<ldlework>
But I sure hope that some day I'm not only able to express compute environments with a data-structure like a Nixlang expression, but whole complex scaling deployments of various components and services
* craige
hints about a review being nice though
<hyperfekt>
ldlework: that username is mean, took me a couple seconds wondering why the mention wouldn't complete
<{^_^}>
[nixos-homepage] @grahamc merged pull request #269 → Added how to update Planet NixOS → https://git.io/fjUP3
<{^_^}>
[nixos-homepage] @grahamc pushed 2 commits to master: https://git.io/fjA6q
* ldlework
write down the name "hyperfekt" on a clipboard next to the number #58832
<ldlework>
hyperfekt: not quite the thing one is looking for, but nice that people are thinking about similar things
<ldlework>
one of the advantages of Terraform is that you can use the output of provisioning one piece of infrastructure as the input another
<ldlework>
And terraform will topologically sort everything and provision stuff in order so that inputs are being fed as they are available
liberiga has joined #nixos
<hyperfekt>
ldlework: What would you hope for instead of Terraform? NixOS and NixOps also maintain state. It's gonna be hard without, living in the real world where things take time and we can't reprovision the universe every time we change it.
<ldlework>
It should query the IaaS provider for that state
<ldlework>
It shouldn't reprovision each time
oborot has joined #nixos
oborot_ has joined #nixos
<ldlework>
But it should be able analyze the real state of the infrastructure and compute hashes from that
<ldlework>
The output from API providers should be consistent
<ldlework>
if I manually delete some instance, the thing should be able to repair the infrastructure
<ldlework>
at the moment Terraform just blows up
ryantrinkle has quit [Ping timeout: 245 seconds]
jluttine_ has quit [Ping timeout: 248 seconds]
<ldlework>
It's incredibly fragile, and, I'm seriously and deeply sorry if I'm wrong about this, but NixOps seems even more fragile with less overall support.
<hyperfekt>
Oh I see what you mean. Transform any infrastructure to the target. That'd be cool.
<ldlework>
I reserve the right to be wrong about that since it's been months since I gave it a run for it's money, but Terraform 0.12 has seen some serious improvements. But to the HCL language itself, the state model is fundamentally the same.
<aanderse>
infinisil: i'm guessing services.postgresql.identMap combined with services.postgresql.ensureUsers might get you what you need
<ldlework>
Anyone here remember when you had to call someone on the phone to get another VPS or dedicated host?
<gchristensen>
lol I do ldlework
<gchristensen>
I remember being on the phone with rackspace in just 2015 doing exactly that at 3am
<ldlework>
We've come so far, I don't understand cloud pessimism.
<ldlework>
"it's just someone else's computer" is possible the most understated statement supported by the english language
mkoenig has quit [Ping timeout: 246 seconds]
<hyperfekt>
damn that sounds tragic. meanwhile i have troubles accepting when provisioning takes more than a second
<ldlework>
heh I mean I still look towards the future, but holy crap
<ldlework>
just a few years ago I was using imperative programming languages to create modules to manage configuration of server states
<ldlework>
IE, puppet, ansible, and friends
<ldlework>
that was over quick :D
<ldlework>
thanks containerization!
<hyperfekt>
I almost got into saltstack because of qubes and i'm happy i didn't because i probably wouldn't have switched to nixos once deeply invested
<ldlework>
I was neck-deep in custom Saltstack modules when I was at Djangocon and saw a DotCloud employee giving a presentation about this new thing they were working on called Docker
<ldlework>
It was like 3 minutes into the talk I was having an existential crisis of how this would literally change how I was doing everything
<ldlework>
I went home and immediately started building a CaaS. That guy giving the talk ended up as my boss. :D
<gchristensen>
nice
<ldlework>
There was like 4 or 5 years when about 80% the people you'd talk to about Docker, who had never deployed it, were just totally and completely negative on the technology
<ldlework>
all the while I was watching executives from silicon valley giants, international banks, and even government employees quietly signing use-deals with the company
<ldlework>
it was never about Docker Inc. It seems people appreciate containers a bit more these days.
<gchristensen>
ldlework: hmm I knew I recognized you, from 2014 when I was doing a lot with docker.
<ldlework>
nix is such an infinitely better dockerfile
<emptyflask>
speaking of that, I'm trying to track down a sha hash mismatch problem between macos and nixos -- and it looks like the source of the problem is that sometimes the .git subdir is counted and sometimes not
polman has joined #nixos
<ldlework>
interesting
<ldlework>
i'm having a similar problem, and git repos are involved so i wonder
<craige>
Back to your earlier point about corporations, ldlework - I know that Atlassian use Nix to manage non-NixOS distros in their infrastructure.
<ldlework>
where's the money
<gchristensen>
groq.com uses Nix for their systems, and has helped improve Nix
<ldlework>
Maybe Nix Inc. could be formed, get some VC, hire some cheap devs to improve UX, develop slick enterprise versions and become a unicorn.
octor_ has joined #nixos
<averell>
lots of disruption potential, but please don't make blockchain jokes, it's a sensitive topic in the community :)
<craige>
That sounds an awful lot like Canonical|Ubuntu and I'd rather be a Debian, than that (as far as analogies go)
m0rphism has quit [Ping timeout: 272 seconds]
jluttine_ has joined #nixos
<ldlework>
craige: worked for Docker
octor_ has quit [Remote host closed the connection]
<craige>
We'll have agree to politely disagree on that. I consider Docker a bin fire :-)
<gchristensen>
be respectful, craige
<ldlework>
it's a billion dollar company with contracts with governments and banks all over to use their enterprise products in various places within their organization
<ldlework>
their product, as a result, improved a lot, despite whatever leftover flaws, and they employ good engineers who need jobs
<gchristensen>
they've done incredible amounts of good
<ldlework>
dunno
knupfer1 has joined #nixos
<ldlework>
lxc was around for a good while and no one really cared about containers
<ldlework>
especially not governments and multinationals
<averell>
except the freebsd hermits
<ldlework>
there are always keen hermits who can identify the goods :)
knupfer has quit [Ping timeout: 264 seconds]
knupfer1 is now known as knupfer
<jackdk>
I have a server which contains my personal website, deployed via `nix copy`. The nix expression that does this includes the output of a hakyll build, which means that each deploy copies across every asset every time, as they're in a new derivation. Has anyone else dealt with this, and how did they get it under control?
slack1256 has quit [Ping timeout: 248 seconds]
<ldlework>
but docker needed VC so they could hire proven successful execs from companies like IBM and VMWare to formulate enterprise strategies and make those connections
octor_ has joined #nixos
<ldlework>
it's hard to understate how valuable aquihiring a tiny tiny garage startup of 4 or 5 geniuses can be too
<ldlework>
which docker also did repeatedly
<ldlework>
i don't mean to fanboi it up, but "dumpster fire" just distorts our ability to reason about reality and make good choices
<gchristensen>
+1
<ldlework>
at some point I remember Docker basically hired everyone who was actively doing unikernel work
<gchristensen>
Docker is so cool because we can use all the in-roads they developed to sneak Nix in
<ldlework>
and Docker Inc still has never announced anything signifcant in that domain
<averell>
what came of that?
<averell>
ah
<ldlework>
who knows what they're quietly up to
<ldlework>
it's been like 5 years
<ldlework>
maybe nothing
<averell>
containers were a pretty nice revolution
<averell>
although i wouldn't invest in docker now :)
<ldlework>
same, but I already cashed out :D
<ldlework>
I doubt they go under or anything but I very much bet that in 10 years you can't look at VMWare and Docker Inc and see a difference
<ldlework>
Once they hired those execs the culture was destroyed completely in my opinion
<ldlework>
NixOS is missing a Hykes to drive the revolution in thought among people who are not even thinking about the problem.
octor_ has quit [Remote host closed the connection]
<ldlework>
Regardless of that guy's flaws he got out there and relentlessly advocated for people to understand despite the endless waves of Hacker News haters and heckler. Despite his real blunders, misbehaviors and embarassments. It worked.
<craige>
I like pie_'s suggestion of a talk ldlework. I was in the audience when Andrew Tridgell gave a talk at linux.conf.au on the technical and security flaws inherent in bitkeeper (sadly the video of the talk was lost). Linus Torvalds was also in that audience and started working that day on git. Gather your thoughts on an issue put it out to the world and who knows who will be inspired :-)
<{^_^}>
[cabal2nix] @unode opened pull request #426 → Enable inlineMarkdown on pandoc-placetable → https://git.io/fjA6b
drakonis has joined #nixos
boxscape has quit [Remote host closed the connection]
<andrewrk>
I did `nix-shell -p gcc9` but how do I call it? I don't see it in PATH. `gcc --version` says 7.4.0
camsbury has joined #nixos
domogled has joined #nixos
<camsbury>
hi all - trying to debug my pulseaudio setup not working on the new laptop - I have hardware.pulseaudio.enable = true in my configuration.nix, but showing as false in the '<nixpkgs/nixos>' repl
<camsbury>
and can't find pacmd, etc.
<camsbury>
I have some hardware imports from nixos-hardware, but is there any way to really tell where the eventual value is coming from or how any kind of precedence is set?
<notgne2>
is maybe the `//` in the config not merging correctly, discarding the entire `hardware` definition with the stuff under `machine.xps`? I think I had a scenario a bit like that
<camsbury>
1500 is the second
<clever>
notgne2: oh, thats it exactly
orivej has joined #nixos
<camsbury>
oh jesus
<clever>
camsbury: line 121, that overwrites the entire hardware key
<camsbury>
its just overwriting the values
<camsbury>
lol
<clever>
so all of hardware = is gone
<camsbury>
lolll
<clever>
camsbury: you want to use mkMerge or imports
reallymemorable has joined #nixos
<clever>
camsbury: or restructure it to be a tree with several roots
<camsbury>
notgne2++
<{^_^}>
notgne2's karma got increased to 1
<camsbury>
laughing
<clever>
camsbury: instead of starting at configuration.nix, and conditionally loading different things
<clever>
camsbury: you start at laptop.nix, or desktop.nix, and then both do imports on a core.nix
<clever>
and configuration.nix symlinks to the right starting point
<notgne2>
personally I have a git repo with a bunch of modules, configuration.nix is the one thing that I don't sync
<clever>
and nothing has to be conditional, because laptop.nix is only used by the laptop
<notgne2>
then I can use the same files in nixops too
<camsbury>
blegh I just have literally 3 laptops on nixos
<clever>
camsbury: use hostnames then
<camsbury>
the idea being they share laptop stuff
oborot_ has quit [Ping timeout: 268 seconds]
oborot has quit [Ping timeout: 268 seconds]
<clever>
camsbury: laptop1.nix, laptop2.nix, laptop3.nix, all 3 have imports = [ ./laptop.nix ];
<clever>
then laptop.nix has imports = [ ./core.nix ];
<clever>
then you put the option in the one that makes the most sense
<camsbury>
okay so basically I should probably never do conditional merges
<camsbury>
and just always use imports
<camsbury>
in my packages and overlays files I'm doing conditional ++ for lists
<camsbury>
which seems fine...
<clever>
but you could also set systemPackages from different modules, and nixos will merge them for you
<camsbury>
ohh okay
<camsbury>
beautiful
<camsbury>
so basically
<camsbury>
gaming.nix would have whatever gaming packages
<camsbury>
and I could just import that one conditionally
<emptyflask>
manveru: I've been looking into why my gemfile entries with git URLs are failing on one computer and not the other, and it looks like for some reason on my nixos machine, the gem is cloned into /nix/store/ but still has the .git subdirectory, which changes the sha hash
laika___ has joined #nixos
<emptyflask>
using nix-prefetch-git manually seems to get it properly without the .git subdir
nexgen2 has quit [Quit: Leaving]
<thefloweringash>
I’ve struggled so much with this. IIUC, when bundler loads a gem with a git source it must be a git repository. So the fetch includes the .git dir via “leaveDotGit”, and .git dirs aren’t deterministic.
<ldlework>
maybe this is the key to my own reproducability problem
<ldlework>
i can't wait to find out
Soo_Slow has quit [Remote host closed the connection]
Soo_Slow has quit [Remote host closed the connection]
Guest55 has joined #nixos
<ldlework>
oh man, diffoscope is gonna take literal hours to run
Supersonic has joined #nixos
orivej has quit [Ping timeout: 245 seconds]
<Guest55>
Hello, I want to make a suggestion- a lot of nixos users are power users- (I know I am, for one). I like to boot directly from ISO. Downloading unetbootin seems a bit of a hassle. May I suggest that you/(we?) guys also offer a unetbootin image. This will solve the problem of burning to USB.
<Guest55>
The process would go like this- the release maintainers use unetbootin to create a bootable image. Typically a flash drive of about a couple MBs over iso size would be required.
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjAP0
<clever>
Guest55: if you run nix-build on this file, it will create a directory with 3 things, a kernel, a initrd, and a fragment of a grub config
liberiga has quit [Ping timeout: 260 seconds]
<clever>
you can then shove that onto any bootable media, and configure whatever bootloader you like to obey that fragment of cfg
<Guest55>
clever: Should I feel disappointed that my effort is wasted or elated that this is so easy?
<clever>
its easy to add to another bootable media
<clever>
and ive answered this kind of question before, about a year ago
<Guest55>
That said, I guess there's still a small usecase for my effort- for example when I started using Nix back around June, I found a lot of reddit and blog posts but none could answer my question. (I didn't ask one myself). Even if one could- I wouldn't have access to nix-build.
<Guest55>
Obviously your way is *the* sane way and much cleaner too
<clever>
Guest55: you can use nix-build without nixos, just install nix on any linux distro
<clever>
but hydra could also be configured to pre-build these files, and offer them for download
<Guest55>
I know, the issue is that's too much work.
<ldlework>
it currently says 20 days
<Guest55>
I guess that would be the right way
<ldlework>
lol
ng0_ has quit [Remote host closed the connection]
* ldlework
wonders if he's using this wrong
<Guest55>
giving these files pre-built
<clever>
ldlework: try just `diff -r` at first
<Guest55>
again, since your solution is the 'clean' way of doing i, I bet there's a lot of ways in which it can be reused, hydra being one you mentioned
ng0_ has joined #nixos
<Guest55>
Ultimately, I'd suggest having at least 1 way in I don't have to 1) Use UnetBootin. 2) Be reduced to a single boot.
<Guest55>
I have a 320 GB HDD and I like carrying around all linux distros (and Windows 10!)
<ldlework>
clever: that was way faster lol
<clever>
Guest55: yeah, thats the kind of thing this nix file was made for
<clever>
Guest55: you can use nix to make a kernel+initrd pair for nixos, then copy them to a drive and add a menu entry
<Guest55>
clever: Yeah I got you, it's just too damn simple
<Guest55>
Thanks for creating this!
<Guest55>
(simple in usage I mean lol)
<clever>
Guest55: the main difference between my initrd and the official iso, is memory usage
<clever>
Guest55: the official iso will mount part of the cdrom when running, and needs the usb/cd inserted to work
<clever>
while my initrd, has the whole rootfs inside the initrd
<clever>
so once it boots, you can unplug the drive!
<clever>
(but it now needs a few extra gig of ram to boot)
<{^_^}>
[nixpkgs] @jonringer opened pull request #67542 → onnxruntime: init at 0.5.0 → https://git.io/fjAPr
<jackdk>
clever: isn't there a `copytoram` option on the official media?
<Guest55>
I don't know much about the initrd stuff and kernel loading process but I feel that having things as a nix expression means no manual fiddling.
<clever>
jackdk: yeah, that does the same thing, but its not optional on my image
<Guest55>
For posterity and anybody intersted in the path I was going down to- I will complete my instructions... For posterity and anybody intersted in the path I was going down to- I will complete my instructions...
<Guest55>
2) Use dd to copy the image to the newly created partition on disk.
<Guest55>
3) Add a grub entry which says: set root=(hd<m>, msdos<n>) (where m and n would correspond to the disk number and partition) (I Don't know how to keep m fixed)
<Guest55>
grub entry goes like this:
<Guest55>
set root = (hd1, msdos3)
<Guest55>
chainloader +1
<Guest55>
wrapped inside menuentry "Nix"{ ... }
<Guest55>
clever: I just found that I had your link bookmarked (via a discourse discussion on your script).
<ldlework>
clever: so it turns out all my differences come from a single dependency I have that comes from git
<clever>
ldlework: dependency paths shouldnt be in the fixed-output drv
<ldlework>
i don't know what that means
<ldlework>
i'm using paket a thing to pull dotnet dependencies, it supports getting them from git
<ldlework>
i have one dependency that comes from git, which it clones and builds
<clever>
it should not be building dependencies in the fixed-output drv
<clever>
it should fetch, and thats it
<ldlework>
i didn't write the software :(
endformationage has quit [Quit: WeeChat 2.5]
Myhlamaeus2 has quit [Remote host closed the connection]
drakonis_ has joined #nixos
drakonis has quit [Quit: WeeChat 2.5]
Guest55 has quit [Remote host closed the connection]
wfranzini has quit [Remote host closed the connection]
drakonis1 has quit [Ping timeout: 245 seconds]
orivej has joined #nixos
asymptotically has joined #nixos
<{^_^}>
[nixpkgs] @petabyteboy opened pull request #67547 → update some coreboot-related packages → https://git.io/fjA1I
drakonis1 has joined #nixos
DariusTheMede has quit [Ping timeout: 246 seconds]
wfranzini has joined #nixos
<{^_^}>
[nixpkgs] @uHOOCCOOHu opened pull request #67548 → valgrind: fix shebang in outputs → https://git.io/fjA1O
Ariakenom has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<Ariakenom>
I'm running nixos with sddm & plasma. I suppose this is a sddm question but in the login screen after boot is there no way to login without mouse?
rauno has joined #nixos
snajpa is now known as SvataGreta
ilikeheaps has joined #nixos
<craige>
I log in all the time without a mouse, Ariakenom - do you mean select users or something else specifically?
<makefu>
Ariakenom: maybe tabbing through all the fields?
Zer0xp has joined #nixos
<Ariakenom>
select user. there is only me but I can't get to typing my pswd
<Ariakenom>
well im not sure if it's select user, since it's only me, but it's plausible
<craige>
Ah, on boot, my single user system already has the password box in focus.
<Ariakenom>
makefu: I tried tab but saw no effect
<ldlework>
how do I make a derivation with an empty src?
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #67549 → Fix sharing in gnome-control-center → https://git.io/fjA1u
<jluttine_>
After upgrading my NixOS, I get the following error from nix-shell that has pinned nixpkgs to older version: qt.qpa.plugin: Could not find the Qt platform plugin "xcb" in ""
<jluttine_>
Any ideas how to fix? Is it possible to run ipython+matplotlib in a nix-shell that has old nixpkgs pinned in a NixOS system that has newer nixpkgs in use..?
Zer0xp has quit [Quit: Bye!]
rauno has quit [Ping timeout: 245 seconds]
Zer0xp has joined #nixos
domogled has joined #nixos
johanot has joined #nixos
chr15 has joined #nixos
<manveru>
jluttine_: qt is... special
<manveru>
you can probably search for solutions with QT_PLUGIN_PATH env var
civodul has joined #nixos
freeman42x has joined #nixos
hyper_ch2 has joined #nixos
Ariakenom has quit [Remote host closed the connection]
Ariakenom has joined #nixos
__monty__ has joined #nixos
fusion809 has quit [Remote host closed the connection]
lurpahi has quit [Ping timeout: 272 seconds]
rauno has joined #nixos
Makaveli7 has joined #nixos
<lightbulbjim>
jluttine_: the way QT apps are packaged has changed recently. You'll need to update them to newer versions so that they can find the QT libs. See https://github.com/NixOS/nixpkgs/issues/65399 for details.
<lightbulbjim>
You could possibly pin QT and the associated wrappers to an older version, but then you would have the same problem with any apps that _have_ been updated.
<lightbulbjim>
It's the QT Rubicon of 2019 ;)
<Ariakenom>
I added a fileSystems entry for the windows partition. but my user doesn't get permission to it
o1lo01ol1o has joined #nixos
captn3m0 has quit [Ping timeout: 272 seconds]
<clever>
Ariakenom: you also have to add a user= to the mount options, standard ntfs stuff
<clever>
2019-08-27 05:40:58 < clever> ldlework: but if a fixed-output drv fails due to a hash mismatch, the $out is renamed to match the hash it actually has
<ldlework>
but it doesn't exist anymore, just a .lock file
mzcore has left #nixos [#nixos]
<clever>
2019-08-27 05:40:58 < clever> ldlework: but if a fixed-output drv fails due to a hash mismatch, the $out is renamed to match the hash it actually has
<clever>
ldlework: so, it got renamed, and you have a different hash in /nix/store/
<clever>
ldlework: simplest thing is to check mod-times, ls -ltrh /nix/store/
<ldlework>
jesus lol ok
<ldlework>
clever: but everything is modified at Dec 1969
<clever>
ldlework: the failed output is the only thing that wont be 1969
<clever>
ldlework: can you pastebin `ls -ltrh /nix/store/ | tail` ?
<Ariakenom>
clever: ntfs with some whitespace
<ldlework>
Option --time (-t) has no "rh" setting (choices: modified, accessed, created)
<ldlework>
heh
<ldlework>
I'm guessing we want modified
<clever>
Ariakenom: then its the ntfs kernel driver, rather then the fuse one, what about `lsmod | grep ntfs` ?
<chr15>
Hey so I want to use this: https://paste.ubuntu.com/p/zZx7qCDYDT/ to implement Weechat-Matrix. I've added this to configuration.nix but I don't know how to run the application.
<clever>
sqlite> .mode column
<clever>
sqlite> select path, narsize/1024/1024 as narmb, datetime(registrationTime,'unixepoch') from ValidPaths where narsize > 256 * 1024 * 1024 order by registrationTime;
<pbb>
clever: wouldn't that build a gcc built to be run on an arm processor?
<clever>
ldlework: this will list everything in /nix/store over 256mb, sorted by when it was created
<clever>
pbb: pkgsCross is the cross-compile framework, to cross-compile from current to something
<ldlework>
clever: this is a little mad, no?
<pbb>
so does it expose a cross-compiler for building arm binaries on x86?
<clever>
ldlework: its a caching mechanism, its expecting that you will paste the new hash into the nix expr, and then it will reuse the product it previously made
<manveru>
alexarice[m]: ah, i was thinking of attrByPath, but if it's just a single level that's kinda pointless
<DariusTheMede>
Does someone manually update hackage-packages.nix every so often?
<ldlework>
clever: yeah but i mean like, tracking down this information
<ldlework>
are you sure we're looking for "ValidPaths"
<clever>
ldlework: you set the limit to 1gig
<clever>
ldlework: you want to go the other direction
<clever>
ldlework: or just remove the limit entirely, or invert the direction
<ldlework>
making there be no limit
<ldlework>
i still see nothing about BindingAttributes
<ldlework>
Only stuff that has ever successfully built
<ldlework>
I don't see any of the names I've been testing with for hours
<clever>
ldlework: add a where path like "%BindingAttributes%"
<clever>
sqlite> select path, narsize/1024/1024 as narmb, datetime(registrationTime,'unixepoch') from ValidPaths where path like "%BindingAttributes%" order by registrationTime;
aveltras has joined #nixos
ThatDocsLady has joined #nixos
<ldlework>
I garbage collected, then built it one more time, then ran your query
<manveru>
ldlework: well, i usually just run with `-k`, then rerun with strace attached to the daemon to see where it writes stuff, but clever tries to teach you the theoretic way :)
<clever>
DariusTheMede: what is the key difference that fixes it?
<ldlework>
clever: lmao
Ariakenom has quit [Remote host closed the connection]
Ariakenom has quit [Remote host closed the connection]
<clever>
ldlework: rm that sucker?
Makaveli7 has joined #nixos
<ldlework>
clever: let's try it!!!
noudle has joined #nixos
m0rphism has joined #nixos
o1lo01ol1o has joined #nixos
<DariusTheMede>
clever: can we fix up the output from cabal2nix somehow or is it best just to use the working derivation?
chr15 has quit [Ping timeout: 272 seconds]
<clever>
DariusTheMede: i think the problem is that the cabal file is asking for the "wrong" library
<clever>
DariusTheMede: and cabal2nix is just obeying that
<DariusTheMede>
clever: I can see that
<clever>
DariusTheMede: you can either fix the cabal file, or use an override like vulkan to force it
<DariusTheMede>
clever: I don't think I can fix the cabal file - AFAIU cabal puts -lsundials_cvode and -lsundials_arkode into the build command it generates - putting -lsundials won't work
<clever>
DariusTheMede: ah, then youll need to fix it in configuration-common.nix, like vulkan = super.vulkan.override { vulkan = pkgs.vulkan-loader; };
<{^_^}>
[nixpkgs] @Moredread opened pull request #67553 → rtmidi: add PulseAudio support → https://git.io/fjAy3
<clever>
pbb: you want this, to change the stdenv entirely
<pbb>
thanks now I'm seeing the cross-compiler
<clever>
pbb: and thats armv6, not 7 or aarch64
<pbb>
yeah I saw that
<davidtwco>
I'm having a strange issue where packages from the nix store on a Ubuntu system are all missing libstdc++.so.6 - stuff like rm. Any thoughts?
<{^_^}>
NuGet/Home#6229 (by tmat, 1 year ago, closed): NuGet Pack is not deterministic
<ldlework>
though the changes were only merged 19 days ago
<alexarice[m]>
does anyone know how things like nixpkgs and home-manager generate the option documentation?
<o1lo01ol1o>
is buildCmd a valid attribute to mkDerivation? I can't remember and can't seem to find documentation. I basically need to run a bash script in a subdirectory as part of a validation step.
<clever>
ldlework: but you shouldnt be doing any builds in the fetch derivation
<clever>
ldlework: it should purely download source, and nothing more
<clever>
o1lo01ol1o: just run that bash and do nothing else? or after the build happens?
<betaboon>
clever ldlework have you seen how i deal with nuget in the eventstore package ?
<o1lo01ol1o>
clever: just run it and do nothing else. It's a hydra job that checks a few things; it sets some local flags and builds and then fails of they don't work. It's super dumb for the momement.
<o1lo01ol1o>
Sorry, the derivation is for a hydra job, the bash script just compiles
<clever>
o1lo01ol1o: then you want pkgs.runCommand "name" {} "bash code"
<ldlework>
betaboon: do you have any interested in helping me restore Paket2Nix?
<DariusTheMede>
if I look in configuration-hackage2nix.yaml then hmatrix-sundials doesn't have a version number so why is nix picking up an old version?
<qubasa>
I want to use nix-instantiate to learn the nix language in detail but executing nix-instantiate --eval --expr 'let s = builtins.forEach [ "a" "b" "c" "d" ] (a: b: a+b ); in { inherit s; } ' will give back { s = <CODE>; } how do I get my result?
<ldlework>
and generate probably the code you're using
<betaboon>
ldlework: i realy have no clue about the .net ecosystem. the solution pointed to above is just what i came up for packaging eventstore.
<ldlework>
betaboon: well i am learning too
<o1lo01ol1o>
clever: I should clarify, I import an existing expression: `foo = import ./default.nix {};` The script in question is in foo/scripts/foo.sh and it needs access to the environment specified in foo, so it needs to happen after foo is built.
<ldlework>
betaboon: but i guess if you just needed some .net app you wanted in nixpkgs and are not actually doing .net dev you are not interested in a Paket2Nix
<{^_^}>
[nixpkgs] @toonn opened pull request #67557 → wire-desktop: 3.9.2895 -> 3.10.2904, mac 3.9.2943 -> 3.10.3133 → https://git.io/fjASM
hyper_ch2 has joined #nixos
<talyz>
alexarice[m]: for nixpkgs it's done in nixos/doc/manual/default.nix - it's seems a bit involved, but from what I can tell it exports the options to xml, runs xsl transformations on the result to extract the descriptions and then compiles them and the rest of the manual..
<ldlework>
maybe toString a derivation gives you the path
<betaboon>
ldlework: i am disabling the online-source and then manually rebuilding a nuget source called "nixos" which is later used by "dotnet restore"
<alexarice[m]>
talyz: ah was hoping there was a nice neat function that made some fancy docs from a set of modules
<betaboon>
ldlework: it might create that source in a relative subdirectory, not sure
abbiya has quit [Quit: abbiya]
<ldlework>
betaboon: oh okay I didn't understand the semantics of the nuget command. It's adding the fetched nuget package to a -Source nixos, but you don't have to explicitly create that source anywhere
<ldlework>
I guess you just add packages to it
<yorick>
oh god, -j48
Profpatsch has joined #nixos
<yorick>
on a ghc build
<betaboon>
yeah seems like it. i have never worked with nuget before packaging eventstore. that's the best solution i came up with
<Profpatsch>
abbradar[m]: Now I wonder: did we meet at cccamp?
o1lo01ol1o has quit [Remote host closed the connection]
<abbradar[m]>
ProfpatschI think no, somehow. last time I saw you was at the first NixCon ;_;
<betaboon>
i am trying to write tests for a new bootloader, could anyone point me in the right direction? should that go into nixos/tests/installer.nix or nixos/tests/boot.nix ? oO
reallymemorable has joined #nixos
reallymemorable has quit [Client Quit]
drgbr has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
Phillemann has joined #nixos
<talyz>
alexarice[m]: ah, sadly it doesn't seem like it, but it's probably not too hard to make a more generalized version from it
<Phillemann>
I have an "installer.sh" file which contains a bash script at the top and some binary stuff at the end. How do I write a derivation for that?
<Phillemann>
Is there something I can look at for reference?
<alexarice[m]>
talyz: I think home-manager has made a seperate package for it so I might give that a try
<talyz>
alexarice[m]: also, I didn't check home-manager ;)
<talyz>
alexarice[m]: right :)
<{^_^}>
[nixpkgs] @lightbulbjim opened pull request #67559 → qt5ct: fix missing app icon and name when running under Wayland → https://git.io/fjA9m
aveltras has quit [Quit: Connection closed for inactivity]
o1lo01ol1o has joined #nixos
noonien has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @sikmir opened pull request #67560 → mkgmap: init at 4289 → https://git.io/fjA9G
<qubasa>
If I execute: builtins.map (a: if a.signature == false && !allowNonSigned then throw "Disable signature checking in firefox if you want this addon" else a ); I get "cannot coerce the partially applied built-in function 'map' to a string" what must I do so that map returns an array of sets?
<hyper_ch2>
I got an email with "You’re invited to join the @NixOS organization on GitHub." and I wonder why
hyper_ch2 has quit [Remote host closed the connection]
<kiwi_82>
Hi folks, did you turn off sysrq on master or is that a regression in the kernel?
hyper_ch2 has joined #nixos
<gchristensen>
kiwi_82: how would we turn off sysrq? I can look for hints
<betaboon>
gchristensen: would you be able to point me in the right direction regarding my question about writing tests for a new bootloader ?
<clever>
qubasa: map takes a function and a list, you only gave it a function
hyper_ch2 has quit [Remote host closed the connection]
<gchristensen>
betaboon: hmmm check out ./nixos/tests/boot.nix and ./nixos/tests/boot-stage1.nix?
domogled has quit [Ping timeout: 245 seconds]
<Ariakenom>
I'm trying to make a windows ntfs partition writeable.
<kiwi_82>
gchristensen sysrq worked on 20-day old master but today I was surprised when OOM killed my NixOS workstation I got no reacton from alt-sysrq-f (oom killer), now I see in the console that sysrq is disabled. So there is a regression in kernel Linux or NixOS.
<betaboon>
gchristensen: i only found references to systemd-boot and grub in ./nixos/tests/installer.nix which led me to believe that is the place to look. is that a wrong assumption ?
bakakuna has quit [Ping timeout: 258 seconds]
bakakuna1 is now known as bakakuna
<Ariakenom>
it mounts, woo. but as read only, boo
<gchristensen>
betaboon: I'd start with the boot.nix ones and see if you can make progress there
<clever>
Ariakenom: the driver your using only supports read-only mode
<clever>
Ariakenom: oh wait, next line down, it does have limited write support
<qubasa>
clever: Right... right I am stupid
<Ariakenom>
clever: heh yeah. i did the same roller coaster when you linked that before
<Ariakenom>
afaict the write support should be enough
<clever>
Ariakenom: the other option is the ntfs3g stuff that uses fuse, but i dont remember how to mount it from configuration.nix
<clever>
Ariakenom: i also tend to not trust any write support for linux+ntfs, id rather use a fat32 middleman then risk the partition becoming corrupt
<Ariakenom>
clever: ok. yeah. that might be most sensible
<betaboon>
gchristensen: looking at boot.nix i am completly lost
<Ariakenom>
the proc/mount seems to include options that isn't in that documentation. ro and relatime.
fendor_ has joined #nixos
<clever>
Ariakenom: those are global ones, check `man mount`
tmaekawa has quit [Quit: tmaekawa]
bakakuna has quit [Ping timeout: 245 seconds]
dansho has quit [Ping timeout: 258 seconds]
selfsymmetric-pa has joined #nixos
<Ariakenom>
clever: right. they're not in fstab either, where I have rw. so I don't know who adds it
fendor_ is now known as fendor
<clever>
Ariakenom: the kernel itself
<clever>
Ariakenom: try umount and then mount -v ?
<yorick>
where is /var/lib made in the activation scripts?
orivej has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @johanot opened pull request #67563 → kubernetes: 1.14.3 -> 1.15.3 (with revert of module systemd dependencies) → https://git.io/fjA97
ThatDocsLady has quit [Read error: Connection reset by peer]
ericsagnes has quit [Ping timeout: 276 seconds]
ThatDocsLady has joined #nixos
ThatDocsLady has quit [Remote host closed the connection]
selfsymmetric-pa has quit [Remote host closed the connection]
<Ariakenom>
clever: my umount isnt working. how scary is force?
lurpahi has joined #nixos
<clever>
Ariakenom: is something currently cd'd into it?
o1lo01ol1o has quit [Remote host closed the connection]
<betaboon>
abbradar[m]: i just saw that you were involved in writing tests for bootloaders. can you point me in the right direction on adding tests for a new bootloader ?
<abbradar[m]>
betaboon: sure! some specifics?
<abbradar[m]>
what's the bootloader, for what platform etc
DariusTheMede has quit [Ping timeout: 258 seconds]
<Ariakenom>
clever: I rebooted
<clever>
Ariakenom: thats the windows way of fixing most problems :D
<Ariakenom>
Indeed!
mudri_ has quit []
<Ariakenom>
sudo umount /mnt/windows && sudo mount -va; doesn't show much
<clever>
Ariakenom: did it say anything about /mnt/windows?
drakonis1 has joined #nixos
<Ariakenom>
/mnt/windows : successfully mounted
mudri has joined #nixos
<clever>
Ariakenom: what if you do `mount -v /mnt/windows` ?
<Ariakenom>
mount: /dev/sda3 mounted on /mnt/windows.
<Ariakenom>
:D
<Ariakenom>
how verbose
<clever>
Ariakenom: can root write to it?
<Ariakenom>
no
<clever>
not sure what to do then
<Ariakenom>
you're probably right that it's safest that way anyway
drakonis1 has quit [Read error: Connection reset by peer]
<rnhmjoj>
infinisil: do you know if members of the nixos org get some permission? i'd be interested in, for example, assigning myself to issues and adding labels, but i don't want commit access.
drakonis1 has joined #nixos
<abbradar[m]>
Ariakenom: do you use ntfs-3g?
<abbradar[m]>
you can check in `mount`
<infinisil>
rnhmjoj: I think it's possible to only get permissions for some things but not commit, maybe you can ask for that
<clever>
Ariakenom: he is currently using the ntfs driver in the kernel
<clever>
abbradar[m]: ^^
<{^_^}>
[nixpkgs] @petabyteboy opened pull request #67566 → nixos/nginx: use mailcap mime.types → https://git.io/fjAH5
<abbradar[m]>
it doesn't support writing at all, I think
<alexarice[m]>
rnhmjoj: infinisil I think not at the moment but gchristensen is working on it?
<abbradar[m]>
there was some "experimental" support but it's dangerous IIRC
<abbradar[m]>
clever: ah, I see! still, I didn't think I managed to write something with it even once
<abbradar[m]>
maybe it supports only modifying existing files with the same size or something?
asymptotically has quit [Remote host closed the connection]
<rnhmjoj>
alexarice infinisil: thank you, i'll ask around
asymptotically has joined #nixos
<alexarice[m]>
rnhmjoj: there's a discourse thread about it
zupo has joined #nixos
<alp>
is there a way (env var or something) to know whether we're building a derivation in a sandboxed env?
fleaz has left #nixos [#nixos]
<clever>
alp: dont think so, but there are files that will be missing, like /home and many others
<fgaz>
Hi all. What is the process/requirements to move myself from ofborg's known user to trusted? I'd like to speed up darwin builds on my prs
<fgaz>
grahamc: ↑
<alp>
clever, ok, I guess this should be enough to hack something together for the scenario I'm concerned about. thanks :)
<clever>
alp: what are you concerned about?
chiefgoat has quit [Ping timeout: 258 seconds]
<alp>
clever, well, it's a colleague, to be precise. wants to skip running tests as part of the derivation building, if sandboxing is enabled. and wants to run them otherwise.
<clever>
alp: that would make the derivation impure!
<alp>
might be a bit tricky to have smth that does the job for osx and linux, but doable I suppose.
chiefgoat has joined #nixos
<alp>
clever, the problem is that some tests require network access apparently.
<clever>
alp: network access to what?
<alp>
good question.
<rnhmjoj>
alexarice: i can't seem to find the thread. do you a link by any chance?
<meck>
Does anyone have any idea why `nixos-rebuild switch` would take a halfhour waiting at `setting up tmpfiles` while `nixos-rebuild boot` is instantaneous, --verbose dosen't say anything?
<gchristensen>
mine takes a long long long time if I have rules cleaning out /tmp/ for example
<gchristensen>
I really regret switching to tmpfiles.d for my own things
<vika_nezrimaya>
Question about flakes - are flakes right now stable enough for a user to try them out? I've downloaded nixFlakes and right now trying to comprehend how to make my app (which requires several Python packages not present in Nixpkgs) to fit in all of this
<ZoomZoomZoom>
I can't get it to run with multiple datadirs, as the nix service file just uses a string and then does "systemd.tmpfiles.rules = [ "d '${cfg.dataDir}' - minio minio - -" ];" which doesn't work for multiple paths.
<{^_^}>
[nixpkgs] @FRidh pushed commit from @borisbabic to master « python3Packages.pre-commit: 1.18.1 -> 1.18.2 »: https://git.io/fjAdd
<betaboon>
judson__: ah yeah you're right.
alex`` has joined #nixos
jollyjester has joined #nixos
gxt has quit [Ping timeout: 260 seconds]
<jollyjester>
hello, can someone open an issue on the nixpkgs page? the brave browser version is outdated and not the latest even on the unstable channel, i currently can't use my github accoutn so i cannot open an issue, anyone willing to help would be appreciated
<infinisil>
ZoomZoomZoom: Ah, that's really just a coincidence though, if the author of that module made sure to escape the shell argument properly even that wouldn't work
<vika_nezrimaya>
jollyjester: I'll take a look at it, and maybe even send a PR to update it :3
<infinisil>
ZoomZoomZoom: You need to change the module itself if you want multiple dataDir's, a PR would be nice too
<{^_^}>
[nixpkgs] @FRidh merged pull request #67452 → vde2: build with `--disable-python` by default → https://git.io/fjAmD
<{^_^}>
[nixpkgs] @FRidh pushed commit from @Ma27 to master « vde2: build with `--disable-python` by default »: https://git.io/fjAdj
gxt has joined #nixos
<ZoomZoomZoom>
infinisil, ok, that's what I thought. Can't this systemd option be overridden somehow in a local config?
<evanjs>
vika_nezrimaya: Man I love how active everbody on nixpkgs is. I'm starting to look at repology like a scoreboard :D (current best - 22 hours - https://repology.org/project/gitkraken/history)
<emily>
is there no way to combine wrapProgram with --argv0? :(
<emily>
the program I'm wrapping looks at argv[0] and thus reports "usage: .foo-wrapped ..."
<emily>
(it's Python which I guess is relevant since Python programs are already wrapped?)
<emily>
(also if this is more appropriate for -dev feel free to let me know, feels more like a nix(pkgs) usage question)
drakonis has quit [Quit: WeeChat 2.5]
<kraem>
are there any ways to count the number of dependencies "top level" packages have under the system-path package? right now i'm just visually examining `nix-store -q --tree /nix/var/nix/profiles/system`
drakonis has joined #nixos
<samueldr>
emily: the question is appropriate for here :)
<emily>
yeah, I think the reason it doesn't work is that the Python stuff *already* does a layer of wrapping
<emily>
and just re-wrapping on top of that is breaking the argv[0]?
<emily>
not sure though
<samueldr>
so yeah, maybe because of... just what you said
<emily>
in that case my question is how to add stuff to the $PATH of a python executable
<FRidh>
emily: wrapPythonPrograms creates both a wrapper and injects code into executables. The snippet of code that is injected sets sys.argv[0]
<FRidh>
e.g. it inserts import sys;import site;import functools;sys.argv[0] = '/nix/store/2rnmilxqjcp5x86szykzvq8q8zg0cnry-python3.7-pytest-4.2.1/bin/pytest';
<FRidh>
that should function regardless of whether it is wrapper again
jollyjester has quit [Quit: leaving]
chloekek has joined #nixos
<emily>
well... it doesn't seem to
<emily>
unfortunately I'm adding a good few derivations so it's hard to give a self-contained example
<{^_^}>
[nixpkgs] @jtojnar opened pull request #67576 → gimp: switch to unstable → https://git.io/fjAF9
<emily>
but when I don't wrapProgram, it identifies as "glasgow", and when I do, it identifies as ".glasgow-wrapped"
<gchristensen>
it is a shame that Matrix hasn't fixed this problem yet
<gchristensen>
" * with nixos-unstable and following configuration for `xserver` - getting no graphical display after logging in. only a black screen with small terminal on top left corner" is what kaychaks[m] said
<kaychaks[m]>
same thing is happening for lightdm. only sddm could load the graphical display of gnome
zupo has joined #nixos
<infinisil>
gchristensen: Which matrix problem? The long message thing?
thc202 has quit [Ping timeout: 250 seconds]
<gchristensen>
yeah
infinisil-m has joined #nixos
evanjs has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
thc202 has joined #nixos
evanjs has joined #nixos
<infinisil>
gchristensen: And that's a problem because?
<infinisil>
I mean IRC just doesn't support this long messages
<ivegotasthma>
infinisil: it could split the message
jgeerds has joined #nixos
<gchristensen>
(and indeed I have a bug report about this very thing)
<infinisil>
Ah I see yeah
<gchristensen>
and I think it is worse that the matrix user has no idea nobody actually sees their message at all, or that there is a reason to send a shorter message before including all the code
fusion809 has joined #nixos
<kaychaks[m]>
<gchristensen "and I think it is worse that the"> oh no
<kaychaks[m]>
, actually I had no idea about it.
<{^_^}>
Invalid command syntax
<kaychaks[m]>
thanks for the information.
<gchristensen>
yeah! so what Matrix sends to IRC in such a case is just this:
<kaychaks[m]>
<gchristensen "17:21 * kaychaks sent a long me"> bummer
<gchristensen>
I agree
<mdash>
kaychaks[m]: FWIW there's also #nix:matrix.org which does not have the same depth of expertise as this channel but is accustomed to long pastes of code etc
<gchristensen>
irccloud does a nice job of this, having the first bit of message be sent and the paste separate thing
<kaychaks[m]>
so if I have to ask / show code snippet then I should paste a link here ?
asymptotically has quit [Quit: Leaving]
<gchristensen>
one way to do it would be to send your first line of text, and then a second message with your paste -- then it would show the context and then the "long message" thing -- which is fine
<gchristensen>
it would provide a teaser for your question and incentive to view your post
<kaychaks[m]>
got it thanks
<mdash>
yeah the main thing is to remember that if you send one or two lines in a message, it gets sent straight across to irc, more than that and it gets linkified
<kaychaks[m]>
sure
<kaychaks[m]>
btw any idea whats going on with gnome3 error with `gdm` and `lightdm` displaymanagers ?
<attente[m]>
has anyone gotten mopidy to work in their configuration.nix? i'm running into https://github.com/NixOS/nixpkgs/issues/63861 (file collision in two python package dependencies)
Soo_Slow has quit [Remote host closed the connection]
xelxebar has quit [Remote host closed the connection]
selfsymmetric-pa has joined #nixos
xelxebar has joined #nixos
Myhlamaeus2 has quit [Remote host closed the connection]
ixxie has quit [Ping timeout: 245 seconds]
orivej has quit [Ping timeout: 268 seconds]
fendor has quit [Ping timeout: 268 seconds]
Myhlamaeus2 has joined #nixos
<judson_>
Experimenting with fixes for chrome/flash versions, I'm doing this in my NixOS configuration: chromium = super.callPackage ./chromium (self.config.chromium or {});
<vika_nezrimaya>
Question: how does Nixops handle 1) users; 2) hardware-configuration.nix?
lally has quit [Remote host closed the connection]
bakakuna has joined #nixos
lally has joined #nixos
vmandela has quit [Quit: Leaving]
<vika_nezrimaya>
looking at clever's deployment.nix, it doesn't handle either of those
<gchristensen>
you have to include hardware-configuration.nix in your nixops expression
<gchristensen>
and your users, too
<gchristensen>
nixops builds the entire remote system locally and sends it over
ris has joined #nixos
<exarkun>
Do I need to tell nixos anything about lvm, if I am trying to install on a system using lvm? Does nixos-generate-config on the system figure out everything it needs to figure out? I don't see anything about lvm in the generated configuration but I see it has picked the lvm-managed dm devices for / and /boot
bakakuna has quit [Remote host closed the connection]
<iqubic>
How can I expand this to include other dependencies that aren't haskell libraries?
bakakuna has joined #nixos
<iqubic>
Namely I want to make it so that cabal and cabal2nix are only available in a nix-shell.
<savanni>
You should be able to add them to buildInputs, as well.
<srhb>
iqubic: mkShell is pretty useful for that..
<savanni>
I'm not used to libraryHaskellDepends being part of an mkDerivation, but this would be my first guess.
<srhb>
But yeah, any buildInputs really..
Ariakenom has joined #nixos
<savanni>
Is this a custom mkDerivation command for the Haskell environment? Or have I just never known about things like libraryHaskellDepends?
<srhb>
savanni: Yeah it's custom to the cabal generic haskell builder.
<iqubic>
savanni: It's only part of mkDerivation because I'm calling it specially from the short two line shell.nix I posted.
<savanni>
Hm. Nice!
<srhb>
iqubic: Reminds me there's also shellFor
<iqubic>
So, how can I extend this?
<superbaloo>
is there a flip/flop (two system images that you swap for update) mechanism builtin w/ nixos? I would think it's the easiest platform to implement that with, but google is no luck :(
<srhb>
superbaloo: What do you mean system images?
<srhb>
superbaloo: Switching (almost) atomically is exactly what NixOS does, but maybe not in the way you're imagining, judging from "images"
<superbaloo>
yeah well, FS
<superbaloo>
I was thinking about building stateless images with nix-build for raspberrypi
<iqubic>
I have no idea how to add cabal and cabal2nix to my shell.nix here.
<vika_nezrimaya>
superbaloo: I think NixOS is even better than A/B partitions, since you can rollback to ANY point in time
<vika_nezrimaya>
for example, if I didn't do a garbage collection on sakura (my laptop) recently, I could boot into a month-old system...
<superbaloo>
I'd like to avoid writing on the raspberry as much as possible
<superbaloo>
because, performance is shit, and sd-card wear out
<exarkun>
switching doesn't involve writing much
<exarkun>
just flipping, what, one symlink? or is it a few?
<superbaloo>
and just building elsewhere?
<superbaloo>
that could work
<vika_nezrimaya>
Yeah, you can use another system as your remote builder! I'm currently working on fixing cross-compiling in Nixpkgs by the way
<exarkun>
sure. building is orthogonal to selecting a profile. you can build off-device (if you figure out how to cross compile and all that crap).
<vika_nezrimaya>
so we have some shared goals
<vika_nezrimaya>
exarkun: I know how to cross-compile!
<exarkun>
easier if you have an aarch64-capable rpi
<exarkun>
vika_nezrimaya: I greatly appreciate you encoding that knowledge into nixpkgs. :)
<vika_nezrimaya>
exarkun: in the manual, you mean?
<vika_nezrimaya>
Nixpkgs is already kinda cross-compatible
<exarkun>
vika_nezrimaya: no I mean just making it magically work so dumb people like me can use it without understanding
<exarkun>
vika_nezrimaya: yea ... emphasis on "kinda" I think?
<exarkun>
I've never managed to cross-compile a whole nixos
<superbaloo>
because no-one expects it, and makes shortcuts
<vika_nezrimaya>
The thread lists a snippet that allows you to cross-compile a configuration.nix and lists packages that are explicitly failing to build... (I need to post some more stuff there - I haven't posted some things yet!)
<vika_nezrimaya>
Right now I'm working on getting my Hydra instance public
<vika_nezrimaya>
it'll build minimal images and track cross-compiling progress
<vika_nezrimaya>
running on an old laptop
<vika_nezrimaya>
so don't expect it to be fast
<vika_nezrimaya>
It'll probably break on the next mass-rebuild and won't be able to keep up
<vika_nezrimaya>
unless my dream of having a whole team combine their processing power under Hydra's rule to churn out these packages becomes real
<iqubic>
What does shellFor do?
<iqubic>
And what is the _: in the second line?
<iqubic>
srhb: Those two questions were for you.
<srhb>
iqubic: It's documented in make-package-set.nix. Basically it collects the haskell _inputs_ from all of packages, eseentially make all the deps available in the shell. The _ is just because I'm not using the package set provided by the function.
<srhb>
iqubic: (ie. haskellPackages)
<iqubic>
I see.
<iqubic>
Cool.
<iqubic>
Will try that out.
o1lo01ol1o has quit [Remote host closed the connection]
<judson_>
And getting an error about called without required argument 'llvmPackages_8'
<judson_>
(sorry, distracted: this is while trying to do a simple overlay for Chromium)
<gchristensen>
you're instnalling from another distro?
<exarkun>
what do I do ... just move the _contents_ aside? unmount the boot partition and let switch-to-configuration write /boot to the root partition?
<exarkun>
yes
<iqubic>
What's the point of shell.nix?
<clever>
exarkun: exarkun what fs is the root partition?
<srhb>
iqubic: It's just convention.
<exarkun>
clever: ext4
<clever>
iqubic: having special expressions for nix-shell by default
<clever>
exarkun: should be fine to just have /boot on / then
<clever>
exarkun: efi or legacy?
<iqubic>
Blergh...
<exarkun>
legacy, if that's bios
<srhb>
iqubic: Why?
<clever>
exarkun: mbr or gpt tables?
<exarkun>
sometimes I hate computers
<iqubic>
srhb: Stuff isn't working the way I want it too.
<exarkun>
clever: I guess mbr but I don't know, how do I check
<clever>
exarkun: `fdisk -l /dev/sdX` and look for the disklabel type
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjANh
<clever>
exarkun: correct
<clever>
exarkun: thats why you should generally avoid editing hardware-configuration.nix
* exarkun
nods
<iqubic>
Issues are occurring.
<srhb>
iqubic: sadface.
<{^_^}>
[nixpkgs] @markuskowa opened pull request #67581 → sdrangel: add airpsy, hackrf, rtlsdr support → https://git.io/fjAAe
<cetjs2>
What is nixos 19.09 release out??
* cetjs2
ждет
<srhb>
cetjs2: Probably in a bit more than a month
o1lo01ol1o has joined #nixos
<srhb>
cetjs2: The release date is not usually set in stone. Freeze happens at the beginning of 09, and then it usually takes about a month, or a little more.
<clever>
vika_nezrimaya: store-uri is used by the c++ to upload build products to the store when the build is done
<vika_nezrimaya>
ok... sounds confusing
<vika_nezrimaya>
I guess I should copypaste that
zupo has joined #nixos
<srhb>
cetjs2: Maybe package requests and ones marked 3.skill:good-first-bug
<clever>
vika_nezrimaya: i too had to rummage thru the source to figure out why both are used
<cetjs2>
srhb thaks
<cetjs2>
*thanks
<vika_nezrimaya>
I suppose I could use deployment.keys to manage my binary cache private key?
<exarkun>
clever: Hm. Regarding partition table... what if it had been gpt instead of mbr?
<clever>
exarkun: legacy on gpt requires an extra bios boot partition, 1mb, no fs, not mounted
<exarkun>
ah
<exarkun>
but no other nixos-config changes? so if I am on a booted, working debian (and indeed it has such a bios boot partition) then it should also turn into a bootable, working nixos
<clever>
exarkun: yeah, no other changes
<clever>
exarkun: grub will detect if its gpt or mbr, and do the right thing
<exarkun>
cool
<{^_^}>
[nixpkgs] @jtojnar opened pull request #67584 → appstream-glib: clean up dependencies → https://git.io/fjAAV
fendor has joined #nixos
rauno has joined #nixos
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 258 seconds]
<vika_nezrimaya>
clever: NixOps seems to magically wait for port 22 to appear, yay!
<{^_^}>
[nixpkgs] @Infinisil merged pull request #66407 → lib/modules: Use options `apply` function even if no values are defined → https://git.io/fj7uS
<exarkun>
maybe it matters - programmatically in what context?
freeman42x has joined #nixos
npmccallum has quit [Quit: npmccallum]
<joko>
exarkun: I'm using readDir in home-manager to read the contents of a directory found in my home directory. I also have a Hydra job which tries to build the packages and readDir fails there due to restrict mode being on
bakakuna1 has joined #nixos
bakakuna has quit [Ping timeout: 248 seconds]
bakakuna1 is now known as bakakuna
<exarkun>
hm I still haven't looked at home-manager
selfsymmetric-pa has quit [Remote host closed the connection]
<joko>
exarkun: it's ok, it could have been something in my /etc/nixos/configuration.nix
<evanjs>
Are kernel config options defined by extraConfig in kernel overrides recursive? e.g. Can I just enable a kernel option and all the items it selects will be enabled during config generation?
<samueldr>
it's using one of the kernel tool to build the configuration, so it should be just as if you answered y or m to the question
<evanjs>
*Interesting. Had to enable ignoreConfigErrors to get things working
<evanjs>
And with structuredExtraConfig, some options are ignored completely. Might just be specifying them incorrectly
cyphase has quit [Read error: Connection reset by peer]
xkapastel has joined #nixos
<samueldr>
it could be that the option requires another one to be answered beforehand
<samueldr>
what I mean is that the question might not be asked by `make config`
<{^_^}>
[nixpkgs] @Ma27 opened pull request #67589 → gnome2.libglade: make python2 build optional → https://git.io/fjAp4
<evanjs>
Hrm. I tried just going through make menuconfig and answering only what was necessary to enable the feature, then checking that diff. Checking the output of nixops deploy does seem to exclude TRANSPARENT_HUGEPAGE (what I'm trying to enable), though.
chloekek has quit [Ping timeout: 272 seconds]
wildtrees has quit [Remote host closed the connection]
fendor has quit [Ping timeout: 245 seconds]
wildtrees has joined #nixos
<evanjs>
(the diff against make menuconfig + my options and the output of `make defconfig`)
<evanjs>
Interesting. Yeah, the option is just skipped completely in the output for make menuconfig.
<evanjs>
Everything else on the memory management options page is present