<aanderse-web>
brand new install of nixos 19.09 from scratch and getting a "Not authorized to control networking." when trying to connect to wifi via network manager. any ideas? internet says i need to be in "netdev" group, but there is no such group
mbrgm_ is now known as mbrgm
<qyliss>
aanderse-web: networkmanager group
aanderse-web has quit [Remote host closed the connection]
aanderse-web has joined #nixos
<aanderse-web>
qyliss++
<{^_^}>
qyliss's karma got increased to 17
<aanderse-web>
thanks!
<gyroninja__>
How do I set an optional argument to a package in configuration.nix. The package is being pulled in by a nixos module, so I am not sure what to do.
<gyroninja__>
Do I have to create a config.nix for the root user?
<qyliss>
What module?
<qyliss>
Does it have a "package" option?
<gyroninja__>
no
<qyliss>
What package and what argument?
<qyliss>
Does it look at Nixpkgs' config?
<gchristensen>
wow, cool! my system is python2-free after jtojnar's PR, plus one PR I'm about to send
<qyliss>
!!
aanderse-web has quit [Ping timeout: 260 seconds]
<gyroninja__>
nice
<{^_^}>
[nixpkgs] @Ma27 merged pull request #75787 → nixos/nixos-option: "See also configuration.nix manpage" in nixos-option manpage → https://git.io/Je7mD
<nh2>
is there a version of `.override` or `.overrideAttrs` that allows me to get at the previous derivation arguments (like check whether `enable*` is true/false)?
<bt`>
I've experienced a suprisingly high rate of failure while compiling packages for my 32bit system since I installed two days ago. Significant packages like xf86-video-intel, mpv, and firefox-esr have failed to compile, and I've not attempted many more than these. Following this I have several questions: Is there any sort of automated testing for non-AMD64 compilation? Is aarch64 in a similar state? Are there plans to remove support
<bt`>
for non-AMD64 cpus?
vld has quit [Ping timeout: 268 seconds]
<gchristensen>
bt`: aarch64 is in very good state
<bt`>
gchristensen: well that's good at least, are there plans to drop 32bit support then, I know many other distributions have done this.
<gchristensen>
armv7l even is in a pretty okay state. few contributors put in work to keep i686 working, and the result is what you're finding
<gchristensen>
(armv7l being 32bit, but not x86)
<xavierzwirtz[m]>
Resolved my issue by swapping to nixpkgs unstable
takeda has joined #nixos
<bt`>
ah okay. I wonder if checking for x86-32 build success could improve derivation quality. Might bit-rot in x86_32 be a indicator of other issues, or is it usually independent?
<colemickens>
when I plug/unplug this zwave/zigbee adapter, sometimes it shows up as /dev/ttyUSB{0,1} and sometimes as /dev/ttyUSB{1,2}. Is this a case where I need to author udev rules to give them consistent names?
<colemickens>
and if so, does someone have an example or a guide? I'm vaguely aware of udev, but I've never done more than activate a rule or copy/paste a line, etc.
<gchristensen>
bt`: we stopped building the full package set on i686 because maintainers didn't seem interested in maintaining it. may I ask why you're using i686?
mexisme_ has quit [Ping timeout: 248 seconds]
orivej has quit [Ping timeout: 258 seconds]
<bt`>
gchristensen: I just use a 32bit x86 machine, nothing to fancy behind the reasoning. I see OfBorg checks PR's for aarch64 compilation in at least one case, is this entirely up to maintainers then? In any case I certainly understand if there aren't benefits to continuing support, but if support is going to be this poor you might consider dropping it entirely.
<DigitalKiwi>
is your machine a toaster
<gchristensen>
bt`: yes, right now we consider i686 to have "limited support". this means we build a minimal system and leave it up to that
<gchristensen>
it may indeed be time to shut it down
<bt`>
DigitalKiwi: I suppose, but it's plenty for me. I'm not complaining to be clear I understand this project is run on a voluntary basis and I'm greatful for all the innovation you've acheived, just trying to understand where things stand. The only thing that has recently posed a issue is software compatibility, and I was running Gentoo before this where it wasn't a issue.
<DigitalKiwi>
the first time arch linux 'dropped support for i686' a lot of people that didn't think they had 64 bit computers found out they really did and switched...and then it was an april fools joke. which means the second time they dropped it was really just an april fools joke gone too far
<gchristensen>
bt`: I do think it is useful to continue supporting i686 in a minimal sense, in case somebody is motivated to improve it in ways useful to them, keeping it in the build farm makes it much less feasible to do that
o1lo01ol1o has quit [Remote host closed the connection]
kleisli has joined #nixos
<gchristensen>
worldofpeace: I'm really not sure, I'm sorry -- I'm not up to speed on buildINputs/nativeBuildINputs, and I don't know the significance of this script in the first place -- just that it looks like it certaily supports python3, and switching is better than it using python2 :)
h0m1 has quit [Ping timeout: 248 seconds]
<gchristensen>
I am happy to follow your lead :)
<DigitalKiwi>
bt`: nice wattage at least ;p
<takeda>
I learned today that fetchgitPrivate was removed last month, what are the alternatives to use it to fetch dependencies in private repos for GoLang? I tried to use builtins.fetchGit, but looks like it is not fully compatible, for example it can't fetch github.com/BurntSushi/toml tag v0.3.1 any suggestion how I can make buildGoPackage work again? If there's a way to make buildGoModule work that would also work
<gchristensen>
takeda: in what way can it not fetch that repository at that tag?
h0m1 has joined #nixos
<worldofpeace>
gchristensen: mind if I fix it in your PR? (👋 python2)
<takeda>
it says
<takeda>
fetching Git repository 'ssh://git@github.com/BurntSushi/toml'error: cannot update ref 'refs/heads/v0.3.1': trying to write non-commit object f8b216ef3a5096277b4fe217e228b15b700be5cd to branch 'refs/heads/v0.3.1'
<gchristensen>
worldofpeace: please :D
<takeda>
actually initially complained about sha256 not being there, but I wrapped it
<worldofpeace>
heh, I probably shouldn't be doing this at this hour. but python2... :)
<gchristensen>
:P
siiky has quit [Ping timeout: 250 seconds]
<gchristensen>
takeda: I wonder why use fetchGit for a public repository?
<takeda>
so I have deps.nix generated by vgo2nix, most deps are public but some are private, the fetchgitPrivate worked with all of them, but it is now gone and I don't know what's the replacement for it
<gchristensen>
takeda: and instead of setting `ref` to `v0.3.1` use `refs/tags/v0.3.1`
<takeda>
yes, that worked, but now some deps have just hash like github.com/PagerDuty/go-pagerduty with rev 94ee1c55dbdb
<takeda>
it doesn't work on that
ddellacosta has quit [Ping timeout: 268 seconds]
<evanjs>
gchristensen: is it actually better to use fetchgit for public? I know fetchGit works better for private, but wasn't aware of any implications re the reverse
<gchristensen>
evanjs: I *think* pkgs.fetchgit will have better caching characteristics? not 100% sure
<takeda>
but is there a way for me to decide what to use from buildGoPackage?
<gchristensen>
takeda: sorry, I don't have a good answer, then, maybe you can examine the source URL and swap in pkgs.fetchgit and builtins.fetchGit based on if it smells like a private URL
acarrico has joined #nixos
<takeda>
the deps.nix generated by vgo2nix is just a nix structure doesn't have any place to reference functions :/
<gchristensen>
you're overriding fetchgit, you can do it there
justanotheruser has quit [Ping timeout: 246 seconds]
selfsymmetric-mu has joined #nixos
<gchristensen>
afaict, worldofpeace, after the PRs I've already submit merge the only remaining py2-dependent packages I'll have are gconf (not sure what to do with that since it isn't really supported anymore?) firefox and chrome
<worldofpeace>
gchristensen: seems firefox adds it unconditonally, I believe for the build, but there's a condition ffversion "61.0" to add python3
<{^_^}>
[nixpkgs] @ehmry pushed commit from @mmahut to master « alertmanager-bot: init at 0.4.0 »: https://git.io/Je5R9
drakonis has quit [Quit: WeeChat 2.6]
<worldofpeace>
well that and some aaliyah song on loop makes the packaging go quicker. I know the feeling of torturing these laptop cpu's, can you say thermal throttle 😇
<gchristensen>
no doubt :D
das_j has quit [Remote host closed the connection]
Scriptkiddi has quit [Remote host closed the connection]
das_j has joined #nixos
Scriptkiddi has joined #nixos
<worldofpeace>
lol, you've unintentionally make a 90s reference without knowing it too
blaggacao has quit [Read error: Connection reset by peer]
blaggacao has joined #nixos
lord| has quit [Ping timeout: 248 seconds]
<gchristensen>
worldofpeace: I guess for a proper test I should *boot* this system :P back soon.
<worldofpeace>
👍️
lord| has joined #nixos
blaggacao has quit [Read error: Connection reset by peer]
blaggacao has joined #nixos
<Ankhers>
I have a script that uses /bin/bash as the interpreter. Is there something I can put in my nixos configuration that would make a link for that? Or is there some other appropriate solution?
<Ankhers>
I would rather not have to go through and change all of the scripts because it is made by android studio and I would not want to have to periodically fix things.
blaggacao has quit [Read error: Connection reset by peer]
<Ankhers>
qyliss: Thanks. Manual symlink it is for right now.
<worldofpeace>
gchristensen: forgot a question mark there, hplip?
<gchristensen>
nah, it has never worked like that I think
<gchristensen>
hplip itself works fine as a print driver, its utilities don't
<gchristensen>
anyway, added dbus-pythno
<jbetz>
what's the recommended nixos version for the best KDE experience?
<gchristensen>
several utilities superficially work, some fail on other reasons. I'm not real interested in fixing them all
<gchristensen>
especially since almost all of them have probably been broken forever
<{^_^}>
[nixpkgs] @fpletz pushed 3 commits to release-19.09: https://git.io/Je50I
<worldofpeace>
gchristensen: hp-check is particularly fun to watch die :)
* gchristensen
hides behind a pillow
<jbetz>
I'm on 19.09 currently, and the worst part has been the display crashing each time it goes into sleep mode. i.e., the wallpapers black out, and then there's a error notification about a missing KDE or QT module.
<worldofpeace>
gchristensen: oop, it might've been a bit scary
<jbetz>
and then the display just freezes on some applications, requiring a logout to get things working again
<worldofpeace>
jbetz: It sounds like the locking dbus program/interface may not be properly wrap and crashes killing your session. but only triggered on dpms operations
<worldofpeace>
* wrapped
<worldofpeace>
jbetz: what's your nixos-version?
<jbetz>
worldofpeace: anything I can do to verify that?
<jbetz>
I'm on 19.09
felixfoertsch has joined #nixos
<worldofpeace>
jbetz: nixos-version is actually a command that will give a string with a nixpkgs commit. that's helpful to check if the issue wasn't already fixed.
<worldofpeace>
jbetz: if you can get logs from the journal would help, the error message is pretty familiar
<jbetz>
ah, okay. in that case: 19.09.1155.bae4d7daa01 (Loris)
<buckley310>
is there any legitimate need for nixos-generate-config to pull in "/dev/zramN" devices? it seems to break if both hardware-config and "zramSwap.enable" are used.
pbb has joined #nixos
hyper_ch has quit [Ping timeout: 268 seconds]
mexisme_ has quit [Ping timeout: 245 seconds]
<CMCDragonkai>
I have a quick question about `buildGoModule`. How do I override the resulting derivation to add an extra go dependency into the Go derivation such that it can be imported.
<keithy[m]>
hi, is command-not-found supposed to work?
<jbetz>
worldofpeace: can't confirm yet whether that fixed it, but it seems to at least be surviving the dim screen cycles. however, after running a rollback, the behavior didn't change, so that might not be telling of much. if it happens again, I'll see if I can get a better handle on where it's coming from using journal logs and let you know. thanks!
blaggacao has quit [Read error: Connection reset by peer]
blaggacao has joined #nixos
jarlg has joined #nixos
alexherbo2 has joined #nixos
<jbetz>
noteworthy is that the dim screen goes through a cycle rather than staying off. it will turn off after however long I configure it for, but then display comes back a few seconds after, so something definitely is going wrong there.
blaggacao has quit [Read error: Connection reset by peer]
blaggacao has joined #nixos
takeda has quit [Ping timeout: 258 seconds]
bvdw has quit [Read error: Connection reset by peer]
blaggacao has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
blaggacao has joined #nixos
jarlg has quit [Quit: WeeChat 2.6]
blaggacao has quit [Read error: Connection reset by peer]
blaggacao has joined #nixos
blaggacao has quit [Read error: Connection reset by peer]
blaggacao has joined #nixos
blaggacao has quit [Read error: Connection reset by peer]
bt` has left #nixos ["ERC (IRC client for Emacs 26.3)"]
mexisme_ has joined #nixos
turq has joined #nixos
mexisme_ has quit [Ping timeout: 248 seconds]
<turq>
how can I use this? https://github.com/ocharles/netwire-classics the error received when running nix-build in any one of the game directories is that mkDerivation is missing from the default.nix file
<turq>
and by use, I mean build
<{^_^}>
[nixpkgs] @buckley310 opened pull request #75921 → nixos-generate-config: don't generate swapDevices for zram devices → https://git.io/Je5ut
felixfoertsch has quit [Quit: ZNC 1.7.3 - https://znc.in]
felixfoertsch has joined #nixos
vld has joined #nixos
`slikts has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @FRidh pushed 43 commits to staging-next: https://git.io/Je5gU
<{^_^}>
[nixpkgs] @FRidh opened pull request #75929 → Staging next → https://git.io/Je5gT
dansho has quit [Ping timeout: 258 seconds]
rauno has joined #nixos
pie_ has quit [Ping timeout: 268 seconds]
sigmundv has joined #nixos
day|flip has quit [Remote host closed the connection]
blaggacao has quit [Ping timeout: 265 seconds]
fusion809 has quit [Remote host closed the connection]
Makaveli7 has quit [Ping timeout: 240 seconds]
makaveli1 has joined #nixos
blaggacao has joined #nixos
<{^_^}>
[nixpkgs] @ijaketak opened pull request #75931 → Fix haskellPackages.dhall_1_28_0 and spago → https://git.io/Je5g1
<lovesegfault>
How does the process around `staging` work?
<lovesegfault>
cc. clever
chloekek has joined #nixos
<etu>
lovesegfault: I might be off about this. But, afaik: Things that cause big rebuilds are put in a branch named staging. Then hydra can pick up that branch and do the some of the rebuilds before those changes are merged into master.
<etu>
lovesegfault: staging isn't rebased on top of master, it's merged with a merge commit. And just because it's new git commits in master before staging is merged doesn't make dependencies change.
<lovesegfault>
FRidh: Ah, nice
<lovesegfault>
etu: Oh, that makes sense
philr has joined #nixos
<vika_nezrimaya>
hi! trying to set up a 1-node Kubernetes cluster on a Raspberry Pi 4. Using nixops for provisioning. It looks like etcd cannot start because it can't find some keys that are supposed to be there
<vika_nezrimaya>
kinda like this: open /var/lib/kubernetes/secrets/etcd.pem: no such file or directory
<meatcar>
whats the best way to trim whitespace on a string? `lib.removeSuffix` is a bit wordy, and requires an exact suffix.
<inferencerules>
i had a botched nixos-rebuild last night that left me with a bad nix store db; i can't seem to find anything about this on the usual sites. according to nix pill #2, this db stores dependency info between all derivations, so it seems like something i should be able to rebuild easily, if i can only find the correct command. the error i get when i try to complete the nixos rebuild (or anything else
<inferencerules>
that touches store) is: "error: executing SQLite statement 'pragma synchronous = normal': database disk image is malformed (in '/nix/var/nix/db/db.sqlite')"
<clever>
inferencerules: usually when things get that corrupt, your only option is to delete the entire /nix and re-run nixos-install from a livecd
zupo has joined #nixos
<vika_nezrimaya>
the good thing about NixOS: if your /etc/nixos lives, one command restores the whole system
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<inferencerules>
meatcar: yes, unfortunately i don't have sqlite installed and trying to install anything in this broken state is throwing the same error
Chiliparrot has joined #nixos
vld has joined #nixos
<meatcar>
inferencerules: lucky for you, sqlite is very light and small, I think you can just download the binary, assuming you have internet access still. https://sqlite.org/download.html
inkbottle has joined #nixos
zebrag has quit [Ping timeout: 246 seconds]
<meatcar>
inferencerules: also lots of software use sqlite, `ls /nix/store | grep sqlite` shows three different versions for me, and I don't have it explicitly installed either.
vld has quit [Ping timeout: 268 seconds]
<meatcar>
scratch that last one, there are no binaries there
<inferencerules>
i did end up having a binary there, good call. unfortunately, another "Error: database disk image is malformed". looks like maybe a reinstall is in order as clever suggests
<inferencerules>
well, running the nixos-install now. still curious how such a thing could happen, but i shut down after running the bad nixos-upgrade and lost the terminal, and couldn’t exactly run it again to see
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<{^_^}>
[rfc-steering-committee] @domenkozar pushed to master « README: clarify on the hangout url »: https://git.io/Je5r3
sigmundv has quit [Ping timeout: 265 seconds]
<siers>
clever, nice, it works
cfricke has joined #nixos
<{^_^}>
[nixpkgs] @mgttlinger opened pull request #75939 → nixos/haskell: Added option to the haskell mkDerivation to build quickjump index → https://git.io/Je5rK
<{^_^}>
[nixpkgs] @davidtwco opened pull request #75940 → wooting: init wootility, wooting-udev-rules and module → https://git.io/Je5r9
<DigitalKiwi>
am trying to package glirc for nixpkgs but am having problems https://gist.github.com/Kiwi/5e1b346ea674e6467144e7282c75118e this is a mess i've come up with so far (i have a shell.nix that can maybe almost build it...but i don't know if that's useful or not. had to use an override??) the goal is to add it to nixpkgs and be able to add this plugin https://github.com/glguy/irc-core/tree/v2/otr-extension which needs meson/ninja which i'm sure just m
<DigitalKiwi>
akes this even more complicated than it already is
<genesis>
gchristensen : i even not understand why it isn't resulting in the starting position "hello"
<DigitalKiwi>
i've updated the gist with a shell.nix
<infinisil>
DigitalKiwi: The package seems to have a dependency version constraint using ==, which is usually not very nice for nixpkgs since versions always update. You can use doJailbreak to remove these equality constraints
<{^_^}>
[nixpkgs] @Ma27 merged pull request #75689 → hydra: Temporarily unbreak hydra with the deprecated lib.closePropagation → https://git.io/JeQ2Z
<infinisil>
I'm gonna have a go at fixing the build, gonna take some time though as my internet is pretty shitty rn
<infinisil>
Oh, I could just ssh into a machine with better internet :o
<DigitalKiwi>
assuming ssh works well enough for that heh
<DigitalKiwi>
5 second typing lag D:
<infinisil>
In my case it's just bandwidth that's bad
<gchristensen>
mosh :)
<DigitalKiwi>
so if the build for haskellPackages gets fixed then that'll make it easier to make a top-level one that uses it and adds the otr plugin or
werner291 has joined #nixos
<infinisil>
DigitalKiwi: How is the otr plugin configured?
<{^_^}>
[nixpkgs] @mgttlinger opened pull request #75942 → nixos/haskell: Added quickjump option to the haskell mkDerivation → https://git.io/Je56B
<infinisil>
DigitalKiwi: Ah yeah, remove it from the list of broken packages in the yaml, apply the doJailbreak patch, and do that PR to the haskell-updates branch
<infinisil>
Then when hackage-packages.nix gets generated next time the broken = true is removed
o1lo01ol1o has quit [Remote host closed the connection]
<DigitalKiwi>
ok
o1lo01ol1o has joined #nixos
zupo has joined #nixos
selfsymmetric-mu has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @teto pushed commit from @peterhoeg to master « ppsspp: 1.4.2 -> 1.9.3 »: https://git.io/Je5in
o1lo01ol1o has joined #nixos
fendor has quit [Read error: Connection reset by peer]
<ocharles>
Is there a convenient way to get the build log for a given `.drv`? I know it's in /nix/var/logs/nix/drvs, but constructing that path is annoying
<infinisil>
DigitalKiwi: Should go to the haskell-updates branch
<ocharles>
ryantm: for one, it uses experimental Nix features to give you early cut off. If you add just a blank line to a file, only that file will be rebuilt (assuming the .hi and .o files are bytewise the same at the end)
<DigitalKiwi>
Oh whoops
<ocharles>
this is designed as a drop-in replacement for GHC, so you can do `cabal build --compiler=ghc-nix` and just use cabal as normal
waleee-cl has joined #nixos
<infinisil>
DigitalKiwi: Make sure to not ping everybody when switching branches :P
pie_ has joined #nixos
<DigitalKiwi>
How do I do that right
<fergus>
\ignore
<gchristensen>
gustavderdrache, etu
<gustavderdrache>
gchristensen: your timing is spooky
<sphalerite>
jirido: depends entirely on which software you put on it :)
<symphorien>
not everything is in the binary cache for 32 bits
<symphorien>
so no firefox
dingenskirchen has quit [Quit: dingenskirchen]
<sphalerite>
jirido: it's not alpine-linux-lightweight, but if you use a small (or no) desktop environment and don't need to evaluate too many big things, you should be fine. But yeah, as symphorien says, 32-bit support isn't so great.
kleisli has joined #nixos
dingenskirchen has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jirido>
lol ok no firefox.. hmm
cfricke has quit [Quit: WeeChat 2.6]
<jirido>
but I can compile FF for it right?
<jirido>
but i guess that would take a year on that mashine
kahiru has quit [Remote host closed the connection]
ng0_ has joined #nixos
ng0 has quit [Ping timeout: 260 seconds]
kahiru has joined #nixos
zupo has joined #nixos
<das_j>
<jtojnar "Janne Heß where do you want to t"> Well I was able to fix it without tracing, but the problem persist. I'm calling traceValSeq directly with the derivation (pkgs.hello)
tpham has quit [Remote host closed the connection]
rootatarch has joined #nixos
<rootatarch>
hey math_lover
<{^_^}>
[nixpkgs] @marsam merged pull request #75864 → sqitchPg: fix build on darwin → https://git.io/Je5Th
<{^_^}>
[nixpkgs] @romildo opened pull request #75949 → volctl: init at 0.6.2 → https://git.io/Je5Pw
zupo has quit [Client Quit]
maths_lover has joined #nixos
<rootatarch>
hey maths_lover
<DigitalKiwi>
infinisil: do i change the branch on github before i git push or after
kleisli has quit [Ping timeout: 250 seconds]
hjjkh has joined #nixos
* DigitalKiwi
thinks i'd change it after the push
<DigitalKiwi>
but uhm, i've made big mistakes before so
captn3m0 has joined #nixos
captn3m0- has quit [Ping timeout: 268 seconds]
<DigitalKiwi>
whoops
<DigitalKiwi>
i do not like this
opthomasprime has joined #nixos
opthomasprime has left #nixos [#nixos]
<ocharles>
Anyone got a sec to help with a quick bash scripting q? How can I do this: rsync --recursive -v --include='*/' --include='*.o' --exclude='*' /nix/store/68q18my5jyvm4jj9yall1mb58z5fiij9-compile-Control.Algebra.hs /nix/store/bmfr6bg27xmyvis7avg12wp59xagd68a-compile-Control.Carrier.Choose.Church.hs foo/
<etu>
gchristensen: hey
<ocharles>
such that `foo/` doesn't contain things like `68q18my5jyvm4jj9yall1mb58z5fiij9-compile-Control.Algebra.hs`?
ng0_ is now known as ng0
<ocharles>
It seems that rysnc command drops the `nix/store` common prefix, but I want it to drop another level
maths_lover has left #nixos ["WeeChat 2.6"]
<{^_^}>
[nixpkgs] @gustavderdrache opened pull request #75950 → oniguruma: set CMake flags for static builds → https://git.io/Je5PM
<cvlad->
Seems like it drops the directory but not the file name, right? The file name is actually mangled into `68q18my5jyvm4jj9yall1mb58z5fiij9-compile-Control.Algebra.hs`, rom what presumably was just `Control.Algebra.hs`? Or perhaps you want to move it to `Control/Algebra.hs`?
<ocharles>
cvlad-: /nix/store/68q18my5jyvm4jj9yall1mb58z5fiij9-compile-Control.Algebra.hs has /nix/store/68q18my5jyvm4jj9yall1mb58z5fiij9-compile-Control.Algebra.hs/Control/Algebra.o, and I want `foo/` to contain `foo/Control/Algebra.o`
<ocharles>
And `/nix/store/bmfr6bg27xmyvis7avg12wp59xagd68a-compile-Control.Carrier.Choose.Church.hs` contains `/nix/store/bmfr6bg27xmyvis7avg12wp59xagd68a-compile-Control.Carrier.Choose.Church.hs/Control/Carrier/Choose/Church.o`, which I want to be in `foo/Control/Carrier/Choose/Church.o`
<gustavderdrache>
etu:
<cvlad->
@ocharles
<ocharles>
I can `cd` into each of these nix store paths and do a single `rsync` there into `foo`'s absolute path, but wondered if I could just do one call
<cvlad->
ocharles : What if you add an extra / at the end?
<ocharles>
Hm, that might be it, let me try
<ocharles>
aw yeah
<ocharles>
Thanks cvlad-!
<cvlad->
Lucky guess -- you're welcome! :)
<ocharles>
:)
knupfer has joined #nixos
kraem has joined #nixos
is_null has quit [Ping timeout: 265 seconds]
siiky has joined #nixos
<chloekek>
Say I have a tarball in a local file. Can I find the hash that I would need to pass to fetchTarball, were I to specify a HTTP URL to the same tarball?
<lordcirth>
,tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected. See: tofu-vim
<chloekek>
That would make fetchTarball download the file to discover the hash.
<lordcirth>
Oh I see. You should be able to just run sha256sum
<chloekek>
Ah, so the hash does not include anything like file permissions or file name.
<{^_^}>
[nixpkgs] @cdepillabout pushed 2 commits to haskell-updates: https://git.io/Je516
tobiasBora has joined #nixos
<tobiasBora>
Hello,
<keithy[m]>
it is, but why 127.0.1.1 what about my ip address
<clever>
keithy[m]: thats localhost, the loopback ip
<tobiasBora>
I'd like to know, is there a way to see if a build is going to recompile new stuff or only use the cache?
<tobiasBora>
(I mean when I run nixos-rebuild switch)
<clever>
tobiasBora: nix-build --dry-run
<keithy[m]>
no thats 127.0.0.1
<clever>
tobiasBora: nixos-rebuild dry-run
<clever>
keithy[m]: the entire 127.*.*.* is loopback
evanjs has joined #nixos
Guest33 has quit [Remote host closed the connection]
<bdju>
https://github.com/poljar/weechat-matrix I would love to see this packaged. It looks like an older lua-based matrix plugin is already packaged. The new supported one uses python and supports encryption.
<keithy[m]>
ive also got dnsmasq loadedin "Server" but for some reason it is querying the gateway router, which is set to use "server" for dns... seem needlessly circular
wolfshappen has joined #nixos
<clever>
keithy[m]: i just use proper bind when i want a dns server, ive never messed with dnsmasq
<keithy[m]>
I only need one function in dnsmasq so its good... just somethings wierd.
<tobiasBora>
clever: ok thanks, I'm trying it!
<tobiasBora>
Also, another question. The thunderbird LaTeXMe extention fails, and I tried to redo the command that this extension does, and indeed there is a problem during the conversion of the DVI (generated from a super simple tex file) into a PNG. The error I get is here http://paste.debian.net/1121782
<tobiasBora>
clever: the problem of this command is that I have in the output a huge number of files that are supposed to be compiled, but most of them are not "real compilation", like nixos-version.drv for example I guess... What I mean here by compilation is something that should, in theory, be taken from a cache, but that is not present for some reasons (for example because a built failed on the cache, or because the
<tobiasBora>
cache is not aligned yet with the channel...) One of my fear for example is that I'm asked to recompile the linux kernel (and it's actually quite possible because last time the linux kernel was not available). Or some big programs, like firefox...
<pie_>
how many different variants of "not evaling the right code" can you come up with
<DigitalKiwi>
pie_: there there, just keep swinging the hammers, you'll get it too
<clever>
tobiasBora: not much you can do for that, you just have to learn what package names are cheap and what are expensive, and skim over the list
<clever>
tobiasBora: one thing that i find helps, is `nixos-rebuild -Q` to hide the build logs, then you can see what it is building
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rootatarch has joined #nixos
<clever>
tobiasBora: and if it hangs for more then a minute on a certain job, look at the job names
<pie_>
DigitalKiwi: ;_; thanks
<tobiasBora>
clever: ok too bad... So it means I cannot hope to automatize it. I'd love a script like "go to the last version of the channel that does not need any 'real' compilation."
<pie_>
clever: does hydra publish build durataions?
<pie_>
durations
<clever>
pie_: i think so, but you need to know which eval its in, which means knowing the nixpkgs rev
<clever>
pie_: enless you look up the build graphs for an attr overall
<clever>
pie_: ive played a bit with namespacing, and read over how nix creates a namespace with only lo, but i havent seen the code yet to create a linked pair of ve- IF's in seperate namespaces
<pie_>
i have to figure out when the right time to run the script is before the container is fully up
rootatarch has left #nixos ["WeeChat 2.7"]
<tobiasBora>
hum... but there should be a way no, at least theoretically? I guess hydra has a list of programs she should build, and then we could just compare the hash we get with the hash hydra gets, and see if it differs or not no?
<tobiasBora>
Also, it's not even clear with the current output I have if I'm going to recompile the kernel http://paste.debian.net/1121787
<clever>
tobiasBora: personally, i have my own hydra setup, to build my entire nixos config, against the latest version of a channel
<clever>
tobiasBora: and only if that build is green, will i know that nixos-rebuild switch can pass without errors
<clever>
and due to hydra pre-building things, it wont have to compile anything
stites has joined #nixos
<clever>
tobiasBora: i dont see any kernel in that pastebin, only linux-modules (a buildEnv) and modules-shrunk (the closure used by the initrd)
<stites>
meerp -- can someone help me read a kernel panic? I tihnk either my motherboard or ram has crapped out on me
<clever>
tobiasBora: and a quick skim over it says that the "biggest" thing your building is a single perl package, the rest are just configs
o1lo01ol1o has quit [Remote host closed the connection]
<clever>
pie_: .environment on all of the services, is set to a set from a let block
<tobiasBora>
clever: ok cool. And indeed I just build this perl package... I'll make a "call-clever.sh" script, that sends you this kind of pastebin whenever I need :P What are the tricks to know if a file is just config or more?
<clever>
pie_: you could also read some of those attrs
<pie_>
clever: wrong kind of inherit :(
<pie_>
clever: i mean i need something like ExecStartPost, except i need it to start when execstart is running...not when its finished
<clever>
tobiasBora: half memorizing the name and purpose of every single package, half looking for version numbers
<clever>
tobiasBora: typically config files dont have version numbers in the names
<clever>
pie_: nixos creates a /etc/containers/foo.conf file, that contains the HOST_ADDRESS=
<clever>
pie_: and EnvironmentFile says to load that, based on the %i (the name after the @)
hmpffff has joined #nixos
<pie_>
hm
o1lo01ol1o has quit [Ping timeout: 240 seconds]
<tobiasBora>
clever: for weechat-bin-env-2.6, you know that -env is not refering to the actual executable? If yes, can I say something like "any file containing -env- is unlikely to be a real build"?
alexcouch has quit [Remote host closed the connection]
endformationage has joined #nixos
jwaksbaum[m] has joined #nixos
dbmikus has joined #nixos
<jwaksbaum[m]>
I'm trying to run a custom kernel on nixos. It fails to boot waiting for the root disk to appear. I know the uuid is right because it boots fine using the regular kernel. Does anyone know why this would happen?
<{^_^}>
[nixpkgs] @brprice opened pull request #76052 → nixos/acme: implement postRun using ExecStartPost → https://git.io/Je5DF
<pie_>
jwaksbaum[m]: can you give additional info? what kind of changes are in the custom kernel?
<pie_>
are there any error messages?
<jwaksbaum[m]>
pie_: the custom kernel is the chromium is kernel. The hardware is a Chromebook, so I'm hoping running the chromium kernel will give better hardware compatiblity, but I don't know exactly the changes.
shibboleth has joined #nixos
<jwaksbaum[m]>
It says it's waiting on /Dev/whatever to appear, then it doesn't appear and it tries to mount anyway and fails
<jwaksbaum[m]>
The only other error message is about failing to bind to some audio thing
<pie_>
hm, i dont know but maybe with that info someone else will be able to say something
<{^_^}>
[nixpkgs] @primeos opened pull request #76053 → [19.09] Ruby security backports for my packages → https://git.io/Je5yI
<pie_>
infinisil: re systemd stuff, i need to time the starting of a service like this: A.service runs it's execStart -> b.service starts up, any idea how to do that?
chloekek has joined #nixos
<pie_>
infinisil: i need to ifconfig up an interface provided by systemd nspawn in execStart, before the execstart finishes
<pie_>
and by before i mean this is a dependency of execstart being _able_ to finish
nehh^ has quit []
<infinisil>
I'm not sure what you mean
<infinisil>
"time the starting"?
<infinisil>
If something needs to happen before ExecStart finishes, just put it in ExecStart?
<pie_>
infinisil: i cant, the final command in execstart that im waiting for blocks
<infinisil>
pie_: See `man systemd.service` for what that does
acowley[m] has joined #nixos
<pie_>
infinisil: i looked at that but no because it only runs after execstart finishes
<infinisil>
It depends on the Type
alexherbo2 has joined #nixos
pistache_ is now known as pistache
<infinisil>
Man that containers.nix module is so messy
<pie_>
apparently its of notify type
<pie_>
which makes a lot of sense
<pie_>
"; however, it is expected that the service sends a notification message via sd_notify(3) or an equivalent call when it has finished starting up."
<infinisil>
Yeah and ExecStartPost mentions something too
<infinisil>
for notify
boogiewoogie has joined #nixos
<pie_>
hm, so does that mean the nspawn returns immedately
dansho has joined #nixos
<infinisil>
I see `--notify-ready=yes`
<infinisil>
In nspawns args
<pie_>
infinisil: it just sets an env var if i read that correctly so useless in this case
<pie_>
i mean the execstartpost comment in the docs
<mrlizard>
Hello. It seems that I have issues with nixos update. Did someone faced similar issue? Whole output of update: https://paste.centos.org/view/2f7bdc04
<infinisil>
pie_: Why would that be an env var?
is_null has joined #nixos
<mrlizard>
Strange that bash wan's to update by recompilation.
<mrlizard>
*want's
<pie_>
ExecStartPost= commands are only run after the commands specified in ExecStart= have been invoked successfully, as determined by Type= (i.e. the process has been started for Type=simple or Type=idle, the last ExecStart= process exited successfully for Type=oneshot, the initial process exited successfully for Type=forking, "READY=1" is sent for Type=notify, or the BusName= has been taken for Type=dbus).
<pie_>
ok whatever that READY=1 thing is
<pie_>
but its still "after it has finished"
lovesegfault has quit [Ping timeout: 252 seconds]
<infinisil>
Read the `Type` section in the same man page, where notify is explained
<mrlizard>
Ah.. it's wireguard issue.
<pie_>
im guessing thats an sd_notify thing
<mrlizard>
Ok. My fault.
MinceR_ has joined #nixos
<infinisil>
pie_: Indeed. But really, I have no idea how to solve the problem, I never used containers
<pie_>
unless youre trying to point something out it still looks to me like everything related to notify still just says it will wait for execstart to finish (or in this case for the notification to get sent) before continuing
<pie_>
its not a container problem its a running things at the right time problem
<pie_>
tl;dr ifconfig ve-something up needs to be run after nspawn but before it times out
<rnhmjoj>
infinisil: hi, could you review, as a code owner, my PR #71473? it's been open for a while...
<boogiewoogie>
hey, I want to use a specific older mesa package via `hardware.opengl.package` option (using mkDefault, see the latest post in https://github.com/NixOS/nixpkgs/issues/75776), however defining this in my configuration gives `error: The unique option `hardware.opengl.package' is defined multiple times`. how would I disable/hide the default option?
<{^_^}>
#75776 (by boogewooge, 3 days ago, open): Numerous games crashing on 19.09 exclusively
<infinisil>
pie_: Hm regarding your issue, I have the feeling your mind is ahead of your hands. It's really hard to know what you want to do, the issue seems to be mostly some parts of your thought process, without explaining the issue for someone not familiar with it
<{^_^}>
[nixpkgs] @brprice opened pull request #76055 → nixos/nginx: drop extra semicolon in return example → https://git.io/Je5SU
Lears has quit [Ping timeout: 258 seconds]
<pie_>
possibly
<infinisil>
pie_: E.g. I'm reading "VPN service that needs internet", "a container is not up until ...", what has this to do with containers? Are you running a vpn inside the container?
<pie_>
infinisil: yes
mexisme_ has joined #nixos
<infinisil>
Like, I can ask a bunch of questions like this, but ideally the issue would tell everybody this already from the beginning
<pie_>
i added those to the bullet points
tom39291 has quit [Quit: WeeChat 2.4]
andreas303 has quit [Remote host closed the connection]
<pie_>
infinisil: it was clear to me because ive been bashing at this for two days now and my brain thought i obviously explained that i ran it in the container
<infinisil>
Yeah :P
<pie_>
also lack of sleep yay
<pie_>
or rest
<pie_>
not sure
<infinisil>
pie_: Does the vpn have a dependency on network-online.target?
<infinisil>
Because it seems like it should
<pie_>
im not sure
<pie_>
infinisil: but i think you may have missed the point, the container will not come up till it has internet access, and it will not have internet access till it has come up
andreas303 has joined #nixos
<pie_>
i tried adding that as a dependency and i think it just broke even more
<pie_>
from what i skimmed out of the docs, network-online just checks if the internal interfaces are configured
<Raito_Bezarius>
What can I do when my Nix Store SQLite database seems corrupted?
<Raito_Bezarius>
I get this error: error: executing SQLite statement 'pragma synchronous = normal': unable to open database file (in '/nix/var/nix/db/db.sqlite')
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<pie_>
which is orthogonal to the issue i think
<infinisil>
pie_: No that's network.target
<Raito_Bezarius>
My server had no space left for some hours, journald was broken and everything broke
<infinisil>
network-online.target is a bit more smart
<infinisil>
See `man systemd.special` for docs on these
<infinisil>
pie_: (you'll want `network-online.target` in `wants` and `after`)
zeta_0 has joined #nixos
<pie_>
infinisil: network online isnt even granular
<pie_>
though might be fine for a single purpose container that doesnt do any fancy networkin
<pie_>
these systems feel like they are way more complicated and unwieldy than they should be
<pie_>
or im just bad
<pie_>
well, i guess these are systems we have as opposed to might have in 10 years
<{^_^}>
[nixpkgs] @raboof opened pull request #76061 → <!-- Nixpkgs has a lot of new incoming Pull Requests, but not enough people to review this constant stream. Even if you … → https://git.io/Je599
oleks has joined #nixos
shapr has joined #nixos
Phlogistique has joined #nixos
averell has joined #nixos
shapr is now known as Guest30294
rager has joined #nixos
trfl has joined #nixos
nore has joined #nixos
otti0815 has joined #nixos
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
<{^_^}>
[nixpkgs] @flokli merged pull request #75721 → nixosTests.networking: Port tests to python → https://git.io/JeQ1N
<duairc>
I guess I can just read the source code of nix myself and see what it's doing
<DanP[m]>
How do I set the kernal boot params? I've got `flip_done timed out` errors all through my logs and I think it's related to some long freezes I've been seeing when disconnecting from my dock.
ng0 has joined #nixos
<DanP[m]>
apparently some people have had some luck setting the kernal boot param `video=SVIDEO-1:d`, and I'm not sure how to do that in NixOS-land
xelxebar has joined #nixos
turbo_MaCk has joined #nixos
<pbogdan>
boot.kernelParams?
eacameron has joined #nixos
<eacameron>
I'm trying to build a rust app the normie way on NixOS. It's complaining about LIBCLANG_PATH not providing the right so. It's not even set. Anyone know how to make this happy?
<DanP[m]>
huh, that didn't show up when searched options for 'kernal'. But thanks, I'll try that
<lordcirth>
eacameron, find where libclang is in /nix/store, and pass it as LIBCLANG_PATH?
zeta_0 has joined #nixos
<eacameron>
lordcirth: I could search /nix/store but I'd rather do something like `nix-build <nixpkgs> -A theRightThing` and pass that
<exarkun>
DanP[m]: make sure you spell it "kernel" and not "kernal".
<pbogdan>
kernel != kernal ;)
cfricke has joined #nixos
<exarkun>
it's like your 2nd grade teacher always said, spelling counts
cfricke has quit [Client Quit]
<lordcirth>
eacameron, might be easiest to just make a derivation for your rust package?
<DanP[m]>
ah, there we go. :the-more-you-know:
<eacameron>
lordcirth: Maybe... but I'd rather not go that route right this moment
drakonis has quit [Quit: WeeChat 2.4]
drakonis1 has joined #nixos
drakonis1 has quit [Client Quit]
<mcbeth>
Or maybe DanP is working on a Commodore machine, where that is the right spelling :)
drakonis has joined #nixos
jirido has quit [Ping timeout: 240 seconds]
zaeph has quit [Ping timeout: 268 seconds]
<sphalerite>
pie_: oh dear x)
<zeta_0>
how do i get ghcWithHoogle to work with ghcmod in home.nix? i am currently doing some web development, and i get the error: `ghcmod: could not find module Yesod`? https://hastebin.com/oyefaxojac.py
<zeta_0>
infinisil: is this ghcmod problem with ghcWithHoogle or all-hies, or is this a problem with hie itself, if its the latter i'll go ask for help in hie channel instead?
<zeta_0>
infinisil: my other comment is a little bit further up if you have not seen it yet
philr has joined #nixos
leotaku has quit [Ping timeout: 268 seconds]
<{^_^}>
[nixpkgs] @raboof opened pull request #76066 → vscode.extensions.scala-lang.scala: init at 0.3.8 → https://git.io/Je5HA
<infinisil>
Taneb: Doesn't look like the redirect-gateway applied. Did you restart the openvpn service?
fendoer has joined #nixos
<Taneb>
I think so
<infinisil>
dansho: ${placeholder "out"} should probably work
robogoat has joined #nixos
<Taneb>
infinisil: restarting again doesn't change anything
<infinisil>
Taneb: Hm well with that `ip route` it looks like the openvpn config doesn't even work
<infinisil>
It should at least have some ips for the tun interface
<infinisil>
You sure it works at all? Maybe check the logs
fenedor has quit [Ping timeout: 268 seconds]
acarrico has joined #nixos
jirido has joined #nixos
<Taneb>
Ah! I restarted it again and it's added some routes
m0rphism has quit [Ping timeout: 268 seconds]
<Taneb>
DNS not working, though
<dansho>
infinisil: huh weird it resoved to some long hash, not a store path: https://paste.ee/p/73Pr9
<infinisil>
dansho: Ah then it doesn't work. That's how placeholder is implemented
<infinisil>
Taneb: What's `ip route`?
<zeta_0>
infinisil: is ghc-mod a part of all-hies? because my web app compiles and runs fine but i am getting this weird error message: ghcmod: could not find module Yesod and hie stops working so it is very frustrating?
<infinisil>
dansho: So yeah, doesn't work, you'll have to use something that can read env vars
<infinisil>
zeta_0: Many things are parts of other things, but no I don't have many clues about ghc-mod, and I don't know the problem here
<infinisil>
Taneb: Is that shortened or the full output?
<zeta_0>
infinisil: ok, i asked in the haskell channel as well, but they don't seem to understand what is going on either, so i guess i am just going to have to do some more research on this and some more trial and error troubleshooting
<Taneb>
infinisil: full. There may be typos
<infinisil>
And I guess the 66 is your subnet
<Taneb>
Yeah, it's what I expect the VPN to assign me
<infinisil>
Taneb: I hope the 178 is a typo in the last line?
<Taneb>
Yeah
<Taneb>
I've edited the gist
<infinisil>
Taneb: And everything other than DNS works?
<Taneb>
There's not much I've tested. I can ping, ssh into, and view in a browser an IP
<Taneb>
(on the other side of the VPN)
<infinisil>
And how does DNS not work?
<Taneb>
I cannot resolve any domain names, whether on the VPN's local network or on the general interent
<infinisil>
Taneb: Can you `dig duck.com @1.1.1.1`?
bvdw has quit [Read error: Connection reset by peer]
<Taneb>
Yes (after disconnecting from the VPN to install dig and then reconnecting)
<Taneb>
Tells me three IP addresses
bvdw has joined #nixos
<zeta_0>
infinisil: if i understand correctly, hie is going to stop using ghc-mod very soon, that's probably why i am getting that error in emacs
m0rphism has joined #nixos
selfsymmetric-pa has quit [Remote host closed the connection]
<infinisil>
Taneb: And what about `dig duck.com`?
<infinisil>
Or another domain
<infinisil>
(without specifying the name server with @)
<Taneb>
"connection timed out; no servers could be reached"
<infinisil>
Taneb: What's in your /etc/resolv.conf?
<Taneb>
search home
<Taneb>
nameserver 192.168.1.254
<Taneb>
options edns0
<Taneb>
That's it
<infinisil>
Ahh
<infinisil>
Right, DHCP sets your router to be the DNS server
<Taneb>
Which is an improvement, this situation is less than ideal for me but workable
<infinisil>
Taneb: What specifically doesn't work now still?
<Taneb>
I would like to be able to write "ssh howl" where howl is a machine on the local network of the VPN server, but when I do that I get the message "ssh: Could not resolve hostname howl: Device or resource busy"
<infinisil>
Taneb: You can do that by mapping howl to an ip in the ssh config
<infinisil>
Host howl\n HostName 192.168.66.X
<gchristensen>
or put the private IPs in public DNS (like howl.yourdomain.com)
<Taneb>
infinisil: I'd rather avoid that, if possible. There's a few I'd like to access, and when I connect to the VPN on my desktop (using the gnome network manager applet thingy) it works fine
<infinisil>
Hm, I'd think networkmanager might have some DNS server inbetween that does this host mapping
m0rphism has quit [Ping timeout: 260 seconds]
phreedom has joined #nixos
xelxebar has joined #nixos
fendoer has quit [Quit: Leaving]
andreas303 has joined #nixos
m0rphism has joined #nixos
gkmngrgn has joined #nixos
<Taneb>
I'm not going to worry about this for now, because it's late and I don't really need this until Monday