<clever>
iqubic: then put mono in the buildInputs and use buildPhase to run it on the src
<Guest36616>
and here I was gonna suggest a bash alias or something :\
<iqubic>
clever: Are there tools that do this for me?
<iqubic>
Like how cabal2nix just creates a nix file for me?
<clever>
iqubic: dont know of any
<iqubic>
Also, I think I'll need to use `mcs` as the compiler
<Guest36616>
!bot help
<clever>
,locate ,bin mcs
<{^_^}>
Couldn't find in any packages
<clever>
,locate bin mcs
<Guest36616>
ah thx
<{^_^}>
Found in packages: mono, mono4, coreclr
<iqubic>
Nice. That's what I'll need.
<Guest36616>
,locate'.*2nix'
<Guest36616>
,locate '.*2nix'
<{^_^}>
Couldn't find in any packages
<clever>
,locate 2nix
<iqubic>
I'm not sure that's how that works.
<{^_^}>
Couldn't find in any packages
<Guest36616>
aw does our good robot boi not do regex?
<Guest36616>
,help
<{^_^}>
Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<Guest36616>
,
<{^_^}>
Special commands: find tell locate - Commands sorted by use count, page 0 (use ,<n> to view page <n>): tofu help libraries unstable unfree library escape'' declarative escape" pr overlay ping profiling ask howoldis pinning imperative which-channel xy wololo exec fancy-uninstall nixlang++ pills -a cache callpackage channels haskell home-manager loot smart-questions stateversion stuck ifd tias paste runtimedeps timer cloak dontask error flakes nur
<PyroLagus>
hm. i wonder how feasible it would be to build a nixos system with overlayfs rather than symlinks and what the limitations and performance implications would be
<gchristensen>
it would involve a lot of layers, and that means a lot of stat calls
<gchristensen>
a problem with overlayfs is it is a series of layers, and Nix represents a full graph
<zeta_0>
i would like this command to be run after every reboot: "light -I", so i am supposed to use: displayManager.sessionCommands in configuration.nix for this, right?
<zeta_0>
it is a hassle always having to reset the brightness over and over again so i want to automate that command?
<PyroLagus>
i suppose the question is how many layers unionfs supports, because the article does point out that you can just flatten the graph
<PyroLagus>
and you don't have to worry about the caching thing either
<PyroLagus>
zeta_0: you could use a one-shot systemd service
<gchristensen>
PyroLagus: the problem remains of the massive stat overhead you'll experience
<gchristensen>
as it has to check every layer
<PyroLagus>
ah
<PyroLagus>
right
<elvishjerricco>
Is there anything in nixpkgs for copying a binary from another derivation to $out, also copying its required libraries, and patchelf'ing the binary to use them? Basically just trying to kill all the unnecessary files in a derivation get solely the binary working.
<gchristensen>
(multiple times, actually, to check for delete markers, etc.)
<zeta_0>
i guess i'll try out displayManager.sessionCommands first, if that does not work i'll try out systemd
gagbo has quit [Ping timeout: 252 seconds]
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.1]
<PyroLagus>
even plan9 seems to be using ordered layers, huh
<PyroLagus>
hmmmm, "To be visible to other software, these files are symlinked into standard locations in the new directory hierarchy under /System/Index" man... it essentially does the same thing as nixos
<PyroLagus>
oh. well.
<PyroLagus>
unless there's some other trick i'm not seeing
rprije has quit [Remote host closed the connection]
kaliumxyz has quit [Quit: leaving]
rprije has joined #nixos
mexisme_ has joined #nixos
<PyroLagus>
"Each program directory (for example, /Programs/KDE) holds version directories (/Programs/KDE/3.4, /Programs/KDE/3.4.2), and a version-neutral directory for settings (/Programs/KDE/Settings), to keep files that would normally be in /etc." rip backwards compatibility breaking updates
rprije has quit [Read error: Connection reset by peer]
<pie_>
PyroLagus, as opposed to not having had a single clue how to textually configure kde stuff this might be progress?
<pie_>
depends on the format i guess
<pie_>
ah wait thats just gobolinux, not kde, eh?
* pie_
had a failure trying to figure out how to configure something in kde at some point and is still salty
<PyroLagus>
another option may just be a script that, given a nix expression, produces a directory with the necessary files generated symlinked that you can then just overlay or bind mount. since the nixos-install and nixos-rebuild scripts do just that plus some extra stuff, it might actually not be all too hard.
o1lo01ol1o has quit [Remote host closed the connection]
<Arahael>
How do I find out how dnsmasq got installed? It's not in the configuration.nix, so something must've pulled it in as a dependency, or I must've nix-env -i'ed it.
<Arahael>
Specifically, this program: /nix/store/lfr3yiqf88vvklx9hjzay1ky687b3mqv-dnsmasq-2.79
<Arahael>
Looks like quite a few applications linked it in.
toppler`` has quit [Ping timeout: 248 seconds]
<PyroLagus>
maybe try nix-store --query
<PyroLagus>
wait no. you need reverse deps
<Arahael>
I ended up re-thinking my approach, the stupid-simple thing to do turned out to be `lsof -i TCP:53`, observing the pid, then `cat /proc/1189/cmdline`
abathur has quit [Ping timeout: 248 seconds]
<Arahael>
That implicated libvirt
<PyroLagus>
ah
ghasshee has joined #nixos
<PyroLagus>
) seems you can get reverse deps with --query using --referrers and --referrers-closure
<{^_^}>
[nixpkgs] @marsam opened pull request #62465 → pythonPackages.pglast: init at 1.4 → https://git.io/fjE0n
<ambro718>
# nixos-rebuild boot building Nix... building the system configuration... error: stack overflow (possible infinite recursion)
<ambro718>
This just started happening, didn't even change the nixos config. I think I only messed with git a bit (deleted some branches and did git gc).
talqu has quit [Remote host closed the connection]
<Orbstheorem>
Hello ^^/, I'm trying to package a go application (https://github.com/juruen/rmapi). I'm following the instructions for packaging Go modules on nixpkgs' manual, but the builder fails complaining it can't find module path for a given dependency <url>. However, I can see that <url> is present and up. I think it's because the secondary builder is trying to pull deps from one of the dependencies of the module I'm
polman has quit [Read error: Connection reset by peer]
polman has joined #nixos
o1lo01ol1o has joined #nixos
realloc has joined #nixos
<realloc>
hi, I am trying out NixOS in a virtual box. It freezes right after booting in the step `running udev`. I tried switching to TTY8, but it does not seem to react. Any ideas?
<hyper_ch>
latest iso?
<realloc>
hyper_ch: Downloaded it a few minutes ago
<hyper_ch>
hmmm, strange
o1lo01ol1o has quit [Ping timeout: 272 seconds]
<hyper_ch>
last time I tried it worked but that was a while ago
<matthewbauer>
realloc: is this 19.03 or 19.09pre?
<hyper_ch>
realloc: nixos makes it easy to generate your own custom iso file :) so I have one for installing purposes that has zfs integrated and and a few installer script for easy deployment
<hyper_ch>
(and especially using swiss german keyboard layout)
<realloc>
hyper_ch: Good to know. Could be very useful
<chris___>
I'm running into an odd issue of after `nix build` it attempts to build a `default.nix` from the resulting nix store path? I'm guessing it might be due to an incorrect call to the `callPackage` function in haskellPackages? Any pointers as to what I could investigate? The nix file I'm using is only 10 lines: https://gist.github.com/60af150ea9d0de4e822a7943defc66c8
<alexarice[m]>
is there a good way to decompress a file fetched with fetchpatch? it uses xz compression
<__monty__>
unxz file.xz?
toppler`` has quit [Ping timeout: 252 seconds]
<alexarice[m]>
__monty__: where does unxz come from, nix can't seem to find it
<__monty__>
The xz program. It's just an alias for xz --decompress.
<__monty__>
Shouldn't nix come with that as a dependency? I thought it used xz for nars.
<alexarice[m]>
__monty__: ah, I was hoping to do this through nix instead of making another derivation for the patch
<alexarice[m]>
__monty__: I'm guessing there's no flag that can be put into fetchpatch to automatically decompress
<__monty__>
Don't know. Check the docs, or the code : )
clacke_movim has left #nixos [#nixos]
<alexarice[m]>
having a look now
<simpson>
alexarice[m]: Instead of `fetchpatch`, you can put the compressed .xz patch directly into the list of patch files, according to the nixpkgs manual.
<simpson>
And then commit the patch into git instead of downloading it over and over.
<alexarice[m]>
simpson: ok I'll try that, thanks
toppler`` has joined #nixos
Matthieu` has quit [Ping timeout: 246 seconds]
o1lo01ol1o has joined #nixos
<alexarice[m]>
simpson: thanks, that worked
<simpson>
alexarice[m]: Excellent, glad to hear it.
Soo_Slow has quit [Remote host closed the connection]
<user____>
How would one go about installing 32 bit libraries, such as libexpat1, libffi6, libfontconfig1 etc?
ym555 has joined #nixos
<simpson>
user____: Probably not. What are you working on which requires them?
<user____>
Running a 32 bit elf
<realloc>
samueldr: the guest 19.03pre is running 4.19.47. Host is running 5.1.6 but that should probably be of no concern. Regarding 19.03 I need to check...
<samueldr>
and when you say "19.03pre", which revision is it?
<samueldr>
because it's going to be `pre` for any build not done through the hydra instances
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<user____>
simpson: Thanks! My installation was successful. When running the application, i got an error while loading shared libraries: libgobject-2.0.so.0 I searched nixpkgs for libgobject and did not see one. What is the recommended way to proceed? manually install it?
zeta_0 has joined #nixos
o1lo01ol1o has quit [Ping timeout: 248 seconds]
<user____>
simpson: I added "glib" and that solved it.
<simpson>
user____: The recommended way, frankly, is to build the package from source, but I'm guessing that it's closed-source.
<samueldr>
are you using a newer VirtualBox release?
<samueldr>
that might be relevant
<realloc>
samueldr: I am using version 6.0.8-2
<realloc>
sooo yes
<samueldr>
I have nothing to back this, and no solution, but I seem to recall that there is or was an issue with recent kernels and newer VirtualBox versions
<samueldr>
that doesn't help you, but just saying that it might be a situation where you're not doing anything wrong
<realloc>
samueldr: Thank you for your help nonetheless. I think I will continue with 19.09pre then for now
o1lo01ol1o has joined #nixos
<samueldr>
though, the fact that 19.09 boots means it's likely there is something that could be done to make 19.03 work
<gchristensen>
samueldr: that background image is straight out of Mac OS Classic!
domogled has joined #nixos
root3 has joined #nixos
<gchristensen>
gosh I am brimming with nostalgia. the whole thing is too perfect
<samueldr>
gchristensen: part of my personal "my setup is better than yours" collection :)
<gchristensen>
what is your WM?
<samueldr>
awesome
<gchristensen>
you bet it is
Guest88 has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
klntsky has quit [Remote host closed the connection]
<o1lo01ol1o>
if i have a unpackCmd = ''tar xvf $curSrc''; in mkDerivation and I expect a source folder /foo to be unpacked from that, how do I cd to /foo in the install phase? I can't seem to figure out where things are
klntsky has joined #nixos
<clever>
o1lo01ol1o: you need to set sourceRoot="foo";
<clever>
oh, but your setting unpackCmd, so unpackPhase should do that for you...
<o1lo01ol1o>
Well, when I do that, it says foo is not a directory
<o1lo01ol1o>
but if I manually unpack it outside nix, it unpacks to foo
<clever>
try unpackCmd = ''tar xvf $curSrc ; ls -l'';
<simpson>
root3: Interesting. As a first attempt at reproduction, I can't see this with `pkgsi686Linux.git` on an amd64 machine, so maybe it's something specific to i686-native building? Wonder how that happened though. FWIW the test failures don't seem trivial, but I am not familiar with this source package's tests.
<clever>
simpson: it may be on a nixpkgs rev that has since fixed it, or not yet broken
<simpson>
user____: Huh, GTK+ *and* Qt? Sounds like quite the difficult package!
Matthieu` has quit [Ping timeout: 248 seconds]
<o1lo01ol1o>
clever: interesting, thanks. I didn't know that was possible.
o1lo01ol1o has quit [Remote host closed the connection]
<root3>
I'm trying it again with the newest version as of ~30 min ago
<root3>
If it's like last try it'll take ~90 min
<root3>
also worth noting is that it tries to build version 2.19 (i think)
Thra11 has joined #nixos
<root3>
when I x-compile git on amd64 it only ever builds 2.21
<clever>
root3: thats more down to your nixpkgs rev, then your arch
<clever>
root3: nix eval nixpkgs.lib.version
<realloc>
is someone here using services.nextcloud?
<Guest88>
Does anyone here know what I'm supposed to do?
<simpson>
Guest88: I would consider avoiding automation of pypi2nix. Ideally it is only run once per package that is being brought into nixpkgs. (Ideally PyPI would export better metadata and we would not have pypi2nix at all~)
<{^_^}>
[nixpkgs] @eadwu opened pull request #62528 → compton: try not to conflict wintypes → https://git.io/fjErX
<simpson>
Guest88: With that in mind, you would probably aim to have a derivation for dvc. What are you working on? What's your ultimate goal?
<Guest88>
Yeah I'd like to setup a derivation for dvc
<Guest88>
That's my ultimate goal
<Guest88>
(To get dvc installed, and do it using nix so I can learn nix along the way)
<root3>
@clever am I running that on my i686 machine or x86-64 one?
<clever>
root3: try both
<Guest88>
I kinda assumed that `nix run pypi2nix` is roughly the same as `nix-env --install pypi2nix && pypi2nix` except it doesn't install pypi2nix globally.
<nh2>
Is ther a good way I can ask the rescue mode's systemd, udev or whatever, what network interfaces will likely be named as under the predictable scheme?
<nh2>
the rescue mode's apt doesn't know `biosdevname`, but also on my desktop I get:
<nh2>
so that seems the other way around somehow
<nh2>
em1
<nh2>
$ sudo biosdevname -i enp0s25
<nh2>
> The new scheme is apparently called "predictable network interface names" but I'm not finding a lot of documentation on how to actually predict the name
<nh2>
so true
<{^_^}>
undefined variable 'The' at (string):254:1
<clever>
nh2: the pattern is fairly obvious, except for the enp vs ens part
<nh2>
clever: and there's also `eno`
<nh2>
clever: you'd think that if somebody made a concept "predictable names" and pushed it on all the Linux deployments out there, they'd also make a simple tool that outputs that predictable name that people can use
<clever>
nh2: the simplest thing i can think of, is to just boot nixos, and use the rescue console to fix it, lol
<nh2>
but it still shows the various different choices of which systemd can pick one
<nh2>
E: ID_NET_NAME_ONBOARD=eno3
<nh2>
For example
<nh2>
E: ID_NET_NAME_PATH=enp3s0f0
<nh2>
the quoted source code also shows which fallback systemd has between those
<nh2>
still aint cool to have to copy-paste that into my script :|
<root3>
I've found that It's supposed to live in ~/.config/nix/nix.conf, which either I am supposed to make manually or it hasn't been produced. I do have it in /etc as a link to the store, where it warns me that I should be editing configuration.nix instead. Should I just create .config/nix/nix.conf?
<stranger___>
i don't understand error messages look cryptic
<clever>
nh2: does `ip link` only show one IF as being up?
<nh2>
clever: yes
<clever>
nh2: something else ive wanted in nixos, is to just give an ip to the 1st link to be online
<clever>
that would solve your problem until somebody plugs in a 2nd cable
<nh2>
clever: yeah, I also feel the current stuff is way too complicated for most use cases. In many cases you can guarantee how things are *operated*, and there should be easy solutions for those cases
<nh2>
clever: also `/lib/udev/write_net_rules`, so sloppily made: No --help text, just always prints `missing $INTERFACE`
<nh2>
I just don't get it: If you know your software is going to be used a lot, why not take the couple minutes to implement `--help`
<etu>
stranger___: midiutils meta maintainer list should be a list
<nh2>
clever: yes still have to watch that one, already on my list
o1lo01ol1o has joined #nixos
<nh2>
clever: maybe we should just make an init system in Haskell or Rust given that we know now reasonably well how the whole booting thing works and how to make static exes
<clever>
:D
<clever>
you must have seen haskell-init then?
<stranger___>
etu: ah. thank you much!
<nh2>
clever: yes I used it for the debugging the chromebook
<nh2>
clever: even just reimplementing systemd's docs with in a way so that it actually works reliably without surprises might be a big benefit
<clever>
nh2: the next thing your missing then, is firmware loading, either run udev, or re-implement it
<clever>
nh2: there is a fun back&forth between kernel and userland to load firmware files, i once implemented it in bash
<nh2>
I am so tired of defining timers and then finding half a year later that they stop running without reasons or logs
<clever>
nh2: systemctl list-timers
<nh2>
clever: I found this bug that in rare situations the timer just stops working, and N/A being shown in list-timers
<clever>
fun!
<nh2>
that kind of stuff just needs to work reliably, if you take over the job of cron it can't stop working after half a year
o1lo01ol1o has quit [Ping timeout: 244 seconds]
<gchristensen>
that is bizarre, I have had great success with timers
<Guest88>
So I've installed pypi2nix with `nix-env -i pypi2nix` and now I'm running `pypi2nix -V 3.5 -e dvc -E libjpeg`
<Guest88>
it seems to be failing on the libjpeg dependency
<nh2>
gchristensen: maybe you haven't run them enough :D of course ith happened only on 1 of 3 identically configured NixOS machines
<Guest88>
Is "libjpeg" what I want if I need to depend on the header files?
<Guest88>
(if this was apt, I'd usually specify the -dev version)
<gchristensen>
I have several dozen of them ...
<clever>
Guest88: when adding something to buildInputs, nix will get the dev version foryou
<nh2>
gchristensen: how often do they run? Mine run once per minute, maybe something overflowed? We couldn't figure it out within 1 day of investigation with #systemd IRC help :(
<Guest88>
Thanks. Hmm, that's not the issue then. I'll keep plugging at it.
<gchristensen>
oh weird
<gchristensen>
mine are usually every 15min or higher, though I have one which runs every minute
<nh2>
gchristensen: though since 3 machines run them at same intervals, I would expect all 3 of them to overflow if that was the problem
mrus has quit [Quit: Surfing the great wave off Kanagawa]
<nh2>
clever: I also have this other interesting problem that `nixos-install` says "building [thing]" but seems capped at around 3 such things per second, slowing down the install a lot
<clever>
nh2: -j5 may help
<nh2>
with the normal `-j1`, it always counts up to 1 (blue currently building), and then jumps back to zero. With e.g. `-j40` it's not faster, but counts up the blue number to somthing high (usually close to whatever I give in `-j`), and then jumps back to 0
<clever>
nh2: weird
<nh2>
due to this "batching" I wonder if some sqlite thing is going on or so
<nh2>
I haven't straced it yet
<clever>
there is a lot of overhead from sandboxing, and the sqlite does enforce waiting on the RPM of the drives
<Guest88>
Mainly I don't know how to add the meta attribute to default.nix
<Guest88>
Because that just pulls in requirements.nix (which I generated from pypi2nix)
<Guest88>
Any comments on the PR would be appreciated
<Guest88>
I'm logging off now. Have a good night everyone.
Guest88 has quit [Quit: Guest88]
o1lo01ol1o has joined #nixos
root3 has left #nixos ["WeeChat 2.4"]
irl25519 has joined #nixos
irl25519 has quit [Client Quit]
zeta_0 has joined #nixos
irl25519 has joined #nixos
irl25519 has quit [Client Quit]
irl25519 has joined #nixos
fenedor has joined #nixos
<zeta_0>
hello guys, right now i am using ghci to learn haskell but it has no autocompletion or any other features, is there a way to integrate haskell repl in emacs?
Matthieu` has joined #nixos
<inquisitiv3>
What's the preferred way to configure applications what have dedicated configuration files? Should I use the same method as I would do on any other Linux distribution, or is there a "Nix way" to handle such configurations?
<zeta_0>
inquisitiv3: are you using nixos?
<clever>
inquisitiv3: if it accepts a `--config /path/to/foo.txt` then you can just use pkgs.writeText to generate the cfg, and pass it the path
<zeta_0>
oh shoot, i thought i was on the #haskell channel my mistake
fendor has quit [Ping timeout: 252 seconds]
<inquisitiv3>
zeta_0: Not currently, but planning on switching. Currently use Nix on Fedora, but I've run into some bugs.
Guest88 has joined #nixos
Guest88 has quit [Client Quit]
toppler`` has quit [Ping timeout: 246 seconds]
<manveru>
inquisitiv3: there's home-manager
<inquisitiv3>
manveru: I'm using home-manager for installing applications. But couldn't find anything about handling application specific configurations when I looked some hours ago.
<inquisitiv3>
Did I miss something_
<inquisitiv3>
?
<manveru>
it has a ton of options for that
<inquisitiv3>
Ooh
<manveru>
`man home-configuration.nix`
o1lo01ol1o has quit [Remote host closed the connection]
stranger___ has quit [Quit: Connection closed for inactivity]
o1lo01ol1o has joined #nixos
Lears has joined #nixos
<inquisitiv3>
I've read trough that manpage. There seems to be only configuration options for applications listed on the page?
<inquisitiv3>
T
<manveru>
then i'm not sure what you mean :)
<manveru>
got an example?
<inquisitiv3>
I want to configure Kakoune (text editor), but there doesn't seem to be any predefined options for that? Or did I miss something when reading the manpage_
<inquisitiv3>
?
<manveru>
no, it doesn't have everything, but you can write any text file using it
<zeta_0>
infinisil: i am very grateful for you helping me in the past, but you need to `STOP` being so `TRIGGERED` with those `PRETENTIOUS` responses, i already searched the internet, i just wasn't sure what was the best method to do repl with haskell, i am still new to the haskell and nix world, have some patience for the beginners!!!
<alex``>
how to install ruby on rails on nixos?
<alex``>
I see multiple links to bundix but it’s broken
<{^_^}>
[nixpkgs] @Hodapp87 opened pull request #62538 → gzdoom: Fix wrong fluidsynth version → https://git.io/fjEiq
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjEia
irl25519 has quit [Quit: My computer has gone to sleep. ZZZzzz…]
<alex``>
manveru: if I understand correctly, "bundix lock" generate Gemfile.lock, "bundix" generate gemset.nix using the lock, and shell.nix uses gemset.nix?
<manveru>
jup
<manveru>
it's a bit complex unfortunately
irl25519 has joined #nixos
<iqubic>
So can I just put haskellPackages.ChessIO into my system.evironmentPackages list?
<clever>
iqubic: if it has binaries, sure
<iqubic>
It does.
<iqubic>
Three different binaries: polyplay, pgnio, cboard
<alex``>
manveru: in the case I have an existing project but deleted the Gemfile.lock, I have to move the "Gemfile" and create one that bundix understand to have the Gemfile.lock back?
InFlames has joined #nixos
orivej has joined #nixos
<manveru>
yeah, atm you still need to have all three files
<manveru>
and the Gemfile can't just be 'gemspec'
<alex``>
generally Gemfile uses gemspec
<manveru>
a gemspec for a rails app doesn't really give you anything
<manveru>
it's not like you'll push it to rubygems or something
<alex``>
right
ym555 has joined #nixos
<iqubic>
clever: I get the error: "attribute not found chessio".
fendor has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @orivej-nixos pushed commit from @orivej to staging « libevent: 2.1.8 -> 2.1.10 »: https://git.io/fjEPk
<iqubic>
Clever: ChessIO does in fact appear to be broken. See this nixos-rebuild switch output: http://dpaste.com/3PA1FKY
<iqubic>
Clever: Do you know what's up with that issue?
rprije has joined #nixos
<nh2>
clever: OK, the OVH server is working now with a reproducible setup, and using static IP configuration. OVH docs I found so far all suggested that this is their way of doing things. I'll do a public writeup of how to do it so people can save some days.
<nh2>
Under grub, would `boot.loader.grub.mirroredBoots` be for that feature? How does one do it with `systemd-boot`, which is what's running on my server now?
<nh2>
I have one unsolved topic though: Ensuring that the EFI System Partition is mirrored so I can continue to boot through RAID1 disk failure.
<gchristensen>
updated more often because fewer things have to be built for it to update
<gchristensen>
so you will end up compiling more stuff more often
<selfsymmetric-mu>
Oh right, I forgot about that. I keep thinking that it has fewer packages. It doesn't, right, but it builds more things locally?
<DigitalKiwi>
ok thank you
orivej has quit [Ping timeout: 245 seconds]
<nh2>
clever: It also seems problematic that NixOS puts kernels onto the ESP. If it didn't, I could probably just copy systemd-boot once on each ESP and be done with it (they would all boot the same `/`). But the way it is now, all my ESPs needs to be updated each time I install a new kernel or change the initrd.
IRCsum has quit [Remote host closed the connection]
Rusty1 has joined #nixos
<nh2>
Also the standard 500 MB ESP seems way too small with nixos's approach of keeping old versions. A single bzimage+initrd is 15 MB, so after 33 updates it'll be full.
selfsymmetric-mu has left #nixos ["gone to the land of dead hiccups and extinguished light bulbs"]
IRCsum has joined #nixos
Matthieu` has quit [Ping timeout: 246 seconds]
alp has quit [Ping timeout: 248 seconds]
<clever>
nh2: the kernels dont have to be on the esp