<pcarrier> @dhess we have our own golang process using journalctl raw output that persists its cursor periodically
<pcarrier> @dhess it's a bit obscure to implement but works really well
<pcarrier> (we didn't want cgo otherwise there's a native lib with a simple API)
<dhess> pcarrier: have you published it, by any chance?
remexre has joined #nixos
<pcarrier> No but I could share the relevant bits in a gist. Remind me in an hour or so if you're interested
<pcarrier> There's incredibly little to the journalctl part to be honest.
<dhess> pcarrier: nah, no worries. If I can't find something that looks well-maintained and open source, I'll probably end up writing something myself in Haskell :) But thanks for the offer.
<catern> hey #nixos
<catern> how would I use a library like nixpkgs.microsoft_gsl from an autotools-based project?
<catern> it's a single-header library
<catern> so all it contains is the file include/gsl/gsl (and a few other single headers)
joshuaks has quit [(Quit: WeeChat 1.9.1)]
<catern> how do I discover that include path with autotools?
<pcarrier> @dhess I'd recommend going straight to the raw format (very redis like) instead of dealing with json which had a bunch of weird corner cases and limitations
<catern> oh wait
<dhess> pcarrier: oh that's useful info. I was considering simply shipping the json via some kind of "-to-cloudwatch" too.
<catern> it's already on NIX_CFLAGS_COMPILE
<catern> but does autotools use that?
<dhess> /s/too/tool/
<pcarrier> @dhess though in the nix+ghc scenario the native lib could make more sense than a pipe. I'm just not a fan of cgo and don't really trust the systemd folks with ABI stability.
<dhess> pcarrier: yeah I was looking at the Haskell libsystemd-journal bindings
<pcarrier> It's a really small lib. I don't know how hard it is to ffi in Haskell but I wouldn't reject binding what you need
<dhess> ok. ocharles is the author, though, and I trust him to do a great job :)
griff_ has quit [(Quit: griff_)]
<dhess> it also seems fairly well-maintained
<pcarrier> journald is buggier than one would expect though. Don't trust features like rate limiting
<dhess> ok
<dhess> thanks for all the useful tips
<pcarrier> Sure thing. We run the logging infrastructure for our cloud platform on it so i learnt a few things along the way
<dhess> pcarrier: ahh interesting. What are you using for storage and analytics, if you can talk about it?
zab has quit [(Ping timeout: 255 seconds)]
<pcarrier> Currently ELK. We're currently investing in Kafka for other properties so I wouldn't be surprised if we used it there as well down the road
<dhess> pcarrier: self-hosted ELK or is this on AWS?
<pcarrier> Our hosting platform is 100% AWS and we use hosted services for everything stateful including ES. Our more recent properties use kubernetes and that's where we run Kafka
<dhess> ok, good to know. I think I'll start out that way as well, see what the initial costs look like
<pcarrier> Our new logging infrastructure (which might or might not ever be used by the hosting platform) is Kafka, Kafka streams for some processing, druid with Kafka ingestion, Kafka connect to s3 and s3 to redshift ingester
magnetophon has quit [(Remote host closed the connection)]
<pcarrier> We puke everything in Kafka and decide what to do with it as we go. We get a replicated low-latency WAL, decoupling, end to end compression, fine tuned retention policies, and Kafka streams + Kafka connect are nice building blocks
magnetophon has joined #nixos
<pcarrier> There's a pretty high operational cost but we sell a monitoring and performance insights service that ingests a looot of stats in near-real-time so that was well worth the cost for us. YMMV
<pcarrier> (By cost I mean "ops" overhead)
<dhess> pcarrier: any experience with CloudWatch Logs?
<pcarrier> Yup, but a long while ago, it probably doesn't reflect on today's product so I'll refrain from commenting
<dhess> ok
<pcarrier> Most of my personal day-to-day is on GCP now
<pcarrier> Stackdriver logging might be nice if you figure out how to integrate with it. The default of puking line-based system and app logs for GKE, almost completely unstructured, combined with a very limited and poorly documented query API and UI, makes it roughly unusable for us
<gchristensen> cloudwatch logs was not good in ~2016
<dhess> gchristensen: what were the biggest 1 or 2 issues you had with it?
<gchristensen> it was almost impossible to actually review the logs
<gchristensen> or make sense of them
joshuaks has joined #nixos
<gchristensen> but it was pretty close to brand new at the time
<dhess> ahh yeah, that bit :) I don't plan to use it for that, just for stashing and alerting. I think I will do the review with ELK or something like that.
<pcarrier> So I spent time with it at Airbnb which loved using all AWS services (can't remember the year, on my phone; pcarrier.ca). We got tons of files in s3 that we could periodically sync wherever, line delimited records, basically worked
<pcarrier> Somewhat high latency at times and very little visibility
<pcarrier> Again years ago.
<dhess> yeah I've seen some concerning things about latency as well
<pcarrier> We also didn't have a good way to select that we cared about vs didn't so mountains of useless data built up
<pcarrier> And back then at least, if an attacker got write access to the s3 bucket you basically lost your audit trails
<dhess> I think now there is a separate policy for writing logs
theseriousadult has joined #nixos
pie_ has quit [(Ping timeout: 260 seconds)]
<pcarrier> An append-only mode with yes-i-will-have-to-contact-support-then-wait-a-week-for-the-rollback-to-be-effective would have been awesome
<gchristensen> it'd be a good idea to make off-site backups anyway
<pcarrier> With aws, ingress is free, egress is not
jb55 has quit [(Ping timeout: 250 seconds)]
<pcarrier> They don't want your data to leave
<gchristensen> yeah
<gchristensen> I know
<pcarrier> I'd rather AWS said "yes it's append only gor everybody but our core oncall team and well-designed-and-audited workflows"
<Unode> If I want to install/test a python module (from pypi) in nix-shell, is there any quick nix-shell incantation that would provide this or do I have to go the usual buildPackage + python.withPackages approach?
<pcarrier> Anybody for my chroot in installer issue? What can I do to be able to nixos-rebuild in a chroot? I run into
<pcarrier> Literally on my phone now
<pcarrier> It's a good phone but I miss my model m ;)
<gchristensen> I've never seen that error
jb55 has joined #nixos
astsmtl has joined #nixos
astsmtl has quit [(Changing host)]
astsmtl has joined #nixos
<gchristensen> pcarrier: any ideas why clone() would fail on your system? https://github.com/NixOS/nix/blob/master/src/libstore/build.cc#L2187-L2196
jmeredith has quit [(Quit: Connection closed for inactivity)]
<pcarrier> IDK it's chroot in a nixos ISO
<pcarrier> Booted with copytoram
<gchristensen> may I ask why you're using chroot anywho?
<pcarrier> Sure. I completely changed my partitioning including adding luks, new EFI partition (larger) and renamed zvols
<pcarrier> None of my existing configs are bootable.
<pcarrier> You could argue I did it wrong but it's done :)
jb55 has quit [(Ping timeout: 268 seconds)]
<gchristensen> ok
<pcarrier> I did move the zvols around (through zpool over luks on a USB3 thumdrive) so all my filesystems are still around
<pcarrier> Well are back.
<gchristensen> I think it is safe, once you've remounted everything as you would a new install, to just run `nixos-install`
jb55 has joined #nixos
<pcarrier> I can't seem to be able to -I nixos-config=something-that-uses-relative-paths-including-for-overlays
<pcarrier> I can hack it to work then revert but I'd like to know I can reliably reinstall my system fast when needed. That includes if I lose my EFI partition and can only restore the zpool
<pcarrier> I kinda purposefully got to a destroyed system state to see if I could easily restore. So far the answer is a resounding nope.
<catern> is there any way in Nixpkgs to handle pulling an autotools-based project from git, where you need to first run "autoreconf --install" before running ./configure && make?
<gchristensen> pcarrier: hrm, I recover machines fairly regularly and quickly.
<pcarrier> @catern you can customize phases. Search for phases in the nixpkgs manual (sorry on phone)
<gchristensen> pcarrier: nixos-config should by default with nixos-install point to /mnt/etc/nixos/configuration.nix. from there, it should already be using relative paths for configs.
<catern> I don't want to customize my phases, though :)
<pcarrier> @catern you could insert between prepare and build for example
<catern> aha
<catern> I just add autoreconfhook to my buildInputs
<catern> easy
<gchristensen> configs / overlays
<hedning[m]> catern: There's autoreconfHook, which you should be able to just pass nativeBuildInput I think
<pcarrier> gchristensen: should as in I shouldn't design my system any other way?
<pcarrier> I guess I can make my /etc file a mere import of /repos/.../dell.nix. fair enough thanks
ris has quit [(Ping timeout: 260 seconds)]
<hedning[m]> catern: ah, duh ^^
<pcarrier> Can I drop into a shell from a failing stage 1? It seems to only offer to continue which panics or reboot which doesn't help
<gchristensen> pcarrier: I suppose so: I always use relative paths in my configs, since Nix handles them in a predictable fashion
<pcarrier> gchristensen: so do I, but my nixos-config is configured through itself to be somewhere else
<gchristensen> pcarrier: reading the source to stage 1 (https://github.com/NixOS/nixpkgs/blob/master/nixos/modules/system/boot/stage-1-init.sh) it looks like you can add the kernel parameter boot.shell_on_fail to allow a shell on failure
<gchristensen> pcarrier: you mind find other helpful parameters there s well
<pcarrier> Thanks!
<gchristensen> you're welcome!
<gchristensen> I love how transparent NixOS is, once you figure out how to traverse nixpkgs
<pcarrier> Reading code on a 6" display lol
<gchristensen> ouch :(
<pcarrier> Damn I wish I could use non legacy mounts in zfs
<pcarrier> Getting my system to /mnt would be a lot nicer
jb55 has quit [(Ping timeout: 268 seconds)]
<gchristensen> yeah it would be :/
<gchristensen> boxofrox: ping
<boxofrox> pong
<boxofrox> what's up?
<gchristensen> boxofrox: I don't think I'll be able to provide you a thorough review this week, please don't consider my feedback "blocking" by any means
<gchristensen> the PR is really great (though I am still having problems building a crate on darwin with it)
<boxofrox> no worries. i wasn't trying to rush you with the fyi earlier. i just wanted to make sure you hadn't missed my post.
<boxofrox> hmmm... I don't have darwin to test on myself.
<gchristensen> I can provide test output
<boxofrox> sure. i'll take a look
<boxofrox> gchristensen: just to confirm, this is building on OSX 10.8?
Mateon1 has quit [(Ping timeout: 248 seconds)]
<gchristensen> no, probably 10.12 or 10.13
Mateon1 has joined #nixos
<boxofrox> k, back to poking around
<pcarrier> Does nixos support maintaining 2 copies for /boot?
<gchristensen> what does that mean pcarrier?
<pcarrier> I guess I should just back it up.
<gchristensen> if that is a question the answer is yes :P
<gchristensen> ("should I back it up")
<pcarrier> It should be easy to rebuild tho
<pcarrier> Anyway. Anybody got a root on zfs on luks configuration.nix? I keep failing the initrd config apparently
<pcarrier> I would push what I have but I don't have git at hand
<pcarrier> (I will make note of that and add it to my installer for sure)
<gchristensen> unfortunately I'm not familiar with any existing config like that
<pcarrier> Gah it's nvme0n1p2
<pcarrier> Ok can I set a label on a luks device? :)
georges-duperon has quit [(Ping timeout: 260 seconds)]
<pcarrier> Alrighty I booted. Thanks everyone!
laggy_wifi has joined #nixos
<gchristensen> pcarrier: w00t I'm glad you're sorted
<gchristensen> pcarrier: now that you've done it once, do you think it'll be easier next time?
<pcarrier> @gchristensen it'll be automated next time
<gchristensen> I'll take that as a yes
mrkgnao has joined #nixos
<gchristensen> pcarrier: sounds like you're talking about >1 system if you're automating this. yeah?
<pcarrier> nope
<gchristensen> cool :)
<pcarrier> gchristensen: I keep at the very least a shell script for everything I'm likely to do 3+ times or _need_ to do urgently. https://github.com/pcarrier/config/blob/master/bin/installer2dell.sh
<gchristensen> right on
<gchristensen> I wrote this up last time I did it -- http://grahamc.com/blog/nixos-on-dell-9560
typeggze1o has joined #nixos
<pcarrier> Lol same gear
<pcarrier> Well I'm on a 9550
<gchristensen> its a pretty good machine, but mine introduces phantom key characters
<pcarrier> Omg I bought the 9550 for windows and would kill for wifi and Bluetooth that just works now
<pcarrier> I mean I would have before I figured it out for nix and presumably won't have to again
<gchristensen> :)
<pcarrier> Got a 5 year warranty so not swapping it with another one.
<gchristensen> I don't know how to setup bluetooth, ... so I haven't
typeggzero has quit [(Ping timeout: 248 seconds)]
<gchristensen> * also haven't needed to
<pcarrier> I use it only for meetings on the go
<pcarrier> Otherwise sennheiser hd 650 on schiit stack and blue yeti at the desk
<gchristensen> right on
<pcarrier> (I'm turning 30 and contractor working from home roughly at the same time, so I decided to finally invest in a nice setup. hence nix)
mrkgnao has quit [(Ping timeout: 255 seconds)]
<adisbladis> pcarrier: Nice :) I'm still on a pair of HD600 but thinking to upgrade to HD800S
<pcarrier> adisbladis: I already feel bad about how much I'm spending but they seem very nice. some might say Beyer T1 is top of that class
<pcarrier> (I ran ABX tests and HD650 is by far good enough for me; I don't have a horse in that high-end race)
andymandias has joined #nixos
typeggzero has joined #nixos
<adisbladis> pcarrier: I tried them all and the 800S are still my favourites :)
uptime is now known as downtime
typeggze1o has quit [(Ping timeout: 248 seconds)]
ryantrinkle has quit [(Ping timeout: 260 seconds)]
<boxofrox> gchristensen: best I can figure, the `-l framework=Security` isn't pulling in the ObjC library, so those symbols aren't found. Not sure if the `defaultCrateOverrides` needs a crate entry to include Security as a `buildInput`. I'll see if I can get my hands on an OSX Sierra VM and try a few things. I'll open a discussion on nest in case pmeunier knows what's missing. I do know the `rust-utils.nix` patch is
<boxofrox> minimal and basically builds pmeunier's Rust projects, atm.
<gchristensen> ok
<gchristensen> thank you :)
<gchristensen> how far out is Pijul 0.9 btw?
typeggze1o has joined #nixos
andymandias has quit [(Ping timeout: 260 seconds)]
<boxofrox> no idea. i'm not that close to the project. lthms might know. only reason I have some involvement is I wanted to build NixOS for my Rust software and `buildRustPackage` wasn't working, and I found carnix/rust-utils pretty easy by comparison to get working.
<gchristensen> ah ok
<gchristensen> yeah, carnix is looking pretty great!
typeggzero has quit [(Ping timeout: 240 seconds)]
<boxofrox> I'm enjoying it.
nuncanada has quit [(Remote host closed the connection)]
nuncanada has joined #nixos
downtime is now known as uptime
<boxofrox> gchristensen: actually, when using nix on darwin, do system libraries need to be specified as `buildInputs`? maybe `nativeBuildInputs` (not sure the difference)? or is there some other means by which nix packages build against system libraries?
<gchristensen> they do need to be specified
Synthetica has quit [(Quit: Connection closed for inactivity)]
sigmundv__ has quit [(Ping timeout: 255 seconds)]
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-17.09-small advanced to https://github.com/NixOS/nixpkgs/commit/b4a0c011e81 (from 2 hours ago, history: https://channels.nix.gsc.io/nixos-17.09-small)
nix-gsc-io`bot has quit [(Client Quit)]
<gchristensen> boxofrox: could you do me a favor? can you host a recent ... fetch? ... of pijul? 0.8.3 can't clone from nest and 0.9 isn't available as a tarball
mrkgnao has joined #nixos
<gchristensen> it is sort of funny to me that there is no way to get a working copy of the tool right now
erictapen has quit [(Ping timeout: 268 seconds)]
nuncanada has quit [(Read error: Connection reset by peer)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
mbrgm has quit [(Ping timeout: 260 seconds)]
griff_ has joined #nixos
endformationage has joined #nixos
griff_ has quit [(Client Quit)]
seanz has joined #nixos
seanz has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
andymandias has joined #nixos
griff_ has joined #nixos
mbrgm has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
jb55 has joined #nixos
griff_ has quit [(Client Quit)]
<boxofrox> gchristensen: 0.8.3's not working? O.o crap, didn't realize pmeunier was gonna shift the nest over to 0.9. `cargo install pijul` will pull 0.9-beta. I can set up a tarball of the latest pijul checkout... I think. One sec
griff_ has joined #nixos
hellrazor has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
<gchristensen> oh maybe 0.8.3 will clone, but its cargo.lock is broken so I can't build it with nix :P
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
griff_ has joined #nixos
griff_ has quit [(Client Quit)]
hellrazo1 has quit [(Ping timeout: 240 seconds)]
<boxofrox> oh, we might be in a predicament, then. 0.9 couldn't pull from nest on Friday. pmeunier is supposed to rebuild the history on the nest repos if I understood correctly.
<gchristensen> haha, I guess bootstrapping a new VCS is hard, messy businss
<boxofrox> so I think you want 0.8.3 until then. does `cargo install --vers=0.8.3 pijul` work? or does it have to be installed with nix?
<gchristensen> mm I can try that
Supersonic112 has quit [(Disconnected by services)]
Supersonic112_ has joined #nixos
Supersonic112_ is now known as Supersonic112
<boxofrox> yea, pijul's been a pain to work with. I like the "eating your own dog food" mentality of the devs, so the software will definitely improve, but... still painful to use at this point. I think it just corrupted by repo ><
<boxofrox> s/by/my/
<gchristensen> cool... ...
<gchristensen> hrm no cargo install fails to get it
<boxofrox> wierd. :/ give me a few minutes to double check my pijul tarball to see if it's corrupted, too. If not, I'll upload it and provide a link.
ryantrinkle has joined #nixos
tempeh has joined #nixos
tempeh has quit [(Client Quit)]
yorick has quit [(Ping timeout: 240 seconds)]
lambdamu has joined #nixos
<gchristensen> yeah now the install problems I'm having are because of bugs in cargo, not pijul
lambdamu_ has quit [(Ping timeout: 240 seconds)]
<gchristensen> ("'failed to acquire jobserver token: Resource temporarily unavailable (os error 11)")
jb55 has quit [(Ping timeout: 255 seconds)]
laggy_wifi_ has joined #nixos
laggy_wifi has quit [(Read error: Connection reset by peer)]
johnsonav has joined #nixos
<boxofrox> i just got that error, too. most of my work has been on my Arch desktop, but I thought I had pijul 0.8.3 running on my server :/
jb55 has joined #nixos
<boxofrox> anyway, tarball is broke. pulling down a fresh copy of pijul master branch, will tarball, upload and send your way :)
<gchristensen> thank you :)
mrkgnao has quit [(Ping timeout: 260 seconds)]
<boxofrox> gchristensen: https://transfer.sh/qAPrG/pijul-0.9-branch.tar.xz 3.6 MB tarball. defaults to master branch. first `cd pijul/pijul` and `cargo install` to get the "latest" 0.8.3 with branch fixes, then you can `pijul checkout pijul-0.9` if you want to play with the beta, but 0.9 won't pull from nest, while 0.8.3 will.
<boxofrox> good thing I checked that tarball. was 600+MB with all the cargo build files I had in there :D
<gchristensen> :D thank you a lot!
<boxofrox> my pleasure :)
<boxofrox> gotta run for a bit to grab some dinner, back in a few.
<gchristensen> I'm off for the night, myseelf
<boxofrox> k, cya tomorrow then o/
Wizek_ has quit [(Ping timeout: 255 seconds)]
jb55 has quit [(Ping timeout: 255 seconds)]
johnsonav has quit [(Ping timeout: 255 seconds)]
dj_goku has joined #nixos
jb55 has joined #nixos
cybrian has joined #nixos
laggy_wifi has joined #nixos
laggy_wifi_ has quit [(Read error: Connection reset by peer)]
joshuaks has quit [(Quit: WeeChat 1.9.1)]
cybrian has quit [(Read error: Connection reset by peer)]
<pcarrier> gchristensen: I don't seem to have much in terms of stats for temperatures or fan speeds or anything like that, would love to understand what's going on (my fans run often loudly). figured anything out of the 9560?
b has quit [(Quit: Lost terminal)]
mightybyte has left #nixos []
<pcarrier> oooh what the hell, it's on cpufreq performance.
<pcarrier> hmmm ondemand isn't even available!?
<clever> it depends on the hardware
<pcarrier> I thought ondemand was universal. here I have performance and powersave.
<clever> i used to use powernowd, which did on-demand in userland, via the userland govenor
<clever> no manual or userland option?
mizu_no_oto has joined #nixos
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
drakonis has quit [(Read error: Connection reset by peer)]
<pcarrier> I could go manual. I was hoping to use ondemand
<pcarrier> I can't figure out how to get nvidia_settings. I use bumblebee.
jmeredith has joined #nixos
<hodapp> hm, can any of you install 'meshlab' on 17.09?
<Ralith> clever: I assume ondemand support on new chips is a matter of time?
<pcarrier> oooh pulseaudio-dlna looks awesome
<pcarrier> might have to learn nix python packaging :P
<hodapp> pcarrier: huh, this does look neat
<pcarrier> right? :D
<pcarrier> I could easily switch my radio streams between headphones & Sonos when I go make tea.
<hodapp> that's what turning up the volume is for
<pcarrier> hodapp: there are walls between my tea maker and my office.
<pcarrier> and those headphones are rated for 500mW
<pcarrier> I also have Sonos in the bathroom :troll:
<pcarrier> liquid DnB for days.
<clever> Ralith: pretty much every chip i have supports ondemand
* hodapp hax0rs pcarrier's system to play Brian Eno on repeat
<Ralith> clever: okay?
<Ralith> that was not the question :p
<Ralith> the 9560 is not a bleeding edge device
babyflakes has joined #nixos
<Ralith> pcarrier: liquid DnB <3
<pcarrier> Ralith: my Sonos is running Liquid DnB from DI.fm roughly 24x7. I just lower the volume at night.
<hodapp> <3 di.fm
mmlb has joined #nixos
Lisanna has joined #nixos
<pcarrier> oops turning into a pumpkin. pulseaudio-dlna will have to wait :D
zzamboni has joined #nixos
Lisanna_ has quit [(Ping timeout: 260 seconds)]
cybrian has joined #nixos
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-17.09 advanced to https://github.com/NixOS/nixpkgs/commit/b4a0c011e81 (from 5 hours ago, history: https://channels.nix.gsc.io/nixos-17.09)
nix-gsc-io`bot has quit [(Client Quit)]
<Ralith> I wonder if DLNA provides mechanisms to synchronize playback from a single source across many sinks
mizu_no_oto has quit [(Quit: Computer has gone to sleep.)]
<pcarrier> Ralith: to the best of my knowledge no
<Ralith> that's frustrating
<pcarrier> yes and no. DLNA is simple
<Lisanna> hey, need some opinions on something: I have a derivation which packages a bash script that has some dependencies. I'm using this package in a custom NixOS module I've written. I've set the systemd.services "path" attribute to contain the package it uses, and that works, but the shell script also calls things like awk, which are not in the path.
<Lisanna> So, even though installing the package also correctly installs all of its dependencies, it has no way of accessing them. Should I try to include all of the package's dependencies in the "path" attr, or is there a better way to do this?
Obscurity[m] has quit [(Ping timeout: 276 seconds)]
Obscurity[m] has joined #nixos
Porkepix has quit [(Ping timeout: 264 seconds)]
Phil[m]1 has quit [(Ping timeout: 276 seconds)]
Phil[m]1 has joined #nixos
wmertens[m] has quit [(Ping timeout: 276 seconds)]
wmertens[m] has joined #nixos
taktoa has joined #nixos
Ralith_ has quit [(Remote host closed the connection)]
<Lisanna> My current idea is to make a wrapper in the derivation which wraps all the shell scripts with the PATH set to include the things it needs
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] aszlig pushed 1 new commit to release-17.09: https://git.io/vbOxi
<NixOS_GitHub> nixpkgs/release-17.09 65522ed aszlig: python/hetzner: 0.7.5 -> 0.8.0...
NixOS_GitHub has left #nixos []
Porkepix has joined #nixos
<fearlessKim[m]> How can I retrieve the store path to the VM "init" program in nixops (I suspect it is similar to nix) ? I would like to use the init path somewhere else in the VM definition
admin[m] has quit [(Ping timeout: 276 seconds)]
admin[m] has joined #nixos
dieggsy has quit [(Ping timeout: 255 seconds)]
zzamboni has quit [(Quit: Leaving.)]
ma27 has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbOpC
<NixOS_GitHub> nixpkgs/master 3eccd0b Orivej Desh: premake5: 5.0.0pre.alpha.11 -> 5.0.0-alpha12...
NixOS_GitHub has left #nixos []
<fearlessKim[m]> I think I got it "init=${config.system.build.toplevel}/init" though the VM now fails with 'Kernel panic - not syncing: Requested init XXX failed (error -2)'
<clever> fearlessKim[m]: that init will have to exist within the rootfs the initrd mounted
<fearlessKim[m]> clever: I wrote an extension to nixops to be able to set kernel/cmdline in the libvirt provider. In the nixops deployment file, I have 'cmdline="root=/dev/sda1 earlycon=ttyS0 console=ttyS0 init=${config.system.build.toplevel}/init";' I am not sure the config.system.build.toplevel is the VM one, might be the host one. I start libvirt with my own kernel but it doesn't use an initramfs
ma27 has quit [(Ping timeout: 276 seconds)]
<clever> fearlessKim[m]: the path to toplevel includes the hostname, and those should be unique
<clever> fearlessKim[m]: that should let you know which toplevel your looking at
<fearlessKim[m]> they do seem different so it must be VM specific
<fearlessKim[m]> I wonder if the /nix/store is mounted though. Doesn't appear in kernel log. If the initramfs is responsible for mounting the store, that might explain why it can't find the init
<fearlessKim[m]> (since I dont have the initramfs)
<clever> fearlessKim[m]: is /nix/store not on sda1 ?
Ralith_ has joined #nixos
<fearlessKim[m]> clever: it should be (dunno how nixops does), here is the booting log http://nixpaste.lbr.uno/w9L_J1Ti?text . I wonder if the readonly in 'VFS: Mounted root (ext4 filesystem) readonly on device' can bother the init
jsgrant has quit [(Read error: Connection reset by peer)]
Harpalus has joined #nixos
<srhb> https://github.com/NixOS/nixpkgs/issues/10183 implies that the issue has been fixed in some commit, but I can't figure out which. 17.09 appears to be affected still at least. Does anyone have a clue? (Missing domain name despite setting the option)
<clever> fearlessKim[m]: the init script can handle the root being readonly
<srhb> Alternatively, is there another way to get dnsdomainname and hostname etc. to report the domain name?
<fearlessKim[m]> I guess the path to the init is wrong then, any advice on how to mount /var/lib/libvirt/images/nixops-2a02ec8a-d903-11e7-aaa8-309c233b770e-client.img in the host ?
mightybyte has joined #nixos
wmertens[m] has quit [(Ping timeout: 276 seconds)]
apocolocyntos[m] has quit [(Ping timeout: 276 seconds)]
madknight has quit [(Ping timeout: 276 seconds)]
<clever> fearlessKim[m]: to start with, run file, blkid, and "fdisk -l" on it
<clever> fearlessKim[m]: and gist all 3
apocolocyntos[m] has joined #nixos
wmertens[m] has joined #nixos
Dr8128 has joined #nixos
madknight has joined #nixos
<mightybyte> How do I run nix-build on the output of cabal2nix?
<srhb> mightybyte: The easiest way iirc is to generate the --shell version of the derivation
<mightybyte> srhb: Ahhh!
<mightybyte> Right. I forgot about that.
<srhb> mightybyte: The default output is suited for inclusion in the haskell package set
<fearlessKim[m]> clever I can't get blkid to output anything but basically the file is QEMU QCOW Image (v3), 10737418240 http://nixpaste.lbr.uno/BaHaIgBI?nix
<srhb> mightybyte: (or with overrides/overlays, with callPackage)
<srhb> mightybyte: Admittedly it's a confusing default :)
<clever> fearlessKim[m]: ah, qcow cant be directly mounted on the host
<fearlessKim[m]> qemu-img can convert the drive so maybe I can conver tit to some other recommended format ?
<mightybyte> srhb: I couldn't get callPackage to work.
<clever> fearlessKim[m]: if you convert it to raw, it will be more easily readable, but that will also double your disk usage
<fearlessKim[m]> qemu-nbd can mount the qcow it seems I'll try
<mightybyte> nix-build -E '(import <nixpkgs> {}).callPackage ./pkg.nix {}'
<mightybyte> gives me
<srhb> mightybyte: It's the one from haskellPackages iirc. Let me check
<mightybyte> error: anonymous function at ... called without required argument ‘mkDerivation’
<clever> mightybyte: you need to use haskellPackages.callPackage
<mightybyte> Ahh, ok.
<srhb> Does anyone have a functioning hostname --fqdn ? As in, it does output the domain (if set)
<srhb> If I shutdown nscd, it works >_>
asuryawanshi has joined #nixos
<fearlessKim[m]> clever: could mount the qcow with libguestfs, seems like the init path is wrong. I edited the libvirt domain with the correct path and the VM correctly booted. I now need to find how to get the VM init path from nixops
<fearlessKim[m]> (btw thanks for the help <3 )
<clever> fearlessKim[m]: oh, and that init path will only be valid after nixops has done the copy-closure
<clever> fearlessKim[m]: it may be better to get it via /nix/var
<clever> /nix/var/nix/profiles/system/init
<fearlessKim[m]> clever: wow cool ! that's just perfect
griff_ has joined #nixos
stanibanani has quit [(Ping timeout: 248 seconds)]
<fearlessKim[m]> indeed that works just fine
jb55 has quit [(Ping timeout: 255 seconds)]
<fearlessKim[m]> I guess I have to provide some parameters that nixops usually set otherwise I can't seem to get an ip, systemd logs calls with "hostname=? addr=? terminal=? "
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis closed pull request #32288: plasma: 5.11.3 -> 5.11.4 (master...plasma_5_11_4) https://git.io/vbOok
NixOS_GitHub has left #nixos []
griff_ has quit [(Ping timeout: 276 seconds)]
mightybyte has left #nixos []
<Lisanna> is there a version of patchShebangs that is a Nix function rather than a bash function?
<Lisanna> seems like there's a whole bunch of useful functions in Nix but implemented in bash rather than Nix itself... like patchShebangs, patchelf, wrapProgram...
<clever> patchelf is in c
<Lisanna> s/implemented in bash/implemented in something only accessible from bash or stdenv/
<Lisanna> My use case right now is I want to use wrapProgram on some scripts to fix their PATH, but the scripts needs to have their shebangs fixed first
<hyper_ch> does it matter what those things are implemented in?
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
zzamboni has joined #nixos
<Lisanna> hyper_ch: yeah, because it means you have to do funky things like this: https://gist.github.com/ledettwy/4e7b38fbd426e9d5d78b25512aff526f
noob9 has joined #nixos
Harpalus has quit [(Quit: Leaving)]
<Lisanna> perhaps I should clarify: these things shouldn't necessarily be /implemented/ in Nix, but accessible from Nix as functions
<Lisanna> where could things like that go in nixpkgs?
<Lisanna> lib I guess?
kreetx has quit [(Ping timeout: 268 seconds)]
<noob9> could I get some help in installing virtualbox?
<etu> noob9: Sure :)
<noob9> I installed the package using `nix-env -i virtualbox` but running it gives an error "cannot access kernel driver, vboxdrv is not loaded..."
<etu> yeah, you probably want to use the service
<etu> virtualisation.virtualbox.host.enable = true;
<etu> In your config
<etu> It will fix all driver related things for you
<etu> And if you're on nixos you probably want to have something like this as well to have permissions for your user to use it: users.extraUsers.etu.extraGroups = [ "vboxusers" ];
<noob9> huh I was sure I tried that and it didn't work earlier
<noob9> well its working now so thanks
noob9 has quit [(Quit: Page closed)]
<fearlessKim[m]> is it possible to login via console into nixops VMs ?
orivej has quit [(Ping timeout: 255 seconds)]
<srhb> fearlessKim[m]: Depends where it's running.
<sphalerite> fearlessKim: like via SSH?
<srhb> fearlessKim[m]: (it's not a feature of nixops, you only get ssh from that)
proteusguy has quit [(Remote host closed the connection)]
<fearlessKim[m]> shrb: linus: nope via console=ttyS0
<fearlessKim[m]> like I have the promt in fornt of me but I am not sure there are credentials as the traditional way is via ssh
<clever> fearlessKim[m]: it will accept the same name/pw as ssh
<clever> you just wont have ssh key pairs
<hyper_ch> and you might want to als use nixpkgs.config.virtualbox.enableExtensionPack = true; but that will often turn into a compile orgy
frankpf has quit [(Ping timeout: 260 seconds)]
<sphalerite> Surely the extension pack isn't needed for serial
<fearlessKim[m]> clever: I don't set any user/password in my VM configuration.
<clever> fearlessKim[m]: then you have no way to sign in until you set one with passwd
<hyper_ch> sphalerite: I don't know what s/he/it wants to use vbox for
<hyper_ch> well, I did migrate now to qemu :)
endformationage has quit [(Quit: WeeChat 1.9.1)]
<hyper_ch> so no vbox recompile hassle orgy anymore
<fearlessKim[m]> hyper_ch: I use qemu too
cybrian has quit [(Read error: Connection reset by peer)]
<hyper_ch> but the only thing that sucks about qemu is that it doesn't have native host->guest folder sharing... neded to setup samba for that
<sphalerite> Oh right I didn't see that someone else was asking about vbox
<sphalerite> I thought it was in response to fearlessKim too
<hyper_ch> sphalerite: :)
<hyper_ch> etu: and you might want to als use nixpkgs.config.virtualbox.enableExtensionPack = true; but that will often turn into a compile orgy
<hyper_ch> and I think qemu is faster than vbox
dmi3y has joined #nixos
<fearlessKim[m]> I will try to create the user from nixops, I've also opened an issue here https://github.com/NixOS/nixops/issues/793
<adisbladis> hyper_ch: CPU/IO is but not graphics iirc
<sphalerite> Virtual box is based on qemu isn't it?
<adisbladis> sphalerite: Nope
<sphalerite> -autocorrect
<hyper_ch> even made a KDE start entry: virsh start WinJur; virt-viewer WinJur & :)
<hyper_ch> adisbladis: graphics?
<adisbladis> hyper_ch: If you want to run gui guests
<hyper_ch> I do run windows in i
<hyper_ch> I think it's faster
<adisbladis> hyper_ch: It might have changed but I have some memory of qemu being slower
<clever> there are some features qemu lacks, like clipboard sync
<hyper_ch> adisbladis: well, I don't use any fancy 3d stuff... it's just some software I need
<hyper_ch> clipboard sync works
<etu> hyper_ch: I'll manage without that. Sadly I don't have a choice at the moment.
<fearlessKim[m]> even with qemu-agent ?
dmi3y has quit [(Client Quit)]
<hyper_ch> no idea what qemu-agent is :)
<etu> hyper_ch: I'm adjusting to the macOS people who decided how we do things
<fearlessKim[m]> it's an insider spy that can help you get more details information about the VM/ do some extra stuff
<fearlessKim[m]> like for instance give you the IP without having to scrap DHCP leases
<hyper_ch> fearlessKim[m]: I installed the spice guest tool in the vm and then clipboard works
asuryawanshi has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis pushed 2 new commits to master: https://git.io/vb3f4
<NixOS_GitHub> nixpkgs/master 911dd47 Tomasz Rybarczyk: psc-package: 0.2.4 -> 0.2.5
<NixOS_GitHub> nixpkgs/master 561de38 adisbladis: Merge pull request #32292 from paluh/master...
NixOS_GitHub has left #nixos []
dmi3y has joined #nixos
<fearlessKim[m]> so I ave added `users.extraUsers.teto = { initialPassword="teto" }` in the VM definition but still can't login. during the deployement, it just rebuilt /nix/store/5ia8y1ayghzmrxm4bn8l5fr7sfg16qr2-libvirtd-ssh-image.drv so I am not sure serial login is supported ?
periklis has joined #nixos
proteusguy has joined #nixos
theseriousadult has quit [(Quit: Leaving)]
laggy_wifi_ has joined #nixos
laggy_wifi has quit [(Read error: Connection reset by peer)]
stuh84[m] has joined #nixos
adisbladis has left #nixos ["ERC (IRC client for Emacs 25.2.1)"]
adisbladis has joined #nixos
dywedir has joined #nixos
Ross has quit [(Ping timeout: 276 seconds)]
Ross has joined #nixos
asuryawanshi has joined #nixos
markus1189 has quit [(Ping timeout: 240 seconds)]
markus1189 has joined #nixos
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
davidak has quit [(Quit: Leaving.)]
ilyaigpetrov has joined #nixos
Tucky has joined #nixos
avn has quit [(Ping timeout: 248 seconds)]
reinzelmann has joined #nixos
<adisbladis> offlinehacker[m]: You should change your name in the maintainers file. Almost pinged the wrong person on a PR.
notdaniel has joined #nixos
Curiontice has joined #nixos
Curiontice has quit [(Remote host closed the connection)]
freusque has joined #nixos
Curiontice has joined #nixos
asymmetric3 has quit [(Ping timeout: 268 seconds)]
Soham has quit [(Ping timeout: 268 seconds)]
<srhb> enabling services.resolved also makes fqdn work. Wat...
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
ertes-w has joined #nixos
pie_ has joined #nixos
dmi3y has quit [(Quit: dmi3y)]
civodul has joined #nixos
Curiontice has quit [(Ping timeout: 248 seconds)]
tusj has quit [(Ping timeout: 250 seconds)]
Avaq has joined #nixos
i-am-the-slime has joined #nixos
davidak has joined #nixos
goibhniu has joined #nixos
<sphalerite> Is there a way to atomically replace a directory with a symlink in a shell script?
zzamboni has quit [(Quit: Leaving.)]
zzamboni has joined #nixos
mrkgnao has joined #nixos
thblt has joined #nixos
Neo-- has joined #nixos
<srhb> No, ln will not remove a directory.
MP2E has quit [(Remote host closed the connection)]
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
<fearlessKim[m]> ok I guess my user is not created because nixops creates the user after loggging in into the VM but as it can't log into it....
Harpalus has joined #nixos
Harpalus has quit [(Client Quit)]
Harpalus has joined #nixos
Dr8128 has quit [(Ping timeout: 260 seconds)]
jensens has joined #nixos
adamt_ has joined #nixos
aloiscochard has joined #nixos
asuryawanshi has quit [(Ping timeout: 276 seconds)]
<vaibhavsagar> hey, is there an easy way to override "syslog.socket"?
Dr8128 has joined #nixos
adamt has quit [(Ping timeout: 240 seconds)]
thc202 has joined #nixos
periklis has quit [(Ping timeout: 260 seconds)]
ThatDocsLady has joined #nixos
cybrian has joined #nixos
ckauhaus has joined #nixos
Harpalus has quit [(Remote host closed the connection)]
coot_ has joined #nixos
i-am-the-slime has quit [(Ping timeout: 248 seconds)]
ris has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
<sphalerite> vaibhavsagar: yes, just set systemd.sockets.syslog = {…};
orivej has joined #nixos
<vaibhavsagar> hmm, I did try something similar and it seemed to have no effect
<sphalerite> vaibhavsagar: you may need to make use of lib.mkForce to ensure that some of the settings override existing ones rather than being passed through
<vaibhavsagar> cool, let me try
zzamboni has quit [(Quit: Leaving.)]
<sphalerite> there may also be issues with overriding services provided by systemd where you can't replace a setting because it doesn't originate from nixos config. Not sure if there's a workaround for that currently.
zzamboni has joined #nixos
adamt_ is now known as adamt
<vaibhavsagar> It worked! Thanks sphalerite!
<sphalerite> great!
ntinos has joined #nixos
ylwghst has joined #nixos
FRidh has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej pushed 2 new commits to master: https://git.io/vb3YW
<NixOS_GitHub> nixpkgs/master 10c2dce Christian Kauhaus: vulnix: 1.3.4 -> 1.4.0...
<NixOS_GitHub> nixpkgs/master f1e1617 Orivej Desh: Merge pull request #32301 from ckauhaus/submit/vulnix-1.4.0...
NixOS_GitHub has left #nixos []
cybrian has joined #nixos
<ckauhaus> :-)
<sphalerite> srhb: re atomic replacement of a directory — is it not possible at all? I know nix's gc uses move to atomically "remove" it (and then recursively remove it properly later on)
<ckauhaus> rename is the only atomic option on POSIX systems
<sphalerite> oh right
<srhb> sphalerite: It's possible to atomically move something, yes
<sphalerite> rename(3): If newpath already exists, it will be atomically replaced, so that there is no point at which another process attempting to access newpath will find it missing. However, there will probably be a window in which both oldpath and newpath refer to the file being renamed.
<sphalerite> so you could replace a directory with a symlink by creating the symlink elsewhere and then rename()ing it?
<ckauhaus> there are still phantom issues with directory listings running
<ckauhaus> these may consist of several syscalls under the hood
<sphalerite> ooh, renameat2() has RENAME_EXCHANGE for exchanging two paths rather than removing the old one
ntinos has quit [(Read error: Connection reset by peer)]
tusj has joined #nixos
ilja_kuklic has joined #nixos
leat has quit [(Quit: WeeChat 1.9.1)]
<sphalerite> but I'm guessing this still isn't possible from a shell script without a specialised program that does the system call
<sphalerite> oh yeah, also, does anyone know a way to await a process's termination without being its parent?
<sphalerite> I usually do `while [[ -e /proc/$pid ]] ; do sleep 1 ; done` but it feels nasty
akfp has quit [(Read error: Connection reset by peer)]
<tilpner> sphalerite - tail --pid=$pid -f /dev/null
<orivej> sphalerite: AFAIK it's impossible to correctly await a foreign subprocess
akfp has joined #nixos
<tilpner> sphalerite - But apparently tail polls via kill -0, so not sure if that's much better
<sphalerite> aww
JosW has joined #nixos
<sphalerite> yeah, just straced it to find exactly that
<tilpner> I hope tail takes extra precautions to catch "$pid dies, another process is created with $pid, all in <1s"
<tilpner> (But I wouldn't bet on it)
<sphalerite> I wonder if there's actually a reason for there not being functionality for that
ylwghst has quit [(Quit: Lost terminal)]
Avaq has quit [(Ping timeout: 260 seconds)]
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
Avaq has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
__Sander__ has joined #nixos
townsend has joined #nixos
leat has joined #nixos
townsend has quit [(Client Quit)]
townsend has joined #nixos
<mbrock> there's a reason for everything in unix: worse is better!
patrl has joined #nixos
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
patrl has quit [(Client Quit)]
cybrian has joined #nixos
adamt has quit [(Ping timeout: 255 seconds)]
jtojnar has quit [(Quit: jtojnar)]
jtojnar has joined #nixos
iyzsong has joined #nixos
dan_b_ has joined #nixos
nwspk has quit [(Quit: Quit: *.banana *.split)]
nwspk has joined #nixos
sigmundv__ has joined #nixos
JosW has quit [(Quit: Konversation terminated!)]
simukis has joined #nixos
FRidh has quit [(Ping timeout: 260 seconds)]
patrl has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
nwspk has quit [(Quit: Quit: *.banana *.split)]
nwspk has joined #nixos
<hyper_ch> question: what does nixos need to make use of a usb 3.1 display port? I did get now an according usb3.1-to-displayPort plug and a displayport splitter so that I can attach 2x hdmi... but whenever I connect, no menus work anymore in kde and no image is transmitted
zzamboni has quit [(Quit: Leaving.)]
ilja_kuklic has quit [(Ping timeout: 248 seconds)]
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
<sphalerite> hyper_ch: does it work on Ubuntu? :p
<hyper_ch> sphalerite: how would I know?
<sphalerite> By trying it
avn has joined #nixos
<hyper_ch> I don't have ubuntu
zzamboni has joined #nixos
zzamboni has quit [(Ping timeout: 260 seconds)]
ma27 has joined #nixos
Avaq has quit [(Ping timeout: 255 seconds)]
proteusguy has quit [(Remote host closed the connection)]
<sphalerite> When I have issues like this I usually try it out on Ubuntu to see if it's nixos-specific (probably easier to solve that way)
townsend has quit [(Ping timeout: 255 seconds)]
<sphalerite> If it doesn't work on Ubuntu it's likely to just not work on Linux at all yet
notdaniel has quit [(Quit: Leaving)]
adamt has quit [(Ping timeout: 260 seconds)]
olto has joined #nixos
obtusetriangle has joined #nixos
<obtusetriangle> i like sex
cybrian has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] peti closed pull request #32112: Revive multiple outputs for Haskell packages. (master...hask-split-out) https://git.io/vbvul
NixOS_GitHub has left #nixos []
m0rphism has quit [(Quit: WeeChat 1.9.1)]
<makefu> i want to fix https://hydra.nixos.org/build/63000479/nixlog/1 but it seems like this is an upstream issue, right?
<obtusetriangle> where is the cybersex sluts
<obtusetriangle> i thought irc had horny sluts
<obtusetriangle> but i cant find them
asuryawanshi has joined #nixos
<makefu> #nixos: You're not a channel operator
asuryawanshi has quit [(Remote host closed the connection)]
asuryawanshi has joined #nixos
coot_ has quit [(Quit: coot_)]
townsend has joined #nixos
obtusetriangle has left #nixos [requested by gchristensen (obtusetriangle)]
FRidh has joined #nixos
<hyper_ch> sphalerite: didn't work on ubuntu... but might be due to optimius
zzamboni has joined #nixos
<Orbstheorem> Hi, I'm trying to run a mono application, but it says I'm missing libgtk-x11. I used nix-locate to find who provides it and though entering in a nix-shell with gnome.gtk, I still get the same error: https://paste.gnugen.ch/raw/7tEj
KABA_ has joined #nixos
<KABA_> Hi, I'm trying to use amdgpu-pro on current nixos master, and with vcunat's latest commit it does build, but I'm only getting a black screen. Here are the journalctl's relevant lines: https://gist.github.com/Mounium/2db8cf66df7ce72848e722b5d99238c4
proteusguy has joined #nixos
<makefu> Orbstheorem: do you have the whole expression for this package?
wolfcub has quit [(Remote host closed the connection)]
jtojnar has quit [(Remote host closed the connection)]
zzamboni has quit [(Quit: Leaving.)]
zzamboni has joined #nixos
olto has quit [(Quit: hasta luego)]
<Orbstheorem> makefu: which package
zzamboni has quit [(Ping timeout: 248 seconds)]
<makefu> ah i see, you've simply started with mono
<joepie91> Orbstheorem: you probably want the `gtk2-x11` dependency specified in your expression
<joepie91> oh
<joepie91> hm, I have no idea about directly executing mono
zzamboni has joined #nixos
<joepie91> you can use a nix-shell I suppose?
<joepie91> something like `nix-shell -p gtk2-x11`
<joepie91> not 100% sure whether that'll work
<makefu> i also do not really know if you can use mono like this ...
<makefu> my guess would be that mono needs to be built with gtk2 support?
Neo-- has quit [(Ping timeout: 248 seconds)]
zzamboni has quit [(Ping timeout: 276 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] LnL7 pushed 3 new commits to master: https://git.io/vb3RF
<NixOS_GitHub> nixpkgs/master 6530159 adisbladis: gox: 20140904 -> 0.4.0
<NixOS_GitHub> nixpkgs/master f019707 adisbladis: vault: 0.8.3 -> 0.9.0
<NixOS_GitHub> nixpkgs/master 97af931 Daiderd Jordan: Merge pull request #32300 from adisbladis/vault-0_9_0...
NixOS_GitHub has left #nixos []
yorick has joined #nixos
zzamboni has joined #nixos
<hyper_ch> how can I find out if display port is even supported by my usb type-c?
zzamboni has quit [(Ping timeout: 260 seconds)]
cybrian has quit [(Read error: Connection reset by peer)]
zzamboni has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nix] edolstra pushed 1 new commit to master: https://git.io/vb30W
<NixOS_GitHub> nix/master f1efb97 Eelco Dolstra: Fix #1635.
NixOS_GitHub has left #nixos []
zzamboni has quit [(Ping timeout: 276 seconds)]
<magnetophon> (how) can I set my locale LANG diferent from the rest of the locale values?
zzamboni has joined #nixos
zzamboni has quit [(Ping timeout: 255 seconds)]
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixpkgs-unstable advanced to https://github.com/NixOS/nixpkgs/commit/74857c9dceb (from 11 hours ago, history: https://channels.nix.gsc.io/nixpkgs-unstable)
nix-gsc-io`bot has quit [(Client Quit)]
zzamboni has joined #nixos
dan_b_ has quit [(Ping timeout: 260 seconds)]
FRidh has quit [(Remote host closed the connection)]
<akfp> KABA_: which commit is that? I'm trying to port it to 17.40 - is it done?
FRidh has joined #nixos
<akfp> and which kernel version?
patrl has quit [(Quit: WeeChat 1.9.1)]
zzamboni has quit [(Ping timeout: 240 seconds)]
pxc2 has joined #nixos
oida has quit [(Ping timeout: 260 seconds)]
zzamboni has joined #nixos
Oida has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixpkgs-unstable advanced to https://github.com/NixOS/nixpkgs/commit/561de381bcc (from 5 hours ago, history: https://channels.nix.gsc.io/nixpkgs-unstable)
nix-gsc-io`bot has quit [(Client Quit)]
<KABA_> akfp: I just used github's current master in the morning, I'm not actually sure but I think it was 4f8a65a163e688a
<KABA_> But e001bd2c2f698037 is the one required to build amdgpu-pro
Avaq has joined #nixos
Wizek_ has joined #nixos
katona has joined #nixos
zzamboni has quit [(Ping timeout: 268 seconds)]
zzamboni has joined #nixos
tmaekawa has joined #nixos
m0rphism has joined #nixos
aloiscochard has quit [(Quit: Connection closed for inactivity)]
zzamboni has quit [(Ping timeout: 248 seconds)]
zzamboni has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
KABA_ has quit [(Quit: Page closed)]
cybrian has joined #nixos
ma27 has quit [(Ping timeout: 255 seconds)]
tmaekawa has quit [(Quit: tmaekawa)]
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-unstable-small advanced to https://github.com/NixOS/nixpkgs/commit/4f8a65a163e (from 4 hours ago, history: https://channels.nix.gsc.io/nixos-unstable-small)
nix-gsc-io`bot has quit [(Client Quit)]
Lisanna1 has joined #nixos
ma27 has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
zzamboni has quit [(Quit: Leaving.)]
zzamboni has joined #nixos
{^_^} has joined #nixos
{^_^} has quit [(Changing host)]
{^_^} has joined #nixos
zzamboni has quit [(Client Quit)]
Synthetica has joined #nixos
Wizek_ has quit [(Ping timeout: 255 seconds)]
dan_b_ has joined #nixos
zzamboni has joined #nixos
Oida has quit [(Ping timeout: 248 seconds)]
oida has joined #nixos
zzamboni has quit [(Remote host closed the connection)]
zzamboni has joined #nixos
cybrian has joined #nixos
Neo-- has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] lsix closed pull request #32294: jenkins: 2.92 -> 2.93 (master...jenkins-2.93) https://git.io/vbODe
NixOS_GitHub has left #nixos []
Wizek_ has joined #nixos
ris has left #nixos ["Konversation terminated!"]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej pushed 2 new commits to staging: https://git.io/vb3wd
<NixOS_GitHub> nixpkgs/staging 4c9f27c Lancelot SIX: libunistring: 0.9.7 -> 0.9.8...
<NixOS_GitHub> nixpkgs/staging 6b42b5a Orivej Desh: Merge pull request #32307 from lsix/update_libunistring...
NixOS_GitHub has left #nixos []
ma27 has quit [(Remote host closed the connection)]
ma27 has joined #nixos
<Wizek> How can I specify Haskell dependencies with profiling enabled?
<Wizek> I'm trying like this at the moment: `nix-shell -p "haskell.packages.ghc821.ghcWithPackages (p: with p; [text hspec lens])" -j4 --run 'ghc Main.hs -prof'`
<Wizek> and ghc tells me `Perhaps you haven't installed the profiling libraries for package ‘lens-4.15.4’?`
dan_b has quit [(Quit: leaving)]
dan_b_ has quit [(Ping timeout: 260 seconds)]
cybrian has quit [(Read error: Connection reset by peer)]
i-am-the-slime has joined #nixos
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-unstable advanced to https://github.com/NixOS/nixpkgs/commit/8c3b96e58c0 (from 20 hours ago, history: https://channels.nix.gsc.io/nixos-unstable)
nix-gsc-io`bot has quit [(Client Quit)]
zzamboni has quit [(Quit: Leaving.)]
zzamboni has joined #nixos
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
Avaq has quit [(Ping timeout: 248 seconds)]
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
aloiscochard has joined #nixos
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
mkoenig has quit [(Ping timeout: 260 seconds)]
zzamboni has quit [(Client Quit)]
zzamboni has joined #nixos
tazjin has quit [()]
mkoenig has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] adisbladis closed pull request #32276: pgtap: 0.97.0 -> 0.98.0, highlight: 3.40 -> 3.41 (master...packageUpdates) https://git.io/vbOZk
NixOS_GitHub has left #nixos []
i-am-the-slime has quit [(Ping timeout: 260 seconds)]
zzamboni has quit [(Remote host closed the connection)]
jluttine has quit [(Ping timeout: 260 seconds)]
frankpf has joined #nixos
ma27 has quit [(Ping timeout: 255 seconds)]
eraserhd has quit [(Quit: leaving)]
iyzsong has quit [(Ping timeout: 248 seconds)]
szicari has joined #nixos
johann__ has joined #nixos
<Wizek> different question: How can I map over all values of a set/hashmap?
<Wizek> and change the values
ryantrinkle has quit [(Ping timeout: 248 seconds)]
<Wizek> something like this: `map (p: p + 2) {a=1; b=2;}`
tpanum has joined #nixos
<the-kenny> Wizek: I think there's lib.mapAttrs (k: v: v+1) ...
<tpanum> I need to add a printer using samba. Is 'gutenprint' drivers sufficient for this? I cannot seem to get it to work :/
<Wizek> the-kenny: that worked, thanks!
<Wizek> nixpkgs.lib.mapAttrs (k: p: p + 2) {a=1; b=2;}
ma27 has joined #nixos
nix-gsc-io`bot has joined #nixos
<nix-gsc-io`bot> Channel nixos-unstable advanced to https://github.com/NixOS/nixpkgs/commit/3eccd0b11d1 (from 9 hours ago, history: https://channels.nix.gsc.io/nixos-unstable)
nix-gsc-io`bot has quit [(Client Quit)]
johann__ has quit [(Quit: Leaving.)]
johann__ has joined #nixos
griff_ has joined #nixos
tpanum has quit [(Ping timeout: 248 seconds)]
magnetop` has joined #nixos
magnetophon has quit [(Ping timeout: 248 seconds)]
i-am-the-slime has joined #nixos
johann__ has quit [(Quit: Leaving.)]
johann__ has joined #nixos
tmaekawa has joined #nixos
jvtbatman has joined #nixos
johann__1 has joined #nixos
pxc2 has quit [(Ping timeout: 268 seconds)]
johann__ has quit [(Ping timeout: 260 seconds)]
magnetop` has quit [(Ping timeout: 260 seconds)]
reinzelmann has quit [(Quit: Leaving)]
slyfox has quit [(Ping timeout: 248 seconds)]
slyfox has joined #nixos
johann__1 has quit [(Ping timeout: 255 seconds)]
johann__ has joined #nixos
jtojnar has joined #nixos
bbarker has joined #nixos
lebel has joined #nixos
<aminechikhaoui> Hi, what would it take to update nix to nixUnstable in a macOS multi-user installation
<aminechikhaoui> is it just `nix-env -i nixUnstable` and then restart the daemon ?
i-am-the-slime has quit [(Ping timeout: 255 seconds)]
i-am-the-slime has joined #nixos
<LnL> don't think that will update the correct profile
peacememories has joined #nixos
<bbarker> Is there a way to override/ add to buildInputs for a dependency used in a nix-shell expression? I'm trying to add pcre to uwsgi here, but this way is obviously wrong since buildInputs isn't a function input: https://pastebin.ca/3944288
Avaq has joined #nixos
drakonis has joined #nixos
zzamboni has joined #nixos
bfrog has joined #nixos
zzamboni has quit [(Remote host closed the connection)]
zzamboni has joined #nixos
jvtbatman has quit [(Remote host closed the connection)]
cybrian has joined #nixos
laggy_wifi_ has quit [(Ping timeout: 260 seconds)]
tmaekawa has quit [(Quit: tmaekawa)]
zzamboni1 has joined #nixos
zzamboni has quit [(Remote host closed the connection)]
cement has joined #nixos
pxc2 has joined #nixos
<aminechikhaoui> LnL: How should that be done then ?
<aminechikhaoui> I would assume as root that should work no ?
<LnL> yes, I think of you login with sudo su -
pxc2 has quit [(Ping timeout: 250 seconds)]
kreetx has joined #nixos
<LnL> otherwise nix-env -p /nix/var/nix/profiles/default ....
i-am-the-slime has quit [(Ping timeout: 276 seconds)]
magnetop` has joined #nixos
i-am-the-slime has joined #nixos
<hyper_ch> how can I test if type-c usb 3.1 port also supports displayport mode?
Slabity has joined #nixos
ryantrinkle has joined #nixos
freusque has quit [(Quit: WeeChat 1.9.1)]
jvtbatman has joined #nixos
<jvtbatman> Anyone having problems with firefox "flickering"? Pages turn white for like 1/10th second each 5 seconds
* goibhniu hasn't seen that
<jvtbatman> Or, it's like 1/2 sec white and 1/2 sec correct page
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej closed pull request #32299: gettext: add setup hook populating GETTEXTDATADIRS (staging...gettext-datadirs) https://git.io/vbObH
NixOS_GitHub has left #nixos []
<goibhniu> could it be that you have some experimental firefox option enabled?
zzamboni1 has quit [(Quit: Leaving.)]
<goibhniu> in about:config
jacob has joined #nixos
<hyper_ch> (one could always upgrade to Chromium)
<jvtbatman> goibhniu: It's a fresh install and I can't find anything such?
cybrian has quit [(Read error: Connection reset by peer)]
<jvtbatman> hyper_ch: The new firefox is imho way better
<hyper_ch> jvtbatman: how so?
<jvtbatman> hyper_ch: Much faster since Firefox 57. Large parts were switched out to concurrently running rust instead of the c++, maybe it's just in my head but it feels so smooth
frankpf has quit [(Ping timeout: 260 seconds)]
<hyper_ch> no idea how ff 57 was
magnetop` has quit [(Ping timeout: 248 seconds)]
<hyper_ch> it's been ages since I used ff
pxc2 has joined #nixos
<jvtbatman> goibhniu: Found such a flag but turning it off fixed nothing :/
<goibhniu> could it be an issue with your GPU?
<jvtbatman> hyper_ch: I really like it since they released the "quantum" version. First time with a problem now, with fresh install of nix
<goibhniu> I'd guess it's not NixOS related, but no idea really.
<jvtbatman> goibhniu: Thought that as well so I explicitly put "intel" as videoDrivers with no luck
<hyper_ch> jvtbatman: optimus?
<gchristensen> MichaelRaskin: what is the name for the << and >> symbols you use?
<hyper_ch> (I blacklisted noveau for my optimus system since I don't have a need for it and had issues)
pxc2 has quit [(Ping timeout: 268 seconds)]
<gchristensen> MichaelRaskin: also, how do you type them? the only fun symbols my keyboard can type is § and ±
<bbarker> any suggestions on overriding buildInputs, as wrongly attempted in https://pastebin.ca/3944288 ?
<jvtbatman> hyper_ch: Isn't optimus some nvidia thing?
<hyper_ch> yeah, where you have intel and nvidia
<hyper_ch> since I don't need the nvidia, I blacklisted the nouveau driver
<jvtbatman> hyper_ch: Oh, I just have intel on this pc
Sonarpulse has quit [(Ping timeout: 255 seconds)]
katona has quit [(Quit: Connection closed for inactivity)]
<jvtbatman> hyper_ch: With blacklisting you mean disabling in /etc/nixos/conf..?
jluttine has joined #nixos
<hyper_ch> hardare-configuration.nix boot.blacklistedKernelModules = [ "nouveau" ];
peacememories has quit [(Quit: My MacBook has gone to sleep. ZZZzzz…)]
leat has quit [(Quit: WeeChat 1.9.1)]
<jvtbatman> hyper_ch: Thanks, guess I'll try that!
<goibhniu> if you don't have an nvidia GPU that shouldn't matter though
<hyper_ch> jvtbatman: if you don't have optimus there's no need for it
<goibhniu> another wild guess ... have you got compositing enabled?
cybrian has joined #nixos
<jvtbatman> Guess it cant hurt to blacklist it anyway?
<jvtbatman> goibhniu: I do! I'll try disabling it!
jvtbatman has quit [(Quit: Lost terminal)]
jvtbatman has joined #nixos
dan_b has joined #nixos
<jvtbatman> goibhniu: That was it... Thanks!
<goibhniu> wow, great!
<hyper_ch> isn't compositing a good thing?
<goibhniu> not when it causes everything to flicker :D
<jvtbatman> hyper_ch: Isn't it just transparancy of windows?
<goibhniu> there might be some way to tweak it though
<hyper_ch> no idea
<etu> jvtbatman: No, it's more that that
<hyper_ch> A compositing window manager, or compositor, is a window manager that provides applications with an off-screen buffer for each window. The window manager composites the window buffers into an image representing the screen and writes the result into the display memory.[1][2]
<jvtbatman> The weird thing is that I set compton to disabled and restarted dispmanager and now I still have transparent terminals with a functioning firefox?
phdoerfler has joined #nixos
nuncanada has joined #nixos
<joepie91> hyper_ch: jvtbatman: the simplified explanation is that 'compositing' means that every application renders to its own 'virtual display', and that all the rendered windows of all the applications are then combined into a single 'final' image, which is what you see on your screen -- this replaces the 'direct rendering' technique where each application renders its own stuff to the same 'real' display in sequence
<joepie91> hyper_ch: jvtbatman: compositing is indeed a necessity for things like window transparency and other post-processing effects (including taskbar thumbnails) as those require the window manager to be able to manipulate drawin windows individually, with knowledge of what image data belongs to what window -- however, that is not the *purpose* of compositing, rather compositing is just a more modern way to handle application rendering, and it also reduces
<joepie91> rendering artifacts (like the weird 'trail of windows' artifact you used to get on Windows 98 and such when you dragged something over a crashed application)
akfp has quit [(Read error: Connection reset by peer)]
<joepie91> s/drawin/drawn/
akfp has joined #nixos
<joepie91> in other words: you want compositing if at all possible, and newer window managers often *only* support compositing; but you *can* switch to direct rendering if your window manager supports it, you just get more rendering artifacts and lose the ability to do any kind of window manager postprocessing, thumbnailing, etc.
townsend has quit [(Remote host closed the connection)]
<joepie91> hyper_ch: jvtbatman: as an analogy; direct rendering is like somebody drawing an entire scene onto a piece of paper at once, whereas compositing is drawing individual pieces on objects of paper and then arranging them on the 'final' piece of paper and glueing them on top of each other to create a scene; this allows you to eg. rotate objects in the scene after the fact, something that would have been impossible if you drew the entire scene at once
pxc2 has joined #nixos
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
<Yaniel> also this way it is possible ti reuse components that were drawn earlier
adamt_ has joined #nixos
adamt_ has quit [(Changing host)]
adamt_ has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
drakonis_ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej closed pull request #32308: nixos/networking: support static resolv.conf (master...static-resolv-conf) https://git.io/vb3rR
NixOS_GitHub has left #nixos []
akfp has joined #nixos
dejanr has joined #nixos
drakonis has quit [(Ping timeout: 250 seconds)]
jensens has quit [(Ping timeout: 260 seconds)]
<jvtbatman> joepie91: Thanks for the thorough explanation! I guess slim with i3 runs compositing then and they just need compton for calculating opacity results?
adamt has quit [(Ping timeout: 255 seconds)]
Neo-- has quit [(Ping timeout: 240 seconds)]
playX has joined #nixos
johann__ has quit [(Quit: Leaving.)]
<playX> Hello everyone, how i can enable mate? I don't find option to use Mate DE
Harpalus has joined #nixos
<goibhniu> hi playX, I don't think anyone has packaged it yet
<playX> in packages have mate-coomon and mate-desktop
markus1189 has quit [(Ping timeout: 240 seconds)]
<goibhniu> oh!
markus1189 has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
patrl has joined #nixos
<playX> some ideas to enable mate?
<goibhniu> playX: did you try services.xserver.desktopManager.mate.enable = true;
<playX> that option doesn't exists
markus1189 has quit [(Ping timeout: 264 seconds)]
markus1189 has joined #nixos
jtojnar has quit [(Quit: jtojnar)]
<goibhniu> playX: what does `nix-info` say?
<goibhniu> I see the option here on nixos-unstable
<playX> that option what u pasted?
<goibhniu> `nixos-option services.xserver.desktopManager.mate.enable`
<playX> maybe later try to update to unstable
bennofs has joined #nixos
<playX> goibhniu: can i watch what new packages in 18.03 when i i'm using 17.09
mrkgnao has quit [(Ping timeout: 255 seconds)]
cybrian has joined #nixos
<playX> Ohh, installed borderlands 2 from steam, it's just closes after pressing start button on gamped :( Any ideas to fix that problem?
<playX> gamepad*
i-am-the-slime has quit [(Ping timeout: 255 seconds)]
<adamt_> playX: Probably missing dependencies
adamt_ is now known as adamt
<goibhniu> you can always try it out and roll back to 17.09 from unstable, if it doesn't work out
<tilpner> playX - Restart Steam in offline mode
<joepie91> jvtbatman: from a quick search, it seems that i3 does not have native compositing functionality and relies on compton for this
<joepie91> jvtbatman: (ie. it would probably revert to direct rendering without compton? I don't know much about i3)
<playX> tilpner: and what next?
<tilpner> playX - That's it
endformationage has joined #nixos
<playX> and i can't play online?
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] andir opened pull request #32313: mstpd: svn 61 -> 0.0.5-g2522c6e (master...mstpd-to-github-0.0.5) https://git.io/vb3jb
NixOS_GitHub has left #nixos []
<tilpner> playX - You can with #31480
<playX> tilpner: what that number mean?
<tilpner> It's a shorthand for the link I posted earlier
playX has quit [(Quit: Page closed)]
playX has joined #nixos
<playX> tilpner: Maybe some other ways to fix this problem? I want to play with my friends
i-am-the-slime has joined #nixos
<tilpner> Yes, hold on
<tilpner> playX - How did you install Steam?
<playX> tilpner: Alright, i'm waiting
adamt has quit [(Ping timeout: 276 seconds)]
<playX> tilpner: writed in configuration.nix "Steam"
<playX> and nixos-rebuild switch
jvtbatman has quit [(Remote host closed the connection)]
<tilpner> Good. Now replace that with the following: (steam.override { extraPkgs = p: [ p.iana-etc ]; })
<playX> i need to type this string at new line?
dan_b has quit [(Ping timeout: 248 seconds)]
<tilpner> Doesn't need its own line, but it would be more readable. Remember to remove the other steam
<playX> tilpner: where i need to paste it? https://pastebin.com/vGitJQUp
<tilpner> environment.systemPackages = [ ... HERE ... ];
<tilpner> But with the parentheses, you can't just remove theme
<tilpner> *them
cybrian has quit [(Read error: Connection reset by peer)]
<playX> tilpner: rebuilding nixos
playX has quit [(Ping timeout: 260 seconds)]
Neo-- has joined #nixos
playX has joined #nixos
<playX> tilpner: Everything works, but my gamepad not working normaly at dinput
<playX> when instaled just steam it works normally, now button A it's B button
patrl has quit [(Quit: WeeChat 1.9.1)]
<goibhniu> playX: is that something you can remap in the game settings?
<playX> just restarted game and gamepad works normaly, strange
erictapen has quit [(Ping timeout: 248 seconds)]
Avaq has quit [(Ping timeout: 250 seconds)]
i-am-the-slime has quit [(Ping timeout: 248 seconds)]
* herzmeister[m] sent a long message: herzmeister[m]_2017-12-04_17:11:38.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/XcsVIOIYfHSzkRaupZCTnYqM>
georges-duperon has joined #nixos
<herzmeister[m]> (And I'm on nixos unstable channel)
lebel has quit [(Remote host closed the connection)]
<gchristensen> herzmeister[m]: does Matrix indicate to you when you're going to send a long message?
<herzmeister[m]> gchristensen: no :) where's it cut off?
<gchristensen> it doesn't show us the message at all
<gchristensen> instead it says " herzmeister[m] sent a long message: herzmeister[m]_2017-12-04_17:11:38.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/XcsVIOIYfHSzkRaupZCTnYqM>"
<herzmeister[m]> all hailz to teh cloud
<das-g[m]> I guess Matrix doesn't have the concept of a "long message". That's probably just the Matrix->IRC bridge.
cybrian has joined #nixos
<herzmeister[m]> (maybe not a bad idea but it should show the beginning of the text at least)
mkoenig has quit [(Ping timeout: 255 seconds)]
<herzmeister[m]> das-g: almost certainly
digitus has joined #nixos
__Sander__ has quit [(Quit: Konversation terminated!)]
erasmas has joined #nixos
danl1240 has joined #nixos
mkoenig has joined #nixos
<gchristensen> someone from the matrix-irc channel says it is user error :( "matrix users knowing the room is bridged to an IRC room should no use them imo"
<herzmeister[m]> sure, always blame the user. 😆
<herzmeister[m]> there is no indicator if the message is too long or not, they want me to look up the restrictions and count chars?
<gchristensen> a few things Matrix does seems a bit user-hostile, but I'm pretty sure not on purpose
<herzmeister[m]> i'm also in the Matrix channel but not in the freenode #matrix-irc, i'll discuss later with the folks maybe
<gchristensen> there is also #irc:matrix.org
magnetop` has joined #nixos
magnetop` is now known as magnetophon
goibhniu has quit [(Ping timeout: 260 seconds)]
<dash> so many choices
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
* hyper_ch thinks matrix is weird
<dash> it is! but so far as i can tell it's the best chance we got to do better than irc
<gchristensen> it is unfortunate they're choosing to alienate the same users they're trying to woo
<dash> oh?
<gchristensen> I mean, it comes off that way to me
<gchristensen> I think the "Long Messages" tend to make it very hard for matrix users to get help, and they have no idea why nobody is replying to them
Neo--- has joined #nixos
playX has quit [(Quit: Page closed)]
<clever> the slack irc gateway handles long messages better, it just spams the channel, lol
<clever> but you have to choose to connect to slack, so you expect that kind of thing, and they can just bend the rules of irc
<dash> there's no single good tradeoff probably :-/
Neo--- has quit [(Remote host closed the connection)]
<gchristensen> dash: no perfect solution, but I think there are great tradeoffs
Neo--- has joined #nixos
davidak has quit [(Ping timeout: 255 seconds)]
<gchristensen> dash: ex: "helloes again, my second (not init-boot) encrypted harddrive (lvm on luks) cannot found during boot by systemd. No clue if bug, missing feature, or my fault, what would I be missing? My config: (truncated, full message at https://matrix.org/_matrix/media/v1/download/matrix.org/XcsVIOIYfHSzkRaupZCTnYqM)"
Neo-- has quit [(Ping timeout: 255 seconds)]
<herzmeister[m]> could be configurable at channel operator or (if feeling liberal) even user level
<herzmeister[m]> yes, something like that would be acceptable
<gchristensen> my current feeling, having seen people be unhappy after nobody helps them after they unknowingly sent a Long Message, is I'd rather not be bridged than offer than crappy experience
<Ralith> herzmeister: it's using multiple lines that triggers it, fwiw
<gchristensen> as Matrix's poor implementation reflects not on them, but on the IRC communities
<gchristensen> which is why I say they're being hostile to the people they're trying to woo: IRC communities
drakonis_ has quit [(Read error: Connection reset by peer)]
<herzmeister[m]> and here i am thinking the line breaks would improve readibiilty. `:-/`
Sonarpulse has joined #nixos
<cransom> i think multiline messages/dumps are why pastebins exist
<gchristensen> herzmeister[m]: they do! and I happily click links to multi-line text to help people. however, it is very hard to motivate people to click a link when they have _no idea_ what is on the other side
<herzmeister[m]> good, then matrix should integrate pastebin, maybe people would trust it more. :)
<gchristensen> given an arbitrary question, what are the chances any given person can help with it? giving context to the long message URL I think is all it needs
<gchristensen> to be substantially better
Neo-- has joined #nixos
<joepie91> gchristensen: isn't what you're suggesting how irccloud does it as well?
Neo--- has quit [(Read error: Connection reset by peer)]
<herzmeister[m]> yes definitely, thats why it should show the beginning of the message at least
<gchristensen> I have no idea how IRCCloud does it
<joepie91> if I'm recalling correctly
<gchristensen> probably because they do something reasonable :P
<joepie91> I do know that the Prosody MUC ("XMPP channel") uses the same approach, and it works there
<joepie91> (though there it's the server that truncates and 'pastebinbs')
<joepie91> pastebins *
<gchristensen> this is how irccloud does it: "<pcarrier> any takers for my problem with nix-env and overlays? it's possibly something stupid. https://www.irccloud.com/pastebin/oEywH2fq/"
* simpson Are you sure you want to paste 22 lines?
<gchristensen> C-Y
<pcarrier> indeed.
<joepie91> simpson: that suffers from the "people do not read dialog boxes" problem
<joepie91> :P
<simpson> joepie91: Dialog box? My IRC client blocks entirely and requests that I enter a relatively rare two-finger salute.
<simpson> Any mispaste of more than two lines is entirely my own fault.
davidak has joined #nixos
<joepie91> simpson: that sounds like a dialog box equivalent to me :D
<gchristensen> IMO Matrix makes many improvements over IRC and I want them to succeed
<gchristensen> I think having integrated pastes is a huge improvement
<simpson> joepie91: I assure you that there are tangible differences between keyboards and mice.
Neo-- has quit [(Ping timeout: 240 seconds)]
<joepie91> nettalk is interesting in that it has a multi-line input box by default and just doesn't send your multiline paste until you hit enter
<joepie91> konversation shows a dialog, including an "edit message first" option that has a separate box where you can remove newlines and stuff
jb55 has joined #nixos
Neo-- has joined #nixos
rihards has joined #nixos
Neo-- has quit [(Remote host closed the connection)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] mikoim opened pull request #32316: logkeys: 2015-11-10 -> 2017-10-10 (master...pkg/logkeys) https://git.io/vbsmb
NixOS_GitHub has left #nixos []
sigmundv__ has quit [(Ping timeout: 260 seconds)]
ryantrinkle has quit [(Ping timeout: 255 seconds)]
Lisanna1 has quit [(Ping timeout: 276 seconds)]
georges-duperon has quit [(Ping timeout: 260 seconds)]
jb55 has quit [(Ping timeout: 240 seconds)]
<herzmeister[m]> so back again to my original question, can anybody help me? so back again to my original question, can anybody help me? my second (not init-boot) encrypted harddrive (LVM on LUKS) cannot be found during boot by systemd.
<herzmeister[m]> No clue if bug, missing feature, or my fault, what would I be missing? My config:
<herzmeister[m]> `filesystems."/hdd-data" = { device = "/dev/mapper/luksrootsda"; encrypted.blkDev = "/dev/sda1"; encrypted.enable = true; encrypted.keyFile = "/root/keyfile"; encrypted.label = "luksrootsda"; fsType = "ext4"; };`
<herzmeister[m]> Also, when I `nixos-rebuild` this with `switch`, my system will reproducibly hard-crash, entering systemd recovery shell, but freeze there. With `boot` I can circumvent that.
bennofs has quit [(Ping timeout: 260 seconds)]
tmaekawa has joined #nixos
cybrian has quit [(Read error: Connection reset by peer)]
tmaekawa has quit [(Client Quit)]
ryantrinkle has joined #nixos
<herzmeister[m]> And I am using nixos unstable channel.
leat has joined #nixos
<Yaniel> you could add the automount/notrequired flags like what is recommended for smb mounts
<Yaniel> that should at least let it continue so that you can debug it further
i-am-the-slime has joined #nixos
grw has joined #nixos
ryantrinkle has quit [(Ping timeout: 260 seconds)]
zraexy has joined #nixos
<MichaelRaskin> gchristensen: re: ‘‘«»’’
<gchristensen> yeah what are those called?
<MichaelRaskin> Unicode name is «left/right pointing double angle quotation mark»
<MichaelRaskin> X11 name is guillemotleft/guillemotright
<MichaelRaskin> I have _extensive_ Level3 remappings using xmodmap.
<MichaelRaskin> I basically can type most diacritics that I can write down, for example…
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
erictapen has quit [(Remote host closed the connection)]
erictapen has joined #nixos
<earldouglas> Has anyone run into "failed to initialize KVM: device or resource busy" during a vbox nixops deploy?
cybrian has joined #nixos
<clever> earldouglas: do you have VM extensions enabled in the bios?
akamaus has joined #nixos
<earldouglas> Yep.
Harpalus has quit [(Quit: Leaving)]
<MichaelRaskin> Sorry for possibly irrelevant remark — I think VirtualBox is generally not too friendly to KVM (for example, VirtualBox hardware acceleration uses CPU virtualisation support in a way incompatible with simultaneous use by KVM), so if your VM needs kvm-inside-VM I would check if KVM is usable inside VBox at all.
<clever> MichaelRaskin: i recently found that hyper-v doesnt work inside virtualbox (the windows version of kvm)
<akamaus> Could anyone help me with overriding gcc inside python.stdenv?
<earldouglas> Nuts. This was working a day ago.
ryantrinkle has joined #nixos
<clever> earldouglas: try doing "rmmod kvm" and see what happens
MP2E has joined #nixos
<earldouglas> That fixed it! Should I blacklist that module?
<earldouglas> Thanks a lot -- this was a major blocker.
dejanr has quit [(Quit: WeeChat 1.9.1)]
<clever> earldouglas: thats strange, because on my end i can have both loaded at once
<earldouglas> For reference, I had to rmmod kvm_intel and then rmmod kvm
<clever> it only blocks if /dev/kvm is in use
<clever> but i'm on the amd kvm module
<clever> the intel one may work differently
<clever> blacklisting kvm will probably solve your problem, and also block qemu from ever having kvm extensions
ThatDocsLady has quit [(Ping timeout: 260 seconds)]
<earldouglas> You had me at "solve your problem". Thanks again.
berce has joined #nixos
<berce> orivej: you changed the base branch of PR #32315 to staging and then back to master. Why back?
ma27 has quit [(Ping timeout: 250 seconds)]
<orivej> berce: you can try and see for yourself: the commit itself has to be rebased to. but it does not build: configureFlags should be deleted too
<orivej> rebased to staging
{^_^} has quit [(Ping timeout: 240 seconds)]
<berce> So I should do 2 things: 1. remove configure flags; 2. rebase to staging. Correct?
<orivej> yes
<berce> Thank you. I 'll do it.
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] FRidh opened pull request #32318: Python: support multiple overrides (master...python-multiple-overrides) https://git.io/vbsWw
NixOS_GitHub has left #nixos []
ma27 has joined #nixos
taktoa has quit [(Remote host closed the connection)]
reinzelmann has joined #nixos
Synthetica has quit [(Quit: Connection closed for inactivity)]
ssmike has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] dezgeg pushed 1 new commit to master: https://git.io/vbslw
<NixOS_GitHub> nixpkgs/master e220f31 Tuomas Tynkkynen: xdaliclock: Fix aarch64 build...
NixOS_GitHub has left #nixos []
bennofs has joined #nixos
jb55 has joined #nixos
reinzelmann has quit [(Quit: Leaving)]
jsgrant has joined #nixos
ilyaigpetrov has quit [(Quit: Connection closed for inactivity)]
ris has joined #nixos
<boxofrox> gchristensen: if you're still trying to install pijul 0.8.3 on nixos, I wrote up my notes for building with cargo on nixos. https://gist.github.com/boxofrox/8b5c71ec29c477108809d868f8acda3e
<gchristensen> I managed to get it! :)
jluttine has quit [(Ping timeout: 248 seconds)]
<gchristensen> thank you :D
<boxofrox> sweet
ylwghst has joined #nixos
ma27 has quit [(Ping timeout: 264 seconds)]
<LnL> gchristensen: boxofrox: let me know if I should start a new eval for carnix
digitalmentat has joined #nixos
<gchristensen> I'll need to fix up my PR a bit first
cybrian has quit [(Read error: Connection reset by peer)]
tj_ has joined #nixos
ylwghst has quit [(Ping timeout: 260 seconds)]
phdoerfler has quit [(Quit: Leaving.)]
cybrian has joined #nixos
<digitalmentat> after getting encrypted suspend-to-ram working for my NixOS system, I'm have the irritating issue of the lock screen crashing, forcing me to go to a virtual terminal to execut loginctl unlock-sessions
<digitalmentat> has anybody dealt with this before? are there any debugging tips?
ylwghst has joined #nixos
ma27 has joined #nixos
JosW has joined #nixos
adamt has joined #nixos
adamt has quit [(Changing host)]
adamt has joined #nixos
<bennofs> encrypted suspend to ram? that sounds interesting... does it encrypt ram contents as well? how do you setup something like that?
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] berce opened pull request #32322: ed: fix output directories. Fixes #32150. Supersedes #32315. (staging...staging) https://git.io/vbsgg
NixOS_GitHub has left #nixos []
<sphalerite> sprunge.us/RVCa :(
ilja_kuklic has joined #nixos
<gchristensen> ouch.
<sphalerite> Only happens with nixUnstable though
<sphalerite> It lives up to its name :D
<sphalerite> Huh now I can't reproduce it
<sphalerite> Bummer. Wanted to debug it
vcunat has joined #nixos
<Dezgeg> check coredumpctl?
<sphalerite> Nope, nothing
<Dezgeg> maybe it needed enabling separately
<sphalerite> Would it make sense to enable writing coredumps to the journal by default?
<sphalerite> Yeah it does
chaker has joined #nixos
biovoid_ has joined #nixos
<pcarrier> there's not antlr4 for python3 in nixpkgs?
<biovoid_> hi
<LnL> just enable it and rebuild, ... oh wait
<pcarrier> LnL: how would I do that?
<LnL> was replying to sphalerite
<LnL> most of the python stuff is packaged manually so it might be missing
<chaker> Hey, what is that best way to work with authenticated request in Hydra? I wrote a script to automate some workflows that I usually do by hand and used the hydra_session cookie for authentication. However, this cookie seems to be refreshed so often. Is there a way to use Basic HTTP authentication or generate some kind of API key?
<biovoid_> im trying to build Nix, but it's throwing an error about a modules.nix in /nix/store
<biovoid_> error: value is a boolean while a set was expected, at /nix/store/643hiir52327kfyd8g8dlmqihksls852-nixos-17.09.2253.559ebb7ed02/nixos/lib/modules.nix:231:62
<biovoid_> afaik i'm not supposed to mess around in /nix/store
<biovoid_> it's even write-protected, so how do i fix this? currently on the live-dvd (on a pendrive)
<LnL> what's the command you are running?
cybrian has quit [(Read error: Connection reset by peer)]
<biovoid_> nixos-install --show-trace
<biovoid_> although i've yet to see --show-trace do anything
<LnL> hmm, maybe there's an error in your configuration.nix?
<akfp> KABA_: That's still 17.10. But thanks for the tip about the xorg version.
ylwghst has quit [(Quit: leaving)]
<gchristensen> who is KABA_
ylwghst has joined #nixos
<clever> i'm guessing thats matrix renaming users again
<vcunat> gchristensen: @mounium on github, I think
<gchristensen> oh
<boxofrox> LnL: a new eval of carnix? I'm a bit new to nix, and not familiar with the process. I latched onto pmeunier's PR for carnix/rust-utils because it "Just Worked(tm)" for building a Rust project I want to use with NixOS.
<gchristensen> boxofrox: no worries, we'll get it :)
* LnL feels like https://xkcd.com/1782/
<LnL> don't really get matrix
<biovoid_> LnL: but why does it refer to a nix file (and line number) in /nix/store?
<clever> LnL: i sometimes use the irc gateway for slack
<LnL> boxofrox: the ofborg pr
<boxofrox> ah
kreetx has quit [(Ping timeout: 248 seconds)]
<sphalerite> LnL: lol
dieggsy has joined #nixos
ylwghst has quit [(Ping timeout: 260 seconds)]
<LnL> biovoid_: is that the only output you get?
<biovoid_> yes
ylwghst has joined #nixos
<LnL> strange, could you share your configuration somewhere?
<biovoid_> sure
mizu_no__ has joined #nixos
bennofs has quit [(Quit: WeeChat 1.9.1)]
<biovoid_> i based it off of someone else's config
bennofs has joined #nixos
ilja_kuklic has quit [(Remote host closed the connection)]
JosW has quit [(Quit: Konversation terminated!)]
<biovoid_> i setup another config last night, but i lost it when i repartitioned the drive it was on (i actually copied the configs to ~, but i rebooted without a second thought when fdisk warned about the kernel only using the new partitions after reboot)
dieggsy has quit [(Quit: ERC (IRC client for Emacs 27.0.50))]
dieggsy has joined #nixos
cybrian has joined #nixos
ckauhaus has quit [(Quit: Leaving.)]
griff_ has quit [(Quit: griff_)]
<sphalerite> Soooo... My Chromebook has two USB ports. One is occupied by a USB stick with the nixos root filesystem on it, the other is occupied by a hard drive which currently houses /tmp. Is there any way I could transfer one of these to a hub, without breaking some running nix builds?
ilja_kuklic has joined #nixos
MightyJoe is now known as cyraxjoe
dywedir has quit [(Ping timeout: 240 seconds)]
Lisanna has quit [(Ping timeout: 260 seconds)]
dieggsy has quit [(Ping timeout: 255 seconds)]
<pbogdan> biovoid_: try 'desktopManager.plasma5.enable = true;' instead of 'desktopManager.plasma5 = true;'
<LnL> biovoid_: don't see anything obvious, I would definitely expect a full traceback
<biovoid_> that didn't fix it, pbogdad, but thanks anyway
<pbogdan> biovoid_: same issue with 'services.avahi.publish' - https://nixos.org/nixos/options.html#services.avahi.publish
<biovoid_> pbogdad i actually fixed 'publish' from 'publishing' in the file i found
<LnL> sphalerite: not that I know of
digitalmentat has quit [(Remote host closed the connection)]
<herzmeister[m]> @Linus: the boot usb stick? i remember one boot option is to load the whole os into memory
<pbogdan> biovoid_: it should be publish = { enable = ... ; } instead of publish = true;
<biovoid_> doesnt "publish = { enable = true }" equal publish.enable = true?
<biovoid_> ah yes
<joepie91> biovoid_: yep
<joepie91> you're missing a ; though
bennofs has quit [(Quit: WeeChat 1.9.1)]
orivej has quit [(Ping timeout: 240 seconds)]
<akfp> in make -jX -lY commands, where is the Y coming from?
<biovoid_> fixing the avahi line fixed that erro
<biovoid_> error
<berce> sphalerite: If you really must do it, try suspend to ram and then move the drive with /tmp.
<sphalerite> herzmeister: yeah, except it's not the installer, it's an actual fully installed nixos on a normal filesystem
pierrec_ has joined #nixos
<sphalerite> berce: nice idea! Don't think I'll risk it, since those pesky LLVM builds take like 6h, but I could see it working
civodul has joined #nixos
<berce> Then try it tomorrow, so you know for next time :-)
<sphalerite> https://github.com/lheckemann/brightnessd little bash/C hack for controlling the screen brightness on my chromebook (just console sessions). Feedback welcome.
<sphalerite> Also if anyone happens to know some magical userspace API (possibly kms-related?) for controlling DPMS (i.e. to switch the screen off) that's some functionality I'd love to add to that
digitalmentat has joined #nixos
<sphalerite> setterm --blank has been very unreliable and I'd quite like to integrate that functionality with brightness control
<sphalerite> I've tried snooping around the Xorg source code, but not found it yet
mizu_no__ has quit [(Quit: Computer has gone to sleep.)]
<berce> sphalerite: interesting read about controlling backlight: https://wiki.archlinux.org/index.php/backlight
<sphalerite> berce: yeah I'm already controlling the backlight
mizu_no__ has joined #nixos
<sphalerite> but setting the backlight brightness to 0 and turning the LCD panel off (via DPMS) aren't the same thing
<berce> I see
dieggsy has joined #nixos
<MichaelRaskin> There are LCD panels where brightness level 0 is still a positive amount of light.
<MichaelRaskin> And there are panels where brightness level 0 is no backlight, but under bright ambient light you can still read the LCD.
<MichaelRaskin> (dpms off means no backlight and no LCD control)
adamt has quit [(Ping timeout: 255 seconds)]
<berce> The same page mentions "vbetool dpms off"
<sphalerite> vbetool doesn't work on ARM :)
mizu_no__ has quit [(Quit: Computer has gone to sleep.)]
<biovoid_> thanks
<Dezgeg> check out from the device tree if it's even possible to toggle the backlight off without disabling other critical power supplies
pierrec_ has quit [(Quit: pierrec_)]
<sphalerite> Dezgeg: chromeos does it. I've also successfully run Weston, which seemed to do it
<sphalerite> now I just want to do it without graphical baggage
<Dezgeg> ok
pierrec_ has joined #nixos
pierrec_ has quit [(Client Quit)]
hiratara has quit [(Ping timeout: 255 seconds)]
pierrec_ has joined #nixos
pierrec_ has quit [(Client Quit)]
jacob has quit [(Quit: Lost terminal)]
hiratara has joined #nixos
simukis has quit [(Ping timeout: 248 seconds)]
simukis has joined #nixos
griff_ has joined #nixos
simukis has quit [(Client Quit)]
cybrian has quit [(Read error: Connection reset by peer)]
<gchristensen> Mic92: ping: re https://github.com/grahamc/ofborg/issues/13 do you think it would work to always discard stdout for all evaluation checks?
<gchristensen> or should I just discard stdout for the package list one
<MoreTea> hm, is there a runInLinuxVM variant that allows you to specify a custom machine config? I want to have a small VM a ZFS module in it.
biovoid_ has quit [(Quit: WeeChat 1.9.1)]
asuryawanshi has quit [(Ping timeout: 258 seconds)]
jensens has joined #nixos
griff_ has quit [(Quit: griff_)]
<MoreTea> I think that the nixos test system is my best shot here
<petersjt014[m]> does anyone have an example of a working config file for the latest version running in a vm (specifically from the minimal iso?)
<petersjt014[m]> grub's giving me a lot of shit
<sphalerite> yay! Managed to get the chromebook to turn its screen off using proptest from libdrm's tests. However, it complains on the journal about a streamclk timeout and turns it back on immediately
georges-duperon has joined #nixos
rihards has quit [(Quit: rihards)]
<ylwghst> petersjt014[m]: what do you get exactly?
berce has quit [(Quit: leaving)]
orivej has joined #nixos
szicari has quit [(Ping timeout: 258 seconds)]
<Mic92> gchristensen: I guess for most outputs and on evaluation error this should be ok.
griff_ has joined #nixos
hiratara has quit [(Quit: ZNC - http://znc.in)]
<akfp> how is firmware loaded / how do I know whether it was loaded?
hiratara has joined #nixos
i-am-the-slime has quit [(Ping timeout: 260 seconds)]
ris has quit [(Read error: Connection reset by peer)]
ris has joined #nixos
akfp has quit [(Read error: Connection reset by peer)]
akfp has joined #nixos
<sphalerite> akfp: firmware is loaded by the kernel, when udev tells it to. You'll typically get messages on the journal indicating firmware loads. Try journalctl -b | grep firmware
<Dezgeg> I don't think udev is related to firmware loading anymore
<sphalerite> it's not? Never mind me then
civodul has quit [(Quit: ERC (IRC client for Emacs 25.3.1))]
Dr8128 has quit [(Ping timeout: 255 seconds)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej pushed 1 new commit to master: https://git.io/vbsQK
<NixOS_GitHub> nixpkgs/master c885b3c Orivej Desh: cataclysm-dda-git: disable parallel building
NixOS_GitHub has left #nixos []
vcunat has quit [(Quit: Leaving.)]
<tobiasBora> Hello,
ilja_kuklic has quit [(Remote host closed the connection)]
<ylwghst> Whatsup :-)
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] orivej closed pull request #32317: darwin.Libsystem: fix build on 10.13.2 (staging...macos-10.13.2) https://git.io/vbscN
NixOS_GitHub has left #nixos []
<tobiasBora> I'm trying to compile NixOs from scratch in order to learn how to bootstrap it and install it on a system where just Debian is installed.
<tobiasBora> (and I don't want to use any binary cache)
<sphalerite> not using a binary cache is a waste of time, you'll get the same results, just that they'll take longer and generate more heat
<sphalerite> bootstrapping nixos is basically installing nix and nixpkgs then doing nix-build -I nixos-config=path/to/configuration.nix '<nixpkgs/nixos>' -A config.system.build.toplevel
<tobiasBora> sphalerite: so maybe I should explain why I want to do that
<tobiasBora> I have a raspberry pi that runs under armv6.
<tobiasBora> The problem is that it's really, really, really slow to compile, and the dezek cache does not contain everything I want (and also, I don't know why, but even if it's supposed to contain what I want... it just compiles it. But I'll focus on this problem later)
chaker has quit [(Quit: WeeChat 1.9.1)]
<tobiasBora> So I would like to spawn several others nixos servers, and use them to compile the stuff on my raspberry pi.
<tobiasBora> To do so, the only idea I have is to "emulate" them with qemu-user-static.
<tobiasBora> For now, using this guide: https://wiki.debian.org/RaspberryPi/qemu-user-static
<tobiasBora> I successfully get raspbian working on my computer
jb55 has quit [(Ping timeout: 248 seconds)]
<sphalerite> clever already made a full setup for using qemu-user to build stuff for foreign arches using nix :)
<tobiasBora> so now, the game is to change this raspbian into a nixos, by compiling nixos
* tobiasBora is reading
<tobiasBora> sphalerite: how am I supposed to use it?
<clever> tobiasBora:build the qemu-user-arm target, and then run the register script under bin, to register it with binfmt-misc
<clever> tobiasBora: then the ekrnel can just run any arm binary
<clever> then use the nix attribute in that file as the nix.package in configuration.nix, and youll have a new option in nix.conf to convince nix-daemon it can actually run arm binaries
* sphalerite sent a long message: sphalerite_2017-12-04_22:57:56.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/TCXtFtprdJzTUIeAJMTWjInS>
<sphalerite> ^ config module I used to make it "just work"
<sphalerite> oh wait, that doesn't include the binfmt stuff
<sphalerite> nvm the "just work" bit :p
<clever> sphalerite: ah, nice, yeah, you have to manualy run register on bootup
<clever> that puts the register script into the default PATH
<sphalerite> it's also for armv7 rather than armv6
* tobiasBora has too many questions in the head
<sphalerite> but yeah, basically you don't need to fuss around with a chroot, you can just build everything transparently in your native nix store :D
<clever> yep
<tobiasBora> that's great !
slack1256 has joined #nixos
<tobiasBora> And you can compile anything like that ?
<clever> tobiasBora: the extra-platforms option takes a space seperated list of platforms
<clever> tobiasBora: so you can just add v6 to the list
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] joachifm closed pull request #32260: osrm-backend: 4.9.1 -> 5.14.1 (master...osrm-bump) https://git.io/vbYg4
NixOS_GitHub has left #nixos []
NixOS_GitHub has joined #nixos
<NixOS_GitHub> [nixpkgs] joachifm closed pull request #30969: osrm-backend: 4.9.1 -> 5.12 (master...osrm) https://git.io/vFtu2
NixOS_GitHub has left #nixos []
<clever> tobiasBora: some things like openssl and the kernel have trouble building for v6, but a good deal does work
<clever> openssl detects that the qemu is v7 capable, and makes v7 opcodes in the v6 build
<clever> then when you run it on a real v6, it crashes
akfp has quit [(Read error: Connection reset by peer)]
<tobiasBora> oh too bad
akfp has joined #nixos
<clever> qemu has some flags that can be used to entirely disable v7 support
<tobiasBora> there is no way to fake armv6 ?
<clever> but then you need to have a second x86 machine for v7 mode
<clever> tobiasBora: also, are you sure you need v6, which model of the pi do you have?
<tobiasBora> well armv6 can be used on armv7 no ?
<clever> yeah
<tobiasBora> clever: I've the raspberry pi 1 model B I think
<tobiasBora> (quite old)
<clever> one minute
<clever> tobiasBora: what does /proc/cpuinfo say?
<tobiasBora> clever: ARMv6-compatible processor rev 7 (v6l)
<clever> tobiasBora: what about the revision and hardware
i-am-the-slime has joined #nixos
ylwghst has quit [(Ping timeout: 260 seconds)]
<herzmeister[m]> soooo like posted earlier i'm still trying to mount a secondary encrypted harddrive at boot time, i can't get it to work. (encrypted system drive works fine). no `/etc/crypttab` file will be created whatsoever, i've tried all options. is it just me? does anyone else use it? (nixos unstable)
<clever> tobiasBora: yeah, the BCM2835 is v6 only
ilja_kuklic has joined #nixos
Mateon3 has joined #nixos
<sphalerite> herzmeister: nixos doesn't use /etc/crypttab
<herzmeister[m]> then how its supposed to be done?
i-am-the-slime has quit [(Ping timeout: 260 seconds)]
<tobiasBora> clever: Hum... So let's go step by step. For now, I'm running on my computer a debian (I may jump to NixOs later, but I want to play with it more before going to NixOs). So I guess that your solution needs a fully working nixos. What is the easiest way to proceed? Should I create a chroot that runs nixOs ?
Mateon1 has quit [(Ping timeout: 255 seconds)]
Mateon3 is now known as Mateon1
<sphalerite> herzmeister: using the boot.initrd.luks.devices options
<clever> tobiasBora: it can be used on debian as well
sigmundv__ has joined #nixos
<clever> tobiasBora: run nix-build on the qemu-user-arm target, and then run register as root
<clever> tobiasBora: then nix-build the nix target in nix-misc, and use that nix to build things for arm, with the right config manualy placed in /etc/nix/nix.conf
<herzmeister[m]> @Luks: this one works fine for my primary encrypted system drive. I have also tried to use it for the secondary drive, if that is what you mean, but then I have problems with the keyfile
<herzmeister[m]> @Linus s/Luks/Linus
<sphalerite> aah so you have your keyfile for the secondary one in the filesystem on the primary one?
<herzmeister[m]> yes
<sphalerite> hm, I'm not sure the current system handles dependencies like that
<herzmeister[m]> automatically for sure it doesn't ;)
erictapen has quit [(Ping timeout: 248 seconds)]
<samueldr> clever: the flags to disable armv7 for qemu are built-time or run-time flags?
<herzmeister[m]> i was trying to use the `filesystems.<name?>.encrypted.` options manually in the configuration file
<tobiasBora> clever: I'm not yet very familiar with Nixos. When you mean "nix-build on the qemu-user-arm target", you mean that I should run "nix-build" on the file https://github.com/cleverca22/nix-misc/blob/master/qemu-user.nix ?
<sphalerite> a somewhat hacky option, which should work if the secondary one isn't needed for boot, is just to stick in a systemd service that will open it up
<clever> samueldr: i think its runtime, but the wrapper for fixing argv[0] can insert them
<clever> tobiasBora: nix-build -A qemu-user-arm -o qemu-user
<clever> tobiasBora: that will open the default.nix file, and build the qemu-user-arm target defined within
<herzmeister[m]> as no crypttab file is being created, i guess the feature is not implemented yet in nixos? or how else would it go low-level
<herzmeister[m]> yeah i was looking at udiskie
ryantrinkle has quit [(Ping timeout: 248 seconds)]
<sphalerite> That could look something like
* sphalerite sent a long message: sphalerite_2017-12-04_23:20:09.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/fXbNabRNCnfhlPHHQVqjkklD>
<sphalerite> yeah or you could udisks fanciness. idk much about that besides that it's what the GUI stuff uses
<herzmeister[m]> but have problems with that udiskie thing too. :-/
<herzmeister[m]> thanks, that systemd option sounds reasonable
<tobiasBora> clever: hum it seems to work, or at least it downloads stuff, not sure what ^^'
<clever> tobiasBora: the real magic is when you do: nix-build '<nixpkgs>' -A hello --argstr system armv6l-linux and it just works
<tobiasBora> clever: just, why do you need toxvpn,
<tobiasBora> hum
<clever> tobiasBora: nix-misc was just a dumping ground for many things
<tobiasBora> ok I understand, sorry I'm not yet very familiar with nix expressions ;) By the way, if it's you that added toxvpn to nix, I wanted to thank you, because I used it a few weeks ago, and the easier way to install it was to use nix ;)
<clever> tobiasBora: i always write nix expressions for every piece of software i make now
<clever> its just simpler to manage all the dependencies
<tobiasBora> clever: oh, you wrote toxvpn???
<clever> yep
<tobiasBora> you're my god ;)
alunduil has joined #nixos
<tobiasBora> Ok, so the first command succeed, and now I tried the "nix-build '<nixpkgs>' -A hello --argstr system armv6l-linux" and it seems to work...
<clever> is it downloading or building?
<herzmeister[m]> ron-paul-itshappening.gif
<tobiasBora> clever: for now it's downloading lot's of stuff. But I'm wondering if it's not basic stuff, like gawk... Is it supposed to also build stuff like gawk or not?
<tobiasBora> whoo it downloads "linux-4.4.10.tar.xz"...
<clever> tobiasBora: at some point, it should begin building, and thats when the real test starts
<tobiasBora> clever: Ok nice. I just hope that my DD won't go out of space before :-°
<tobiasBora> oups
<tobiasBora> clever: I'm sorry, the news are bad: "error: a ‘armv6l-linux’ is required to build ‘/nix/store/c1p9xpbvb8f2r40pww2q0pc0mwfvvq3v-bootstrap-tools.drv’, but I am a ‘x86_64-linux’
<tobiasBora> but it tries to build building path(s) ‘/nix/store/abzyqd4j5i6za05n20h27h5l5hg1apgz-bootstrap-tools’
<clever> tobiasBora: you need the `build-extra-platforms = armv6l-linux` line in /etc/nix/nix.conf
<sphalerite> gchristensen: yay!
<clever> tobiasBora: and you need to use the patched nix from nix-build -A nix, then ./result/bin/nix-build
<sphalerite> And set the option in /etc/nix/nix.conf or on the command line
<sphalerite> oh right clever mentioned that just a minute ago
<tobiasBora> So to sum up, after I ran the "nix-build -A qemu-user-arm -o qemu-user", I edit the file /etc/nix/nix.conf, and then I run "nix-build -A nix". Finally, I use "./result/bin/nix-build '<nixpkgs>' -A hello --argstr system armv6l-linux" and I cross my fingers.
<clever> tobiasBora: also, after building qemu-user, you need to run ./qemu-user/bin/register
<tobiasBora> as root right
<sphalerite> you may also want to add -o nix when building nix so you don't clobber the result symlink with hello afterwards
<sphalerite> (and correspondingly use ./nix/bin/nix rather than ./result/bin/nix)
<tobiasBora> ok, now I don't have any error, it seems to build the bootstrap tools.
<sphalerite> this will take a long time probably :)
alunduil has quit [(Quit: leaving)]
<tobiasBora> ok
jb55 has joined #nixos
<sphalerite> there was a mass rebuild on 17.09 the other day, so my chromebook had to rebuild everything
<tobiasBora> what is a mass rebuild?
<tobiasBora> your chromebook is arm also?
<sphalerite> where a dependency near the root of the dependency tree changes and everything has to be rebuilt
<sphalerite> yes
<sphalerite> I did it with max-jobs 1 and build-cores 1 because of some race conditions that seem to occur in the build of gcc, it took about 32h in total (this is on a 1.8GHz ARMv7 CPU).
<tobiasBora> when was it? Because my raspberry pi just decided to compile everything a few days ago, and I'm wondering if it cannot be explained by that
<tobiasBora> omg
<sphalerite> qemu-user on my i7-7700HQ laptop tends to be about 4× slower than the chromebook when operating with a single core
<sphalerite> so depending on how beefy your x86_64 machine is, you may be looking at several days of building
<sphalerite> of course, if you're spawning a bunch of servers to do the building you can cut down on those times significantly
<tobiasBora> sphalerite: mine can go between 2.5Ghz to 3Ghz.
<tobiasBora> wow wait
<tobiasBora> with the i7 HQ, you are 4x slower than the 1.8GHz chromebook?
<tobiasBora> So do I have any chance to beat the raspberry pi?
<sphalerite> yes
<tobiasBora> (because if it's slower than building on the rpi, why would I use my computer ?)
cybrian has joined #nixos
<sphalerite> although that's single core on each. It has hyperthreading that gives it twice as many threads so it ends up being about 2× slower when I'm building fully parallel
<sphalerite> you can use both at the same time :)
<clever> tobiasBora: more ram, and more cores, and you are likely to have many x86 machines
<clever> tobiasBora: so you can just parallelize the job over every x86 machine, and every core
<sphalerite> (the chromebook is quad-core too)
<tobiasBora> clever: the parallelization can be done efficiently without any user intervention ?
dan_b has joined #nixos
<clever> tobiasBora: just configure nix build slaves and bump up the -j count
pxc2 has quit [(Ping timeout: 276 seconds)]
<sphalerite> yeah the biggest problem you'll have probably is big derivations that take long to build individually and so can't be distributed across multiple machines (LLVM, linux kernel, …)
davidak has quit [(Read error: Connection reset by peer)]
davidak has joined #nixos
<sphalerite> I've been planning to get a bunch of desktops booting from the network and acting as build slaves
<sphalerite> clever: would you happen to have any suggestions for a storage arrangement with those? It's 8-10 machines (not sure how many of them are working) that would all be connected to a switch along with the server that's serving their netboot image
georges-duperon has quit [(Ping timeout: 260 seconds)]