<gchristensen>
it is fine, but first we should open an issue describing the problem and get the original commiter involved (unless the breakage is significantly more severe than 50+ packages)
andreas303 has quit [Remote host closed the connection]
<Ankhers>
Would anyone know why I would not be able to find servers on my home network? From other machines I can just ssh / ping / whatever to `foo.local'.
ixxie has quit [Ping timeout: 268 seconds]
<ghutzriop>
hi, I'm trying to update a package to the current github master branch using an overlay( https://pastebin.com/qGYb8zLk ), but I'm getting an error just from having this in ~/.config/nixpkgs/overlays. the error is "undefined variable 'fetchFromGitHub'"
<ghutzriop>
gchristensen, thanks. but now I'm getting a "value is a function while a set was expected, at /nix/var/nix/profiles/per-user/root/channels/nixos/lib/modules.nix:141:29"
<worldofpeace>
I really need to follow up on my own PRs better :D
<genesis>
worldofpeace : would not be so difficult to write a desktop entry parser, i've some difficulty to get start because of regex
<genesis>
what do you think ?
<genesis>
just i say the desktop-file-edit could be a tempory workaround, but the input set for the edit function would be what we need to let unchanged
<worldofpeace>
genesis: I avoid regex because it makes my days... longer
<genesis>
^^
<qyliss>
lol
<worldofpeace>
lovesegfault: I just tried it testing wise
<genesis>
a few regex to parse and validate the desktop file seems to be the better solution btw;
<worldofpeace>
trace: warning: The option `services.xserver.displayManager.extraSessionFilesPackages' defined in `/home/worldofpeace/Code/nix/vms/gnome-vm/configuration.nix' has been renamed to `services.xserver.displayManager.sessionPackages'.
<genesis>
excuse me to insist again but can some people look at #45347 , i'd that merge for my mother
<worldofpeace>
you know, I just realized I should slightly be personaly offended by bash. The shape in this icon has been desecrated
<gustavderdrache>
it's a fantastic job control language, but the stringly-typed data structures make it hard to organize data as effectively as i sometimes want :|
<worldofpeace>
I think that's well put gustavderdrache
<worldofpeace>
bash dun care about data tbh
<gustavderdrache>
i can't entirely lay the blame at its feet since it does aim for posix conformance
emmanuelrosa[m] has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
reallymemorable has joined #nixos
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
Supersonic112 is now known as Supersonic
<lovesegfault>
colemickens: sway pkg needs a fix
ddellacosta has quit [Ping timeout: 258 seconds]
orivej has joined #nixos
<colemickens>
<lovesegfault "colemickens: sway pkg needs a fi"> In my overlay or in nixpkgs? Is it addressed by the sway wrapper PR? What's broken?
<lovesegfault>
colemickens: in the overlay, unstable added some attr that needs to be set for sessionPackage to work
<lovesegfault>
one second
<lovesegfault>
worldofpeace: what was that PR?
<colemickens>
Okay. I'm doing a pass on it right now.
gustavderdrache has quit [Quit: Leaving.]
<worldofpeace>
lovesegfault: I believe this has been fixed, providedSessions in passthru
<lovesegfault>
```
<lovesegfault>
trace: Package, 'sway-aa8fe584215d29e31ae5af43afa81db03b997f85', did not specify any session names, as strings, in
<lovesegfault>
'passthru.providedSessions'. This is required when used as a session package.
<lovesegfault>
colemickens: Do you see the solution worldofpeace linked?
bvdw has quit [Read error: Connection reset by peer]
<colemickens>
I'm guessing that the pkg is expected to provide a session file and that my overlay pkg doesn't place that file.
<colemickens>
I think worldofpeace linked to the module change and I need to find the equivalent pkg change.
bvdw has joined #nixos
<colemickens>
I'll probably just copy the sway default.nix back ontop of my overlay copy just to refresh it from upstream.
<lovesegfault>
colemickens: He linked to the pkg (sway) change, no?
<colemickens>
Should be fixed in an hour or less assuming the other pkgs are okay too.
<lovesegfault>
colemickens: <3
<colemickens>
I haven't scrolled much but he linked the module... not hte pkg.
<colemickens>
The overlay doesn't actually ship the module, only the pkg.
<lovesegfault>
Oh, right
LLED2_3 has quit [Ping timeout: 268 seconds]
<worldofpeace>
all the package needs is passthru.providedSessions = [ "sway" ];, or some fuzzing the module for the package in the option to have this attr
<worldofpeace>
the package in the option being, sessionPackages = [ sway ] needs to have a providedSessions attribute
<lovesegfault>
Setting up a new phone is so much work
<lovesegfault>
Specially since I don't want to transfer all the garbage from my old one
<lovesegfault>
I want Nix for my phone
vld has joined #nixos
<evanjs>
lovesegfault: part of the reason I seriously appreciate how far iCloud has come. Apple person or not, incremental backups, iCloud Drive, and not including photos in your main backup are huge steps forward
<lovesegfault>
evanjs: It's an iPhone, but I didn't want to inherit all the 1000's of pictures, docs, etc etc
y2oioX has joined #nixos
<evanjs>
And I just don't worry about transfers anymore at all. On another note, I would be really interested in trying nix on mobile as well. And yeah I try to keep what I can haha
<evanjs>
It's okay, I have 512GB...
<evanjs>
I actually got as far as installing the iSH beta and curl'ing the nix install into sh... but then it complained about missing pthread something haha
<lovesegfault>
tssss flexing :P
<y2oioX>
How do I tell nixos to not wait for "random: crng init done" (dmesg line), I tried the "security.rngd.enable = false;" in configuration.nix but doesn't seem to have any effect. Thus, I cannot ssh to the virtualbox nixos without first pressing 'a' five times at its prompt to cause the dmesg line to appear(ie. randomness lacking)
<evanjs>
*checks again*
<evanjs>
GC Warning: pthread_getattr_np or pthread_attr_getstack failed for main thread
<{^_^}>
#45345 (by vmandela, 1 year ago, merged): nixos/install-grub: include child configs in grub menu
pie_ has joined #nixos
mexisme has joined #nixos
vld has joined #nixos
thePirateKing has joined #nixos
slack1256 has quit [Remote host closed the connection]
<infinisil>
y2oioX: Btw, I can confirm that kernel version 5.3 does *not* solve the crng problem, only >=5.4 seems to do that
<y2oioX>
nice
<infinisil>
Well, 5.3.11 doesn't work, 5.4.5 does. Somewhere inbetween is the line
<y2oioX>
how easy would it be to revert just one commit from 5.4.5 let's say? and just rebuild all kernel. that is, telling configuration.nix to do all that
<y2oioX>
dno how to use nesting.clone error: The option `boot.kernelPackages' has conflicting definitions, in `<unknown-file>' and `/etc/nixos/configuration.nix'.
y2oioX has quit [Quit: Leaving]
mrlizard has quit [Ping timeout: 246 seconds]
palo has quit [Ping timeout: 260 seconds]
palo1 is now known as palo
<infinisil>
mkForce on the definition
<infinisil>
boot.kernelPackages = lib.mkForce ...
kolby_ has joined #nixos
sjkelly1 has quit [Remote host closed the connection]
Guest13048 has quit [Quit: Lost terminal]
LarryTheCow has quit [Remote host closed the connection]
<thomashoneyman>
has anyone set up shellcheck on nix scripts before?
<thomashoneyman>
along the lines of writeScriptBin + some later hook which lints the produced script with shellcheck?
reallymemorable has quit [Quit: reallymemorable]
<thomashoneyman>
i might only be looking for some setup hook i can run after the script has been written
<sphalerite>
thomashoneyman: you can use writeTextFile's checkPhase (see pkgs/build-support/trivial-builders.nix)
<colemickens>
Is Gimp no longer being built? :(
dansho has quit [Remote host closed the connection]
dansho has joined #nixos
<colemickens>
lovesegfault: I guess now that I've re-read, if I understand correctly, there wasn't actually anything that needed to be done. but I also don't use a login manager so IDK if that scenario works all the way
<thomashoneyman>
sphalerite: thanks! checking this right now
<lovesegfault>
colemickens: Hmm
<lovesegfault>
yeah, gimp build is borke
<lovesegfault>
*d
<colemickens>
It builds for me locally.
<lovesegfault>
For me too, but hydra isn't building it
<thomashoneyman>
sphalerite: in the check phase i'm not sure where to find the text written by writeTextFile
<colemickens>
that's landed in nixos-unstable or are you ahead?
<thomashoneyman>
sphalerite: running `ls -a` in the checkPhase just shows env-vars, none of the script
<colemickens>
(just wondering how I didn't hit it)
<lovesegfault>
colemickens: I'm on nixos-unstable
<thomashoneyman>
i thought it would be in the directory by the time checkPhase runs
<colemickens>
lovesegfault: I'm kinda tired but I'm confused how I'm not hitting this too. Might have to pick this up tomorrow.
<thomashoneyman>
for example, setting the writeTextFile destination to "/bin/my-script" should apparently be the same as writing it to $out/bin/my-script
<thomashoneyman>
but in the check phase there is no bin directory
<lovesegfault>
colemickens: Definitely go rest and look at it when you have time :)
<clever>
thomashoneyman: i dont think writeTextFile runs checkPhase normally
<colemickens>
I hate leaving stuff broken, esp since people actually use the dang thing :P I hate it very much when things like that are broken, esp since it probably means blocking upgrades for people.
<clever>
thomashoneyman: can you paste your nix expr?
<clever>
it never did a cd, so the ls is in $NIX_BUILD_TOP
<thomashoneyman>
yea
<thomashoneyman>
i could reuse $n, i suppose, or do it manually
<clever>
yeah
<thomashoneyman>
wait
<thomashoneyman>
NIX_BUILD_TOP?
<thomashoneyman>
this output still leaves me confused
<clever>
thats the temp dir that your cd'd into when the script starts
<thomashoneyman>
ah, ok
<colemickens>
Oh right, it's only evaluating that when you have a login manager set. I can try that.
<thomashoneyman>
i'd still expect the ls to show bin/run-local
<clever>
thomashoneyman: thats what you get from `ls $out`
<thomashoneyman>
oooooohhhhhhh
<thomashoneyman>
of course
<lovesegfault>
colemickens: Ah, yes
<lovesegfault>
I have GDM, FWIW but I doubt it's specific
mexisme has quit [Ping timeout: 248 seconds]
<thomashoneyman>
clever: i might be good with "${shellcheck}/bin/shellcheck $out${destination}.sh" then
* colemickens
grumbles at having to enable an xserver option ;)
<clever>
thomashoneyman: writeShellScriptBin will stick a #! at the top, and shellcheck it for you
kleisli has joined #nixos
<clever>
ah, correction, it runs the shell with -n, not shellcheck
<thomashoneyman>
clever: i'm planning to write a little utility
<thomashoneyman>
writeCheckedScript or something
<thomashoneyman>
which just does the same thing as writeScript but shellchecks it
<thomashoneyman>
thanks for your help, that command is useful
<thomashoneyman>
working now
<thomashoneyman>
using $out${destination}
<lovesegfault>
colemickens: I really hate that DM's are under xserver
eri has joined #nixos
<kolby_>
How do I override a package that's used for bootstrap? I'm trying to override configureFlags for gmp and libffi using an overlay. I get an error saying that stdenv-linux doesn't allow refering to the following paths.
hax404 has quit [Ping timeout: 246 seconds]
mexisme has joined #nixos
hax404 has joined #nixos
felixfoertsch has quit [Quit: ZNC 1.7.3 - https://znc.in]
felixfoertsch has joined #nixos
<lovesegfault>
colemickens: Did you try to trigger it?
<colemickens>
lovesegfault: here's the thing - there's a big rewrite coming to the nixos module AND base nix packaging that requires bigger changes to the overlay.
<lovesegfault>
There is?
<colemickens>
plus, holiday plans, means I went ahead and leap-frogged to where nixos-unstable will be in a week or so when it updates.
<{^_^}>
#75247 (by Elyhaka, 1 week ago, merged): sway: refactor with a wrapper
atlas_ is now known as nixhubd
<colemickens>
so I'm working on a new branch that builds against my nixpkgs that now contains that commit
<lovesegfault>
Oh, sweet!
<colemickens>
I'm going to test this config at least with gdm. I'll push this as soon as the wrapper/rewrite hits nixos-unstable, and hopefully all will be good for everyone.
<lovesegfault>
colemickens: sounds good :)
<lovesegfault>
speaking of, gebaar-libinput was upstreamed
Heirlung has quit [Read error: Connection reset by peer]
<colemickens>
oh yeah, I saw that, nice; thanks! I should update the pkg def in the overlay
<lovesegfault>
🎉
Heirlung has joined #nixos
mexisme has quit [Ping timeout: 260 seconds]
is_null has joined #nixos
<dansho>
when i load this derivation with nix-shell i get two copies of include in NIX_CFLAGS_COMPILE, is that normal? https://paste.ee/p/wk0LN
<colemickens>
lovesegfault: :/ I just did gdm.enable like old days and wound up with a bad display-manager.service
<lovesegfault>
O.o
<colemickens>
display-manager.service: Service has no ExecStart=, ExecStop=, or SuccessAction=. Refusing.
<colemickens>
Ya. It's on by default but the displaymanager unit isn't emitted correctly unless you enable xserver specifically too :(
<lovesegfault>
Hm
<lovesegfault>
lame-o
nixhubd is now known as atlas_
vidbina has joined #nixos
<keithy[m]>
sigh i'm still banging my head agains dnsmasq
<keithy[m]>
and resolv
thePirateKing has quit [Quit: Leaving]
mexisme has joined #nixos
daveyjones01 has joined #nixos
daveyjones01 has quit [Client Quit]
marcusr has joined #nixos
<colemickens>
lovesegfault: I don't know what's going on now
<colemickens>
I can see in GDM's logs that it is parsing and not rejecting sway.desktop.
<lovesegfault>
O.o
mexisme has quit [Ping timeout: 246 seconds]
<colemickens>
I cansee in the logs that it is running in wayland mode. And yet, it doesn't show "Sway". but it shows xterm.
<lovesegfault>
did you set senssionPackages?
mexisme has joined #nixos
thomashoneyman has quit [Quit: Lost terminal]
lovesegfault has quit [Quit: WeeChat 2.7]
<colemickens>
lovesegfault: yes, the rewrite I linked to includes the sessionPackages = [swayPackage]; line, and like I said, it's in the XDG Share location that gdm looks in (and is why it mentions it in the log)
<colemickens>
I'm a bit flabbergasted.
lovesegfault has joined #nixos
<colemickens>
oh weird, the wayland session is crashing out and then its restarting the x version, hence why sway is missing.
xkapastel has joined #nixos
leotaku has joined #nixos
<keithy[m]>
my router hates me... all dns queries are going to the router and being forwarded to dnsmasq. Direct query of dnsmasq returns local lan/etc/host results as requested. Indirect query of local lan, via router, no result.
<keithy[m]>
So clearly not all dns queries are being forwarded afterall, it is making an exception for local queries.
Maxdamantus has quit [Quit: Back in around 2 million seconds]
<keithy[m]>
the reason appears to be that dnsmasq has decided that if it is asked directly for a local lookup it provides it from etc/hosts - great!
<keithy[m]>
but if it is asked indirectly from the router, it has decided that the dns-server to ask about local addresses is... the router!
_daveyjones has joined #nixos
ambro718 has joined #nixos
drakonis has quit [Quit: WeeChat 2.6]
domogled has quit [Ping timeout: 265 seconds]
<keithy[m]>
* but if it is asked indirectly from the router, it has decided that the dns-server to ask about local addresses is... the router! adding no-resove is better so now the logs show the response being sent back... but I never get them.
<colemickens>
lovesegfault: tldr: gdm+wayland only works if all of gnome3 is enabled
mexisme has quit [Ping timeout: 248 seconds]
<gyroninja__>
I'm trying to override git with a patch, but when I do `home-manager switch` it wants to redownload like all my system
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
<gyroninja__>
Does a package using fetchgit cause it to need to be rebuilt? Even if we are specifying hashsums for the output?
ixxie has joined #nixos
<lovesegfault>
colemickens: Oh, yeah, you need some random shit
<lovesegfault>
now I remember
<colemickens>
anyway, sway starts from gdm now
<colemickens>
I'm going to test if these pkg changes work with current nixos-unstable, if so I'll push
opthomasprime has left #nixos [#nixos]
fintohaps has joined #nixos
<fintohaps>
Hey 👋 Glad to join the community while I'm testing out NixOS 😁
<fintohaps>
I have a question about a process that's running, but it may not be a NixOS specific thing. There's some command running called "tags.so" that's maxing out my CPU and I can't seem to find out what it is based off of googling it. Anybody have an idea of what it is?
<colemickens>
lovesegfault: I've pushed it, since it works without the wrapper changes in nixpkgs. You should be able to use it through gdm now (thats how I'm currently using it)
<dansho>
buildInputs is for runtime dependencies?
<jschievink>
have there been any recent changes to Nix that change how the fetchgit sha256 is calculated?
leotaku has quit [Ping timeout: 265 seconds]
orivej has quit [Ping timeout: 265 seconds]
<lovesegfault>
colemickens: testing!
<lovesegfault>
colemickens: Looks good!
cosimone has quit [Quit: Terminated!]
<lovesegfault>
pycurl build failed
<jschievink>
(I'm asking because the hash of my OpenOCD git checkout changed twice without me changing the git revision)
<jschievink>
and MITM seems very unlikely
<lovesegfault>
colemickens: rebuild successful!
lovesegfault has quit [Quit: WeeChat 2.7]
<gyroninja__>
Was the revision set to a tag?
lovesegfault has joined #nixos
<lovesegfault>
worldofpeace: Hmm, lollypop is still borked
Guest33385 has joined #nixos
cosimone has joined #nixos
the_pumpkin_man[ has joined #nixos
leotaku has joined #nixos
is_null has quit [Ping timeout: 260 seconds]
zaoboa has joined #nixos
goldengoof[m] has joined #nixos
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
<zaoboa>
Just a heads-up that master and its changes to session borked my XMonad desktop and gpg settings. So watch out. Now trying to figure out how to fix it.
<dansho>
is doing src = ./.; bad? it seems to pull .git into the nix store
<zaoboa>
It looks like a deprecation notice suggested to switch to desktopManager.session* but PR #53843 talks about windowManager.session*
<epsilonhalbe>
replacing old 'dhall-1.24.0'installing 'dhall-1.28.0'error: path '/nix/store/30m1l7ss2b8qbxa7ajx26pfvxzmvlghk-dhall-1.28.0' does not exist and cannot be created
<epsilonhalbe>
probably not - I just started using nix
<epsilonhalbe>
I'll read up on cache configuration, give it a try and come back
<LnL>
the channel expressions provided by hydra directly like that don't include the full build instructions IIRC
<LnL>
so those only work when substituting
<johanot>
i'm trying to cross compile for my raspberry pi 3 (aarch64), but I'm very new to the nix-cross-machinery. I thought I could substitute most things from cache.nixos.org, but for some reason i'm now compiling gcc. My localSystem is x86_64.
dicearr has quit [Client Quit]
domogled has quit [Ping timeout: 258 seconds]
<epsilonhalbe>
LnL - yeah the caches were the problem - thx for pointing me in the right direction
reanimus has quit [Quit: Logging off...]
reanimus has joined #nixos
epsilonhalbe has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 260 seconds]
chloekek has joined #nixos
kolbycrouch has quit [Read error: Connection reset by peer]
<NixOsFun>
Hi gays. What package I need to install to add zip support to krusader?
drakonis has joined #nixos
mexisme has quit [Ping timeout: 260 seconds]
<angerman>
If something ends up being merged into master, does that end up in 19.09, 19.03? What would need to happen to have it end up in there?
<__monty__>
An explicit backport afaik.
chiefgoat has quit [Read error: Connection reset by peer]
<johanot>
angerman: you can ask for a backport in the original PR or you can open a new PR against "release-19.09" yourself, that's what I normally do :)
<angerman>
fair, I'll probably have to do that then.
<Fare>
I upgraded my NixOS to 19.09 then nixos-unstable, and in both cases, the X server won't start.
<Fare>
Maybe because grub now uses efi instead of vesa for initializing the video? And/or because there is no intelfb driver in the kernel? I'm not sure what's happening exactly, but no X for me :-(
EuAndreh[m] has quit [Quit: authenticating]
GiGa has quit [Quit: Leaving]
Fare has quit [Ping timeout: 260 seconds]
<chloekek>
I installed Nix but now the nix command is only available if I my shell is sh, not with bash.
<chloekek>
Oh, I think it’s an SSH issue.
chiefgoat has joined #nixos
<chloekek>
I will make bash the login shell and try again.
<MichaelRaskin>
Also, have you changed hardwareScan settings?
dbmikus has joined #nixos
<luc65r[m]>
Hi, I'm new to this NixOS, and I'm trying to install it
magnetophon has joined #nixos
<genesis>
luc65r[m] : you're welcome
<luc65r[m]>
I did enable ohmyzsh, but I have a default zsh
<luc65r[m]>
genesis: thanks !
<genesis>
which distro did you use luc65r[m] ?
<luc65r[m]>
Arch, Gentoo, Alpine and KISS
<genesis>
oki, was using mainly gentoo, was so dead :D
sigmundv_ has joined #nixos
<srhb>
luc65r[m]: How did you enable ohmyzsh?
<luc65r[m]>
In configuration.nix, like in the manual
<srhb>
luc65r[m]: Can you confirm that /etc/zshrc contains the sourcing of the oh-my-zsh init?
<lassulus>
luc65r[m]: is the line source $ZSH/oh-my-zsh.sh in /etc/zshrc?
<lassulus>
whupps, to slow
<srhb>
luc65r[m]: Also, you're saying you're trying to install NixOS, where in the process are you exactly?
cosimone has quit [Quit: Terminated!]
<luc65r[m]>
I didn't have that…
<srhb>
luc65r[m]: Are you booted into the final NixOS system yet, or still in the installer?
<luc65r[m]>
I'm on the final NixOS system
<srhb>
luc65r[m]: Okay, good. So you altered /etc/nixos/configuration.nix, added programs.zsh.ohMyZsh.enable = true; and then you did what to activate the new system?
Fare has joined #nixos
nyowa has joined #nixos
<Fare>
OK, X is back after I removed config.boot.vesa = true.
<Fare>
But it looks like the contents of xservers.libinput isn't heeded anymore.
<luc65r[m]>
nixos-rebuild switch
mexisme has joined #nixos
<srhb>
luc65r[m]: As root? Did the command succeed?
<luc65r[m]>
yes
<srhb>
luc65r[m]: Can you pastebin your configuration.nix then please?
<luc65r[m]>
oh can I do that?
<srhb>
luc65r[m]: There's no automated process for it, if that's what you're asking :)
<lassulus>
I think there is programs.zsh.enable = true; missing
<srhb>
Yep, that's it :)
<luc65r[m]>
Do I write programs.zsh.enable = true; or I add zsh in environmenc.systemPackages ?
<lassulus>
the former
fenedor has quit [Ping timeout: 265 seconds]
<srhb>
luc65r[m]: The programs.zsh.enable option will ensure it's in systemPackages automatically.
<lassulus>
programs.zsh.enable will add the /etc/zshrc and other files
<srhb>
And will enable the rest of the module logic that the ohmyzsh module relies on
<luc65r[m]>
Thanks a lot!
<luc65r[m]>
But I have other problems
<srhb>
Ask away.
kleisli has joined #nixos
<luc65r[m]>
I did add NOPASSWD in my configuration.nix in the sudo configuration, but I have to type my password
<srhb>
luc65r[m]: You probably don't want to use configFile, but extraConfig
<srhb>
Or extraRules, even better.
<sphalerite>
or security.sudo.wheelNeedsPassword
Fare has quit [Ping timeout: 260 seconds]
<luc65r[m]>
That works! Thank you
<luc65r[m]>
I have two more things
<duairc>
Does anyone have any experience using hydra as a substituter/binary cache?
<luc65r[m]>
I use nvidia non-free drivers, how can I change the resolution in the TTY ?
<duairc>
I've been trying for days but I keep getting this problem: warning: substituter 'https://...' does not have a valid signature for path '/nix/store/...'
<sphalerite>
luc65r[m]: you can't because nvidia doesn't implement KMS
<duairc>
I don't know enough about nix signatures to know where the problem is
<sphalerite>
luc65r[m]: only way to achieve that is to use nouveau if it works with your card.
<duairc>
But I've set trusted-public-keys on my local machine and secret-keys on the remote
<sphalerite>
duairc: did you set secret-keys before building the stuff you're trying to get from it?
<duairc>
sphalerite: Yes, I did, but just in case I ran nix-store sign-paths --all
<sphalerite>
duairc: then it's probably a caching issue — try removing ~/.cache/nix/binary-cache*
<duairc>
I did that too :/
<duairc>
Like the only thing I can think of is that it's something stupid like the public key somehow doesn't match the private key
<sphalerite>
try removing root's as well
<duairc>
I did
<srhb>
It may be easier to use the ttl options to avoid caching issues
<sphalerite>
damn!
<duairc>
Is there some way to rederive the public key form the private key?
<worldofpeace>
lovesegfault: I just don't think I'll have the bandwidth to fix it for a bit
<sphalerite>
duairc: narinfo-cache-positive-ttl
<sphalerite>
duairc: try running it with --narinfo-cache-positive-ttl 1
NixOsFun has quit [Remote host closed the connection]
ixxie has quit [Ping timeout: 240 seconds]
<duairc>
I don't know if this complicates things, but for what it's worth the actual building is happening on another machine defined in nix.buildMachines. I wasn't clear on which machine needs to do the actual signing
<duairc>
But in any case nix sign-paths --all should work, right?
<srhb>
duairc: Yes.
<duairc>
I might just generate new keys and try again, maybe my public and private got mixed up somehow, that's the only think I can think of
<srhb>
duairc: Sounds possible, yeah.
<duairc>
But are there any tools for inspecting if a particular build is signed, and if so, by what key? Is there some artefact on the filesystem of that?
<luc65r[m]>
Ok… And last thing, I did enable AwesomeWM, but when I do startx, I land on a weird environment. Do I need a .xinitrc?
<srhb>
luc65r[m]: Yes, read the description of services.xserver.displayManager.startx.enable
<srhb>
luc65r[m]: Basically all the automagic only happens if you use a real display manager, with startx you're on your own
<MichaelRaskin>
Fare:does manually running xinput commands have the desired effect for input configuration?
<MichaelRaskin>
Just to make sure which level of configuration being ignored it is…
<MichaelRaskin>
Also, is it about a touchpad?
Fare has joined #nixos
<luc65r[m]>
Thanks a lot srhb sphalerite lassulus !
<srhb>
luc65r[m]: Sure thing, and welcome to NixOS again. :)
<keithy[m]>
<Fare>
looks like the entire xorg configuration generation was disrupted sometimes between 19.03 and now, and my xorg.conf (or lack thereof) is missing many pieces.
<Fare>
who's on top of that?
<srhb>
Fare: We probably need more details than that.
<keithy[m]>
I am still confused as to why the hostname ends up in /etc/hosts as 127.0.1.1
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<srhb>
keithy[m]: Why is that confusing?
<keithy[m]>
because then it gets served to dns as 127.0.1.1
<srhb>
Uh.. What DNS?
<keithy[m]>
dnsmasq
<keithy[m]>
so apparently there is no place to actually set the ip of a host
<keithy[m]>
I thought 127.0.1.1 was a fallback if no ip was set
<sphalerite>
duairc: nix path-info --sigs
<srhb>
keithy[m]: It gets set if the hostname is nonzero. The reasoning is in c57892462bf22e537a2cd5ee5e6f88b0b59a2d33
<keithy[m]>
nice hash what does it mean?
<keithy[m]>
so I have to set two hostnames, a dummy one and a real one
<Fare>
when trying to use libinput, a /etc/static/X11/xorg.conf.d/40-libinput.conf is generated, but it has no entry for any of my many options like tapping=false
<srhb>
Fare: Do you have a config that can reproduce this?
<Fare>
19.03 used to do the right thing
<keithy[m]>
thx for that, srhb:
<keithy[m]>
gtg
dellie__ has joined #nixos
dirkx has quit [Quit: Leaving.]
<Fare>
srhb: I do, but it's currently long and private :-/
<srhb>
Fare: Okay, I guess I can try and see if I can reproduce it blindly.
ixxie has joined #nixos
delli3_ has quit [Ping timeout: 260 seconds]
<Fare>
in this case, try xserver.libinput = {enable=true;accelProfile="adaptive";tapping=false;}
<Fare>
the enable=true does cause the 40-libinput.conf to be created, as opposed to say the equivalent thing for synaptics if I use synaptics instead.
<Fare>
But there is no entry for tapping or accelprofile
<srhb>
Fare: Right, it looks like that's added to the main xorg config.
<Fare>
but there is no xorg config
<srhb>
Fare: That's passed as a command line option iirc
<sphalerite>
there is, just not in /etc :D
<sphalerite>
I think
<srhb>
Yeah
<Fare>
where can I find the xorg config?
<srhb>
Fare: systemctl cat display-manager
<srhb>
I think..
<sphalerite>
it's DM-dependent :/
<Fare>
nope, that's after X is started, it seems
brnzz has quit [Quit: Ping timeout (120 seconds)]
<srhb>
Ah, yeah, looks like it..
<sphalerite>
if X is running, you can use pgrep -a X to find the path to the config
brnzz has joined #nixos
<sphalerite>
if you're using sddm, you can look for the wrapper script in /etc/sddm.conf and the wrapper script contains the path to the config file
<srhb>
or `nix-store -qR /run/current-system | grep xserver.conf
<sphalerite>
or that lol
bvdw has quit [Read error: Connection reset by peer]
<Fare>
thanks
<Fare>
OK, I can see options being defined there indeed
bvdw has joined #nixos
<Fare>
... but they don't seem to make a difference anymore :-/
<Fare>
(they used to work on 19.03)
<srhb>
Sounds like you might want to file an issue.
<sphalerite>
Fare: #73785 might be related to your issue
<sphalerite>
although that's been reverted now so… idk
vidbina has quit [Ping timeout: 260 seconds]
<srhb>
The revert is quite fresh though
<srhb>
It landed in the channel sometime yesterday.
<srhb>
Fare: ^ might want to check if you have that
* Fare
runs a nix-channel --update (using nixos-unstable)
<sphalerite>
oh, the revert isn't in unstable AFAIK
<srhb>
Oh, unstable. Didn't check if it's in there yet
<srhb>
It's not.
<Fare>
how do I check?
<Fare>
so, you guys use directly nixpkgs-unstable for your nixos?
<srhb>
I do not.
<Fare>
I used to use 19.03, and trying to upgrade to 19.09 was... eventful.
<sphalerite>
it won't be reverted on unstable AFAIK
<srhb>
So you decided to go directly to 19.09? :-P
<srhb>
Er, unstable
<sphalerite>
Fare: eventful how?
<srhb>
Yeah, it doesn't look like it..
<Fare>
sphalerite, X stopped working
<Fare>
I eventually removed config.boot.vesa, but only after trying unstable first.
<sphalerite>
Fare: huh, was there anything in the release notes about that?
<Fare>
should I be going back to 19.09 ?
<srhb>
Fare: You can try :)
<Fare>
sphalerite, I didn't see anything related to vesa in the release notes.
<srhb>
19.09 was definitely one of the more breaky upgrades for me too, but I guess it's very dependent on individual configs. The release notes just aren't sufficient to cover all the cases.
<sphalerite>
:(
<Fare>
Can I get something retroactively added to the release notes?
<srhb>
Fare: Quite possibly.
<Fare>
I don't know what boot.vesa used to do, or if it even changed, but the nomodeset it includes causes X to fail to start.
<Fare>
also, I doubt a hardcoded mode 0x317 is optimal for everyone, but hey, I don't feel like hacking it right now.
<srhb>
I think it probably requires a bit more understanding of the issue to write a comprehensible releas enote.
<Fare>
Actually, I don't have a comfortable way to hack nixos
<Fare>
(I also really don't like nixos syntax.)
<Fare>
(The mostly equivalent language jsonnet has much nicer syntax, for instance)
<srhb>
That sounds like Objective Truth. :-)
domogled has joined #nixos
<MichaelRaskin>
Dunno, I think Nix syntax is pretty fine.
<srhb>
I also quite like it.
<MichaelRaskin>
BTW, I do use Nixpkgs-unstable as a base of my system, and I think I have never really used Nixpkgs/NixOS releases
<MichaelRaskin>
(I still run that not-quite-NixOS system, but I use mainline NixOS X.org config generator)
bhipple has joined #nixos
jgeerds has joined #nixos
<MichaelRaskin>
BTW, if you want to check xorg.conf, xserver.exportConfiguration might help
red[evilred] has joined #nixos
<red[evilred]>
I gave a friend who is a gentoo user and every time a package breaks in portage he switches it to nix
<MichaelRaskin>
Ah, so Nix-on-Gentoo setup?
o1lo01ol1o has quit [Remote host closed the connection]
<srhb>
Are the libinput values essentially equivalent to synaptics ones? I have all sorts of neat magical constants that I have no idea what mean.
<red[evilred]>
Yes
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<MichaelRaskin>
I would not be so sure the exact calibration is always the same
o1lo01ol1o has joined #nixos
xkapastel has quit [Quit: Connection closed for inactivity]
<srhb>
Yeah, probably not. Let's see what happens..
reallymemorable has joined #nixos
wucke13 has joined #nixos
<Fare>
how do I tell nixos-rebuild to use a checkout I hack instead of a channel from ~/.nix-channels ?
<clever>
Fare: -I nixpkgs=/path/to/nixpkgs
gustavderdrache has joined #nixos
oida has quit [Remote host closed the connection]
oida has joined #nixos
magnetophon has quit [Remote host closed the connection]
<Fare>
Thanks a lot to you all... rebooting with a new configuration, let's see how it works...
Fare has quit [Remote host closed the connection]
NoctisLabs has joined #nixos
bitmapperphone has quit [Read error: Connection reset by peer]
zupo_ has joined #nixos
zupo has quit [Ping timeout: 260 seconds]
FareWell has joined #nixos
<FareWell>
I rebooted with my new nix configuration from nixpkgs-unstable. X works, but my stumpwm and the utilities it runs at startup all fail, claiming they can't connect to display :0.0, which never happened before.
delli3_ has joined #nixos
xkapastel has joined #nixos
<FareWell>
Although at one point I had to remove the ~/.Xauthority that was somehow garbled and owned by root, possibly due to my using sudo the wrong way at some point.
<FareWell>
I see that there is -nolisten tcp on the command line in the journalctl, even though I have enableTCP = true in my common.nix -- or is that being overridden away from using xserver = { ... } in my local configuration.nix ?
dellie__ has quit [Ping timeout: 260 seconds]
<MichaelRaskin>
How are they combined? Module require or attrset override?
<FareWell>
I am not sure how they are combined... I just added ./common.nix in the imports of my local configuration.nix
Chiliparrot has joined #nixos
<MichaelRaskin>
Then this is via module system, and local configuration's lack of mention of enableTCP should not override anything
<FareWell>
Also, after I explicitly add enableTCP = true in the local configuration.nix, I get a new error, and new to me, where stumpwm now says: Connection failure to X11.0 server display 0: No protocol specified
<FareWell>
First time I see this.
FareTower has joined #nixos
<FareTower>
stumpwm works again after I change the default from :0 to unix/:0 on this line of my ~/.stumpwm: (stumpwm:stumpwm (or (uiop:getenvp "DISPLAY") "unix/:0"))
drakonis has joined #nixos
FareWell has quit [Ping timeout: 268 seconds]
<FareTower>
Which begs the two questions of why enableTCP is now not heeded somehow by sddm, and how stumpwm fails to get the current display from either environment or command-line
FareTower has quit [Remote host closed the connection]
justanotheruser has joined #nixos
magnetophon has joined #nixos
<drakonis>
lassulus: thanks for merging, but really though, the lack of these files were what made doom's music sound not how it should
<drakonis>
like it came from a roland sc-55
bitmapperphone has joined #nixos
<lassulus>
alright, I guess you are right, I never realized, but actually I seldom play with the original music :D
<lassulus>
thanks for fixing!
FareTower has joined #nixos
mexisme has joined #nixos
orivej has joined #nixos
magneticduck has joined #nixos
FareTower has quit [Ping timeout: 260 seconds]
acarrico has quit [Ping timeout: 265 seconds]
kolbycrouch has quit [Ping timeout: 260 seconds]
dm9 has quit [Quit: WeeChat 2.7]
chloekek has quit [Ping timeout: 260 seconds]
meck has quit [Quit: WeeChat 2.7]
kolbycrouch has joined #nixos
bitmapperphone has quit [Remote host closed the connection]
<mica[m]>
Is anyone using a fully color managed work flow on NixOS?
reallymemorable has quit [Quit: reallymemorable]
medvid has joined #nixos
reallymemorable has joined #nixos
cosimone has joined #nixos
reallymemorable has quit [Client Quit]
ashesham` has quit [Ping timeout: 252 seconds]
PLPD-Bot has joined #nixos
cosimone has quit [Quit: Terminated!]
Phillemann has joined #nixos
<Phillemann>
I've written a new NixOS service. What's an easy way to test if it works? I mean if a proper systemd script is generated and the system is started when I enable it?
thomashoneyman has joined #nixos
<drakonis>
Phillemann: rebuild your system using your nixpkgs clone
gambpang has quit [Quit: ZNC 1.6.3+deb1+xenial0 - http://znc.in]
<Phillemann>
Hmmmmm
<infinisil>
Phillemann: You can build a vm with `nixos-rebuild build-vm`
<infinisil>
Then `-I nixpkgs=$PWD` to use the checkout
<infinisil>
And `-I nixos-config=$PWD/config.nix` to specify a custom configuration.nix to test with
kleisli has quit [Ping timeout: 265 seconds]
gambpang has joined #nixos
<Phillemann>
error: assertion failed at /home/philipp/nixpkgs/pkgs/top-level/impure.nix:82:1
<Phillemann>
Ah, sorry.
gambpang has quit [Quit: ZNC 1.6.3+deb1+xenial0 - http://znc.in]
<Phillemann>
infinisil: Seems to work. Although I need to read up on actually running this VM now :D
<infinisil>
Phillemann: result/bin/run-vm or so
gambpang has joined #nixos
<evanjs>
Might take a look at my touchpad issue again today. Anybody have an idea of what to check first? No clue if it’s nix-related
<evanjs>
Last things I figured were that adding myself to inputs worked around the issue, and that the touchpad wasn’t initialized properly
<evanjs>
evtest showed that it was functioning as intended, though. Output for movement, clicks, etc
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<Fare>
my, upgrading Linux is such a shitshow. Not even NixOS can paper over that.
<gchristensen>
okay!
o1lo01ol1o has quit [Ping timeout: 268 seconds]
cosimone has joined #nixos
magneticduck has quit [Quit: Leaving]
<Fare>
Apparently, nixos now adds default entries in sudoers for wheel, that override previous entries I defined in security.sudo.configFile
<duairc>
srhb: sphalerite: I regenerated my keys, nix sign-paths'd everything again, verified with nix path-info --sigs that that actually did something, rebuilt the configuration on my local machine with the new public keys... "still substituter 'X' does not have a valid signature for path 'Y'"
<Fare>
OK, remains the issue with libinput parameters not being properly applied anymore.
<Fare>
Maybe the Xorg options must be repeated in each and every relevant section?
<MichaelRaskin>
I think the problem it the removal of MatchTouchpad
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
kleisli has quit [Ping timeout: 265 seconds]
cosimone has quit [Remote host closed the connection]
xkapastel has quit [Quit: Connection closed for inactivity]
cosimone has joined #nixos
<Fare>
I don't know or understand the Xorg section merging / prioritizing algorithm. Maybe things are discarded rather than merged when the matching criteria are less specific?
gcoakes[m] has joined #nixos
<esclear>
Huh. I can't manage to install the python matplotlib package, because that pulls in the sphinx package, which in turn pulls in SQLAlchemy.
<esclear>
Unfortunately during the build, it starts to run tests, one of which fails for SQLAlchemy.
<Fare>
note that in addition to the entry in xorg.conf, there are entries in xorg.conf.d/40-libinput.conf
kleisli has joined #nixos
<Fare>
my guess is that the latter replace the former, and since they have no options, the options don't apply
<Fare>
maybe when MatchTouchpad is added, the entry becomes more specific, hence takes priority
<Fare>
... how do I experiment without rebooting every time?
<MichaelRaskin>
Rebooting or reloading X?
<MichaelRaskin>
systemctl restart display-manager from a root login in tty2 sounds like a plan
<MichaelRaskin>
Not performed by nixos-rebuild for obvious reasons
<Fare>
in this case, I suppose, without reloading X
<Fare>
in a secondary X server :1
<MichaelRaskin>
Oh sigh, it was so easy before systemd…
fusion809 has joined #nixos
<keithy[m]>
Fare did you resolve your sudo issue there are some issues discussing it.
<keithy[m]>
if you just want wheel to have NOPASSWD there is an option for that
<keithy[m]>
otherwise you can trivially create and use a sudo group which is immune to wheel interference.
o1lo01ol1o has joined #nixos
<keithy[m]>
* Fare: did you resolve your sudo issue there are some issues discussing it.
<keithy[m]>
wrt syntax, C is consistent and smalltalk, nothing else has impressed me in 40 years, until nix.
<keithy[m]>
the important thing is eliminating exceptions to the rules.
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
domogled has quit [Ping timeout: 268 seconds]
reallymemorable has joined #nixos
<keithy[m]>
and if possible eliminate static typing
<evanjs>
.... well... I guess my touchpad works once again.... note entirely sure why it stopped in the first place
reallymemorable has quit [Client Quit]
Fare has quit [Ping timeout: 260 seconds]
<MichaelRaskin>
keithy[m]: have you seen Scheme? APL (or J)?
reallymemorable has joined #nixos
<keithy[m]>
yes
<keithy[m]>
to Scheme
fendoer has quit [Ping timeout: 268 seconds]
<keithy[m]>
feels like forth,
<keithy[m]>
last used forth aged 11
<keithy[m]>
I just realised it was my forth programming language fancy that
<MichaelRaskin>
Heh
<keithy[m]>
I still struggle to take anything seriously that cant improve on the readability of smalltalk.
<keithy[m]>
since 80% of your costs are down to maintenance, I like to code in sentences.
<keithy[m]>
anyone who invents a langiage less readable that smalltalk is causing pain in the world.
<keithy[m]>
and that would be all of them.
<MichaelRaskin>
If your concept of sentence is COBOL-oritiented and rigid…
<keithy[m]>
why be restricted to thisFunctionThatDoesStuffWith(a,b,c,d) when you could have
<keithy[m]>
between: a and :b
<keithy[m]>
a description for each argument
<keithy[m]>
between(a)and(b)
<keithy[m]>
just looked at a few jsonet slides, and saw +:: and prayed I never have to use it
<MichaelRaskin>
No, a _description_ for each argument would be named parameters
<MichaelRaskin>
(actually supported in a lot of languages)
<evanjs>
MichaelRaskin: J omg. I need to try euler again with J
<keithy[m]>
named parameters are named parameters
<keithy[m]>
description is the function of what you do with them
<keithy[m]>
and named parameters can have "parameter names"
<keithy[m]>
variables I think they are called
<keithy[m]>
nevertheless I find nix pretty good.
<keithy[m]>
and I will reserve judement until I find myself tracking down a bug that is caused by me accidentally calling a variable after a reserved word. (pet hate)
ddellacosta has quit [Ping timeout: 258 seconds]
<MichaelRaskin>
Hehe, in Common Lisp that kind of bug is impossible
<keithy[m]>
argghh ||| is a delimeter in jsonnet!
reallymemorable has quit [Quit: reallymemorable]
<keithy[m]>
perhaps pl/1 was on to something
domogled has joined #nixos
<keithy[m]>
* perhaps pl/1 was on to something, or im thinking of something else...
<keithy[m]>
* perhaps pl/1 was on to something, or im thinking of something else... APL
reallymemorable has joined #nixos
<martyet-o>
oh jeeez.. anyone on 19.09 with KDE desktop experiencing disturbing kdewallet (which im not using) asking in loop for some password (which i dont even know) when reconnecting to different wifi spot?
<martyet-o>
its pretty annonying and after searching through google, nothing found :-(
<martyet-o>
switching to anything else (xfce, mate, whatever else..) is totally fine
<martyet-o>
even if i disable kdewallet, it seems its not accepting that change and keeps turned on :(
kleisli has quit [Ping timeout: 265 seconds]
__monty__ has quit [Quit: leaving]
domogled has quit [Ping timeout: 260 seconds]
ixxie has quit [Ping timeout: 260 seconds]
cosimone has quit [Quit: Quit.]
<rotaerk_>
https://nixos.wiki/wiki/Nvidia is this accurate? the only solution to getting it to toggle between the integrated GPU and the nvidia card is to use bumblebee, and that this has problems with vulkan?
<rotaerk_>
and performance problems in general
<MichaelRaskin>
Well, I use Nouveau and offload and it works
<rotaerk_>
hmm... I don't think nouveau is an option for me; it was what was enabled by default, and with it, my laptop kept crashing periodically; hasn't happened since I switched to proprietary
Thra11 has quit [Quit: WeeChat 2.6]
Fare has joined #nixos
<Fare>
keithy[m]: I worked around the sudoers issue, didn't actually "solve" it: I configured wheel to have nopassword, instead of a single user to have no password.
<Fare>
I don't understand why the latter wheel entry was adding the password to the magic user.
<Fare>
keithy[m]: I have something like that in configFile, but that was not working.
<Fare>
Is Nouveau supposed to work on an i915 ? I thought it was for NVidia only.
domogled has joined #nixos
<Fare>
MichaelRaskin: are you still here? I don't understand #55520
<Fare>
is "modesetting" a driver that somehow supersedes intel?
<MichaelRaskin>
modesetting is a driver that uses the KMS
<MichaelRaskin>
And intel driver seems abandoned
<MichaelRaskin>
And after an Xorg update some usecases have started breaking
<rotaerk_>
I'm generally having trouble getting the graphics cards working right with my laptop
rotaerk_ is now known as rotaerk
<MichaelRaskin>
Fare: But if disabling VESA was enough to get X, maybe you do not have anything manually set in the configuration anyway
_daveyjones has left #nixos ["WeeChat 2.6"]
<rotaerk>
enabling bumblebee made it so that the system froze when it tried to start X server
<rotaerk>
I'll try optimus_prime
<Fare>
what's the KMS ?
<MichaelRaskin>
Kernel mode setting
<MichaelRaskin>
The kernel direct rendering driver handles the video mode changes
<clever>
another benefit of KMS (in theory) is that the kernel can switch back to text mode when it has a panic attack
hyperfekt has quit [Quit: bouncer quit]
<clever>
so you can actually see the panic msg
hyperfekt has joined #nixos
<clever>
prior to kms, the panic was printed to the text console, but xorg had a strangle-hold on the gpu, so you never saw the error, and the machine just "hangs" silently
<Shell>
also, mode setting doesn't wind up in xorg so we can have shiny things like Wayland :)
satoshi has quit [Ping timeout: 250 seconds]
<clever>
yeah, any program can now benefit from the modesetting
<MichaelRaskin>
fbpdf, for example
chiefgoat has quit [Ping timeout: 265 seconds]
<rotaerk>
"couldn't find RGB GLX visual or fbconfig"
domogled has quit [Quit: domogled]
domogled has joined #nixos
elux has joined #nixos
hyperfekt has quit [Quit: bouncer quit]
<elux>
hi there
hyperfekt has joined #nixos
<elux>
just wondering if anyone here has heard of using nixos+i3 along with elementaryOS's ui, or popOS's ui? both of those distros have some nice gui programs which would be nice to adopt
<Fare>
> OK, so removing the videoDrivers=["intel"] line works.
<Fare>
Back in the days, I could launch a second X as :1 on vt8, but I'm not sure how to do it with current NixOS
zaeph has joined #nixos
<mica[m]>
elux: depends on what GUI programs you're talking about
<Fare>
trying to mimic what pgrep -a X does leads to lockups and/or multiple Xs on the same vt, etc.
<MichaelRaskin>
The most insulting is to get «permission denied»
<MichaelRaskin>
Apparently one would need to rewrite the systemd unit to a multi-instance version
philr has joined #nixos
coco has joined #nixos
misterwhatever has joined #nixos
<sphalerite>
elux: yes, there are some efforts in that direction
<coco>
beginner question: i installed lots of software for a specific user on a machine with "nix-env -i", and now i want to install the same software for some user on another machine. what file do i need to transmit, and what do i need to run to install it there? in other words, how are user-level packages declared?
inkbottle has joined #nixos
<sphalerite>
coco: the situation's a bit unfortunate with nix-env -i
<coco>
sphalerite: how would i do it declaratively then?
<elvishjerricco>
Is there a function in lib that removes duplicates from a list?
reallymemorable has quit [Quit: reallymemorable]
<sphalerite>
coco: what you could do is run `nix-store -q --references ~/.nix-profile` on the one machine, and then `nix-env -i` all those paths on the other machine. That will only work if they're in the binary cache though.
<sphalerite>
elvishjerricco: unique
zebrag has quit [Ping timeout: 260 seconds]
<elvishjerricco>
sphalerite: Ah, thanks
reallymemorable has joined #nixos
<coco>
sphalerite: if i want to install packages on the user-level in a way that i can easily reproduce it elsewhere, how would i do that?
<sphalerite>
,declarative coco
<sphalerite>
aw no, the bot's down!
<sphalerite>
gchristensen: any idea what's up with that?
<sphalerite>
coco: There are multiple ways of managing declarative profiles. 1) Attrset, compatible with imperative use of nix-env https://git.io/fAQHW ; 2) buildEnv, providing more control over the paths that are linked into the profile https://git.io/fp0aU ; 3) home-manager, providing nixos-like config for your ~ https://github.com/rycee/home-manager
{^_^} has quit [Remote host closed the connection]
<coco>
sphalerite: thanks
elux has quit [Quit: leaving]
<rotaerk>
how would one approach diagnosing glxinfo saying "Error: couldn't find RGB GLX visual or fbconfig", and glxgears saying "Error: couldn't get an RGB, Double-buffered visual"
o1lo01ol_ has joined #nixos
<rotaerk>
when I google these errors, I get suggestions to reinstall stuff
<sphalerite>
rotaerk: what graphics chip do you have?
<rotaerk>
I also just added a hardware.nvidia.optimus_prime section, though this problem existed before and after that
<thomashoneyman>
perhaps due to the use of `rec`?
<sphalerite>
rotaerk: hm ok… try changing it to just ["modesetting"] for now
<thomashoneyman>
as shown in the small paste, i would like package A' to be the same as package A with one attribute overridden
<rotaerk>
okay, brb
<thomashoneyman>
ie. "rec { a = ..., dev = { a' = a.overrideAttrs ... } }"
<gcoakes[m]>
Does anyone here have a working example of using coc.nvim with home-manager? I have everything installed, but the language servers I have installed don't seem to be working.
<thomashoneyman>
but what actuall happens is that A == A'
<sphalerite>
thomashoneyman: overrideAttrs changes what's passed to mkDerivation
<sphalerite>
thomashoneyman: not what's passed to the function.
<rotaerk>
sphalerite, hmm, with modesetting, it no longer gives the error. although, before, when I was using noveau, my system would crash periodically, until I switched to intel+nvidia
<thomashoneyman>
yeesh
<thomashoneyman>
sphalerite: that makes sense
<sphalerite>
thomashoneyman: maybe you want to use callPackage instead of import, and override instead of overrideAttrs
<rotaerk>
so what did setting it to modesetting do?
<rotaerk>
glxgears also runs
<rotaerk>
though I'll have to wait and see if it starts crashing again
<sphalerite>
rotaerk: modesetting is the X driver you usually want to use with intel graphics nowadays, but it's probably now using the intel graphics for everything. Check glxinfo to confirm that
FareWell has joined #nixos
Fare has quit [Ping timeout: 258 seconds]
<FareWell>
Trying to run X with startx, and I only managed to lock up my laptop so I had to force reboot.
<thomashoneyman>
sphalerite: i probably do. still unfamiliar with callPackage so i'll look into it a bit
<rotaerk>
what if I did [ "modesetting" "nvidia" ] ? I couldn't do just [ "nvidia" ]; I think X wouldn't start at all when I did that
<rotaerk>
also what am I looking for in the glxinfo output to see if it's using intel for everything
<sphalerite>
rotaerk: that is, modesetting is a generic X driver that doesn't care what graphics chip or corresponding kernel driver it is
<sphalerite>
rotaerk: as long as the kernel driver implements the relevant API. i915, the intel one, does. nouveau also does. nvidia's proprietary driver doesn't.
<rotaerk>
server glx: SGI; client glx: Mesa Project and SGI; OpenGL: Intel Open Source Technology Center
<rotaerk>
relevant API for what?
zaeph has joined #nixos
<rotaerk>
for modesetting?
<sphalerite>
rotaerk: now, if you don't want to actually use your nvidia graphics you can probably stop there. Otherwise, you'll need some software to support the hybrid graphics stuff. I'm not sure of the current state of affairs, but last time I used nvidia hybrid graphics bumblebee worked well
<sphalerite>
yes, modesetting
<rotaerk>
I tried enabling bumblebee, and when X tried to start, the whole thing froze
<sphalerite>
bumblebee should be a matter of setting hardware.bumblebee.enable = true; and running graphics-intensive stuff with primusrun
<sphalerite>
hm, maybe try that again now that your videoDrivers is set to modesetting only?
<sphalerite>
If that doesn't work I don't know, sorry
<rotaerk>
yep, bumblebee + modesetting had the same result as before: it says the X server has started, but then nothing comes up
<rotaerk>
my main problem is that I have no idea how to begin to diagnose the problems with these things...
<sphalerite>
yeah, hm…
<sphalerite>
maybe there's something in the X logs. Which DM are you using?
<rotaerk>
like with the nvidia proprietary driver, when glxinfo and glxgears fail, how to diagnose *why*
<rotaerk>
or when X server doesn't start with bumblebee ... *why*
<rotaerk>
I'm using xmonad
<rotaerk>
err
<sphalerite>
display manager, the bit where you log in :)
<rotaerk>
whatever the default is; I didn't set that explicitly
<srhb>
rotaerk: I didn't scroll all the way up, but the simple bumblebee setup (just enabled, using intel/modesetting for the xserver driver) usually only fails in that way if the firmware is misconfigured
<srhb>
rotaerk: That is, set to either the discrete or internal, not dynamic or whatever it's called.
<rotaerk>
hmm would that be configured in the BIOS?
<sphalerite>
aaah
<srhb>
Yes
<rotaerk>
lemme check that... brb
<srhb>
rotaerk: Be advised this means you can only boot into X with the NixOS config matching your firmware.
<sphalerite>
srhb: shouldn't modesetting+bumblebee also work with the firmware set to integrated-only?
<sphalerite>
just that bumblebee won't start?
<srhb>
Yeah, that sounds right.
ambro718 has quit [Ping timeout: 268 seconds]
kleisli has joined #nixos
<rotaerk>
i see nothing in my BIOS configuration options for that
<srhb>
rotaerk: That sounds very surprising for an optimus-enabled machine
<srhb>
rotaerk: Are you sure it is?
<rotaerk>
not certain, though i googled the nvidia GPU and it says it is
<FareWell>
after more rebooting than I'd like... it looks like changing the MatchIsTouchpad "on"
<FareWell>
will not help much
<srhb>
rotaerk: This is not a laptop?
thomashoneyman has quit [Quit: Lost terminal]
<FareWell>
there must be some conflict between multiple sections of xorg.conf
<rotaerk>
i linked the laptop above, srhb
<rotaerk>
hp Spectre
<FareWell>
I suspect the problem is a conflict between xorg.conf and /etc/static/X11/xorg.conf.d/40-libinput.conf
<srhb>
rotaerk: Huh, maybe it really isn't configurable on that model. I've never had a laptop with such a setup. Anyway, presumably it can only do hybrid then, and you should be fine, sorry for the red herring.
bvdw has quit [Read error: Connection reset by peer]
<rotaerk>
that's the BIOS config screen, srhb
<FareWell>
Ahem. I tried to comment out the 40-libinput part in libinput.nix, but it's not removed by nixos-rebuild switch -I ..../nixpkgs
<srhb>
rotaerk: Yeah, I think you're right. :)
bvdw has joined #nixos
<rotaerk>
but still, bumblebee fails ... is there a log i can check that would explain why
aleph- has quit [Changing host]
aleph- has joined #nixos
<FareWell>
even if I introduce a syntax error in said file, that doesn't affect nixos-rebuild... wtf?
<rotaerk>
heh, okay, with modesetting, the crashing i was getting is back
drakonis has quit [Quit: WeeChat 2.6]
<rotaerk>
i need to switch back to [ "intel" "nvidia" ]
<rotaerk>
FareWell which file
<elvishjerricco>
Wtf. I'm making a nixos module, and I can do `foo.bar = {};`, but `foo.bar.baz = {};` says the option isn't defined. `options.foo.bar = mkOption { type = submodule { baz = ...; }; };`
<FareWell>
aha, I need to use -I nixpkgs=/path/to/nixpkgs, not -I /path/to/nixpkgs
<elvishjerricco>
Er, `options.foo.bar = mkOption { type = attrsOf (submodule { baz = ...; }); };`
<srhb>
rotaerk: Still attempting bumblebee? If so, you shouldn't have nvidia in there iirc.
delli3_ has quit [Ping timeout: 260 seconds]
<rotaerk>
bumblebee+modesetting and bumblebee+nvidia both fail
<srhb>
elvishjerricco: Assuming those ellipsis also have mkOption in there?
<FareWell>
at which point, it says that console is an invalid option... is that defined in 19.09 but not unstable?
<rotaerk>
so I'm abandoning bumblebee entirely
<elvishjerricco>
srhb: Yea, and also there's more typos in that code :P But I think I figured it out
<srhb>
rotaerk: The "correct" config is bumblebee + intel/modesetting. Not nvidia. But OK :)
<srhb>
The even correcter config is hatemail sent in Nvidias direction.
<rotaerk>
bumblebee doesn't work at all, and nouveau or modesetting (without bumblebee) result in periodic crashing of X
<srhb>
Great..
Henson has joined #nixos
delli3 has joined #nixos
<elvishjerricco>
Agh, more confusion. Gotta be some kinda typo but I can't see it
<rotaerk>
and nvidia on its own doesn't start X
<rotaerk>
but [ "intel" "nvidia" ] starts X and doesn't cause crashes
<rotaerk>
but then glx* stuff doesn't work
<srhb>
elvishjerricco: You may have elided too much, I guess... It looked like you wanted foo.bar.$NAME.baz = ...
<rotaerk>
also, weird, while booting, my laptop just gave me a warning against using non-HP battery
<rotaerk>
the battery is what came with the laptop from hp...
<elvishjerricco>
srhb: Yea, so I have `foo = mkOption { type = attrsOf (submodule { bar = mkOption { type = attrsOf (submodule { ... }) }; }); };`, but I can't set `foo.x.bar.y.option`
<Henson>
how does one play CSS encrypted DVDs in NixOS?
<elvishjerricco>
I get `option foo.x.bar defined in /foo.nix does not exist`
<FareWell>
rotaerk: did you try nouveau ?
<rotaerk>
yeah, random crashing, though glx* works
<rotaerk>
haven't found a setup that works in all ways
<FareWell>
:-/
<FareWell>
I'm reminded of the bad old days, mixing and matching X server and X configuration files on RedHat to run with a Matrox server.
zaeph has quit [Ping timeout: 248 seconds]
<FareWell>
Henson: doesn't it "just work" ?
<FareWell>
nixos supports libdvdcss
<srhb>
elvishjerricco: submodules need an option key, don't they?
<elvishjerricco>
Been doing this a long time and I still make that mistake :P
<srhb>
elvishjerricco: But I think your shape is still not quite what you want. Assuming baz is a str what you'd get is.. `config.foo.firstsub.bar.secondsub.baz = "foo";`
<syd>
I'd like to build my nixos configuration from another machine , just to check if it will work ahead of time with a few different configurations.
<srhb>
(Or maybe it is, not sure)
<syd>
but I can't seem to figure out what the expression is that I want
<syd>
I currently have `import <nixos> { configuration = import ./quintus.nix; }`
<Henson>
FareWell: I guess I was using the wrong DVD player. I'm used to using Xine, and with that encrypted DVDs don't work. But with VLC they seem to work just fine.
<syd>
but then I get `error: The option `services.deluge.authFile' defined in `/home/syd/src/nix-depot/modules/media.nix' does not exist.`
<syd>
and that's strange
<syd>
because I can see that that option is defined
<srhb>
syd: The name nixpkgs comes from there. That's too old
<syd>
trying `nix-channel --update nixpkgs` as root
<srhb>
syd: You're probably confusing root and user channels.
<syd>
jup that was it!
<syd>
woot, so the `nix-info` command saved me. Thanks srhb!
vika_nezrimaya has joined #nixos
chloekek has quit [Ping timeout: 240 seconds]
Arahael has quit [Ping timeout: 246 seconds]
reallymemorable has quit [Quit: reallymemorable]
<evanjs>
does grub.useOSProber automatically check other drives for boot entries? Or rather, how do I make sure e.g. gentoo is added to my grub menu on NixOS?
<evanjs>
Do I have to manually add it to extraEntries or etc?
bhipple has quit [Ping timeout: 258 seconds]
<syd>
hm, I managed to get the thing working, thanks srhb!
<syd>
now, I am making an `all.nix` file that contain all my nixos system configs in an attrset
<syd>
I'm trying to use `nix-build all.nix` and that finishes immediately
<syd>
but when I run `nix-build all.nix -A quintus`, it builds the `quintus` attribute