<eacameron>
Anyone have an idea why starting the default display manager doesn't do anything? Booting up new install takes me to terminal login and starting the service manually does nothing
sbdchd has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @Infinisil pushed 2 commits to release-18.03: https://git.io/fA1lQ
<clever>
all attributes in config are also in options
<clever>
you appear to have a systemd mising
<steveeJ>
I directly copied your line, the error remains
<clever>
id just use what i linked above from kexec then
<steveeJ>
agreed
<clever>
multi-user.target is the way nearly every single service is ran on bootup
<haslersn_>
What can I do if I want to contribute to a GPL-licensed project but I don't want the code which I contribute to be GPL-licensed?
<haslersn_>
In particular: I thought about adding nix completions to fish
<emily>
haslersn_: you can license your contributions under a license that is compatible with but more permissive than the GPL
<emily>
haslersn_: e.g., you could use the CC Zero declaration to make them (effectively) public domain.
<emily>
haslersn_: well... more specifically you can hopefully release whatever portions of your contributions you have copyright interest in
<emily>
it's a bit complicated because I guess the changes themselves are likely to count as a derivative work
<emily>
(at least if they're patches to existing code -- greenfield code is more likely to be all your copyright)
<haslersn_>
As far as I can tell now, the contribution will just be a single nix.fish file which I write from scratch
<jasongrossman>
haslersn_: I would love nix completions for fish. Thank you.
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fA1lh
<steveeJ>
thanks clever, sshd works now :)
<emily>
right. so you can probably license it whatever you want, and they can incorporate it so long as that license is GPL-compatible.
<emily>
you can also dual license it under GPL (for incorporation into fish) and whatever other license you want (for incorporation into other projects).
<haslersn_>
An alternative would be to make it a standalone-project with its own nix package
<jasongrossman>
haslersn_: If you don't mind, it would be ideal to package it with both fish and nix, to advertise each to the other.
<haslersn_>
You mean, the completions package has fish as a dependency which comes with the completions as well, so you have the completions twice?
kreetx has joined #nixos
<jasongrossman>
haslersn_: OK, on second thougts I was being idiotic. Of course you want to package it with fish and then nix will pull it in with fish.
kreetx has quit [Ping timeout: 252 seconds]
<haslersn_>
I would package it with both, but for another reason: you can use the standalone-package if it's not yet merged into fish-shell
<haslersn_>
The standalone one would then probably be called fish-nix-completions-unstable
<jasongrossman>
haslersn_: Yes, good point, especially since we'll need to wait (b) for the fish people to package it and (c) for nix to package the new version of fish.
<jasongrossman>
haslersn_: And (a) for you to write it. :-)
wpcarro has quit [Ping timeout: 272 seconds]
mekeor has joined #nixos
<haslersn_>
Can I just distribute my code under both, MIT and GPL license? That way, they (fish-shell) wouldn't refuse to pull it in and I could still make the standalone project MIT licensed
Akii has joined #nixos
redj has quit [Read error: Connection reset by peer]
hamishmack has joined #nixos
<{^_^}>
[nixpkgs] @Ericson2314 pushed commit from @peti to release-18.09 « haskell-stm: update overrides that use version 2.4.5.0 to version 2.4.5.1 »: https://git.io/fA14f
benmachine has quit [Read error: Connection reset by peer]
silver_ has quit [Read error: Connection reset by peer]
<adisbladis>
haslersn_: What's so bad about GPL licensing you are trying to avoid it? Keeping the same license as upstream is pretty much always preferable.
<joepie91>
haslersn_: if it's your own code, then yes
<joepie91>
(licensing-wise anyway)
<cement>
except for one specific GPL license type, the rest state that you can't use GPL licensed stuff inside of non-GPL'd stuff
<adisbladis>
It's unlikely any upstream will accept your changes if you start doing license weirdness.
<joepie91>
dual-licensing is hardly 'license weirdness', it's a fairly common practice
<haslersn_>
adisbladis yes, but it's not weirdness if I silently distribute it under MIT
<adisbladis>
joepie91: Currently all fish completions are GPL
<haslersn_>
The pull request would not contain the MIT license
<adisbladis>
haslersn_: No you are free to do that if you _also_ contribute it under GPL
<joepie91>
.... ok?
aanderse has quit [Quit: aanderse]
<cement>
I'm getting a gtk error whenever I try to open pulse audio control
blankhart has quit [Ping timeout: 252 seconds]
<haslersn_>
adisbladis: Yes, I would be willing to do that.
kreetx has joined #nixos
<adisbladis>
joepie91: I was under the impression that haslersn_ was going to try to upstream under dual licenses.
<adisbladis>
haslersn_: Then by all means :) Fish completions would be great \o/
aanderse has joined #nixos
<adisbladis>
haslersn_: Are you aware that fish can generate completions from manpages? That can be a good starting point
<haslersn_>
Yes but are there man pages for the nix commands?
<cement>
lolno
blankhart has joined #nixos
<cement>
docs on nix are pretty terrible
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 is now known as Supersonic
kreetx has quit [Ping timeout: 244 seconds]
<haslersn_>
Also, the commands nix-env, nix-build and nix-shell aren't consistently designed
lassulus_ has joined #nixos
<joepie91>
adisbladis: ah, in that sense.
<joepie91>
dunno, I'd personally not really care (from the perspective of *being* the upstream) so long as the project license was in the list somewhere
endformationage has joined #nixos
lassulus has quit [Ping timeout: 252 seconds]
lassulus_ is now known as lassulus
<adisbladis>
I guess I'm just being a bit sensitive on the topic.. Have seen so much licensing stupidity at work recently
<joepie91>
adisbladis: where do you work that this is a common topic :P
<adisbladis>
joepie91: In the magical world of blockchains
<adisbladis>
The most "fantastic" example is a white paper for a client project where the term "open source" is redefined..
<cement>
can someone help me with my problem? I tried to open pulseaudio's volume controls using "pavucontrol" and got a gtk error
<adisbladis>
It's super hacky, but at least you can work around the issue
<adisbladis>
pavucontrol should probably set GDK_PIXBUF_MODULE_FILE in it's wrapper
<cement>
hm
<joepie91>
adisbladis: I vaguely recall one particular project (non-blockchain) that tried to define 'open-source' as 'if you e-mail us and sign an NDA, you will get insight into the source'...
<cement>
I enable pulse audio through my configuration
<joepie91>
alongside the usual smattering of projects that claim to be 'open-source' but then exclude particular usages in their nonstandard license, of course
<cement>
not sure if that actually includes pavucontrol
<adisbladis>
cement: pavucontrol is not in pulseaudio, it's a standalone thing
jbaum98 has quit [Quit: Connection closed for inactivity]
<cement>
right, ok
<adisbladis>
joepie91: "Open source: A source that is made thoroughly accessible for the public"
<cement>
then it's probably pointing somewhere weird
<adisbladis>
So yeah, not really open source at all
<adisbladis>
cement: Did that command work for you?
<cement>
it did, thanks
<cement>
probably need to re-nix-env pavucontrol
<joepie91>
adisbladis: right :P
<cement>
right, what was I thinking, reinstall is never the solution with nix
init_6 has joined #nixos
<cement>
bleh... I don't wanna do the change to the wrapper... but I should because ecosystem
<catern>
does anyone know of a Kubernetes controller that some core Nix committer was developing to integrate k8s with Nix? It was still in development when the person mentioned it here; you attached labels to pods which described a Nix store path (or maybe a package?) and it would deploy those store paths to the hosts when the pod was deployed there
<{^_^}>
[nixpkgs] @vbgl opened pull request #46726 → ocaml ≤ 4.05: mark as broken on Aarch64 → https://git.io/fA1Rs
worldofpeace has quit [Ping timeout: 246 seconds]
patrl has quit [Ping timeout: 244 seconds]
<__red__>
Quick Question - I'm adding a package to nixpkgs and it needs to be a github revision instead of a release. If I use the revision in mix-env -qa it looks like this:
<__red__>
is it appropriate to shorten that version?
<adisbladis>
__red__: It's common to add a human readable date rather than a rev
<adisbladis>
Something like "openspin-unstable-2018-09-16"
<__red__>
date as in the date I make the package, or the date that the last checkin was made to their repo?
<adisbladis>
__red__: Last commit date
<__red__>
rgr
<__red__>
is the "unstable" mandatory? I mean I could go back to their last release (2 years older) but I would have to patch the crap out of their buildsystem
<__red__>
or I guess I could start with unstable and then bug them to do a release
<adisbladis>
__red__: It's not mandatory, just a relatively common practice
<adisbladis>
__red__: Commit sha is fine too if that's what you want
jmeredith has quit [Quit: Connection closed for inactivity]
<adisbladis>
__red__: Stable releases are time-based (twice a year) and will get conservative bugfixes/security upgrades.
<adisbladis>
__red__: Unstable packages will end up in a "stable" release in the next release (when the branch-off point from master happens)
<__red__>
is the -unstable considered a part of the packagename or a part of the version#
<__red__>
?
<__red__>
openspin-unstable-20180212
<__red__>
for example
martingale has joined #nixos
kreetx has joined #nixos
martingale has quit [Ping timeout: 245 seconds]
JonReed has quit [Ping timeout: 252 seconds]
<adisbladis>
__red__: A version attribute is not required. Packages names/versions are parsed from the name attribute using builtins.parseDrvName. What I would do is https://ptpb.pw/fRQH/nix
jbo has quit [Ping timeout: 264 seconds]
blankhart has joined #nixos
kreetx has quit [Ping timeout: 272 seconds]
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
work_ has joined #nixos
Kelppo has quit [Ping timeout: 272 seconds]
blanko5 has joined #nixos
<blanko5>
Аllɑh iѕ doіnɡ
blanko5 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<ldlework>
lest we forget
<adisbladis>
They never get to the part _what_ is being done... The suspense is killing me.
<ldlework>
adisbladis: your mom
* ldlework
badum tish!
<ldlework>
ldlework--
<adisbladis>
¯\_(ツ)_/¯ their loss
<emily>
adisbladis: the whole rant is actually pretty funny.
counting1ort has joined #nixos
<emily>
at one point it says "filesystem does not write without permission of" -- actually, I feel like completing it would be a bad idea.
countingsort has quit [Ping timeout: 252 seconds]
<haslersn_>
Is there a way to tell if I'm inside 'nix run' ?
<{^_^}>
nix#2020 (by ElvishJerricco, 24 weeks ago, closed): nix run doesn't set IN_NIX_SHELL
brejoc has joined #nixos
<elvishjerricco>
haslersn_: I just use SHLVL
<elvishjerricco>
Works with more things that way
<haslersn_>
I have a program that prints the loaded packages when inside nix shell. I guess for supporting "nix run", I can work around the issue
orivej has quit [Ping timeout: 244 seconds]
tomberek has joined #nixos
kreetx has joined #nixos
martingale has joined #nixos
kreetx has quit [Ping timeout: 252 seconds]
endformationage has quit [Quit: WeeChat 1.9.1]
martingale has quit [Ping timeout: 252 seconds]
Kelppo has joined #nixos
Kelppo_ has joined #nixos
init_6 has quit [Ping timeout: 244 seconds]
init_6_ has joined #nixos
cement has quit [Ping timeout: 245 seconds]
Fare has quit [Ping timeout: 245 seconds]
Fare has joined #nixos
Kelppo has quit [Read error: Connection reset by peer]
Kelppo_ has quit []
pyvpx has quit [Ping timeout: 244 seconds]
cement has joined #nixos
<{^_^}>
[nixpkgs] @vbgl opened pull request #46727 → ocamlPackages.odoc: init at 1.2.0 → https://git.io/fA1Ey
pyvpx has joined #nixos
rauno has joined #nixos
<kandinski>
Hi, I've tracked down a bug both in 18.09 (I think, I don't know whether it upgraded well) and 18.03 (for sure) that is not present in Debian Stretch (latest). In order to post the issue on the tracker, I'd like to find the versions of network-manager, network-manager-openvpn and network-manager-openvpn-gnome that I'm running.
<kandinski>
Hmm, of course, reading the source on the release-18.03 branch. Thanks, rubber duck!
kreetx has joined #nixos
kreetx has quit [Ping timeout: 252 seconds]
cross has quit [Ping timeout: 272 seconds]
cross has joined #nixos
Fare has quit [Ping timeout: 272 seconds]
<{^_^}>
[nixpkgs] @samdroid-apps opened pull request #46728 → vim-tsx: init at 2017-03-16 → https://git.io/fA1uc
martingale has joined #nixos
<sphalerite>
kandinski: you can also work out which versions your system depends on by running nix-store -qR /run/current-system | grep ...
martingale has quit [Ping timeout: 252 seconds]
sb0 has joined #nixos
sir_guy_carleton has joined #nixos
jasongrossman has quit [Ping timeout: 252 seconds]
blankhart has quit [Quit: WeeChat 1.9.1]
spear2 has quit [Ping timeout: 240 seconds]
<kandinski>
sphalerite: I've nailed down the issue, but I don't know whether it's NixOS's or upstream's. Is it ok to submit it to the nixpkgs tracke, and leave it for the maintainer to contact upstream?
<kandinski>
(if I were to bet, I'd bet it's ustream's)
<sphalerite>
kandinski: absolutely :)
xdroop4 has joined #nixos
<xdroop4>
Αllah ⅰѕ dⲟⅰᥒɡ
<xdroop4>
suᥒ ⅰs not doiᥒg Aⅼⅼah is doіnɡ
xdroop4 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<kandinski>
sphalerite: do you have any suggestion on how to improve this?
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @vbgl to release-18.09 « pythonPackages.libnacl: fix build on darwin »: https://git.io/fA12t
<sphalerite>
whose bright idea was it anyway to cause the system to go into emergency mode if local-fs.target fails when it was previously already reached >_>
<{^_^}>
[nixpkgs] @Mic92 merged pull request #46726 → ocaml ≤ 4.05: mark as broken on Aarch64 → https://git.io/fA1Rs
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
spear2 has quit [Ping timeout: 245 seconds]
martinga_ has joined #nixos
martinga_ has quit [Ping timeout: 246 seconds]
hamishmack has joined #nixos
<steveeJ>
has anyone come across `/nix/store/574xmi6n0ys0qdxk9q2lh7fk5ipjsg85-libkrb5-1.15.2/lib/libgssapi_krb5.so.2: undefined reference to `glob@GLIBC_2.27`` when building golang sources?
ambro718 has joined #nixos
thc202 has joined #nixos
<sphalerite>
Any particular reason nixos generates an /etc/fstab rather than systemd mount units?
<ambro718>
did anyone have success with Steam in Wine? I am getting an obscure "content servers unavailable" error when trying to download a game. There was a wine bug about some certificate verification but is fixed in the wine versions I tried (3.14, 3.16). I'm thinking it's somehing with NixOS packaging.
<Myrl-saki>
ambro718: I'm also having that with normal Steam.
<Myrl-saki>
s/normal/native/
hotfuzz_ has quit [Ping timeout: 252 seconds]
<ambro718>
Myrl-saki: I don't get that with native steam, I just started a download with no problems there.
<Myrl-saki>
I'm considering reinstalling Steam.
<Myrl-saki>
FWIW.
<Myrl-saki>
ambro718: Have you tried reinstalling Steam for your Wine env?
<ambro718>
Myrl-saki: I have once but I will try again with wine 3.16 (I updated the nix package manually).
<Myrl-saki>
ambro718: Please tell me if you get success with it.
sir_guy_carleton has quit [Quit: WeeChat 2.0]
weechat4 has joined #nixos
Thra11 has joined #nixos
mayhewluke has quit [Ping timeout: 272 seconds]
weechat4 has quit [Client Quit]
mayhewluke has joined #nixos
weechat4 has joined #nixos
<c15ade4>
ambro718: steam itself has their own version of wine
<ambro718>
Myrl-saki: nope, tried a clean wine install and also again in windows xp mode, same problem
<ambro718>
c15ade4: specifically intended for running steam and games?
<c15ade4>
yes
weechat4 is now known as mingc
<c15ade4>
they did an announcement about it
<ambro718>
though seems doubtful it would work in nixos
<ambro718>
I will try normal wine in an ubuntu vm and see if I also get this error. If I don't it must be a NixOS related issue.
kreetx has joined #nixos
<ambro718>
huh, though maybe the steam wine would work in the environment that nixos makes for native steam
<typetetris>
When I use data-files in a cabalfile and use cabal2nix, nix-build doesn't install those files. What do I need to do, to get that working?
<c15ade4>
ambro718: you just run the linux version of steam
<ambro718>
c15ade4: and it would run windows only games in wine?
<Myrl-saki>
ambro718: I've tried it and it works for me.
<sir_guy_carleton>
graphical.target reached after 20.015s in userspace
<sir_guy_carleton>
yu
<Myrl-saki>
sir_guy_carleton: FWIW, you're only booting for 22s.
<lesh>
hm, this might not be helpful to resolve the issue, but I'm wondering why whould graphical UI or anything like that depend on dhcp? dhcp could resolve after you have your desktop visible, that's the whole idea with tree dependencies and systemd no?
<Myrl-saki>
sir_guy_carleton: Firmware is BIOS/UEFI.
<Myrl-saki>
sir_guy_carleton: loader is bootloader.
<sir_guy_carleton>
oh, that's good
<lesh>
ah I see, you have graphical in 20 seconds, nvm :)
<sir_guy_carleton>
it's not really an issue, i just want to make sure the money i spent for an ssd was put to good use and there wasn't any particular problem with nixos
<Myrl-saki>
I personally think that this is the correct way.
<Myrl-saki>
When I'm logged in my system, it *better have everything setup*.
<sphalerite>
I think systemd-analyze critical-chain display-manager.service might be more relevant for perceived startup time though
<Myrl-saki>
On Arch Linux, I had problems since dhcpcd is not finished when I logged in on the system. Firefox started when my IP wasn't assigned yet and restarting it was the fastest way to get it working.
<lesh>
Myrl-saki: yes sounds ok, but then I wouldn't expect multi-user to block on dhcp (given it's a personal machine and not a server)
<Myrl-saki>
lesh: This is dhcpcd, not dhcpd.
<lesh>
Myrl-saki: yes, that's what I ment, block on dhcp resolution
<Myrl-saki>
lesh: Can you explain more on why it shouldn't?
<Myrl-saki>
sphalerite: That's weird. I have display-manager started, but X11 still doesn't output anything until dhcpcd is done.
<Myrl-saki>
Actually, scratch that.
<Myrl-saki>
I believe that's lightdm's behavior?
<lesh>
Myrl-saki: two reasons, you can work on your network connection (wifi/dhcp etc) in parallel with loading your desktop, and if there are any network issues, this shouldn't delay your desktop from loading
<Myrl-saki>
lesh: That's what I'm saying
<Myrl-saki>
05:59 <Myrl-saki> On Arch Linux, I had problems since dhcpcd is not finished when I logged in on the system. Firefox started when my IP wasn't assigned yet and restarting it was the fastest way to get it working.
<sphalerite>
that sounds like a reason to block display-manager on dhcpcd, not a reason *not* to
<Myrl-saki>
sphalerite: My latest message?
<sphalerite>
yes
<Myrl-saki>
sphalerite: Yeah, I'm saying that multi-user should wait for dhcpcd.
<sphalerite>
Myrl-saki: you mean display-manager?
<Myrl-saki>
sphalerite: Either should be okay.
Mateon1 has quit [Ping timeout: 252 seconds]
<sphalerite>
Myrl-saki: no, because multi-user doesn't actually mean running anything in particular
Mateon2 has joined #nixos
<sphalerite>
it just means "this set of units is done starting up"
<Myrl-saki>
True.
<lesh>
Myrl-saki: oh, I missed that line, yeah, I guess this is a personal preference depending on how you use your machine, I want to get my desktop unrelated to the network state, which seems also how this works traditionally on other os desktops
<sphalerite>
and display-manager doesn't wait for that
<Myrl-saki>
Hm
klntsky has quit [Remote host closed the connection]
georges-duperon has joined #nixos
<Myrl-saki>
lesh: Yeah, I agree! It's really preference in the end. The first thing that I start on my desktop is Firefox, and it didn't(doesn't) handle the IP resolution properly.
<Myrl-saki>
sphalerite: I'm actually a bit confused now. Hm.
Mateon2 is now known as Mateon1
sb0 has quit [Ping timeout: 244 seconds]
<Myrl-saki>
sphalerite: I mean, as with what lesh has said, the desktop should be unrelated to the network state, so display-manager should start parallel with dhcpcd, and this seems to be the default behavior.
<sphalerite>
Myrl-saki: right
<{^_^}>
[nixpkgs] @Mic92 pushed to master « go_1_11: add url to flaky test »: https://git.io/fA1wZ
<sphalerite>
Myrl-saki: what would be nicer to solve your firefox thing is to wait for network-online.target before starting *firefox*
<Myrl-saki>
sphalerite: But in my case, I get a black screen while dhcpcd is starting up, even though systemd-analyze said that display-manager started parallel to dhcpcd.
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @bobvanderlinden to release-18.09 « go: skip flaky os/exec TestExtraFiles »: https://git.io/fA1wC
<Myrl-saki>
This might be lightdm-only behavior though.
knupfer1 has joined #nixos
<lesh>
sphalerite: right! obviously, depend ff on network.target
<sphalerite>
I blame lightdm :p I don't know if this occurs with other display managers or not but lightdm is weird so I'm assuming it only occurs with lightdm :D
<lesh>
sphalerite: if I remember correctly, this works great in theory, until you want to depend on system target with a user target that would start ff, at least few years ago this wasn't possible
<Myrl-saki>
I won't be surprised if that's the case. :P
<Myrl-saki>
Is there something along the lines of systemd-wait? :P
<sphalerite>
lesh: nah I'm not thinking "make firefox a systemd unit", I'm thinking "wrap it with a script that waits for network to be online first"
knupfer has quit [Ping timeout: 252 seconds]
<sphalerite>
Myrl-saki: that's what I've justasked in #systemd ;)
<Myrl-saki>
There's systemd-notify
<Myrl-saki>
But it seems to be user -> systemd
<Myrl-saki>
sphalerite: Report back to us if you find success! :P
<sphalerite>
there's systemctl start --wait, but that requests a start and requires the appopriate permissions
<lesh>
sphalerite: hahaha right. even sudo within a systemd started script was an issue somehow, it was really nasty unfortunately, such an obviously useful usecase
Dedalo has joined #nixos
<rycee[m]>
sphalerite: Not sure if you are talking about a user service but from I know the `--wait` option doesn't work because the dbus user session is not set up the way systemd expects.
<rycee[m]>
(i.e. being socket activated)
orivej has quit [Ping timeout: 252 seconds]
<sphalerite>
rycee[m]: well this is for the system bus so it should be fine I think
<rycee[m]>
sphalerite: Right, then I agree that it is likely to work better :-)
sigmundv has joined #nixos
Dedalo has quit [Ping timeout: 252 seconds]
kreetx has joined #nixos
<sphalerite>
oh wait no, not --wait. --wait waits for the service to stop again
<Myrl-saki>
sphalerite: TBF, the correct behavior is "firefox shouldn't die because of this."
<Myrl-saki>
die being unusable, not die like process kill.
<Myrl-saki>
Anyways, I'll go disable dhcpcd on my laptop then reboot.
DigitalKiwi has joined #nixos
<steveeJ>
when I build it in a nix-shell with glibc_2_27 available the build works. is this a problem with the go toolchain or with the package?
<steveeJ>
however, the executable is dysfunctional and just prints this error: ./terraform-provider-libvirt: relocation error: /nix/store/9r71d59rdkidbbcmwqm686fyy7qfw6p9-glibc-2.27/lib/libc.so.6: symbol _dl_exception_create, version GLIBC_PRIVATE not defined in file ld-linux-x86-64.so.2 with link time reference
<Myrl-saki>
Interesting.
v0|d has joined #nixos
<Myrl-saki>
If I boot with dhcpcd.service disabled, I actually get thrown to the getty for a while.
<v0|d>
Hello, I need a little help w/ nvidia on x86-64. It get this: cudaGetDeviceCount returned 35 -> CUDA driver version is insufficient for CUDA runtime version Result = FAIL
tzemanovic has quit [Remote host closed the connection]
kreetx has joined #nixos
silver has joined #nixos
goibhniu has joined #nixos
kreetx has quit [Ping timeout: 252 seconds]
tzemanovic has joined #nixos
Kelppo has joined #nixos
<gchristensen>
anyone have a Canon printer working under CUPS?
stanibanani has joined #nixos
<srhb>
gchristensen: Welcome home, hope you had a nice vacation. :)
tzemanovic has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @volth opened pull request #46739 → fetchgit with leaveDotGit: increase determinism, fix hashes → https://git.io/fA167
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
rihards has joined #nixos
<infinisil>
tilpner: what is unfortunate?
sb0 has joined #nixos
<gchristensen>
srhb: yeah! quite good! lots of time without computers or cell phones or electricity
<srhb>
gchristensen: Glad to hear it. :)
<gchristensen>
hard to come back, especially when CUPS is what I come back to X)
<srhb>
:D
stanibanani has left #nixos [#nixos]
<sphalerite>
let us all take a moment to think about people less fortunate than ourselves, like gchristensen who currently has to deal with printing
jD91mZM2 has joined #nixos
<tilpner>
infinisil - The eval invocation syntax conflicts with blockquote syntax in some clients (IRCCloud, but possibly others)
<sphalerite>
tilpner: I daresay that's irccloud's fault :)
<gchristensen>
sphalerite: LOL
<tilpner>
That's one way to look at it
<infinisil>
Ah yeah, I think that's fine, happens occasionally
aminechikhaoui has quit [Read error: No route to host]
simondodsley_2 has joined #nixos
<simondodsley_2>
Αⅼⅼɑһ іѕ dоⅰnɡ
simondodsley_2 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<sphalerite>
gchristensen: you may have missed the latest spam fashion while you were away ^
<gchristensen>
oh?
<gchristensen>
oh no I saw that before :|
<srhb>
Yeah, we've suffered while you were away :P
martingale has joined #nixos
<infinisil>
Sigyn has been doing great work though
<sphalerite>
hm, running `runHook unpackPhase` in `nix-shell '<nixpkgs>' -A clang_6` complains about $src not pointing to the source and exits the shell. Especially the exiting bit is annoying, can I disable that somehow?
<teto>
is the https://nixos.wiki/wiki/Vim still up to date regarding how to add vim plugins ? seems like there is now an update.py instead (whic doesnt work here)
revtintin has joined #nixos
<sphalerite>
teto: the bit with knownPlugins?
<teto>
sphalerite: the "Adding new plugins" section
_ris has quit [Ping timeout: 260 seconds]
<teto>
if I run the command I get attribute `'pluginnames2nix' missing`
<das_j>
I've been long for a longer time, how are the 18.09 builds going?
mayhewluke has quit [Ping timeout: 252 seconds]
mayhewluke has joined #nixos
winem_ has quit [Ping timeout: 264 seconds]
Thra11 has quit [Ping timeout: 240 seconds]
<LnL>
teto: the generator was recently replaced, but maybe there are still some bugs
<azazel>
hi guys, I see that "sudo" really runs "/run/wrappers/bin/sudo". If I have to reference it from a systemd script, how I would do that?
<teto>
LnL: the new one can only be better in my point of view . If I call update.py from nixpkgs root, I get undefined variable 'supertab' at /home/teto/nixpkgs/pkgs/misc/vim-plugins/default.nix:419:25
<LnL>
I've not used it myself yet so don't know, ^ Mic92
tasse13 has joined #nixos
<tasse13>
Аⅼlah is dοⅰng
tasse13 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<sphalerite>
azazel: why would you need sudo in a systemd script? I'd recommend starting with the necessary privileges then dropping them rather than dropping them at the beginning then reescalating them…
blankhart has joined #nixos
<azazel>
sphalerite: it's to run a process as another user as part of a one time running "upgrade script", specifically I need to run posstgreSQL' psql as "postgres" user and run some php scripts (from nextcloud) as the nginx user
<sphalerite>
azazel: any reason not to set the unit's user to that user?
<azazel>
sphalerite: because the users are two, and I have to do some contextual chowns, which I need to run from root
<sphalerite>
azazel: hm since root doesn't need the setuid bit set, I guess you could just put it in the packages for the service
<sphalerite>
then just use sudo as it appears on $PATH in the unit
<azazel>
yes, I just did that, thanks
Dedalo has joined #nixos
martingale has quit [Ping timeout: 240 seconds]
<azazel>
sphalerite: btw, I did systemd.services.foo.path = [ pkgs.sudo ];
<sphalerite>
azazel: that's what I meant :D
haslersn has joined #nixos
<haslersn>
Hi, sometimes if my girlfriend restarts her display-manager or wakes the notebook from suspend, the fans get very load for a view seconds and the notebook then reboots. When doing that reboot, it shows warnings about the CPU temperature. In journalctl, there are no logs from the time of waking from suspend. What could be the reason for that to happen? (NixOS)
<srhb>
haslersn: Ugh, things like that are terrifying to debug, if it really is a failure to throttle the cpu causing an overheat shutdown.
<azazel>
haslersn: what model is her laptop, have you checked nixos-hardware if it has support for it?
<azazel>
haslersn: I did commit support for mine there and part of if was about controlling sleep states and CPU temperatures ( I own a Lenovo X1)
iyzsong has joined #nixos
<haslersn>
It's a Lenovo ThinkPad E480
<azazel>
never heard of a TP E480...mmm
<haslersn>
It's only sold to students
<haslersn>
Or also business, I think
<haslersn>
In nix-hardware, there are some ThinkPads that have a cpu-throttling fix
<Myrl-saki>
What's nix-hardware?
<jD91mZM2>
Speaking of hardware support, what hardware would you people recommend for buying a new laptop?
<Myrl-saki>
jD91mZM2: Anything not released in a year. :P
<azazel>
haslersn: mine has issues in Linux with cpu temperatures so I had to implement... that module that you just posted ;-)
<haslersn>
azazel: Yeah I saw your commit
<azazel>
there is the chance that yours suffers from that too... and maybe it has a bios similar to mine with sleep states optimized for windows
<azazel>
haslersn: what does running "dmesg | grep ACPI.*supports" says?
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
<haslersn>
with the . being a placeholder?
<teto>
How is one supposed to change the installation folder for cmake projects? Usually export out=$PWD/toto is enough but for cmake the value is $prefix instead of $out and it's being overwritten. aka cmake/setup-hook.sh: CMAKE_INSTALL_PREFIX="$prefix $cmakeFlags"
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
<azazel>
haslersn: no, it's just part of the regular expression, you have to execute it as it is
tzemanovic has quit [Remote host closed the connection]
hakujin has joined #nixos
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
tzemanovic has joined #nixos
tzemanovic has quit [Remote host closed the connection]
jD91mZM2 has quit [Quit: WeeChat 2.0]
<sphalerite>
teto: you mean for a nix-shell build?
<teto>
yep
jbo has joined #nixos
hakujin1 has joined #nixos
<teto>
apparently I can export dontAddPrefix but this is still an extra step
Anton-Latukha has quit [Ping timeout: 246 seconds]
<azazel>
haslersn: so, the cpu-throttling-bug.nix module can help keeping throttling low especially when you run cpu-intensive tasks, althought I dont't know if that will help with your resume issue... I'm looking it up on the 'net, but have found no similar cases to yours
<azazel>
which kernel are you running? Run "uname -a"
<haslersn>
Yeah I also tried to google it first
<haslersn>
Linux ayguenef-nixos 4.17.19 #1-NixOS SMP Fri Aug 24 11:07:17 UTC 2018 x86_64 GNU/Linux
<azazel>
It isn't an old one...
knupfer1 has quit [Remote host closed the connection]
knupfer has joined #nixos
<haslersn>
The problem is that in "journalctl -b -1 -e" the last thing is "kernel: PM: suspend entry (deep)", so no logs from before the crash
Ariakenom has quit [Read error: Connection reset by peer]
Ariakenom has joined #nixos
<haslersn>
azazel: However, thank you for your effert!
<haslersn>
effort
<azazel>
haslersn: You're welcome,maybe give the cpu throttling module a try
<haslersn>
How to include it? Or just copy and paste?
civodul has joined #nixos
endformationage has joined #nixos
<azazel>
haslersn: just copy it alongsite your configuration.nix add ./cpu-throttling-bug.nix
<azazel>
to the "imports" list in configuration.nix
_deepfire has quit [Remote host closed the connection]
<azazel>
haslersn: if you start using more modules from nixos-hardware the best thing to do is to clone the repo, copy the configuration and add a configuration for your laptop there, maybe submitting it as pull request when you see that it works well
copumpkin has joined #nixos
fendor has joined #nixos
Anton-Latukha has joined #nixos
fendor has quit [Client Quit]
<makefu>
hey, anybody uses hibernate (suspend-to-disk) on nixos and has a working config?
fendor has joined #nixos
waynr has joined #nixos
Anton-Latukha has quit [Remote host closed the connection]
<Aleksejs>
makefu: I use suspend/resume on my laptop on lid close/open
<teto>
Mic92: does the script generate all plugin derivations from scratch or only for the new/modified ones ? if I have to prefetch every plugin, this will be long xD
Thra11 has joined #nixos
<eliasp>
according to https://nixos.wiki/wiki/Nixpkgs/Create_and_debug_packages#Using_nix-shell_for_package_development I should be able to run `nix-shell path/to/cloned/nixpkgs -A bc` and then have a shell with `bc` available, right? executing `bc` in the shell just tells me to use `nix-env -iA nixos.bc` to install it, although it seems the required paths are fetched/copied when invoking the shell
<eliasp>
what am I doing/getting wrong about this?
<Myrl-saki>
eliasp: -p bc
<Myrl-saki>
eliasp: nix-shell actually just puts you in an environment to build the specified package.
<eliasp>
Myrl-saki: thanks a lot, that worked (although I got scared at 1st a bit by this warning: `warning: dumping very large path (> 256 MiB); this may run out of memory`) :)
<eliasp>
ok
<Myrl-saki>
eliasp: Internally -p, makes a derivation where the specified package is a build input.
<eliasp>
Myrl-saki: hmm, ok - but: it seems `-p bc` refers to the channel/`<nixpkgs>`-provided bc-version, not the one present in "path/to/nixpkgs" specified as 2nd arg, so I might rather invoke the shell with `-E` and point it towards the expression I want?
<Myrl-saki>
nix-shell -I nixpkgs=/path/to/nixpkgs -p bc
<Myrl-saki>
That one should work.
<eliasp>
Myrl-saki: fantastic, it worked!
<eliasp>
Myrl-saki: thanks a ton
<Myrl-saki>
No worries, just paying it forward. :P
<Myrl-saki>
Learned a lot from this channel.
<eliasp>
I have so much stuff I want to do/improve/update in nixpkgs, so I'm just trying to get comfortable with it up to a level where my contributions also make sense and don't end up being ugly hacks ;)
<jluttine>
any help how to write a package derivation for an application that is built with stack? https://github.com/carp-lang/Carp
<haslersn>
rather nix-shell -E '(import /path/to/nixpkgs {}).bc'
<Myrl-saki>
haslersn: That doesn't work. That gives you an environment to build bc.
<Myrl-saki>
haslersn: Is this in hackage or no?
<Myrl-saki>
jluttine: *
<jluttine>
Myrl-saki: at least can't find it there
<Myrl-saki>
Me neither.
<Myrl-saki>
Hm. Wait.
<Myrl-saki>
Does stackage mean "stable hackage" or "stack hackage"
periklis has joined #nixos
<Myrl-saki>
jluttine: FWIW, it has a cabal file.
<Myrl-saki>
So you could try to build it the cabal way.
periklis has quit [Remote host closed the connection]
<jluttine>
Myrl-saki: ok! any help how to write a package derivation for an application that is built with cabal? :)
ixxie has quit [Ping timeout: 246 seconds]
<jluttine>
any examples in nixpkgs i could look at?
<haslersn>
Myrl-saki: correct. Then: nix run -f /path/to/nixpkgs bc
<Myrl-saki>
There are, but I don't find them nice.
<eliasp>
Myrl-saki: hmm, it looks like `nix-shell -I path/to/nixpkgs -p $PKG` also just gives me the default version of $PKG, but not the one based on my modifications in nixpkgs
<eliasp>
haslersn: ah, will try this one
<Myrl-saki>
eliasp: nixpkgs=/path/to/nixpkgs
<Myrl-saki>
And yeah, nix run works.
endformationage has quit [Ping timeout: 272 seconds]
__lawlesseel__ has quit [Remote host closed the connection]
<d1rewolf>
is there a solution in nixos that which allows you to remotely share a currently running desktop?
lawlesseel has joined #nixos
etu1 has joined #nixos
<etu1>
Aⅼlaһ ⅰs ԁoing
etu1 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<{^_^}>
[nixpkgs] @rvolosatovs opened pull request #46751 → vimPlugins: vim-plug, fixes and updates → https://git.io/fA19m
kiloreux has quit [Ping timeout: 252 seconds]
brejoc has joined #nixos
<srhb>
Uh oh, now they're stealing our usernames.
<__red__>
So, I'm trying to test my package for the pull request using the sandbox but whenever I try to nix-build -A packagename it's just giving me the pre-compiled path
<d1rewolf>
what's with this "allah is doing"?
<__red__>
is there a way to nuke the version from /nix/store so I can re-compile it with sandboxing on
<__red__>
?
<Neo-->
d1rewolf, just about to asked the same question
<d1rewolf>
I presume it's a bot attack on freenode, but...why?
<d1rewolf>
and what exactly *IS* Allah doing?
<d1rewolf>
and to whom?
<d1rewolf>
;)
<__red__>
also nix.useSandbox=true - can I specify that on the commandline for nix-build?
<d1rewolf>
how can I see what files are contained within pkgs.i3?
<haslersn>
__red__: --sandbox I think
<srhb>
__red__: You can use --check to rebuild it
<srhb>
__red__: If you're using nix-build at least.
sbdchd has joined #nixos
<{^_^}>
[nixpkgs] @rvolosatovs opened pull request #46752 → Gotools etc. update, remove goimports → https://git.io/fA19w
<srhb>
__red__: I'm not sure whether enabling the sandbox from the command line really works, but I recommend having it enabled always.
sigmundv has quit [Ping timeout: 252 seconds]
<Neo-->
__red__, did you specify the path where your nixpkgs is? -I nixpkgs=/path/to/nixpkgs
<Neo-->
or -f /path/to/nixpkgs
<__red__>
so, it won't recompile
<__red__>
lookie
<__red__>
(2 lines)
<__red__>
[root@evil:~/projects/nixpkgs]# nix-build -A openspin --sandbox
<d1rewolf>
how does one find the path of a drv via the repl? for example, the drv which builds pkgs.i3?
<eliasp>
d1rewolf: are you looking for this? `nix path-info nixpkgs.i3` → `/nix/store/32ji24rc06sir3jzbjn0jp5j07b2hk8z-i3-4.15`?
<adisbladis>
d1rewolf: i3.drvPath ?
<teto>
well that new vim updater rocks.
ambro718 has joined #nixos
ris has joined #nixos
_ris has quit [Ping timeout: 252 seconds]
<d1rewolf>
eliasp: I think so. I was trying to determine the relative path of the i3 binary to use in string interpolation of pkgs.i3
<d1rewolf>
like {pkgs.i3}/bin/i3
<d1rewolf>
thanks guys
<adisbladis>
d1rewolf: Oh. You are not looking for the drv path but the output path?
<adisbladis>
You can do "${i3}" in the repl too
<d1rewolf>
adisbladis: ah, ok. great, thanks!
<{^_^}>
[nixpkgs] @redvers opened pull request #46753 → openspin: init at 2018-02-26 → https://git.io/fA19H
<ambro718>
There is something wrong with the Wine packaging on NixOS. When I install Steam using Wine 3.16-staging, I cannot download any game, I always get an error "content servers unreachable" like this: http://pasteall.org/pic/show.php?id=5288a1616ebc8046fc54333c717dfda7
<ambro718>
the same Wine version on Ubuntu does not have this problem.
<__red__>
yay - PR submitted, wish it luck :-)
<__red__>
and thanks for your help guys
orivej has quit [Ping timeout: 272 seconds]
jasonp10 has joined #nixos
<jasonp10>
Αⅼⅼaһ іѕ dοiᥒg
jasonp10 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<__red__>
hmm, failed on darwin. darwin stdenv not have g++ ?
aminechikhaoui has joined #nixos
georges-duperon has quit [Ping timeout: 252 seconds]
<srhb>
__red__: iirc you should not rely on having g++ directly, but some alias that is platform independent (if you think it'll otherwise build on other c++ compilers)
<srhb>
I forget the alias...
<Dezgeg>
gcc -> cc, g++ -> c++
<srhb>
Ah, simple, thanks Dezgeg
<__red__>
g++ is how the original packagers hard-coded it in the Makefiles :-/
ixxie has joined #nixos
<srhb>
__red__: time to employ substituteInPlace :-)
<eliasp>
when `nix-store --delete […]` tells me, the path can't be deleted because it's still alive, how do I found out why it is still alive? I'm pretty sure, it actually isn't and I could simply use `--ignore-liveness`, but I want to understand better the mechanisms in the background which determine the liveliness…
<__red__>
so do I delete my PR and try again?
<srhb>
__red__: No no, just add commits to the PR.. You can edit the title to mark it WIP
<sphalerite>
__red__: you can modify your existing PR
<__red__>
since I don't have darwin, can I still get the automated testing without ten thousand PRs being opened?
<__red__>
and anytime I push that branch - the PR will automatically update?
<srhb>
__red__: darwin builds are not automatic in the CI unfortunately. It happens when a trusted users triggers ofborg. Because sandboxing on OS X is not perfect, unfortunately.
<srhb>
__red__: Yep.
<eliasp>
adisbladis: LnL: thanks!
<adisbladis>
Use what LnL said. Deleting store paths manually is rarely a good idea :)
<teto>
eliasp: I ran a --delete once, that was not pretty, don't do it.
<__red__>
so, for the more experienced nix integrators here - this is the package's default Makefile
<srhb>
Going by line numbers can break easily on version bumps though
periklis has quit [Remote host closed the connection]
<adisbladis>
If you are going for perfection I'd make a proper patch
<adisbladis>
At least you can easily spot when a patch breaks
<__red__>
it may be easier to just write some patchfiles than sed -e '5,22dd' Makefile > Makefile.new ; mv Makefile.new Makefile
<srhb>
__red__: sed has an -i switch for in-place
<srhb>
But yeah, sure, patches are prettier
<__red__>
oh, I didn't know that
<__red__>
huh
<srhb>
I just never can be assed to create them.
<srhb>
Like, if only I understood the format well enough to write them by hand, but no...
* srhb
flails
<adisbladis>
srhb: git clone, make the change, git diff > whatever.patch
<srhb>
adisbladis: So much work! :-)
<adisbladis>
:3
ndrei has joined #nixos
ndrei has left #nixos [#nixos]
ndrei has joined #nixos
ndrei has quit [Client Quit]
<infinisil>
What would cause a program to exit with "[1] 12825 abort result/bin/vis"
ndrei has joined #nixos
ndrei has quit [Client Quit]
<infinisil>
In the strace i only see some
<infinisil>
open("/nix/store/yydnhs7migvlbl48wpsxan1yvq2icbr9-glibc-2.25-49/lib/locale/en.utf8/LC_TIME", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
<Dezgeg>
probably it calls abort();
<infinisil>
Can't find any aborts in the source code, the program is `cli-visualizer`
<Dezgeg>
try a debugger then
<infinisil>
I get this abort in all versions, and it seems to be impure, because when I fixed it some time ago in #30388, it worked for sure
<infinisil>
How could I compile something from gcc with debug symbols and stuff for gdb?
<infinisil>
Do i need to patch the makefile?
<sphalerite>
infinisil: make sure -g is passed to gcc
<sphalerite>
infinisil: for extra useful debug info, -g3 -ggdb
<infinisil>
Alright, but how do?
<sphalerite>
infinisil: NIX_CFLAGS? :D
<infinisil>
Ahh sweet
* infinisil
tries
<{^_^}>
[nixpkgs] @copumpkin opened pull request #46756 → top-level/release.nix: add patchShebangs test → https://git.io/fA1QO
<__red__>
thanks, found a second Makefile... working it
<infinisil>
"Reading symbols from result/bin/.vis-wrapped...(no debugging symbols found)...done."
<infinisil>
:/
<srhb>
__red__: By the way (just throwing ad-hoc information at you) it's usually best not to override standard phases like that. If someone were to add a patch to `patches` now, it wouldn't work because you've overwritten the automagic of that phase. Instead, something like preBuild would probably be better suited. :)
<srhb>
or even just postPatch if you think that phase is better suited.
revtintin has quit [Quit: WeeChat 1.9.1]
<infinisil>
Maybe it's NIX_CFLAGS_COMPILE
<__red__>
thanks - I'll absorb as many best practices as possible
<infinisil>
Or actually, it's g++ it's compiling with
<srhb>
__red__: Sure thing :)
<infinisil>
so maybe NIX_CXX_FLAGS_COMPILE or NIX_CXX_FLAGS
Peppard23 has joined #nixos
<Peppard23>
Aⅼlah iѕ ԁοⅰᥒg
Peppard23 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<__red__>
okay - did another push
<srhb>
__red__: Already triggered.
<__red__>
I really wish I could see the full build log
<__red__>
but my corporate filtering software won't let me hit the URL
<srhb>
__red__: :(
<infinisil>
Yeah I can't get debug symbols :/
<srhb>
__red__: Looks like it builds though
<__red__>
Yay!
<__red__>
okay, so I'm going to soft reset my HEAD, switch to postPatch and push again
<__red__>
brb
<infinisil>
Damnit gdb how do you work
<infinisil>
I'm 100% certain i compiled it with debugging info now
<infinisil>
But it still says it can't find any debugging symbols
<sphalerite>
infinisil: with nix?
<infinisil>
Yeah
<Dezgeg>
with dontStrip as well?
<infinisil>
Oh nope
<infinisil>
Ah yep that was it!
<infinisil>
Thanks
<infinisil>
Dezgeg: Now only problem: gdb can't find the source because the path in the sandbox was /build/<name>
<infinisil>
I guess i could copy the source to $out first and compile with that
knupfer has quit [Ping timeout: 240 seconds]
<Dezgeg>
or hope you figure out the problem without having the gdb showing the source lines
<Myrl-saki>
No wonder why I had a build that worked on 18.03 but not on 18.09...
smolboye has joined #nixos
i-am-the-slime has joined #nixos
<Myrl-saki>
FINALLY IT WORKS.
<Myrl-saki>
FFS.
<Myrl-saki>
infinisil: Thanks for pointing out sandbox and /build.
<infinisil>
Lol sure
<infinisil>
Hah, I fixed it by changing some C++ code I have no idea about
<infinisil>
Changed that line from
<infinisil>
loc = std::locale{loc, "", std::locale::ctype};
<infinisil>
loc = std::locale();
<infinisil>
To that
<__red__>
once a PR is complete, I guess I just switch back to my master and leave the old branch alone right?
<__red__>
I don't delete it or do anything else to it
<Ralith>
if the PR got merged then you might as well delete it
<__red__>
hmm, so if I wanted to nixos-rebuild switch from my local nixpkgs... I figured there'd be a flag for that like -f? I don't see it in the man page
<srhb>
__red__: -I nixpkgs=path/to/checkout
<srhb>
__red__: Which is the most universal "my nixpkgs lives here..." flag. It essentially prepends the argument to $NIX_PATH before passing it to the command.
<__red__>
thanks :-)
sigmundv has joined #nixos
<__red__>
time to start on my next package :-D
<srhb>
__red__: woo :)
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
<infinisil>
So, the problem is certainly something about the program not finding such locale files
<infinisil>
It tries to open /nix/store/yydnhs7migvlbl48wpsxan1yvq2icbr9-glibc-2.25-49/lib/locale/en_US.UTF8/LC_TIME
<infinisil>
which doesn't exist
<infinisil>
It also reads from /run/current-system/sw/lib/locale/locale-archive
<infinisil>
which I have no idea what it's for, but this seems to be the source of the impurity
<infinisil>
Dezgeg: Would you perhaps have an idea?
<Dezgeg>
yes there is a locale-related ticket open
<__red__>
Okay - question. There is a massive SDK for a the parallax propeller which is the only codebase that contains the code-flashing sources. The official docs tell you how to compile ONLY that loader code so the package is designed to either output EVERYTHING, or just the loader.
<__red__>
so, it would seem to me to be reasonable to create a package for just the loader, but it seems like seroius overkill to download a couple of hundred meg of sourcecode to just grab 4 files (2x .c, 2x .h)
<__red__>
(their documentation actually tells you to wget the .h and .c files from github and manually compile them)
martingale has quit [Ping timeout: 252 seconds]
Fare has joined #nixos
<andi->
4x fetchurl?
<__red__>
yup
<__red__>
I can do all that - no problem
<__red__>
and link to the docs as the maintainers will rightly say "wtf?"
<__red__>
so - if I go ahead and add the graphical IDE package later, I would make that smaller package a dependancy
<__red__>
and them modify it to not compile that app so it doesn't 'collide'
<steveeJ>
it was intentional for opening discussions about it
<{^_^}>
[nixpkgs] @demyanrogozhin opened pull request #46760 → particl-qt: init at 0.16.2 → https://git.io/fA1db
<steveeJ>
besides, are we supposed to PR against the nixpkgs-channels repo at all?
<samueldr>
nope, against nipkgs only
<samueldr>
well, nixpkgs, nixpkgs isn't one
* samueldr
can't even mistype a typo
Fare has quit [Ping timeout: 244 seconds]
jD91mZM2 has quit [Quit: WeeChat 2.0]
<samueldr>
since your fork isn't against nixpkgs proper, it may not be possible to directly open a PR? you may need to fork nixpkgs, add it as a remote, then push your branch (unless there are cleaner github tricks)
<steveeJ>
samueldr: from that repo I could only PR against the nixpkgs-channels repo
<emily>
it's a real slow-burner, I wish the spam killers let it stay around long enough to get to the funny ones.
Anton-Latukha has joined #nixos
<ambro718>
Wine (wow64) is quite broken on NixOS, more so than it seems. The registry is missing all the COM class registrations for 32bit that should be in HKEY_CLASSES_ROOT\Wow6432Node\CLSID. I think the problem is an error during wineprefix creation because "wine" is a shell script wrapper: /nix/store/5qmynzrm1xc8z9qbdf3c7ay08kynqqz0-wine-wow-3.16-staging/bin/wine: not an ELF binary... don't know how to load it
<ambro718>
I think because wine is failing to load the wine executable here (the 32-bit executable), the COM classes are not populated.
kreetx has quit [Ping timeout: 240 seconds]
<ambro718>
the 64bit executable (wine64) is not wrapped and the COM classes are populated fine for 64-bit.
<{^_^}>
[nixpkgs] @matthewbauer pushed to staging-18.09 « Revert "Merge pull request #46108 from obsidiansystems/cross-patch-shebangs" »: https://git.io/fA1bN
martingale has quit [Ping timeout: 252 seconds]
fenedor has joined #nixos
<rawtaz>
ambro718: nice debugging there. if you look at the derivation, can you find a place to change something in it so that it tries to run it as a shell script instead?
<ambro718>
rawtaz: the derivation wraps wine if gstreamer support is enabled, that's easy to see. I'll have to look what wine itself is trying to do. But I suspect it will need to find an executable.
<rawtaz>
well i dont know that stuff (yet) but surely whatever is calling the binary should be able to call a shell script?
<ambro718>
perhaps it's not trying to call it but do something more interseting with it, I'll check right now
<ambro718>
we will need to fix this so that wine finds an executable
<ambro718>
but I'll first build a wine without this wrapper to confirm the hypothesis
jimki20 has joined #nixos
<jimki20>
Allaһ ⅰѕ dοіng
jimki20 has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<rawtaz>
ambro718: great :))
<rawtaz>
ambro718: ok, i misunderstood, thought it was something in Nix that called a file that was a shell script, wondered why that was a problem.
<ambro718>
nope, the derivation wraps the "wine" executable which wine doesn't like because it actually needs an ELF file there to do funky things with.
Anton-Latukha has quit [Read error: Connection reset by peer]
monotux has quit [Quit: prutt]
<rawtaz>
silly wine
Anton-Latukha has joined #nixos
<ambro718>
I trust it has good reasons, binary compatibility layers are hard :)
monotux has joined #nixos
Ariakenom_ has joined #nixos
teser has joined #nixos
Ariakenom has quit [Ping timeout: 240 seconds]
<teser>
hi, how can I enable the rtl8814au module in order to be able to use my wireless card. I've put it into the "boot.kernelModules = [ ... "rtl8814au" ];", but the built kernel contains the 8814au.ko module; if I reboot my computer now, will the module be enabled and working?
<ambro718>
or maybe not, perhaps it's fine if it always points to "wine" executable, judging from the content of the wrapper scripts like winecfg.
<catern>
does anyone know of a Kubernetes controller that some core Nix committer was developing to integrate k8s with Nix? It was still in development when the person mentioned it here; you attached labels to pods which described a Nix store path (or maybe a Nix derivation?) and it would deploy those store paths to the hosts when the pod was deployed there
kreetx has quit [Ping timeout: 252 seconds]
martingale has quit [Ping timeout: 272 seconds]
<avn>
ambro718: I use wine in nixos a lot, but never used gtreamer with it.
<ambro718>
avn: I just picked the "full" package in order to make sure it has everything, but turner out it's broken because gstreamer is enabled.
<avn>
ambro718: and I used 32bit wine only
simukis has quit [Quit: simukis]
<sphalerite>
teser: yes, it should
<sphalerite>
teser: as long as the module does indeed support the wireless card and reports this correctly
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @xeji to release-18.09 « pythonPackages.dendropy: fix build »: https://git.io/fA1xu
martingale has joined #nixos
tzemanovic has joined #nixos
tzemanovic has quit [Client Quit]
<waynr>
okay i've installed a nix-info using nix-env -i but it just displays 'system: "x86_64-linux",' and seems to be hanging
<rawtaz>
installed? i never had to install it, it was in the system by default :)
<waynr>
yeah i think my multi user installation is just f'd
<infinisil>
rawtaz: It's only installed by default in recent nixos versions
<infinisil>
It's rather new
<waynr>
i installed nix on this computer a couple months ago
<waynr>
well, on all my computers a couple months ago...i don't use this particular computer very often
<rawtaz>
infinisil: i see :)
CuriousErnestBro has left #nixos ["Leaving"]
<teser>
sphalerite, I'm using the alfa AWUS036ACH wifi card, which is not directly supported, but I would like to test whether the existing module works. However, is there a list of supported cards for the existing module, does it support my card?
spear2 has quit [Read error: Connection reset by peer]
<sphalerite>
teser: the module's source will contain a dfeinition of what it supports — usb and/or pci ids as appropriate. I think there's a way through sysfs to bind a driver to another device disregarding what the driver claims to support, but I can't guarantee that that won't also melt your cat or whatever
Czen has quit [Remote host closed the connection]
Czen has joined #nixos
ambro718 has quit [Quit: Konversation terminated!]
<gchristensen>
would it be annoying if search.nix.gsc.io included the linux kernel?
<sphalerite>
seems like an odd addition? But I wouldn't mind
<samueldr>
gchristensen: do you know if you can filter it out by default? or would that negate the idea of adding it?
<teser>
sphalerite, where is the list of usb/pci IDs that it supports - https://github.com/zebulon2/rtl8814au - I was looking at it, but there's too much code to find it without knowing what to look for?
<Dezgeg>
searching for MODULE_DEVICE_TABLE usually helps
graphene has quit [Remote host closed the connection]
fenedor has quit [Read error: Connection reset by peer]
graphene has joined #nixos
teser has joined #nixos
<teser>
sphalerit, Dezgeg I've rebooted my machine and now the module is loaded into the kernel and the device is recognized by the lsubs, however there is no interface for it. For example, the lsusb reports the following:
<teser>
Bus 001 Device 008: ID 0bda:8812 Realtek Semiconductor Corp. RTL8812AU 802.11a/b/g/n/ac 2T2R DB WLAN Adapter
tzemanovic has quit [Remote host closed the connection]
<__red__>
if I'm packaging up another application that has as a dependency something I already submitted in the PR - I have to wait until it's been approved before I build the next package
<__red__>
otherwise the automated build testing is going to fail.
<__red__>
as it won't be able to resolve the dependency
<sphalerite>
__red__: in cases like that I usually put them both in one PR (in separate commits)
<__red__>
oh well, too late :-)
<__red__>
I can still work on it, just can't raise the new PR until the first is committed I guess. No big.
<__red__>
Learning curve ;-)
teser has quit [Ping timeout: 246 seconds]
sbdchd has joined #nixos
<{^_^}>
[nixpkgs] @xeji opened pull request #46761 → gtk-doc: don't build with dblatex by default → https://git.io/fAMeR
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @pacien to release-18.09 « exim: parametrise package »: https://git.io/fAMez
mayhewluke has quit [Ping timeout: 240 seconds]
mayhewluke has joined #nixos
kreetx has joined #nixos
thc202 has quit [Ping timeout: 240 seconds]
Kelppo has joined #nixos
<Myrl-saki>
How to disable sandbox for a derivation?