<{^_^}>
→ 52858e42 by @bkchr: kde_frameworks: 5.44 -> 5.45
<{^_^}>
→ 39ba6636 by @bkchr: ki18n: Adds patch that fixes infinite loop in po files generation
<{^_^}>
→ b162c2c0 by @peterhoeg: Merge pull request #38960 from bkchr/kde_frameworks_5_45
jtojnar has joined #nixos
jperras has joined #nixos
<zybell_>
another way:by patching the kernel install in /proc a 'session' link analog to 'self' now, but aimed at the sessionleader of the current process. That would make /proc/session/cwd point at the current $HOME and /proc/session/cwd/.nix-profile the root of fixed path to installed fhs progs. Then fill /usr with symlinks into that path. Instant FHS!
<gchristensen>
wat
<gchristensen>
I mean we can make a fhs given some paths
<zybell_>
That idea is to make a FHS *user specific*! That is what the user installed is there in *the ver* the user installed.
<gchristensen>
I'm not sure why we'd want that
keith` has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
<zybell_>
When a user runs a prog and that prog neds say wine in fhs, it will find the version of wine the user has configured.
<zybell_>
and which the user expects to run with.probably.
epl692 has left #nixos ["John 16:33 NIV—“I have told you these things, so that in me you may have peace. In this world you will have trouble. But take heart! I have overcome the world.”"]
<gchristensen>
that isn't how nix is designed to work
<zybell_>
it will combine the nix way as far as possible with the expectations of designers of binary sw.
<justanotheruser>
how do I see which channel a package is installed from?
<zybell_>
it is for running binary only
Sonarpulse has quit [Ping timeout: 240 seconds]
<iqubic>
Is it just me or is the default version of ghc LESS RECENT in the unstable branch than the stable on.
<gchristensen>
iqubic: sometimes that can happen ...
mahalel_ has joined #nixos
<iqubic>
ghc in the 18.03 repo is 8.2.2 and the unstable version is 8.0.2
<zybell_>
justanotheruser:nix why-installed(?2.0 command)should give you a profile.Then you look which channel the profile is at.But actually it makes no sense to ask which channel a package is from. see ^
jackdk has joined #nixos
<Drakonis[m]>
iqubic: no?
jtojnar_ has joined #nixos
<zybell_>
(Iqubics lines I mean)
<Drakonis[m]>
nixpkgs is at the latest
jtojnar has quit [Ping timeout: 256 seconds]
jtojnar_ is now known as jtojnar
warbo has quit [Ping timeout: 255 seconds]
vidbina has quit [Ping timeout: 276 seconds]
fragamus has joined #nixos
dcol has quit [Ping timeout: 256 seconds]
thc202 has quit [Ping timeout: 276 seconds]
<hodapp>
"CUDA driver version is insufficient for CUDA runtime version" ugh, it's always fucking CUDA
<hodapp>
ran into this because I had to use a newer nixpkgs master in order to get something to build...
pie_ has quit [Ping timeout: 264 seconds]
dbmikus_ has joined #nixos
<justanotheruser>
zybell_: well a profile can use multiple channels?
<hodapp>
uhm, getting "warning: unknown setting 'signed-binary-caches'" trying to upgrade to 18.03 followed by it just stalling indefinitely
<gchristensen>
that warning is just a warning
<hodapp>
that's great and all, but why is it just stalling after that and doing nothing?
<hodapp>
CPU isn't being loaded, no I/O appears to be ongoing, no network activity
jperras_ has joined #nixos
jperras has quit [Ping timeout: 276 seconds]
jrolfs__ has joined #nixos
Fare has joined #nixos
<hodapp>
yeah, running it in strace now, it's still seemingly doing nothing
<zybell_>
justanotheruser:especially any version/derivation can be availabe on multiple channels,or installed because of dependency,the channel changed in meantime and because of reproducibility you can not differentiate the possible cases.
<hodapp>
all I did was "nix-channel --add https://nixos.org/channels/nixos-18.03 nixos" and then "nixos-rebuild switch --upgrade".... was there some other magic required for an upgrade to not fail silently?
nuncanada has quit [Quit: Leaving]
<boomshroom>
hodapp: Did you nix-channel --add as root? nixos-rebuild implicitly uses root's nixos channel.
<hodapp>
boomshroom: yes.
<zybell_>
18.03 likes it if the nix command(s) are version 2.0. But normally they fail with fanfare.
<boomshroom>
hodapp: You might have to do a `sudo nix-channel --update` or include `--upgrade` to nixos-rebuild to update the channel.
<hodapp>
I did 'nixos-rebuild switch --upgrade'
iyzsong has joined #nixos
<hodapp>
zybell_: is this nix 2.0 dependency documented anyplace? an upgrade page?
<zybell_>
i know about it from here irc.
<hodapp>
I searched around first to see if there were any special instructions for 18.03
<zybell_>
If you have ALSA sound expect to have to unmute it once from cmdline (gui doesnt work in this case) after update.
jperras_ has quit [Read error: Connection reset by peer]
jperras has joined #nixos
jrolfs__ has quit [Ping timeout: 240 seconds]
astsmtl has quit [Ping timeout: 256 seconds]
astsmtl has joined #nixos
astsmtl has quit [Changing host]
astsmtl has joined #nixos
<hodapp>
yeah, if this upgrade ever decides it's not going to stall indefinitely while doing absolutely nothing, I'll handle that
jperras_ has joined #nixos
jperras has quit [Ping timeout: 264 seconds]
jrolfs__ has joined #nixos
Lnavry has joined #nixos
kreisys has quit [Read error: Connection reset by peer]
Yaniel has quit [Excess Flood]
olto has quit [Ping timeout: 264 seconds]
kreisys has joined #nixos
jrolfs__ has quit [Ping timeout: 264 seconds]
olto has joined #nixos
sigmundv__ has quit [Ping timeout: 240 seconds]
blonkhart has joined #nixos
iyzsong has quit [Ping timeout: 240 seconds]
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
blonkhart is now known as blinkhard
zybell_ has quit [Ping timeout: 264 seconds]
zybell_ has joined #nixos
Supersonic112 has joined #nixos
Supersonic has quit [Disconnected by services]
mbrgm has quit [Ping timeout: 260 seconds]
Supersonic112 is now known as Supersonic
srdqty has quit [Ping timeout: 264 seconds]
mbrgm has joined #nixos
srdqty has joined #nixos
iyzsong has joined #nixos
spear2 has joined #nixos
iqubic` has joined #nixos
iqubic has quit [Ping timeout: 240 seconds]
iqubic` is now known as iqubic
joehh has quit [Ping timeout: 240 seconds]
hamburglar[m] has joined #nixos
jgt has joined #nixos
ryantrinkle has quit [Ping timeout: 276 seconds]
<mightybyte>
Anyone know why the home-manager install doesn't get rustc cached?
jrolfs__ has joined #nixos
Ridout has quit [Quit: Lost terminal]
jrolfs__ has quit [Ping timeout: 256 seconds]
<samueldr>
hodapp: the "stalling" is nix 2 running on a nix 1 daemon, they don't like to talk, but it's (when hung) downloading and compiling as it would
<samueldr>
(this should've probably been added in the 18.03 release notes)
<samueldr>
it will come back after the "usual download and compilation" time (it probably came back since it's been an hour since you asked)
<disasm>
how do I escape ${foo[@]} in a multiline string? e.g. I want the script to have ${foo[@]} in it.
<boomshroom>
disasm: Is foo a list and you want to get an element from it? I don't think nix supports that operator and you'd have to use a builting get function.
<disasm>
boomshroom: foo is a bash array and I want the script to really have ${foo[@]} in it. It's for looping through an array using `for i in ${foo[@]}; do echo $i; done;`
<disasm>
I tried \${foo[@]} but that failed.
<boomshroom>
disasm: Ah, I think it should just be \${foo[@]}
<disasm>
yeah, tried that, no luck
<disasm>
error: syntax error, unexpected '@', at
<disasm>
found it: `''${core_pid[@]}`
<disasm>
the two '' before $ escapes it
chessai has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vpLYI
<{^_^}>
→ 4eb5bdd9 by R. RyanTM: znapzend: 0.17.0 -> 0.18.0
<pirateking>
boomshroom: essentially yes. the end result is making as many databases as i need when i need them, and they need to be accessible simultaneously, and on the same machine
<{^_^}>
→ 4ae08796 by @matthewbauer: Merge pull request #39029 from r-ryantm/auto-update/znapzend
<achambe>
oh
schoppenhauer has joined #nixos
<boomshroom>
pirateking: The database server shouldn't have to move. It probably could be just invoked multiple times with different databases.
johnw has joined #nixos
chisui has joined #nixos
<pirateking>
boomshroom: i'll take a look again. though neo4j is one database per server. maybe i can use a derivation as a thing to clone when i need to breed more graphs
<pirateking>
achambe: thanks i'll take that into consideration
<fearlessKim[m]>
how to debug a "stack overflow (possible infinite recursion)" ? I have no idea where it is
<boomshroom>
pirateking: You can't just install neo4j once and run the server multiple times?
semilattice has joined #nixos
<pirateking>
boomshroom: nope, you have to install a fresh copy of neo4j each time you want a database
<achambe>
lol
semilattice_ has joined #nixos
<achambe>
that sounds wrong
semilattice_ has quit [Remote host closed the connection]
<achambe>
neah, i think the nix containers is the best option from what you said
<achambe>
yeah*
<pirateking>
achambe i'll have to look that up. containers sound nice but i need to get them to work with an app. email graph, projects graph, contacts graph, reference file graph, reminders graph etc
<achambe>
its not like docker containers or anything
<achambe>
its an instance of nixos running in a private area
<achambe>
using the same config
<pirateking>
ok that sounds more appealing
<pirateking>
i'm about to get off for the night, but can i create dupes of containers programmatically as well? if so it sounds really nice for some other experimentation
<achambe>
yeah
iyzsong has quit [Ping timeout: 260 seconds]
<achambe>
there are imperative nixos containers
<achambe>
and declarative ones
<achambe>
both take in a nix config
<achambe>
sort of how mutableUsers work if you are familiar with that
<adisbladis[m]>
achambe: You'll save yourself a lot of grief by using the nix build abstractions
<ryantm>
adisbladis[m]: Until you run into some issue with it: I currently am trying to track down why a buildGoPackage binary stopped getting it's interpreter set.
<achambe>
well, go has reproduceable byte for byte builds now with vgo, but In general I agree.
astiles_ has quit [Remote host closed the connection]
<ryantm>
night
<achambe>
good night, thanks for the help
<adisbladis[m]>
I _think_ that go2nix should already work with vgo
dbmikus_ has quit [Ping timeout: 264 seconds]
odi has joined #nixos
astiles has joined #nixos
jonjitsu has joined #nixos
<adisbladis[m]>
Because the whole `go build` step is manual when generating the expressions. But I'm not sure how GOPATH works with vgo
<adisbladis[m]>
achambe: Have you used vgo yet?
<achambe>
for a few things
<iqubic>
is the nix expression language turning complete?
<elvishjerricco>
iqubic: Yep
<elvishjerricco>
`fix = f: let x = f x; in x` :P
<adisbladis[m]>
achambe: And what's your experience like so far?
<iqubic>
Alright. How does that prove my point?
<jonjitsu>
not sure if this is the right channel, what is the simplest way of setting up my own private nix pkgs on a non nixos distro?
<achambe>
adisbladis[m]: it seems to solve my annoyances with most other tools. being able to verify the hashes in a binary is pretty cool.
<achambe>
unfortnately it is totally broken on my terraform plugin though
<achambe>
maybe go2nix won't be haha.
<adisbladis[m]>
I guess I'll spare a few minutes today to playing with vgo and how it relates to nix :)
<elvishjerricco>
jonjitsu: Delete your nix-channel, then somewhere set `export NIX_PATH="nixpkgs=/path/to/my/nixpkgs:$NIX_PATH"`
<achambe>
adisbladis[m]: if anything, it should be possible to convert a go.mod file to a set of nix expressions
astiles has quit [Ping timeout: 264 seconds]
<iqubic>
It seems a whole lot less useful than cabal2nix
<jonjitsu>
elvishjerricco, will that make unavailable the standard nix packages? I just want to be able to add custom packages to my local ecosystem similar to gentoo overlays
<elvishjerricco>
oh
<elvishjerricco>
then you want nixpkgs overlays :p
<{^_^}>
→ 642c8a8d by @peterhoeg: nixos ddclient: support multiple domains and run via systemd timer
<{^_^}>
→ 740bafa9 by @peterhoeg: Merge pull request #36864 from peterhoeg/f/ddclient
odi has quit [Ping timeout: 264 seconds]
joehh has joined #nixos
jonjitsu has quit [Ping timeout: 264 seconds]
mahalel_ has quit [Ping timeout: 240 seconds]
<zybell_>
boomshroom:try to run SHELL=... sublime
<boomshroom>
zybell_: If I'm running it through the desktop entry, do I need to create a new desktop entry? I can't just set it in a config file somewhere?
<zybell_>
Try first from xterm. it may not work.
Fare has quit [Ping timeout: 248 seconds]
<zybell_>
especially if perl is the shell you want to run
odi has joined #nixos
<zybell_>
boomshroom does it work?
<boomshroom>
zybell_: What? No! I do Ctrl-b and get `[Errno 2] No such file or directory: '/bin/bash'`
<boomshroom>
zybell_: It didn't work.
Ariakenom has quit [Read error: Connection reset by peer]
<zybell_>
then the sledgehammer method: strace -o fileonfsGBfree -f -s 4567 sublime
<adisbladis[m]>
achambe: Well the output hash would depend on compiler version. So fixed-output would not be viable.
<zybell_>
boomshroom `less +/bin/bash fileonfsGBfree` should show you where the setting comes from
<boomshroom>
zybell_: Where does "fileonfsGBfree" come from and what does it mean?
<boomshroom>
Is it refering to the sublime text binary?
<zybell_>
read my line before that line too
fragamus has joined #nixos
ma27 has quit [Ping timeout: 256 seconds]
<zybell_>
boomshroom another way were you could run sublime from the shell you want to start in sublime, if you dont do it already,what I suspect.
hamishmack has quit [Ping timeout: 256 seconds]
jacob_ has joined #nixos
<{^_^}>
[nixpkgs] @avalent opened pull request #39136 → The Storage Performance Development Kit (SPDK) → https://git.io/vpLnW
alex`` has joined #nixos
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #39137 → nixos lcdproc: module for lcdd/lcdproc → https://git.io/vpLn8
<{^_^}>
[nixpkgs] @avalent closed pull request #39136 → The Storage Performance Development Kit (SPDK) → https://git.io/vpLnW
jrolfs__ has joined #nixos
<{^_^}>
[nixpkgs] @avalent opened pull request #39138 → spdk: init at 18.04-pre → https://git.io/vpLnP
freeman42x]NixOS has quit [Ping timeout: 260 seconds]
jensens has joined #nixos
reinzelmann has quit [Ping timeout: 276 seconds]
<Aleksejs>
hi, I still can't run teamviewer12 on 18.03 after upgrade. It says that teamviewerd is not running but I checked it and restarted it like 100 times in different combinations - still the same error
ma27 has joined #nixos
<Aleksejs>
I googled it and tried solutions that are offered - basically they offer to restart daemon which I did
jackdk has quit [Ping timeout: 246 seconds]
<Aleksejs>
when I run teamviewer --daemon status it shows "teamviewerd.service: Can't open PID file /run/teamviewerd.pid (yet?)" however pid file exists: "-rw-r--r-- 1 root root 5 Apr 19 10:52 /run/teamviewerd.pid"
<clever>
Aleksejs: re-run it under strace and see what the error is
<sphalerite>
Aleksejs: check if teamviewerd is running using pgrep -a teamviewerd, and if not remove the pid file (sudo rm /run/teamviewerd.pid) then try again
<{^_^}>
[nixpkgs] @peterhoeg pushed to master « beep: fix for CVE-2018-0492 »: https://git.io/vpLRH
<Aleksejs>
sphalerite: it does; clever: it shows a bunch of "openat(AT_FDCWD, 0x7ffea9623810, O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)"
<clever>
Aleksejs: the 2nd argument should be a filename, and strace should have translated it to a string
freeman42x]NixOS has joined #nixos
<clever>
that error also makes sense if the pid file doesnt exist yet
<clever>
its opening it in read-only mode, so it wont be created
<{^_^}>
[nixpkgs] @peterhoeg pushed to release-18.03 « beep: fix for CVE-2018-0492 »: https://git.io/vpL0v
<Aleksejs>
well when I stop teamviewerd, pid file disappears. When I start it, it appears
xy2_ has quit [Ping timeout: 264 seconds]
<clever>
that sounds entirely normal
reinzelmann has joined #nixos
<Aleksejs>
yes
<Aleksejs>
but I don't understand why teamviewer doesn't see it
<clever>
try stracing that next?
sanscoeur has quit [Ping timeout: 240 seconds]
civodul has joined #nixos
<Aleksejs>
there are some "ioctl(13, VFAT_IOCTL_READDIR_BOTH, 0x224000) = -1 ENOTTY (Inappropriate ioctl for device)"
<Aleksejs>
and also "stat64("/home/aleksejs/.local/share/teamviewer12/dosdevices/c:/TeamViewer/mscoree.dll", 0x33f4b8) = -1 ENOENT (No such file or directory)"
<{^_^}>
→ f6e8ceaf by @poelzi: add tests for multi search
<{^_^}>
→ b712d467 by @poelzi: Allow multiple search experssions in nix search
<{^_^}>
→ 6f907b75 by @edolstra: Merge pull request #2090 from poelzi/multi_search
chisui has joined #nixos
xCuri0 is now known as kuri0
knupfer has quit [Remote host closed the connection]
<neonfuz>
compiling my kernel segfaults while building uim
knupfer has joined #nixos
ma27 has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @xeji opened pull request #39141 → nixos/release-small.nix: disable predictable-interface-names test → https://git.io/vpLz3
<rizary>
is there anyone here develop nodejs on nixos and using hydra? I am planning to make nodejs backend using nixos, but can't find the example of hydra configuration.
asuryawanshi has quit [Remote host closed the connection]
asuryawanshi has joined #nixos
<sphalerite>
coorey[m]: have a backup in case there was anything there you want to keep :)
<jophish>
Ah, Ralith: playing with vulkan?
mojjo has joined #nixos
ma27 has quit [Ping timeout: 248 seconds]
humanoyd has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mojjo>
hi! Does anybody know how to use melpa instead of melpa-stable when building emacs via `(pkgs.emacsPackagesNgGen pkgs.emacs25).emacsWithPackages (epkgs: ...);`
reinzelmann has quit [Quit: Leaving]
<mojjo>
Particularily I'd need the use-package package in a less outdated version (the one on melpa stable is 2 years old..)
<hlavaty>
hi i have a fileSystems."/var/lib/foo" entry in configuration.nix. the disk failed and now on boot the machine goes into rescue mode. how do i disable the disk so that the machine starts normally again and i can ssh in an upload and activate new configuration?
<Akii>
hlavaty: can't you just mount the disk in rescue mode?
Myrl-saki has joined #nixos
<Akii>
or do you only have access over SSH?
<hlavaty>
Akii: no they failed
peacememories has joined #nixos
<hlavaty>
is there no way to disable that systemd device in the rescue mode?
vaninwagen has quit [Ping timeout: 256 seconds]
humanoyd has quit [Ping timeout: 265 seconds]
tmaekawa has quit [Quit: tmaekawa]
<Myrl-saki>
So, uhh
<Myrl-saki>
After the initial "bootstrapping" process, Nix is actually very good even as source-based.
<hlavaty>
Akii: i have physical access to the machine but nixos config is installed via ssh
<hlavaty>
i might just plug in new disks with the same label, hopefully it will boot normally and then start modifying configuration.nix via ssh again
<Akii>
hlavaty: ah I thought you could mount the disk, change the config. But that would require a system rebuild; I had not considered that
<hlavaty>
yeah
tmaekawa has joined #nixos
sigmundv__ has joined #nixos
tmaekawa has quit [Client Quit]
<greska[m]>
Hello, I just downloaded nixos to install it in QUBES os. I am a complete new user to nixos and I cannot get the internet to work for the install. Nixos sees the hardware designated from qubes but I cannot get it to connect. Any ideas?
<{^_^}>
[nixpkgs] @flokli opened pull request #39146 → notmuch: build with gmime2 for now → https://git.io/vpL64
Ariakenom has quit [Ping timeout: 256 seconds]
sanscoeur has joined #nixos
satreu has joined #nixos
periklis has joined #nixos
ciastek has joined #nixos
<Myrl-saki>
error: Nix daemon out of memory
<Myrl-saki>
help
<satreu>
hello guys, i need some advise... currently i am trying to build a package in nixos but i think it is the wrong way of doing it. I want to compile some java-source with ant and then deploy it as a local Website.
<Myrl-saki>
That's for import
<achambe>
Myrl-saki: how much ram in your machine?
<Myrl-saki>
4 GB
<Myrl-saki>
The import is 1.2 GB
<achambe>
ouch
<achambe>
hmm
<achambe>
there are a few issues in the bug tracker about this
<Myrl-saki>
tmpfs 967M 7.0M 960M 1% /run
<Myrl-saki>
Is this the suspect?
<achambe>
why is 1 percent use suspect?
<Myrl-saki>
960M
joehh has joined #nixos
<Myrl-saki>
I increased the size
humanoyd has joined #nixos
<Myrl-saki>
Still says OOM
<Myrl-saki>
Sigh, time to strace
<Myrl-saki>
I have no idea anymore. My RAM usage is okay.
<satreu>
is there some way to define services etc. in nixos?
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Myrl-saki>
Uhhh... is there a memory checker?
<Myrl-saki>
Say, I have 4 GiB of RAM, is there a soft limit for processes?
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #39147 → tox: libtoxcore and assorted tox client updates → https://git.io/vpLXU
<Myrl-saki>
It spiked at around 3 GB, but never reached 4.
jrolfs has joined #nixos
<Akii>
hlavaty: well if all goes wrong you could probably edit fstab manually (if that is still a thing; my linux-fu is rusty)
<Myrl-saki>
I'll never know.
iyzsong has quit [Ping timeout: 256 seconds]
<jophish>
hmm, hardeningDisable = ["bindnow"] doesn't seem to be working for my haskell package
freeman42x]NixOS has quit [Ping timeout: 255 seconds]
ma27 has joined #nixos
<achambe>
adisbladis[m]: vgo list -m spits out a list of urls and a list of import paths
<achambe>
I mean list of import paths and a list of tags/versions
<{^_^}>
[nixpkgs] @thorerik opened pull request #39148 → Update license for VSCode → https://git.io/vpL17
<adisbladis[m]>
achambe: I'm gonna have a quick look at vgo now :)
<jophish>
I'm debugging something with nix-shell, however I can't remember the incantation to get runHook in scope
drkna has joined #nixos
<jophish>
when I run `exec $configurePhase` runHook is missing
<drkna>
hello guys, wanna ask a question. I have tried to install a ruby gem onto NixOS with nixfiles, however, the `gem list` and `gem env` shows it is not linked with the gem path..
<drkna>
anyone knows how to do this?
<achambe>
drkna: have you checked the nixpkgs user manual section on ruby?
<drkna>
yup, i have followed the "packaging ruby" page and created my nixfiles for the gem
<domenkozar>
although that probably refers to original store..
<domenkozar>
store path*
<yorick>
domenkozar: fetchGit actually looks at my .git dir
<domenkozar>
yeah that's the point :)
<domenkozar>
g2g
Wharncliffe has joined #nixos
<hodapp>
etu[m]: no, I do it regularly but I just upgraded to 18.03
vaninwagen has joined #nixos
ma27 has quit [Ping timeout: 256 seconds]
<etu[m]>
hodapp: You didn't keep any old generation? :D
ottidmes has joined #nixos
iyzsong has quit [Ping timeout: 256 seconds]
<octe>
how do i find the location of a package from name?
<octe>
a package without a binary
<hodapp>
hmm... nix-locate?
<hodapp>
or in my case I usually just look directly in /nix
<{^_^}>
[nixpkgs] @DerGuteMoritz closed pull request #33734 → stumpwm: Work around dependency load errors during build → https://git.io/vNOAN
<octe>
yeah i can do a find /nix for the file i want but that seems inefficent
<octe>
might also be multiple version of the package
pie_ has joined #nixos
<ottidmes>
In a final attempt to use my GPU with a later kernel than 4.4 (the latest kernel supported by the closed source driver), I managed to get it working under kernel 4.15 with the modesetting driver. Is modesetting a valid choice as a day-to-day driver?
knupfer has joined #nixos
<Dezgeg>
yes
<ottidmes>
octe: you could just ask nix to give you the path if you have the name
<octe>
how?
<Dezgeg>
I think modesetting nowadays also replaces the "intel" videoDriver but not totally sure about that one
<hodapp>
...but the CURL line is just " 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0"
<hodapp>
so it's just sitting there
<srhb>
hodapp: Downloads for me. 1.5 GiB though :|
<ottidmes>
Dezgeg: Ah, good to know. It was so annoying to be stuck at 4.4. Sure 3d preformance is terrible right now, but I don't game on NixOS and video rendering and my 4k display work just as well as with the proprietary driver
<jophish>
Aha, it was the dynamic executable which got me
<octe>
ottidmes, what's the command for that?
<hodapp>
if I download it with CURL it produces a 0 byte file
<hodapp>
if I do it from the browner it's fine
<srhb>
hodapp: Yeah, nvidia is doing something weird...
<tilpner>
octe - nix-build --no-out-link "<nixpkgs> -A hello
<srhb>
hodapp: Add it manually
<tilpner>
octe - Goes together well with an alias
<hodapp>
it's *always* fucking CUDA that is my problem
<etu[m]>
hodapp: Redirects?
<tilpner>
octe - * nix-build --no-out-link "<nixpkgs>" -A hello
<etu[m]>
Yeah, they do a redirect
<etu[m]>
Curl doesn't follow redirects by default
<ottidmes>
octe: What tilpner said would do what you want (I was trying something with nix-instantiate for you)
<srhb>
etu[m]: fetchurl does though..
<hodapp>
so... nixos 18.03 just shipped with a broken derivation for CUDA?
<hodapp>
seems unlikely.
<srhb>
hodapp: It appears so.
rauno has quit [Remote host closed the connection]
<srhb>
hodapp: Or they changed something recently. We can check on Hydra.
<etu[m]>
srhb: Yeah, it should... Well that's odd :(
<Dezgeg>
it probably doesn't build cuda as it's unfree
<hodapp>
all I know is that I want all things CUDA-related to go die in a fire
<srhb>
I sympathize :P
<srhb>
I don't get it though. If I use the curl parameters that fetchurl does it looks like it's working just fine. Wat.
acarrico has joined #nixos
<LnL>
ottidmes: yes, it's not redistributable so there's no way to download it automatically
seanparsons has quit [Read error: Connection reset by peer]
<srhb>
hodapp: wait.. I think this was all false alarm
<srhb>
hodapp: The curl indicator may be broken though.
<srhb>
hodapp: We just assumed the fetchurl was broken because a plain curl doesn't redirect. But fetchurl does.
<srhb>
hodapp: More patience would have shown that it just works
<ottidmes>
LnL: You mean the JDK? I assumed as much, something to do with licensing. I don't mind it, my biggest annoyance with the current workaround though, is that it will error the whole build. Would be better to ask at the end when the rest has been built, even better if it was interactive, so it could just continue once I have done adding it to the store
<hodapp>
srhb: yeah, somehow I wasn't going to wait for an hour for it to do something while it gave no indication whatsoever
<LnL>
you can use -k to continue building stuff if there's a failure
pirateking has joined #nixos
<Myrl-saki>
nix-shell "$(nix-instantiate '<nixpkgs>' -A ghc)" --option substituters ''
<Myrl-saki>
nix-shell -p ghc
<Myrl-saki>
Have different results
<ottidmes>
LnL: Cool, then I will add that flag in future nixos-rebuild calls! :D
jtojnar_ has joined #nixos
<WilliamHamilton>
hi, could someone suggest a correction of http://lpaste.net/4076259757332103168 ? It's intended as an overlay to install an additional haskell package
vandenoever has quit [Ping timeout: 256 seconds]
<Myrl-saki>
Oh. Different derivers.
<Myrl-saki>
Well, this makes me even more confused.
<ottidmes>
Myrl-saki: nix-shell <derivation> and nix-shell --packages <pkgname> have different goals / semantics. It should be described in the manpage of nix-shell
<Myrl-saki>
ottidmes: I nix-copy-closured from a different /nix/store.
<Myrl-saki>
ottidmes: I copied both the deriver and the output.
<WilliamHamilton>
but I get "error: attribute 'constraints-emerge' in selection path 'nixos.haskellPackages.constraints-emerge' not found"
<ottidmes>
WilliamHamilton: Try dropping the nixos. prefix
<Myrl-saki>
I have it in my nix store, but the path is not valid..
Wharncliffe has quit [Ping timeout: 256 seconds]
<ottidmes>
Myrl-saki: Your first command basically builds a environment to build ghc, while the second creates a shell environment with ghc available
<pirateking>
there's a concept in business that's been bugging me about searching for packages in linux - "performance based hiring". when there are so many options, you don't want to spend forever trying everything out if you need to have a job done by someone else (in this case, software and tools). databases can make this process a breeze to query and get a good decision and start on something instantly for any problem you need to solve
<pirateking>
is there any way to query packages by date created, date added to nix, date updated, date updated by nix, whether they build, and number of stars on github?
<Myrl-saki>
ottidmes: Oh.
<Myrl-saki>
ottidmes: Ugh, that makes a lot of sense.
knupfer has quit [Ping timeout: 264 seconds]
<Myrl-saki>
ottidmes: How do I nix-shell the output?
<srhb>
WilliamHamilton: You probably also need some executableHaskellDepends
<ottidmes>
Myrl-saki: Not sure what you mean with your question, could you eloborate?
pie_ has quit [Ping timeout: 240 seconds]
<Myrl-saki>
ottidmes: I want a nix-shell with the output available.
<pirateking>
when i do `sudo nix-env -qa --json -sPc --description --meta -A nixos.evilwm`, what I'm wanting to know is how much it's helped people. i get no answers
<Myrl-saki>
ottidmes: The output, being the drv's output.
<pirateking>
oh whoops you weren't talking to me :P
MP2E has quit [Remote host closed the connection]
hlavaty has quit [Remote host closed the connection]
<Myrl-saki>
ottidmes: nix-shell /nix/store/gpx67mf293442444hcvasmsws7a8agdr-ghc-8.2.2.drv -A out
<Myrl-saki>
Seems to work?
<ottidmes>
Myrl-saki: Ah, this seems to work: nix-shell -p $(nix-instantiate --eval --expr 'with import <nixpkgs> {}; toString hello')
<Myrl-saki>
Ohhh
<Myrl-saki>
You can -p an output?
<ottidmes>
Myrl-saki: Yep
<pirateking>
to make it easier to come to an answer here, i'll change my question. does anyone know of any discussions on package fields and data, or if it's not a job for nix, then what might be used that i can take output from nix to query something that has this extra meta data?
<pirateking>
and vice versa
alex`` has joined #nixos
Rusty1_ has joined #nixos
vandenoever has joined #nixos
<ottidmes>
Myrl-saki: Seems to be undocumented though, but yeah it works. I just tried with ghc and it returns me the nix store path rather than the /run/current-system/sw/bin one
<srhb>
pirateking: It's really difficult to do with Nix in the general case, aside from using git, because there's no 1-1 file-package correspondence
<WilliamHamilton>
srhb: thanks, that works indeed; what threw me off was the invocation: I was under the impression that the overlay system picked the overlays automatically, when I put them under the `~/.config/nixpkgs/overlays/` folder (and it worked in the past, I think). So I was trying to install it with `nix-env -iA nixos.haskellPackages.constraints-emerge`
<Myrl-saki>
ottidmes: Btw, I believe they're semantically different too.
<srhb>
WilliamHamilton: Let me find the whole default location shenanigans for you...
<WilliamHamilton>
srhb: do you have any info on automatic overlays?
<pirateking>
srhb: thanks
<Myrl-saki>
ottidmes: I mean, semantically the same, but executionally different?
<Myrl-saki>
ottidmes: -p requires loading everything into memory or something of that sort. "Dumping very large path..."
<srhb>
WilliamHamilton: nixpkgs/pkgs/top-level/impure.nix -- search for overlays
<ottidmes>
Myrl-saki: Who are they? out path vs pkgname?
<WilliamHamilton>
srhb: you are a saint :)
<Myrl-saki>
ottidmes: Err, wait, give me a few.
<srhb>
WilliamHamilton: It's... Too complicated in my opinion. I prefer just calling functions directly :P
<Myrl-saki>
ottidmes: Vs nix-shell -p haskell.compiler.ghc822
<WilliamHamilton>
srhb: I checked what I was missing: I didn't add the `.nix` extension to the file! now it works even in the other way
<srhb>
WilliamHamilton: Great :)
<manveru>
pirateking: so kinda like a nix package store with reviews and such?
<WilliamHamilton>
srhb: oh but the overlay system is wonderful: if you just put that file in the ~/.config/nixpkgs/overlays/` folder, then it becomes part of <nixpkgs> and you can issue simpler commands!
<srhb>
WilliamHamilton: Yes, but.. Say you also had overlays in other places, or you used nix.overlays, or nixPath = [ "overlays=..." ];
<WilliamHamilton>
srhb: thanks for the help, as always :)
<srhb>
WilliamHamilton: I have no idea which ones take effect and which don't
<pirateking>
manveru: almost. my purpose is similar to that of reviews. but more basic. just more meta data. i can use my own intuition and wisdom to drive my decisions. ..i find myself hunting for the same kind of information everytime just to make a decision. i'd rather not do this manual work
<Myrl-saki>
ottidmes: Of course, I get proven wrong once again. I'm guessing something went wrong with the nix copy.
<srhb>
tilpner: Right, and so does impure.nix.. But I have to look it up each time. Instead I just invoke nixpkgs manually and add the overlays explicitly. :)
<srhb>
It's a bit of duplication, but it's immediately apparent which overlays are in effect.
<tilpner>
My <nixpkgs> is a file that adds the overlays
<srhb>
tilpner: Yeah, that's essentially what I do. :)
<Myrl-saki>
Is there a way to nix-copy-closure with SSH key?
<srhb>
Myrl-saki: It always does, but I assume you mean with your user's key instead of root?
<Myrl-saki>
srhb: Yep.
<manveru>
pirateking: atm i think it's just a simple json file generated by nix. in theory it might be possible to figure out for each package what .nix files constitute it, and when they were last updated. of course it already includes URLs for github and such, so you could get stars from there
<srhb>
Myrl-saki: Perhaps you can use NIX_SSHOPTS?
<srhb>
Myrl-saki: Or maybe nix copy has a better option.
<Myrl-saki>
srhb: Thanks, where's that documented?
<srhb>
Myrl-saki: NIX_SSHOPTS? man nix-copy-closure
<pirateking>
manveru: yeah it looks like i'll have to make it myself. i wanted to make sure i wasn't spending time considering things that others have already considered. or worse, started a project already half done. but it looks like it hasn't been done yet
<Myrl-saki>
srhb: Thanks. My / failed me.
<srhb>
Myrl-saki: I tend to just be a trusted user on each of the machines though.
lohfu has joined #nixos
<srhb>
Myrl-saki: I think that makes any root access unnecessary.
<manveru>
pirateking: there was one other project that checked what packages are up to date, and also one to update automatically (which works quite well now). but that doesn't seem very helpful for your case?
<srhb>
Myrl-saki: (That is, `nix-copy-closure --to sarah@foo path` just works)
<pirateking>
it might be. do you think you could dig up any one of those or both?
<jophish>
Is it possible to override rust packages with buildRustPackage?
<jophish>
cargo isn't able to find the version of a package I need
<pirateking>
manveru: beautiful - thank you so much!
<manveru>
we definitely could use something a bit... more usable for normal users :)
<pirateking>
mhm :)
<seanparsons>
Does anyone here work on projects using npm with bundled binary dependencies? I've got something that in this case bundles 7za, tries to run it and it pops instantly because patchelf hasn't been run on it. I'm wondering if there's a sensible way to handle these cases.
lo_mlatu has quit [Ping timeout: 260 seconds]
rauno has joined #nixos
<manveru>
seanparsons: yarn2nix can handle such cases
<manveru>
though it might require some fiddling still
<Myrl-saki>
srhb: Still requires an SSH key though, right?
<srhb>
Myrl-saki: Yes, but just for the calling user. if ssh foo works, that should too.
<seanparsons>
manveru: Hmmm.
<Myrl-saki>
srhb: Makes sense. I forgot how to use SSH keys without ssh-agent though.
joehh has quit [Ping timeout: 248 seconds]
<Myrl-saki>
srhb: under config, I believe?
<srhb>
Myrl-saki: I don't think you need to do anything? As long as the key is present in the default location
<seanparsons>
Cheers for that though, I'll need to have a bit of an investigation.
<srhb>
eg. ~/.ssh/id_rsa
<srhb>
Or whatever. :P
<manveru>
seanparsons: it works via the pkgConfig attribute, i'm at work atm so can't provide a ton of info :|
jonjitsu has joined #nixos
<manveru>
but that should be enough to get you some examples
alex`` has quit [Ping timeout: 248 seconds]
darlan has joined #nixos
darlan has quit [Client Quit]
<Myrl-saki>
srhb: I also have that, yeah.
<srhb>
Myrl-saki: Hm, not sure what else is required. I didn't think anything.
<Myrl-saki>
srhb: I think you have ssh-agent?
<Myrl-saki>
srhb: I mean, there's nothing wrong with having it. Lmao
<srhb>
Myrl-saki: I do, but I'm pretty sure it worked without as well.
<srhb>
I might be mistaken though
<Myrl-saki>
srhb: Host 192.168.2.11
<Myrl-saki>
IdentityFile ~/.ssh/lappy_rsa
<Myrl-saki>
;
<Myrl-saki>
srhb: That's in my ~/.ssh/config
<Myrl-saki>
srhb: I believe that is required, but I may be wrong.
<srhb>
It probably is if you're using a nonstandard name like that.
pkill9 has joined #nixos
lord| has quit [Quit: WeeChat 2.1]
<Myrl-saki>
srhb: True, makes sense, but I believe it's standard practice to have a unique key per machine?
<srhb>
Myrl-saki: That... Sounds cumbersome.
<srhb>
I mean, sure, more security is more better.
<Myrl-saki>
srhb: Haha, true. :P
<Myrl-saki>
srhb: I'm not so fond of going full security, I just assumed that everyone else does it that way. I guess I was mistaken.
<srhb>
I had just one private key for the longest time, now I have a few divided by clients really.
<Myrl-saki>
srhb: And yes, it doesn't only sound cumbersome, it is cumbersome.
<Myrl-saki>
srhb: I also have some other system quirks, such as having a static IP for my wlan.
<Myrl-saki>
srhb: So if I'm in a different place, I have to `systemctl disable dhcpcd` then manually start a new dhcpcd.
<srhb>
Fun!
<Myrl-saki>
srhb: TBF, my laptop is mostly a build machine, so it's not too bad.
<Myrl-saki>
Advantages of being a shut-in. :D
tertle||eltret has quit [Quit: Connection closed for inactivity]
<srhb>
Still waiting for those new thinkpads coming this summer (I hope...,) then I can claim the same..
peacememories has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Myrl-saki>
srhb: What's new about them? Coffee lake?
<srhb>
Myrl-saki: That and maybe a numpad-free 15 inch.
<srhb>
Anymoo. We should probably not veer into that topic because I can talk about my crippling indecisiveness on this topic for... Well, a long time. Ostensibly I've been buying a new Thinkpad for three years.
<Myrl-saki>
srhb: How about the XPS 15?
<srhb>
:-)
jperras has joined #nixos
<sphalerite>
I got an XPS 15 just about a year ago, and I almost regret it a bit. It's a lovely machine, but my chromebook made me realise it's overpowered (and overpriced) for most of what I do
<{^_^}>
[nixpkgs] @dtzWill opened pull request #39150 → rw: init at 1.0 → https://git.io/vpLNJ
<srhb>
I like having a real build machine.
<srhb>
If I had a ton of money I'd set up a real PC at home to act as a remote nix builder and buy a lighter laptop, but...
humanoyd has quit [Quit: WeeChat 2.1]
<sphalerite>
that sounds like a cheaper option than getting a powerful laptop
<srhb>
Well.. Yeah, it might be, really. But also gaming in bed! Important things!
<sphalerite>
hahaha
<srhb>
(An eGPU could probably take care of that...)
<Myrl-saki>
I wanna have a build farm of Threadrippers.
<etu[m]>
Why? They was terrible :D
leat has quit [Ping timeout: 268 seconds]
<sphalerite>
srhb: laptop + streaming :p
<srhb>
sphalerite: I've considered it. My demands might be low enough that the latency is fine
<etu[m]>
A job I had was to build Threadripper clusters for universities. We build 400 blade servers in a week or so by hand and ran hardware tests and such to make sure they worked.
<sphalerite>
that sounds kind of cool.
<{^_^}>
[nix] @edolstra pushed 2 commits to 2.0-maintenance: https://git.io/vpLNb
<{^_^}>
→ 497325a7 by @edolstra: Remove some release-critical jobs
<{^_^}>
→ f7454db4 by @edolstra: Remove Fedora / Ubuntu releases that are no longer in Nixpkgs
<etu[m]>
And later on for my next job I've bought one server like that from the very same company for a quite good price. We had 4 Threadrippers, 12 cores each and 64GiB of RAM. I compiled the kernel in less than 60 seconds which was quite cool to be (running Gentoo at the time).
<etu[m]>
cool to me*
<gchristensen>
I like to buy my powerful hardware by the hour
<gchristensen>
then when it goes bad I can just hand it off
<talyz>
etu[m]: weren't those bulldozers?
Lears has quit [Remote host closed the connection]
<Myrl-saki>
...
<Myrl-saki>
I have 16 GB of RAM
<etu[m]>
oh, yes. You're right.... I'm confusing things now.
fuzzy-id has joined #nixos
<Myrl-saki>
And the nix-daemon says OOM.
<Myrl-saki>
Does nix-daemon require swap? lol
<fuzzy-id>
i try to run terraform, but it says it cannot find the aws provider although i installed terraform-full
<etu[m]>
gchristensen: We needed to run maya all the time, every day, all of the days, all of the hours. And maya licenses are expensive. So it was cheaper to pack as much power as possible in one box.
<gchristensen>
yeah no doubt
Fare has joined #nixos
<Myrl-saki>
This is so weird. This happened twice now.
<etu[m]>
gchristensen: Did you have a link where it was possible to buy nixos stickers btw? Can't find it in my history :(
<Myrl-saki>
Or maybe I have o remount, who knows.
<gchristensen>
it is pricey, but 16 cores, 128g ram, 2.4tb nvme, and 2x10gbps nics for $1.75/hr :)
periklis has quit [Ping timeout: 256 seconds]
odi_ has quit [Ping timeout: 256 seconds]
<gchristensen>
etu[m]: ping me in a few hours? I'm writing slides for a talk in an hour
<steveeJ>
hey, I just stumbled upon an issue with latexindent which is already reported yet unsolved: https://github.com/NixOS/nixpkgs/issues/35537 any chance someone has an idea why this happens?
<magnetophon>
During nixos-rebuild I keep getting "Cannot write: No space left on device", while "df" says I have 13Gb free on / How do I troubleshoot further?
leat has joined #nixos
<steveeJ>
magnetophon: which filesystem are you using?
<fuzzy-id>
magnetophon: is this during installation?
<sphalerite>
magnetophon: is your /boot maybe out of space?
<sphalerite>
hahaha onslaught of questions
jonjitsu has quit [Ping timeout: 248 seconds]
aarvar has joined #nixos
<magnetophon>
steveeJ: ext4
<steveeJ>
magnetophon: what does df -i say?
<unacceptable>
magnetophon: did you recently delete something big?
<magnetophon>
fuzzy-id: df doesn\t mention a separate boot
ihar has quit [Quit: WeeChat 2.0]
<fuzzy-id>
my question was actually whether you are currently installing ;)
<magnetophon>
steveeJ: df -i says I have 3.3mB free!
<magnetophon>
unacceptable: that isn\t used by nixos-rebuild, is it?
<unacceptable>
No idea, it's just your only entry that's even remotely full
<magnetophon>
steveeJ: I gues M doesn't stand for MB in this context...:)
<steveeJ>
magnetophon: alright, in that case my suspicion of no free inodes is disproved
<symphorien>
magnetophon: with ext4 a fraction of the fs is reserved for use by root. ~10% by default. other users get ENOSPACE
<symphorien>
sorry more 5%
<symphorien>
this might be it
<magnetophon>
symphorien: df says I have 20% free
<symphorien>
hum misread sorry
<magnetophon>
sphalerite: it's trying to unpack a big font pkg, so that rules out /boot issues, doesn't it?
<sphalerite>
yeah
Ariakenom has joined #nixos
<sphalerite>
magnetophon: OH! It's probably your /tmp
<sphalerite>
try unmounting it and redoing it
<sphalerite>
iirc stuff gets built in /tmp
<fuzzy-id>
yeah, 4G is too less for /tmp on nixos!
Lears has quit [Remote host closed the connection]
<magnetophon>
umount: /tmp: target is busy.
<unacceptable>
lsof
<Myrl-saki>
There's really something wrong.
<magnetophon>
so I gues I have to change my config to use regular disk instead of tmpfs for tmp, rebuild without the font, and then try again, right?
<sphalerite>
unacceptable: yeah or just make it a bigger tmpfs, adding swap if necessary
<zybell_>
How much RAM has that machine?
mojjo has quit [Ping timeout: 276 seconds]
<ottidmes>
srhb: If you ever want to try the desktop streaming to laptop approach, this works pretty well: https://parsecgaming.com/
<srhb>
ottidmes: Thanks :)
<fuzzy-id>
magnetophon: you could also try 'umount -l /tmp' and rebuild. i'm not sure how exactly it behaves though.
<magnetophon>
zybell_: 8G
<zybell_>
you have over 12G committed to various tmps
Ariakenom has quit [Ping timeout: 265 seconds]
<magnetophon>
so in general tmpfs for /tmp on a 8G nixos machine is a bad idea, right?
<magnetophon>
zybell_: 12G? how do you figure?
alaricsp has quit [Ping timeout: 260 seconds]
coot has quit [Quit: coot]
<magnetophon>
fuzzy-id: umount -l /tmp gives no errors, rebuilding now
<magnetophon>
and it finished rebuilding. thanks everybody!
coot has joined #nixos
alexteves_ has joined #nixos
<magnetophon>
I commented out # tmpOnTmpfs = true; and will rebuild again. If I understood fuzzy-id correctly, I should leave it that way, right?
<srhb>
magnetophon: fwiw I regularly run out of memory while building with 16GiB and zfs.
<srhb>
I ran out much more often with tmpOnTmpfs
<ixxie>
So, it seems my XPS 9350 sometimes randomly switches its SATA BIOS setting from disabled to RAID or ADHC. I was wondering if there is any chance whatsoever NixOS is somehow doing that, since googling this issue yielded nothing.....
spear2 has quit [Ping timeout: 260 seconds]
<srhb>
Usually I just lower the amount of builders when stuff like that happens
<zybell_>
that means normal RAM use (compiler) will cut into available space on tmps. /tmp + /run/wrappers + /dev/shm + /run look at Size
<magnetophon>
srhb: thanks, I sometimes go into swap when building.
<{^_^}>
[nixpkgs] @shlevy pushed to master « rdkafka: Build with openssl support. »: https://git.io/vpLp9
<vaibhavsagar>
silly Nix question: I'm getting the "called without required argument 'mkDerivation'" error because I'm expecting callPackage to take care of it
fragamus has joined #nixos
<srhb>
vaibhavsagar: Is it in scope?
<vaibhavsagar>
and if I pass an argument for mkDerivation, then the derivation complains about the next input
<vaibhavsagar>
is there something else I should be doing instead?
alaricsp has joined #nixos
odi_ has joined #nixos
<magnetophon>
sphalerite: and thank you especially for the correct answer!
MichaelRaskin has joined #nixos
<sphalerite>
\o/
<ottidmes>
vaibhavsagar: Should it not be stdenv.mkDerivation? Not sure if mkDerivation even exists as a separate top-level attribute
<sphalerite>
oh hey MichaelRaskin
aarvar has quit [Ping timeout: 255 seconds]
<sphalerite>
MichaelRaskin: I worked out the sauerbraten thing. It turns out it does some nasty hacks that actually break the glibc headers, and it is sauerbraten's fault, not glibc's or gcc's
aarvar has joined #nixos
<vaibhavsagar>
ah, I see
<{^_^}>
[nixpkgs] @shlevy pushed to release-18.03 « rdkafka: Build with openssl support. »: https://git.io/vpLh8
<derchris>
i tried to convert an pypi package to nixos using pypi2nix. there are a few dependencies for the package, one of which is pillow. but when trying to build pillow it failed with missing zlib headers
<unacceptable>
derchris: did you provide libzip into that env?
<derchris>
tried -E zlib or a nix-shell -p zlib to package it without success
<derchris>
i thought it was zlib
<sphalerite>
it is
<derchris>
will try wiht libzip
maingo has quit [Quit: Connection closed for inactivity]
<ij>
Is it posible to get a working shell for a node project with nixpkgs/npm2nix?
<{^_^}>
[nix] @edolstra pushed 2 commits to 2.0-maintenance: https://git.io/vptkA
<{^_^}>
→ e30bd350 by @masaeedu: Wrap thread local in function for Cygwin
<{^_^}>
→ c67476fb by @edolstra: Bump version to 2.0.1
pmeunier has quit [Client Quit]
pemeunier has joined #nixos
kreetx has joined #nixos
<kreetx>
hi! question: why is NIX_PATH in .profile and not in .bashrc? There is no nix in path when logging in via ssh, since it appears not to start a login shell
<betaboon>
hello #nixos, i am deploying containers into a virtualbox using nixops with nixos18.03 and recently i am (more often than not) running into a problem "Cannot find device host0", which i guess is the step that is trying to rename host0 to eth0, any thoughts on that ?
lo_mlatu has joined #nixos
humanoyd has joined #nixos
dbmikus has quit [Quit: WeeChat 2.1]
dbmikus has joined #nixos
<{^_^}>
[nixpkgs] @wmertens pushed to zramSwap-one-dev « zramSwap: remove mentions of old kernels »: https://git.io/vptqj
Lears has quit [Remote host closed the connection]
Lears has joined #nixos
jperras has joined #nixos
kreisys has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @taku0 opened pull request #39157 → oraclejdk: 8u161, 8u162 -> 8u171, 8u172 (backporting to 18.03) → https://git.io/vptm2
asuryawanshi has joined #nixos
<srk>
is there a nix-copy-closure for copying closures locally? I need to copy a container closure to zfs dataset
orivej has quit [Ping timeout: 248 seconds]
<srk>
hmm, nix copy maybe
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jperras_ has joined #nixos
jperras has quit [Ping timeout: 240 seconds]
Fare has quit [Ping timeout: 240 seconds]
lexspring has joined #nixos
jon_ has joined #nixos
<{^_^}>
[nixpkgs] @catern opened pull request #39158 → pythonPackages.dataclasses: init at 0.5 → https://git.io/vptYw
alaricsp has quit [Ping timeout: 265 seconds]
jonjitsu has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @taku0 opened pull request #39159 → oraclejdk: 8u161, 8u162 -> 8u171, 8u172 (backporting to 17.09) → https://git.io/vptYM
sanscoeur has quit [Remote host closed the connection]
sphalerite has joined #nixos
<sphalerite>
argh weechat is being janky, didn't notice I didn't auto-rejoin here >_>
orivej has joined #nixos
kreisys has joined #nixos
<betaboon>
wb then :)
Lears has quit [Remote host closed the connection]
__Sander__ has quit [Quit: Konversation terminated!]
<lexspring>
hello, i'm trying to get simple shell script to work : "nix-shell ./makeFHSchroot1.nix ; cd /somedir" ; but cd part don't happen after nix-shell
<lexspring>
is there a way to get this work, or tell nix-shell to execute particular shell commands ?
knupfer has quit [Read error: Connection reset by peer]
<betaboon>
lexspring: did you try "nix-shell --run" ?
blankhart has quit [Read error: Connection reset by peer]
jensens has quit [Ping timeout: 248 seconds]
<frostspanner>
How do you specify a different overlay path from the default in configuration.nix?
blankhart has joined #nixos
<betaboon>
lexspring: can you share the content of that nix-expression ?
alaricsp has joined #nixos
<lexspring>
betaboon, contents of my fhs.nix : https://hastebin.com/epubajiyis.pl , i tried to replace runScript = "bash"; with runScript = "cd /tmp"; or runScript = "nano"; but it didn't seem to execute those
<iqubic>
Are there any fonts out there that give me Braille characters?
Fare has quit [Ping timeout: 240 seconds]
<betaboon>
lexspring: i dunno what you're trying to acchieve there.
<betaboon>
and i never used buildFHSUserEnv, so dunno on that part
<Ralith>
jophish: on and off
astiles has joined #nixos
<iqubic>
What's the right font set to install if I want noto-sans?
<lexspring>
here is what I trying to achieve : I would like to execute my pkgs.buildFHSUserEnv function as usual, which works for me. and it drops me to a shell. next thing I want to automate and exececute particular shell script commands. which doesn't work for me now and i'm unsure how to perform that. I would like to perform both operations in one step\script\command\line
coot has joined #nixos
panther420[m] has joined #nixos
<lo_mlatu>
Hi all, is there a way to 'pin' a package without installing it?
<lo_mlatu>
I mean, let garbage collect ignore the package
vaninwagen has quit [Ping timeout: 256 seconds]
tgunb has joined #nixos
astiles has quit [Remote host closed the connection]
<tgunb>
is nixos.org down?
astiles has joined #nixos
<Akii>
no
jrolfs_ has quit [Ping timeout: 248 seconds]
rauno has quit [Remote host closed the connection]
<tilpner>
lo_mlatu - nix-store --add-root perhaps
alaricsp has quit [Ping timeout: 256 seconds]
<{^_^}>
[nixpkgs] @dtzWill opened pull request #39160 → tree-wide: fix instance of "meta.maintainer" -> "meta.maintainers" → https://git.io/vptZ4
<clever>
tilpner: nix-build creates such roots by defaul
<clever>
t
<tilpner>
Right, I sometimes forget that because I aliased --no-out-link everywhere
<gchristensen>
dtz: is eval broken?
talyz has quit [Read error: Connection reset by peer]
etu has quit [Read error: Connection reset by peer]
<{^_^}>
[nix] @shlevy pushed to 2.0-maintenance « ssh-ng: Don't forward options to the daemon. »: https://git.io/vptZK
astiles has quit [Ping timeout: 265 seconds]
<betaboon>
lexspring: sorry can't help with that
xcmw has joined #nixos
<dtz>
well, I was seeing check-meta failures --but AFAICT this isn't anything new? So if eval worked yesterday, AFAIK still works :P.
talyz has joined #nixos
<dtz>
(although having something test everything with check-meta enabled does seem like a good idea...?)
coot has quit [Ping timeout: 256 seconds]
etu has joined #nixos
<lo_mlatu>
tilpner looks like what I need, thank you
civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
<tilpner>
lo_mlatu - nix-build '<nixpkgs>' -A hello should work too
<tilpner>
That will create a result symlink. As long as that link exists there, hello will not be gc'ed
ryanartecona has quit [Quit: ryanartecona]
jrolfs_ has joined #nixos
<lo_mlatu>
tilpner: yes I know that, I just don't want the outlink :)
<lo_mlatu>
Oh I see, texlive-combined-full is the name attribute of texlive.combined.scheme-full
astiles has joined #nixos
hotfuzz has quit [Quit: WeeChat 2.0]
<dtz>
hmm, might be hidden due to using callPackage instead of callPackages or something, at least the ones I'm looking at seem different enough in terms of organization that maybe they're not found
jrolfs has quit [Ping timeout: 265 seconds]
nick_l has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @dotlambda pushed commit from @bjornfor to release-17.09 « wildmidi: 0.3.9 -> 0.4.2 »: https://git.io/vptcP
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Fare has joined #nixos
<tgunb>
MichaelRaskin: did you mean nixos.libreoffice?
ryanartecona has joined #nixos
<tgunb>
i updated the channel and did nix-env -iA nixos.libreoffice right before my answer.
jon_ has quit [Remote host closed the connection]
TonyTheL1on has quit [Quit: leaving]
<MichaelRaskin>
Maybe the channel is too far back, though.
TonyTheLion has joined #nixos
jon_ has joined #nixos
vandenoever has joined #nixos
vandenoever has quit [Changing host]
vandenoever has joined #nixos
<MichaelRaskin>
I remember nixos-unstable being stuck for a couple of weeks recently…
fragamus has joined #nixos
<MichaelRaskin>
I mean nixos.libreoffice.libreoffice when nixos channel catches up
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mkaito has quit [Read error: Connection reset by peer]
mkaito has joined #nixos
mkaito has quit [Changing host]
mkaito has joined #nixos
Guanin has quit [Remote host closed the connection]
jluttine has quit [Ping timeout: 248 seconds]
Fare has quit [Ping timeout: 268 seconds]
blankhart has quit [Ping timeout: 264 seconds]
<Myrl-saki>
clever: The binfmt qemu takes 2 seconds to load up.
<Myrl-saki>
clever: Got any idea why?
<Myrl-saki>
I tried the qemu package, and it's instant for me
<sphalerite>
Myrl-saki: so if you call qemu-user explicitly it's instant but if you call it through binfmt_misc it's slow?
<Myrl-saki>
sphalerite: The static(?) qemu-user takes 2 seconds, the nixos package is instant.
<sphalerite>
what do you mean by the nixos package/
<Myrl-saki>
sphalerite: The one from nixpkgs.
<sphalerite>
but that's something completely different, isn't it? That's qemu-system?
<Myrl-saki>
sphalerite: I mean
<Myrl-saki>
sphalerite: I think clever uses a different source for the qemu, since the versions are different.
<adelbertc>
has anyone seen issues like `Error relocating /nix/store/w81d7rbysi0rmf9c47njqbxjnllqv9cq-python-2.7.14/lib/libpython2.7.so.1.0: __vfprintf_chk: symbol not found` ? im hitting this running some python code in a Nix-built/provisioned Docker image
<Myrl-saki>
I'll try updating it right now.
<acowley>
Is putting an `overlays = [ (import ./blah.nix) ];` in config.nix something that is expected to work? As in, the overlay definitions become part of nixpkgs.
ryanartecona has quit [Quit: ryanartecona]
<sphalerite>
adelbertc: apparently it has something to do with FORTIFY_SOURCE. I don't know any details though
<sphalerite>
acowley: no. Overlays are a layer above config
<acowley>
sphalerite: Drats. Okay, thanks.
<sphalerite>
I don't know why that is…
<acowley>
I had wanted to do some pathExists checking to control which overlays are imported
<sphalerite>
you could do that in the overlays themselves.
<sphalerite>
or make one overlay which combines them conditionally
tgunb has quit [Quit: Leaving.]
blankhart has joined #nixos
nuncanada has joined #nixos
doublehp has joined #nixos
wangoe has joined #nixos
ThatOtherGuy has quit [Ping timeout: 276 seconds]
warbo has joined #nixos
warbo has left #nixos [#nixos]
<doublehp>
fresh debian; "nix-env -i freecad" returned OOM; I have 2GB of RAM (no swap); how much swap should I add ? will really help or is it likely i have a bug ?
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vptaB
<{^_^}>
→ 757f026f by @matthewbauer: Merge pull request #38726 from r-ryantm/auto-update/hwdata
<{^_^}>
→ 8564ceac by R. RyanTM: hwdata: 0.310 -> 0.311
<mightybyte>
Anyone know why `nix-env -i ripgrep --dry-run` isn't getting a cached version of rustc?
sehqlr has joined #nixos
<acowley>
mightybyte: There was a rustc build failure on hydra for darwin
jeaye is now known as jeaye-hiring
ixxie has joined #nixos
<nek0>
hi folks, maybe someone here knows the solution to this problem: whenever I invoe `git add -p` I get the error message `Can't locate Git.pm in @INC (you may need to install the Git module)`
<acowley>
sphalerite: If I go the route of doing the overlay merging myself in one master overlay, is there a recommended library function for composing them? Should I just ape stage.nix's use of `foldl' (flip extends)`?
<nek0>
what o I need to install?
<sphalerite>
nek0: strange, git add -p works for me. Do you maybe have something in your git config that might affect it?
<{^_^}>
[nixpkgs] @thoughtpolice pushed commit from @taku0 to release-17.09 « oraclejdk: 8u161, 8u162 -> 8u171, 8u172 »: https://git.io/vptwi
<acowley>
The convention with overlays is that we have a function whose arguments are `self: super: `, but the nixpkgs manual on emacs configuration (https://nixos.org/nixpkgs/manual/#sec-emacs-config) does customization with a function whose arguments are `super: self: `. What's up with that?
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to staging: https://git.io/vptwd
<{^_^}>
→ f43da349 by R. RyanTM: wayland: 1.14.0 -> 1.15.0
<{^_^}>
→ 5bc59008 by @matthewbauer: Merge pull request #38691 from r-ryantm/auto-update/wayland
<acowley>
Is that a real difference or a documentation error? If it's real, can we have an overrideScope2 that swaps the order so I can develop some muscle memory at writing such functions?
<panther420[m]>
hi im new
<panther420[m]>
so whats cool about nixos?
jgt has joined #nixos
<sphalerite>
acowley: looks like a nasty mess, especially since it *also* uses a rec set.
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vptrL
<{^_^}>
→ 156dfd6e by @magnetophon: faust: add more faust2appl scripts
<{^_^}>
→ 48c24a8b by @matthewbauer: Merge pull request #38644 from magnetophon/faust
<dtz>
panther320: yessir
<sphalerite>
panther420[m]: declarative configuration, system configuration history and atomic rollbacks, parallel installation of multiple versions or configuartions of packages, and many other things.
<{^_^}>
→ a0f5d776 by @grahamc: Merge pull request #39157 from taku0/oraclejdk-8u171-release-18.03
drakonis has joined #nixos
<sphalerite>
gchristensen: we've got so much commit activity on nixpkgs by now that the messages are getting a bit spammy. Any chance we could only have the "PR merged" without the corresponding "commits pushed" messages, and still get "commits pushed" messages if they're done independently of a PR?
<nek0>
sphalerite: is there anything I can do? I updated my channels, but the problem still persists.
<boomshroom>
Unfortunately, RedoxOS recomendeds using the prebuilt deb toolchain rather than building from scratch. How would I wrap a deb package so it would work?
<krey>
ah, I was missing propagetBuildInputs...
jon_ has quit [Ping timeout: 248 seconds]
<manveru>
boomshroom: there's some tool to unpack the .deb file, then you have to fix the rest of the insanity...
xcmw has joined #nixos
<adisbladis[m]>
boomshroom: Unpack it yourself and a whole lot of patchelf
<adisbladis[m]>
boomshroom: Look at the spotify derivation
<boomshroom>
manveru: One thing is that I was able to build binutils and gcc, but I was having trouble with newlib because it contained a rust crate barried within the C code.
<manveru>
i thought redox was all rust
<manveru>
can't say i've looks into it yet :)
jgt has joined #nixos
<boomshroom>
manveru: The OS is, but it still has a C toolchain to port existing applications (like Nix, wink wink).
<manveru>
of course you'll have to replace the bootstrap.sh, but that's not too hard
wangoe has quit [Ping timeout: 240 seconds]
<drakonis>
at least they're not insisting posix is a saint that can do no wrong
<drakonis>
boo debootstrap doesn't chroot
jon_ has joined #nixos
pkill9 has joined #nixos
<boomshroom>
I'd also like to mention that building with rust-nightly can be challenging since it was removed from the official repos.
jgt has quit [Ping timeout: 256 seconds]
<drakonis>
boo no libreoffice icons or gtk, so i can't figure out which thing i'm opening
<boomshroom>
Honestly, I don't care how up to date or "nightly" it really is, I just want access to nightly features that are fully implemented in the latest stable release, but blocked anyways.
humanoyd has quit [Quit: WeeChat 2.1]
<boomshroom>
Adding xargo to the mess just makes it more anoying.
<drakonis>
hmm
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<drakonis>
have you tried tossing it into a container
<drakonis>
not docker though
<boomshroom>
drakonis: My main objective is to get the toolchain running as normal derivations so I could make a stdenv that targets redox.
<drakonis>
oh i see
Fare has quit [Ping timeout: 264 seconds]
<drakonis>
what's a thing to repartition devices
<boomshroom>
drakonis: Gparted? fdisk?
<drakonis>
i need to resize a partition in a flashdrive
<drakonis>
gparted isn't allowing me to alter it
<boomshroom>
drakonis: Are you root?
<drakonis>
no, user
<drakonis>
running with sudo though
linarcx has joined #nixos
<drakonis>
ah i see how it is with libreoffice, gtk and qt integration is through plugins, not some build option
ixxie has quit [Quit: Lost terminal]
jrolfs_ has quit [Ping timeout: 260 seconds]
linarcx has quit [Remote host closed the connection]
<drakonis>
and i see how it is, its done through environment variables, oh boy oh boy
<drakonis>
brb
jon_ has quit [Ping timeout: 248 seconds]
<nur0n0>
what is the purpose of 'buildInputs'? is it used solely to make sure that certain other derivations are present in the filesystem? from what I can tell so far, it seems that nix does not attempt to deal with paths ( in the sense of setting environment variables) and this must be done manually. is that correct? Ive been packaging a few projects and
<nur0n0>
I still don't have a good sense for what work I can rely on Nix to do and what work must be done manually
jon_ has joined #nixos
drakonis has quit [Remote host closed the connection]
sehqlr has joined #nixos
vaninwagen has joined #nixos
<zybell_>
boomshroom:It may be that you first need an older version of the toolchain. Check that that crate was there forever.
<boomshroom>
nur0n0: buildInputs refers to packages that should be linked into the build environment for a derivation. Things like the binaries should be in the PATH, and the headers and libraries should be available to the compiler.
<Myrl-saki>
clever: I have something that can potentially make mmap faster.
<Myrl-saki>
clever: Err, qemu-arm.
<boomshroom>
zybell_: Of the redox toolchain? It should be able to be built with just a standard C-toolchain and xargo wrapping nightly-rust.
<boomshroom>
Funny how I check the website for a major RedoxOS contributer and see "The Case for Non-Binary Genders" directly above "You Are (Probably) Doing Login Systems Wrong".
<nur0n0>
boomshroom: binaries have been working fine, but nix does not appear to deal with header files and libraries at all. I keep having to set NIX_CFLAGS_COMPILE and the like. I just wonder whether this is the standard way of doing things?
<boomshroom>
nur0n0: That shouldn't be nessisary. Can you post the derivation you're working on?
akfp has joined #nixos
<akfp>
I'd like to programmatically spin up a hetzner cloud server with nixos. On hetzner cloud I can boot using ubuntu/fedora/debian/centos. Then it's possible to attach nixos 18.03 iso as a mounted drive. Is there a nicely scripted way I can quickly "invert" the installation, installing nixos and removing the existing installation, apply some configuration.nix and be up, in a few minutes... Would be nice to have a cookbook for that.
<MichaelRaskin>
boomshroom: both are about perils of getting types wrong, right?
boomshroom is now known as boomshroom-away
<Myrl-saki>
clever: Back.
<Myrl-saki>
clever:
<Myrl-saki>
* Test if requested memory area fits target address space
<Myrl-saki>
* It can fail only on 64-bit host with 32-bit target.
<Myrl-saki>
clever: I think 32-bit qemu would run faster on a 32-bit "host".
phreedom has quit [Remote host closed the connection]
<disasm>
akfp: nixops has hetzner support. You can also use something like nixos-in-place
phreedom has joined #nixos
chisui has joined #nixos
jperras has quit [Ping timeout: 256 seconds]
<sehqlr>
hey all, I've been trying to install 18.03 on a PC for two days and I'm stuck. The partition that I mounted to /mnt/boot is not recognized as an EFI partition even though I've set it as a FAT32 and set the boot flag. Am I missing something?
<gchristensen>
nixpkgs unstable could break your computer to the point that rollbacks don't work
<drakonis>
hmm
<drakonis>
aight then
<Myrl-saki>
sphalerite: strace was mmapping/munmapping so hard, so I had to check the qemu source. I think it was probing for a 32-bit compatible memory address.
<gchristensen>
nixpkgs-* channels aren't tested to be fit for an os at all. whereas, the nixos-* channel is.
<drakonis>
it is redownloading everything again
<gchristensen>
that means he nixos-channel is tested for things like "doesn't erase your bootloader"
sigmundv__ has quit [Ping timeout: 248 seconds]
<drakonis>
seems a little excessive
<drakonis>
i'm running efi :v
<Myrl-saki>
sphalerite: So, a stupid fix I did was to just use i686-linux for qemu.
<drakonis>
but could be worse, it could brick my motherboard :v
<drakonis>
thanks old efivars versions
<sphalerite>
Myrl-saki: fair enough haha
<sehqlr>
MichalRaskin: Thanks for the tip, I got to a different error further down the line :)
<Myrl-saki>
sphalerite: But the thing that confuses me is that it works fine for a non-static build. :/
<sehqlr>
now I'm getting a 'Failed to create EFI Boot variable entry: no such file or directory'
stranger___ has quit [Quit: Connection closed for inactivity]
<LnL>
gchristensen: thought, could we add a file to the nixpkgs-* channel pipeline that triggers a warning when it's used for the module system
<sphalerite>
Myrl-saki: low-level black magic ♥
<samueldr>
sehqlr: known issue, give me 10s to fetch the issue
<Myrl-saki>
sphalerite: "Do not touch for sake of sanity"
<MichaelRaskin>
sehqlr: it might still boot, actually
<gchristensen>
LnL: yyyyes I think we could!
<drakonis>
there's a module system now?
<LnL>
gchristensen: if we could the we should :)
fuzzy-id has quit [Remote host closed the connection]
<sphalerite>
drakonis: there has been for a long time in nixos
<tilpner>
^ that's the original repository. I stripped the zfs parts, and there's some more information there
ryantrinkle has joined #nixos
<akfp>
ok, I'll give it a shot. it requires nix to be installed first on my machine though.. I don't even have that.
zybell_ has quit [Ping timeout: 265 seconds]
mkoenig has quit [Remote host closed the connection]
<tilpner>
Oh
<tilpner>
You can install it anywhere, you don't need NixOS to build this image
blankhart has quit [Ping timeout: 265 seconds]
<akfp>
I know.. the installation of nix in single-user-mode is a bit disingenious (IMHO). asking me to create /nix as root and then using it in single-user-mode.. that directory should be under $HOME in that case.
<tilpner>
I think you can do that with Nix 2.0, but you might have to install/build it manunally
szicari has joined #nixos
<tilpner>
*manually
jrolfs has quit [Ping timeout: 240 seconds]
ma27 has quit [Ping timeout: 265 seconds]
Fare has joined #nixos
linarcx has quit [Remote host closed the connection]
<joepie91>
akfp: the hashes of the binary builds include the store paths of dependencies, which are /nix-prefixed :)
<joepie91>
so you *can* change the store path, but it'll mean you won't get anything from the binary cache
<joepie91>
(and will have to build everything locally)
kim0 has joined #nixos
<tilpner>
Except if you have mount namespaces c.c
boomshroom is now known as boomshroom-away
ciastek has quit [Quit: Connection closed for inactivity]
szicari has quit [Quit: szicari]
astiles has quit [Remote host closed the connection]
astiles has joined #nixos
szicari has joined #nixos
astiles has quit [Remote host closed the connection]
astiles has joined #nixos
Rusty1_ has joined #nixos
michas_ has joined #nixos
jrolfs has joined #nixos
<Ralith>
it would be neat if nix transparently handled mount namespacing somehow
<Ralith>
maybe have single-user installs always be used through a special shell
<petersjt014[m]>
lemme see If I know enough to offer anything useful...
szicari has quit [Quit: szicari]
<akfp>
tilpner: there's only one problem with the kexec stuff: if the ssh key was given as an env-var, then there could be a generic build of this. now everyone needs to build and host their own.. otherwise - looks fantastic!
bennofs has joined #nixos
<tilpner>
akfp - You can do that, I just didn't think it would be useful
<tilpner>
akfp - In fact, clevers original has something like that
<petersjt014[m]>
Yeah, I got thru like 1/3 or something of the nix katas, so can't really say much :|
<petersjt014[m]>
except that the formatting is very tidy
<thoughtpolice>
michas_: I don't have time for a review I'm afraid, but that looks like a cool turn-key little tool, actually!
<jeaye>
Anyone have updated docs on working with React Native on NixOS?
<petersjt014[m]>
the OS or the javascript tangle?
<petersjt014[m]>
(or something else)
<norfumpit>
how would one get emacs 26 going on nixos
endformationage has joined #nixos
<jeaye>
petersjt014[m]: React Native is the JS framework.
<phry>
is there a way to import an "options" property from another module file? I want to clone a few options from another module over without having to redefine every single one of them. but on import, the propery seems to have bben handled differently
<akfp>
tilpner: ah, perfect
jgt has joined #nixos
<akfp>
tilpner: booting and installing worked perfectly.. I think it installed nixos in .. like 20 seconds!
<tilpner>
Are you sure that wasn't the kexec?
<tilpner>
It should take a little longer than that
<tilpner>
If it didn't reboot, you didn't see the installed system
<akfp>
oh wait.. it didn't install anything :-)
<tilpner>
Did you interrupt it/rebooted manually?
<tilpner>
It does things in the background, just give it a few more minutes
jgt has quit [Ping timeout: 260 seconds]
<akfp>
ok.. I ran curl ... && ./kexec_bundle without chmod 755 ./kexec_bundle.. probably failed during cloud init.
<akfp>
doing it manually now..
obadz has quit [Quit: WeeChat 2.0]
magnetophon has quit [Remote host closed the connection]
blankhart has joined #nixos
<akfp>
it pegged the cpu at 100% for 2 minutes, unpacked lots of stuff.. is it done? does it reboot automatically?
<tilpner>
Yes, just leave it alone
<tilpner>
But 2 minutes? That sounds too long
sehqlr has quit [Quit: Konversation terminated!]
sehqlr has joined #nixos
<akfp>
yes it's done or yes it's still working?
<tilpner>
It's done after it rebooted
jrolfs has quit [Ping timeout: 248 seconds]
<akfp>
almost exactly 2 minutes on the hetzner graphs.
<akfp>
this is the ultra-tiny 2G server though.
<tilpner>
I used that one too
<akfp>
nothing is happening.. 0-1% cpu
<tilpner>
journalctl -u performInstall
<tilpner>
(If you want to login and check what it's doing)
<akfp>
error: unrecognized flag '-j'
<akfp>
try 'nix --help' for more information
<tilpner>
Huh
<jmsb>
does anyone have a recentish example of developing react native on nix?
<akfp>
must be "nixos-install --no-root-passwd -j 4" .. only place with a -j?
<akfp>
I tried.. it fails with -j4.. and works without it.
blankhart has quit [Ping timeout: 256 seconds]
<akfp>
yes, I just installed a fresh nix from the nixos website.
obadz has joined #nixos
<obadz>
heeelp
<obadz>
need to edit my passwd/shadow files, tried grub init=/nix/var/nix/profiles/system/sw/bin/bash like I usually do, but keyboard doesn't work on that computer in that mode
<obadz>
keyboard works fine in bios/grub
<obadz>
any recommandations?
astiles has quit [Remote host closed the connection]
<akfp>
jmsb: that's nice.. thanks!
<tilpner>
obadz - Could try a live USB stick
sehqlr has quit [Remote host closed the connection]
astiles has joined #nixos
<obadz>
tilpner: hmmm true. I guess that hurts the ego though..
<tilpner>
Sounds better to me than debugging keyboard problems via IRC :)
<obadz>
I suppose you're right
<obadz>
sad
drakonis has quit [Remote host closed the connection]
sehqlr has joined #nixos
<akfp>
tilpner: can confirm that it all works splendidly! thanks!
<tilpner>
Yay
Acou_Bass has quit [Quit: byeeeeeeeeeeeeeee]
<akfp>
running python -m SimpleHTTPServer 8000 & on the server that has nix is a nice one-liner, and the server will spit out the IP address when cloud-init fetches the bundle, so then it's easy to later ssh root@<IP>..
astiles has quit [Ping timeout: 260 seconds]
ryanartecona has joined #nixos
pirateking has joined #nixos
hiratara has quit [Ping timeout: 265 seconds]
jsgrant has quit [Remote host closed the connection]
mightybyte has left #nixos [#nixos]
Acou_Bass has joined #nixos
asuryawanshi has quit [Ping timeout: 260 seconds]
knupfer has quit [Ping timeout: 265 seconds]
<akfp>
tilpner: is there a place hetzner will place the configured ssh keys? is /ssh_pubkey a well-known location for this or a random place you came up with?
<tilpner>
akfp - clever came up with that, and I have not used it
hiratara has joined #nixos
<tilpner>
akfp - My reasoning was that nobody would prebuilt images. So if everyone builds the image anyway, they might as well hardcode their ssh key
<tilpner>
*would trust prebuild images
ryanartecona has quit [Ping timeout: 248 seconds]
<pirateking>
i'm getting my feet wet in nixos containers. i'm following the manual https://nixos.org/nixos/manual/#ch-containers. but when as root i command the shell `nix-info` (and other commands) i get "error: opening lock file '/nix/var/nix/db/big-lock': Read-only file system". is the manual outdated?
<akfp>
my thinking is that most cloud services lets you chose which keys to start a VM with, and those keys are probably placed either in env vars or on disk somewhere.. so picking those up would be nice.
<achambe>
those keys are usually placed in a metadata server
<akfp>
tilpner: I might have one prebuilt image, but I have multiple environments - dev, pre, pro. I'm not going to use the same keys.
<achambe>
lookup cloudinit iirc
cacatoes has left #nixos ["WeeChat 1.6"]
obadz has quit [Quit: WeeChat 2.0]
obadz has joined #nixos
<achambe>
pirateking: they container is managed from the host
<achambe>
the nix store is mounted into the container as read only
<tilpner>
akfp - I agree, that would be nice
MichaelRaskin has quit [Ping timeout: 248 seconds]
<pirateking>
achambe: how do i become the host?
<achambe>
i mean, whichever machine you are running the containers on
<pirateking>
yes it's on my laptop
<achambe>
hmm, not sure why you are getting that error then
sehqlr has quit [Remote host closed the connection]
sehqlr has joined #nixos
<pirateking>
achambe: maybe the db isn't made yet. lemme follow the guide a bit more
xeji has quit [Quit: WeeChat 2.0]
day|flip has joined #nixos
<Ralith>
is there documentation for manually driving a package's build process from start to end somewhere?
jperras has quit [Ping timeout: 260 seconds]
<LnL>
nix-shell '<nixpkgs>' -A hello --run genericBuild
<LnL>
like that?
phreedom has quit [Ping timeout: 268 seconds]
phreedom_ has joined #nixos
<Ralith>
LnL: one step at a time, I mean
<Ralith>
I'm trying to debug why libcxxStdenv is apparently missing some important headers
<pirateking>
i haven't gotten to that part in my nix journey, but it should be in some build.sh file, right?
<Ralith>
pirateking: there is a *lot* of indirection and magic
<LnL>
Ralith: in that case look at the nixpkgs manual it explains the stdenv phases
<LnL>
the unpackPhase function should just do the right thing if it's overridden
judson has quit [Remote host closed the connection]
Wharncliffe has joined #nixos
<Ralith>
pirateking: yes, aside from that cite (which appears to contain the requisite magic incantation) that just says "run the phases," which doesn't address the problem
endformationage has quit [Ping timeout: 260 seconds]
endformationage has joined #nixos
jperras has joined #nixos
kisik21 has joined #nixos
joehh has joined #nixos
alex`` has quit [Ping timeout: 264 seconds]
dcol has joined #nixos
jrolfs_ has quit [Ping timeout: 240 seconds]
<dcol>
anyone know how I can make emacs daemon get new packages? If I add some emacs packages to my configuration.nix they get added if I start the standalone emacs client but when I try the systemd daemon the packages are not there. I managed to get it to work by deleting the path in nix-store so that it forces nixos-rebuild to build the package anew, but that feels very contrived..
dbmikus has quit [Quit: WeeChat 2.1]
stabbmesiter has joined #nixos
Synthetica has quit [Quit: Connection closed for inactivity]
endformationage has quit [Ping timeout: 240 seconds]
thefloweringash has joined #nixos
matklad has quit [Remote host closed the connection]
endformationage has joined #nixos
pirateking has quit [Quit: Page closed]
<Ralith>
LnL: do you participate in maintenance of the llvm package? I think I've got a lead on exactly what's broken here
joehh has quit [Ping timeout: 260 seconds]
jcb_ has joined #nixos
endformationage has quit [Ping timeout: 256 seconds]
jperras has quit [Ping timeout: 268 seconds]
jcb_ has quit [Client Quit]
jmsb has left #nixos ["ERC (IRC client for Emacs 25.2.1)"]
endformationage has joined #nixos
kisik21 has left #nixos [#nixos]
stabbmesiter has quit [Quit: Leaving]
ryanartecona has joined #nixos
sehqlr has quit [Remote host closed the connection]
jgt has joined #nixos
sehqlr has joined #nixos
orivej has joined #nixos
ryanartecona has quit [Ping timeout: 248 seconds]
jensens has quit [Ping timeout: 248 seconds]
<dhess>
Anyone seen this error with hydra-server.service on a recent Nixpkgs?
<jcb_>
Hi, I am trying to build taffybar-1.0.1 and have run into this issue: https://github.com/NixOS/nixpkgs/issues/38531. I am wondering how often hackage-packages caches are updated on the nix servers?
Fare has joined #nixos
<jcb_>
the taffybar on nixos-unstable is still version 1.0.0, which is broken due to a dependency issue in gtk-traymanager-1.0.0
<jcb_>
I can successfully build taffybar-1.0.1 with stack, but I would prefer to directly manage this package in configuration.nix
<jcb_>
I am fairly new to nix, and am hoping there is a convenient way for me to override the package
<jcb_>
Or perhaps, you know, some nixos developer can run the hackage-update script on the nix server and pull in the latest packages
astiles has joined #nixos
<jcb_>
given that that is the solution proposed in the above issue, which is marked closed
ertes has quit [Ping timeout: 265 seconds]
boomshroom-away is now known as boomshroom
sehqlr has quit [Remote host closed the connection]
astiles has quit [Ping timeout: 276 seconds]
sehqlr has joined #nixos
<petersjt014[m]>
Is there a specific reason installing nix on the WSL uses the unstable channel?
<gchristensen>
nix on its own always uses unstable by default
<petersjt014[m]>
Kinda makes sense--can't make as many assumptions about the env if it's not nixos
<gchristensen>
typically people only use the nixos-* channels (the only one we have a stable channel for) when using nixos which needs the extra testing