<maralorn>
clever: That‘s cool. But why does the example from the help not work? Have I misconfigured my system? Does that example work on normal systems? Do I need to be in the nixpkgs folder for it to work?
<clever>
maralorn: what is the contents of $NIX_PATH ?
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @Ma27 pushed to structured-attrs « treewide: use flat lists for *Inputs, fix env usage »: https://git.io/JeVHm
NoctisLabs has quit [Quit: WeeChat 2.6]
NoctisLabs has joined #nixos
<NoctisLabs>
When I start Firefox on i3 it sort of glitches out and only renders on half the screen. Switching it to floating and back to tiled fixes it. Anyone else had this issue?
philr has quit [Quit: WeeChat 2.6]
<vika_nezrimaya>
How to use nixpkgs.obs-wlrobs with nixpkgs.obs-studio? Please note that I am an OBS noob and both packages are currently in my configuration.nix
<{^_^}>
[nixpkgs] @globin pushed to structured-attrs « treewide: fix types for mkDerivation params »: https://git.io/JeVHn
wikiemol has joined #nixos
mojjo has quit [Remote host closed the connection]
stagboot has quit [Quit: Leaving]
<wikiemol>
Hello, I am a bit confused about the instructions provided in section 7.2 Using a script as a builder here https://nixos.org/nixos/nix-pills/working-derivation.html#idm140737316245424. Maybe I am just misunderstanding, but when I followed the steps, the environment variables were not printed in the nix repl, but the derivation did build (a
<wikiemol>
<hash>-foo was put into my nix store with "foo" as text). When I left the repl, and ran declare -xp, the environment variables were edited. I thought this was supposed to be sandboxed when the build was running? Am I misunderstanding something?
<pie_>
does nobody on nixos use cgroups? am i just missing something?
Ariakenom has quit [Quit: Leaving]
<infinisil>
wikiemol: They probably weren't printed because the repl doesn't output stderr. It should print it with e.g. nix-build
<infinisil>
wikiemol: "When I left the repl, and ran declare -xp, the environment variables were edited" -> What do you mean they've been edited?
<wikiemol>
My Path looks longer, but actually, maybe I am just imagining it
<wikiemol>
I don't have proof
<wikiemol>
infinisil If building the derivation in the repl isn't supposed to change anything outside of the build, then I would fully believe my memory is just bad, I just wanted to verify my understanding and make sure that nothing was supposed to be changed.
<infinisil>
Yeah it really shouldn't be able to change anything outside the build :)
selfsymmetric-mu has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JeVH4
<infinisil>
The only commands that could change stuff outside the store are ones that do more than just building derivations, such as nix-env or nixos-rebuild
<atlas_>
I called Apple Support and they don't know how I could make the system volume (which includes /) writable. I guess nix won't work on macOS >=10.15 unless nix either implements a variable nix root dir or we find a way to break into the read-only volume.
<{^_^}>
nix#2925 (by mroi, 22 weeks ago, open): /nix will not be writable on macOS Catalina
ris has quit [Ping timeout: 246 seconds]
<atlas_>
clever: I don't want my nox dir overwritten by updates or something wtf
<wikiemol>
infinisil Okay thanks! So, I ran nix-build /nix/store/ac0qjiha6vh79ghka07cdy021f75xdf6-foo.drv and I didn't see the output. Is that what you meant?
<infinisil>
wikiemol: Nix won't build things it already built by default. Run that same command with --check to force it to build it again
<infinisil>
(without --check it should only have output a single line, the resulting store path)
selfsymmetric-mu has quit [Remote host closed the connection]
<wikiemol>
infinisil That worked!!
<wikiemol>
Thanks!
<infinisil>
:)
<clever>
wikiemol: run `nix-store -l /nix/store/foo.drv`
<pie_>
infinisil: im dont really know how systemd works, i havent looked at anything cgroup specfic to systemd, no
<infinisil>
Hm yeah not sure if it could be used to have user processes start in it
<pie_>
infinisil: if it means anything, i dont want to run firefox as a service
wikiemol has joined #nixos
Osmiate has quit [Remote host closed the connection]
<wikiemol>
,locate nm
<pie_>
the random things ive seen done are spaw a shell in a cgroup and start stuff from there, or i think some permission delegation stuff that just lets people directl add stuff to a cgroup
<pie_>
not currently supported by systemd, such as net_prio. "
<pie_>
infinisil: oh wait "The libcgroup package, which was the main tool for cgroup management in previous versions of Red Hat Enterprise Linux, is now deprecated. To avoid conflicts, do not use libcgroup tools for default resource controllers (listed in Available Controllers in Red Hat Enterprise Linux 7) that are now an exclusive domain of systemd. This leaves a limited space for applying libcgroup tools, use it only when you need to manage controllers
<pie_>
sooooo im not sure what one is supposed to do here
<pie_>
infinisil: right , so run some script on startup?
<pie_>
ill look at that SE post
<pie_>
why is this so confused heh
<infinisil>
Probably yeah, or maybe cgmanager can start a process in a cgroup to start with
<gyroninja>
Hydra is failing the package (different versions though) on 19.09 and unstable with "do not know how to unpack source archive"
mexisme has joined #nixos
<gyroninja>
is the unpack unpure or something?
<clever>
gyroninja: you probably have to add xz to the nativeBuildInputs
jgt has quit [Ping timeout: 246 seconds]
mexisme has quit [Ping timeout: 246 seconds]
<wikiemol>
is there an equivalent of nix-build --check for failed builds?
<wikiemol>
To avoid getting the message "error: some outputs of '/nix/store/vc1y7p66r7bfw5bqxaj8xzbpd2d29w6i-simple.drv' are not valid, so checking is not possible"
work_ has quit [Quit: Connection closed for inactivity]
<pie_>
infinisil: looks like systemd-run might be what i want
<infinisil>
wikiemol: nix-build without --check
<infinisil>
pie_: Ah neat, that does sound like it might work
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen1 has joined #nixos
<wikiemol>
ah, so the build failing is 100% my fault :(
o1lo01ol1o has joined #nixos
<pie_>
infinisil: actually, this doesnt look like it has anything directly to do with cgroups. im not sure whats going on and i dont like the vibes im getting, im not usually one to say that :P
<pie_>
infinisil: so this looks like...it will make a one-off service for me?
<pie_>
and then i have to use the usual systemd resource allocation mechanisms to deal with that (?)
dingenskirchen1 is now known as dingenskirchen
felixfoertsch23 has joined #nixos
felixfoertsch has quit [Ping timeout: 252 seconds]
<clever>
gyroninja: unxz is already in the PATH of that derivation
<clever>
gyroninja: and the .drv nixpkgs generates, is identical to the one hydra has
<clever>
gyroninja: its likely just a random failure, and the hydra job needs to be restarted
<gyroninja>
Well it looks like the issue has been arround for 2 weeks
<clever>
gyroninja: likely, none of the inputs have changed, so hydra hasnt bothered to try again, since doing the exact same thing twice is the definition of insanity
<gyroninja>
and it doesn't say it's a propgated failure so I assume it is not a cached result
<clever>
thats weird
ddellacosta has quit [Ping timeout: 245 seconds]
<clever>
gyroninja: only other thing i can think of, is things like sse2 features
<{^_^}>
[nixpkgs] @leocp1 opened pull request #73149 → rasdaemon: init at 0.6.4 → https://git.io/JeV7t
stagboot has quit [Quit: Leaving]
jedai42 has quit [Ping timeout: 240 seconds]
o1lo01ol1o has joined #nixos
dansho has joined #nixos
nexgen has joined #nixos
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 is now known as Supersonic
o1lo01ol1o has quit [Ping timeout: 240 seconds]
<wedens[m]>
clever: `/nix/store/fg5kgrs619jldskbcqhp20j6l1wnwkv5-nixos-system-wedens-pc-19.09pre-git` that's what I have in the link :/
<clever>
wedens[m]: then it was built from a git clone, and the .version-suffix file is missing
<wedens[m]>
clever: it was built with NIX_PATH=nixpkgs=...
<clever>
wedens[m]: but what was it pointing to?
<clever>
a url or a dir?
<wedens[m]>
github url
<clever>
ah, those also lack .version-suffix
<clever>
no easy way to find the nixpkgs rev directly
<clever>
but if you guess, and use nix-diff and nix-instantiate, you can see how much differs between whats built and your guess
<clever>
then go forwards/backwards in history
iyzsong has quit [Ping timeout: 240 seconds]
jgt has joined #nixos
<wedens[m]>
alright. thanks for suggestions
<wedens[m]>
clever: my theoretical use case is: I know that in some specific generation some package was working and I'd like to use nixpkgs revision of this generation to pin the package
<clever>
wedens[m]: thats much simpler if it was built from a channel
jgt has quit [Ping timeout: 246 seconds]
<wedens[m]>
just when I was trying to get rid of channels :D
LysergicDreams has quit [Ping timeout: 276 seconds]
Ralith_ has quit [Ping timeout: 276 seconds]
LysergicDreams has joined #nixos
<clever>
wedens[m]: you can also point NIX_PATH to a tarball that was made for a channel, but those are a bit more tricky to find
<clever>
wedens[m]: this generates a vim, that has a vimrc baked into it, along with plugins
<clever>
with a minor change (make it not a module, add with import <nixpkgs>{};), i can nix-build it, and ./result/bin/vim to test the changes (or use it on somebody elses machine)
Guest84 has quit [Remote host closed the connection]
palo has quit [Ping timeout: 240 seconds]
palo1 is now known as palo
mexisme has joined #nixos
revtintin has joined #nixos
<atlas_>
Guest95680: What do you mean? If this is what you mean: what is "prefix"?
mexisme has quit [Ping timeout: 252 seconds]
<atlas_>
whoops wrong hl Guest84
<atlas_>
"The installation path can be specified by passing the --prefix=prefix to configure. The default installation directory is /usr/local." implies that "prefix" is /usr/local, but I don't have anything like /usr/local/etc/profile.d/nix.sh
rauno has joined #nixos
jgt has joined #nixos
boredbigu has joined #nixos
jgt has quit [Ping timeout: 245 seconds]
kolbycrouch has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
boredbigu has left #nixos [#nixos]
dingenskirchen has quit [Quit: dingenskirchen]
dingenskirchen has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace pushed to master « cadence: generate wrap commands, use qt5's mkDerivation »: https://git.io/JeV50
<{^_^}>
[nixpkgs] @worldofpeace pushed to release-19.09 « cadence: generate wrap commands, use qt5's mkDerivation »: https://git.io/JeV5u
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JeV5a
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-19.09: https://git.io/JeV5V
<{^_^}>
[nixpkgs] @vcunat pushed 9 commits to release-19.09: https://git.io/JeV5D
<{^_^}>
[nixpkgs] @Izorkin opened pull request #73151 → php74: init at 7.4 RC5 → https://git.io/JeV5y
Heirlung has joined #nixos
wikiemol has joined #nixos
o1lo01ol1o has joined #nixos
kolbycrouch has joined #nixos
o1lo01ol1o has quit [Ping timeout: 265 seconds]
mexisme has joined #nixos
<wikiemol>
Hello, I am trying to follow the steps in this section of this nix pill https://nixos.org/nixos/nix-pills/working-derivation.html#idm140737316196736 and I am getting stuck. I am on "x86_64-darwin" (OS X Mojave). The main problem is that collect2 is not finding nm. I tried adding builtins = builtins to the derivation set, to no avail. I also tried
<wikiemol>
builtins_unwrapped = builtins_unwrapped. I even tried adding $binutils/bin to my PATH in the bash script in the tutorial. But that didn't work either. What are some ways I could debug this?
<{^_^}>
[nixpkgs] @mmahut opened pull request #73154 → datovka: init at 4.14.0 → https://git.io/JeVFa
chloekek has joined #nixos
LysergicDreams has quit [Ping timeout: 240 seconds]
wikiemol has quit [Remote host closed the connection]
silver has joined #nixos
<bdju>
what all is needed for interacting with android phones on NixOS? I've got adb enabled but don't seem to have the fastboot command. I see there's a package called android-udev-rules that I might need... can't find anything on fastboot or android-tools which is what contains it on some distros
<ashkitten>
,locate bin fastboot
<{^_^}>
Couldn't find in any packages
<rnhmjoj>
ashkitten: are you looking for the adb stuff?
<ashkitten>
rnhmjoj: was trying to see if it's provided by a package that the bot here can see
jgt1 has quit [Read error: Connection reset by peer]
<rnhmjoj>
it should be in androidenv.androidPkgs_9_0.platform-tools
<ashkitten>
for bdju
<ashkitten>
and yeah it's in platform-tools according to `readlink $(which fastboot)`
<bdju>
alright, thanks
<bdju>
okay platform-tools in my config is apparently not valid and the way I search nix packages freaks out over the - even with quotes
<MichaelRaskin>
I expect the problem to be with unfreeness of Android SDK parts
chreekat has quit [Ping timeout: 245 seconds]
<bdju>
guix has this stuff packaged
bvdw has quit [Read error: Connection reset by peer]
<flokli>
jtojnar: Ericson2314 jtojnar: got success getting meson to statically compile? looking at pkgs/stdenv/adapters.nix, it seems we only need to add "-Ddefault_library=static" to mesonFlags, however, meson docs suggest it should come immediately after "meson", not after mesons "build" subcommand (which is where mesonFlags end up in)…
<pie_>
i dont know much about systemd, does anyone know why `systemd-run --user --scope -p MemoryHigh=300M firefox ` does successfully run but doesnt seem to create a cgroup, and no manner of resource limiting seems to be imposed on firefox?
<pie_>
systemd-cgls lists a .scope but systemd-cgtop doesnt show a cgroup
<adisbladis>
pie_: Iirc user units don't support resource limiting
<pie_>
adisbladis: .....FFFFUUUU xD
<pie_>
i figured the daemon would handle stuff so i dont have to hav root
<pie_>
all the other cgroup management things are marked deprecated in favor of systemd so how a i supposed to do this *scratches head*
<pie_>
adisbladis: wair
<pie_>
wait, if thats true, why the heck doesnt it give me any warnings or anything
<pie_>
systemctl --user status ... shows `CGroup: /user.slice/user-1000.slice/user@1000.service/tempname.scope` , so i guess its just cgtop that doesnt show some thigns for some unoverridable reason
<pie_>
so yeah why doesnt this work
civodul has joined #nixos
<pie_>
its supposed to start swapping if i pass MemoryHigh right?
<infinisil>
NinjaTrappeur: I think `nix-build nixos/release.nix -A manualHTML.x86_64-linux`
<__monty__>
pie_: Isn't MemoryLimit the one you want? That's what I've been using.
cosimone has joined #nixos
gxt has joined #nixos
<pie_>
__monty__: i dont want it to get killed
<pie_>
i might be mixing that up with max
<pie_>
i think i read emorylimit is deprecated in favor of memoryhigh
<__monty__>
Ah, MemoryMax replaces MemoryLimit
<jtojnar>
flokli I am mostly interested in nixos, so I never tried that
LysergicDreams has quit [Read error: Connection reset by peer]
LysergicDreams has joined #nixos
<flokli>
jtojnar: but on NixOS you can also have static binaries and libraries ;-)
<NinjaTrappeur>
Indeed infinisil, thanks!
<flokli>
I assume we might want to introduce a second list for flags to meson before the "build" command (or however this is called in meson nomenclature)
<__monty__>
Question about setting up bonding. Somehow wpa_supplicant ends up configured wrong: http://ix.io/214y
pointfourone has quit [Quit: WeeChat 2.6]
jmeredith has joined #nixos
cosimone_ has joined #nixos
cosimone has quit [Ping timeout: 250 seconds]
ZIOCARLO45 has joined #nixos
ZIOCARLO45 has left #nixos [#nixos]
iyzsong has quit [Ping timeout: 265 seconds]
Soo_Slow has joined #nixos
dansho has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed commit from @risicle to nixos-19.03-small « jhead: add patches for CVE-2019-1010301, CVE-2019-1010302 »: https://git.io/JeVAm
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed commit from @risicle to nixpkgs-19.03-darwin « jhead: add patches for CVE-2019-1010301, CVE-2019-1010302 »: https://git.io/JeVAm
<ryantm>
The nixpkgs manual has a section about building all Haskell packages with profiling on. https://nixos.org/nixpkgs/manual/#miscellaneous-topics Is it possible to partially enable profiling, so I don't need to recompile everything?
<nh2>
flokli: I'm trying out building a hell-world project statically with Meson.
<nh2>
That is with meson `0.45.1` as delivered by Ubuntu.
<nh2>
`meson builddir -Ddefault_library=static` doesn't do anything: `meson configure` afterwards continues to show `default_library shared`.
<nh2>
First point: On Ubuntu 18.04 meson is certainly bugged.
<nh2>
Updating to meson `0.51.2` from nixpkgs fixes that; `meson configure` now shows what I've set.b
<nh2>
That doesn't give me a static exe yet, because it probably talks only about libraries, but is good to know that we shouldn't try make things work with old meson versions.
chiefgoat has joined #nixos
<flokli>
nh2: are we using old meson versions in nixpkgs?
<nh2>
flokli: no, at leat not in 19.09, that's where I got the working `0.51.2` from
<nh2>
flokli: `-Dc_link_args=-static` results in a static exe
<flokli>
nh2: but shouldn't it be a `meson --default_library=static build ${mesonFlags}`?
<nh2>
flokli: I don't think `--default_library=static` is legal, you meant `-Ddefault_library=static` like above, right?
<nh2>
Also I think those two are different things:
<nh2>
If you build *libraries*, then `-Ddefault_library=static` makes them be static (.a files).
<nh2>
If you build *executables*, then `-Dc_link_args=-static` makes them be statically linked (`ldd` shows `not a dynamic executable`).
cosimone_ has quit [Quit: Quit.]
domogled has joined #nixos
<flokli>
nh2: seems you are right
<flokli>
And we don't need something separate from mesonFlags
acarrico has quit [Ping timeout: 265 seconds]
asymmetric has joined #nixos
<nh2>
flokli: I just found that your approach of using `--arg=value` only works with some subset of options: https://mesonbuild.com/Builtin-options.html#universal-options. Everything that's in `Universal options` seems to accept that, but the stuff further down in `Compiler options` does not
acarrico has joined #nixos
Thra11_ has quit [Ping timeout: 240 seconds]
chiefgoat has quit [Ping timeout: 265 seconds]
asymmetric has quit [Ping timeout: 246 seconds]
asymmetric has joined #nixos
<gentauro>
how do you make a local sandbox with both `gtk2` and `gtk3`? I'm trying to build an old version of FF (48.0.2 to be specific)?
<gentauro>
if I add both `stdenv pkgconfig gtk2 gtk3 glib gobjectIntrospection ...` I get: "collision between ... -gtk+3-3.24.5 and gtk+-2.24.32"
drakonis has joined #nixos
<gentauro>
Shouldn't it be this that nix/nixos should be really good at?
drakonis has quit [Read error: Connection reset by peer]
<nh2>
clever infinisil: I need another nix-shell tip like you gave me recently. It seems `-p git` gives me shell with git in it, but `nix-shell -E '(import <nixpkgs> {}).git'` gives me a shell in which I can _build_ git. (How) can I use `-E` to get a shell with git in it?
<gentauro>
providing both gtk2 and gtk3
drakonis has joined #nixos
<symphorien>
you don't need stdenv in buildInputs
oscarvarto has joined #nixos
<symphorien>
and I just tried nix-shell -p gtk2 gtk3 and it seems to work
<edef>
nh2: nix-shell -p 'foo bar' baz is shorthand for nix-shell -E 'with import <nixpkgs> {}; mkShell { buildInputs = [ foo bar baz ]; }'
<nh2>
Looks like you can also just replace the `-E` with `-p` and it'll do what I need (thus a shortcut for what you wrote). But it is not documented in the man page apparently
<{^_^}>
nix#454 (by cstrahan, 4 years ago, closed): BUG: nix-shell requires bogus -p argument to make --expr work
<flokli>
I need to check why it's not working in my special case, though ;-)
<edef>
nh2: so either way you're in a shell for building some derivation, just -p makes you a synthetic one
<nh2>
that is, this works: `nix-shell -p '(import <nixpkgs> {}).git'`
<edef>
nh2: yes
<edef>
nh2: look at the substitution i showed you above
<edef>
nh2: what you're doing is equivalent to nix-shell -E 'with import <nixpkgs> {}; mkShell { buildInputs = [ (import <nixpkgs> {}).git ]; }'
<symphorien>
ah the error is in buildEnv
Thra11_ has joined #nixos
<nh2>
edef: yes it makes sense now that I understand it, but before I didn't assume that the argument _literally_ spliced in the given expression this way as a string, especially because the man page says ` -p packages`
<symphorien>
remove the line with env=buildEnv....
<edef>
nh2: apparently the modern, C++ version does actually parse it in isolation
<edef>
nh2: the old perl version does a literal perl splice
<edef>
nh2: er, literal string splice*
<edef>
nh2: so `nix-shell -p ']; foo = ['` was valid
<edef>
nh2: apparently it rejects that now — progress!
<gentauro>
symphorien: oh, thanks that worked :)
<gentauro>
I think I copy/pasted that line from some online example
<nh2>
edef: indeed
<symphorien>
I dont't really understand what this line was meant to do
<gentauro>
why isn't it needed?
<symphorien>
I would reverse the question: what was it needed for ? :þ
<gentauro>
:D
<gentauro>
bugs created by copy/pasta
<{^_^}>
[nixpkgs] @ikervagyok opened pull request #73165 → factorio: update all versions → https://git.io/JeVpw
reallymemorable has joined #nixos
jgeerds_ has quit [Ping timeout: 240 seconds]
gxt has joined #nixos
<gentauro>
btw, is there anyway to downgrade `sed` for a specific sandbox?
<gentauro>
I can read that `sed` changed their interface from 4.2 and currenlty on NixOS is 4.7 ...
<andi->
Maybe I am not running int o that since I am using a fork of nix
<evanjs>
Okay so I think we figured the problem out with a different glibc version — it was a rust project so we just overrode the sysroot to the host system rather than nixpkgs
rtjure has quit [Quit: rtjure]
fendor has joined #nixos
tsrt^ has joined #nixos
<evanjs>
But it brought up another question: does anybody have an idea of when we might be updating glibc? re #46890
<evanjs>
Looks like aarch64 is the blocker based on the most recent comment
<gentauro>
how do I actually overrule gcc with a local sandbox?
bvdw has quit [Read error: Connection reset by peer]
<gentauro>
I keep getting `gcc-7.4` even though I add `buildInputs = [ gcc8 ]` ...
<gentauro>
spooky
<flokli>
Izorkin: okay, sounds reasonable. Can you update the commit message of the existing commit, and add a second one introducing the samba-smdb - samba-nmbd dependency?
<flokli>
I wouldnt call it revert, at least not in the first line of the commit msg. Something like "samba: remove redundant dependency on network.target\nThis reverts commit xxxx. We already set after=network.target on samba.target.
<flokli>
Izorkin: define it's not working. What are the symptoms? Error logs? Also, please push the new commit messages, and let's move discussion back into the issue.
phaebz has quit [Remote host closed the connection]
phaebz has joined #nixos
rtjure has joined #nixos
<karetsu>
I'm having a dbus issue - my user dbus goes inactive (dead) shortly after login [not sure if its meant to do this], I can't use notify-send as the messages just disappear into a void with no error but no notification and I cannot unlock any admin panels in gnome-control-center with a permissions error [yes, my user is wheel] - is this dbus or accountsservice at fault?
rtjure has quit [Client Quit]
rtjure has joined #nixos
rtjure has quit [Client Quit]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
doelie has joined #nixos
<doelie>
hi. is there a reason why binutils doesn't export objcopy? any other package that exports it?
<doelie>
binutils maps to binutils-wrapper it seems
<__monty__>
,locate bin objcopy
<{^_^}>
Found in packages: binutils-unwrapped
stepcut has joined #nixos
<stepcut>
Is there a way to get nix-env -q to show all the emacs packages that are available? Or any query at all that will show that nix-mode is available?
<karetsu>
stepcut: doesn't `package-activated-list` in emacs show nix-mode?
<stepcut>
karetsu: Well, at this stage I have not even installed emacs or nix-mode..
<adisbladis>
gentauro: That's not what stepcut asked for though, it was to search in emacs package names from nixpkgs
noudle has quit []
<gentauro>
adisbladis: ahhh, didn't see that
<stepcut>
gentauro: well.. I am trying to do things like use nixops deploy to setup a server that emacs available with nix-mode and other packages already installed. So I need to know the nix attribute names. In my case, `emacsPackagesNg.melpaStablePackages.nix-mode`. But I'd be great if I could get `nix-env -q` to find that attribute path for me
<adisbladis>
stepcut: You could do something like: nix-instantiate --eval --expr 'builtins.toJSON (builtins.attrNames (import <nixpkgs> {}).emacsPackages)' | jq -r '.' | jq -r
<stepcut>
(also, I am currently using an old nixpkgs-stable snapshot when Ng was still Ng)
<adisbladis>
It's not pretty, but works in a pinch
<gentauro>
stepcut: but you know from what adisbladis said that all of the melpa pkgs are somehow `scraped`. Could you download/search for `nix` and the prependt the static `emacsPackagesNg.melpaStablePackages`?
cosimone has joined #nixos
<wikiemol>
I am kind of confused as to the point of nix-shell. What are the use cases for it? I thought it was to test go into a "sterile" environment where only the environment variables from the .nix file are loaded. But this doesn't seem to be the case
rauno has quit [Ping timeout: 246 seconds]
<stepcut>
adisbladis: that's pretty horrible. :) Sounds like I am correct in thinking that `nix-env -q` is pretty terrible -- it is slow and can't find half the packages in nix anyway
<adisbladis>
stepcut: That's by design.. A full eval takes a long time
<gentauro>
wikiemol: I use it to install stuff in a local sandbox (binaries only accesible from that nix-shell)
<stepcut>
gentauro: there are many things I *could* do, the question is whether `nix-env -qa` whose entire purpose is to do that for me, can actually do it
<stepcut>
gentauro: and the answer seems to be 'no'
<wikiemol>
symphorien --pure still seems to bring in the environment variables from my regular environment
<symphorien>
it brings in your bashrc
<symphorien>
but it should not bring in arbitrary env vars
<stepcut>
adisbladis: yup! if the query included all the emacs, haskell, python, etc, libraries by default it would take foreveeeeeer.
<wikiemol>
Ooh, is there a way to get it to not bring in the bashrc?
<symphorien>
it executes bash
<stepcut>
adisbladis: but.. it would still be nice if something like, `nix-env -qaP emacsPackages.\.*` worked
<symphorien>
so --run "the invocation of bash with right option to have it skip bashrc" might do
<adisbladis>
stepcut: Though we could certainly change it to include emacs packages.. But that's a few thousand extra packages to eval :/
<__monty__>
stepcut: Have you tried emacsPackages.\*?
<red[m]>
@wikiemol - ever had different packages that can’t coexist on the same machine, either because they share the same name or have dependencies that are incompatible?
<red[m]>
It solves that problem
<wikiemol>
symphorien Gothcha, that makes sense
<__monty__>
wikiemol: That *is* the case when you use nix-shell --pure. Or should be at least : )
<mightybyte>
Anyone know how to set system-wide ulimits in nixos?
Thra11_ has joined #nixos
<mightybyte>
I haven't found anything in the docs or Appendix A
<__monty__>
The impurity is often useful so you can use some tools from your environment when testing a package.
<adisbladis>
++
<adisbladis>
Develop in impure shell, run tests in pure shell
<stepcut>
__monty__: I have tried many variations of that, with -qa and -qaP. But maybe I have not gotten the right combination of things
<__monty__>
stepcut: It may not be possible. Just know it works for haskellPackages. -qaAP nixpkgs.haskellPackages
<wikiemol>
Interesting, thanks everyone. I think I must have done something strange when I was experimenting. I opened a new terminal and there are far less environment variables now.
<wikiemol>
When I go into nix-shell --pure
Chiliparrot has joined #nixos
<__monty__>
stepcut: Actually it doesn't with the wildcard.
<__monty__>
And *-qaPA
<stepcut>
adisbladis: right -- that is why it would be nice if there was I way I could specify that I want to eval those (and only those) packages..
<stepcut>
__monty__: oh! That seems to be the key, the -A flag
<__monty__>
Oh, so it works? Great : )
<stepcut>
__monty__: works better than anything I've tried. trying to figure out if I can get it to match on just nix-mode or if I need to use grep
<Thra11_>
Has anyone ever tried writing an overlay which imposes a minimum version on a package? I.e. it checks super.version, and only overrides it if it's below the minimum. I can't see any reason why you *couldn't* do this, I'm just wondering whether it's a good idea.
<__monty__>
stepcut: You can, but I prefer putting that output in a file and then grepping that because it needs to evaluate nixpkgs entirely every time.
<adisbladis>
Thra11_: Sure, why not?
<stepcut>
this works, nix-env -qaA 'emacsPackagesNg' | grep nix-mode, but maybe there is a 'better way'
<stepcut>
also, it gets mad if I try to use -A and -P at the same time
<stepcut>
oh, the P has to go before the A
<__monty__>
stepcut: Yeah, that's my *-qaPA comment.
<stepcut>
yeah
<__monty__>
You can try specifying 'emacsPackagesNg.nix-mode' and add --description
<stepcut>
it looks like when using -A you can not use a regex?
<__monty__>
I don't think you ever can when selecting an attribute?
<stepcut>
__monty__: not that I am aware.. just checking :)
<__monty__>
,locate nix-mode
<{^_^}>
Found in packages: nix1
<__monty__>
Ah, I guess nix-locate doesn't recurse into emacsPackagesNg either.
<__monty__>
I guess you could do an overlay where you override dontRecurseIntoAttrs, call it "emaxpkgs" in NIX_PATH and then use `-f '<emaxpkgs>'` with locate or just nix-env for querying.
drakonis_ has quit [Ping timeout: 246 seconds]
<gentauro>
didn't know aoubt `nix-shell --pure`. It's kind of nice
dansho_ has joined #nixos
<gentauro>
so you can ensure that you can't call stuff that is not part of the project :o
<manveru>
gentauro: note that --pure may not be fully pure :)
drakonis_ has quit [Ping timeout: 276 seconds]
<manveru>
since for example stuff like bash will still read your ~/.bashrc
growpotkin has joined #nixos
eightyone has joined #nixos
rtjure has joined #nixos
ng0 has quit [Ping timeout: 260 seconds]
jluttine has quit [Ping timeout: 240 seconds]
tungdil has joined #nixos
acarrico has quit [Ping timeout: 240 seconds]
<tungdil>
Hi! I'm having a problem using NixOS in GnomeBoxes. The mouse does not work properly (mouse cursor gets stuck or jumps to host system). Any hint?
ng0 has joined #nixos
Thra11_ has quit [Ping timeout: 240 seconds]
<betaboon>
does anyone know if the installer boots in efi mode ?
<adisbladis>
betaboon: It can
stepcut has quit [Remote host closed the connection]
<samueldr>
it should
pointfourone has joined #nixos
<betaboon>
argh this is driving me nuts. I'm just working non the tests for #58121. "problem" is, that the initial installation in the tests are done booted in bios-mode. thus no efivars are available and that just has a chain of impacts ...
<wikiemol>
If I want to use the version of nix that is currently in master, what are the steps to do so?
jmeredith has quit [Quit: Connection closed for inactivity]
<wikiemol>
I saw some mention of this on the website https://nixos.org/nix/download.html in the continuous builds section, but it seems the documentation is sparse
<betaboon>
samueldr: yeah thats what i am struggling with. let me elaborate a bit: initial nixos-install in tests is run in bios-mode -> thus no efivars -> thus it is required to install into the fallback location -> thus installer-tests need to check for bootloader config to be located there. systemd-boot just installs itself to fallback location anyway.
<samueldr>
imo, here systemd-boot is rude :)
<samueldr>
and I do see your issue
<samueldr>
I'm checking out your branch to have a better look
Thra11_ has joined #nixos
<betaboon>
samueldr: i am considering enforcing refind to be installed into the fallback location. but that feels weird
<samueldr>
yeah, that's not a good solution
<samueldr>
though having the install as removable flag moved under the more generic boot.loader.efi namespace, and then installing there when it is true, would be fine I think
<betaboon>
but if i dont do that, the location for the configuration file will be different depending on whether it has been installed in bios-mode or efi-mode. thus there will be a hidden assumption on the tests.
<samueldr>
ooh, I see
<samueldr>
I hadn't thought about how refind confs are relative to the binary
<betaboon>
yeah. i looked into getting it to a different location. but that wont work.
<{^_^}>
[nixpkgs] @Ma27 pushed to structured-attrs « glib: patch python3 shebangs for gdbus-codegen »: https://git.io/Jewvu
<samueldr>
still, wouldn't the assumption hold true in all cases, that if it's installed a removable, it's in /EFI/BOOT, otherwise in (IIRC) /EFI/NixOS/ ?
gxt has quit [Ping timeout: 260 seconds]
<samueldr>
maybe I'm not seeing what the issue is
<betaboon>
moving the installassremovable option to the generic efi-options seems good. but it seems like that wont be possible due to systemd-boot not supporting that kind of option
<betaboon>
brb in 5 minutes
<samueldr>
yeah, I wouldn't put the work on making all EFI bootloader respect that option in that PR anyway
jedai42 has quit [Ping timeout: 265 seconds]
jluttine has joined #nixos
zupo has joined #nixos
<hyperfekt>
can anyone here read makefiles and tell me what i have to do to the qemu derivation to get it to install with debug symbols?
LysergicDreams has quit [Ping timeout: 265 seconds]
<gentauro>
19:18 < manveru> since for example stuff like bash will still read your ~/.bashrc
<betaboon>
samueldr: now talking to you and learning about the existence of installasremovable i would like to do the following: add installasremovable to refind options -> have that option enforce installing to fallback location -> in the installer tests use installasremovable and check for refind.conf in fallbacklocation -> removing the hidden assumption from tests and being a good citizen on efi-mode
<betaboon>
installations. opinions ?
<samueldr>
the main thing to remember is --pure is mainly pure through environment variable isolation; all other impurities subsist :)
pointfourone has quit [Quit: WeeChat 2.6]
<samueldr>
betaboon: sounds good
mexisme has quit [Ping timeout: 252 seconds]
<samueldr>
"removing hidden assumption from tests", does that mean fixing what seems to be the issue with the initial boot being in bios mode?
<betaboon>
samueldr: nice. thanks for bouncing ideas :D had been locked in a circle for an hour.
<betaboon>
samueldr: yes.
<samueldr>
good
<{^_^}>
[nixpkgs] @Ma27 pushed 2 commits to structured-attrs: https://git.io/Jewvy
<betaboon>
when trying to install refind from within bios-mode there is no other way than installing to fallback location IMHO
<betaboon>
samueldr: it's just weird that systemd-boot seems to be so rude
<edef>
[nix-shell:~]$
<samueldr>
betaboon: to me too, though it feels likely to be part upstream behaviour, part "first EFI bootloader in nixpkgs" legacy
o1lo01ol1o has joined #nixos
<samueldr>
time eludes me, and I've not made myself favours in the matter with starting work on other nixos stuff... but I wanted to study all bootloader things in nixpkgs, and better consolidate things
<betaboon>
samueldr: when you have a plan of action and would like some help just hit me up
<betaboon>
unemployed for the rest of the year so i got some time on my hands XD
<samueldr>
no real plan :) the first step was to take inventory of everything bootloadery (including u-boot, and raspberry specific options) and make a kind of feature matrix
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 5 commits to nixos-19.09-small: https://git.io/JeVhT
<jakubracek>
Hi, having a problem with my nixos. Can't really login. Broke display-manager somehow and TTY just keeps asking me for login over and over again. So I decided to chroot into the partition from the live usb but I can't really do that. I've mounted my root into /mnt but nixos-enter keeps saying that the /mnt is not a nixos installation?
<clever>
jakubracek: if you `ls /mnt/` what files do you see?
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #73173 → gcc: Clean up cross configure flags and name prefix → https://git.io/JewfI
<jakubracek>
I ran 'lsblk', saw my partition named nvme0n1p2. So I did `cryptsetyp luksOpen /dev/nvme0n1p2 enc-pv`. `lvscan` and `vgchange -ay` now mounted `mount /dev/nixos-vg/root /mnt`
<jakubracek>
I can see there normal root. like bin, boot, home, nix, tmp, var' ...
<jakubracek>
`ls /mnt` outputs normal root. like bin, boot, home, nix, tmp, var' ...
cosimone has quit [Quit: Quit.]
acarrico has joined #nixos
<samueldr>
ah!, betaboon, I was doing something else on the side, and got confused with the tests... not pushed to the PR branch yet, right? :)
<jakubracek>
I am only really interested in chroot in there and run `nixos-rebuild --switch` but when trying to `chroot /mnt` then I get `/run/current-system/sw/bin/bash` No such file
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #73175 → gcc: Clean up cross configure flags and name prefix for 19.09 → https://git.io/Jewfc
<betaboon>
samueldr: nope not pushed yet. just working in what we discussed. always takes some time. installer-tests take "forever" XD
<samueldr>
I was going to have a look at your test, but couldn't figure out where it was :)
ddellacosta has joined #nixos
<betaboon>
samueldr: will give you a heads up when i pushed it
LysergicDreams has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 16 commits to nixos-unstable-small: https://git.io/JewfC
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pinkien>
loooks like bad type of partition, but gparted says its fine
<clever>
pinkien: what partition did you mount to /mnt/boot/ ?
<NoctisLabs>
Is there any way to add flatpak repos and applications via configuration.nix?
chiefgoat has quit [Read error: Connection reset by peer]
<babariviere[m]>
does your system support efi ?
<mica[m]>
pinkien: efi needs a very specific partition type with specific flags set
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<pinkien>
its kinda new amd heater, i guess its efi friendly
<pinkien>
fuuu :/
<mica[m]>
NoctisLabs: write a shell script in your .nix file to do it
<pinkien>
gdisk says EF00 EFI System
<jakubracek>
@clever Yeah I think it didn't do more than that. I've the system for over a year now but I still had some permission issues which I think the installer did setup incorrectly
<clever>
pinkien: what partition did you mount to /mnt/boot/ ?
<pinkien>
ah, now i see
<pinkien>
i mounted wrong partition
<drakonis>
pinkien: it stopped being a heater a while ago i guess
<pinkien>
drakonis so far i feel right now.. its better than regular heater
<pinkien>
but of course, im just joking in general
Neo-- has quit [Ping timeout: 252 seconds]
Thra11_ has quit [Ping timeout: 265 seconds]
jedai42 has joined #nixos
<betaboon>
samueldr: does the possibility exist that the bootloader is reinstalled at a later point? I'm asking due to following scenario: initial install is done in bios-mode using installAsRemovable -> later installAsRemovable is set to false -> bootloader is installed in second location !?
<betaboon>
samueldr: in those kind of situation it might be possible to move the bootloader.
<samueldr>
I'd have to review the conditions for bootloader installation, but IIRC changing options for the bootloader imply a reinstall
<samueldr>
(of the bootloader)
<clever>
pinkien: what partition did you mount to /mnt/boot/ ?
<samueldr>
otherwise you couldn't change from grub to systemd-boot, or the reverse
<pinkien>
clever i mounted originally EF02, BIOS partition
<pinkien>
EFI was supposed to be the boot one, unencrypted with luks
<pinkien>
i guess
<pinkien>
building passed.. lets burn it with reboot now :]
pinkien has quit [Remote host closed the connection]
<betaboon>
samueldr: but will chaning the bootloader remove the old one?
<samueldr>
betaboon: I don't think so
<samueldr>
I don't know that we have the ability to know it should or shouldn't remove a bootloader
<samueldr>
but, assuming efibootmgr can touch variables, it will set the new one as the default one
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #73173 → gcc: Clean up cross configure flags and name prefix → https://git.io/JewfI
chiefgoat has quit [Remote host closed the connection]
Chiliparrot has joined #nixos
Neo-- has quit [Ping timeout: 252 seconds]
chiefgoat has joined #nixos
stepcut has joined #nixos
<evanjs>
So has anybody else been getting weird errors with nixos-option lately? Like usually it'll give me an idea of where things go wrong
<evanjs>
But lately it's just 'an error occured when looking for attribute names'
<evanjs>
with `--verbose` I can go up and I think I have an idea of where things break, but certain messages seem a tiny bit less helpful now without e.g. `--verbose`
<pittma>
I'm experiencing some fontconfig weirdness with my Ubuntu-with-Nix laptop. Maybe the version of fontconfig that Nix is using doesn't match what is installed via Ubuntu? It seems to be unhappy with the content of /etc/fonts/font.conf.
<pittma>
I'm seeing, for instance "Fontconfig error: "/etc/fonts/fonts.conf", line 7: invalid attribute 'version" when trying to start the desktop irccloud app (which then crashes / core dumps). If I navigate to wher ethe AppRun lives and run it directly, all's well.
<pittma>
I can see that Nix is running it with its own bash which is what led me down the road of mismatched fontconfig versions.
<emily>
I would guess that the core dump is unrelated to the fontconfig error, fwiw
<emily>
unfortunately I don't have concrete advice beyond that
o1lo01ol1o has quit [Remote host closed the connection]
wikiemol has quit [Remote host closed the connection]
<pittma>
It works when run it directly i.e. `./AppRun` from within the nix-store directory, I presume because the necessary environment is in order (from my shell)
<pittma>
*I run it directly
<pittma>
But that script is running it from Nix's installation of Bash: `#!/nix/store/zavn4np1jvm79f0rafkv0p1mrag09qkz-bash-4.4-p23/bin/bash`
<pittma>
ah, yeah, nixpkgs has 2.10 and 2.12 available and I've got 2.13 installed.
chiefgoat has quit [Read error: Connection reset by peer]
o1lo01ol1o has joined #nixos
<clever>
,profiling
chiefgoat has joined #nixos
<{^_^}>
Use NIX_COUNT_CALLS=1 and/or NIX_SHOW_STATS=1 to profile Nix evaluation
bvdw has quit [Quit: bvdw]
bvdw has joined #nixos
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed commit from @offlinehacker to nixpkgs-19.03-darwin « facetimehd-firmware: 1.43_4 -> 1.43_5 »: https://git.io/Jewf1
<Thra11_>
I want to override `version` in an overlay. Do I have to manually override all the other bits which used the original version attribute, such as src urls, or can I make it use the new version wherever the overridden version was used?
<Thra11_>
symphorien: Ok. Thanks. Presumably if you did want to let other people just change the version of a package, you could make version an argument to the function instead (but you'd still have to override src hashes and so on, so maybe it's not as helpful as it first seems)
<symphorien>
yes
philr has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 3 commits to nixos-19.09-small: https://git.io/JewUP
<NoctisLabs>
Can someone point me to an example of adding a shell script to configuration.nix?
<EsperLily>
hmm ok, I'm surprised it has its own package at all
<gchristensen>
> :p pkgs.which.meta
<{^_^}>
{ available = true; description = "Shows the full path of (shell) commands"; homepage = "https://www.gnu.org/software/which/"; license = { fullName = "GNU General Public License v3.0 only"; shortName ...
<gchristensen>
which is also a shell builtin for some shells
<boogiewoogie>
https://hastebin.com/inunevipil.nix <-- it seems like it's trying and failing to build "facetimehd" and thus failing the system rebuild? I don't know that package, is anyone else having this problem? channel is 19.09
<clever>
EsperLily: bash can do the same thing, on its own
<EsperLily>
sure, just not with the name `which`
<clever>
boogiewoogie: its some firmware for apple hardware
<EsperLily>
I'm just looking into this because of a PR that adds `which` as a build input
<boogiewoogie>
clever: is it mandatory? seems weird to to build it on this non-apple machine :p
<clever>
boogiewoogie: theres already a pr somewhere to fix it
<NoctisLabs>
Thanks, I'll take a look,
<gchristensen>
boogiewoogie: it is turned on if you have enableAllFirmware set to true
<boogiewoogie>
clever: oh beautiful! almost expected that!
<Thra11_>
Is it possible you might need firmware for apple hardware if running linux on apple hardware?
<boogiewoogie>
gchristensen: ahh, I see! makes sense
<Thra11_>
(Not sure what sort of hardware we're talking about in this case)
<clever>
Thra11_: facetimehd is only needed if you want to use the facetime (webcam) on a macbook, i think
<samueldr>
iMac probably too
<Thra11_>
clever: So probably not commonly needed, but people do run linux on those?
LysergicDreams has quit [Ping timeout: 265 seconds]
<boogiewoogie>
clever: thanks for passing the info! not in a hurry though, so it's all good :) I'm trying to update without enableAllFirmware for now. can't even quite remember why I enabled it in the first place :p
<NoctisLabs>
@clever Thanks a lot for sharing this config, I feel like if I chew through this I can learn a lot.
philr has quit [Ping timeout: 240 seconds]
fusion809 has quit [Remote host closed the connection]
LysergicDreams has joined #nixos
abathur has joined #nixos
Soo_Slow has quit [Quit: Soo_Slow]
<EsperLily>
I forget, where can one find builds of the manual from the unstable channels?
<{^_^}>
[nixpkgs] @Ericson2314 opened pull request #73185 → gcc: Fix cross after I accidentally changed build target → https://git.io/JewTQ
<{^_^}>
[nixpkgs] @Ericson2314 merged pull request #73185 → gcc: Fix cross after I accidentally changed build target for 19.09 → https://git.io/JewTQ
_reanimus_ has quit [Quit: Logging off...]
<{^_^}>
[nixpkgs] @Ericson2314 pushed 2 commits to release-19.09: https://git.io/JewT5
xkapastel has joined #nixos
reallymemorable has quit [Quit: reallymemorable]
testy has joined #nixos
shreyansh_k has joined #nixos
<shreyansh_k>
Hello, can you confirm if nixops supports AWS's Paris (eu-west-3) region? I'm trying to use it but it keeps throwing the error: "Exception: invalid EC2 region ‘eu-west-3’".
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #73160 → pythonPackages.qrcode: Fix "No module named pkg_resources" error → https://git.io/JeVN7
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to release-19.09: https://git.io/JewTN
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #73159 → pythonPackages.qrcode: Fix "No module named pkg_resources" error → https://git.io/JeVNQ
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JewTA
silver_ has joined #nixos
shreyansh_k has quit [Quit: Konversation terminated!]
shreyansh_k has joined #nixos
laudecay has joined #nixos
zmacs has quit [Quit: Connection closed for inactivity]
CMCDragonkai has joined #nixos
<CMCDragonkai>
If I use a netboot image for a long period of time, where does the systemd journald logs go?
<gchristensen>
ram
testy has quit [Read error: Connection reset by peer]
<CMCDragonkai>
Do they just accumulate in memory? Is there an easy way to point to a network location in nixos config?
<CMCDragonkai>
Is there a limit to how many records it keeps?
testy has joined #nixos
<CMCDragonkai>
to prevent it eating up all the memory
testy has quit [Max SendQ exceeded]
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #73135 → nixosTests.lightdm: port to python → https://git.io/JeVMB
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JewTj
<{^_^}>
[nixpkgs] @worldofpeace merged pull request #73136 → nixosTests.sddm: port to python → https://git.io/JeVMg
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/Jewke
silver has quit [Ping timeout: 268 seconds]
mexisme has joined #nixos
jgt1 has joined #nixos
<{^_^}>
[nixpkgs] @flokli opened pull request #73186 → dpdk: build with meson, odp-dpdk: -> 1.22.0.0 → https://git.io/JewkL
<clever>
CMCDragonkai: `man journald.conf`
<clever>
CMCDragonkai: it will limit itself to keep a certain amount free, not use more then a certain amount (both absolute, and percentage), and gc automatically based on those params
<clever>
CMCDragonkai: it will not use more then 4gig for the logs, and not use more then 10% of the space on whatever fs is mounted to /var/log/journal
<clever>
and if thats a tmpfs, it will be limited to 50% of ram, so the journal is limited to 5% of ram
<{^_^}>
[nixpkgs] @svanderburg pushed to master « xcodeenv: bump default Xcode version to 11.1, default iOS SDK to 13.1. Add troubleshooting section to docs »: https://git.io/JewkO
<Thra11_>
Is it possible to apply an overlay when building an sd-image?
<clever>
Thra11_: set nixpkgs.overlays in the nixos config
is_null has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @svanderburg pushed to release-19.09 « xcodeenv: bump default Xcode version to 11.1, default iOS SDK to 13.1. Add troubleshooting section to docs »: https://git.io/Jewks
<otwieracz>
I am trying to "refresh" Bacula service in NixOS. I've got a case where, depending on configuration, multiple directories should be created.
__monty__ has quit [Quit: leaving]
<otwieracz>
What's expected way of ensuring those directories exist, with valid permissions?
<otwieracz>
Should I do this from the service definition implementation somehow, or rather attach this as pre script in systemd?
<otwieracz>
I'd bet for the later, but just wanted to make sure.
<clever>
otwieracz: if you set PermissionsStartOnly, then the User= only affects ExecStart, so ExecPreStart gets ran as root
<clever>
otwieracz: you can then freely use mkdir and chown to prepare things
<otwieracz>
But that's again in pre-script.
<otwieracz>
Not "by nixos" somehow.
<clever>
otwieracz: the "new" way is to use systemd tempfiles, but that name feels backwards and ive not looked into how to do it yet
Okinan has joined #nixos
<otwieracz>
tmpfile?
<otwieracz>
It wasn't called something something session?
<clever>
otwieracz: the way i personally do it, is i just define a users.users.foo, give it a home dir, and tell nixos to createHome = true;
<otwieracz>
uhm
<otwieracz>
hmm
<otwieracz>
Maybe.. But maybe not. I am not sure yet :)
<otwieracz>
Either way, thank you!
<otwieracz>
That might be useful approach, with createHome.
<otwieracz>
I've got one more question.
<otwieracz>
I am changing Bacula "workflow" right now - it will, at least in my opinion, reassemble `samba` approach where nix maps ({ "foo bar" = true; } are used to generate valid config instead of current apprach, where there are few predefined options + general reliance on `extraConfig` being huge multiline string.
<otwieracz>
But this will be breaking change.
<otwieracz>
What's the policy about such changes, assuming I'd like to have them merged to upstreaM?
civodul has quit [Quit: ERC (IRC client for Emacs 26.3)]
<otwieracz>
And sorry, but I've got one more question :) How can I merge two maps in nixos?
<elvishjerricco>
How would I go about finding the PCI device for the USB controller containing my mouse and keyboard
<elvishjerricco>
?
eightyone has quit [Quit: "]
<niso>
elvishjerricco: i'd use lsusb
<elvishjerricco>
niso: Not sure how to go from the lsusb output to a PCI device
phaebz` has joined #nixos
phaebz` has quit [Remote host closed the connection]
<clever>
00:16.2 USB controller: Advanced Micro Devices, Inc. [AMD/ATI] SB7x0/SB8x0/SB9x0 USB EHCI Controller (prog-if 20 [EHCI])
<clever>
Kernel driver in use: ehci-pci
<clever>
elvishjerricco: with this, i can figure out which driver is using a given card
<{^_^}>
[nixpkgs] @blitz opened pull request #73188 → spike: init at 1.0.0 → https://git.io/JewIm
laudecay has quit [Ping timeout: 265 seconds]
<clever>
--w------- 1 root root 4096 Nov 10 19:47 /sys/bus/pci/drivers/ehci-pci/unbind
<clever>
elvishjerricco: and i think if you `echo 00:16.2` to this file, it will just detatch the driver from the card, causing it was doing to stop (and all usb devices to unplug)
<clever>
elvishjerricco: you can use the bind find to re-connect things
<clever>
just note, if you disconnect the usb card from the usb driver, you loose your keyboard, so test over ssh!
phaebz has quit [Ping timeout: 240 seconds]
laudecay has joined #nixos
<elvishjerricco>
clever: I'm not really following. How did you go from lsusb output to identifying which PCI device is behind a specific USB device?
<clever>
elvishjerricco: i dont know of a way to directly do that
<clever>
elvishjerricco: but the above steps let you basically disconnect any pci device from its driver
<clever>
elvishjerricco: causing whatever usb devices are behind it, to vanish
<red[m]>
So, I'm trying to write a program in pony which links to the termbox C library. So, I'm doing nix-shell -p ponyc termbox
<red[m]>
So nix downloads the library - but I'm not entirely sure how to make my program "see" the termbox library
cosimone has quit [Quit: Quit.]
<red[m]>
now in code, I can specify a path... but obviously hard-coding a nix store path is a bad idea. Can anyone help me understand how I should go about addressing this is nixos?
MichaelRaskin has left #nixos [#nixos]
<clever>
red[m]: nix automatically adds a -L flag for you
<clever>
red[m]: so you just use the normal -ltermbox you always use
Ariakenom has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @nixos-channel-bot pushed 3 commits to nixpkgs-19.09-darwin: https://git.io/JewUP
<red[m]>
clever: apparently it isn't. Can you tell me the mechanism that it should work via so I can maybe debug it?
<clever>
red[m]: its tied into cc-wrapper, and only impacts gcc, so ponyc may ignore it
<red[m]>
oh wait - found the issue
shibboleth has quit [Quit: shibboleth]
<clever>
red[m]: $NIX_LDFLAGS and $NIX_CFLAGS_COMPILE
<red[m]>
apparently, if I run tmux inside of a nix-shell it doesn't work
<red[m]>
but if I run nix-shell inside of tmux, it does