<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/JJUct
<jared-w>
or is that actually belonging to AmandaC and I am a dumb?
<jared-w>
ah ok no, I can still read. Phew.
<porphyrogenetos>
what exactly makes nix decide to build a package from source? I was messing around with adding an optional flag to emacs (for libvterm support) and when i "nix-env -iA nixos.userPackages" i thought it would pull all the binaries from nixpkgs except for emacs and its dependencies. Instead what happened is it looks like its building everything in userPackages from source
<AmandaC>
porphyrogenetos: it depends what there's cached in the public cache
nate48423 has quit [Remote host closed the connection]
nate48423 has joined #nixos
<porphyrogenetos>
I see, so far its only rebuilding racket, but it alarmed me since its always pulled the binaries for racket. Could just be that racket was removed from the cache before i started messing with emacs.
<simpson>
As a general rule, the binary cache doesn't have anything removed from it. Instead, likely something about your configuration non-obviously changes the system somehow.
<simpson>
Maybe the better question and answer would be: When does Nix decide to *not* build a package from source? The binary cache is only populated with packages from a subset of nixpkgs, and only some particular configuration options.
<gchristensen>
hazel[m]: I'm sorry for your loss
<hazel[m]>
whuh
cyraxjoe has quit [Quit: I'm out!]
<gchristensen>
hazel[m]: the hair you lost to antiquated and user-hostile encryption software
<gchristensen>
(not saying there is anything better ... I just wish there were.)
<li_matrix>
<gchristensen "Li: eh?"> we dont know, and the availability heuristics pushes people to believe bigger chips will come that break the crypto later
<gchristensen>
sorry, I just found that sentence interesting from your link
<li_matrix>
Id bet my life nobody will ever break an ed25519 key
Rusty1 has joined #nixos
<li_matrix>
actually, we know that the master 'encryption' key will not be used for the actual payload encryption, thats what the ratchet is for
justanotheruser has quit [Ping timeout: 244 seconds]
<li_matrix>
the signing key is by far the most important, it could just sign a temporary encryption pubkey
cyraxjoe has joined #nixos
* gchristensen
is out of his league
<{^_^}>
[nixpkgs] @fkautz opened pull request #91978 → purple-matrix: set c compiler to cc, fixes build on osx → https://git.io/JJUCd
<simpson>
li_matrix: I would *not* bet a life, FWIW, on the odds that there will be improvements in the index-calculus techniques that have been used against elliptic-curve constructions in the past.
<simpson>
Your point is a good one: Any particular key is not likely to be broken. But that doesn't mean that we should bet everything on just one key, if we can avoid it.
dbmikus has quit [Ping timeout: 240 seconds]
<li_matrix>
<simpson "Your point is a good one: Any pa"> except that people bet their lives every time they get on the road, they trust in the engineering.
<gchristensen>
(strawman?)
<simpson>
Yeah, cars aren't at risk of suddenly having internal combustion engines proven to not exist tomorrow.
<li_matrix>
if we understood our predicament, it would drive us mad with terror
<gchristensen>
I think I do understand the predicament and it is why I don't use gpg
<li_matrix>
youd need a counterparty to use it with, lol
<gchristensen>
I'm not sure what you're trying to say
abrar has joined #nixos
<li_matrix>
but ill import my ed25519 key into the next framework, be it gpg or ssh or whatever
<li_matrix>
<gchristensen "I'm not sure what you're trying "> just joking, gpg use is kind of rare now
<gchristensen>
ah, okay :). people send me gpg mail :(
<li_matrix>
if you understood your predicament youd be driven mad.
<gchristensen>
I mean... I feel like I have a pretty good handle on it ...
<gchristensen>
but do say more :)
<li_matrix>
i daily have to deal with this, the madness coming over
<li_matrix>
know a lot is also a curse
<gchristensen>
yup
* simpson
refrains from posting "The Call of Cthulhu" introduction
<li_matrix>
they make is sound so romantic to be a genius: in actuality it just feels like youre surrounded by a bunch of retarded psychopaths
<gchristensen>
li_matrix: that sort of language isn't appropriate
<li_matrix>
thats why they call those 'dark poets'
<li_matrix>
- Nietzsche
<li_matrix>
To me, the masses seem to be worth a glance only in three respects: first as blurred copies of great men, presented on bad paper with worn out printing plates, then as the resistance against the great men, and finally as working implements of the great. For the rest, let the devil and statistics carry them off!
<gchristensen>
this is almost certainly off topic lol
felixfoertsch has quit [Ping timeout: 246 seconds]
felixfoertsch has joined #nixos
<craige>
I suspect this is a FAQ but how do you reproduce Hydra evaluation errors locally? Everything builds fine locally but Hydra still reports evaluation issues. Thanks!
<li_matrix>
<gustavderdrache "https://gist.github.com/gustavde"> `pkgs'` needs a `system` ... `crossSystem` is optional, but `system` isnt
<gustavderdrache>
i guess i should elaborate: i'm looking to cross-compile some code for mingw, but for some reason both attempts at getting a w32api derivation out of nixpkgs fails to build
zebrag has joined #nixos
<infinisil>
gustavderdrache: pkgsCross.mingwW64.<attribute> is the now-recommended way to do cross
proofofkeags has quit [Remote host closed the connection]
slack1256 has quit [Remote host closed the connection]
pjt_tmp has joined #nixos
<li_matrix>
why is calling `nix make-content-addressable` triggering the inner sandbox to `mkdir /nix/var` ?
<li_matrix>
<li_matrix "why is calling `nix make-content"> using a recursive-nix derivation that is
orivej has quit [Ping timeout: 246 seconds]
proofofkeags has joined #nixos
proofofkeags has quit [Remote host closed the connection]
Yaniel has quit [Ping timeout: 246 seconds]
proofofkeags has joined #nixos
proofofkeags has quit [Remote host closed the connection]
waleee-cl has quit [Quit: Connection closed for inactivity]
Yaniel has joined #nixos
<li_matrix>
<li_matrix "using a recursive-nix derivation"> I passed `"--extra-sandbox-paths" "/nix/var/nix/"` to it and now I think its trying to write to the daemon socket, and fails
gustavderdrache has quit [Quit: Leaving.]
proofofkeags has joined #nixos
zanc has quit [Ping timeout: 246 seconds]
zanc has joined #nixos
proofofkeags has quit [Ping timeout: 264 seconds]
jmarhee has quit [Ping timeout: 264 seconds]
pr06lefs has quit [Ping timeout: 260 seconds]
ddellacosta has quit [Ping timeout: 246 seconds]
neonfuz2 has joined #nixos
<neonfuz2>
Hello
<neonfuz2>
I'm trying to enable trackball scrolling on my logitech marble mouse, but it isn't working
<neonfuz2>
I'm using services.xserver.inputClassSections
<neonfuz2>
I don't see a generated xorg.conf file in /etc/X11, I'm not sure where places it
endformationage has quit [Quit: WeeChat 2.7.1]
<{^_^}>
[nixpkgs] @veprbl opened pull request #91984 → zstd: build programs and tests against the shared library → https://git.io/JJUR8
nuncanada has quit [Read error: Connection reset by peer]
<srhb>
neonfuz2: Not sure how to do that, but if you set services.xserver.exportConfiguration, the xorg.conf will be conveniently symlinked into /etc/X11/xorg.conf
<srhb>
neonfuz2: Otherwise it's a bit hard to get at (it's just some path in the store)
<neonfuz2>
oh thanks that will help
<neonfuz2>
yeah in general I find it hard to find some internal things like that with nixos
<srhb>
neonfuz2: For many things (but not xorg) you can usually `systemctl cat someservice` and find the configuration file in the ExecStart directive or similar.
<neonfuz2>
I'm not sure how to dig through the current system and find paths to the packages really
<neonfuz2>
ah
<srhb>
neonfuz2: It's all pointed at through /run/current-system
<srhb>
neonfuz2: But _where_ exactly takes some practice to dig at :)
<neonfuz2>
yea
<neonfuz2>
harder than regular linux unfortunately
<srhb>
neonfuz2: Yes, there's a cost.
<neonfuz2>
I guess there's benefits to the layout too though
<srhb>
For sure :)
<neonfuz2>
there could probably be better tools for exploring
<{^_^}>
[nixpkgs] @DamienCassou pushed 2 commits to master: https://git.io/JJUuq
simba2 has quit [Ping timeout: 272 seconds]
hellrazo1 has joined #nixos
simonthesourcere has joined #nixos
<simonthesourcere>
I've got nixos installed on one M2 drive and windows on another. I couldn't get the auto detection in Nixos to detect my windows drive so currently I have to physically pull out the Linux drive to boot windows as for some reason changing the UEFI settings won't help. So how can I configure grub to add my windows drive?
<simonthesourcere>
The M2 connection on my motherboard seems to be loosening up for every time I do it, which is every day after work when I need windows to play computer games :P
user_0x58 has quit [Quit: Leaving]
<srhb>
simonthesourcere: Are you uefi booting?
cole-h has quit [Quit: Goodbye]
<simonthesourcere>
yes
<simonthesourcere>
and I've used the autoprober which doesn't seem to detect windows on my 2nd drive
marusich has quit [Ping timeout: 260 seconds]
<srhb>
simonthesourcere: Maybe you can set it up manually using extraEntries?
<srhb>
simonthesourcere: Experimenting from the grub command line should be possible.
<simonthesourcere>
You mean I reboot into grub and try to figure out the settings from there?
<srhb>
simonthesourcere: Yes, looking up a guide for chainloading windows 10 uefi first :)
FRidh has joined #nixos
<simonthesourcere>
than you that's a good idea, I'll try that out
<srhb>
simonthesourcere: It will involve identifying the efisys partition in which the windows 10 bootloader lives in grub syntax.
<clever>
those appear to be the most important parts
<simonthesourcere>
one more question before I head off to my morning meeting, what irc clients do people use and is there one I can configure with nix expressions?
<clever>
i just use irssi, i still havent moved that client to a nixos machine, lol
<simonthesourcere>
I haven't used irc in ages but using this web client I realize that's a mistake
<clever>
its on an old gentoo box that i have been neglecting for years
<simonthesourcere>
I see, maybe just irssi + home-manager should do the trick
<clever>
i think irssi may want to write to its own config file if you use /set and /save
<clever>
so it may get upset about home-manager making things read-only
<{^_^}>
[nixpkgs] @bjornfor merged pull request #91720 → nixos/doc/manual: Fix parted's set subcommand for esp partition → https://git.io/JJvqs
<srhb>
I'm using weechat. It's OK out of the box, except it's a horrible box of mutating spaghetti.
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @konrad to master « nixos/doc/manual: Fix parted's set subcommand for esp partition »: https://git.io/JJUzn
<simonthesourcere>
I got to go to my standup meeting, thanks for the help tho. Much abliged
<simonthesourcere>
or sitdown meeting now when we all wfh
<clever>
and i'm off to bed
<srhb>
Sleep well clever :)
<simonthesourcere>
sleep well
<clever>
-rw-r--r-- 1 root root 4.9K Aug 9 2019 /boot/grub/i386-pc/keylayouts.mod
<clever>
the required module is already installed...
MidHotaru has joined #nixos
marusich has quit [Quit: Leaving]
<clever>
ls: cannot access '/boot/grub/layout/': No such file or directory
<clever>
and the dir is missing, so nixos isnt creating it either
<clever>
install-grub.sh will likely need patching
<jneplokh>
So after some troubleshooting, I realized port 443 is not open. However, I have `networking.firewall.allowedTCPPorts = [ 80 443 ];` Port 80 is open, but for some reason 443 is not? Why could this be?
<srhb>
jneplokh: Is anything listening on that port?
lsix has quit [Client Quit]
<jneplokh>
Nextcloud is supposed to be, but when I type in the url with https it doesn't work :/
<srhb>
jneplokh: Check: "sudo ss -tlnp | grep 443"
<jneplokh>
Huh that's strange, it doesn't show anything
<srhb>
jneplokh: So it's not listening on that port at all
o1lo01ol1o has quit [Remote host closed the connection]
o1lo01ol1o has joined #nixos
<ocharles>
Does anyone know why my self-hosted Hydra instance doesn't fetch from cache.nixos.org? I currently have: `Below is the build log of derivation /nix/store/p6vd0y4rqah2d202sl91bccs1bsylbs5-perl-5.30.1.drv`, yet if I ssh into the very same machine and run `nix-store -r /nix/store/p6vd0y4rqah2d202sl91bccs1bsylbs5-perl-5.30.1.drv` it downloads from cache.nixos.org.
<ocharles>
So the machine is certainly set up to use it, but Hydra seems to be doing something special
cosimone has joined #nixos
<ikwildrpepper>
ocharles: you can use: services.hydra-dev.useSubstitutes
<ikwildrpepper>
or services.hydra.useSubstitutes if you use hydra from nixpkgs
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<ocharles>
Cool! I'll check that out. Why would that not be on by default?
<ocharles>
Ah ok, the description explains why
MidHotaru has quit [Quit: Connection closed for inactivity]
o1lo01ol1o has joined #nixos
<{^_^}>
[nixpkgs] @ocharles opened pull request #91999 → haskellPackages.ghcWithPackages: fix for GHC 8.10 → https://git.io/JJUwh
o1lo01ol1o has quit [Ping timeout: 260 seconds]
sangoma has quit [Read error: Connection reset by peer]
phreedom has joined #nixos
<{^_^}>
[nixpkgs] @ckauhaus pushed 0 commits to security-updates: https://git.io/JJUrw
<aloiscochard[m]>
greetings, I'm running into issue when calling `sbt` (scala build tool) from python (it always use java8) as opposed to running it directly from my shell where it take the proper (JAVA_HOME specified) version of java, here is a complete log showing the issue: https://pastebin.com/Zfzyh4H7 ... does anybody would have an idea of what is going on here?
<aloiscochard[m]>
I can't reproduce same issue on other OS (ie. MacOS)
orivej has quit [Ping timeout: 258 seconds]
<fps>
hmm, is there a quick way to find out what services are enabled by default in a system?
<talyz>
fps: you can run systemctl without any arguments, which will list all active units, and scroll down to the .service-file section
mallox has joined #nixos
smatting has joined #nixos
<fps>
talyz: yeah, that's an indirect way of answering the question :) but better than nothing ;)
sangoma has quit [Ping timeout: 258 seconds]
zupo has joined #nixos
sangoma has joined #nixos
<fps>
hmm, ok so what does this unit correspond to remote-fs.target
<fps>
i searched man configuration.nix for "remoteF"
orivej has quit [Quit: No Ping reply in 180 seconds.]
m0rphism has joined #nixos
<hyperfekt>
can someone confirm that minecraft isn't building on current unstable and master because curl is not found in the source derivation? i'm not believing this error...
orivej has joined #nixos
<srhb>
hyperfekt: Builds for me.
eoli3n has quit [Remote host closed the connection]
<fps>
hmm, how precisely does the --max-freed option of nix-collect-garbage work?
<fps>
is it the maximum free diskspace after running it or the maximum space it is allowed to free?
Izorkin has joined #nixos
<dutchie>
"keep deleting bytes until at least bytes bytes have been deleted" from man nix-store
<fps>
oh ok, if i wanted to free up a certain amount of space on the disk (to leave me room to "breathe") i would have to do the calculation myself. i guess that can be done with df, cut and bc ;)
orivej_ has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos
ambro718 has joined #nixos
meh` has joined #nixos
felixfoertsch has quit [Ping timeout: 260 seconds]
o1lo01ol1o has joined #nixos
zanc has quit [Ping timeout: 240 seconds]
orivej has quit [Ping timeout: 265 seconds]
orivej has joined #nixos
meh` has quit [Read error: Connection reset by peer]
zanc has joined #nixos
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<LambdaDuck>
Which nix tools accept nix expressions on the command line? I know nix-shell does with -p, but can you give an expression to nix-build in some way?
<srk>
yes, via -E
zanc has quit [Read error: Connection reset by peer]
<ma27[m]>
sth like `nix-build -E 'with import <nixpkgs> {}; /* type your expr here */'` ?
<ma27[m]>
`nix-instantiate -E` exists as well if you need that.
<ma27[m]>
btw if you need an interactive session for building you can also use the `:b` command in `nix repl`.
<LambdaDuck>
ma27[m]: Ah, right. Thanks! I was just going to ask why it didn't find nixpkgs.
sgrunert has joined #nixos
<LambdaDuck>
What I really want is a way to convert "nix-shell -p someExpression -p somethingElse" to a list of store paths for there packages. I think the nix-build -E 'with ...' should work.
<LambdaDuck>
Yes. Looks good. Thanks!
teto has joined #nixos
zanc has quit [Ping timeout: 256 seconds]
alp has quit [Ping timeout: 272 seconds]
zanc has joined #nixos
sgrunert has quit [Remote host closed the connection]
<chiiba>
I've configured `gnupg.agent w/ enable & pinentryFlavor = gtk2`, but gpg always errors our w/ `No pinentry`. I can work around it using `pkill gpg-agent && gpg-agent --daemon --pinentry-program "$(readlink $(which pinentry))"`, but how would I go about figuring out a legitimate fix?
lawr3nce has joined #nixos
pr06lefs has joined #nixos
smatting has joined #nixos
<{^_^}>
[nixpkgs] @offlinehacker opened pull request #92009 → whirlpool-gui: init at 0.10.1 → https://git.io/JJU9l
ckauhaus has quit [Quit: WeeChat 2.7.1]
knupfer has joined #nixos
knupfer1 has joined #nixos
lawr3nce has quit [Ping timeout: 240 seconds]
sangoma has quit [Ping timeout: 264 seconds]
zaeph has quit [Ping timeout: 240 seconds]
<gchristensen>
anyone have an example of an overlay overriding pythonPackages?
<immae>
gchristensen: I do have some workaround to implement that yes
<gchristensen>
may I see it? :) (also, needing workarounds is a bit unfortunate)
<FRidh>
noticed now with flakes that this is especially needed
lawr3nce has joined #nixos
zaeph has joined #nixos
<gchristensen>
FRidh: all of them
<srk>
does it allow overriding buildPythonPackage as well?
<FRidh>
srk: yes and no. For that, it would be better if we get rid of python.pkgs, and instead do the haskell way where you have a package set, and the interpretr is part of the set.
orivej has quit [Quit: No Ping reply in 180 seconds.]
<srk>
I see
felixfoertsch has quit [Quit: ZNC 1.7.5 - https://znc.in]
<srk>
we've had some issues with that before, overlays touching that were seemingly ignored
<FRidh>
yes, overriding python is tricky, because you need to have something like python = let self = python.override { ... , inherit self }; in self;
orivej has joined #nixos
<FRidh>
if you forget the `inherit self;` then `python.withPackages` wont work
<srk>
noted.
alp has joined #nixos
sangoma has joined #nixos
cosimone has joined #nixos
<{^_^}>
[nixos-search] @garbas pushed to complete-option-name « replimplemented suggestion field »: https://git.io/JJUHL
<FRidh>
Making the changes that are needed / would be good will most definitely be backwards incompatible which is a problem
felixfoertsch has joined #nixos
<FRidh>
which is why I am so hesitant to make these changes
lawr3nce has quit [Remote host closed the connection]
<gchristensen>
FRidh: a company I know of would like to create and maintain a set of packages unique to python2 in nixpkgs to restore a packages when they're upgraded beyond python 2 support. I think I've mentioned this before but I don't think I got much of an answer on what you thought. what do you think? :)
<FRidh>
gchristensen: I recall you and adisbladis once mentioned the python 2 packages set / overrides
<gchristensen>
I think their intention would be to maintain them separately so as to not clutter up the py3 stuff with py2 gunk -- and when the day comes to excise py2 it'd be just `rm -rf` on a whole directory
<FRidh>
the python 2 set has in these last month or so diverged significantly from the python 3 set. Many python 2 packages still work, but not as many as 3.
<gchristensen>
yeah
<FRidh>
Right. Yes, probably we should do that now. python-packages.nix is full of `if isPyk3 then ... else ..;`
<gchristensen>
yeah
<gchristensen>
so you're on board? :)
<FRidh>
and I typically create separate expressions anyway
<gchristensen>
yeah
<FRidh>
yes, I am. Just need someone willing to make the changes ;)
<gchristensen>
they'll do it :P maybe you could make the change to 1 or 2 packages to set a template to start from? this way it would certainly match what you'd expect, and from there they can just churn it out
Neo-- has joined #nixos
<gchristensen>
or you could just say what you're thinking and that would be fine too
<FRidh>
Right. Yes, probably the initial overlay and a couple of core packages
<FRidh>
but are they willing to go through the nixpkgs queue of PR's?
<gchristensen>
like, submitting PRs? for sure
<gchristensen>
they don't have push access and I don't feel right "lending" my push access
<gchristensen>
it is something they want to do and upstream, as long as you/we want them to. and if you/we want them to, they want to do it the most normal and proper way possible
<{^_^}>
[nixos-search] @garbas merged pull request #74 → Add option name completion field → https://git.io/JfyiI
<{^_^}>
[nixos-search] @garbas pushed commit from @adisbladis to master « Suggestions for search field (#74) »: https://git.io/JJUHQ
<{^_^}>
[nixos-search] @garbas pushed 0 commits to complete-option-name: https://git.io/JJUH7
m4ts has quit [Quit: bye]
m4ts has joined #nixos
<gchristensen>
FRidh: how do you want to do it? :)
<FRidh>
gchristensen: well, I am fine with having it in nixpkgs. But indeed, it then just ends up in the normal PR queue, if that's fine for them, good. This is really something that would be good to have in a flake eventually.
<gchristensen>
cool
<gchristensen>
do you want for them to just go ahead and start, or would you like to do the first package?
<FRidh>
need to create an overlay first, so let's wait with that for now
<turion>
jakobrs: My first step would be adding the nixos-hardware overlay with your specific machine or graphics card, and see whether that helps
<turion>
My current setup to debug pythonPackages.cassandra-driver is 1. nix-shell -A python37Packages.cassandra-driver 2. genericBuild 3. setuptoolsCheckPhase
<turion>
But 3. always says "Ran 0 tests in 0.000s", so something is wrong here.
<turion>
What phase do I really need to run?
<turion>
Or have I maybe missed some phase somehow?
<turion>
$installCheckPhase and installCheckPhase don't do anything at all
felixfoertsch has quit [Ping timeout: 260 seconds]
<jakobrs>
There's no subfolder for my machine specifically, but I'll try enabling whatever looks relevant and see if it works.
<turion>
It seems like in python, genericBuild already removes the test files, so it runs a few steps to many. Where can I look up which phases I really need?
<{^_^}>
[nix] @edolstra merged pull request #3777 → Run the tests in parallel → https://git.io/JJUNf
<{^_^}>
[nixpkgs] @dominikh opened pull request #92052 → x265: build a single shared library for all bit-depths → https://git.io/JJUAt
<sheeldotme>
I'm unable to get nixos to boot from a mdraid setup. Does anyone know if there's any additional configuration required past `nixos-generate-config` to make the installation bootable?
<turion>
I have some raid setup that works, but I don't remember whether I needed to tweak hardware-configuration.nix. Did you mount your complete raid before you called nixos-generate-config?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sangoma has joined #nixos
cosimone has quit [Quit: Quit.]
<sheeldotme>
turion I did, but I don't see anything raid specific in the hardware-condition. I'm wondering if the issue is a result of the raid 0 configuration. (the boot partition being striped).
<sheeldotme>
Maybe if I mirror that things will be a little easier.
virus_dave has joined #nixos
fendor_ has joined #nixos
fendor has quit [Ping timeout: 256 seconds]
turion has quit [Ping timeout: 258 seconds]
<eacameron>
Is there a nix or nixpkgs function to take a string and replace invalid characters such that it becomes a valid name for a derivation?
proofofkeags has quit [Remote host closed the connection]
proofofkeags has joined #nixos
fendor__ has joined #nixos
nuncanada2 has joined #nixos
Czen1 has joined #nixos
zupo has joined #nixos
fendor__ has quit [Remote host closed the connection]
orivej has quit [Quit: No Ping reply in 180 seconds.]
<{^_^}>
[nixpkgs] @FRidh pushed commit from @fomichevmi to master « tremor: fix src link »: https://git.io/JJUpl
<turion>
sheeldotme: Still it would be strange if you RAID 0 a boot partition... not sure a boot loader can handle that. Maybe the config I posted is what you need after all
veleiro has quit [Read error: Connection reset by peer]
erasmas has joined #nixos
<sheeldotme>
turion I don't, I'm just in the process of reconfiguring the raid setup. If I mirror the boot partition would I still have to set both as grub devices or would I add /dev/md0?
lordcirth has quit [Ping timeout: 260 seconds]
<sheeldotme>
Maybe grub wouldn't know how to handle the raid so maybe I shouldn't mirror that partition and just have grub installed on both?
sangoma has quit [Read error: Connection reset by peer]
gustavderdrache has joined #nixos
<cransom>
my /boot, isn't mirrored per se, but grub is set to write to both of them so they are effectively in sync
<{^_^}>
[nixos-artwork] @samuelgrf opened pull request #54 → Add dark gray variant of "nineish" wallpaper → https://git.io/JJTv9
nate48423 has quit [Ping timeout: 256 seconds]
cole-h has joined #nixos
<tmplt>
I just updated my local release-20.03-small channel and can no longer rebuild: `error: Invalid package attribute path `nextcloud19'`. I dont use nextcloud.
MidHotaru has quit [Quit: Connection closed for inactivity]
nwm_ has quit [Read error: Connection reset by peer]
nate4842 has joined #nixos
nate48423 has quit [Ping timeout: 244 seconds]
nwm_ has joined #nixos
nate4842 has quit [Remote host closed the connection]
nate48423 has joined #nixos
litschio is now known as litschi
orivej has quit [Quit: No Ping reply in 210 seconds.]
fendor_ has quit [Read error: Connection reset by peer]
orivej has joined #nixos
sheeldot_ has joined #nixos
fendor has joined #nixos
<KarlJoad>
I have 2 GPUs in my desktop, with 1 monitor plugged into each. I can only get output on the main monitor/GPU. Both are nVidia. Is there something I need to do to make this work?
quinn has joined #nixos
sheepfleece has quit [Quit: Lost terminal]
lordcirth has quit [Read error: Connection reset by peer]
sheeldotme has quit [Ping timeout: 264 seconds]
<slabity>
KarlJoad: Try using xrandr's `--setprovideroutputsource` option
zeromorphis_ has joined #nixos
<slabity>
For me, I use `xrandr --setprovideroutputsource 1 0` to get my second GPU working with a monitor
<slabity>
Might need to play with the options
<slabity>
Although I use 2 AMD gpus, so I'm not sure if Nvidia will work the same way :\
zeromorphism has quit [Ping timeout: 256 seconds]
ilya-fedin has joined #nixos
lordcirth has joined #nixos
zupo has joined #nixos
<KarlJoad>
slabity: When you use just `xrandr`, do you have 2 screens or just 1?
<slabity>
KarlJoad: When I first login to my session, `xrandr` only shows the displays from one GPU at first. After I perform the above command, `xrandr` displays both GPU's displays and I can freely configure it rom there
orivej has quit [Ping timeout: 256 seconds]
<slabity>
Also, apprently there can also be some strange behavior with multiple GPUs if you enable CSM in your BIOS. Not sure why, but my motherboard would not let me use both GPUs at the same time when I had CSM enabled. Had to switch to pure UEFI
<{^_^}>
[nixpkgs] @grahamc merged pull request #91344 → ZFS: Ask for stage 2 encryption passwords using systemd-ask-password → https://git.io/JfxrN
<notgne2>
zanc: you can use recursion for that, so maybe a bit like `let xxx = i: max: acc: if i < max then xxx (i + 1) max (acc ++ [ i ]) else acc; in xxx 0 10 []`
<{^_^}>
Found in packages: lsof, busybox, bashCompletion
orivej has quit [Ping timeout: 240 seconds]
<notgne2>
well that's significantly shorter than my solution
<{^_^}>
[nixpkgs] @vcunat opened pull request #92086 → cc-wrapper: ensure existence of some $out/nix-support/* → https://git.io/JJTIt
orivej has joined #nixos
pinecamp has quit [Remote host closed the connection]
domogled has joined #nixos
alp has joined #nixos
<zanc>
> :p builtins.genList (x: x + 1) 10
<{^_^}>
[ 1 2 3 4 5 6 7 8 9 10 ]
meh` has joined #nixos
<{^_^}>
[nixos-artwork] @samueldr merged pull request #54 → Add dark gray variant of "nineish" wallpaper → https://git.io/JJTv9
<{^_^}>
[nixos-artwork] @samueldr pushed 2 commits to master: https://git.io/JJTIW
<maier>
I tried out the proprietary 'nvidia' drivers for my xsession, but now on startup or on ctrl+alt+F1 the font in the terminal looks very blurry. is this normal?
camsbury has joined #nixos
<sphalerite>
maier: yes. nvidia doesn't like implementing the same API as everyone else, but at the same time didn't want to write their own vtconsole.
<sphalerite>
maier: so the nvidia graphics driver only supports a low-resolution console, which will usually be scaled up to fit your monitor.
<sphalerite>
nvidia's drivers are kind of awful.
<{^_^}>
[nixpkgs] @FRidh pushed to master « python.pkgs.zstandard: use libzstd instead of python zstd »: https://git.io/JJTIS
<{^_^}>
[nixpkgs] @FRidh pushed commit from @danieldk to master « python3Packages.tokenizers: init at 0.8.0 »: https://git.io/JJTLV
alp has quit [Ping timeout: 256 seconds]
pinecamp has joined #nixos
pinecamp has quit [Remote host closed the connection]
hellrazo1 has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
<{^_^}>
[nixpkgs] @vcunat closed pull request #92086 → cc-wrapper: ensure existence of some $out/nix-support/* → https://git.io/JJTIt
<maier>
sphalerite++
<{^_^}>
sphalerite's karma got increased to 98
<maier>
thanks, I'll just go back to nouveau or whatever's the default then :)
orivej has quit [Ping timeout: 246 seconds]
<sphalerite>
maier: that won't be fun for anything requiring 3D performance though, because nouveau (unless there have been some major unexpected changes since I last checked) doesn't support power management, i.e. the card will always run at the lowest clock rate.
<sphalerite>
maier: there's just no winning with nvidia. :p
<slabity>
Nouvuea's not much better... Nvidia doesn't like to provide documentation, which makes performance really poor
lsix has quit [Ping timeout: 260 seconds]
orivej has joined #nixos
<sphalerite>
Both of the problems are nvidia-the-company's fault in the end though >_<
<slabity>
If you want performance or proper power management you will need Nvidia's proprietary drivers. You'd just need to live with the issues caused
<slabity>
notgne2: I believe you can use GNOME's wayland implemention on Nvidia. They provide egl-streams support
<slabity>
But since everyone else only provides GBM support, then yea... Not many wayland options
<notgne2>
I heard kwin was getting it too, but I'm a sway/i3 user, and sircmpwn seems to have taken a much stronger stance against nvidia
<slabity>
Can't blame em. GBM was perfectly fine and I don't see any benefit to egl-streams
eoli3n__ has quit [Ping timeout: 240 seconds]
Dadou has joined #nixos
<Dadou>
Hello, is it possible to, in a nix-shell, to get a link on a file from nixpgs ?
<notgne2>
I think there was some proposed vulkan extensions that would make wlroots able to function on a vulkan backend (and therefore in theory gaining nvidia functionality), but that's going to take long enough to standardize let alone get nvidia to implement
<notgne2>
though the situation is way beyond what I have the understanding of, I'm just annoyed I can only use sway on my laptop (though NixOS makes it very easy to keep my configs the same except for a few differences which I've made a wayland toggle for)
<slabity>
Dadou: What do you mean by 'get a link on a file from nixpkgs?'
<slabity>
As in find the location of a binary?
<Dadou>
More like callpackage ${nixpgs/my/expression.nix}
<{^_^}>
[nixpkgs] @Pablo1107 opened pull request #92093 → insomnia-designer: init at 2020.2.2 → https://git.io/JJTtR
<tpw_rules>
i'm trying to make a derivation for a package that vendors like 20 other libraries for no apparent reason, and gives the option to not vendor them. is it appropriate to try and reduce the vendoriness as much as possible, or should i just keep the vendored versions?
KarlJoad has quit [Remote host closed the connection]
<slabity>
tpw_rules: Could you have the libraries act as inputs to your derivation? That was it can be decided at build-time what libraries to pass in and which to keep default?
<tpw_rules>
another factor is that if it can't find a library, it just vendors it. there's no way to have it fail if it can't find one
<tpw_rules>
which means it's questionable if the inputs you give it will make a difference
<LnL>
anybody around that uses the stateless root setup?
<LnL>
was trying it out but seems like plymouth is trying to use /run/currrent-system before that's created
<slabity>
LnL: Stateless root setup?
<simpson>
tpw_rules: Avoiding vendored code can be good; it means less likelihood of missing out on security updates, for example. But often the vendored version is the only version which works, and that's just the way it has to be.
<LnL>
restoring a blank snapshot for / or using tmpfs
<tpw_rules>
simpson: it works in any case. i think it was all originally vendored as a matter of convenience. i'm not sure if "vendoring" is the right word, because they're all just submodules, not prebuilt or code copies
szicari has quit [Remote host closed the connection]
szicari has joined #nixos
<LnL>
oh no it's a different issue
tsrt^ has quit []
<maier>
why are sometimes previous generations of my nixos system remaining after `sudo nix-collect-garbage -d`?
<maier>
energizer: uhm, I'm not really understanding what I should paste. I see e.g. generation 15 - 19 in my boot loader after having done the GC and `reboot`?
<energizer>
maier: share whatever you're seeing that makes you think something is weird
<slabity>
I think it's a bit difficult to paste something to a website from your bootloader :P
shibboleth has joined #nixos
<slabity>
maier: Do you see the `menuentry` for those generations in `/boot/grub/grub.cfg` (near the bottom)
<slabity>
I think the issue is that the bootloader config doesn't get updated with nix-collect-garbage. You may need to do a rebuild again to get the old ones removed after you run that
Dadou has quit [Remote host closed the connection]
<maier>
slabity: right on, just tried it and I only have the one "NixOS" generation in my boot loader. Thank you :)
<maier>
slabity++
<{^_^}>
slabity's karma got increased to 7
daf90 has joined #nixos
kreyren has quit [Remote host closed the connection]
Robert7 has joined #nixos
<daf90>
is this a good channel for nix support, or just nixos?
<infinisil>
daf90: Both :)
<energizer>
nix too
ixxie has quit [Ping timeout: 256 seconds]
<daf90>
cool. i installed multi-user nix on artix and have just installed fontconfig as a global pkg. wondering what the best way is to do the config files. fonts.conf was placed in /nix/var/nix/profiles/default/etc/fonts. Should i just symlink to that directory, or is there a better way?
<slabity>
daf90: That is probably the simplest way unless you want Nix to be managing your /etc/ files for you
lopsided98 has quit [Quit: Disconnected]
orivej has quit [Quit: No Ping reply in 180 seconds.]
lopsided98 has joined #nixos
orivej has joined #nixos
sheeldotme has joined #nixos
sheeldotme has quit [Client Quit]
alp has joined #nixos
mallox has quit [Quit: WeeChat 2.8]
ju5tevg3niy has joined #nixos
sheeldotme has joined #nixos
maier has quit [Quit: leaving]
sheeldotme has quit [Client Quit]
Robert7 has quit [Ping timeout: 245 seconds]
ju5tevg3niy has quit [Remote host closed the connection]
nate48423 has quit [Ping timeout: 244 seconds]
absoluutely has joined #nixos
<tpw_rules>
has anybody made a derivation which involves the cmake file(DOWNLOAD)? it seems like it's just being ignored and acting as everything is fine even though the file isn't downloaded. of course it works fine in a nix-shell --pure
<tpw_rules>
does including libcxx as a build input automatically set everything up for LLVM? that seems to be the case, but this package needs to link with it despite being built with g++
<samueldr>
so by reading more carefully, it seems there is no way to have the default behaviour (dhcp on all interfaces) while having a bridge too?
<samueldr>
and I see "+1 for more deprecation on networking.useDHCP", but I have no idea what the alternative would be??
lsix has quit [Quit: WeeChat 2.8]
<samueldr>
should I get educated in networking to setup a simple server?
<cmk_zzz>
I have a shell.nix file. I want to use libcurl-gnutls and I see that there is an optional gnuTls support in the curl package. How do I specify this in my shell.nix?
<jneplokh>
I'm sure this has come up a lot, but I have also seen a bunch of different solutions, so I thought I would ask. What's the best practice for password files? For example, Nextcloud needs an admin pass, but not sure what's the best place to store the file. What do you guys do?
pr06lefs has quit [Ping timeout: 260 seconds]
lopsided98 has quit [Quit: Disconnected]
<nh2[m]>
jacob: I currently store mine in the mutable state `/var`, or using the nixops "keys" feature when using nixops
<nh2[m]>
Is there a way to first-class override a `config` (the NixOS system config) value? I have at hand a `config` value and want to have a copy of it that has `boot.loader.grub.device` set to something else. Like .override or .extend, but for system configs.
<samueldr>
>> Whether to use DHCP to obtain an IP address and other configuration for all network interfaces that are not manually configured. Using this option is highly discouraged and also incompatible with networking.useNetworkd. Please use networking.interfaces.<name>.useDHCP instead and set this to false.
<infinisil>
nh2[m]: You can just add another module that has that value set with mkForce
<samueldr>
why is it discouraged?
<samueldr>
it feels like someone knows something that I should know, but doesn't want to tell
laerling has joined #nixos
laerling has quit [Changing host]
<infinisil>
nh2[m]: E.g. if `configuration.nix` is the base, the module { imports = [ ./configuration.nix ({ lib, ... }: { boot.loader.grub.device = lib.mkForce "foo"; }) ]; } should do that
<nh2[m]>
samueldr: I think they want you to enable `.useDHCP` for each individual network interface, instead of using a global bool for all network faces. I believe that's all
<samueldr>
but why?
<samueldr>
why should I care?
<samueldr>
I want dhcpcd to just get the addresses, wherever they are
<samueldr>
in this particular instance (I think) the interface keeps its name
<samueldr>
and that makes it a hardware-dependent value!!
<samueldr>
if I have two different hardware machines that have different hardware device names for their network access, now they need two different configurations rather than share
<nh2[m]>
samueldr: I'm not sure about it, but one reason I can imagine is that it'd be ambiguous to have `networking.interfaces.eth1.useDHCP = false` and `networking.useDHCP = true`, you'd always have to define in which case which ones take precedence. But I think it could be made so that there's an assert that checks that if you use the global option, you're not allowed to use any of the specific options. I do agree that it's quite
<nh2[m]>
annoying to have to list the hardware-dependent interface names in the config in all cases
<samueldr>
and right now it's irritating because I'm being artificially stopped by the nixos config to do what was doing fine beforehand
<samueldr>
so because there is a bridge, and because dhcpcd doesn't give IPs anymore, I'm forced to change the config in an unergonomical way, while the change doesn't affect me at all
<nh2[m]>
jacob: should be, yes, you just have to make sure that the file system permissions of that dir/file are so that you those allowed to read it can read it
<jneplokh>
Yeah, I'll make it that only `wheel`/root can read it
<nh2[m]>
infinisil: or would I literally do `imports = [ ./qemu-vm.nix ` in `qemu-vm.nix`? It seems a bit odd
<CRTified[m]>
jacob,nh2 I'm also storing secrets in `/var`. I wrote a module that uses `preStart` to generate the secrets on demand, e.g. for services that can not read a passwordFile (but can include configs)
<infinisil>
nh2[m]: The hacky way would be to call lib.evalModules again with the same arguments as in nixos/lib/eval-config.nix
<infinisil>
Where you get the arguments like baseModules, extraModules, etc. from the module arguments at the top ({ pkgs, lib, baseModules, ... })
cosimone_ has joined #nixos
zanc has quit [Read error: Connection reset by peer]
dermetfan has quit [Ping timeout: 244 seconds]
<nh2[m]>
then the throwaway VM will fail.
<nh2[m]>
infinisil: I'm pretty blocked on how I should do this. I found the bug in the VM creator: It uses a throwaway VM (`runInLinuxVM`) to create a disk image with grub on it. It uses the "real" VM's `config` inside that throwaway VM to say where the grub is to be installed. But the issue is that in the throwaway, grub must always go to /dev/vda, as only the grub disk. But if the "real" vm wants grub to be installed on e.g. /dev/vdb,
mbrgm_ has joined #nixos
Thra11 has quit [Ping timeout: 265 seconds]
<nh2[m]>
infinisil: So I thought somehow overriding `config` might do it, but maybe there's also a completely different way to solve this
cosimone has quit [Ping timeout: 256 seconds]
mbrgm has quit [Ping timeout: 260 seconds]
mbrgm_ is now known as mbrgm
Thra11 has joined #nixos
sheeldotme has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<nh2[m]>
etrigan63: I like your suggestion, I also commented now. I think it will need some people to step up to do that work though. If you become a NixOS user and want to contribute in this direction, I'd be happy to review that stuff