<baum>
ottidmes: i assume the following works on your machine too?: echo $PWD | xargs nix-shell
<ottidmes>
yes
<ottidmes>
for you it does not work?
jrolfs has joined #nixos
<ik3123>
ottidmes: not believing it was the virtio block that was the issue, there's a section in the wiki where there's a provisioning boilerplate using them. https://nixos.wiki/wiki/Virtualization_in_NixOS
<baum>
ottidmes: nope, on neither of my machines, if i add --verbose i can see that nix evaluaes nixpkgs, however it doesn't switch to the correct shell
<baum>
*evaluates
<ottidmes>
ik3123: yeah, I know of a few people using it in VMs that way, but still, it could be related somehow, its a hard problem to pinpoint
Mr_Keyser_Soze has quit [Remote host closed the connection]
<kini>
What's nixos's policy for updating the default linux kernel version?
epicmetal has joined #nixos
<ashkitten>
does anyone have firefox nightly working
ng0 has quit [Quit: Alexa, when is the end of world?]
<samueldr>
kini: not written, but "current LTS at release time"
<samueldr>
(not written AFAIK)
<kini>
cool, makes sense, thanks
<samueldr>
si it's going to be updated to 4.19, eventually
<samueldr>
especially important as current-non-LTS sometimes gets EOL'd before the stable version of nixos is dropped :)
endformationage has quit [Ping timeout: 240 seconds]
<kini>
I guess the kernel devs set one LTS per year or so, so that means autumn releases of nixos will probably have the same kernel versions as the spring release of the same year
<samueldr>
I don't know what's their policy on LTS versions
<kini>
I don't know either, just guessing based on history
<rsully>
is kernel version tied to nixos closely or is there a way to use newer kernels quicker? or is it unsupported
<samueldr>
as long as you're going with released kernels, easy as it can be
<samueldr>
kernel updates are backported, including newly released versions
<rsully>
is stateVersion to track a nix release, or does that property not really do anything? sorry I have not poked around much, very overwhelming
<samueldr>
,stateVersion rsully
<{^_^}>
rsully: Setting stateVersion to the latest release doesn't update anything and can only break your setup at best. If you want to regardless, Ctrl-F for "stateVersion" in https://nixos.org/nixos/manual/release-notes.html to see things that need to be manually upgraded with the new value
Mr_Keyser_Soze has quit [Ping timeout: 246 seconds]
<simpson>
rsully: To add onto that: The purpose of stateVersion is to be set *once*, when you install. You can change it after reading release notes. The entire goal is to enable most folks to cleanly upgrade between releases without worrying about having to upgrade Pg tables or other stateful data.
<rsully>
is the proper way to update to use nix-rebuild and it will update it for you?
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fpD95
jperras has quit [Ping timeout: 250 seconds]
jperras has joined #nixos
nDuff has joined #nixos
cbarrett has quit []
<{^_^}>
[nixpkgs] @ivan opened pull request #51644 → erlang: disable parallel building → https://git.io/fpDHS
<nDuff>
When are paths used in a derivation added to the store? I have some code trying to iterate over a list to make only a subset of the source tree we're in relevant to a build (subpackages=[ "one/two" "three/four" ]; extraSources = map (x: ../../. + x) subpackages;), but I end up with paths that point to my local source tree, not to mirrors of that content in the store -- which is nasty, since I'm creating symlinks to the result of those locations (and *those
<nDuff>
symlinks* into my home directory as a user are ending up in the store).
<dramforever>
Seems like everything is *being* moved into just 'nix', *gradually*
<dramforever>
So stuff like nix-env are still there
<iqubic>
How long does: "Building system configuration" take?
cyounkin_ has joined #nixos
<dramforever>
Shouldn't take long
<dramforever>
I mean, shouldn't take long until more log appears
<ottidmes>
iqubic: would depend on your configuration and the system you are building it on, would it not
<iqubic>
It would.
<iqubic>
Oh, there it goes, working on something else.
Mr_Keyser_Soze has quit [Ping timeout: 245 seconds]
cyounkins has quit [Ping timeout: 250 seconds]
astrofog has quit [Remote host closed the connection]
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
graphene has quit [Remote host closed the connection]
graphene has joined #nixos
Radivarig has quit [Ping timeout: 252 seconds]
Mr_Keyser_Soze has joined #nixos
erratic has joined #nixos
the-kenny has quit [Ping timeout: 250 seconds]
cyounkin_ has quit [Remote host closed the connection]
the-kenny has joined #nixos
mmlb has quit [Read error: Connection reset by peer]
mmlb has joined #nixos
Mr_Keyser_Soze has quit [Ping timeout: 246 seconds]
<iqubic>
Wow. Still building and compiling stuff.
<dramforever>
Compiling?
<iqubic>
Geez, how long is this gonna take.
<iqubic>
dramforever: Does "nixos-rebuild switch --upgrade" compile stuff? I thought it did.
<iqubic>
I might be wrong.
<nDuff>
...if there's compiling happening, presumably something about your box is changing the derivations in such a way that the upstream cached versions can't be used.
<dramforever>
Shouldn't compile much
<iqubic>
I see. I doubt stuff is being compiled than.
<iqubic>
But it is definately building a ton of stuff.
<iqubic>
Is building different than compiling?
<dramforever>
Oh that's actually normal
epicmetal has quit [Ping timeout: 244 seconds]
<iqubic>
Oh, looks like it is now building Texlive.full. I see why that's taking a while.
<dramforever>
NixOS also builds configuration files and the like
<meizikyn>
kinda, building in nixos is mostly decompressing binary content
<dramforever>
Hmm
<meizikyn>
compiling is... compiling
<iqubic>
No. Actually this is building texlive drvs
<iqubic>
It's building all the plugins and packages.
<dramforever>
I don't know about textlive
<dramforever>
*texlive
bobnormal has joined #nixos
<ldlework>
hi dramforever
<iqubic>
It's a version of LaTeX.
klntsky has quit [Ping timeout: 256 seconds]
<dramforever>
I mean the internals, like what is build when installing and the like
<dramforever>
ldlework: hi
<dramforever>
iqubic: maybe you can ctrl-c it, then 'nixos-rebuild dry-build', then paste the log
sir_guy_carleton has joined #nixos
<dramforever>
And someone might be able to tell if it's normal
<bobnormal>
Suddenly having issues with arduino IDE - calls 'avr-g++' which appears via 'strace' to search for various locales then fail to execute, returning 127 (file not found). No idea what change caused this. Ideas?
<{^_^}>
'avr-g's karma got increased to 1
<dramforever>
lol
* ivan
learns SELECT * FROM builds WHERE (stoptime - starttime) > 7200 ORDER BY job;
<bobnormal>
The only locale in the glibc directory it is searching is en_GB ... although I have tried reinstalling glibc with nix-env -i glibc and also have en_US.UTF-8 set as i18n.defaultLocale in configuration.nix
<bobnormal>
Going to try setting that to "en_GB" and rebuilding/rebooting. Seems an obtuse fix for a weird issue though. Not sure of the nix way to resolve this.
bobnormal has quit [Quit: Leaving]
<dramforever>
'Suddenly having' <- It's not happening before?
<dramforever>
Like, it worked on NixOS before but not now?
aleph- has quit [Ping timeout: 250 seconds]
Mr_Keyser_Soze has joined #nixos
<dramforever>
bobnormal: Can you try 'file -L $(which avr-g++)' or something like that?
<{^_^}>
avr-g's karma got increased to 1
<dramforever>
ldd $(which avr-g++)
<{^_^}>
avr-g's karma got increased to 2
<dramforever>
I'm not actually sure what happened, but I think it's got something to do with shared libraries
bobnormal has joined #nixos
rprije has quit [Ping timeout: 250 seconds]
<dramforever>
I feel like the locale stuff are just glibc trying to print an error message, but meh, dunno
<dramforever>
This is incorrect -> /lib64/ld-linux-x86-64.so.2
<ottidmes>
infinisil: its also not possible to track say the last N user names that said something, and just check if its in that list?
hakujin has quit [Ping timeout: 250 seconds]
<ottidmes>
infinisil: or do you only have access to things prefixed with ,
<dramforever>
bobnormal: Binaries packaged for other distributions don't generally work on NixOS
<dramforever>
Where did you get this arduino ide?
<infinisil>
ottidmes: I could do that yeah
<bobnormal>
dramforever: ok, it was working previously. i think maybe it was arduino IDE's self update
<dramforever>
(and avr-gxx)
<dramforever>
oh
<bobnormal>
dramforever: i guess i should delete its whole tree and reinstall it
<bobnormal>
dramforever: thanks
<dramforever>
self update hmm
<dramforever>
you're welcome
<bobnormal>
dramforever: its trying to make things easy for people... 'board manager' and 'library manager' built in. i guess the former destroyed the toolchain with incorrect dynamic linking paths.
<infinisil>
ottidmes: I'm restricted by the interface, which is that I only receive every message in the form { channel, message } and can only send in the form { channel, message }
<ottidmes>
infinisil: ah ok, minimal as it is, it should be enough to implement such a feature, keeping track of the N last users that said something
<dramforever>
Maybe only register '<nick>++' messages that are alone on a line only?
<{^_^}>
'<nick>'s karma got increased to 1
<dramforever>
Dang it
<infinisil>
Ah, I can just add some more blacklisted names
<dramforever>
People don't usually randomly say 'g<plus><plus>' without context
<infinisil>
avr-g, and g should do it
<dramforever>
That sounds like not worth the effort
<infinisil>
Why not?
<dramforever>
I mean 'keeping track of the N last users that said something'
<dramforever>
I forgot to put a mention
<ottidmes>
its a proper fix at least, there are so many context you want to write ++ on these channels...
<infinisil>
Ah yeah, a simple blacklist will cover 99.9% of these mis-increases while being super easy to implement
<bobnormal>
OK that fixed it. nix-env -e arduino && mv ~/.arduino15 ~/.arduino15-old && nix-env -i arduino
<dramforever>
the 'nix-env -e' should be completely unnecessary
<dramforever>
And 'nix-env -i'
<dramforever>
Also 'nix-env -i glibc' shouldn't really do anything
rprije has joined #nixos
<bobnormal>
thanks, anyway is there somewhere to document this for future users?
Mr_Keyser_Soze has joined #nixos
<dramforever>
Document 'Don't let Arduino IDE auto-upgrade?'
pxc has quit [Ping timeout: 252 seconds]
Mr_Keyser_Soze99 has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @vbgl pushed commit from @marsam to master « flow: 0.86.0 -> 0.87.0 »: https://git.io/fpDdk
<iqubic>
I wonder if anyone still codes in c++ anymore
<iqubic>
Looks like that's been blacklisted.
<Ralith>
...?
<iqubic>
g++
<iqubic>
And that's also working as expected.
<ottidmes>
i++
<{^_^}>
i's karma got increased to 1
<dramforever>
That sounds rare enough
rprije has quit [Ping timeout: 268 seconds]
iqubic_ has joined #nixos
iqubic has quit [Ping timeout: 256 seconds]
<iqubic_>
Well, this is weird. I upgraded my system to 18.09 and did a reboot of my laptop. Now emacs is complaining about not being able to find the font specified in the config file.
iqubic_ is now known as iqubic
<iqubic>
So I have no idea what's going on here.
dramforever has quit [Ping timeout: 256 seconds]
<iqubic>
Is there a way to check if a certain font is installed on my machine?
<Ralith>
iqubic: fc-list, fc-match
<iqubic>
I just ran this:
<iqubic>
fc-match "knack nerd font"
<iqubic>
It came back with this:
<iqubic>
FreeMono.ttf: "FreeMono" "Regular"
<iqubic>
What the heck?!?!
<iqubic>
I can't tell if knack nerd font is installed or not any more.
<gpstrucker>
Hello. I just started using NixOS. Pidgin keeps crashing. What irc client do you folks recommend/use?
drakonis_ has quit [Read error: Connection reset by peer]
epicmetal has quit [Read error: Connection reset by peer]
rsully has quit [Quit: rsully]
<Myrl-saki>
gpstrucker: Weechat.
<Myrl-saki>
Or irssi.
<Myrl-saki>
If you're comfortable with your terminal, that is.
<gpstrucker>
Thank you, I'll give those a try. Terminal is not a problem for me.
<iqubic>
Does anyone know how to fix my font issues?
<gpstrucker>
hmmm, just did a nix-env -qa weechat and it returned no matches
<{^_^}>
[nixpkgs] @aszlig pushed to master « nixos/tests: Fix wrong arch in runInMachine test »: https://git.io/fpDbG
<gpstrucker>
Did a nix search and found weechat-bin-env. Is that what I need?
<iqubic>
I think so.
<gpstrucker>
Ok
<iqubic>
Yes, that is indeed what you want.
<gpstrucker>
Installing now. I appreciate the help. Now off to learn how to use it LOL
<gpstrucker>
:quit
gpstrucker has left #nixos [#nixos]
<iqubic>
So I think my fonts are screwed up now.
<iqubic>
Well, this is weird. I upgraded my system to 18.09 and did a reboot of my laptop. Now emacs is complaining about not being able to find the font specified in the config file.
<iqubic>
My fonts were working before that reboot.
<iqubic>
I want to use "Knack Nerd Font"
<Ralith>
try nuking ~/.cache/fontconfig
<Ralith>
chromium likes to mangle it sometimes
<iqubic>
I'm not using Chromium. I'm using firefox.
<iqubic>
I don't even have chromium installed right now.
hamishmack has joined #nixos
<Ralith>
or don't, up to you 🤷
hakujin has joined #nixos
<iqubic>
Would that help me fix this issue.
<iqubic>
It's Emacs, a text editor, that is complaining about fonts.
mkoenig has joined #nixos
<mdash>
iqubic: emacs uses fontconfig too IIRC
<iqubic>
So?
<iqubic>
I'm just concerned that I have permenantly screwed up my fonts.
<iqubic>
Should I nuke the fontconfig file, will that fix this stuff?
<iqubic>
I literally jus upgraded to 18.09 today.
reinhardt has joined #nixos
Mr_Keyser_Soze has quit [Remote host closed the connection]
jrolfs has joined #nixos
rprije has joined #nixos
jrolfs has quit [Ping timeout: 246 seconds]
rprije has quit [Remote host closed the connection]
rprije has joined #nixos
vk3wtf has quit [Ping timeout: 268 seconds]
mmlb has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @gnidorah opened pull request #51650 → me_cleaner: init at 1.2 → https://git.io/fpDxN
rawreraw has quit [Ping timeout: 245 seconds]
crmlt has joined #nixos
rawreraw has joined #nixos
<{^_^}>
[nixpkgs] @hedning opened pull request #51651 → cyrus-sasl: Remove private libs from .la files (fixes spice build) → https://git.io/fpDpb
jrolfs has joined #nixos
tv has quit [Quit: WeeChat 2.0]
jrolfs has quit [Ping timeout: 250 seconds]
tv has joined #nixos
cbarrett has joined #nixos
crmlt has quit [Remote host closed the connection]
agander has joined #nixos
pxc has joined #nixos
[Leary] has joined #nixos
metastance has joined #nixos
Ariakenom has joined #nixos
tv has quit [Quit: WeeChat 2.0]
Lears has quit [Ping timeout: 240 seconds]
tv has joined #nixos
tv has quit [Client Quit]
tv has joined #nixos
periklis has quit [Remote host closed the connection]
periklis has joined #nixos
ThatDocsLady has joined #nixos
Tucky has joined #nixos
rawreraw has quit [Ping timeout: 268 seconds]
johanot has joined #nixos
freeman42x has joined #nixos
rawreraw has joined #nixos
rawreraw has quit [Client Quit]
rawreraw has joined #nixos
freeman42x has quit [Ping timeout: 252 seconds]
rawreraw has quit [Client Quit]
rawreraw has joined #nixos
nix` has joined #nixos
civodul has joined #nixos
crmlt has joined #nixos
crmlt has quit [Remote host closed the connection]
hotfuzz_ has quit [Ping timeout: 268 seconds]
bobnormal has quit [Quit: Leaving]
tpanum has joined #nixos
<tpanum>
When launching Inkscape with just a WM active, I get 'Gtk::IconThemeError' and it crashes instantly. How do I get Inkscape to pickup the themes correctly?
hke has quit [Read error: Connection reset by peer]
rawreraw has quit [Quit: WeeChat 2.3]
rawreraw has joined #nixos
hke has joined #nixos
goibhniu has quit [Ping timeout: 246 seconds]
sigmundv has joined #nixos
goibhniu has joined #nixos
<Izorkin>
How to fix error. Packages - nginx unit - https://github.com/NixOS/nixpkgs/pull/51612 - Error load php extension - Unable to load dynamic library '/nix/store/...-php-apcu-5.1.11/lib/php/extensions/apcu.so (/nix/store/...-php-7.2.12/lib/php/extensions//nix/store/...-php-apcu-5.1.11/lib/php/extensions/apcu.so.so: cannot open shared object file: No such file or directory))
mmlb has quit [Read error: Connection reset by peer]
mmlb has joined #nixos
<bkchr>
Hey. Since a long time I having problems with rust's cargo and git dependencies, sometimes they just timeout with a ssl error. Does someone else observed something similar on Nixos?
<{^_^}>
[nixpkgs] @Mic92 merged pull request #51618 → crashplan: mark as broken → https://git.io/fpMV4
<CMCDragonkai>
How come `buildOcaml` results in `mkdir: cannot create directory '/nix/store/hpvrspda6i7mg755jdpzfzra48v7v3lx-ocaml-demo-0.0.1': Read-only file system`?
<CMCDragonkai>
but if I use `mkDerivation` normally this error doesn't occur?
<srhb>
avn: That's not enough to go on. (Aside from it being a terrible idea to use publically assigned ranges internally, it should be fine) -- interface, routes, etc. still needed. But my guess is that your problem is on the VMWare side, not NixOS.
<avn>
srhb: is totally internal, I am on phone with my collegue, who are inside ;)
<avn>
And I agree by all points ;)
dmc has quit [Changing host]
dmc has joined #nixos
<avn>
btw, would vmware guest tools is redistributable? would be nice to have network drivers out of box
rawreraw has joined #nixos
<srhb>
avn: I thought it was included in recent kernels? I may have just made that up, but check. :)
<Mic92>
Where is nativeBuildInputs evaluated? Or how does it pick the right architecture? There are some cases where binaries are passed via function and non-native binaries end up in nativeBuildInputs
<{^_^}>
#49526 (by symphorien, 5 weeks ago, open): nativeBuildInputs gives non native dependencies when used with overrides in all-packages.nix
<symphorien>
Apparently the magic happens in callPackage if the argument is implicit
<{^_^}>
[nixpkgs] @joachifm pushed commit from @dtzWill to release-18.09 « tor: 0.3.4.8 -> 0.3.4.9 »: https://git.io/fpyl1
<{^_^}>
[nixpkgs] @pmiddend opened pull request #51654 → libopusenc: init at 0.2.1 → https://git.io/fpy8Z
<{^_^}>
[nixpkgs] @joachifm opened pull request #51655 → uwsgi: fix build when withSystemd = false → https://git.io/fpy8W
vidbina has quit [Read error: Connection reset by peer]
<jophish>
Ericson2314: to be honest, I think having such a deep meaning for target is banannas from the start
<jophish>
there's no really good reason why the binaries my compiler can produce should depend at all on what platform the compiler itself is running on
<jophish>
LLVM has the right idea here
<jophish>
GHC can be excused I think because of template haskell complications
vidbina has joined #nixos
crmlt has joined #nixos
<typetetris>
It would be really nice, if someone could check if I got that right: I write a nix expression, which evaluates to a derivation, which is simply a special set. nix-instantiate then turns that into a store derivation, whose store path ends in `.drv` and nix-store --realise builds, whatever is described by that derivation, and puts the output paths in the store.
crmlt has quit [Client Quit]
hke has quit [Quit: bye]
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
aramiscd has quit [Ping timeout: 246 seconds]
<tilpner>
typetetris - nix-instantiate does both the "evaluates to a derivation" and "turns that into a store derivation"
crmlt has joined #nixos
<typetetris>
tilpner - thanks, what about evaluating a nix expression in the nix repl, does that also produce the store derivation, if the result is a derivation ( the special set value) ?
<tilpner>
I think it does
<typetetris>
Ok, thanks.
<fendor>
my terminal does not seem to support unicode, i repeatedly get the error, <stdout>: commitBuffer: invalid argument (invalid character) when trying to print unicode characters
<etu>
Which terminal?
<fendor>
etu, konsole
<etu>
And which application?
<srhb>
fendor: Is your locale UTF-8 supporting?
<fendor>
srhb, i dont think so, it seems to be set to "C" whatever that is
<etu>
fendor: That's probably your issue
<fendor>
etu, happened with stack2nix and `stack bench`, criterion output
<srhb>
fendor: Right, then it should break like you're seeing. :)
crmlt has quit [Remote host closed the connection]
<fendor>
srhb, what value should the locale have?
<srhb>
I don't know what localization you want re. language and number formats, but *that* plus UTF-8.
<fendor>
srhb, I am trying to set it in /etc/nixos/configuration.nix explicitly and reboot, I dont think I have it set in either bashrc nor fish.config
abcrawf has quit [Remote host closed the connection]
fendor has quit [Remote host closed the connection]
<locallycompact>
/build/.attr-0: line 15: stack: command not found
<tilpner>
Maybe add nix to contents of stackBaseImage
<gchristensen>
locallycompact: extraCommands doesn't run inside the container, it can only modify the layer files directly -- not use the programs you've added
fendor has joined #nixos
<fendor>
srhb, i think, now i found the culprit and it wasnt actually my mistake! i think...
<locallycompact>
gchristensen, I need to prebuild a bunch of things with stack into the docker
<gchristensen>
locallycompact: you could install rio via adding haskellPackages.rio to the `contents` list
<locallycompact>
but then stack won't pick that up
alex`` has joined #nixos
<srhb>
fendor: Rolling your own involves using localegen with, say, en_US, en_US.UTF-8, de_DE and de_DE.UTF-8 if I understand your requirements correctly. I'm sure google has some opinion on how to do this, as it's probably used quite often.
<locallycompact>
and it will be at a random version unrelated to the stack manifest
<srhb>
Though apparently not as often as en_DK.UTF-8 :P
<gchristensen>
locallycompact: Nix's dockerTools isn't behaving like Docker. the point of Nix's dockertools is to use Nix to build the software, and place the result of it in a container. I'm not sure youc an really get what you need out of Nix here -- you won't be able to access the internet during the build, for example
<tilpner>
Building glibcLocales for en_DE* or en_AU* fails with "unsupported locales detected"
<fendor>
srhb, KDE knows en_AT, so maybe nixos just doesnt have many german people :P
<srhb>
fendor: It's really not very NixOS specific. KDE probably just does the naive thing and says "OK, I can construct the string en_AT -- no problem!"
<srhb>
fendor: That does not make it available in the locale definitions on the system. I bet it's the same on many other distros, since really they'll be using approximately the same locale archive that we are (namely the one from glibc)
<fendor>
srhb, it actively offers it, I dont understand why it would do that
<srhb>
Hm, okay, me neither.
<fendor>
also it kind of works, since the time is displayed in the correct format then
<srhb>
Regardless, what you need to is to actually create it if you want to use it correctly :-P
fendor has quit [Remote host closed the connection]
<tilpner>
When you do "locale", it prints different ones for different uses
<tilpner>
Perhaps you want LC_TIME to be de_DE, and en_US elsewhere?
<FireFly>
locales.. both a blessing and a curse, hehe
fendor has joined #nixos
<tilpner>
i18n.extraLocaleSettings
<tilpner>
Example from man: { LC_MESSAGES = "en_US.UTF-8"; LC_TIME = "de_DE.UTF-8"; }
<fendor>
tilpner, options does not seem to know that
<tilpner>
Try it anyway
<srhb>
I think tilpner's suggestion is by far the simplest.
<srhb>
All the options should make it pretty easy to set your own frankenlocale.
<locallycompact>
gchristensen, that's confusing but thanks
<fendor>
the option: i18n.extraLocaleSettings does not exist :(
<tilpner>
Huh, maybe it's 18.09+
<srhb>
fendor: It's implemented in terms of environment.sessionVariables, so you can just do that.
<srhb>
Or upgrade. :-P
<tilpner>
Not just sessionVariables, you'd also need to set /etc/locale.conf
<srhb>
Oh, of course, yes.
<srhb>
Still I don't think that'll matter much to most applications..
<fendor>
tilpner, I tried to add it to my /etc/nixos/configuration.nix, right?
<tilpner>
Yeah, and your nixpkgs is probably too old
<srhb>
Seems it's in unstable only :)
<tilpner>
You can upgrade, or look at the module from unstable and do what it does
<srhb>
fendor: You should probably not just go to unstable for one feature. There are caveats.
<srhb>
Many, many caveats.
<fendor>
srhb, what is unstable ad that sense? 19.03, or unstable channel? because i already have the unstable channel
<srhb>
fendor: It will eventually become 19.03 yes.
<srhb>
fendor: (It's master, once it's passed the NixOS test set)
<srhb>
fendor: Since the config option does not exist for you, your NixOS is presumably tracking a stable channel. "Upgrading" in this sense would mean your entire NixOS system now tracks unstable. This may be undesirable.
<fendor>
srhb, undesirable since things break? ^^
<srhb>
Well, even less warranty than normal, at the very least.
<srhb>
fwiw I seldom encounter breakage that I can't fix, but ymmv.
<fendor>
ok, maybe ill just endure the US locale for now
hakujin has joined #nixos
fendor has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @erictapen opened pull request #51664 → pythonPackages.tlslite-ng: init at 0.7.5, migrate tlslite references to it. → https://git.io/fpyVf
WhittlesJr has joined #nixos
<{^_^}>
[nixpkgs] @FRidh pushed 325 commits to python-unstable: https://git.io/fpyV3
<{^_^}>
[nixpkgs] @FRidh pushed to master « buildPython*: add updateScript to passthru »: https://git.io/fpyVQ
<avn>
etu: does it attempt install EFI grub, instead bios one? (is a VM with single ext4 partition, I have no reasons to play with EFI)
<etu>
avn: Well, do you have a GPT partition?
iyzsong has quit [Read error: Connection reset by peer]
<avn>
etu: GPT assume EFI-only?
rawreraw has quit [Ping timeout: 250 seconds]
<etu>
avn: "GPT is modern and has many advantages over MBR. However, there are also some issues with GPT booting in Legacy BIOS mode. GPT is part of the EFI specification, of course it will work best in UEFI mode. But maybe it will not be compatible and can not boot up on the BIOS computer, see more here." -- https://www.aioboot.com/en/gpt-legacy/
<etu>
avn: So yeah, I think it is
dbmikus has joined #nixos
<clever>
etu: if you make a bios boot partition (1mb, no fs, never mounted), then grub can do a legacy install just fine on gpt
<clever>
the bios boot partition is not the /boot partition
<clever>
and you can still choose to either have a /boot partition, or a /boot directory on the / partition
Ariakenom has quit [Ping timeout: 268 seconds]
<iqubic>
So I just updated to 18.09 yesterday, and now my fonts are screwed up.
<iqubic>
Emacs is now complaining that it can't find "Knack Nerd Font"
<Ericson2314>
I've been poking at Meson with the absurdly long term goal of rewriting the GCC build system so building all the libraries/components separately can be trivially supported
<Ericson2314>
After which it would be easier to justify getting rid of the target specific ifdefs and making them ifs
hakujin has quit [Ping timeout: 250 seconds]
<Ericson2314>
On the other hand, nixpkgs's default LLVM no longer has all the backends as people didn't like the slow build times
<avn>
clever: ty ;) we just redone with old-style mbr ;)
ThatDocsLady has quit [Ping timeout: 252 seconds]
freeman42x has joined #nixos
ntd has joined #nixos
vidbina has quit [Read error: Connection reset by peer]
ntd has quit [Quit: ntd]
pie___ has quit [Ping timeout: 264 seconds]
<sphalerite>
Ericson2314: ooooh
oida has quit [Remote host closed the connection]
<sphalerite>
jophish: does template haskell really require a special status for the target arch conceptually, if the compiler were to support multiple targets anyway?
klntsky has joined #nixos
Phillemann has joined #nixos
<Ericson2314>
Template Haskell should be changed too
<Ericson2314>
I started writing a GHC proposal for that, I'd anyone wants to collaborate
Phillemann has left #nixos ["WeeChat 2.3"]
<Ericson2314>
*if anyone wants to
ntd has joined #nixos
locallycompact has quit [Ping timeout: 250 seconds]
hakujin has joined #nixos
<jabranham>
Can I get nix-env --list-generations to tell me what changed between generation X and generation X-1?
<tilpner>
Is the store path filename schema stable? Should I expect the hash in /nix/store/rgmc4d3spji36n2l1sicm80yq79dpcc2-hello-2.10 to have different sizes in exotic configurations?
<jabranham>
gchristensen: I don't understand how that does what I want. I want to see what changed between generations but Gabriel's is focused on diffing derivations, right?
<gchristensen>
ah, I see. nix-diff can do it too, and shows pretty in-depth levels of what changed
endformationage has joined #nixos
<{^_^}>
[cabal2nix] @ttuegel opened pull request #394 → Include cabal-version field inferred by hpack → https://git.io/fpyM3
<gchristensen>
like it can show you that not only did PkgFoo upgrade from 1.0.0 to 1.0.1, but also how the build instructions changed
<qyliss^work>
generations are just derivations, also
<jabranham>
gchristensen: awesome. I'll keep that in mind for the future. For now, nix-diff.sh seems to have done the job. In that it told me enough that the issue I'm running into is of my own making, not because something updated and broke stuff. :-/
<gchristensen>
neat
cranjery has joined #nixos
erasmas has joined #nixos
Ariakenom has joined #nixos
sigmundv has quit [Ping timeout: 250 seconds]
<yorick>
tilpner: storePathHashLen is hardcoded in store-api.cc
<yorick>
.h*
<tilpner>
Oh, sorry, should have checked there :/
<tilpner>
yorick++
<{^_^}>
yorick's karma got increased to 1
rsully has joined #nixos
civodul has joined #nixos
Ariakenom has quit [Read error: Connection reset by peer]
lnikkila has joined #nixos
lnikkila has quit [Client Quit]
pxc has joined #nixos
<tilpner>
Who would I ask about fair usage policies of cache.nixos.org?
<fusion809>
Hi, what option shows all the metadata belonging to a specific installed package (e.g. version, description, maintainer, upstream URL, etc.)? The Nix equivalent to the RPM command rpm -qi package and the pacman, pacman -Qi package
<gchristensen>
tilpner: how can I help?
Mr_Keyser_Soze has joined #nixos
<tilpner>
gchristensen - How costly are .narinfo queries, and how many of them are acceptable to perform concurrently? Are there burst/daily limits I should try to keep within?
<tilpner>
(There's no way to check thousands of .narinfos at once, is there?)
<gchristensen>
right, there is not
<gchristensen>
narinfo queries cost about $0.0004 per thousand, and there is no specified policy w.r.t. burst / daily limits. it sounds like you have a Big Project in mind, maybe you could share some info privately? the nixos foundation does, of course, appreciate support to help cover costs, something we can talk about too
Radivarig has joined #nixos
__Sander__ has quit [Quit: Konversation terminated!]
<tilpner>
It's not a big project, I'm just thinking about a potential nix-weather
<gchristensen>
gotcha
<gchristensen>
we'd probably want to calculate and cache the data, but go hack and have fun :P
<tilpner>
Yes, caching helps a little, but not all that much if you iterate on nixpkgs revisions
Mr_Keyser_Soze has quit [Ping timeout: 250 seconds]
<gchristensen>
sure, go for it :P `nix-env`'s `--query` option has the flag `--status` which checks to see if the drv is cached, and that will issue around 50k requests with a loose enough query
<gchristensen>
tilpner: cache.nixos.org is pointing to Fastly as a CDN, with S3 behind it. you certainly won't take it down.
<tilpner>
gchristensen - I don't expect to, but I don't want Fastly to get angry at you, or rate-limit me
<tilpner>
gchristensen - I just commented-in the querying code, it currently queries at 840 rqs/s
<gchristensen>
we get a lot of requests / day, it shouldn't be a problem :)
<tilpner>
Okay :)
<averell>
I need some advice, if this is the correct channel; I'm trying to get nvidia-docker to work, and it basically does but one problem remains: https://github.com/NixOS/patchelf/issues/44 (after patchelf touches a .so, ldconfig cannot read it anymore). These docker containers mount nvidia libs (and runs an ldconfig on them), so that doesn't work anymore. Now I have a few options (rebuild nvidia drivers, fix
<{^_^}>
patchelf#44 (by darealshinji, 4 years ago, open): ldconfig warns about truncated libraries
<averell>
libs/fix patchelf, fix ldconfig). I would prefer a patchelf/lib fix, but I don't know if that is possible.
<fusion809>
jabranham: thanks :)
<averell>
I basically have to move .dynstr back to the first LOAD segment. does anyone know the constraints on that, or if it is possible?
<gchristensen>
tilpner: we seem to average ~185k requests/hr
fendor has quit [Ping timeout: 245 seconds]
<tilpner>
That's just 52/s
pie___ has joined #nixos
<Ankhers>
Can home manager be used to setup postgres service?
<gchristensen>
yeah, indeed
<tilpner>
So if I send 840/s, even if just in bursts, it would at least be visible in the fancy graphs I asusme you have :)
<gchristensen>
it might be, sure
<gchristensen>
not sure anyone'd care
<tilpner>
"Not sure" is bad, I need a "sure that nobody cares" :P
<tilpner>
But I can of course throttle it
<gchristensen>
I think you're over worried
<gchristensen>
if there is a problem, we'll reach out. you won't be the only person issung hundreds of requests/s.
hakujin has joined #nixos
<tilpner>
Not worried anymore after your initial reply. Thank you!
<gchristensen>
looking here, hthere are sections of around 2kreq/sec
betaboon has joined #nixos
pie___ has quit [Ping timeout: 252 seconds]
civodul has quit [Quit: ERC (IRC client for Emacs 26.1)]
<cranjery>
I'm trying to build a derivation using cmake with a dependency on opencv, but the cmake include directory isn't getting passed to the compiler. Does anyone know how to fix this?
ahawkins has quit [Quit: Connection closed for inactivity]
rhalff has quit [Remote host closed the connection]
feep has joined #nixos
<feep>
yo
<feep>
I can use ping
<feep>
but when I do strace ping, it does not work
<yorick>
feep: ping is setuid
<yorick>
strace it as root
<feep>
oh okay
<feep>
yorick: it also does not work in nix-shell
fusion809 has quit [Remote host closed the connection]
<yorick>
feep: nix does not do setuid packages (that would be a security risk). nixos installs a setuid wrapper for ping
<feep>
it doesn't seem to actually be setuid
<yorick>
/run/wrappers/bin/ping
<feep>
yes
<yorick>
oh, then it's setcap
<feep>
(it's 0511)
<feep>
yorick: capset is indeed the call that fails in strace
<yorick>
feep: getcap /run/wrappers/bin/ping
<feep>
cap_setpcap,cap_net_raw+p
<feep>
getcap does not exist in the nix-shell
<feep>
is that a problem?
Mr_Keyser_Soze has quit [Remote host closed the connection]
<yorick>
obviously you can't do weird strace stuff to things that elevate privileges
<yorick>
feep: you want to use a ping from a nix shell?
<feep>
yes
<yorick>
well you can't
<gchristensen>
nix-shell -p ping won't get you a ping binary with capabilities, it has to be done using the one in /run/wrappers/bin/ping, or with sudo
Mr_Keyser_Soze has joined #nixos
<feep>
well I need to.
<feep>
I have an app that uses fhs paths that calls ping :p
<feep>
this is not exactly an uncommon usecase.
<gchristensen>
we'll have to find a different solution, then!
<feep>
gchristensen: to be clear, the ping in /run/wrappers does not work inside the nix-shell.
<feep>
oh, but with a different error, it says socket: operation not permitted
<feep>
all these files have root:root, but in nix-shell they show as nobody:nogroup
<{^_^}>
[nixpkgs] @yegortimoshenko pushed 0 commits to yegortimoshenko-patch-1: https://git.io/fpydv
<cyounkins>
I'm working on this package and am getting "Can't load '/root/nixpkgs/net-snmp-5.8/perl/blib/arch/auto/NetSNMP/default_store/default_store.so' for module NetSNMP::default_store: /root/nixpkgs/net-snmp-5.8/perl/blib/arch/auto/NetSNMP/default_store/default_store.so: undefined symbol: netsnmp_ds_get_boolean at /nix/store/cmajym0bi9jyxw0fx8y9w7kl256wa9gl-perl-5.28.0/lib/perl5/5.28.0/x86_64-linux-thread-multi/DynaLo
<cyounkins>
ader.pm line 193." I understand that the symbol is not available in the file, but how can I know if it's supposed to be in default_store.so or another library?
<feep>
gchristensen: when you run the nix-shell, what's the user/group of the /run/wrapper fiels?
<feep>
files*
<gchristensen>
nix-shell doesn't change anything about /run/wrapper
pxc has quit [Ping timeout: 260 seconds]
<feep>
ah
<feep>
does buildFHSUserEnv?
<gchristensen>
ah, yes, because it puts you in to a chroot
vidbina has joined #nixos
agander_ has joined #nixos
<feep>
can I make it let me use those files anyway somehow?
meizikyn has joined #nixos
agander has quit [Ping timeout: 250 seconds]
<gchristensen>
hmm an interesting question
Mr_Keyser_Soze has quit [Ping timeout: 250 seconds]
lsyoyom has joined #nixos
ninjin has quit [Remote host closed the connection]
<tilpner>
They are at /host/run/wrappers/bin
agander_ has quit [Client Quit]
<gchristensen>
oh!
<gchristensen>
tilpner++
<{^_^}>
tilpner's karma got increased to 11
<feep>
still nobody nogroup.
<feep>
because ... it just symlinks to /run/wrappers
ninjin has joined #nixos
<ottidmes>
any reason why _module.args.utils is not just part of lib? its a source of infinite recursion for me so I am going to see if I can just place those utils in lib
<tilpner>
ottidmes - That will not work unless you also fix every module expecting utils to be ther
<feep>
/host/run/wrappers/wrappers.*/ping google.de just does not work at all
<feep>
ah, because it asserts it's in the wrapper dir
<feep>
and they're still nobody/nogroup
<ottidmes>
tilpner: I am well aware, but there are not that many
<tilpner>
ottidmes - It's possible it was put there specifically to avoid inf-rec. lib might be overlay-able, utils probably isn't
<ottidmes>
tilpner: seems very unlikely to me, that same overlayed lib is used in the modules using utils, but am doing some tests
vidbina has quit [Ping timeout: 250 seconds]
<cyounkins>
OK I found the symbol in /root/nixpkgs/net-snmp-5.8/perl/blib/arch/auto/SNMP/SNMP.so using `nm`. So this package (net-snmp) is running a test by invoking a perl script which cannot find a dynamically linked library that it just built. Any pointers?
<ottidmes>
tilpner: also the one things utils depend on... is lib :P
<{^_^}>
[nixos-homepage] @sternmull opened pull request #254 → "the the" -> "the" → https://git.io/fpyFG
<feep>
aaa
Mr_Keyser_Soze has joined #nixos
fragamus has joined #nixos
ckauhaus has quit [Quit: WeeChat 2.2]
<aanderse>
i see references to renaming a deployment with nixops, but was wondering if there is any way to rename an actual machine managed by nixops?
klntsky has quit [Remote host closed the connection]
<aanderse>
i assume if i simply rename the variable nixops will think i've deleted one machine and added another
jrolfs has joined #nixos
slack1256 has joined #nixos
fragamus has quit [Ping timeout: 246 seconds]
Mr_Keyser_Soze has quit [Ping timeout: 246 seconds]
<{^_^}>
[nixos-homepage] @edolstra pushed 2 commits to master: https://git.io/fpyN5
halfbit has quit [Ping timeout: 260 seconds]
slack1256 has quit [Ping timeout: 250 seconds]
dbmikus__ has joined #nixos
dbmikus_ has quit [Ping timeout: 250 seconds]
vidbina has joined #nixos
<clever>
aanderse: the only solution i can see is to `nixops export`, rename it in the json, and `nixops import`, then it will think it always had that name
<clever>
aanderse: but ive not tested that, and dont know what side-effects may occur
c0bw3b_ has joined #nixos
<{^_^}>
[nixpkgs] @c0bw3b opened pull request #51673 → polkit: add patch for CVE-2018-19788 → https://git.io/fpyAq
pxc has joined #nixos
hedning has joined #nixos
<aanderse>
clever: thanks. still looking.
<aanderse>
ah... yeah just found it... should have read --help before the manual >_<
<{^_^}>
[nixpkgs] @c0bw3b opened pull request #51675 → [18.09] polkit: add patch for CVE-2018-19788 → https://git.io/fpyxl
<elgoosy>
after upgrading from 18.03 to 18.09, i keep finding packages that do not work (lib x was not found..) until i reinstall that package. How can force to relink all libraries used by my packages?
<aanderse>
yup, rename worked fine
<tilpner>
elgoosy - Are you using NixOS or just Nix?
<tilpner>
(I assume NixOS, because you mentioned version numbers 18.03 and 18.09)
<tilpner>
Are you using declarative package installation methods, or nix-env?
endformationage has quit [Ping timeout: 250 seconds]
<elgoosy>
tilpner, nixos, yes
<elgoosy>
i trie to configuration.nix for all my packages, but sometimes install them -iA
<{^_^}>
[nixpkgs] @ck3d opened pull request #51676 → vaapi-intel: added missing library libXfixes for x11 → https://git.io/fpyxj
<elgoosy>
so maybe those with -iA stayed stale right?
<tilpner>
Yes, that's my suspicion
<ottidmes>
elgoosy: yep, happened to me too when upgrading, you got to update or remove them
<elgoosy>
thanks
jtojnar has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @pmiddend opened pull request #51677 → pro-office-calculator: init at 1.0.6 → https://git.io/fpypz
<typetetris>
So, next year my employer will start to use nix (to a small amount).
<gchristensen>
woot!
<typetetris>
:)
Thra11 has joined #nixos
Thra11 has quit [Client Quit]
<ottidmes>
I have to do some more tests but it seems I finally solved my issue with _module.args and infinite recursion, the problem being pkgs was also defined via _module.args, so I could not use any _module.args in a config defining something that affected pkgs (i.e. an overlay), making _module.args pretty useless to me
<tilpner>
Yeah, you have to be very careful with module._args :/
Mr_Keyser_Soze has quit [Ping timeout: 272 seconds]
<typetetris>
Can get nix-shell to pass `--norc` to the invoked bash?
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
srl295 has joined #nixos
andromeda-galaxy has joined #nixos
<typetetris>
tilpner: thanks.
pxc has quit [Quit: WeeChat 2.3]
<typetetris>
tilpner: Somehow tab completion doesn't work in the shell starting then, do you know why that is so? (The shell claims to be interactive as "$-" contains an "i").
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
locallycompact has joined #nixos
<Synthetica>
I'll see if I can bisect the commit responsible for the difference in nix-diff
<locallycompact>
the documentation has loads of information for if you fetchFromGithub and have a shell script but doesn't really say what to do if you fetchFromGithub and the rpeo already has nix stuff in it
<locallycompact>
like I just want to say x = buildNixThingFromGithub
<locallycompact>
{ blahdeblah }
<tilpner>
locallycompact - The answer depends on whether you want to include the result in nixpkgs
<locallycompact>
why does that matter
<tilpner>
Because nixpkgs forbids IFD
<tilpner>
,ifd
fragamus has joined #nixos
<locallycompact>
what's that
<tilpner>
Oh well, it's dead again. IFD is import-from-derivation, so what you would do if you did "import (fetchFromGitHub { ... })"
<ottidmes>
tilpner: I am doing some crazy stuff right now with _module.args, so the patch seems to work fine, so I no longer have to be careful to prevent accidental infinite recursion :D
<infinisil>
,IFD
<{^_^}>
import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2
<tilpner>
infinisil - I tried that in PM, it didn't answer!
<tilpner>
But case insensitivity might be a good change
vidbina has quit [Ping timeout: 246 seconds]
<infinisil>
Oh darn, PM's don't work indeed
<tilpner>
They did a while ago, I think?
<infinisil>
Oh!
hedning has quit [Quit: hedning]
<infinisil>
tilpner: Everybody has a separate state in PM's
elgoosy has joined #nixos
<gchristensen>
:o
<tilpner>
Ahh, right
<infinisil>
Because otherwise anybody could change all definitions in PM's without anybody noticing
<infinisil>
I'll have to make it read only or so to fix that
<tilpner>
That explains why ,find/,locate were working, but not ,IDF
<gchristensen>
infinisil: can we append ". This is a problem because it requires evaluating some, building some, and then evaluating the build result."
cnidario has joined #nixos
<infinisil>
gchristensen: Sounds good to me
<gchristensen>
ok, how? :P
<infinisil>
,help
<{^_^}>
Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<tilpner>
,foo infinisil
<tilpner>
Lies!
<gchristensen>
,IFD = import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2
<{^_^}>
IFD redefined, was defined as: import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2
<tilpner>
,foo = Foo!
<{^_^}>
foo defined
<gchristensen>
,IFD = import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2. This is a problem because it requires evaluating some, building some, and then evaluating the build result.
<{^_^}>
IFD redefined, was defined as: import-from-derivation (IFD) is when you evaluate nix from a derivation result, for example `import (pkgs.writeText "n" "1 + 1")` will evaluate to 2
<gchristensen>
,ifd
<gchristensen>
,ifd = See: ,IFD
<infinisil>
Huh
hedning has joined #nixos
<infinisil>
,ifd = See: ,IFD
<{^_^}>
ifd redefined, was defined as: See: ,IFD
<infinisil>
I should fix that too lol
graphene has quit [Remote host closed the connection]
<infinisil>
These tiny outages are annoying
<disasm>
is there a way to detect if it's a dry run in nixops backends?
<Izorkin>
How to need to merge PR 51314 and 51612?
<disasm>
locallycompact: helps if you tell us what the problem is :)
<locallycompact>
I think I did
<locallycompact>
I want to just do something like
<locallycompact>
x = buildNixThingFromGithub { blahdeblah }
<locallycompact>
no builder.sh no make or shell just all the nix stuff is there already
<tilpner>
And you didn't tell us your paste fails
<tilpner>
*us how your
<locallycompact>
error: cannot coerce a function to a string, at /nix/store/ndx1c6zh8d8v101x81iy34xs5nff42gi-nixpkgs-19.03pre160749.3a393eecafb/nixpkgs/pkgs/stdenv/generic/make-derivation.nix:180:11
<locallycompact>
(use '--show-trace' to show detailed location information)
<disasm>
yeah, that's what I was getting at tilpner :)
<disasm>
locallycompact: so the next thing you want to do is --show-trace
<locallycompact>
error: while evaluating the attribute 'buildCommand' of the derivation 'docker-image-foundation.tar.gz' at /nix/store/ndx1c6zh8d8v101x81iy34xs5nff42gi-nixpkgs-19.03pre160749.3a393eecafb/nixpkgs/pkgs/stdenv/generic/make-derivation.nix:180:11:
<locallycompact>
while evaluating the attribute 'contents' of the derivation 'docker-layer-foundation' at /nix/store/ndx1c6zh8d8v101x81iy34xs5nff42gi-nixpkgs-19.03pre160749.3a393eecafb/nixpkgs/pkgs/stdenv/generic/make-derivation.nix:180:11:
<locallycompact>
cannot coerce a function to a string, at /nix/store/ndx1c6zh8d8v101x81iy34xs5nff42gi-nixpkgs-19.03pre160749.3a393eecafb/nixpkgs/pkgs/stdenv/generic/make-derivation.nix:180:11
<disasm>
and do not paste multiple lines in IRC :)
<tilpner>
Although I guess you could also try import (fetchFromGitHub { ... }) {}
<tilpner>
That would use their nixpkgs instead of yours
jrolfs has quit [Ping timeout: 246 seconds]
<tilpner>
Which might be a better version due to having whatever they're currently missing
<locallycompact>
oh yeah nice that's it
<locallycompact>
thanks!
jrolfs has joined #nixos
jrolfs has quit [Read error: Connection reset by peer]
alex`` has quit [Remote host closed the connection]
alex`` has joined #nixos
<spacekookie>
I'm trying to mount a vfat partition in the nixos net installer and mount is telling me "unknown filesystem". What obvious thing am I missing?
jrolfs has joined #nixos
cnidario has quit [Remote host closed the connection]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<tilpner>
Does /proc/filesystems contain vfat? (It does here)
<averell>
i'm adding an optional parameter to a derivation (glibc), and even when the result should be identical, everything gets rebuilt, because my flag is added to the build environment. how can i prevent that?
m0rphism has joined #nixos
<tilpner>
What parameter?
<gchristensen>
don't add it
<averell>
it's a new flag, i want to include a patch, but not rebuild everything dependent on glibc, so i added glibc_patched as a new package with the flag true, and the default being false
<tilpner>
Only add the flag when it's true?
<tilpner>
You can try system.replaceRuntimeDependencies if it's just for your local system. Won't help with preventing a mass-rebuild of nixpkgs of course
<{^_^}>
[nixpkgs] @adisbladis opened pull request #51682 → cni/cni-plugins: Remove go references → https://git.io/fpSt8
<cyounkins>
I have a derivation that is OK with "nix-shell --pure -I nixpkgs=/root/nixpkgs -A net_snmp", "genericBuild" but not with "nixos-rebuild -I nixpkgs=/root/nixpkgs build" - any suggestions on differences in the environment?
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
<eyjhb>
Is there any easy way to enable debugging for `udev` package?
locallycompact has quit [Ping timeout: 246 seconds]
orivej has quit [Ping timeout: 244 seconds]
<ottidmes>
develCuy: BTW I managed to reduce the errors I got from installing palemoon to just this one: Missing file(s): bin/browser/extensions/{972ce4c6-7e08-4474-a285-3208198ce6fd}/chrome.manifest
<sphalerite>
c0bw3b_: many shells also support Ctrl-L ;)
<develCuy>
hi ottidmes!
<develCuy>
great progress so far!
<ottidmes>
develCuy: adding ac_add_options --disable-devtools managed to remove the others
<develCuy>
ottidmes: is the file chrome.manifest present somewhere in the source or the build folder?
<ottidmes>
develCuy: I removed the build from disk in the mean time, but yes, I remember it was present
<cyounkins>
my package in systemPackages has binaries in the store but doesn't get symlinked in /run/current-system/sw/bin - Why would that be?
<ottidmes>
cyounkins: the package does not have $out/bin/<name> or it is not set executable?
<cyounkins>
ottidmes: Hmm. Well I have `-r-xr-xr-x 1 root root 17920 Jan 1 1970 /nix/store/n6r3bvwpsy82nddsmbfpqp91nhmb1iqm-net-snmp-5.8/bin/snmptest`
<develCuy>
yeah, but it is looking for the file somewhere else, it looks like an extension
<cyounkins>
I'm dumb, had 'build' instead of 'switch'...
<develCuy>
every extension has a chrome.manifest AFAIK
<develCuy>
I think that folder bin/browser/extension/{972ce4c6-7e08-4474-a285-3208198ce6fd} might be missing
<ottidmes>
develCuy: yeah, so I doubt any chrome.manifest from a different location would do, most likely totally unrelated files
<ottidmes>
develCuy: pretty sure it is, but what enable/disable flag do we need to add to either remove it from the list of expected files, or to make sure it exist
<develCuy>
ottidmes: look for this file: browser/extensions/{972ce4c6-7e08-4474-a285-3208198ce6fd}.xpi
<ottidmes>
cyounkins: it had to be something simple like that, such things happen, like not understanding why something is not working and finding out you have been editing a copy of the file
<ottidmes>
develCuy: I shall rebuild it then, because like I said, I have since purged it form my /tmp
rauno has quit [Ping timeout: 268 seconds]
<develCuy>
ottidmes: that extension is the default theme, I think we should remove the branding option
<bgamari>
domenkozar, have you run into limited tmpfs sizes when using runInLinuxVm?
<Ralith>
anyone using emacs lsp-mode with advice for how to wire it up to shell environmenets?
<ottidmes>
develCuy: nice find, then I will rebuild with that set, what do you suggest, either: ac_add_options --disable-official-branding or resetting ac_add_options --enable-application=palemoon back to ac_add_options --enable-application=browser?
<develCuy>
ottidmes: I think you can try each option in that order
m0rphism has quit [Quit: WeeChat 2.2]
<develCuy>
so first disable branding and compile
m0rphism has joined #nixos
<ottidmes>
develCuy: guess I have to remove: export MOZILLA_OFFICIAL=1, too then?
<ottidmes>
I could test first with it still there
<develCuy>
yes
<ottidmes>
its building, so in about 35min we will know...
Dezgeg has quit [Ping timeout: 264 seconds]
* develCuy
crossing fingers, pushing the couch
liff has quit [Ping timeout: 252 seconds]
ent has quit [Ping timeout: 250 seconds]
<ottidmes>
Ralith: not using Emacs, but maybe this is useful: https://direnv.net/
johnw_ has joined #nixos
<eyjhb>
Is there any reason to keep `systemd-udev-settle.service` enabled?
m0rphism has quit [Quit: WeeChat 2.2]
johnw has quit [Ping timeout: 252 seconds]
hakujin has joined #nixos
<ottidmes>
eyjhb: depends on your system, I read that you indeed what to prevent it from being pulled if possible, because its very slow, but some things might require it: https://freedesktop.org/wiki/Software/systemd/Optimizations/ (point 1)
cyounkins has quit [Remote host closed the connection]
<eyjhb>
Just currently hating the **** out of it, because it times out if my USB 3.0 Dock is inserted
<eyjhb>
And udev basically says F you, when you try to debug it at all....
endformationage has joined #nixos
<{^_^}>
[nix] @LnL7 opened pull request #2566 → s3: make scheme configurable → https://git.io/fpS3C
<ottidmes>
eyjhb: another answer "My advice is to mask it with sudo systemctl mask systemd-udev-settle.service and reboot. If your system comes up normally, you don't need it; if not, you can always unmask it." seems fair enough: https://ask.fedoraproject.org/en/question/81006/slow-boot-time-udev-settle/
<eyjhb>
Except, you cannot just mask it on nixos, right?
ninjin has quit [Remote host closed the connection]
ensyde has quit [Ping timeout: 272 seconds]
ninjin has joined #nixos
vk3wtf has joined #nixos
hedning has quit [Quit: hedning]
<gchristensen>
LnL: nice!
<eyjhb>
ottidmes: Might just disable it that way then ! - Also, is there a way I can check the connected device at this address ? `'/devices/pci0000:00/0000:00:14.0/usb4/4-3/4-3.1/4-3.1:1.0'`?
<LnL>
was still working on the docs
<eyjhb>
As it seems like that is the reason => `systemd-udevd[471]: seq 1504 '/devices/pci0000:00/0000:00:14.0/usb4/4-3/4-3.1/4-3.1:1.0' is taking a long time`
<LnL>
gchristensen: is your comment about the value or phrasing of the message?
<gchristensen>
LnL: the phrasing, I think just sticking to "https/http" should be sufficient? maybe caches will have sensitive information - its a judgement call for the user
<ottidmes>
eyjhb: udevadm info --query=all /devices/pci0000:00/0000:00:14.0/usb4/4-3/4-3.1/4-3.1:1.0
elgoosy has quit [Ping timeout: 250 seconds]
<eyjhb>
`Unknown device, --name=, --path=, or absolute path in /dev/ or /sys expected.`
<ottidmes>
eyjhb: probably /sys/devices
__monty__ has quit [Quit: leaving]
<ottidmes>
it is for me at least
<eyjhb>
Hmm.. It is the DisplayLInk vendor stuff, that times out because it does not respond... How would I even be able to change this?
jrolfs_ has quit [Quit: WeeChat 2.3]
<eyjhb>
I cannot even see, what service calls the displaylink stuff, or why it does, because udev doesn't want to give me any debug information in either syslog or journalctl -b
logzet has joined #nixos
<ottidmes>
eyjhb: try and search for a forum about the product, or maybe try stackexchange (has been mixed success for me, either you hit the jackpot and get a great answer, or you just get complaints and end up answering it yourself)
<eyjhb>
Somewhat niche and new, so nothing so far
<eyjhb>
And none-free.. So... Even the log files are encrypted
<ottidmes>
eyjhb: you can test rules with udev, did you look into that already?
<eyjhb>
Depends what you mean by testing?
<ottidmes>
eyjhb: see if they fire correctly
halfbit has joined #nixos
<eyjhb>
How would you go about that, during the boot process
graphene has quit [Remote host closed the connection]
<ottidmes>
eyjhb: not sure exactly, only needed to look into this stuff once myself, but if I remember correctly, you could use udevadm too to test your rules for a device, so you can test and see if your debug script works and such
irdr has quit [Ping timeout: 246 seconds]
graphene has joined #nixos
<ottidmes>
eyjhb: I have this as a shell function: sudo udevadm test $(udevadm info --query=path "$@")
<logzet>
I want to package an NodeJS application (only for myself for now, I may create a PR in the future). There seem to be a few packages on Nixpks, is this frowned upon, or should non-development packages not be added to pkgs/developmen/node-packages?
Ariakenom has joined #nixos
apaul1729 has quit [Ping timeout: 268 seconds]
<adisbladis>
logzet: It's a judgement call. Which package is it?
<ottidmes>
I just modified my NixOS implementation to allow overlays for lib, but before I clean it up and make a pull request, any reason why this would not be a good idea?
<logzet>
Even if there are, I think a PR would be a good idea for documentation purposes
hedning has joined #nixos
<ottidmes>
good point
<tilpner>
ottidmes - Did you account for misc/lib.nix?
<ottidmes>
ah, forgot the * in my find check just now ;)
c0bw3b_ has quit [Remote host closed the connection]
<tilpner>
Also, do you add a new overlay mechanism, or allow existing overlays to alter lib?
<tilpner>
(I would strongly oppose the former, but am already using the latter)
<ottidmes>
tilpner: I do exactly what nixpkgs overlays do
<tilpner>
That doesn't answer my question. I'll just wait for the PR
<ottidmes>
then your question is not clear to me, "do you add a new overlay mechanism", if taken literally, then yeah, but that is kinda the point, or if you mean, new compared to existing overlay mechanism, then that is why I mentioned nixpkgs overlays, but sure code answers things the best
endformationage has quit [Ping timeout: 250 seconds]
<tilpner>
Do you add nixpkgs.libOverlays, next to nixpkgs.overlays, or allow nixpkgs.overlays to also set lib?
<adisbladis>
logzet: I'm lazy so I'd be tempted to put the package in node-packages anyway
<adisbladis>
logzet: Look in to node2nix/yarn2nix too
cyounkins has quit [Remote host closed the connection]
<ottidmes>
tilpner: right now the former, but I am considering the latter, but not sure if its feasible
<ottidmes>
tilpner: my current prototype does not even have any relation to NixOS, its just pure standalone lib, using either <nixpkgs-lib-overlays> (not sure about that naming, it at least unambiguous) or ~/.config/nixpkgs/lib/overlays{,.nix}
<tilpner>
Oh no, that is terrible :c
<ottidmes>
tilpner: what is, that approach, or the naming?
<tilpner>
What happens if you do self: super: { lib = super.lib // { foo = 42; }; }, and then nix-instantiate '<nixpkgs>' --eval -A lib.foo ?
<eyjhb>
ottidmes: But I have NO idea, what I should supplythe function with
<tilpner>
Put that in your ordinary nixpkgs-overlays
hedning has quit [Quit: hedning]
timon37 has joined #nixos
<ottidmes>
tilpner: your example was my first test case, and work just fine
<tilpner>
So then why do we need nixpkgs-lib-overlays?
<ottidmes>
tilpner: but I agree, if I can get it to work to leverage pkgs.lib overlays, than there is no need for anything seperate, hangs on that though
<ottidmes>
tilpner: ^ that
<yegortimoshenko>
jtojnar: does gnome application list refresh for you after nixos-rebuild?
<tilpner>
ottidmes - Please specify the challenges of "if I can get it to work to leverage pkgs.lib overlays"
<tilpner>
Oh, inf-rec, of course
<ottidmes>
right
ntd has quit [Quit: ntd]
dbmikus__ has quit [Quit: WeeChat 2.3]
erasmas has quit [Quit: leaving]
<Ralith>
ottidmes: hmm, doesn't seem to reach lsp-mode, somehow