<manveru>
tazjin: what's in your `sudo systemctl list-timers`?
mk-fg has joined #nixos
mk-fg has quit [Changing host]
mk-fg has joined #nixos
Lisanna has joined #nixos
<Lisanna>
ugh... is there a way to make forgetting to setup the proxies during nixos-install *not* a death sentence when you boot and try to nixos-rebuild?
<Lisanna>
it's the darn nix-daemon, nixos-rebuild isn't passing any local proxy information to it apparently
<manveru>
Lisanna: you can try with NIX_REMOTE="" and HTTP_PROXY?
<manveru>
:)
<manveru>
i think nix 2.0 doesn't use that variable anymore though...
<Lisanna>
oh fun...
michas_ has quit [Ping timeout: 252 seconds]
ebzzry_ has quit [Ping timeout: 260 seconds]
<manveru>
ah, maybe still possible with NIX_REMOTE=local
bpa has quit [Quit: Leaving]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 opened pull request #36399: guile: Fix cross more, including 1.8 (staging...no-stdenv-cross-adapter) https://git.io/vAFYv
NixOS_GitHub has left #nixos [#nixos]
dan_b has quit [Ping timeout: 256 seconds]
cement has quit [Ping timeout: 245 seconds]
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
erasmas has quit [Quit: leaving]
fresheyeball has joined #nixos
digitus has quit [Quit: digitus]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] shlevy opened pull request #1956: Store the schema version in the database. (master...safe-schema-upgrade) https://git.io/vAFOE
NixOS_GitHub has left #nixos [#nixos]
<vil>
I've been trying to modify the net_snmp derivation to build the python modules, and I've gotten as far as getting them built, but I can't figure out how to add them to the PYTHONPATH like the normal python packages are
<vil>
is there another derivation that does this that I can use as an example?
<ottidmes>
vil: I've got almost no experience with Python, let alone on NixOS, but shouldn't you use "withPackages" to make them available in scope of your Python installation? For example: python3.withPackages (pyPkgs: with pyPkgs; [ numpy toolz ] ++ yourCustomPyPkgs)
<vil>
ooh, that looks promising
<vil>
I've been trying to figure out what the magic is that causes pythonPackages.* to just appear in PYTHONPATH when you pull them in
<vil>
the only mention of it I found is one line in the docs that says something to the effect of "anything that depends on a python interpreter will be added"
<vil>
but I made my net_snmp depend on python27 and no luck
<vil>
the module is ending up in the result lib/python*/site-packages as it's supposed to
<ottidmes>
vil: Sorry, thats all I got, hopefully somebody more experienced with Python on NixOS will be able to help
<vil>
np, thanks for the lead!
<vil>
I'm not too concerned about this package being "optimal" since I'm just using it in a nix-shell
<sonarpulse>
johnw: I think --keep-going and --keep-failed are at least accepted
<sonarpulse>
if not documented
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/staging b5cf1d2 John Ericson: Merge pull request #36399 from obsidiansystems/no-stdenv-cross-adapter...
<NixOS_GitHub>
nixpkgs/staging 76bc1d3 John Ericson: guile: Fix cross more, including 1.8
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 2 new commits to staging: https://git.io/vAFsm
NixOS_GitHub has left #nixos [#nixos]
logzet_ has quit [Remote host closed the connection]
vidbina has quit [Ping timeout: 256 seconds]
ebzzry_ has joined #nixos
LnL has quit [Ping timeout: 265 seconds]
Synthetica has quit [Quit: Connection closed for inactivity]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 opened pull request #36401: stdenv cross adapter: Get rid if `selfNativeBuildInput` (staging...no-stdenv-cross-adapter) https://git.io/vAFs5
NixOS_GitHub has left #nixos [#nixos]
<digitalmentat>
grahamc, have you ever seen the hydra-queue-runner wedge for no good reason at all?
<gchristensen>
why do you ping grahamc specifically?
<digitalmentat>
for some reason grahamc is the alias I think you are identified by
<digitalmentat>
I'll stop doing that
<gchristensen>
ok no worries
<gchristensen>
usually when people ping grahamc it is because they saw me in an issue :)
<gchristensen>
I never have but I really haven't operated Hydra much
<digitalmentat>
oh okay
<digitalmentat>
we're getting murdered by numerous Hydra bugs at the moment
<gchristensen>
ouch
<gchristensen>
Nix 2?
<MP2E>
Nix 2 is killing me as well, can't fetch any binary caches over SSH no matter what I try :/
<gchristensen>
cc shlevy
<gchristensen>
Nix 2 has been in development for a long time
<digitalmentat>
no, we're pinned to nixpkgs-17.09
pkill9 has quit [Ping timeout: 260 seconds]
<digitalmentat>
so using Hydra from November
<gchristensen>
there is inevitably going to be troubles. the important thingis fixing them asap
<gchristensen>
digitalmentat: hmm I know hydra.nixos.org is running Nix 2, so I wonder if some things have been resolved
<digitalmentat>
it's very possible
<digitalmentat>
the issues we're seeing are mostly related to hydra-queue-runner stability and github pr building stability
ebzzry_ has quit [Ping timeout: 256 seconds]
<gchristensen>
maybe one day ofborg will be ready to fill that need :)
<shlevy>
digitalmentat: I haven't run my own hydra for a while, havne't seen that particular issue... But there are definitely stability issues :|
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/staging 8d58008 John Ericson: stdenv cross adapter: Get rid if `selfNativeBuildInput`...
<NixOS_GitHub>
[nixpkgs] Ericson2314 pushed 2 new commits to staging: https://git.io/vAFGl
<NixOS_GitHub>
nixpkgs/staging 8f54d93 John Ericson: Merge pull request #36401 from obsidiansystems/no-stdenv-cross-adapter...
NixOS_GitHub has left #nixos [#nixos]
<digitalmentat>
shlevy, okay, thanks
justan0theruser has quit [Quit: WeeChat 1.9.1]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] obadz pushed 1 new commit to master: https://git.io/vAFGK
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master ee9a785 obadz: nixos doc: give context and definition re NIXOS_LUSTRATE...
justanotheruser has joined #nixos
digitalmentat has quit [Quit: Leaving]
markus1199 has quit [Ping timeout: 240 seconds]
ebzzry_ has joined #nixos
vidbina has joined #nixos
markus1189 has joined #nixos
lverns has joined #nixos
__monty__ has joined #nixos
<__monty__>
Is it possible to nix-env --delete-generations <all but the last two>?
<rawtaz>
bdimcheff: the astroid package is using an old insecure webkitgtk. i tried overriding stuff in it using https://pastebin.mozilla.org/9079278 but it failed on "Checking for webkitgtk-3.0... no" with "webkitgtk not found.". is there any way this package can be updated soon, to make use of more current webkitgtk thats not insecure?
ebzzry_ has quit [Ping timeout: 268 seconds]
<gchristensen>
rawtaz: it sounds like the update has to be done upstream
hellrazor has joined #nixos
Guest52462 has quit [Ping timeout: 260 seconds]
<bdimcheff>
I think upstream astroid is broken
<bdimcheff>
It's a totally unsupported webkitgtk
<rawtaz>
gchristensen: indeed :)
<bdimcheff>
I haven't looked in the last few months though
rauno has quit [Ping timeout: 240 seconds]
<rawtaz>
bdimcheff: yeah.. broken in what way? you mean that it depends on that old webkit and using a newer one is something that astroid (completely isolated from nix) has to be updated to support?
<bdimcheff>
I believe so, yes
<rawtaz>
darn
<gchristensen>
yes, we point it at the insecure webkitgtk because it doesn't work with any which are maintained
<rawtaz>
then that probably means that compiling it manually isnt much of an option either then (unless you want the insecure webkit)
<rawtaz>
very general question: for something like astroid, what other toolkits would be proper stuff to use? qt?
<gchristensen>
astroid basically needs a browser
<rawtaz>
ah right. its the browser we're talking about here, not the window toolkit like GTK
<rawtaz>
thanks
ryanartecona has quit [Quit: ryanartecona]
<gchristensen>
yeah
<gchristensen>
and browsers require constant patching to keep up with the bugs found
<rawtaz>
very much so..
sonarpulse has quit [Ping timeout: 240 seconds]
lverns has left #nixos [#nixos]
gammarray has joined #nixos
gammarray has quit [Remote host closed the connection]
ottidmes has quit [Ping timeout: 276 seconds]
LnL has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to glibc-2.27: https://git.io/vAFcc
<NixOS_GitHub>
nixpkgs/glibc-2.27 cb13826 Shea Levy: Merge branch 'staging' into glibc-2.27
NixOS_GitHub has left #nixos [#nixos]
__monty__ has quit [Quit: leaving]
LnL has quit [Ping timeout: 268 seconds]
Supersonic has quit [Disconnected by services]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
markus1199 has joined #nixos
markus1189 has quit [Ping timeout: 240 seconds]
d4g_ has joined #nixos
__monty__ has joined #nixos
<__monty__>
I'm having trouble specifying texlive in my package overrides I'm getting this error: http://ix.io/Tzz, this is my config.nix: http://ix.io/Tzy
d4g has quit [Ping timeout: 240 seconds]
__monty__ has quit [Quit: leaving]
<Rick[29SMPittsbu>
Hi how are all of you doing ;D
<Rick[29SMPittsbu>
Finally find a program I can use the client with, that will fit my needs.
<Rick[29SMPittsbu>
so that is awesome :D
<Rick[29SMPittsbu>
So I just installed Nix on Windows Subsystem on Linux, that was nice
<Rick[29SMPittsbu>
it doesn't work perfectly, you have to tinker a little with a file to make it work better. But no complaints.
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/glibc-2.27 d61fe2d Shea Levy: Temporarily make acl part of the jobset for hydra debugging
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to glibc-2.27: https://git.io/vAFWG
<Rick[29SMPittsbu>
I still need to get around to learning nix declarative language. Now compared to python and other high level languages what would you guys say is the difference?
vidbina has quit [Ping timeout: 260 seconds]
ql-to-nix has joined #nixos
<ql-to-nix>
Do Bad Things happen if the builder for a derivation ultimately uses nix-build to build some other derivation?
sirkha has quit [Quit: Leaving]
pxc has quit [Ping timeout: 260 seconds]
mmlb has quit [Ping timeout: 240 seconds]
mmlb has joined #nixos
mbrgm has quit [Ping timeout: 240 seconds]
thc202 has quit [Ping timeout: 240 seconds]
Lisanna has quit [Quit: Lisanna]
mbrgm has joined #nixos
<awygle>
is nix pretty fundamentally based around having a central repository of packages?
AluisioASG has quit [Remote host closed the connection]
AluisioASG has joined #nixos
<ql-to-nix>
<--- nix noob, but I don't think so. You can ask nix to build any nix expression you want. That expression doesn't have to depend on nixpkgs
<aminb>
does anyone know how to see the list of updates before actually doing `sudo nixos-rebuild switch`?
pxc has joined #nixos
ql-to-nix has quit [Ping timeout: 260 seconds]
<aminb>
it'd be nice if I could see what packages I'm gonna be upgrading before the actual upgrade
mduggie has joined #nixos
<mduggie>
hey, i'm trying to add some modules to my config that I found on nixpkgs. What's the recommended approach for doing this?
<mduggie>
should I clone the repo and link to the module I'm trying to add in the imports statement?
<frio>
you can maybe use --test aminb, but off the top of my head there isn't anything that does that
<frio>
maybe if there was some kind of nix-diff to diff two derivations?
<frio>
basically, given some builds and tests in the middle, github/nixpkgs --> a build --> a snapshot --> a "nix channel" --> your computer via a tarball when you update your channels --> your configuration.nix
<frio>
unless the module you're looking at was added 2 days ago, hasn't been thru that process/you haven't upgraded in a while, if you read it on nixpkgs it's pretty straightforward to use it
<mduggie>
cool, sounds good. so if I'm trying to get znc up and running, I'd just have to add the line `services.networking.znc.enable = true`?
<mbrock>
the fix was merged in November, but I still get that same error using nixUnstable from the current nixos-unstable channel
ChongLi has quit [Ping timeout: 252 seconds]
shikiryogy has joined #nixos
Itkovian has quit [Read error: Connection reset by peer]
<mbrock>
ah, false alarm, there was something weird about my git state
<shikiryogy>
Hi guys, I can't connect from a docker container to the host, I tried several ways (192.168..., 172....) doesn't seems to work, should I do something additional on Nixos?
pxc has joined #nixos
ChongLi has joined #nixos
vaninwagen has joined #nixos
pxc has quit [Ping timeout: 240 seconds]
zzamboni has quit [Read error: Connection reset by peer]
nico202 has quit [Ping timeout: 240 seconds]
asuryawanshi has quit [Remote host closed the connection]
asuryawanshi has joined #nixos
<Mic92>
shikiryogy: does your container has the ip command?
<shikiryogy>
yes
MichaelRaskin has left #nixos [#nixos]
<Mic92>
shikiryogy: ip r
leat has joined #nixos
<Mic92>
there is a default gateway in the first line
<Mic92>
it contains the ip address of the host
LnL has joined #nixos
<Mic92>
ok, assuming you know that. your firewall could be messed up also.
zzamboni has joined #nixos
<Mic92>
tcpdump -i docker0 -n on the host is then your friend, or tcpdump -i <container-interface>, where container-interface is the name of the to bridged attached container vether interface
jmeredith has quit [Quit: Connection closed for inactivity]
<Mic92>
the -n is always important to suspress dns resolution
jensens has joined #nixos
sigmundv has joined #nixos
<hyper_ch>
sphalerit: sphalerite: howdy
Itkovian has joined #nixos
<hyper_ch>
Mic92: I did just add the sqlite change to my homeserver manually... and there's like a gazillion of things that are being recompiled... I wouldn't have thought it would be such a big deal
<shikiryogy>
Mic92: ok, after the tcpdump what should I try?
zzamboni has quit [Client Quit]
zzamboni has joined #nixos
<Mic92>
shikiryogy: produce the traffic that does not come through.
<Mic92>
and take a look if you see it on the bridge
mduggie has joined #nixos
<shikiryogy>
Mic92: I don't see anything there, shouldn't the ping from the container show at least?
<Mic92>
hyper_ch: that's why do not add sqlite to master directly and it goes to staging first.
<Mic92>
otherwise master would never finish with all these rebuilds
<hyper_ch>
Mic92: as said, I didn't think it would be such a big deal to just add two compile options :9
<Mic92>
sqlite is one of the most used software libraries.
pxc has joined #nixos
<hyper_ch>
but why does it compile other stuff as well when just sqlite has changed
<Mic92>
shikiryogy: So you do not see it on both the container interface on the host and the bridge using the tcpdump command with the -n?
<Mic92>
mduggie: it uses the same version that is in nixpkgs at the time.
<dmj`>
anyone have a derivation for facebook’s jest library?
<mbrock>
I'm trying to install a package defined by `pkg.overrideAttrs (_: { src = ~/src/pkg; })` but I get an error "do not know how to unpack source archive /nix/store/xxx" -- weirdly, that store path exists and seems fine
<shikiryogy>
Mic92: I executed the tcpdump command on the host, pinged from within the container, and nothing shows on the tcpdump
<mduggie>
at what time, Mic92? when it's built?
pxc has quit [Ping timeout: 264 seconds]
zzamboni has quit [Quit: Leaving.]
<Mic92>
mduggie: when you say nixos-rebuild switch, your local nixos channel points to a specific commit -> nixos modules are from the same channel/source and the version of the corresponding is somewhat hard coded.
<Mic92>
*corresponding package
<Mic92>
you can override versions though with overlays or packagesOverrides
<Mic92>
shikiryogy: ok, what does ping tell you exactly?
zzamboni has joined #nixos
<shikiryogy>
PING 172.19.0.1 (172.19.0.1): 56 data bytes
<shikiryogy>
64 bytes from 172.19.0.1: icmp_seq=0 ttl=64 time=0.083 ms
<Mic92>
shikiryogy: so it actually works, but you do not so anything?
<Mic92>
shikiryogy: try tcpdump -i any -n icmp on the host next
<mduggie>
ah i see. thanks for clarifying Mic92. is there a way to reference the current version number, like `pkgs.znc.version`?
achambers has quit [Quit: Page closed]
<Mic92>
mduggie: you can use `nix repl` in nix 2.0 or the nix-repl package in the current stable nixos to explore packages:
<Mic92>
mduggie: nix repl
<Mic92>
nix-repl> :l <nixpkgs>
<Mic92>
nix-repl> znc.versi<TAB>
<Mic92>
<TAB> is literally pressing tab
<mduggie>
i mean for use in a .nix file, not for user installations
<mduggie>
nvm
<mduggie>
didnt mean to press enter lol
<mduggie>
I decided to actually read your messages, and they answered that question
<mduggie>
thank you for the help
ebzzry_ has joined #nixos
nico202 has joined #nixos
Intensity has quit [Remote host closed the connection]
<NixOS_GitHub>
[nixpkgs] tiramiseb opened pull request #36410: gnomeExtensions.system-monitor: fix this package and upgrade to v33 (master...gnome-shell-extension-system-monitor) https://git.io/vAFPw
NixOS_GitHub has left #nixos [#nixos]
pxc has quit [Ping timeout: 256 seconds]
__Sander__ has joined #nixos
ZoomZoomZoom has joined #nixos
ebzzry_ has quit [Read error: Connection reset by peer]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] tiramiseb opened pull request #36411: gnomeExtensions.impatience: fix the package and upgrade to version 0.4.5 (master...gnome-shell-extension-impatience) https://git.io/vAFXW
<NixOS_GitHub>
[nixpkgs] edolstra pushed 1 new commit to master: https://git.io/vAF13
<NixOS_GitHub>
nixpkgs/master 1deb779 Eelco Dolstra: nixUnstable: Point to Nix 2.0
NixOS_GitHub has left #nixos [#nixos]
vaninwagen has quit [Quit: Connection closed for inactivity]
roberth has joined #nixos
freeman42x]NixOS has joined #nixos
alexteves has joined #nixos
Thra11 has joined #nixos
Arcaelyx has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
vaninwagen_ has joined #nixos
<vaninwagen_>
Hi, i need to work on a python project on my nixos machine. i get my python using nix-shell and a default.nix with a stdenv derivation including python and all the c-libs i need. unfortunately LD_LIBRARY_PATH is not updated correctly (i.e. for geos). is there a preferred way to set it other than manipulating it manually?
seanparsons has quit [Read error: Connection reset by peer]
humanoyd has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] edolstra closed pull request #1955: rand() -> random(), since we use srandom(). (master...fix/rand-random) https://git.io/vAFtB
<NixOS_GitHub>
nixpkgs/glibc-2.27 cc5e442 Shea Levy: libxshmfence: Bump to 1.3....
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to glibc-2.27: https://git.io/vAF5I
NixOS_GitHub has left #nixos [#nixos]
thblt has joined #nixos
ma27 has joined #nixos
<hyper_ch>
*homeserver is still happily recompiling just about everything*
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ma27 closed pull request #36296: nixos/gnome-keyring: add option to enable the daemon as systemd service (master...run-gnome-keyring-independently) https://git.io/vAHZz
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/release-18.03 5868f9a Nikolay Amiantov: primusLib: build only for x86 Linux...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to release-18.03: https://git.io/vAF5F
simukis has joined #nixos
<manveru>
is there some way to run the quake3 test so it's displayed?
<manveru>
i think it used to in the past...
ma27 has quit [Quit: WeeChat 2.0]
knupfer has quit [Remote host closed the connection]
ma27 has joined #nixos
<mpickering>
I managed to get qgis to build successfully on OSX last night. Now when running the executable it fails to load some dynamic libraries. The problem is that at the end of installation, the installed tries to symlink them all into the output directory. I should instead stop this happening and make them somehow reference the correct nix path?
ma27 has quit [Client Quit]
pxc has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to glibc-2.27: https://git.io/vAFdM
<NixOS_GitHub>
nixpkgs/glibc-2.27 6ac7b19 Shea Levy: gnat: Mark broken.
NixOS_GitHub has left #nixos [#nixos]
<BlessJah>
gchristensen: gotcha of borg: when you use authorized_keys:command="borg serve --apend-only", borg init will save that to config; I wanted to have crontab on receiving end do the cleanup
<arximboldi>
that is fantastic, thanks a lot tilpner!!!
<nick_l>
Using NixOps, how do I compute the id of a particular security group?
crst has quit [Remote host closed the connection]
zzamboni has joined #nixos
<pmeunier>
Hi! I'm using the unstable channel, and it seems NIX_REMOTE is not set anymore, causing nix-env to not work as user. Anything I can do to investigate further/fix the problem?
<tilpner>
arximboldi - Another way would be to upstream configurability of these features (provided they work well), but leave them off by default
<arximboldi>
yes, I guess that would be ideal, I'm not sure I am confident enough with the Nix language to make a proper change there... but it'll look into it. for now I am just playing with this for a prototype
vidbina has quit [Ping timeout: 240 seconds]
zzamboni has quit [Client Quit]
zzamboni has joined #nixos
<LnL>
pmeunier: nix 2.0 doesn't require that anymore
<ottidmes>
pmeunier: I believe the fix is to set environment.sessionVariables.NIX_REMOTE = "daemon"; for now, for those programs that still use 1.11, like nix-repl, I think
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
alex`` has joined #nixos
zzamboni has quit [Ping timeout: 256 seconds]
davidlt_ is now known as davidlt
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to staging: https://git.io/vAFNN
<NixOS_GitHub>
[nixpkgs] shlevy pushed 1 new commit to glibc-2.27: https://git.io/vAFNx
<NixOS_GitHub>
nixpkgs/glibc-2.27 6dd3f49 Shea Levy: Merge branch 'staging' into glibc-2.27
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to master: https://git.io/vAFAe
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master b6a2833 Nikolay Amiantov: megatools: 1.9.98 -> 2017-10-26...
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti closed pull request #23023: haskell: envForPackages (hs: with hs; [A B]) should contain all dependencies of A and B, but not A, B themselves (master...export-haskell-inputs) https://git.io/vDFKs
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master df6e6d9 Peter Simons: Merge pull request #36393 from ElvishJerricco/haskell-shell-for...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 9adb4d2 Will Fancher: Haskell: Added haskellPackages.shellFor
<NixOS_GitHub>
[nixpkgs] peti pushed 2 new commits to master: https://git.io/vAFAJ
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to release-18.03: https://git.io/vAFAt
<NixOS_GitHub>
nixpkgs/release-18.03 0e44dd1 Nikolay Amiantov: megatools: 1.9.98 -> 2017-10-26...
NixOS_GitHub has left #nixos [#nixos]
zzamboni has joined #nixos
colescott has joined #nixos
zzamboni has quit [Ping timeout: 240 seconds]
zzamboni has joined #nixos
zzamboni has quit [Client Quit]
clefru has joined #nixos
mjvoge02 has quit [Ping timeout: 240 seconds]
mjvoge02 has joined #nixos
zzamboni has joined #nixos
freeman42x]NixOS has quit [Ping timeout: 240 seconds]
pxc has joined #nixos
zzamboni has quit [Client Quit]
zzamboni has joined #nixos
orivej has quit [Ping timeout: 256 seconds]
ma27 has quit [Quit: WeeChat 2.0]
orivej has joined #nixos
pxc has quit [Ping timeout: 276 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 399edfe Fatih Altinok: nodejs: 9.6.1 -> 9.7.1
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vAFpC
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vAFp4
NixOS_GitHub has left #nixos [#nixos]
Itkovian has quit [Read error: Connection reset by peer]
Zoom__ has joined #nixos
ZoomZoomZoom has quit [Read error: Connection reset by peer]
lexnoproblem has joined #nixos
Itkovian_ has joined #nixos
<lexnoproblem>
hey, how is it going?
kim0 has joined #nixos
rauno has joined #nixos
<lexnoproblem>
I would like to specify a version of nixpkgs for my nix-shell command, e.g. nixpkgs=https://github.com/NixOS/nixpkgs-channels/archive/nixos-14.12.tar.gz , but https://github.com/NixOS/nixpkgs-channels/archive/ is 404
fleaz has quit [Ping timeout: 260 seconds]
tnias has quit [Ping timeout: 260 seconds]
tnias has joined #nixos
fleaz has joined #nixos
troydm has quit [Ping timeout: 260 seconds]
MinceR has quit [Ping timeout: 260 seconds]
vaninwagen_ has quit [Ping timeout: 269 seconds]
vaninwagen_ has joined #nixos
MinceR has joined #nixos
troydm has joined #nixos
<aminechikhaoui>
-I nixpkgs=channel:nixos-14.12 if you have nix 2.0
Itkovian_ has quit [Read error: Connection reset by peer]
<lexnoproblem>
17.09 standart default mainstream. not a 2.0 I guess
<__monty__>
How do I specify the texlive.combine function in a myEnvFun?
<srhb>
__monty__: There's no difference that I know of. But mind that you're in a list.
<srhb>
So you probably want (texlive.combine ...)
<lexnoproblem>
srk, I though maybe list can't contain dynamic items, and those items could be pre-evaluated in 'let' section, but I could be wrong
<srhb>
__monty__: That is, [ f x ] is a list of two elements, not a list of f applied to x
<__monty__>
srhb: Ah, seems like that was it. Only other thing was missing () around the texlive in inherit texlive scheme-basic. Why are those necessary?
<lexnoproblem>
sorry for chaning topic, but : I would like to specify a version of nixpkgs for my nix-shell command, e.g. nixpkgs=https://github.com/NixOS/nixpkgs-channels/archive/nixos-14.12.tar.gz , but https://github.com/NixOS/nixpkgs-channels/archive/ is 404
<srhb>
__monty__: inherit (from) thing-to-inherit
<srhb>
Otherwise it's just the surrounding scope
<lexnoproblem>
which nixpkgs working url I can specify for nix-shell -I for 17.09
alex`` has quit [Quit: WeeChat 2.0.1]
logzet has joined #nixos
<__monty__>
srhb: Ok, guess I'll just accept that, not obvious from the manual.
<srhb>
lexnoproblem: nixpkgs=https://github.com/NixOS/nixpkgs-channels/archive/nixos-17.09.tar.gz for instance
<srhb>
"the set used in the function call to the function defined in ../tools/graphics/graphviz inherits a number of variables from the surrounding scope (fetchurl ... yacc), but also inherits libXaw (the X Athena Widgets) from the xlibs (X11 client-side libraries) set."
<lexnoproblem>
srnb, thx, that works, badly there is no url to browse available versions like https://github.com/NixOS/nixpkgs-channels/archive/ , and 17.09 still continue to incude updates, I would like some frozen static version of 17.09 that won't be updated-changed
<srhb>
lexnoproblem: Then you should refer to a specific commit. They're automagically made by github for each branch, so it's just the branch list
<__monty__>
srhb: Yeah, I found it after I knew it was special syntax but it's barely a paragraph worth of explanation, not exactly obvious when you're just looking for the definition of inherit.
<srhb>
__monty__: Ok :)
<__monty__>
srhb: Is your texlive setup availabe online?
<srhb>
__monty__: Afraid not. :/
anton_ has quit [Remote host closed the connection]
anton_ has joined #nixos
Itkovian has quit [Remote host closed the connection]
<srhb>
lexnoproblem: (So find the commit on a branch like 17.09, then the url is .../archive/commithash.tar.gz)
<__monty__>
Ok, you said you started from scheme-basic, do you just run latex, find out what package the missing files are from, add the package to your config.nix, wash, rinse and repeat? Is there a way to find out what's in scheme-basic, scheme-medium, etc.? And maybe a way to more easily find all the various scheme-*?
<gchristensen>
BlessJah, LnL: the original planning of it depended upon lots of community members volunteer their computers whenever they could, to provide build space... but they'd all be "speaking" / reporting under the same name. This made me think of "Locutus of Borg", and when I make things that are potentially spammy / annoying I try to make it very obvious that it is my fault so people can come ask me to turn it off
<gchristensen>
asap... thusly grahamcofborg
<__monty__>
srhb: Ah, thanks, was missing the qualifications.
<infinisil>
srhb: why the wrapfig?
<srhb>
infinisil: Arbitrary package example outside of the scheme-x'es :)
<infinisil>
Ah
<infinisil>
srhb: the pkgs. in there aren't needed though btw
<__monty__>
Hmm, you're right.
<srhb>
infinisil: No, indeed, but it's portable to expressions of type { pkgs, ...}: like that.
<infinisil>
true dat
<srhb>
And I prefer to be explicit :)
<BlessJah>
gchristensen: how would you prevent someone injecting malicious code?
<BlessJah>
nixos doesn't test builds reproducibility (does it?)
<__monty__>
So I'm missing lmodern.sty, the lmodern package is at the top level of nixpkgs so including it in the list after scheme-basic isn't working. Do I just specify it as another package or should I find a latex package instead?
rauno has quit [Ping timeout: 260 seconds]
<ij>
How do I compute a store path's closure with the new nix?
<ij>
The closure tree.
<srhb>
__monty__: It's called lm in the tex packageset
<gchristensen>
BlessJah: (1) all the builds happen inside the tight sandboxing Nix provides (2) ofborg only builds PRs where certain people have instructed it to do so
<lexnoproblem>
can anyone tell me not to specify nix-shell -i interpreter in shell\python script based on this example of what I have https://hastebin.com/oqixutediz.vbs
<__monty__>
srhb: How can I find this for myself? I grepped pkgs.nix for lmodern, do I need to try all possible prefixes?
<lexnoproblem>
lol, sorry, I mean can anyone tell me HOW. *typing mistake
<srhb>
__monty__: The names are from upstream, so I generally search the tug things for the file I need to find the package in question
xcmw has quit [Ping timeout: 264 seconds]
zzamboni has quit [Quit: Leaving.]
<srhb>
__monty__: (ie: google to the rescue)
<__monty__>
srhb: Hmm, the search on tug.org doesn't give me the package name. Do you then go through all the links or is there a better way to formulate the query?
zzamboni has joined #nixos
<srhb>
__monty__: I can't really give you a generic answer. It's the first search result for me on Google when I search for lmodern.sty, but my search bubble is highly specialized and honed through years of training :-P
<srhb>
And also, the only indication that it's called lm is in the _url_ of the page.
pxc has joined #nixos
<srhb>
TeX packaging is really not very good.
zzamboni has quit [Client Quit]
zzamboni has joined #nixos
<__monty__>
Ah, google turns it up. I used the search on tug.org and that's a ddg site-specific search which doesn't turn up the package name unfortunately.
<__monty__>
I'll manage, thanks : )
ma27 has joined #nixos
<srhb>
__monty__: If it's in the bin cache, you could try nix-locate (from the nix-index package)
<srhb>
__monty__: It'll try to match filename to package
asuryawanshi has quit [Remote host closed the connection]
<fearlessKim[m]>
is the latest (master) doc put on the website. I see some outdated info in the manual, would like to quickly if it changed on master
pxc has quit [Ping timeout: 260 seconds]
asuryawanshi has joined #nixos
__monty__ has quit [Read error: Connection reset by peer]
<srhb>
fearlessKim[m]: I think you'll have to grab it from hydra
<fearlessKim[m]>
shrb: it would be neat to push the updated doc on the website
<nick_l>
gchristensen: I got it to work, but now I have the reverse problem which is that "./script" works, but "bash script" doesn't or is that to be expected?
<nick_l>
gchristensen: ever read "In the beginning was the command line"?
<gchristensen>
nope... :)
<nick_l>
gchristensen: I don't know anyone who is into automation who doesn't like that.
<gchristensen>
I'm on it, thanks!
__monty__ has joined #nixos
ma27 has quit [Quit: WeeChat 2.0]
<infinisil>
Hmmm.. I'm currently writing this script to verify urls in nixpkgs, but I have some problems that are hard to solve with bash
<nick_l>
infinisil: you can build a binary as part of your build script, right?
<nick_l>
infinisil: so, build e.g. a boost C++ program, and then just call that.
<infinisil>
I'm not sure, because I would've liked it to be a script that everybody could run
<infinisil>
My next best language would be haskell, which needs lots of dependencies and things
<infinisil>
But since this script problably wouldn't be actually run by a lot of people it might be okay
<nick_l>
infinisil: if these only need to run on a local machine, I couldn't care less.
iyzsong has joined #nixos
ma27 has joined #nixos
<nick_l>
infinisil: if it becomes part of the closure on a remote machine, I would start caring.
<infinisil>
The goal in the end would be to have this script run every week or so and make a PR with url updates automatically, so yeah this probably only runs on one machine
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<nick_l>
infinisil: well, then there is really no point in caring about dependencies.
<infinisil>
true, thanks, gonna write it in haskell instead \o/
<nick_l>
infinisil: the reason for not using Haskell would be stability.
<lexnoproblem>
excuse me, sorry for bothering you, buy how I can get this nix-shell + expression + interpreter get to work toghether https://hastebin.com/fadesequxo.scala ?
<nick_l>
infinisil: do you really want to fix library breakage all the time? :)
<infinisil>
nick_l: I could just use stack
<infinisil>
or the nix package set which is i think derived from stackage's set
<__monty__>
nick_l: Why would haskell not be stable?
<nick_l>
__monty__: perhaps things have changed in the past 5 years, but I don't believe people change.
<nick_l>
__monty__: and the same people are still in control
<nick_l>
__monty__: and they don't care about introducing breakage in compilers or not.
<srhb>
Pah, naysayer! Go go team haskell :-)
<nick_l>
__monty__: I also think GHC is a low-quality compiler, but that only indirectly relates to the breakages.
<__monty__>
Uhm, WAT?
<nick_l>
Haskell is a nice language; it just has no good implementation.
<infinisil>
What's so bad about ghc?
* srhb
breathes into a bag
ma27 has quit [Quit: WeeChat 2.0]
<nick_l>
infinisil: the fact that the authors don't understand the features they have created themselves would be one.
<nick_l>
infinisil: careless development practices would be another.
ma27 has joined #nixos
<nick_l>
infinisil: deviating from the Haskell standard they have written themselves would be yet another.
<gchristensen>
perhaps we can take a break from this ghc chatter to help lexnoproblem :)
<nick_l>
I also have no time for this stuff.
<lexnoproblem>
thnx gchr!
xcmw has joined #nixos
ma27 has quit [Client Quit]
<nick_l>
You can use GHC if you want; I will even use your programs. Just don't expect me to cheer for something I don't think is very good.
<infinisil>
(I am in #nixos-chat too for offtopic stuff :))
<gchristensen>
two people putting big nix expressions in nix-shell shebangs...
zzamboni has joined #nixos
<nick_l>
gchristensen: guess we found a killer feature.
<LnL>
lexnoproblem: can you show us the output of nix-shell -p nix-info --run nix-info
<gchristensen>
mis-feature imo :P
<gchristensen>
nick_l: nix* commands should fail if the `-E` parameter is > 30 characters ;)
<lexnoproblem>
gchristensen is that non supported feature? looks like desirable functionality imho, as an dev will be happy to distribute only 1 file, with 1) install required nix packages, 2) contain and run actual program script
<nick_l>
gchristensen: I got rid of it, so you have one convert.
<gchristensen>
that is what lex is doing according to the paste
ma27 has joined #nixos
alex`` has quit [Ping timeout: 264 seconds]
<nick_l>
gchristensen: that was not my understanding, but I think it does add to the confusion that there are multiple files in that paste.
zzamboni1 has joined #nixos
zzamboni has quit [Read error: Connection reset by peer]
<lexnoproblem>
my question : 1 ) is -E and -i args are supported in one line shebang for nix-shell 2) if it is not suppoted feature, maybe nix lang lib have function like 'runCommand' so I can specify interpreter from my -E 'expression'
<__monty__>
srhb: There's a couple graphicx packages but it seems like graphicx itself is missing, any idea where to look?
periklis has joined #nixos
<gchristensen>
lexnoproblem: it would be nice if you had a second file here, which contained your expression
<lexnoproblem>
1 min, I will try that
<nick_l>
lexnoproblem: my understanding of the man page is that it is unspecified whether or not 1) should work.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master 9edd4c8 Nikolay Amiantov: gajim: add plugin installer...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to master: https://git.io/vAbky
<nick_l>
lexnoproblem: using ./yourshell.nix which would then contain your expression does work, however.
<nick_l>
lexnoproblem: I would also have expected 1) to work (didn't test it).
alex`` has joined #nixos
<srhb>
__monty__: I recommend using nix-locate graphicx.sty
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/release-18.03 01f9a70 Nikolay Amiantov: gajim: add plugin installer...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] abbradar pushed 1 new commit to release-18.03: https://git.io/vAbkA
<srhb>
__monty__: (Which, through convoluted means, will tell you that it's part of the graphics latex package)
ma27 has quit [Quit: WeeChat 2.0]
<__monty__>
srhb: Thank you, I couldn't run nix-index because I ran out of memory.
<lexnoproblem>
nick_l I expect it to work too, but reality different)
<rawtaz>
hm, i installed astroid which installs webkitgtk - shouldnt that mean that x11 is installed too as a dependency? ther's no `startx` or `x11` on the system
<srhb>
__monty__: :|
<nick_l>
lexnoproblem: sure, but our expectation is not based on any manual, other than "sane API design".
<clever>
rawtaz: just because something is downloaded into the store doesnt mean it gets added to $PATH
<nick_l>
lexnoproblem: Open an issue in the nix repository asking for clarification.
<ottidmes>
Anyone experienced with kexec? I got it working fine in my initrd script, but when I try to run it from a fully booted system, my monitor output freezes and then goes to sleep. I kexec'ed in the same kernel/initrd as I booted into, just to test it out
<clever>
ottidmes: was X running?
<ottidmes>
clever: Probably
xcmw has quit [Ping timeout: 260 seconds]
<clever>
ottidmes: the newly booting kernel expects the GPU to be in text mode, and the video driver is confused
<rawtaz>
clever: indeed. what's the proper way to start x11 on nixos?
<clever>
ottidmes: kexec --console-vga may help, or you need to stop X first
<clever>
rawtaz: services.xserver.enable = true;
<ottidmes>
clever: Aha, since it is a server I shouldn't run X, but thinking about it, I might not have set it explicitly to off, and it probably is on by default
Zoom__ has quit [Quit: Leaving]
<rawtaz>
clever: right, but manually when youre in the terminal?
<gchristensen>
rawtaz: some people do both. I prefer all in the configuration.nix.
<ottidmes>
clever: the netboot-minimal config is self contained right? so it runs from memory?
<clever>
ottidmes: yeah
<ottidmes>
clever: Otherwise I have a crazy plan B, grub4dos with syslinux memdisk :P
<rawtaz>
gchristensen: i guess it's quite rare that you have software you want for just one user, at least on desktop systems, probably more common on servers
<lexnoproblem>
rawtaz depends on number of users, roles, and software ; per user approach can be written in user.nix file too
<rawtaz>
right
periklis has quit [Ping timeout: 240 seconds]
<gchristensen>
rawtaz: for anything I dont want in my global config I use shell.nix and nix-shell
<nick_l>
I have a file allvars.nix which contains {imports= [./foo.nix];} foo.nix contains {a = "hello"}. In yet another file I do "with (import ./allvars.nix)". Why isn't variable "a" available in a let block below that with statement?
<gchristensen>
nick_l: they're probably stuck in the `imports`variable
Rusty1_ has joined #nixos
<lexnoproblem>
rawtaz as for temporary software tests I use nix-env, as for main portable config I use system level configuration.nix with user level config.nix+overlay
<rawtaz>
i'll have to dig into those options a bit to see what suits me. i guess the're explained in the manual
<clever>
nick_l: imports is a special attr in nixos, nix alone doesnt treat it specially
<gchristensen>
for temporary software you'd probably be better off with nix-shell
<clever>
nick_l: so "with (import ./allvars.nix)" creates an attr called imports, that contains the path to foo.nix
<nick_l>
clever: but this was within a nixops evaluation.
<rawtaz>
gchristensen: thanks
<clever>
nick_l: imports only works if the file containing it is within another imports field
<clever>
nick_l: import doesnt obey imports
<ottidmes>
clever: I checked after reboot, it actually does not run X server, as it should, so I am going to try your other suggestion
Guest83256 is now known as TweyII
<lexnoproblem>
sorry to say, I've separeted myprog.py with env.nix in to seperated files, so I could #!nix-shell -E env.nix -i interpreter ,,, but still doesn't work with error: expression does not evaluate to a derivation (or a set or list of those) https://hastebin.com/bexosolisa.py
<nick_l>
clever: so, how do I express "import the combinations of two files"?
TweyII is now known as Guest48865
Guest48865 has quit [Changing host]
Guest48865 has joined #nixos
<clever>
nick_l: cant you just add imports = [ ./allvars.nix ]; to the main nixos config?
<nick_l>
clever: I would then have to do that everywhere.
Guest48865 is now known as TweyII
<nick_l>
clever: I don't want to repeat myself.
<BlessJah>
gchristensen: oh, if it's just PR then it's not a problem
<nick_l>
clever: should I put the paths in an paths.nix file then?
<clever>
nick_l: have you seen the special defaults in nixops?
<nick_l>
clever: doesn't ring a bell
<clever>
nick_l: defaults = { imports = [ ./allvars.nix ]; }; within a nixops deployment file will apply to every machine in the deployment
<BlessJah>
gchristensen: it's a pity that builds reproducibility is not tracked though, with nix it'd very easy to reproduce past builds and compare (granted that they're cached somewhere)
<nick_l>
clever: are you saying that the variable "defaults" is available?
<gchristensen>
BlessJah: it could be done, but features must be prioritized based on my limited available time.
<clever>
nick_l: if you try to create a machine in nixops called defaults, its config will be merged into every other machine in the file
<nick_l>
clever: ok, but that's not applicable to this situation.
<clever>
nick_l: how are you having to repeat things when just adding imports = [ ./allvars.nix ]; to the nixos configs?
MercurialAlchemi has quit [Ping timeout: 260 seconds]
thblt has quit [Remote host closed the connection]
pxc has joined #nixos
Itkovian has joined #nixos
<nick_l>
clever: the problem was that this was only making available the "imports" variable as discussed before.
<nick_l>
clever: not resolve the imports in imports.
<clever>
nick_l: if you add the file to imports (not import), then imports will be searched automatically and recursively
<nick_l>
clever: OK
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<nick_l>
clever: but that doesn't allow me to make the binding available in a let block.
<nick_l>
clever: I can hack around that with a useless attr set, though.
<BlessJah>
gchristensen: is there easy way to compare seemingly identical closures?
<clever>
nick_l: nixos will merge whatever values that sets, and make them available in the config attr passed to every module
<gchristensen>
BlessJah: diffoscope for binaries and nix-diff for derivations are cool. also nix-build --check
<gchristensen>
nick_l: sounds like you've jumped right in to the deep end :)
vidbina has quit [Ping timeout: 268 seconds]
<gchristensen>
I wonder if there is some X-Y going on
rindvieh has quit [Remote host closed the connection]
<nick_l>
clever: ok, so assuming allvars.nix contains the binding foo="hello", I will be able to refer anywhere where config is in scope to config.foo and obtain the value "hello"?
szicari has joined #nixos
<lexnoproblem>
gchristensen, sorry for bothering, I've separeted my nix expression to env.nix, see https://hastebin.com/bexosolisa.py , after that new error occured expression does not evaluate to a derivation . I'm not sure how to build nix expression correctly, but it worked with -E flag before. however not in\with env.nix file.
<clever>
nick_l: only if you also define options.foo to be a valid option using mkOption
<clever>
nick_l: allvars.nix can optionally do that
<nick_l>
clever: are you tricking me into using all the Nix features? ;)
<__monty__>
srhb: Are you sure graphics is the right package? There's both a graphics and a graphicx package for latex normally.
<clever>
nick_l: those are nixos features
<nick_l>
clever: is it? Because I don't necessarily need to use NixOS in the deployment target.
<clever>
nick_l: what are you trying to do exactly?
<nick_l>
clever: currently, I would be happy with the foo option, which I was now figuring out, which should be easy.
<lexnoproblem>
Success! my issue solved for now. I removed -E in my shebang line and changed it to #!nix-shell ./env.nix --pure -I https://github.com/NixOS/nixpkgs-channels/archive/nixos-17.09.tar.gz -i python3.5 now it works\executes perfectly! thank u guyz. So -E flag assumes non-interactive execution, removing it allowed me to execute expression form file and interpteter after
<lexnoproblem>
that's a victory
Thra11 has quit [Read error: Connection reset by peer]
<lexnoproblem>
I wish my web irc client had and emoji support to represent festive fireworks ;`)
<nick_l>
clever: I don't think this is exactly what I am looking for, because I am looking for making available all terraform.nix variables in a Nix file without having to manually create options. Perhaps I can just map over all the attributes to to do mkOption, though.
<srhb>
Oh, though it might just be in scheme-basic I suppose
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] matthewbauer opened pull request #36419: git-sizer: init at 1.0.0 (master...init-git-sizer) https://git.io/vAbmY
NixOS_GitHub has left #nixos [#nixos]
<srhb>
__monty__: Oh, it is. But replacing basic with minimal shows that it is indeed in graphics.
<__monty__>
srhb: I was thinking maybe graphicx overrides part of graphics (the tex packages).
periklis has joined #nixos
zzamboni1 has quit [Read error: Connection reset by peer]
<clever>
nick_l: you can create an option of type attribute set, then nix just lets you stick almost anything in there
<ottidmes>
clever: Thank you very much for the pointer, knowing it was the GPU, after some googling, I solved it by setting nomodeset in my kernel params
<nick_l>
clever: after this modification I now get mymachines.nix 1:1 called without required argument config. .
<clever>
nick_l: this allows most of nixos to just throw things into system.build.whatever, and then later read config.system.build.whatever in another file
<fendor>
how can install a different language pack for libreoffice?
jtojnar has joined #nixos
<nick_l>
clever: when doing nixops info
NixOS_GitHub has joined #nixos
<clever>
nick_l: the config param is at the nixos level, not the nixops level
<fyuuri>
In nixos 17.09 the qutebrowser version is really old. Is there a way to only install single packets from the testing channel?
<ottidmes>
fyuuri: There is, but it will download the whole closure of that package, i.e. its dependencies and their dependencies in turn, so some small package can get pretty big with regards to disk space needed
<fyuuri>
Ok. thank you :): I might just install that using nix-env and override the systemwide old version for the current user
<makefu>
however it actually is possible with nixpkgs to override the version in your nixpkgs with overlays and explicitly importing the `default.nix` from unstable.
<nick_l>
clever: I defined an option, but no way to actually access it.
<makefu>
nick_l: options.services.my-service.enable = lib.mkEnableOption "my-service"; then you can set services.my-service.enable = or use this attribute in your config
<ottidmes>
I would suggest you use unstable/testing if that works for you, but if your package diverges from what is being done upstream, I would suggest the approach of makefu, copying the package over and referring to it in your nixpkgs overlay
<srk>
more of a skeleton
loonquawl has joined #nixos
rindvieh has joined #nixos
ElGoreLoco has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] phile314-fh opened pull request #36421: mongodb: Add authentication support (master...mongodb-auth) https://git.io/vAbst
<the-kenny>
Hm. Anyone any problems with rebar (or any Erlang) stuff recently? Rebar fails to build with `init terminating in do_boot ({cannot get bootfile,start_clean.boot})` but I can't pinpoint why.
<nick_l>
I have rec{imports = ["./tfvars.nix"]; options = {foo = mkOption{..;default = foobar;..}}, but foobar is again not in scope.
<nick_l>
There is a foobar ="bar" binding in tfvars.nix.
<clever>
nick_l: quoting it breaks the path magic in nix
<clever>
nick_l: can you gist all of the involved files?
<nick_l>
clever: quoting? I am not quoting anything.
<ottidmes>
nick_l: "./tfvars.nix"
<mfiano>
What is the difference between fusermount and fusermount3 commands? Both seem to be their own binary and I don't see a different in usage
<nick_l>
clever: although, I don't mind using the option system that much, I really want to just make all the variables from tfvars.nix available.
<clever>
nick_l: you want to just omit the default, tfvars.nix will set the value
mrkgnao has quit [Ping timeout: 260 seconds]
<nick_l>
clever: huh, why would it do that?
xcmw has joined #nixos
<clever>
nick_l: thats what imports does
<clever>
nick_l: also, you need to adjust the type on line 13 of the pastebin some
<clever>
nick_l: try type = types.attrs;
orivej has quit [Ping timeout: 248 seconds]
<nick_l>
clever: but I want the value to really just be a string.
<clever>
nick_l: then tfvars.nix should set it to a string
<nick_l>
clever: that's why I set the value like that (also not the "a" at the end)
<nick_l>
clever: that's why I set the value like that (also note the "a" at the end)
<clever>
then you need to define 2 options, one with the type of attrs, and one with a type of str
<nick_l>
clever: and why I reference "value".
<clever>
and then you refer to the attrs one in the default of the str one
<nick_l>
clever: so, why can't I just reference the value like I am currently doing?
<nick_l>
clever: the logic being "I import the variable", and I reference an attribute of that variable.
<clever>
nick_l: it needs its own mkOption, or nixos will refuse to allow it
vidbina has joined #nixos
pxc has joined #nixos
<clever>
nick_l: also, you have to reference it via config, config.udp_security_group.value;
<nick_l>
clever: that's circular, I am talking about setting the default.
<nick_l>
clever: OK, I think I get it.
<the-kenny>
Okay, now I'm confused. A change to `buildRubyGem` triggers a build error in rebar-2.5.1 where the *only* dependency for is Erlang. Reverting fced35fa44098be0296d8b42166583bd5e505141 fixes the error in the rebar build.
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master e59b91b Vladimír Čunát: nixos/release: fix evaluation if missing x86...
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vAbnf
<clever>
you can set the default value to refer to the second option with type attrs
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-18.03: https://git.io/vAbnI
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/release-18.03 aa84b58 Vladimír Čunát: nixos/release: fix evaluation if missing x86...
<nick_l>
clever: but how can I do this without any options? I am not interested in options at all.
<nick_l>
clever: all I want is to make the .value available to some module.
<clever>
nick_l: define a single option of type attrs, then store everything inside of that
<nick_l>
clever: all I want is to make the .value available to some file*.
<clever>
nick_l: options.stuff = mkOption { ... }; and then in every file, do stuff.thing1 = ... and stuff.thing2 = ...
loonquawl has quit [Quit: loonquawl]
<nick_l>
clever: at that rate it seems easier to just generate a larger .nix file outside of nix.
<clever>
nick_l: nix code can also be used to auto-generate the options, just use map and listToAttrs
<nick_l>
clever: aren't you already doing that in your repository>
<clever>
nope
pxc has quit [Ping timeout: 264 seconds]
<nick_l>
I would love to read an operational semantics of Nix and specifically how this imports thing works in general.
<NixOS_GitHub>
[nixpkgs] matthewbauer opened pull request #36423: emacs: use unstable melpa for dependencies (master...emacs-use-unstable-melpa) https://git.io/vAbcv
NixOS_GitHub has left #nixos [#nixos]
erasmas has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] matthewbauer closed pull request #36345: emacs: give melpa and melpa-stable own scope (master...emacs-melpa-fix) https://git.io/vAQFv
NixOS_GitHub has left #nixos [#nixos]
<nick_l>
clever: that's not how it would appear in a programming languages paper.
<nick_l>
clever: but it seems to "work".
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jfrankenau opened pull request #36424: nixos/i18n: add option for extra locale settings (master...i18n-extra-locale) https://git.io/vAbcl
NixOS_GitHub has left #nixos [#nixos]
lexnoproblem has quit [Quit: Page closed]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] YorikSar opened pull request #36425: Fix fetching SSH keys on GCE (master...gce-keys-fix) https://git.io/vAbc0
NixOS_GitHub has left #nixos [#nixos]
rindvieh has joined #nixos
<infinisil>
nick_l: import is just a function that takes a path to a file and returns the nix value that file contains
<nick_l>
makefu: that's not a minimal module with just a single option.
<infinisil>
nick_l: what's the error?
<nick_l>
infinisil: the attribute udp_* is missing.
<nick_l>
infinisil: in the default assignment.
<clever>
nick_l: can you paste the exact error?
<makefu>
nick_l: minimal would be {lib, ... }: { options.example.enable = lib.mkEnableOption "example"; } but i do not see the point of that
<infinisil>
nick_l: remove the rec
<infinisil>
nick_l: your config is actually the one defined at the bottom
<clever>
nick_l: aha, yeah, the rec is to blame
<infinisil>
not the argument at the top
<nick_l>
pastebin.com/jNmtaxXM
<nick_l>
How can rec make things worse?
<infinisil>
nick_l: do you know how rec works?
<clever>
nick_l: the config on line 15 refers to line 24, which is an empty set
<nick_l>
I thought it could never do that in that it would only allow more behaviour.
<nick_l>
clever: it's supposed to be like letrec.
<infinisil>
nick_l: rec has very clear semantics, it doesn't just "allow more behaviour"
<nick_l>
What's the point of the config = {} block anyway?
<infinisil>
nick_l: you don't need it
<makefu>
nick_l: config can actually change your host configuration
<tilpner>
Hey garbas, first time trying to use pypi2nix, got errors and error reporting failed with 414 Request-URI Too Large. Any log I can send?
<nick_l>
infinisil: why would someone want to have it?
<nick_l>
OK, so it finally compiles.
xcmw has quit [Ping timeout: 264 seconds]
<infinisil>
nick_l: what makefu said, it's actually the thing that sets options
cement has joined #nixos
clefru has quit [Quit: Leaving]
<garbas>
tilpner: right, i knew we are going to hit that at some point :)
<garbas>
tilpner: can you rerun the command with -v flag and copy the output
<ottidmes>
nick_l: https://nixos.wiki/wiki/NixOS_Modules I use config in a module without options sometimes when I need mkMerge or want the config to be optional (via mkIf)
<tilpner>
garbas - Oh, I was hoping there was some temporary log location. I don't want to rerun it just yet, it takes too long. I'll try running it on the failed subpackages though
<garbas>
tilpner: also provide your pypi2nix version (pypi2nix --version)
<tilpner>
This was with 1.8.1
<tilpner>
"Failed to build Pillow cryptography cffi"
<garbas>
tilpner: if you manage to drill down the problem that would be good also and then report a smaller problem
<tazjin>
I'm having the displeasure of trying to build a Go project & link it statically in a derivation, does anyone here know how to get `-tags netgo` to work in Nix?
<tilpner>
I just tried 'pypi2nix -V "3.6" -e Pillow'. That gives the same error (414), am I using it wrong?
<tazjin>
the brilliantly designed go compiler does this thing where it tries to overwrite files in its install location by default ...
<garbas>
tilpner: does that help you?
cinimod has joined #nixos
<tilpner>
garbas - Ah, copying the extraBuildInputs makes Pillow build, but that was not why I pinged you
* tilpner
checks for pypi2nix issues mentioning the 414, which he should have done earlier :/
* tilpner
didn't find any
__monty__ has quit [Remote host closed the connection]
<garbas>
tilpner: please open an issue. not to forget about it
<nick_l>
"The option `accessKeyId' defined in vars.nix does not exist. ". How can it not exist if I have it defined there? There is no rec keyword in sight this time.
<ottidmes>
If you want to use modeset=0, do you really have to specify the GPU driver? I see radeon.modeset=0, noveau.modeset=0, and i915.modeset=0, but does modeset=0 without a prefix also set these? I now have nomodeset, but I read that it is outdated and modeset=0 should be used instead
<cinimod>
I am trying to install sundials via nix - in the sundials documentation it says you install the software by % cmake -DCMAKE_INSTALL_PREFIX=/home/myname/sundials/instdir -DEXAMPLES_INSTALL_PATH=/home/myname/sundials/instdir/examples /home/myname/sundials/srcdir
<cinimod>
Clearly at the very least I need to change /home/myname but is there some standard nix variable I should use? $out?
logzet has quit [Ping timeout: 245 seconds]
<cinimod>
Searching for cmake in the nixpkgs doc didn't reveal anything
logzet has joined #nixos
davidlt has quit [Remote host closed the connection]
<ottidmes>
cinimod: You should search nixpkgs itself instead, I found a relevant example: export cmakeFlags="-DCMAKE_INSTALL_PREFIX=$prefix/opt $cmakeFlags" (not sure what $prefix, I would have expected $out, but maybe that var is not yet available in preConfigure)
alex`` has quit [Ping timeout: 260 seconds]
<cinimod>
I did search nixpkgs but there are too many examples :)
<cinimod>
It's not clear which to copy
<ottidmes>
cinimod: Seems $out is available (I saw it in other packages at that point in the config)
<ottidmes>
cinimod: I would guess, just keep it as close as the docs of sundials says, so just do the export, re-export the existing cmakeFlags and add your own where everything before instdir would be $out
<nick_l>
Where does nixops even get that it's an "option".
davidlt has joined #nixos
<nick_l>
I certainly don't want it to be an "option".
<clever>
nick_l: why are you using inherit then?
<nick_l>
clever: I just want x = x;
dnovosel has joined #nixos
<nick_l>
clever: which is inherit x; afaik
<clever>
nick_l: that is setting a config entry called data_subnet, so there must be a matching options.data_subnet to define its type
<nick_l>
clever: all I want is to import a bunch of bindings as if they are defined in a let block.
<nick_l>
clever: that's all. Is that really too much to ask of Nixops?
knupfer has joined #nixos
<ottidmes>
nick_l: That is unrelated to Nixops, more a lack of understanding how NixOS works
<nick_l>
So, how can I get this to work?
alex`` has joined #nixos
<ottidmes>
nick_l: You want to use these attrs elsewhere in your config?
ElGoreLoco has quit [Quit: ElGoreLoco]
<nick_l>
ottidmes: yes
<ottidmes>
nick_l: Outside of the scope of this current file, i.e. other imported modules?
<nick_l>
ottidmes: yes
<nick_l>
ottidmes: and without indirections.
<ottidmes>
nick_l: Then you will have to define a module that defines these attributes (options), but if you want it to be just an untyped attrset, you will still have to define an option to be an attrset and then you can define whatever in that an access it via that option name: config.my-stuff.<random-attr-you've-set>
lexnoproblem has joined #nixos
<nick_l>
ottidmes: that's still an indirection.
<nick_l>
Why would I not just generate some nix code via some trivial script then instead?
<nick_l>
That at least has simple to understand semantics.
<ottidmes>
nick_l: That's subjective
<nick_l>
ottidmes: I would love to see a module which allows injection of arbitrary amounts of variables.
<nick_l>
ottidmes: then there is something to compare.
<ottidmes>
nick_l: What you might consider simple, I might not, but either way, the semantics of the NixOS module system is that you cannot define arbitrary config attributes
<jwynn6>
anyone have any ideas why i get a good install but then a bootloop that never makes it to grub? I'm running zfs, a grub mdadm array on an older HP z600.
<ottidmes>
nick_l: Import that, and you can define go nuts with, attrs.foo = 5; and then access it elsewhere with config.attrs.foo
<clever>
jwynn6: efi or legacy booting?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] mt-caret opened pull request #36428: satysfi: init at 2018-03-07 (master...satysfi) https://git.io/vAbRx
NixOS_GitHub has left #nixos [#nixos]
<nick_l>
ottidmes: that solution still has more indirections then just compiling a big let block into every nix file generated by some trivial templating engine.
pxc has joined #nixos
<nick_l>
ottidmes: but thank you.
Lisanna has joined #nixos
<jwynn6>
clever: legacy booting
<clever>
jwynn6: what is boot.loader.grub.device(s) set to?
<jwynn6>
clever: i made sure the mbr boot flag is set, the drives are gpt
<lexnoproblem>
stupid notice of (imho) lack of proper documentation: when invoking nix-shell it is unclear what kind of nix expressions it expect to get and what kind of return value it expects to receive from that nix expression (if it expects a return value at all). I mean 1+1 is valid nix expression to, is it fit to be invoked with nix-shell? unclear
<jwynn6>
clever: ["/dev/sdb" "/dev/sda"]
<lexnoproblem>
*too
<clever>
jwynn6: and both those drives have bios boot partitions that are not mounted/formated?
<ottidmes>
nick_l: Whell I would consider introducing a template engine to be quite an indirection, and I do not think that approach would end up simplifying things. This is a clear case of the difference between simple and easy, that solution might be easy, but it could get complicated quickly, figuring out how everything interacts
<jwynn6>
yup. nixos-install gives me good returns on both
<jwynn6>
clever: ^
mounty has quit [Ping timeout: 260 seconds]
<ottidmes>
lexnoproblem: I believe you could do that with nix-instantiate, while nix-repl and nix-build expect expressions reducing to a derivation (low level description of a package)
<clever>
jwynn6: if you run something like this against the bios boot partition, do you see a string like 00000120 6e 67 00 2e 00 0d 0a 00 47 65 6f 6d 00 52 65 61 |ng......Geom.Rea|
<ottidmes>
lexnoproblem: I mean, the error pretty much explains that: nix-shell -E '1+1' gives error: expression does not evaluate to a derivation (or a set or list of those)
<ottidmes>
lexnoproblem: As you can see in clever's example, with --eval you can run an expression that does not evaluate to a deriviation, without it, it will complain with the same error. Both nix-shell and nix-build do not have this
<clever>
jwynn6: the strings command is part of binutils, but i wouldnt run it on a partition
<jwynn6>
clever: not sure how to get the output you're after..
<clever>
jwynn6: what partition is the bios boot partition?
<jwynn6>
oh i misse a message
<jwynn6>
/dev/sda2 or /dev/sdb2
<ottidmes>
nick_l: Ow, if you really want to just have some variables conceptually in a let block above certain files, why not just do: with (import ./my-vars.nix);
<clever>
jwynn6: hexdump -C /dev/sda2 | head -n30
<clever>
jwynn6: at offset 0x120, do you see an the text from grub?
<clever>
jwynn6: yeah, that looks like a normal grub stage 2
leat has joined #nixos
<lexnoproblem>
ok when I nix-shell -E '1+1' error: expression does not evaluate to a derivation (or a set or list of those) - it clearly tells what it want, i'm ok with it. I know stdenv.mkDerivation . but for example I invoke runCommand and it evaluates to correct derrivation. But return value of derivation is unclear from runCommand syntax. the only way new user could grasp that fact, not from official doc,
<jwynn6>
clever: it really seems like it should work fine, but it feels like grub never loads and it just forces the system to restart.
endformationage has joined #nixos
<clever>
jwynn6: is legacy booting enabled in the bios?
<jwynn6>
clever: i'll doublecheck but i thought so
mounty has joined #nixos
s33se has joined #nixos
s33se has quit [Client Quit]
s33se has joined #nixos
kitttn has joined #nixos
rogue_koder has joined #nixos
<kitttn>
hello, guys, a question. I am trying to make my first package. After i do nix-shell myfile.nix, dependencies are downloaded and i enter the shell. After i do the unpack phase, which is just extracting src=fetchFromGitHub, the archive is being inflated to my directory, that's ok. But i can't remove that inflated directory! Permission is denied, and the rights are read (execute)
shikiryogy has quit [Quit: Leaving]
<kitttn>
so, how can i remove that dir?
MinceR has quit [Ping timeout: 240 seconds]
logotags has quit [Ping timeout: 256 seconds]
<clever>
kitttn: chmod -R +w thatdir
<jwynn6>
clever: I'm gonna go with yes, legacy boot is enabled. It's a bios from 2010.
<lexnoproblem>
show output of "ls -la; whoami ; groups;" from your work dir
<clever>
jwynn6: ah, then it should just work, which drive is it set to boot from?
MinceR has joined #nixos
<clever>
jwynn6: try using something like f12 to manually boot from each drive
<kitttn>
@clever thanks. sudo rm -rf also worked, i was just so surprised that i can't remove dirs from my home dir, sorry :/
<jwynn6>
clever: sda
<kitttn>
also a question
marusich has quit [Quit: Leaving]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dezgeg pushed 1 new commit to unstable-aarch64: https://git.io/vAbz1
<NixOS_GitHub>
nixpkgs/unstable-aarch64 42bb9eb Tuomas Tynkkynen: nixos/qemu-vm: Add virtualized display + HID devices on AArch64
NixOS_GitHub has left #nixos [#nixos]
<clever>
kitttn: its more that the directory lacks the +w bit by default, and even though you have permission to give yourself write, rm -rf cant delete it
<kitttn>
@clever good to know, thanks!
<clever>
kitttn: also, fetchFromGitHub does the unpacking in this case, and unpackPhase is just a recursive cp
logotags has joined #nixos
<clever>
kitttn: if you only want to read the source, you could just ls -lh $src
<kitttn>
another problem. i am trying to override $out to test installPhase in nix-shell; using `nix-shell --command "export out=~/some/writable/dir; return" <myfile.nix>; `
<kitttn>
after i enter the shell, echo $out shows that dir
<jwynn6>
clever: nope, didn't work, just tried manually selecting each drive and still in the boot loop. could i be something about using an mdadm array to mirror the boot partitions?
<kitttn>
but installPhase still points to /nix/store
<clever>
jwynn6: if grub was unable to read /boot, it would drop into a rescue shell
spear2 has joined #nixos
<kitttn>
or i can just do nix-build, and it will not break my system?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master eab479a Vladimír Čunát: nixos release: more evaluation fixes...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to master: https://git.io/vAbgt
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/release-18.03 82bb0a1 Vladimír Čunát: nixos release: more evaluation fixes...
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to release-18.03: https://git.io/vAbgm
NixOS_GitHub has left #nixos [#nixos]
<clever>
kitttn: just do nix-build, it cant break things
<clever>
kitttn: every $out is unique, based on the inputs, so it will never be able to change a path that is already in use elsewhere
hotfuzz_ has joined #nixos
knupfer has joined #nixos
<kitttn>
yeah, i am just using the same inputs while editing and testing the same .nix file
greglearns has joined #nixos
<clever>
the strings in the .nix file are part of that hash
<kitttn>
wow, didn't understood it, thought about like inputs for functions, like arguments, thanks!
<kitttn>
will try with nix-build then. It makes generalBuild() (or smth like that) as i understood, which includes installPhase, right?
<clever>
kitttn: the buildInputs are flattened down to a simple string, and that string is what gets hashed
<kitttn>
and until i mark the package with nix-env -i, it will be still collected as garbage, right?
hotfuzz has quit [Ping timeout: 240 seconds]
<clever>
kitttn: yeah, the default is to run genericBuild, which handles unpack, patch, configure, make, and install
<greglearns>
for NixOps, there is something wrong with nixos-17.09.2681.59661f21be6-x86_64-hvm-ebs (ami-3eb40346) image (?). The `ebsInitialRootDiskSize = 20` (20 GB) but the instance only gets a 2GB main partition:
<greglearns>
[root@onebox:~]# df -H │latform_name":"bigcommerce1","username":"cdxbl1hxdew9dpw26g43s32rc Filesystem Size Used Avail Use% Mounted on │ndk6hk"}} devtmpfs 404M 0 404M 0% /dev
<clever>
kitttn: the last result nix-build made will be pointed to by a result symlink, which protects it from GC
<clever>
kitttn: anything older then that is garbage and can be deleted
greglearns has quit [Client Quit]
<clever>
kitttn: but if you perfectly undo a change to the .nix file, the hash will match up, and nix will reuse some of the "garbage", making it not-garbage again
greglearns has joined #nixos
<greglearns>
a
<greglearns>
for example: /dev/disk/by-label/nixos 2.1G 2.0G 0 100% /
<clever>
greglearns: what does fdisk say about the drive?, and did you change ebsInitialRootDiskSize before or after the deployment?
<kitttn>
so assuming i am making 10 unsuccessful builds and 1 last successful, i will make 10 garbage nix-store links, right?
<greglearns>
even though NixOps IS correctly creating a 20GB EBS volume.
<greglearns>
Who do I tell about this nixops / ami problem?
<clever>
kitttn: any build that fails will automatically be garbage
<kitttn>
ah, that is good)
digitus has joined #nixos
<clever>
kitttn: and the garbage collector will delete those failed builds first
<greglearns>
I changed ebsInitialRootDiskSize before the initial deploy. (and even tried a few times)
<kitttn>
thanks, will continue to building then :)
<clever>
kitttn: so if you do `nix-collect-garbage --max-freed 10M` it will delete 10mb worth of garbage, starting with the failed builds
<kitttn>
if anything goes wrong, i will be back, thanks for help :)
<clever>
kitttn: using small GC's like that lets you get X mb free, without destroying too much and loosing your dl cache
<kitttn>
will try it, thank you!
<clever>
greglearns: what does fdisk say about the root drive?
<greglearns>
clever: Disk /dev/nvme0n1: 20 GiB, 21474836480 bytes, 41943040 sectors Units: sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disklabel type: dos Disk identifier: 0x373af4ea Device Boot Start End Sectors Size Id Type /dev/nvme0n1p1 2048 4194303 4192256 2G 83 Linux
<clever>
greglearns: i heard something about the auto-resize not supporting nvme drives at the moment
<greglearns>
clever: Disk /dev/nvme0n1: 20 GiB, BUT notice the "/dev/nvme0n1p1 .. 2G"
<greglearns>
what is the auto-resize?
<clever>
greglearns: the ami contains a 2gig disk image, on the first bootup, it has to auto-resize its own partitions and the FS within
<clever>
greglearns: that doesnt work when using an nvme drive
<clever>
greglearns: the fix is only going to be in the 18.03 channel, when it comes out later this month
<greglearns>
oh.
orivej has joined #nixos
<clever>
i even see your username on the comments there! heh
<greglearns>
clever: because I did notice that the deploy says it is using 17.xx
<greglearns>
clever: yeah, I made that comment, but then I never heard from anyone.
<greglearns>
clever: any idea who I should poke? do people know about the problem? because I made a comment on the commit, but that seems like the wrong place.
<clever>
greglearns: comments on commits are very easy to miss
<clever>
greglearns: its best to open an issue within nixpkgs
<greglearns>
I've never opened an issue on nixpkgs... any recommendations for how to make sure the nixops people see it? language I should use?
<clever>
greglearns: hmmm, maybe the nixops repo would be better then, so the nixops people see it right away
<cinimod>
The sundial docs say -DCMAKE_INSTALL_PREFIX=/home/myname/sundials/instdir
<cinimod>
So what should I put for /home/myname?
<ottidmes>
cinimod: I do not know the program behaves, it cannot write anything to the /nix/store, so if it tries to write state to it, it will fail, but if it is just the location where it installs the binaries and such, you would use simply: ''-DCMAKE_INSTALL_PREFIX=''${out}'', no need for this instdir, your $out is the installation directory (here I assume instdir == installation directory)
wangoe has joined #nixos
<cinimod>
Yes it's the installation directory
<ottidmes>
cinimod: And I remember you also had an examples directory, if you want to keep that, it should probably go to $out/share/examples
<mbrgm>
is it possible to use string interpolation/variable substitution in a `path` type option?
<ottidmes>
clever: BTW, I am looking at writing a custom initrd right now, any reason to go with Haskell and not just a bash script? Other than, well... because Haskell :P
<ottidmes>
mbrgm: Yeah, you can just use + and strings, just make sure the leftmost element is of type path
<clever>
ottidmes: mostly just preference and if you want to see errors at compile-time or boot time, lol
pkill9 has joined #nixos
<ottidmes>
mbrgm: An example from my config: <nixcfg> + "/${hostName}/config/openssh/root/id_ed25519.pub"
<clever>
ottidmes: you can also do <nixcfg/${hostName}/config/openssh/root/id_ed25519.pub> i think
<ottidmes>
clever: That alone is indeed reason enough! I have had it happen to me plenty of times that I had to retry because of some typo or refactoring that was incomplete
oleks has quit [Remote host closed the connection]
zzamboni has quit [Quit: Leaving.]
<mbrgm>
ottidmes: thx! the leading slash in the "/${..." feels a bit odd though
<kitttn>
Hey, guys. I am trying to install an app, which is being cloned from github, and then after running it, continue to installing itself. The problem is it is using bash, and trying to access it via #!/bin/bash
<ottidmes>
clever: let foo = "nixos"; in <nixpkgs/${foo}> fails as well, I just tried in nix-repl
<ottidmes>
mbrgm: There is, pkgs.substituteAll on the top of my head
pxc has joined #nixos
<kitttn>
infinisil: so i anyway should change some bash scripts to remove that /bin/bash at the beginning?
<mbrgm>
i.e. I have a file that I can turn to a string with `fileContents`, but I want to have interpolation in it?
<clever>
kitttn: substituteAll only works for tokens @wrapped@ with @'s
<ottidmes>
mbrgm: There is also substituteInPlace (but I think that is only available as part of your nix-build script)
<mbrgm>
ottidmes: ok, I'll check that later. thx so far!
greglearns has quit [Ping timeout: 260 seconds]
<infinisil>
kitttn: Ah yeah, they seem to use just a simple bash script, you could just change it manually from #!/bin/bash to #!/usr/bin/env bash
<infinisil>
Well the installation process probably won't work on nixos without some effort though
<infinisil>
oh and the bash script is used for the installation
<ottidmes>
kitttn: You probably want to use patchShebangs
<ottidmes>
kitttn: But I have not read all your messages, so might be unrelated to your question
<kitttn>
infinisil: ottidmes: thank you for help! Will try and play a bit with installation :)
<cinimod>
clever: I finally understood what you were telling me
sonarpulse has joined #nixos
<ottidmes>
kitttn: BTW, if you do not already know, if you just want to run e.g. bash via #! in some script in NixOS you can just use #!/usr/bin/env bash, but if you plan to package it you should probably use #!${stdenv.shell} or #!${bash}/bin/bash
<infinisil>
kitttn: Having checked out flutter's repo, they seem to be using a bash script in place of a proper makefile
ryanartecona has quit [Quit: ryanartecona]
<kitttn>
was guessing about this one: #!${bash}/bin/bash , but thanks for pointing that out!
<infinisil>
Oh no, and they are fetching the dart compiler itself in that script and then use that
__Sander__ has quit [Quit: Konversation terminated!]
<kitttn>
yeah, and it doesn't work because need libPath, as i understood :)
<infinisil>
Oh no, it's even their own version of dart as far as i can tell
ryantm has quit [Ping timeout: 248 seconds]
alex`` has quit [Quit: WeeChat 2.0.1]
<kitttn>
infinisil: also they require cloning the repo, not just fetching some revision. Any ideas how to do it using src variable in mkDerivation?
<kitttn>
infinisil: yes, it is their library, you are right
<infinisil>
kitttn: yeah you can use fetchFromGitHub and set deepClone = true, which is slower, but keeps the .git directory
<kitttn>
wow, thank you!
<infinisil>
it's most likely going to be annoying to make this work, especially with the nix sandbox
<kitttn>
infinisil: i guess, it is not the easiest first package to build on Nix, right? :D
<infinisil>
Yeah, it's definitely one of the harder ones to package
<infinisil>
Properly at least
<ottidmes>
kitttn: If there is a deb package out there for it, it might be easier to take that, and patch it
<kitttn>
ottidmes: no, there is no deb package; The only way to install is via gh + their script :)
<kitttn>
infinisil: at least, after i build it (if i will) other packages would be far more easier to build :)
Myrl-saki has joined #nixos
<infinisil>
kitttn: Yeah
mrkgnao has joined #nixos
<infinisil>
There's buildFHSUserEnv which would make it simple but not nice
vidbina has quit [Ping timeout: 256 seconds]
<ottidmes>
kitttn: One improper way to package it would be have it be two phases, first phase is outside of nix, runs that scripts that downloads the deps and puts them in the nix store as sources, and then the second phase would be the normal build process that now can be done with the sources in place
<ottidmes>
infinisil: But I doubt that would circumvent the sandboxing, or does it?
Tucky has quit [Quit: WeeChat 2.0.1]
<infinisil>
ottidmes: true it wouldn't
ryantm has joined #nixos
<infinisil>
But maybe it's not too bad actually, i think their dart version is the normal dart, not something special
<infinisil>
Then it's not too hard to translate the script to nix
<ottidmes>
I am doing that two phase trick for a package that requires to mount a disk image
<kitttn>
infinisil: there is even a dart package in nix repo already
<infinisil>
Yeah
<kitttn>
ottidmes: wow, running that 2-step build is super-cool idea! A bit pity that i don't have my Arch :)
<infinisil>
To have the package in nixpkgs though (where everybody could profit from it), you'd need to package it properly
<mpickering>
On the last stage of a build, the package symlinks a bunch of stuff into the output destination, is this ok to do?
jmeredith has joined #nixos
<kitttn>
infinisil: i guess, it is to early to speak abut packaging at this moment :)
<kitttn>
but i will definitely read (and ask) how to package properly
pkill9 has quit [Ping timeout: 248 seconds]
asuryawanshi has joined #nixos
<ottidmes>
mpickering: Should be OK, as long as they do not point outside the package
<cinimod>
I have no idea what nas-1.9.4 or why it appears
<cinimod>
And googling did not turn up much
endformationage has quit [Quit: WeeChat 1.9.1]
<cinimod>
It seems to be an entirely unrelated package
asuryawanshi has quit [Remote host closed the connection]
<infinisil>
mpickering: Well that's how dependencies work, what stuff is it?
asuryawanshi has joined #nixos
<mpickering>
A bunch of Qt libraries and so on
<ottidmes>
kitttn: That disk image package does actually do the following, have a Nix package that generates the script that generates the disk image. That script once run calls nix-build to build a package based on its inputs to be used as its dependencies, and that then produces the disk image
<mpickering>
I went through this morning manually trying to patch the dylib and executable to write in the right paths but seemed like a long road
<mpickering>
If they point outside the package, is that ok or not?
jensens has quit [Ping timeout: 276 seconds]
<infinisil>
mpickering: Yeah that might be something for qt to work properly, lots of packages have something in its $out/lib folder
<gchristensen>
sphalerit / sphalerite you seem to be a bit ... misspelled ...
<srhb>
truncerite
<gchristensen>
hah
<infinisil>
mpickering: With nix, if something in $out refers to another /nix/store path then this becomes a dependency of it, nothing bad about that
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jtojnar has quit [Quit: jtojnar]
aarvar has joined #nixos
fendor has quit [Quit: Leaving]
<ottidmes>
cinimod: I always use Arch Linux to base my package on, it never fails to deliver for me, I have been surprised multiple times to seem some obscure software being packaged on Arch Linux
<ottidmes>
cinimod: Ah, yeah, I have been told when I first started to package to just always modify the sha if you do not know it, to prevent a clash, so I never actually ran into that problem before :P
<nick_l>
cransom: then I don't really get why anyone is excited about NixOS in the cloud.
<sphalerite>
hyper_ch: sort of
<nick_l>
cransom: it seems utterly useless.
<cransom>
nick_l: ok?
Fare has joined #nixos
<hyper_ch>
sphalerite: you had issues with qemu, zfs and available ram right?
<sphalerite>
yeah, but the arc_max thing fixed it
<hyper_ch>
sphalerite: boot.kernelParams = [ "zfs.zfs_arc_max=6442450944" ]; didn't know you set it also
<sphalerite>
hyper_ch: I don't want to set it permanently though, just one-off for starting the VM (which won't be running most of the time)
blankhart has joined #nixos
<sphalerite>
in fact I prefer the `head -c 6G /dev/zero | sort > /dev/null` option because that doesn't even persist within that boot :D
mounty has joined #nixos
Myrl-saki has joined #nixos
<cransom>
nick_l: the only other references i see for ec2-net-utils is people having ported it to other distros 4 years ago. if that is a necessary feature (its just hot plug rules and route table setup, i don't know why you need that specific tool set), you should keep on with amazonlinux
<blankhart>
let's say i want to understand pkgs.runCommand. where is this documented? it is used but not explained in the Nix and Nixpkgs manuals
<clever>
sphalerite: thats very similar to `echo 1 > /proc/sys/vm/drop_caches`
<hyper_ch>
sphalerite: how do you run that automagically before starting a qemu vm?
<sphalerite>
clever: but I don't want to drop everything, just enough to start the VM
<sphalerite>
hyper_ch: I could use a wrapper script, but in practice I just run it manually since it's not so frequent
shlevy has joined #nixos
<sphalerite>
^R6G enter
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] maximedenes opened pull request #36438: pythonPackages.antlr4-python3-runtime: init at 4.7.1 (master...antlr-python) https://git.io/vAbyk
NixOS_GitHub has left #nixos [#nixos]
<clever>
sphalerite: i find that 64gig of swap makes the kernel just not complain
<nick_l>
cransom: it's just that I have this almost complete NixOps setup, but there is just this tiny detail that I got wrong.
<clever>
it will let me allocate ram i dont yet have, and then zfs mostly gives up blocks before the swap can kick in
<hyper_ch>
if I'd use 64gb swap now I'd be down to 50GB free disk space
xcmw has quit [Ping timeout: 264 seconds]
<blankhart>
what does "rec" do? it isnt really possible to search for this in the manuals; there are hundreds of matches of 'rec'
<cransom>
nick_l: what would ec2-net-utils fix?
<gchristensen>
blankhart: it makes this work: { foo = "bar"; baz = "${foo}"; }
<nick_l>
cransom: its feature set would allow me to call AssignPrivateAddresses at run-time.
* hyper_ch
went outside today and he say many, many clouds
<sphalerite>
clever: not sure I have that space to spare outside my zfs, and I've heard bad things about swap files on zfs
<cransom>
blankhart: what gchristensen said, short for 'resursive'. you can lookup variables in same scope you are in.
<pierron>
shlevy: What do you mean? building from source for nixpkgs-mozilla?
<clever>
sphalerite: thats why i have a dedicated swap partition, outside zfs
<dmj`>
Anyone use nix with npm?
<hyper_ch>
sphalerite: use a seperate partition for swap
<pierron>
shlevy: I would propbably not fix it, but I would accept contributions.
<clever>
sphalerite: and yeah, zfs cant be shrank, so such a change requires a reinstall, my laptop is currently stuck with a 40mb swap
<sphalerite>
clever hyper_ch: I'd prefer using that space for my zfs! :p
<hyper_ch>
sphalerite me too :)
<hyper_ch>
clever: you have 40mb more swap than me
<pierron>
shlevy: I had managed to build rustc locally at one point, but it took me a while to figure out all the command line arguments to get it working :/
<clever>
hyper_ch: i forgot about units when i configured justdoit.nix to make a 40gig swap, lol
<pierron>
shlevy: since, I remove my /home by accident.
<hyper_ch>
clever: lol
<hyper_ch>
clever: btw, you're already up again or didn't go to bed yet?
<LnL>
that's always a mystery with clever :p
<gchristensen>
we should make a clevertracker
<cransom>
nick_l: i don't see any reason you couldn't script that yourself. you wouldn't do it on box because follow nixops deploys would blow it away. you'd make api calls, update the deployment, deploy.
<NixOS_GitHub>
[nixpkgs] dywedir opened pull request #36440: iwd: 2017-12-14 -> 0.1 (master...iwd) https://git.io/vAbS3
NixOS_GitHub has left #nixos [#nixos]
<sphalerite>
grep -h clever\t ~/.weechat/logs/irc.freenode.#nix* | sort | cut -d $'\t' -f 1 | ask gchristensen for some fancy thing that can visualise event dates
<gchristensen>
hah
zzamboni has joined #nixos
shlevy_ has joined #nixos
<gchristensen>
this ... would be not so difficult ....
<shlevy>
pierron: I mean building rustc in mainline nixpkgs segfaults during tests :)
<gchristensen>
for his privacy though I won't
<shlevy>
pierron: Against glibc 2.27
shlevy_ has quit [Client Quit]
<nick_l>
cransom: that resource is managed outside of NixOps, so that's not a problem.
noam has quit [Ping timeout: 252 seconds]
<nick_l>
cransom: it is a problem that even in the web-interface of Amazon I just get back eni-<some numbers>: secondary-addresses in red. And that's all.
<nick_l>
cransom: so, apparently it was too much for the dear people from Amazon to write a full sentence about what is wrong.
<fyuuri>
Is there a package like texlive-full in Nix?
<gchristensen>
texlive.combined.scheme-full
<sphalerite>
fyuuri: texlive.combined.scheme-*
TweyII has quit [Ping timeout: 245 seconds]
<fyuuri>
thx :). But why the scheme?
tmaekawa has joined #nixos
zzamboni has quit [Quit: Leaving.]
* gchristensen
has no idea
<sphalerite>
I think those are actually the TeXLive names. Not sure though
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mog>
can anyone help me with some udev sillyness? i am trying to get a rule to fire when i add or remove a bluetooth keyboard but im not seeing it match on my rule
<mog>
clever, i made that change to call it directly and rebuild switch complained "/home/mog/bton.sh is called in udev rules but not installed by udev"
<clever>
mog: ah, i think there was something to auto-detect running things not in the default PATH
<clever>
it may be falsely complaining about absolute paths
<Dezgeg>
maybe NAME is a special variable then as well? so not ATTRS{name}=="foo" or ENV{NAME}=="foo" but NAME=="foo"
<mog>
yes thats where i would expect to see errors about not being able to cat to a file or something, but im just seeing it add the keyboard to xinput
<mog>
i dont think any udev rule ive tried has matched yet
Fare has quit [Ping timeout: 248 seconds]
<clever>
mog: ok, run this, then reconnect again: udevadm monitor
<clever>
blankhart: how big is "du --max=0 shared server -c -h" ?
<infinisil>
well damn, i did not know this
noam has quit [Read error: Connection reset by peer]
noam has joined #nixos
andersk has quit [Remote host closed the connection]
<infinisil>
Huh, so let in bindings are fully bidirectional
<blankhart>
big. maybe because i have old .stack-work directories in there
<logzet>
Hi, are the KDE pim DBus interfaces packaged anywhere?
<logzet>
I am trying to package an application that requires KF5KdepimDBusInterfaces in its CMakeLists.txt
coconnor has quit [Remote host closed the connection]
<ottidmes>
cinimod: Ah your back, if you want, I have it, I have packaged sundials for you
cinimod has quit [Ping timeout: 240 seconds]
davidlt has quit [Ping timeout: 252 seconds]
<infinisil>
clever: I always thought `let a=1; b=2; in` was just sugar for `let a=1; in let b=2; in` but apparently it's desugared to something like `with rec { a=1; b=2; };`
<infinisil>
let in bindings are a lie
<clever>
infinisil: thats likely also why inherit works inside there
<infinisil>
clever: Actually I think it might be possible as of recently, trying to find the PR
jtojnar has quit [Ping timeout: 240 seconds]
<fyuuri>
should I return the state to the version of my installation medium? I installed a bunch of stuff in the meantime though
<fyuuri>
It was a 17.03 medium
<clever>
fyuuri: if nothing has broken, then your probably just not using the software that is likely to break, so the new version is probably also good
MichaelRaskin has joined #nixos
<fyuuri>
Everything works perfectly
<clever>
then you can just leave it at the current value
<fyuuri>
I did the channel upgrade as soon as I installed the system
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jerith666 opened pull request #36447: thunderbird: fix typo in desktop item genericName (master...thunderbird-typo) https://git.io/vAbpu
<blankhart>
what does this mean? substituter 'https://cache.dmj.io' does not have a valid signature for path '/nix/store/l713q53x0h6lngb58g85qrsh9rr8n3fb-source'
<fyuuri>
thx!
<blankhart>
specifically why is it a warning rather than an error
<blankhart>
(it is the only output message that seems to indicate why nix can't build a dependency)
<clever>
blankhart: did you add his public key to your nix.conf file?
<MichaelRaskin>
It is not an error, because a local build could in principle be attempted
<infinisil>
Found it
<blankhart>
yes
<clever>
blankhart: what is the contents of your nix.conf, and what output does nix give when ran?
<infinisil>
clever: Something I don't get: "Note that using a distinct version of Nixpkgs with NixOS may be an unexpected source of problems" Why are these different? Shouldn't everything use this pkgs set in this option?
pxc has quit [Ping timeout: 240 seconds]
<clever>
infinisil: i think thats more about the services from <nixpkgs/nixos> expecting version A and your custom pkgs set have version B, and now everything breaks
<blankhart>
maybe that is a question for dmj
<clever>
infinisil: and yeah, that attribute has no way to change service definitions
<clever>
blankhart: yeah, his docs suck, he never made the public key public!
<infinisil>
clever: Oh right..
simukis has quit [Quit: simukis]
<clever>
infinisil: i can only see it being useful in nixops, when you want to force all 10 machines to share a single instance of nixpkgs, so nix doesnt have to eval the same version 10 times
<infinisil>
that would indeed be a valid usecase for me
<clever>
infinisil: the memoize builtin also fixes that
<blankhart>
thanks all! i will attempt a local build for now
vcunat1 has joined #nixos
vcunat1 has quit [Client Quit]
vcunat1 has joined #nixos
<infinisil>
clever: So does nixops actually evaluate each machine's config individually?
<clever>
infinisil: it evals all of them in the same process, one at a time
<infinisil>
Hmm
<clever>
infinisil: 110 instances of nixos (100 nixos containers over 10 ec2 machines) takes several minutes and several gig of ram
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<infinisil>
clever: I think a good solution to this would be to have a nixops module, which stands above the nixos module, so that the declaration can be something like { machines = { a = ...; b = ..; }; } with the option machines of type `attrsOf nixosConfig`
<clever>
infinisil: yeah, the require trick in nixops is much more limited then imports
<cransom>
at least a couple years ago, lots of instances deployed in a single nixops deploy was very hard on memory and would bail inconveniently.
<infinisil>
so that it doesn't just go through each attribute, but actually evaluates the whole thing as a single value
<infinisil>
clever: require trick?
<clever>
infinisil: you can make a tree of deployment files with require, the same as imports, but you cant define custom options and pass them around between deployment files
<clever>
infinisil: it works identically to just listing several .nix files at `nixops create` or `nixops modify`, except you can manage the list with git
<clever>
there is no way to pass sets between the files
<clever>
but nixos definitions on each machine are merged as you would expect
<infinisil>
clever: i see, would love for nixops to adapt a module system
<ottidmes>
I wonder, does it leverage persistent data structures, I mean you generally will have the same nixpkgs for all your machines, the overlays will just differ, it would be nice if the base nixpkgs could then be shared memory-wise
<clever>
infinisil: in the example i linked, 2 files define nixos config for report-server
<infinisil>
Would also solve a problem I've had with machines that depend on each others config..
<infinisil>
which is currently not possible afaik
fragamus has joined #nixos
<infinisil>
ottidmes: i think it would only eval it once
<clever>
infinisil: oh, i think you can cheat, nodes.machine1.config would let you store things in a random machine
<ottidmes>
infinisil: with nixops you should be able to refer to other machine's configurations
rihards has quit [Quit: rihards]
<clever>
and machine1 = { ... }: hmmm, no nodes is passed to the nixos instances, not the deployment files
<infinisil>
Yeah I can refer to other machines configs, but not recursively
<infinisil>
i think
<ottidmes>
infinisil: I have not tested it, but I assume it is no different than the rest, so you can refer to each other, but you cannot refer to an attribute that is dependent on the attribute you are defining
MP2E has joined #nixos
jtojnar has joined #nixos
<infinisil>
Yeah I have tested this and am pretty sure it doesn't work, because it evaluates each machine separately
<ottidmes>
infinisil: That is why add an indirection to my configs, so I can define my overlays and use my overlays
BlessJah has quit [Ping timeout: 265 seconds]
<ottidmes>
infinisil: Hmm, thats too bad, I have not really tested nixops that much, I have a module of things I want to share between machines
<thoughtpolice>
Oh wow, is that nixpkgs.pkgs option the one I've been wanting forever now? I have a set of machines that I keep in sync with master, and I want them always locked to the same revision, but keeping channels in sync or wrapping nixos-rebuild (with -I) is somewhat annoying...
<infinisil>
ottidmes: that's totally possible normally
<infinisil>
My module is kinda special though
<infinisil>
I could share it here, it's kinda nice actually (if it worked :()
<ottidmes>
infinisil: Yep, I am using it outside NixOps. My NixOps is just a map that includes my hosts, 17 lines of code at the moment
<ottidmes>
infinisil: Please do, I am curious what you are talking about
<infinisil>
Have a look at the option declarations
<infinisil>
Maybe I made a mistake though, hmm
<ottidmes>
infinisil: So this is the special bit: default = builtins.attrValues nodes;
<infinisil>
Yeah this is a special bit, it allows any client to use the server for remote port forwarding
<infinisil>
But it only gets enabled for the ones that set the client option
cfricke has joined #nixos
coconnor has joined #nixos
BlessJah has joined #nixos
<infinisil>
So in the end you only have to set mine.rpf.client.* and the server will be configured as well with all the clients that set this option
<infinisil>
each client that set it should get a webserver proxy and a subdomain, or an ssh remote port forward
<ottidmes>
infinisil: You have a lot of lib.* calls while you have with lib; at the top
<infinisil>
oh right, yeah it went through a bunch of refactorings to try avoid the infinite recursions
<infinisil>
couldn't do it though
<ottidmes>
infinisil: But really what you want is that submodule available for all your hosts in all your hosts? I have that
zzamboni has quit [Quit: Leaving.]
<infinisil>
Nah that's not the problem, the problem is enabling the server option without it not infinitely recursing
Ross has joined #nixos
chisui has joined #nixos
<infinisil>
The module gets imported for every machine no problem
<infinisil>
Oh you know what, the nodes attribute includes the current machine itself too, so evaluating all nodes for the current config is resursive.. that might be it
<ottidmes>
infinisil: I am quite sure I have what you want
<ottidmes>
infinisil: I cannot get access to the whole client configs, but I do get access to the module I am interested in, of each client, so I can define things on my server for all my clients, and have that info defined in each client instead of hardcoded in the server
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 1 new commit to release-17.09: https://git.io/vANJf
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/release-17.09 30a782c Anthony Cowley: nix-info: apply SC1117 suggested fix...
alex`` has joined #nixos
asymmetric has joined #nixos
<mbrgm>
ottidmes: I tried substituteAll, which you mentioned above... but the problem seems to be that nix does not evaluate problems in the attribute set I'm passing
zzamboni has joined #nixos
<mbrgm>
seems to be due to lazy evaluation
<ottidmes>
mbrgm: If you do not use pkgs.substituteAll, it will not be evaluated, where do you use it?
cinimod has joined #nixos
logotags has quit [Read error: Connection reset by peer]
logotags has joined #nixos
xcmw has joined #nixos
hamishmack has quit [Quit: hamishmack]
hamishmack has joined #nixos
hamishmack has quit [Client Quit]
<mbrgm>
ottidmes: I'm using substituteAll.. however the problem seems to happen when a key in an attrset contains dots..
<mbrgm>
the there's no environment variable for that property
<mbrgm>
oh, there is an environment variable, but it does not get replaced
cinimod has quit [Ping timeout: 240 seconds]
lexnoproblem has quit [Quit: Page closed]
alex`` has quit [Quit: WeeChat 2.0.1]
xcmw has quit [Ping timeout: 268 seconds]
<mbrgm>
ok, got it working by replacing the dots with underscores in the attribute names
<ottidmes>
mbrgm: I have not actually ever needed to use pkgs.substituteAll, I thought it did string replacements, or does it do both, @name@ is string replacement and $name via env vars?
<rawtaz>
whats the command to release and get a new dhcp ip address?
zzamboni has quit [Read error: Connection reset by peer]
<rawtaz>
i can take the if down and up again but thinking there's a nicer way
zzamboni has joined #nixos
chisui has quit [Ping timeout: 260 seconds]
<infinisil>
ottidmes: yeah that sounds like what i need, i think the problem would be solved by just declaring each client in the server separately though
<infinisil>
Or at least remove the server itself
<ottidmes>
infinisil: I have a attribute called host (submodule), hosts (listOf submodule), and hostmap (attrsOf submodule) accessible when evaluating a particular config
<MichaelRaskin>
dhclient -r
<MichaelRaskin>
If you use dhclient, though
<rawtaz>
its a vanilla nixos install, i have no idea. i thought dhclient commands where now taken over by that pesky `ip` one like most other good old networking commands seem to be :/
<rawtaz>
thanks, will try that
<MichaelRaskin>
Maybe dhcpcd, actually
<MichaelRaskin>
dhcpcd -k
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] YorikSar opened pull request #36450: nixos/virtualisation: Increase GCE image size to 1.5G (master...enlarge-gce-image) https://git.io/vANkm
NixOS_GitHub has left #nixos [#nixos]
<rawtaz>
yeah that was it, i used -n
zzamboni has quit [Quit: Leaving.]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/release-18.03 17ff10a Andrew Childs: gdb: Flexible target matching for darwin...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 1 new commit to release-18.03: https://git.io/vANk6
<infinisil>
ottidmes: ohhh i think i know how
<infinisil>
ottidmes: the submodule represents a nixos module evaluation?
<ottidmes>
infinisil: Yeah, submodule is like any other module
<mog>
so i have my udev script working but it doesnt have the commands in path it should
<rawtaz>
which is the default one?
<mog>
how do i add a package to the run path for it
<ottidmes>
infinisil: What did you mean then?
<infinisil>
ottidmes: but it's a bit tricky to declare such an option, but I've done it before
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c opened pull request #36451: A note about scheme-* in texlive.combine (master...texlive-doc) https://git.io/vANIU
NixOS_GitHub has left #nixos [#nixos]
<tnks>
hmm... I have a caching server. And if I do a "nix copy --from" then "nix-build" won't rebuild the project.
<MichaelRaskin>
rawtaz: xpra is a separate story
<ottidmes>
infinisil: I will just show you what I have, although it might be missing stuff to competely understand what is going on
<infinisil>
ottidmes: was just trying to confirm that its a nixos submodule and not something else
<tnks>
but if I nix-collect-garbage first, and do a "nix-build" without first an explicit "nix copy --from" it pulls some of the project from the cache, but not all of it.
<tnks>
like something is not hashing right?
<tnks>
it's kind of confusing.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 1 new commit to release-18.03: https://git.io/vANI3
<NixOS_GitHub>
nixpkgs/release-18.03 bf5e7c4 Matthew Bauer: git-sizer: init at 1.0.0...
NixOS_GitHub has left #nixos [#nixos]
<rawtaz>
MichaelRaskin: how? the options list say its a display manager, same as the other ones
<NixOS_GitHub>
[nixpkgs] LnL7 pushed 1 new commit to release-17.09: https://git.io/vANIC
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/release-17.09 df69194 Domen Kožar: haskellPackages.foundation: disable tests on darwin...
<MichaelRaskin>
The name suggests that it is X persistent remote access related, though — it might replace the role of the display manager, but provide different functionality
<rawtaz>
okay. i guess the options info isnt entirely telling the full story then :9
<rawtaz>
:)
<infinisil>
ottidmes: hmm no this doesn't seem like what i want
<rawtaz>
but which is the default when you dont configure any of them specifically? services.xserver.displayManager.slim.enable is said to be false by default in the options list, yet the manual says that slim is the default display manager. is it inconsistent or am i missing some understanding of those conflicing pieces of info?
<infinisil>
ottidmes: i don't want to declare each of my host in a special file, i want to use nixops' nodes argument to refer to the other hosts directly
<ottidmes>
rawtaz: I would recommend as your display manager, gdm when you use Gnome, sddm when you use KDE, lightdm for the rest (slim is dead and xpra is some remote X thing)
<rawtaz>
ottidmes: thanks!
<rawtaz>
im using i3wm <3
<ottidmes>
rawtaz: Good choice, have used it in the past, with lightdm, so that should work
xeji has joined #nixos
aborsu has joined #nixos
hamishmack has joined #nixos
<rawtaz>
what do you use nowadays if i may ask?
<rawtaz>
i read about someone using emacs the other day heh
<ottidmes>
infinisil: Sure, but I think our goals are the same, and this achieves it for me in plain nixos, without the need for nixops, and makes using nixops easy as well, since I just reuse what is there
acarrico has joined #nixos
<MichaelRaskin>
StumpWM. Common Lisp, and don't accept these Elisp surrogates.
<ottidmes>
infinisil: You could also just evaluate your configs yourself first :P
<MichaelRaskin>
(I actually use StumpWM with some custom modules)
<ottidmes>
rawtaz: I use bspwm, big fan
cement has quit [Ping timeout: 240 seconds]
<ottidmes>
MichaelRaskin: What is your opinion of Clojure?
<MichaelRaskin>
I think the first question is whether JVM is a plus or a minus for you.
<rawtaz>
ottidmes: what do you like with bspwm that i3 doesnt provide?
<MichaelRaskin>
Clojure is also quite opinionated about the restrictions to mutability.
<ottidmes>
MichaelRaskin: ClojureScript is pretty mature as well
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
cfricke has quit [Quit: WeeChat 2.0]
<MichaelRaskin>
A fine language, though. I generally use JVM only when I have no other choice, but I did use Clojure when I needed to evaluate a few graph databases where the sanest option was via Java API. Obviously, Java is not a language I wanted to write.
slyfox has quit [Ping timeout: 256 seconds]
<mpickering>
another basic question. Once a file is copied into the $out directory in installation, that's it right? it can't be touched again with tools like patchelf?
<ottidmes>
rawtaz: First of, i3 is more beginners friendly, I started out with it, went to xmonad and after getting fed up with certain bugs learned of bspwm
slyfox has joined #nixos
<michiel_l>
mpickering: Sure it can be touched.
ryanartecona has quit [Quit: ryanartecona]
<ottidmes>
mpickering: During your building process you can do whatever with $out, afterwards it is readonly though
<mpickering>
how can I understand the error "error: install_name_tool: can't open file: /nix/store..." when I inspect the path there is a file at that location
orivej has quit [Ping timeout: 240 seconds]
<coconnor>
what should the owner be of the secret-key-file provided to nix for signing builds?
<mpickering>
it's hard to debug as the build takes over an hour
<ottidmes>
rawtaz: I do not think there is much in bspwm that cannot be done in i3 and vice versa, there are some differences, but what made me choose bspwm is that with i3 it is more common to stick with the default config and tweak it a bit, while with bspwm automating and creating your own behaviors with the CLI is the common way of doing things
<rawtaz>
ottidmes: indeed. question: if you have a window in full screen mode, and want to configure such that another window (matching a title, for example) can be shown "on top" (i.e. "over") that full-screen one?
<rawtaz>
that cant be done in i3 easily. the fullscreen one is what shows, period.
<fyuuri>
I just changed the hostname in configuration.nix and everything stopped working. What is dependent on the hostname?
<ottidmes>
MichaelRaskin: I really like the philosophy and general ideas behind Clojure, but got fed up with the language due to being dynamically typed and having to question my whole code base at some point due to assumptions being broken that then propagate and can no longer be easily tracked to the place it went wrong.
<coconnor>
fyuuri: Are you able to check journalctl -xe ?
<MichaelRaskin>
ottidmes: I actually write a lot of Common Lisp, which is also dynamically typed. Of course, SBCL does pretty aggressive type inference for optimisation, and warns if it foresees problems…
zzamboni has joined #nixos
<ottidmes>
MichaelRaskin: I much prefer a compiler yelling at me all the time then finding out I made a typo in some code after some time. I only ever had that feeling of not being able to trust my code with JavaScript and Clojure, not with PHP for example
<fyuuri>
coconnor: yes. After setting the hostname back to the old one. Everything works
<coconnor>
fyuuri: can you post the journal output from the previous switch?
<ottidmes>
rawtaz: you can place stuff over a fullscreen window, but only when it is in floating mode
<MichaelRaskin>
ottidmes: I like dynamical typing for many cases, and type inference catches a large neough share of problems.
<fyuuri>
I will run that again
<rawtaz>
ok, peculiar.. i changed display manager in configuration.nix from lightdm to gdm, then ran `nixos-rebuild test`. for some reason the display manager isnt showing up on tty7 - instead i just see a couple of old boot messages there. seeing "NOT restarting the following changed units: display-manager.service" in the output, but doesnt help restarting that one manually.
<rawtaz>
so my graphical login is gone. even after i changed it back to lightdm and ran `nixos-rebuild test` again
<coconnor>
fyuuri: I recently had an issue with nscd that caused (what looked like) hangs. Curious if related
<rawtaz>
(that one worked before)
<tnks>
nice. I just needed to do a "nix copy-sigs"
<MichaelRaskin>
Restarting/not-restarting X is a complicated story
<mog>
how would i run a nix-shell script from udev?
hiratara has quit [Ping timeout: 252 seconds]
<ottidmes>
rawtaz: Could be some state changed, that will not be rolled back
<fyuuri>
Mar 07 21:48:55 nixos_air xsession[931]: No protocol specified
<fyuuri>
Mar 07 21:48:55 nixos_air xsession[931]: urxvt: can't open display :0.0, aborting.
<fyuuri>
coconnor ^
<rawtaz>
so best bet is to just reboot i guess?
mounty_ has joined #nixos
<ottidmes>
rawtaz: I always reboot for such things
<rawtaz>
ok
<fyuuri>
coconnor : the old hostname was the standard nixos and I changed it to nixos_air
<mfiano>
Something is wrong with my mime types. All files in pcmanfm or thunar are either text or octet streams. file associations are screwed because I can only associate applications from amoung those 2 types. Anyone know what could cause this?
<coconnor>
fyuuri: aside from looking like the xserver failed not much I can say. Any earlier errors? Have you tried ye-old reboot?
sdemos has joined #nixos
<coconnor>
fyuuri: nixos-rebuild boot will switch to a config on next boot.
<ottidmes>
Is systemConfig actually used as kernel param, I see it in grub, but I see a lot of examples where they only use init+kernelParams, so I doubt it is needed
hiratara has joined #nixos
<fyuuri>
coconner: resetting the hostname repairs it. I can not start X-server aps from the fish shell after changing the hostname
pxc has joined #nixos
<mounty_>
Is there a channel for discussing Hydra ?
<mounty_>
#hydra is something else.
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] primeos pushed 1 new commit to master: https://git.io/vANmO
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master fa051cd Michael Weiss: aiohttp-cors: 0.6.0 -> 0.7.0...
<coconnor>
is #hydra more on the "hail hydra" side?
<MichaelRaskin>
I think everything in Nix ecosystem defaults to here…
<coconnor>
mfiano: what desktop config are you using? KDE, Gnome, custom?
<mfiano>
i3
<MichaelRaskin>
The /topic lists the exceptions…
<coconnor>
mfiano: what is the value of services.xserver.desktopManager.default ?
<coconnor>
or specifically environment.pathsToLink
<mfiano>
coconnor: "none". I don't use a DM, only WM
<fyuuri>
coconnor: yes. I will try that again. Does using rebuild boot make a difference here? Is it ok if I reboot after a rebuild switch?
<mfiano>
coconnor: So what could be causing those xml files to not be there?
<mfiano>
Am I missing some package?
<coconnor>
mfiano: maybe... one of the other share dirs in XDG_DATA_DIRS might contribute the missing mime files. However, that does not resemble my desktop
<fyuuri>
Ill use boot and reboot
<mfiano>
coconnor: From my understanding of gnu/linux, the shared-mime-info package is supposed to provide those xml mime types
<mfiano>
But I'm not sure the package name in nixpkgs
<coconnor>
mfiano: try adding pkgs.shared_mime_info to your systemPackagtes
<mbrgm>
I'm trying to auto-generate a DNS zone file from nix(ops) config... however, one point I did not manage to implement is an auto-incrementing serial number (so zone transfers get through properly)... is this somehow possible by (standard) nix means or can't it be done?
acarrico has joined #nixos
<blankhart>
nixpkgs is giving surprising behavior with a build for ghcjs - this is the derivation http://lpaste.net/363256
<blankhart>
using this, it tries to build the ghcjs module with ghc 7.10.3
<blankhart>
switching to ghcjsHEAD correctly uses ghcjs to build
<blankhart>
i find the organization of nixpkgs difficult to navigate (as a beginner) so am not sure why this would be happening
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c closed pull request #36451: A note about scheme-* in texlive.combine (master...texlive-doc) https://git.io/vANIU
NixOS_GitHub has left #nixos [#nixos]
<ottidmes>
blankhart: That is how ghcjs is defined: ghcjs = packages.ghc7103.callPackage ... (in pkgs/top-level/haskell-packages.nix)
<blankhart>
why would ghcjs be aliased to ghc?
<blankhart>
as opposed to ghcjs.
thblt has quit [Ping timeout: 256 seconds]
szicari has quit [Quit: szicari]
<ottidmes>
blankhart: Its not, check the file, I truncated it, here is a bit more: ghcjs = packages.ghc7103.callPackage ../development/compilers/ghcjs
<ottidmes>
blankhart: It is just using ghc 7103 to build it
acarrico has quit [Ping timeout: 256 seconds]
Myrl-saki has quit [Read error: Connection reset by peer]
erictapen has joined #nixos
cement has joined #nixos
<blankhart>
ottidmes: thank you. i'll have to think harder about why changing from ghcjs to ghcjsHEAD affects the build. it seemed to me it was trying to compile ghcjs dependencies with ghc
<ottidmes>
mbrgm: I found it an interesting question, I checked home-manager, which needs a new id for each generation, it seems to use a file to track the ids
leat has quit [Ping timeout: 245 seconds]
shabius_ has joined #nixos
leat has joined #nixos
ilhart has joined #nixos
vidbina has joined #nixos
shabius has quit [Ping timeout: 240 seconds]
ilhart has quit [Ping timeout: 260 seconds]
xcmw has joined #nixos
<ottidmes>
mbrgm: Anyway you probably want to look at activationScripts if you did not already
<mbrgm>
ottidmes: yeah I was thinking about that too... basically I thought about keeping a state file on the machine and then generating the serial number from it in a pre-activation script
jtojnar has quit [Quit: jtojnar]
<ottidmes>
mbrgm: If a sha256 or part of it was allowed, you could just use the hash of your zone file store path, otherwise you would have to resort to stateful things, which means either activationScripts or systemd
davidlt has quit [Remote host closed the connection]
davidlt has joined #nixos
civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
<mbrgm>
ottidmes: thought about taking 32 bits of the sha, too... but it seems the serial must be incremented, not only changed, for slave servers to pick up
jtojnar has joined #nixos
xcmw has quit [Ping timeout: 260 seconds]
loonquawl has joined #nixos
<rawtaz>
is putting an xrandr command in ~/.xprofile the best way to configure the screen resolution for your GUI (using i3wm), or is there a more proper way to do it in NixOS?
<rawtaz>
hm maybe the services.xserver.resolutions is a better place
<ottidmes>
rawtaz: I had it in my i3 config and now have it in my bspwm config
<rawtaz>
hmm ok. as a command that was executed on start of i3 i suppose
<tnks>
I'm running Nix+Debian, and I tried the zoom-us package. Is it working for people on NixOS?
jxf has joined #nixos
<tnks>
I'm curious if it's just not working for me because of a Nix/Debian interaction.
<blankhart>
what should my next step be if Nix tells me that building one of my dependencies failed, with no further information?
<rawtaz>
blankhart: what's the last thing you saw before that message? can you pastebin the output?
<ottidmes>
blankhart: scroll up :P whenever I have that it is something else going bad further back
<coconnor>
rawtaz: there is also the services.xserver.xrandrHeads option set
<rawtaz>
blankhart: might be that you're just seeing a bunch of downloads before that. if that's the case, run the command again and youll probably get better output
<symphorien>
jxf: use nix-env -iA
<symphorien>
!-A
<ottidmes>
nix-env -f '<nixpkgs>' -iA spotify
<rawtaz>
coconnor: ah yeah. and even the services.xserver.virtualScreen
<jxf>
symphorien: OK, so I did `nix-env -f '.nix-defexpr' -iA spotify`, but received `attribute 'spotify' in selection path 'spotify' not found`. I feel like I'm badly misunderstanding something about how packages work.
<clever>
jxf: there is no default.nix inside .nix-defexpr, thats why it failed
<clever>
jxf: you need to nix-env -iA nixos.spotify
<symphorien>
or nix-env -iA nixpkgs.spotify on non-nixos
<symphorien>
or nix-env -f '<nixpkgs>' -iA spotify
<rawtaz>
blankhart: i have no idea. let's hope someone with more skills can take a look
<blankhart>
the project builds with stack; my experience with nix so far has been long builds from source and then terse failures, but i want to believe
<blankhart>
thanks for looking rawtaz
acarrico has quit [Ping timeout: 260 seconds]
<rawtaz>
blankhart: what packages have you had problems with? are yo uon stable or unstable?
endformationage has joined #nixos
<blankhart>
unstable
<blankhart>
it really depends
<symphorien>
also nix 2 or nix 1 ?
<rawtaz>
might go stable then :) you can still pull in unstable packages when you want the very latest version
vcunat1 has quit [Read error: Connection reset by peer]
<symphorien>
on nix 2 you have a nix log command to have the build logs
<blankhart>
nix 2
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jtojnar opened pull request #36457: gnome3: minor update (master...gnome3-update) https://git.io/vANnK
NixOS_GitHub has left #nixos [#nixos]
<clever>
symphorien: 1.11 also has `nix-store -l /nix/store/path` for successfull local builds
<symphorien>
well on nix 1 you juste have to scroll up :þ
<jxf>
clever + symphorien: Okay, that seemed to work. I was able to run `nix-env - spotify --command spotify`! Thanks.
<infinisil>
clever: I always get suprised on saturday night when it starts running, since it's literally the only automatic background thing that takes some power on my machine
<clever>
infinisil: mine runs every day at midnight
<infinisil>
(Unlike on macOS or Windows where there's tons of stuff happening in the background)
plakband has joined #nixos
<dtz>
yeah I use that on my servers haha :3
<clever>
5gig free on my primary desktop
<jxf>
So I got spotify working on non-nixos, but audio doesn't actually work (`ALSA lib pcm_dmix.c:1099:(snd_pcm_dmix_open) unable to open slave` when playing songs in Spotify). I think I need to grab some audio support. Is there way to ask Nix to install, e.g., pulseaudio? I tried `nix-env -iA nixpkgs.pulseaudio` but that didn't seem to work.