learning stuff :c no idea
o1lo01ol1o has joined #nixos
<manveru>
,locate bin context
<{^_^}>
Found in packages: texlive.combined.scheme-medium
<pie__>
i think we need to start a tip jar for infinisil :P
hex3 has joined #nixos
hex3 has quit [Remote host closed the connection]
<infinisil>
Haha nah, the work for ,locate is being done by bennofs[m]'s nix-index :)
srid63908 has quit [Ping timeout: 246 seconds]
hex2 has quit [Ping timeout: 246 seconds]
hex3 has joined #nixos
<pie__>
infinisil can share the tip jar then :p
alexherbo2 is now known as alex``
<{^_^}>
[nixpkgs] @alyssais opened pull request #63373 → linux: multiple TCP-based remote denial of service vulnerabilities → https://git.io/fjaMP
<pie__>
i think it might be summarizable as another little embedded overlay system inside the nixpkgs tree , with a consistent user facing API
<pie__>
at least that would be the goal. some people might have something to say about influencing eval time, idk... (as long as nothing core depends on it its prooobably fine i guesss?) - and i havent benchmarked
hex3 has quit [Ping timeout: 246 seconds]
xkapastel has quit [Quit: Connection closed for inactivity]
Soo_Slow has joined #nixos
ambro718 has quit [Quit: Konversation terminated!]
<{^_^}>
[nixpkgs] @timokau pushed commit from @r-ryantm to master « linbox: 1.6.2 -> 1.6.3 (#63288) »: https://git.io/fjaDW
toppler has joined #nixos
<primeos>
arianvp: Just added a few extra notes to the issue from what I remembered (regarding the session file for Sway), in case it helps a little bit ;)
<codedmart>
slabity: OK so not sure if this is helpful, but I was using xsettingsd. I took that out and tried using gsd-* settings daemon stuff from gnome. Everything is tiny then. UI, fonts, cursor.
<codedmart>
I have tried a number of different settings, but it seems I end up with weird combinations.
<codedmart>
Not sure what the best approach to handle HiDPI is these days and especially in nixos.
<codedmart>
Like I said gnome works fine when I use gnome. But I want those same settings in xmonad.
<{^_^}>
[nixpkgs] @alyssais pushed 5 commits to release-19.03: https://git.io/fjaDd
<colemickens>
The best approach for hidpi is undoubtedly wayland
<colemickens>
I can prettyl Super+Ctrl+Alt Plus or Minus and step up/down my entire PER-monitor scale factor by .1 increments.
<codedmart>
xmonad does not work with wayland.
<codedmart>
HiDPI works fine in gnome with xorg.
<qyliss>
dfranke_: (person who just updated linux here) I didn't bother opening an issue because the change was so simple
<colemickens>
I mean "fine" as long as you don't want fractional, or per-monitor scaling.
<colemickens>
Maybe I should package waymonad for the overlay...
<dfranke_>
codedmart: I set `services.xserver.dpi = 192` and that's worked fine for me.
<codedmart>
I don't want fractional or per monitor.
<codedmart>
dfranke_: what card do you use? I have that set and see no change.
<dfranke_>
codedmart: some boxes are embedded Intel GPUs, others are NVidia Quadros with the proprietary drivers.
<dfranke_>
WFM on both.
morr has joined #nixos
<codedmart>
dfranke_: Do you use xmonad? Scaling is fine out of the box?
* colemickens
wonders if ~/.Xresources has been editted
<dfranke_>
codedmart: I use i3, but I used to use xmonad and it worked then too.
<codedmart>
It has been many times.
acarrico has joined #nixos
<dfranke_>
though on the box where I used to do that I still have `displayManager.sessionCommands='xrandr --dpi 192'`, because the above option didn't exist yet.
lovesegfault has quit [Ping timeout: 252 seconds]
<dfranke_>
So I guess try that spelling and see if it changes anything.
<codedmart>
dfranke_: Sorry stepped out for a min. Not sure if you responded. I am really at a loss here. If I remove all settings except dpi stuff in configuration.nix. Fonts look fine and non gnome gtk apps seem fine, but all gnome apps the fonts are good but UI and cursor is tony.
<codedmart>
s/tony/tiny/
<bsima>
infinisil: i see
selfsymmetric-mu has left #nixos ["gone to the land of dead hiccups and extinguished light bulbs"]
ivan has joined #nixos
aquad has joined #nixos
selfsymmetric-mu has joined #nixos
fendor has quit [Read error: Connection reset by peer]
<bsima>
> :p builtings.tryEval (abort "hi")
<{^_^}>
undefined variable 'builtings' at (string):255:1
<bsima>
> :p builtins.tryEval (abort "hi")
<{^_^}>
evaluation aborted with the following error message: 'hi'
<codedmart>
Is there a global xresources or xdefaults? I don't see any options when searching nixos options.
<bsima>
so i should prefer 'throw' to 'abort' when reporting errors in nix code
aquad has quit [Remote host closed the connection]
<bsima>
codedmart: does 'services.xserver' have what you want?
<codedmart>
bsima: I am not seeing what I am looking for there.
<codedmart>
Oddly enough I have set `GDK_SCALE` in sessionCommands. But the cursor is still small until I open lxappearance. Then the cursor is the correct size in xmonad blank screen and in gnome apps.
srid63908 has quit [Read error: Connection reset by peer]
srid63908 has joined #nixos
acarrico has quit [Ping timeout: 245 seconds]
<colemickens>
Are you running xmonad with dbus-launch?
<colemickens>
That's probably related.
<colemickens>
Or using a login manager?
<colemickens>
GDK_SCALE does not affect the cursor size...
<codedmart>
colemickens: I am using gdm.
<lfish>
hello, when updating from 18.09 to 19.03 nixos stopped recognizing the wi-fi drive, I'm on a lenovo thinkpad v310, any ideas what may be causing that? (I rolled back to 18.09 and wifi is working just fine)
<colemickens>
hmmm
<codedmart>
Yeah I am confused by how hard this seems.
<codedmart>
And why opening lxappearance changes the size of the cursor to the size I want even after I close lxappearance.
<colemickens>
I'm betting it's doing the same thing as the sway workaround, and calling setting some values in an in-memory gsettings backend that only apply to that session. (hence why its in my sway startup, etc).
<codedmart>
What workaround is that? I couldn't find that.
<colemickens>
Though, I don't really fully understand why, I'd have assumed that lxappearance would also write to the INI files that are supposed to still be respected under X11? Unless your config files are bad, but the config that lxappearance applies is good?
<colemickens>
It's what finally fixed my cursor woes after months of a STUPIDLY small cursor, it was so frustrating, heh.
<colemickens>
but again, from the text, it seems like it shouldn't affect you, but the symptoms are too similar to shake this feeling
<codedmart>
What did you set those values 2?
<codedmart>
Trying that from the command line I am not seeing results.
<colemickens>
(you'd need to relaunch, I think?)
<codedmart>
OK brb
<colemickens>
from my ~/.config.gtk-3.0/settings.ini: among other things: "gtk-cursor-theme-size=128
<colemickens>
but even that ticks me off, it should be based on the scale of the output that the window resides on, it's really baffling how the GTK folks haven't cleaned this up IMO
<codedmart>
Hmm... nope
ajs124 has joined #nixos
<colemickens>
are you still running your xsettings daemon?
<codedmart>
Probably should I not?
<codedmart>
Do you run anything
<colemickens>
idk, gtk won't read those files if the xsettings daemon is running is the only reason I asked.
<colemickens>
I don't run anything other than that gsettings import script on sway start to fix my cursors in gtk3 apps.
<colemickens>
I don't really know what the "state of the art" is for xsettings in non-gnome x11 wms :(
rprije has joined #nixos
<colemickens>
It certainly seems annoying. Would be great to write this up for nixos.wiki at least when you do get it :)
<codedmart>
I really don't understand what is going on now. Somewhere along the lines I have messed everything all up. Gnome is off now and xmonad is off.
<codedmart>
I can't believe how hard this seems to be just to get HiDPI working properly across a few different destop/window managers.
goibhniu1 has quit [Ping timeout: 248 seconds]
LandShark1488 has joined #nixos
<monokrome>
hiii
mexisme_ has quit [Ping timeout: 245 seconds]
<monokrome>
Does anyone know why my install fails saying "Failed to probe partition scheme: /boot"
<monokrome>
I'm guessing maybe it has to do with having a LUKS root, but I've never seen it before this?
<jackdk>
did you mount an unencrypted /boot as well?
<monokrome>
nope
<monokrome>
The root, boot, and home are all on a LUKS volume
<monokrome>
using LVM
<monokrome>
I have ` preLVM = true; ` in the filesystem config for the LUKS
<monokrome>
I have 3 other laptops w/ that same structure and they worked fine O_o
<jackdk>
okay, well you've gone beyond what I know about this stuff then. Sorry.
<monokrome>
ah, ok
<monokrome>
well thanks :)
<codedmart>
colemickens: Do you have your configs somewhere. Maybe I will try sway to see if wayland is nicer to me along side gnome.
<colemickens>
You don't necessarily need to use the overlay, the way I do there, if you just want to use Sway. But if you want the bleeding edge, https://github.com/colemickens/nixpkgs-wayland will show you how to utilize it (since I use my own function from lib.nix)
<codedmart>
Sway wasn't added to my gdm list.
<colemickens>
We haven't packaged a session file for it and AFAIK upstream doesn't either.
<codedmart>
Ah I see
<codedmart>
OK
<colemickens>
from a TTY: `systemctl stop display-manager; sway`
<colemickens>
should get you going, I think?
<colemickens>
I don't remember what nixpkgs sway module does for the config file, but the upstream git repo has a config file you can copy as a default. I guess you're familiar with xmonad, but if you don't have urxvt you may not get far.
alp has joined #nixos
<codedmart>
Yeah I used i3 for a bit a while ago. It is more nixos stuff I am trying to understand at this point.
<colemickens>
for the most part, Sway just works when run from the TTY directly, but if you're using services.sway.enable, then you get a little bit of a wrapper that runs it in a dbus session, etc.