<Church->
Hey question, with services like nginx and mariadb, do they only search in the paths localized to their nix store path? Or do they fall back to /etc/ too?
drakonis_ has quit [Ping timeout: 252 seconds]
jasongrossman has joined #nixos
aleph- has joined #nixos
strobelight has quit [Quit: gtg]
drakonis_ has joined #nixos
<aleph->
Christ, I want to slit vlc's imaginary throat.
<aleph->
Sigh
<{^_^}>
[nixpkgs] @benpye opened pull request #49320 → miniflux: init at 2.0.12 → https://git.io/fxHoA
<bpye>
Woo, first PR :)
<aleph->
gg
<bpye>
What's wrong with VLC?
<slack1256>
Some ...forums on the internet have an irrational hate for it.
<aleph->
bpye: Chromecast rendering support doesn't work.
<aleph->
And it's driving me batty, since it fails for chromium too, but for different reasons I suspect
<aleph->
Since gnomecast works with firewall disabled.
<bpye>
Oh I see, isn't it's Chromecast support reverse engineered anyway?
<bpye>
My point being I think it might be Google's fault
lassulus_ has joined #nixos
drakonis_ has quit [Ping timeout: 276 seconds]
<aleph->
bpye: Works on other distros
drakonis_ has joined #nixos
<aleph->
Only failing on Nix currently.4
lassulus has quit [Ping timeout: 246 seconds]
lassulus_ is now known as lassulus
<colemickens>
Actually Chromecast broke in Google Chrome for me as well.
<colemickens>
It had been working fine in `google-chrome-*` but never in Chromium (I assume it has something to do with the Chromecast keys, but I thought they were only needed in the receiver side). Anyway, it doesn't work in VLC in NixOS for me either. Looks like it's working in Windows+VLC though
jasongrossman has quit [Ping timeout: 252 seconds]
ryantrinkle has quit [Ping timeout: 240 seconds]
<colemickens>
If I disable the NixOS firewall then Chrome works, Chromium/VLC still don't though.
<colemickens>
Looks like VLC is built with libupnp, which I think is what Chromecast uses for discovery.
<colemickens>
or maybe it uses mdns, I can't remember.
drakonis1 has joined #nixos
drakonis2 has joined #nixos
drakonis_ has quit [Ping timeout: 250 seconds]
jasongrossman has joined #nixos
drakonis1 has quit [Ping timeout: 252 seconds]
dramforever has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
jasongrossman has quit [Ping timeout: 246 seconds]
orivej has quit [Ping timeout: 272 seconds]
dramforever has quit [Ping timeout: 256 seconds]
drakonis has quit [Quit: WeeChat 2.2]
sveitser has joined #nixos
drakonis has joined #nixos
<drakonis>
hm, i can't seem to be able to change my kde default settings
<clever>
dramforever: ive also memorized how it works, and never knew it was in the docs, lol
<jophish>
What 'fetch*' command should I use to unpack a tarball?
<clever>
jophish: does it have to be unpacked in the store, or just at build time?
<jophish>
clever: I'd like for it to be unpacked in the store
<clever>
fetchzip can do that then
<jophish>
I believe that if I put a tarball in the 'src' attribute it's unpacked
<jophish>
oh, grat
<jophish>
great*
<jophish>
thanks again clever
<clever>
its a bit poorly named, fetchzip is just a modified form of fetchurl, that has added zip support, and will auto-unpack
<clever>
the reason it auto-unpacks, is because a lot of libraries for generating a zip include wildly varying timestamps, and then the hash of your zip keeps changing, so nix cant work
<clever>
so it instead unpacks, and operates on the hash of the files within
<dramforever>
This seems still undocumented: Is store-paths.xz the closure of a channel's packages?
<jophish>
clever: are you sure fetchzip works with gzipped tarballs?
<dramforever>
Or just the packages themselves...?
<jophish>
I've prefetched the file with nix-prefetch-url --unpack
<ekleog>
was there some kind of trick to run nix-channel --update in a VM? I've checked my store does have an overlayfs and I have virtualization.writableStore = true; :/
<clever>
ekleog: and does nix-channel --list show it as having a channel?
<clever>
ekleog: which user are you on?
<ekleog>
clever: I'm running `sudo nix-channel --update` ; `sudo nix-channel --list` returns nixos <-> channel:nixos-18.09, like my host
<clever>
and does --update fail or suceed?
jtojnar has joined #nixos
<ekleog>
fail with “getting attributes of path /nix/store/[...]-unpack-channel.sh: No such file or directory” (which I can confirm doesn't exist)
<clever>
does it exist on the host?
<ekleog>
neither
<clever>
*looks*
<ekleog>
I'm having the VM with a nix.package = pkgs.nix.overrideAttrs (a: { patches = [ ... ]; }); maybe that has an impact? it really shouldn't make a file disappear, though
<ekleog>
(testing a patch to the nix-daemon is my reason to spawn said VM)
<clever>
ekleog: is this on master?
<ekleog>
(can continue in PM if you want details about the patch)
<clever>
nix should build it when you refer to <nix/unpack-channel.nix>, which nix-channel does internally
<clever>
can you test builtins.toFile in `nix repl` ?
<ekleog>
on the VM: “cannot link '/nix/store/.tmp-link-1413-[...]' to '/nix/store/.links/[...]': Operation not permitted”
<clever>
ekleog: is auto-optimize on?
<ekleog>
(same when I run it as root)
max3raza has quit [Ping timeout: 244 seconds]
<ekleog>
yeah, I just imported the configuration from my laptop and added virtualisation.{memorySize,diskSize,writableStore} and nix.package options
bollu has quit [Quit: Ping timeout (120 seconds)]
jcrben has quit [Ping timeout: 244 seconds]
<clever>
ekleog: try turning off auto-optimize and seeing if that helps
jcrben has joined #nixos
bollu has joined #nixos
jw358 has quit [Ping timeout: 244 seconds]
johnnyfive has quit [Ping timeout: 244 seconds]
Enzime has quit [Ping timeout: 244 seconds]
leothrix has quit [Ping timeout: 244 seconds]
Cheery has quit [Ping timeout: 244 seconds]
lejonet has quit [Ping timeout: 244 seconds]
primeos has quit [Ping timeout: 244 seconds]
tylerjl has joined #nixos
julm has quit [Ping timeout: 244 seconds]
silver_hook has quit [Ping timeout: 244 seconds]
jeaye has quit [Ping timeout: 244 seconds]
troydm has quit [Ping timeout: 272 seconds]
<ekleog>
oooh yeah maybe hardlinks are ill-supported by overlayfs
Cheery has joined #nixos
primeos has joined #nixos
julm has joined #nixos
troydm has joined #nixos
redj has quit [Read error: Connection reset by peer]
vidbina has joined #nixos
<ekleog>
clever: so with auto-optimize off it manages to `builtins.toFile`, but now nix-channel --update fails with not finding /nix/store/[...]-nixos-18.09.[...].[...].drv
<ekleog>
which, indeed, exists neither on the host nor on the guest
max3raza has joined #nixos
<clever>
not sure on what one, anything obvious if you add -vvvv ?
silver_hook has joined #nixos
<ekleog>
knowing there is a warning: “nix search path entry /nix/var/nix/profiles/per-user/root/channels does not exist, ignoring”, which I assume is normal given it's exactly what I'm trying to make it download
<ekleog>
(looking into -vvvv)
<jophish>
Hmm, I'm trying to boot nixos on a de10-nano fpga (it has an arm core)
<dramforever>
'This seems still undocumented: Is store-paths.xz the closure of a channel's packages?'
<dramforever>
Answer: No.
<jophish>
I'm not quite sure about how to debug this, is there some kind of super verbose boot I can do?
lejonet has joined #nixos
<ekleog>
clever: so the nixos-18.09.[...].[...] part of the .drv path failing appears to be the same as in https://releases.nixos.org/nixos/18.09/nixos-18.09-[...].[...]/binary-cache-url ; apart from that I don't see anything looking like an error :/
Mateon1 has quit [Remote host closed the connection]
<ekleog>
dramforever: if you think this should be documented, can you open an issue to nixpkgs with the information? (or even better, a PR :))
Mateon1 has joined #nixos
<clever>
jophish: i dont see any obvious signs its actually running stage1 or stage2, yet udev somehow loaded
<dramforever>
ekleog: Maybe. But apparently I just tasked myself to set up a cache.nixos.org mirror. There's way more to document than just that.
<clever>
jophish: my only thoughts are to add a -v to the #! line, and maybe set -x, then boot that new image
<ekleog>
dramforever: oh if you can document the whole process it'd be really great indeed :)
<jophish>
oh right, this is before stage1
<dramforever>
And it would be great if I could avoid using Nix itself
<dramforever>
Because I don't maintain the servers
<clever>
jophish: line 67 of the above file prints the stage1 header
grenade has joined #nixos
<clever>
jophish: everything prior to that is just setting env vars and testing for files in the initrd, nothing that should hang
Izorkin has quit [Ping timeout: 240 seconds]
<dramforever>
So using Nix would probably mean a docker image :(
<bpye>
Do I need to do anything to get a new package PR reviewed or should I just wait until someone looks over it?
<rawtaz>
thats a good question. i dont know what the general expectationsa re
<rawtaz>
are*
<srhb>
Normally putting the PR up is just fine. if it gets no attention, try finding someone who's a good candidate for reviewing, or ping this channel. :)
lord| has quit [Quit: WeeChat 2.2]
lord| has joined #nixos
<rawtaz>
how long do you think it makes sense to wait before pinging srhb ?
pie_ has quit [Ping timeout: 252 seconds]
johnw has quit [Read error: Connection reset by peer]
<{^_^}>
0. Sorry. 1. If you complain about PR being in limbo, always provide a number («I have an unmerged PR (#23924)…»), bot will print PR title and it might bring you an impulse-merge. 2. Please look up who commits similar work and ping these people (or request-review), keeping up with all the PRs is not really possible.
<rawtaz>
s/complain/notify/
jtojnar has joined #nixos
Thra11 has joined #nixos
<{^_^}>
[nixpkgs] @c0bw3b opened pull request #49328 → [18.09] installation-device: set GC initial heap size to 1MB → https://git.io/fxHQa
<dramforever>
I'm (still) trying to mirror cache.nixos.org. How can I get a list of 'most useful' store paths?
<r5d>
lassulus: yes
<dramforever>
One way is to go into the channels and use store-paths.xz, but (sadly) that does not contain the closure, so it requires a traversal
<r5d>
need to set print-mode to text
<jophish>
clever: would you be able to help me a little bit with building not-os (for armvl on the aarch64 machine). I'm trying to get a rootfs image and a kernel and command line (or extlinux config)
<tilpner>
dramforever - A mirror could easily contain all of store-paths, it's not that muchd
<tilpner>
dramforever - If you want to mirror only what you need, how about printing your system closure and mirroring just that?
<dramforever>
All store paths or all of store-paths.xz?
<dramforever>
We have quite a bit of disk space and bandwidth. That's not a big problem
<dramforever>
I'm trying to build a general mirror. Intention is that anyone close to us and use nix can use it as a replacement for cache.nixos.org
o1lo01ol1o has quit [Ping timeout: 250 seconds]
<dramforever>
tilpner: Is all of store-paths.xz a closure? I checked nixos-18.09-small and it seems that some paths has references not listed in nixos-18.09-small/store-path.xz
<tilpner>
I don't know how it is generated
<tilpner>
Although if you want it to be a general-purpose mirror, you might want to mirror a lot more than just one version of nixpkgs artifacts, and that could quickly become TBs
<jophish>
clever: ah, dist_test looks like what I want
<clever>
jophish: yep
<jophish>
I'm racing building not-os and nixos, I'll put on whatever's done first
<clever>
jophish: i think thats a native arm build, not a cross
<clever>
i did have my hydra building that on an rpi, but it hasnt been booting reliably lately
periklis has quit [Ping timeout: 276 seconds]
<jophish>
clever: are you sure, because at the moment it's 'waiting on locks or build slots' while the nixos build (armv7l) is proceeding
<clever>
jophish: that can happen if you hit the max-jobs on nix.conf, or are depending on something another nix cmd was already building
<jophish>
yeah, I think it's the latter
<jophish>
they both depend on some armv7l package
justbeingglad has joined #nixos
<jophish>
at the moment the nixos builder is on armv7 perl
<clever>
sadly, not-os depends on perl at runtime
<dramforever>
tilpner: TBs doesn't sound *too* bad
<dramforever>
I feel like we could use a cache eviction strategy to get rid of too old .nar.xz files
<dramforever>
If anyone wants them just go back to the definitive cache.nixos.org
mayhewluke has quit [Ping timeout: 246 seconds]
<jophish>
clever: oh, why's that?
<dramforever>
How fast does the binary cache grow then?
<dramforever>
I feel like someone behind the s3 buckets know this...
<nh2>
Are there written-down guidelines on what I should do as a new commiter before merging PRs, beyond testing them? E.g. I remember that I was asked to --signoff commits before, probably I should check for that? Also, which merging strategy should be used?
mayhewluke has joined #nixos
<clever>
jophish: for setup-etc.pl and managing /etc/
adamantium has quit [Ping timeout: 240 seconds]
<jophish>
ah
<jophish>
a pain to rewrite
<{^_^}>
[nixpkgs] @manveru opened pull request #49334 → buildRubyGem: Provide manpages for Ruby gems → https://git.io/fxH55
<clever>
jophish: yeah
<clever>
jophish: i could make /etc/ a static symlink to a storepath, but then its a bit more tricky to build up /etc, and every /etc file is world readable
<clever>
jophish: openssh refuses to use a world-readable private key
<clever>
and resolv.conf becomes static as well, so your dns cant be based on dhcp
<infinisil>
nh2: most just use the standard merge (creates a merge commit)
<{^_^}>
[nixpkgs] @bjpbakker opened pull request #49335 → [18.09] xorg.xf86videovmware: fix for GFX support → https://git.io/fxH5p
<infinisil>
nh2: wait for ofborg to finish the checks, make sure the branch is correct (staging with lots of rebuilds), potentially check some dependents to still work (e.g. For a pr with a library update i sometimes check if the packages that use this library still compile and run)
<srk>
dramforever: it depends :) on the commit activity and also how much stuff you build on your hydra - make sure you track disk usage in some manner
<dramforever>
Oh not my hydra
<dramforever>
I'm literally trying to mirror cache.nixos.org
<dramforever>
Fastly is a bit... 'slowly' here in China, I have to say
<dramforever>
Or did you mean hydra.nixos.org? I assumed not
bohadi has joined #nixos
<srk>
cache contains quite a lot of stuff from older releases as well, squid might work to cache the ones you hit
<__monty__>
It's a blanket statement from those maintainers. Not a blanket statement about maintainers.
<dramforever>
Yes, I meant 'the server maintainers'
<dramforever>
__monty__: That's also what I thought you meant
<leotaku>
I just noticed that pacman, the arch package manager, was in the nixpkgs tree. What would one possibly do with that package?
<__monty__>
leotaku: It's probably similar to why nix is in the arch repositories.
silver has joined #nixos
<leotaku>
__monty__: But nix can/is intended to be used on any linux system.
<__monty__>
And why wouldn't you be able to use pacman on another linux?
<leotaku>
__monty__: You can surely use packman on other distros, but I don't think that is its intended purpouse.
<leotaku>
also, how would I use packman on nixos?
Izorkin has joined #nixos
<__monty__>
Not sure you can sensibly do so concurrently with nix. Maybe to migrate from nixos to arch? Or using the nixos pacman to repair an arch system?
o1lo01ol1o has joined #nixos
<leotaku>
__monty__: Maybe somehow with the "FHSUserEnv" stuff? But I think thats read-only...
<dramforever>
Maybe you can make a usable arch chroot with it
<__monty__>
I'm not sure pacman needs an FHS if you provide the --root.
ThatDocsLady has joined #nixos
bohadi has quit [Quit: WeeChat 2.2]
o1lo01ol1o has quit [Read error: Connection timed out]
justbeingglad has quit [Ping timeout: 272 seconds]
<jD91mZM2>
How do you find the $OUT_DIR after a rust compilation?
<feep_>
I have a folder with a makefile based project in
<feep_>
how do I make binaries happen.
feep_ is now known as feep
<feep>
is I guess the core issue.
<clever>
feep: create a file called default.nix, with the following contents: with import <nixpkgs> {}; stdenv.mkDerivation { name = "name"; src = ./.; }
<clever>
then just run nix-build
<feep>
ah I didn't have the with import
<feep>
I was copying something that said to just do { stdenv }:
<clever>
if the file starts with { stdenv }: then it must be loaded with callPackage, from another file
<feep>
ah
<feep>
also nix build just gave a permission error, which obfuscated the fact that I wanted nix-build
<clever>
with import <nixpkgs> {}; callPackage ./otherfile.nix {}
<jophish>
Is there a way to get a list of all paths which contributed to a build?
<feep>
strace, lol
<jophish>
(assuming I have a derivation)
<clever>
feep: what was the error?
<feep>
error: getting status of /nix/var/nix/profiles/per-user/root/channels/nixos: Permission denied
<clever>
feep: your ~/.nix-profile isnt setup right
<clever>
or wait no
<clever>
thats the channel
<clever>
feep: what does ls -lh say about that path?
<feep>
it's a symlink to /nix/var/nix/profiles/default
<clever>
feep: ls -lh /nix/var/nix/profiles/per-user/root/channels/nixos:
<nikivi>
I assume it's something rust related due to `Compiling rustc_data_structures v0.0.0 (file:///private/var/folders/5l/5s7qw5ld1xxglm0094_hyh_c0000gn/T/nix-build-rustc-1.30.0.drv-0/rustc-1.30.0-src/src/librustc_data_structures`
<feep>
and I may have totally misread it
<nikivi>
but can't see what actual package is being build
<clever>
nikivi: `ls -ltrh /tmp/`, the names of all packages building are in ah you already found that dir
<feep>
apparently it meant *userid* 1000, lol.
<clever>
nikivi: and then compare the procs using cpu in `top` or `htop` with the owners of the nix-build-* dirs
<feep>
"ownership" is not an unambiguous word here.
<clever>
feep: oooo!, that makes a lot more sense!
<feep>
clever: I'll make a pr to change it to "bad user id"
<nikivi>
it just has `com.apple.launchd.geNRz5a3H7` inside /tmp
<clever>
nikivi: apple has to be different :P
<clever>
nikivi: its somewhere in /private/var/folders/5l/5s7qw5ld1xxglm0094_hyh_c0000gn/T/
<clever>
dont ask me why the dir has to be so complex :P
<feep>
clever: which reminds me, did you guys rebase master at some point? my old repo was not in line with the remote one...
<feep>
might just have been cause I'd fetched shallow though
<clever>
feep: which repo?
<feep>
nixpkgs
<clever>
nixpkgs master is never rebased, would break too much
<feep>
okay, so probably my fault then
<feep>
anyway, please review and accept (the github version of "please like, share and subscribe" :D)
<jophish>
clever: what would be a good wantedby for this timer service?
<clever>
2018-10-28 04:15:27 < clever> Arahael: you can define a custom timer by just setting systemd.services.foo = { script = ''....''; startAt = "nightly"; }; in configuration.nix i believe
<clever>
jophish: i think that handles the wantedby for you
<clever>
jophish: you will obviously want a shorter interval, and maybe test it on a working machine first, systemctl list-timers
<{^_^}>
[nixpkgs] @eadwu opened pull request #49339 → pythonPackages.zstd: init at 1.3.5.1 → https://git.io/fxHAk
iyzsong-x has joined #nixos
freeman42x has joined #nixos
<{^_^}>
[nixpkgs] @eadwu opened pull request #49340 → pythonPackages.distribute: init at 0.7.3 → https://git.io/fxHAl
<feep>
clever: yay!
<jophish>
oh, that's handy
iyzsong has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @eadwu opened pull request #49341 → pythonPackages.almost: init at 0.1.5 → https://git.io/fxHA0
<{^_^}>
[nixpkgs] @eadwu opened pull request #49342 → pythonPackages.trueskill: init at 0.4.5 → https://git.io/fxHAu
<hyper_ch>
so, how to add a new package to nixos or rather nixpkgs?
boxscapeR has joined #nixos
<boxscapeR>
does anyone have example config of home-manager + zsh? I'm not quite sure how to use the zsh submodules
<feep>
how do I make nix build log the build to stdout?
<clever>
feep: i just use `nix-build` when i want logs
<clever>
feep: there is also `nix-store -l /nix/store/foo` to view the logs of something that has previously passed
<feep>
oh, nix-build != nix build?
<feep>
ahh
<feep>
yes that is much better
<clever>
yeah, `nix build` is the newer interface
<clever>
but it needs some more work
<feep>
hm, nix-build doesn't do a parallel build
<feep>
despite having maxJobs set in configuration.nix
<feep>
oh, apparently I need enableParallelBuilding
<clever>
there is both make and nix level parallism
<clever>
enableParallelBuilding is for the make level
<feep>
yes
<clever>
max-jobs is at the nix level
<feep>
that is the one I want
<feep>
ah
boxscapeR has quit [Remote host closed the connection]
<clever>
cores is also at the make level
boxscapeR has joined #nixos
<clever>
so if enableParallelBuilding is on, make will obey cores from nix.conf
<boxscapeR>
oops, lost internet. Actually I think I figured out how to use the submodules though
<clever>
and then nix itself will build max-jobs seperate packages, in parallel
<feep>
yep, that's the one I needed!
<feep>
hm
<feep>
what do I do with buildscripts that #!/bin/bash?
<clever>
feep: run patchShebangs on the dir they are in
<feep>
I do not have commit rights on this repo and I don't want to keep a local branch
<clever>
i'm going to get some sleep, send me a PM if you get it all working, i'm interested in fpga stuff as well
pointfourone has quit [Remote host closed the connection]
<jophish>
Thanks clever for all the help!
<jophish>
I had a working system on an arria 10 board
<jophish>
but it's a bit outside hobbyist scope :)
<srid>
anyone here actually develop using nix on *chromeos*?
<Dezgeg>
is it those fpga boards?
<jophish>
Dezgeg: hmm?
<simpson>
srid: Sadly, Chromebooks and similar are not general-purpose computers. The avenue I'd probably explore is Nix inside Crouton: https://github.com/dnschneid/crouton
<srid>
simpson: doesn't chromeos v70 come with ability to run arbitrary linux apps (isolated inside a debian container)? i was wondering if anybody run run like that. the upcoming Google Pixel Slate is tempting as a portable nix dev environment.
<thefloweringash>
I use nixos on a chromebook
<hyper_ch>
thefloweringash: with zfs?
<srid>
s/run run/run nix/
<thefloweringash>
hyper_ch (IRC): f2fs, did I miss zfs in the context?
<srid>
thefloweringash: ah, so you wiped chromeos and installed nixos on it. i actually would love to retain chromeos mainly for a company-backed hardware support (bluetooth, etc.)
<srid>
... bluetooth, pen input, trackpad etc..
<thefloweringash>
I dual boot, nixos on microSD, chromeos on the internal storage
<srid>
one day I hope to see something like `nix-darwin` but for chromeos (`nix-chromeos`?).
<srid>
thefloweringash: which chromebook do you have?
<simpson>
srid: Oh, I guess I haven't caught up. Interesting, thanks.
mayhewluke has quit [Ping timeout: 240 seconds]
<thefloweringash>
A Saumsung Chromebook Plus (v1, not the v2), aka "Kevin"
<srid>
simpson: yea, chromeos today can run android and linux apps :)
<hyper_ch>
thefloweringash: so you don't run zfs then?
mayhewluke has joined #nixos
<srid>
there is hope for 'linux on desktop' after all.
<thefloweringash>
hyper_ch (IRC): nope, f2fs
<hyper_ch>
thefloweringash: thx
<srid>
thefloweringash: does pen input work on nixos?
<simpson>
srid: Oh! That feature isn't general-purpose. I'm imagining the effort required to build a custom ChromiumOS image. And the hardware itself is not exactly suitable for development work.
<thefloweringash>
yep. Not sure which software is best to use with it, but Xournal can definitely see the pen as some kind of xinput device
<Neo-->
does adding -I nixpkgs=/path to nix-shell work just for regular packages or should it effect things like stdenv as well? specifically, I'm toying around with allowUnfree but even setting it to true in check-meta doesn't seem to work for nix-shell
<thefloweringash>
aaand I just realised I horribly misread the context, even though it was carefully enclosed in stars. sorry!
copumpkin has joined #nixos
<thefloweringash>
I definitely sympathise with wanting to keep chromeos around, it does have a decent desktop experience. I've never had much luck with the linux vm though. Every time I try it my chromebook freezes and reboots shortly after. Maybe too much memory pressure?
<boxscapeR>
With something like nix-prefetch-git, why does it output the data in what looks like JSON rather than something you could use in a .nix file
<boxscapeR>
?
<jophish>
clever: sysreq ? doesn't do anything after the hang
<jophish>
but it does before
<srhb>
boxscapeR: You can use it in a nix file easily with builtins.fromJSON
<Alphare>
Hi all! I'm thinking of checking out NixOS. I have a new SSD, and I was wondering if there was a way of installing NixOS directly to that SSD from my current Linux install?
<{^_^}>
[nixpkgs] @bendlas pushed to master « Revert "chromium: make gcc8 build available via buildWithGcc flag" »: https://git.io/fxQJT
<Alphare>
well, first install, I home I don't brick my other OSes
<aanderse>
aszlig: thanks for your post on discourse. not sure what i'm doing wrong, but switching to the new configuration doesnt pickup fileSystem changes... test script: https://pastebin.com/wEvR4q5i and relevant output: https://pastebin.com/uxsatuy5 note how no mounts listed under /. any ideas?
<jophish>
I like how similar u-boot and linux are though
<jophish>
not sure what it is at the moment, but I think i should be using socfpga_defconfig
mystfox has quit [Remote host closed the connection]
<nh2>
aszlig: do you have a recommendation on how to best make your patch compatible with 17.09? My Hetzner disconnect repro is on that
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos
agander_ has joined #nixos
Alphare has quit [Remote host closed the connection]
drakonis has joined #nixos
<jophish>
well, setting that config creates some drivers, however the kernel fails to boot!
<nh2>
in 17.09, we have like `ipv4.addresses = [ { address = ...; prefixLength = ...;} ]` and no `routes` at all
agander has quit [Ping timeout: 246 seconds]
agander_ is now known as agander
<nh2>
or no wait, I think it's `networking.interfaces.<interface>.ip4 = [ { address = ...; prefixLength = ...;} ]`
endformationage has joined #nixos
<{^_^}>
[nixpkgs] @prusnak opened pull request #49351 → python.pkgs.Pyro4: new package (at version 4.74) → https://git.io/fxQkE
erictapen has quit [Ping timeout: 240 seconds]
<hodapp>
wow, odd... had mounted a CIFS share which had an ISO image, and had mounted that ISO image too, was copying files off the mounted ISO, hit ^C and it just stalled
<hodapp>
couldn't kill it, not even kill -9; dmesg was complaining regularly that cp was hung for over 120 seconds
IRCsum has joined #nixos
IRCsum has quit [Max SendQ exceeded]
IRCsum has joined #nixos
aleph- has joined #nixos
alhariel has joined #nixos
IRCsum has quit [Max SendQ exceeded]
<aleph->
Afternoon folks
IRCsum has joined #nixos
<aleph->
Question when writing to a file via environement.etc."FILE_NAME".text = '' '';, can I embed string interpolation inside the text I write to file? Ie. have "${pkgs.PKG_NAME}/PATH/IN/PKG/"
<sphalerit>
aleph-: yes
<{^_^}>
[nixpkgs] @talyz opened pull request #49352 → nomachine-client: init at 6.3.6_1 → https://git.io/fxQIB
agander has quit [Read error: Connection reset by peer]
<aleph->
Cool, thanks sphalerit
agander has joined #nixos
<aszlig>
aanderse: ah, fileSystems are a bit special, because they have VM overrides (priority 10), so you need to mkOverride with a priority below that
<nh2>
second: the `('ipv6', 'addresses')` stuff didn't expand to real json keys for me, should that work there? (I just rewrote it explicitly locally)
<aszlig>
oh right... defaultGateway6 doesn't have a prefixLength
<aszlig>
hm, maybe /PL works there
erictapen has joined #nixos
<aszlig>
ah, never mind... it actually wouldn't make sense either way
<aszlig>
and yeah, the ('ipv6', 'addresses') also need to be made nested dicts... at the time i was writing this i thought it would be py2nix()ed...
<aszlig>
which it is, but it's first serialized to json
<nh2>
ok that's all clear then
<nh2>
aszlig: I currently get a `RTNETLINK answers: Invalid argument` with the added v6 command: `ip -6 route replace default via "fe80::1" proto static`
<aanderse>
aszlig: thanks, will try :)
<aszlig>
aanderse: for example: fileSystems = lib.mkOverride 5 { ... your fs ... };
<aszlig>
oh, no...
<aszlig>
that's wrong
<aszlig>
fileSystems = lib.mkVMOverride { ... }
<aszlig>
you need to have the same prio because you probably want them to be merged
<aanderse>
oh, ok
<aanderse>
so just mkVMOverride without a priority because i'm assuming it matches?
<aszlig>
nh2: hm... may i boot up properly fist? :-D
<aszlig>
too much action in the morning =)
<nh2>
aszlig: sure :D
<aanderse>
ah, examples in tests/
<aanderse>
cool
<nh2>
aszlig: if I add `dev eth0` it works
<aleph->
Questions writing this config to file via environment.etc."/nginx/sites-available/lobste.rs/".text and I keep getting an error that use_file_cache doesn't exist when setting it on line 88, doesn't make sense, can only assume it thinks it might be a shell variable? Source I'm trying to write to file is here:
<tilpner>
aleph- - You should probably use services.nginx.* instead of forcing the existence of that file
<tilpner>
(Unless you don't have services.nginx enabled, and start it manually. That might work)
<{^_^}>
[nixpkgs] @tazjin opened pull request #49353 → nixdoc: init at 1.0.1 → https://git.io/fxQtq
<tilpner>
A lot of it can be copied into services.nginx.virtualHosts.<name>.extraConfig, and then you can still have the module handle letsencrypt for you
<aleph->
tilpner: Oh really? I think some of it I'm going to have write to /etc/, unless there's an option for setting mime.types or other misc files I've missed.
<tazjin>
Can a trusted user ask the bot to build `nixdoc` on #49353? Thanks!
<aleph->
Second being, how best to structure the rest of the server blocks? Should I stick those in services.nginx.config? Should I set them in Virtualhosts.extraconfig?
<{^_^}>
[nixops] @nh2 opened pull request #1032 → hetzner: Don't use localCommands, set interface device of default gateway → https://git.io/fxQmm
<nh2>
aszlig: OK here you go, this is the cleaned up PR that seems to work for me: ^
<{^_^}>
[nixpkgs] @globin pushed commit from @fadenb to master « graylogPlugins: update »: https://git.io/fxQmu
chrisaw has quit []
chrisaw has joined #nixos
<aanderse>
anyone else find that the fancy log.html files that nixos tests produce don't work well with firefox? the javascript doesn't work on my machine...
drakonis_ has joined #nixos
<nh2>
aanderse: link I could try to confirm?
<aanderse>
nh2: run any test on your own machine then open with firefox
<tilpner>
aleph- - Sorry, can't assist right now. Perhaps tomorrow
drakonis1 has quit [Ping timeout: 250 seconds]
IRCsum has quit [Remote host closed the connection]
<hyper_ch>
anyone has packaged perl stuff for nixos and who is in here?
<{^_^}>
[nixpkgs] @globin pushed commit from @r-ryantm to master « python36Packages.mysql-connector: 8.0.12 -> 8.0.13 »: https://git.io/fxQYk
goibhniu has joined #nixos
<{^_^}>
[nixpkgs] @aanderse opened pull request #49354 → filesystems: escape spaces in fstab with \040 → https://git.io/fxQY4
erictapen has quit [Ping timeout: 244 seconds]
<Neo-->
any ideas why wouldn't nix-shell want to read my .config/nixpkgs/config.nix file? I have put allowUnfree into it but starting up nix-shell doesn't work (except with env variable NIXPKGS_ALLOW_UNFREE)
<haitlah>
I'm back again :P. Having successfully made a full nix CI that will replace our current travis one
<haitlah>
But I am facing troubles with stack2nix
<haitlah>
When running locally, everything is fine, but in my alpine+nix CI runner, I get a: <stdout>: commitBuffer: invalid argument (invalid character)
<Alphare>
I'm having an issue with installing packages. Every time I try, I get the following error: "error: opening lock file '/nix/var/nix/profiles/per-user/root/profile.lock': Permission denied"
<Alphare>
I think fish is not resolving my user profile correctly
<Alphare>
is there a file somewhere where I need to specify which profile to use?
<{^_^}>
[nixpkgs] @Mic92 opened pull request #49358 → nixdoc: init at 1.0.1 → https://git.io/fxQGq
<andi->
gobject introsprection, need it more specific?
crmlt has quit [Remote host closed the connection]
<gchristensen>
python36Packages.pygobject3 looks like the one! thank you!
<nh2>
aszlig: I thought about backwards compat first as well but we probably don't need to care since 17.09 already hasn't got security updates for half a year so it's really unsafe to use on Hetzner anyway
<{^_^}>
[nixpkgs] @Infinisil pushed 3 commits to release-18.09: https://git.io/fxQZP
crmlt has joined #nixos
<nh2>
aszlig: I made another comment, removing the extra route certainly helps, but I'm not sure that then your case of multiple IPs is covered
<elvishjerricco>
Can I run `hackage2nix` myself so that I can make a change to `configuration-hackage2nix.yaml`, update `hackage-packages.nix` myself, and use that change in `configuration-ghc-8.0.x.nix` all in a single PR?
<aszlig>
nh2: yeah, IIRC ikwildrpepper has mentioned at some point in the past that we aim to support the last two recent stable releases, so in this case 18.09 and 18.03, but not sure whether this is the official policy (do we even have one?)
<elvishjerricco>
Or is the automated `haskell-updates` branch the only thing that's allowed to touch `hackage-packages.nix` in all circumstances?
<aszlig>
nh2: well, i do have a few server with additional subnets, but they're in production so i can't test that
crmlt has quit [Remote host closed the connection]
<aszlig>
nh2: but it doesn't mean that it won't be the case someday
<nh2>
aszlig: right, and you can already now ask them to add 2 NICs when ordering. So I'm thinking the `iface_attrs[iface] = {` should be in `if iface != defgw_dev:`
<aszlig>
nh2: yeah, that probably makes sense
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<LnL>
well yes, if you have arbitrary types they can't be merged
<infinisil>
Can confirm, use attrsOf over attrs wherever possible
<infinisil>
There's been a couple of "bugs" because of options being attrs instead of attrsOf
jackdk has joined #nixos
<LnL>
yeah, maybe they should be the same but it's probably not ok to change that now
<makefu>
steveeJ: the more precise you are with your types, the better they will merge. make modules of the type submodule and implement the options you need to get most out it
<{^_^}>
[nixpkgs] @Mic92 pushed 16 commits to release-18.09: https://git.io/fxQWG
MoreTea has quit [Quit: WeeChat 1.9]
voiceftp has quit [Remote host closed the connection]
<sophiag>
does anyone know where the torrc file is located?
voiceftp has joined #nixos
vk3wtf has joined #nixos
vidbina has quit [Ping timeout: 245 seconds]
<infinisil>
sophiag: What's that?
<infinisil>
Oh, tor config. Probably somewhere in ~, so you can just `file` or so to find it
<infinisil>
Maybe
Mateon3 has joined #nixos
<drakonis>
i'm trying to figure out why pipenv won't pull in pyqt5
<sophiag>
infinisil: i think i found the problem...
<sophiag>
it seems to generate it after you launch tor for the first time
<drakonis>
i may as well just package the missing dependency instead of using pipenv
<sophiag>
but when trying to do so i get "You cannot run Tor Browser from a read-only file system. Please copy Tor Browser to another location before trying to use it."
<sophiag>
srhb: actually nixos.torbrowser was giving me trouble. i just had to swap it out for the bundled one
<sophiag>
makefu: well, the reason i need to edit this file is really absurd...
chris-martin has joined #nixos
<srhb>
sophiag: Hum, strange.. :)
<sophiag>
i'm abroad and need to purchase health insurance to avoid a tax penalty from the us govt but the site blocks foreign ips
<sophiag>
and the reason i'm travelling is to pay out of pocket for better health care!
<sophiag>
#thefuture
<chris-martin>
(sorry, repeating my question because I've been offline:) Is the cache having problems? http://ix.io/1qk7
<sophiag>
ugh! i guess it blacklists exit nodes too??
<sophiag>
i'm actually impressed new york state even knows tor exists
<chris-martin>
I can't seem to build much of anything on EC2 today, it just hangs at "copying path '/nix/store/wprwklil0487mwqivldmix866fy72i7f-gsettings-desktop-schemas-3.28.0' from 'https://cache.nixos.org'..."
<sophiag>
or not..i edited torrc but it's still bouncing me through mumbai..
<srhb>
chris-martin: Works fine here. But I just noticed your message in #unregistered -- you're getting a cloudfront front to cache.nixos.org? I thought we switched over to fastly
<srhb>
chris-martin: That seems fishy
<srhb>
chris-martin: Might want to investigate your DNS. :)
<chris-martin>
sorry I'm not much of an irc user, I am confused here
<samueldr>
(the index page probably wasn't updated when moving to fast.ly)
<srhb>
chris-martin: (You got redirected to #nixos-unregistered before because you weren't authenticated at the time -- now you're in #nixos proper)
<chris-martin>
I just started this instance from an older nixos AMI
<samueldr>
(don't remember the moment of the move)
<srhb>
chris-martin: Still the ipv4 curl seems to succeed...
<chris-martin>
correction, it isn't an older AMI, it's the latest 18.09
crmlt has quit [Ping timeout: 252 seconds]
<srhb>
chris-martin: Can you try downloading the nar in question by hand and seeing what's going on? I suspect it's just ridiculously slow from the behaviour you're describing
<srhb>
Which does indicate some problem getting to the cache
boxscapeR has quit [Remote host closed the connection]
<georges_>
I would like to add a runtime dependency to docker-compose on git. How would I proceed?
<georges_>
because docker-compose needs git and I have the following error "ERROR: failed to init repo at /var/lib/docker/tmp/docker-build-git190939800: : exec: "git": executable file not found in $PATH"