<gchristensen>
we're doing pretty well patching on quickly and thoroughly, but a bit more organizing and tooling around sending and tracking patch status is definitely needed
<dhess>
gchristensen: is vulnix still a thing?
<gchristensen>
dhess: yes, and it is getting better and ckauhaus semi-regularly uses it to find a list of packages to patch
<phdoerfler>
gchristensen: are you doing this as a volunteer?
<dhess>
cool
<gchristensen>
phdoerfler: yes, NixOS is almost entirely volunteer-based
rogue_koder has quit [Remote host closed the connection]
rogue_koder has joined #nixos
<phdoerfler>
gchristensen: thanks for your hard work :)
<gchristensen>
:) you're welcome!
<gchristensen>
(if you know of any companies that would like to regularly fund work on NixOS, we can easily put it to work)
<phdoerfler>
the google group there seems a bit dead, right?
<gchristensen>
it is, but once we're back sending mail it'll be used again
<phdoerfler>
ok
Supersonic112 has joined #nixos
MP2E has quit [Remote host closed the connection]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dtzWill opened pull request #33589: make-derivation: Don't add host-suffix to fixed-output derivations names (master...fix/fixed-output-cross-name) https://git.io/vNIOE
NixOS_GitHub has left #nixos [#nixos]
<pjan_>
Is there an (easy) way to see what settings (nix.config) nix(pkgs) is using? (background, I'm using nix-darwin's nix module, but can't figure out whether it's using those settings or not. $NIX_CONF_DIR keeps pointing to the original /etc/nix directory)
jrolfs__ has joined #nixos
<gchristensen>
I'm sorry I let the mails stop
nithor has joined #nixos
<gchristensen>
pjan_: it is using /etc/nix/nix.conf
jrolfs has joined #nixos
alexteves has quit [Ping timeout: 265 seconds]
MP2E has joined #nixos
<LnL>
why the confusion, are you having a problem with nix.conf?
randomstrangerb has quit [Ping timeout: 252 seconds]
asuryawanshi has joined #nixos
<phdoerfler>
The only company I know of that uses nixos would be iohk.io
randomstrangerb has joined #nixos
<phdoerfler>
No idea if they'd wanna fund things, though
<gchristensen>
:)
jrolfs__ has quit [Ping timeout: 255 seconds]
jrolfs has quit [Ping timeout: 265 seconds]
<dhess>
Dezgeg: around?
<phdoerfler>
I know I can configure nixos to auto update itself. Where do I see if that's working? And is there a way to configure how often this happens?
alexteves has joined #nixos
<pjan_>
LnL, if I print out the contents of `/etc/nix/nix.conf`, it's not the once set in `nix-darwin` (or the nix.conf that the store references in the latest nix-darwin module, which does contain the settings defined in nix-darwin)
<pjan_>
*ones
rcshm has quit [Ping timeout: 246 seconds]
<LnL>
oh, and you didn't get a warning that there's an existing file?
asuryawanshi has quit [Ping timeout: 276 seconds]
<pjan_>
not that I'm aware of. Didn't see any
<LnL>
sounds like a bug then
<LnL>
could you create an issue with some info about the layout/symlinks of /etc /etc/static and the active system
<infinisil>
And there you can see that all it does is run nixos-rebuild switch at the given dates
cybrian has quit [Remote host closed the connection]
<phdoerfler>
I see
<phdoerfler>
thank you :)
<infinisil>
np :)
<pjan_>
@LnL, yeah, getting this. So we need to remove the /etc/nix/nix.conf before running this? Every time?
cybrian has joined #nixos
<LnL>
pjan_: you only have to do that once, then nix-darwin knows it's safe to replace because it's a symlink
orivej has joined #nixos
slack1256 has joined #nixos
jrolfs has quit [Ping timeout: 248 seconds]
jrolfs__ has quit [Ping timeout: 276 seconds]
<slack1256>
Is there anyway to do cgroups (v2) on the fly?
<slack1256>
I don't want to be mkdir that hierarchy every other bootup
<andi->
I just started playing with nixops on AWS.. I was able to create a spot instance but it is stuck at "Reconfigure the system from EC2 userdata on startup" (seen in the serial log). SSH connectino fails. Ping also fails. Security group alls all connections.. Never really used AWS before... :/ Any idea?
rycee has quit [Ping timeout: 276 seconds]
Obscurity[m] has quit [Ping timeout: 276 seconds]
matrixkrav has quit [Ping timeout: 276 seconds]
M-fishy has quit [Ping timeout: 276 seconds]
aspiwack[m] has quit [Ping timeout: 276 seconds]
spietz has quit [Quit: WeeChat 1.9.1]
Arcaelyx has quit [Read error: Connection reset by peer]
orivej has quit [Ping timeout: 265 seconds]
Arcaelyx has joined #nixos
matrixkrav has joined #nixos
rycee has joined #nixos
Obscurity[m] has joined #nixos
M-fishy has joined #nixos
aspiwack[m] has joined #nixos
infinisil has quit [Quit: ZNC 1.6.5 - http://znc.in]
infinisil has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] corngood opened pull request #33590: mesa: set sysconfdir to /run/opengl-driver/etc (staging...mesa-sysconf) https://git.io/vNI37
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] obadz pushed 1 new commit to master: https://git.io/vNI3F
<NixOS_GitHub>
nixpkgs/master 3b36db7 Frederik Rietdijk: citrix_receiver: 13.7.0 -> 13.8.0
NixOS_GitHub has left #nixos [#nixos]
Pepe_ has joined #nixos
liori has joined #nixos
hamishmack has quit [Quit: hamishmack]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] puffnfresh opened pull request #33591: SunVox: init at 1.9.3b (master...package/sunvox) https://git.io/vNIsC
NixOS_GitHub has left #nixos [#nixos]
Inst_ has quit [Ping timeout: 248 seconds]
orivej has joined #nixos
m0rphism has joined #nixos
nithor has quit [Ping timeout: 255 seconds]
thc202 has quit [Ping timeout: 252 seconds]
jrolfs__ has joined #nixos
jrolfs has joined #nixos
asuryawanshi has joined #nixos
klntsky has quit [Ping timeout: 272 seconds]
nithor has joined #nixos
swayf has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
orivej has quit [Ping timeout: 248 seconds]
asuryawanshi has quit [Ping timeout: 240 seconds]
ryanartecona has joined #nixos
rcshm has quit [Ping timeout: 265 seconds]
jrolfs__ has quit [Ping timeout: 260 seconds]
jrolfs has quit [Ping timeout: 276 seconds]
<ertes>
is there an easy way to override just a single field of a haskell package? use case: setting doCheck = false
<dtz>
infinisil: naw, progress thing shows up as all users. But your nix-daemon needs to be 1.12 too, so just dropping into a nix-shell or installing with nix-env won't do the trick
<NixOS_GitHub>
[nixpkgs] yegortimoshenko opened pull request #33596: wine: wineStaging -> wine-staging, add wine-development, use full scheme for wine-stable (master...20180108.033429/wine) https://git.io/vNInx
<NixOS_GitHub>
[nixpkgs] ghuntley opened pull request #33597: duosec: use root uid as sshd uid has been retired (master...patch-1) https://git.io/vNIc8
stepcut has quit [Remote host closed the connection]
jrolfs has quit [Ping timeout: 265 seconds]
jrolfs__ has quit [Ping timeout: 264 seconds]
<dhess>
hmm why can't I get Nixpkgs to honor a postPatch attribute in a kernel package
<dhess>
it just ignores it
<dhess>
oh I put it in the wrong place
rogue_koder has quit [Ping timeout: 252 seconds]
rogue_koder has joined #nixos
<dhess>
how interesting. I put the postPatch in a nonsensical place, so it had no effect on the derivation.
Inst has joined #nixos
<dhess>
Meanwhile my Hydra just refused to re-run the job and I couldn't figure out why. Basically, it ws telling me that I had not actually changed the derivation and there was nothing to do!
schoppenhauer has quit [Ping timeout: 265 seconds]
jrolfs__ has joined #nixos
<samueldr>
to re-run it while in the non-sensical place or after fixing it?
schoppenhauer has joined #nixos
<dhess>
the first time, when I put it in the wrong place and it had no effect on the derivation, Hydra refused to re-evaluate the job despite my clicking "Evaluate job"
<dhess>
and I thought Hydra was broken somehow
<dhess>
but that should have been a hint to me that what I had done was not actually changing anything.
<samueldr>
:)
<dhess>
now that I've fixed it it's running just fine
rogue_koder_ has quit [Remote host closed the connection]
rogue_koder_ has joined #nixos
sary has joined #nixos
ryanartecona has quit [Quit: ryanartecona]
jb55 has quit [Ping timeout: 264 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Ericson2314 opened pull request #33603: stdenv: Force `doCheck` to be false when we are cross compiling (master...cross-check) https://git.io/vNI4b
NixOS_GitHub has left #nixos [#nixos]
orivej has quit [Ping timeout: 265 seconds]
jrolfs__ has joined #nixos
jrolfs has joined #nixos
MercurialAlchemi has joined #nixos
aloiscochard has joined #nixos
reinzelmann has joined #nixos
samtoday has quit [Ping timeout: 240 seconds]
aarvar has quit [Ping timeout: 264 seconds]
randomstrangerb has quit [Ping timeout: 246 seconds]
orivej has joined #nixos
randomstrangerb has joined #nixos
asuryawanshi has quit [Remote host closed the connection]
<hyper_ch>
hmmm, I tried to upgrade nixos on the server but it doesn't build. Now how can I install throught the configuration.nix new services for the current generation/revision?
freusque has quit [Quit: WeeChat 1.9.1]
freusque has joined #nixos
atrx has quit [Read error: Connection reset by peer]
sigmundv__ has joined #nixos
<makefu>
hyper_ch: but keeping the "old" rev running? i think you will need to cherry-pick the commits onto the revision you want
<hyper_ch>
makefu: no, I don't want to upgrade since it's broken
<hyper_ch>
but I can't add new stuff because it uses the new channel info
<hyper_ch>
and hence breaks again
freusque has quit [Ping timeout: 240 seconds]
<hyper_ch>
how can I "downgrade" the channel rev to what the current generaetion is?
rogue_koder_ has quit [Quit: Konversation terminated!]
jrolfs has joined #nixos
jrolfs__ has joined #nixos
orivej has quit [Ping timeout: 248 seconds]
<makefu>
nixos-rebuild switch --rollback
<makefu>
or nix-channel --rollback nixpkgs , depending on what you want to do
<hyper_ch>
nix-channel --rollback nixpkgs
<hyper_ch>
invalid channel generation number ‘nixpkgs’ at /run/current-system/sw/bin/nix-channel line 229.
<hyper_ch>
makefu: again, I trun: nixos-rebuild boot --upgrade
<hyper_ch>
makefu: upgrade went wrong
<hyper_ch>
makefu: I edit configuration.nix add new servier
<hyper_ch>
makefu: I run nixos-rebuild boot
<hyper_ch>
makefu: it fails to build because it still attempts the use the previous channel info form the previous failed --upgrade attempt
<hyper_ch>
how can I roll back to current generation's channel and get that service installed?
MercurialAlchemi has quit [Ping timeout: 248 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vNIuc
<NixOS_GitHub>
nixpkgs/master e5d9d61 Joachim Fasting: electrum: 3.0.4 -> 3.0.5...
NixOS_GitHub has left #nixos [#nixos]
MercurialAlchemi has joined #nixos
leapingfrogs has joined #nixos
<jluttine>
makefu: yep, thanks. probably more stuff is needed, but it's just weird i can't make it find boost even though it's added as a build input. i added pkgconfig too, but it didn't help. yesterday i tried figuring out how cmake tries to find boost i couldn't figure it out. tried to study cmake a bit as i'm not familiar with that. but couldn't solve the issue.
leapingfrogs has quit [Remote host closed the connection]
atrx has joined #nixos
<makefu>
i also checked for the boost version from nixpkgs and on stable at least it is 1.62
asuryawanshi has joined #nixos
Itkovian has joined #nixos
bobjohn has joined #nixos
bobjohn has quit [Client Quit]
<jluttine>
yep. my nixpkgs checkout has 1.65.1 so that shouldn't be the issue
la_putin has joined #nixos
ech0_42 has joined #nixos
jrolfs__ has joined #nixos
jrolfs has joined #nixos
jensens has joined #nixos
periklis has quit [Ping timeout: 268 seconds]
dan_b has joined #nixos
Itkovian has quit [Read error: Connection reset by peer]
m0rphism has quit [Quit: WeeChat 1.9.1]
asuryawanshi has quit [Remote host closed the connection]
dan_b has quit [Ping timeout: 260 seconds]
stphrolland has joined #nixos
bkchr has joined #nixos
<srhb>
Hmm, I thought nix-copy-closure --to somehost --include-outputs /nix/store/...-system-...drv would copy the entire system closure as well as the derivation. It doesn't. What does that flag do then?
hamishmack has joined #nixos
goibhniu has joined #nixos
<srhb>
If I nix-store --query --outputs /nix/store/...-system-...drv it clearly produces the actual output, but it's not copied over.
dywedir has joined #nixos
thc202 has joined #nixos
jrolfs has quit [Ping timeout: 240 seconds]
Itkovian has joined #nixos
jrolfs__ has quit [Ping timeout: 268 seconds]
asuryawanshi has joined #nixos
<srhb>
What I'm trying to achieve is ensuring that nix-store -qd /run/current-system actually always works on the target host.
asuryawanshi has quit [Ping timeout: 260 seconds]
<joko>
Hello, is there a way to access a shell while making a derivation and control the build process? Sth like a shell where I enter unpack, patch etc.
moet has joined #nixos
<srhb>
joko: nix-shell '<nixpkgs>' -A yourAttribute
orivej has joined #nixos
<joko>
srhb: I was looking for nix-build --run-env
<joko>
ok, I'm running it, but it seems unpackPhase does not create a temporary directory, is this correct?
<srhb>
joko: Yup.
orivej has quit [Ping timeout: 240 seconds]
<dhess>
srhb: I'm so sorry your Gitlab PR got clobbered :(
<srhb>
dhess: No problem, I fixed it :)
<srhb>
It was a trivial merge (if I did it right ;-))
<dhess>
srhb: oh wow sweet!
<srhb>
Or well, rebase
<dhess>
I didn't get a notification, hmm
<srhb>
Probably rebase weirdness
<srhb>
Ugh, clearly I didn't resolve them all.
<dhess>
I guess GitHub doesn't notify on a push update
<srhb>
I'll have to take a look when I'm off work
<dhess>
anyway that's great to hear
<dhess>
it looks like just the manual?
<srhb>
Indeed.
<srhb>
I thought I fixed that.
<srhb>
Oh, it's a new conflict.
<dhess>
ohh
fendor has joined #nixos
samtoday has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 2 new commits to release-17.09: https://git.io/vNIVw
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/release-17.09 5431160 elseym: mattermost: create role and db with postgres superuser...
leapingfrogs has quit [Remote host closed the connection]
i-am-the-slime has joined #nixos
__Sander__ has joined #nixos
oida has joined #nixos
FRidh has joined #nixos
betaboon has joined #nixos
ThatDocsLady has joined #nixos
atrx has quit [Ping timeout: 260 seconds]
pjan_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<srhb>
So, it seem that if I nix-copy-closure both the system and the system derivation to a remote host and then switch, nix-store -qd /run/current-system will work for _new_ system generations, but not for existing ones.
pjan_ has joined #nixos
<srhb>
I'm not sure how this association works. It's very weird,.
klntsky has quit [Remote host closed the connection]
klntsky has joined #nixos
kriztw_ is now known as kriztw
<sphalerite>
grahamc: (and anyone else who has a Dell 9560) do you have bluetooth working?
mkoenig has quit [Remote host closed the connection]
<sphalerite>
When I do "scan on" in bluetoothctl, I get "Failed to start discovery: org.bluez.Error.NotReady"
<sphalerite>
"power on" also doesn't work: "Failed to set power on: org.bluez.Error.Failed" and in the journal "Bluetooth: hci0 urb ffff929cc3545780 failed to resubmit (113)" and "Failed to set mode: Failed (0x03)"
asuryawanshi has joined #nixos
<joko>
Is it possible for nix-build to comply with nixpkgs.config.packageOverrides of the system?
<sphalerite>
a quick search suggests that this is related to missing firmware, but I haven't found anything corresponding to this bluetooth controller, only for broadcom stuff. The bluetooth device is a 0cf3:e300 (internal USB)
<srhb>
How do I tell Nix store that the deriver is now present for a given store path?
mkoenig has joined #nixos
<joko>
srhb: I don't get it, nix-env etc. should realise without any extra command that a path is available, right?
jrolfs has quit [Ping timeout: 240 seconds]
jrolfs__ has quit [Ping timeout: 260 seconds]
<srhb>
joko: I thought so, but clearly not.
<srhb>
I have a very poor understanding of the "which is the deriver for this out path" mechanism
asuryawanshi has quit [Ping timeout: 276 seconds]
<sphalerite>
joko: the easiest way to do that is by putting it in a separate config file and putting it in the system config as `nixpkgs.config = import path/to/config.nix;`
<sphalerite>
and also symlinking that config.nix file to ~/.config/nixpkgs/config.nix
sigmundv__ has joined #nixos
<sphalerite>
Although nowadays you should be using overlays instead of packageOverrides :)
<srhb>
Maybe it's related to the gc root somehow...
<fendor>
why do i get messages about Repeated allocation of very large block?
<sphalerite>
srhb: iirc it's kept in the database
<joko>
sphalerite: many thanks! With overlays I noticed the following: I was overriding a package (e.g. lvm2), but the rest of the packages were not using the new package as a dependency, I had to override their lvm2 dependency attribute. Is this intended or I was doing sth wrong?
<sphalerite>
joko: it should affect all things that depend on lvm2
<srhb>
sphalerite: Hm.. Can it be regenerated?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] edolstra pushed 2 new commits to master: https://git.io/vNI6A
<NixOS_GitHub>
nixpkgs/master f674cb5 Peter Hoeg: kernel: enable writeback throttling to avoid IO queues filling up
<sphalerite>
srhb: the only way I know of is realising the derivation
<srhb>
sphalerite: That only works the first time.
<srhb>
sphalerite: Which is the weirdness.
moet has quit [Ping timeout: 240 seconds]
<srhb>
(As in, if the output is already in the store, you can't seem to add a deriver)
<sphalerite>
Can't you delete the output path and realise it again?
<bennofs[m]>
The deriver is a .drv, right?
pjan_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite>
I think that makes sense
<srhb>
sphalerite: No, it's the active system. Yes, it's a .drv
<sphalerite>
Benno Fünfstück: yes
<sphalerite>
Ah, that is problematic x)
<srhb>
Yes. Even the older ones are problematic, since obviously I don't want to get rid of the gc paths
<sphalerite>
Why do you actually need the deriver set?
<srhb>
Because I want to be able to nix-diff the current deployment with the incoming one
<srhb>
Or any of the older system generations with any of the others :)
<sphalerite>
Aaaah
<srhb>
And it _works_ in my current deployer, but only for newer deployments that were copied over along with the derivation
<srhb>
Which also seems incredibly brittle
<sphalerite>
The only way I know of is manually messing with the database
<srhb>
sphalerite: Yeah.. Meh :-)
asuryawanshi has joined #nixos
<sphalerite>
I think it's also possible for different derivations to produce the same output path
<srhb>
How so? Isn't the derivation actually considered an input?
<sphalerite>
So it's not necessarily one-to-one
alexteves_ has joined #nixos
<srhb>
Well, that's okay in this case regardless.
<srhb>
As long as the output is identical.
<sphalerite>
I don't know exactly how it works but it involves fixed-output derivations
alexteves_ has quit [Client Quit]
<srhb>
I wonder if this whole issue is somehow related to how nix-copy-closure doesn't seem to recognize the system output as an actual output.
<srhb>
Like it would with, say, hello
alexteves has joined #nixos
<sphalerite>
I don't think so..?
<srhb>
meh, then I have two issues. :-)
<sphalerite>
Well
<sphalerite>
What do you mean by nix-copy-closure not recognising the system output as an actual output?
<srhb>
nix-store --query --outputs on a /nix/store/...system...drv derivation (eg NixOS) will output the actual system out path, but nix-copy-closure --include-outputs will not copy it
MP2E has quit [Remote host closed the connection]
<srhb>
Hm, maybe the term outputs is overloaded..
<srhb>
But if I nix-copy-closure --include-outputs /nix/store/...hello...drv it will indeed copy the output.
<NixOS_GitHub>
[nixpkgs] peti closed pull request #33405: Make GHC work on armv7l-linux (master...ghc-armv7l) https://git.io/vNe5m
NixOS_GitHub has left #nixos [#nixos]
r0bby has quit [Ping timeout: 255 seconds]
r0bby has joined #nixos
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] peti pushed 1 new commit to master: https://git.io/vNI1H
<NixOS_GitHub>
nixpkgs/master 83a5765 Peter Simons: LTS Haskell 10.3...
<musicmatze>
morning! When will the next unstable land? The current unstable is 5 days old and does not yet contain the fixes for spectre/meltdown ... If it can be expected to land today I won't have to build my own kernels on two devices... if not, I guess I have to...
Neo-- has joined #nixos
samtoday has quit [Quit: Quit]
<hyper_ch>
Mic92: online?
<betaboon>
hello #nixos, i have a question regarding nixops: sometimes during 'deploy' it gets stuck and i have no idea where to start on trying to debug the problem. any ideas ?
coot has joined #nixos
dywedir has quit [Ping timeout: 240 seconds]
Inst has joined #nixos
<LnL>
musicmatze: you can use the kernel from 17.09, but I recommend doing it the other way around. it's very easy to overlay packages from unstable without switching the entire system
<musicmatze>
LnL: ah, I can do this with an overlay? I never thought of that! What I would have done: Checkout my system generation in my nixpkgs clone, cherry-pick the kernel update and rebuild the system with my local nixpkgs clone! Using an overlay is way smarter, though! Do you have an example how to do it?
<LnL>
but instead of pkgs.linuxPackages you can use something like let stablePkgs = import <nixos-17.09> {}; in stablePkgs.linuxPackages;
<LnL>
if you have a "nixos-17.09" channel
asuryawanshi has quit [Remote host closed the connection]
<musicmatze>
I have not. I guess I would override the kernel package to use a newer kernel (4.14.15 I guess) by overriding the whole kernel package definition...
<hyper_ch>
weird..... on the server with systemd bridges the actual host can't get a network connection anymore... but the vms do just fine
jrolfs has joined #nixos
jrolfs__ has joined #nixos
orivej has joined #nixos
fendor has joined #nixos
orivej has quit [Client Quit]
orivej has joined #nixos
iMatejC has quit [Remote host closed the connection]
<musicmatze>
I don't know what I am doing wrong, but I fail to override the kernel in my configuration.nix to use 4.14.12 :-/
randomstrangerb has quit [Ping timeout: 248 seconds]
randomstrangerb has joined #nixos
iMatejC has joined #nixos
srdqty has joined #nixos
jrolfs has quit [Ping timeout: 248 seconds]
jrolfs__ has quit [Ping timeout: 265 seconds]
<hedning[m]>
musicmatze: this should do it: boot.kernelPackages = pkgs.linuxPackages_latest;
patrl has joined #nixos
fyuuri has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
FRidh has joined #nixos
asuryawanshi has joined #nixos
<musicmatze[m]>
hedning: nope because the unstable channel is not released with a new kernel yet. Also I run 4.14 on my workstation and latest only on my thinkpad
zzamboni has quit [Quit: Leaving.]
Ivanych has joined #nixos
iyzsong has joined #nixos
<hedning[m]>
musicmatze: ah, yeah, unstable has been stuck for a while :( unstable-small does the trick, though there might be some compilation involved
asuryawanshi has quit [Ping timeout: 252 seconds]
zzamboni has joined #nixos
jrolfs has joined #nixos
FRidh has quit [Ping timeout: 252 seconds]
pjan_ has joined #nixos
jrolfs__ has joined #nixos
fendor has joined #nixos
<hyper_ch>
something strange has changed and I don't know where or how
fendor has quit [Remote host closed the connection]
<gchristensen>
makefu: funny that nixos-on-your-router (started as a joke) is now A Channel :D
* etu
is looking forward to the nixos-phone-channel
ma27 has quit [Ping timeout: 255 seconds]
ilyaigpetrov has joined #nixos
<makefu>
etu: i think it is not that far fetched
<pie_>
manveru, dont i need to update or something?
<hoverbear>
gchristensen: Transmission is amazing, qbittorrent is also ok
<makefu>
i'd love to have nixos installed on my smartphone
<etu>
makefu: I know, I backed the Purism Librem 5 for that reason ;)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peti pushed 2 new commits to master: https://git.io/vNIh4
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 397515c Peter Simons: Merge pull request #33607 from mpickering/purescript...
<NixOS_GitHub>
nixpkgs/master 889d618 Matthew Pickering: Jailbreak purescript package...
patrl has quit [Quit: WeeChat 1.9.1]
<LnL>
hoverbear: I just use 2 channels instead of one, but the basic idea is the same
<manveru>
pie_: i thought it was, but seems like it's not :(
<hoverbear>
LnL: I'd like it if I didn't have in my setup README "Now run these commands" instead of just having it declared
<hoverbear>
LnL: So if that works then maybe that's a good solution.
<gchristensen>
I think andi- knows the status w.r.t. intel / amd updates, manveru, pie_
<hoverbear>
LnL: I think it will do a network fetch every `nixos-rebuild` though? :(
<LnL>
the disadvantage is that you loose some reproducibility, you can’t rollback the unstable part when something breaks
raynold has quit [Quit: Connection closed for inactivity]
<andi->
pie_: update to a stable kernel asap. that is about what you can do right npow
<hoverbear>
LnL: When doing nix-channel stuff you get that?
<andi->
pie_: there are intel and amd micorcode updates (officially unreleased, leaked through redhat/suse/...) but those require kernel changes to work... e.g. lfence and the amd update code is yet to be found in a released kernel
ryantrinkle has quit [Ping timeout: 276 seconds]
<manveru>
andi-: so it's in 17.09?
ma27 has joined #nixos
<gchristensen>
17.09 has patched kernels
<gchristensen>
it has the best we have to offer right now
<LnL>
hoverbear: yes, nix-channel —rollback will downgrade both channels and you can update them separately if neccecary
<hoverbear>
Hm okay
ckauhaus has joined #nixos
<hoverbear>
Now if only I could say in my nixos config "Hey have this channel available"
<LnL>
yeah, that would be neat
<gchristensen>
I'm not sure that is possible though?
<hoverbear>
Kind of silly that it isn't xD
<hoverbear>
I'm currently just doing `unstable = import <unstable>` in my configuration.nix but that means in order to use it you need to add the channel
<hoverbear>
It also means that updating my unstable packages involves a nix-channel update *then* a nixos-rebuild --upgrade
coot_ has joined #nixos
<pie_>
so, more dumb questions, how do i update my kernel
<LnL>
I'm fine with that and I generally only need stuff from unstable in the last few months of the release
<gchristensen>
pie_: nixos-rebuild boot --upgrade, then reboot
<hoverbear>
LnL: I'm kind of at the point where I am questioning if I should just run on unstable
<NixOS_GitHub>
[nixpkgs] andir opened pull request #33610: Mark syncthing012 and syncthing013 as insecure (master...syncthing-insecure) https://git.io/vNLeI
NixOS_GitHub has left #nixos [#nixos]
coot_ has quit [Quit: coot_]
self has joined #nixos
vidbina has joined #nixos
<hoverbear>
joepie91: So if I do that and run `nixos-rebuild switch` twice will it do one network request or two?
<joepie91>
hoverbear: I *think* it will grab it on every rebuild but I forgot if there was any sort of cacheing
<hoverbear>
Hmm
fenton has quit [Ping timeout: 260 seconds]
<LnL>
there's a default ttl of something like 10min IIRC
<infinisil>
It's 1 hour, cc hoverbear LnL joepie91
FRidh has joined #nixos
<hoverbear>
🤔
<hoverbear>
And running nixos-stable and nixpkgs-unstable is not a good idea/
stepcut has joined #nixos
adc has quit [Quit: WeeChat 2.0.1]
<LnL>
depends what you use nixpkgs-unstable for
<joepie91>
hoverbear: I run NixOS stable with a small selection of packages from unstable; notably games and other things for which I really want or need the latest version
<infinisil>
nixpkgs shouldn't be used for nixos
<infinisil>
i mean the nixpkgs-* channels
<hoverbear>
infinisil: What if I need things available only in nixpkgs-unstable like kitty?
<infinisil>
hoverbear: use nixos-unstable
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] peterhoeg pushed 1 new commit to master: https://git.io/vNLvE
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master d2cf59d Peter Hoeg: optar: init at 20150210
<hoverbear>
Hm so nixos-unstable has the same things as nixpkgs-unstable?
<etu>
hoverbear: yes, but with more tests before the channel updates
<NixOS_GitHub>
[nixpkgs] peterhoeg closed pull request #33610: Mark syncthing012 and syncthing013 as insecure (master...syncthing-insecure) https://git.io/vNLeI
NixOS_GitHub has left #nixos [#nixos]
peacememories has joined #nixos
NixOS_GitHub has joined #nixos
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nixpkgs] yegortimoshenko pushed 1 new commit to master: https://git.io/vNLJs
<seequ>
Is there a good way to set up hibernation?
jrolfs has joined #nixos
ma27 has quit [Ping timeout: 240 seconds]
<goibhniu>
seequ: IIRC it should work out of the box as long as you have a swap partition
<seequ>
Oh?
jrolfs__ has joined #nixos
ericsagnes has quit [Ping timeout: 265 seconds]
betaboon has quit [Quit: WeeChat 1.9.1]
<seequ>
Apparently so.
ixxie has joined #nixos
FRidh has quit [Quit: Konversation terminated!]
FRidh has joined #nixos
reinzelmann has quit [Quit: Leaving]
<ckauhaus>
short question w.r.t. my recently opened PR https://github.com/NixOS/nixpkgs/pull/33611/ - The tools now builds on both Linux and Darwin. I've put "platforms.unix" in meta.platforms
<ckauhaus>
is this correct or should I list linux and darwin explicitly?
<gchristensen>
that seems good, ckauhaus
<ckauhaus>
ok, great
<hoverbear>
So it seems like running nixos-unstable channel instead of a mix isn't so bad since you can always rollback
<gchristensen>
it depends if you need timely security updates or not
ericsagnes has joined #nixos
<seequ>
Hm, stable gets them sooner?
<gchristensen>
yes
<nixy>
Security updates should get pushed to both no? Or is this about binary availability?
<seequ>
I guess it has to do with master being broken.
<gchristensen>
the channel won't advance until all the tests pass, and sometimes the unstable channel can have broken tests for days-to-weeks
<hoverbear>
Yeah but it doesn't really seem like mixing stable and unstable is able to be done in a purely declarative way :S
<gchristensen>
NixOS 17.09 has had meltdown / spectre patches for days now, and Unstable still hasn't advanced to contain them for all kernels
<nixy>
hoverbear: What do you mean?
<etu>
hoverbear: Should be perfectly fine to do?
<hoverbear>
nixy: I can't pull in stable.xterm and unstable.firefox in a configuration.nix without doing some imperative commands
<hoverbear>
Namely adding unstable as a nix-channel and updating the channel
<self>
how do I configure xserver to run rootless via starts without display manager?
<nixy>
That's fair, but IMO adding (and maintaing) the channels isn't too bad
<hoverbear>
I spent a few hours attempting it but didn't have much luck. :(
<mpickering>
What exactly is "import-from-derivation"?
<hoverbear>
The `import (fetchTarball ...` method would be fine but I think that will update my packages even when I do `nixos-rebuild` without the upgrade flag, which is undesirable
<mpickering>
:q
bkchr has quit [Ping timeout: 240 seconds]
tpanum has joined #nixos
<tpanum>
I'm trying to compile some gocode to run on a 32-bit system. I'm getting: `gnu/stubs-32.h: No such file or directory`. I've tried installing glibc without any luck. Anybody have a clue on how to fix it? :/
<NixOS_GitHub>
[nixpkgs] dtzWill opened pull request #33615: python-modules: fix httpbin by adding all the deps (master...fix/httpbin) https://git.io/vNLYE
NixOS_GitHub has left #nixos [#nixos]
jensens has quit [Ping timeout: 240 seconds]
<sirkha>
hi, trying to workout how nix-env works with channels. When i use nix-env -qaP <pkgname> it gives me results in the form of nixos.<pkgname>, nixos-unstable.<pkgname> and nixpkgs.<pkgname>. I am using channels nixos (17.09) and nixos-unstable. Is my understanding that the packages with those channel name at the beginning come from those channels correct? Where do the packages that begin with nixpkgs come from?
jrolfs__ has joined #nixos
jrolfs has joined #nixos
randomstrangerb has quit [Ping timeout: 246 seconds]
self has quit [Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org]
randomstrangerb has joined #nixos
<gchristensen>
sirkha: can you run nix-info?
aarvar has quit [Remote host closed the connection]
bfrog has quit [Ping timeout: 276 seconds]
fuzzy_id has joined #nixos
<fuzzy_id>
which tools do you guys use to calculate the `sha256` attribute for a source?
<srhb>
fuzzy_id: nix-prefetch-* or just read the error.
<srhb>
fuzzy_id: (I'm lazy...) :-)
vidbina has quit [Ping timeout: 240 seconds]
bfrog has joined #nixos
<sirkha>
gchristensen, i think it is interesting that it shows me channels(jkha) for nixos-unstable, because, as jkha, nix-channel --list is blank
<sirkha>
oh
<sirkha>
nevermind
<gchristensen>
srhb: there is a factoid for you to use here, `!tofu` :)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dtzWill opened pull request #33616: pythonPackages.structlog: patch to fix tests, build (master...fix/py-structlog-tests) https://git.io/vNLOx
NixOS_GitHub has left #nixos [#nixos]
<LnL>
does tofu mention nix-build --hash
<gchristensen>
I don't think so, LnL, it could use expanding and also maybe a link to a wiki article explaining more in depth
<LnL>
I'll add an example after work, if I didn't forget by then :)
<gchristensen>
:)
marusich has quit [Quit: Leaving]
<sirkha>
so if i have a local nixpkgs clone and I add a symlink to it in .nix-defexpr as 'local' packages will show up as local.<pkgname>? Is that safe/recommended/discouraged?
vidbina has joined #nixos
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dash has left #nixos ["User left"]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] fpletz pushed 1 new commit to master: https://git.io/vNLsV
<NixOS_GitHub>
nixpkgs/master 33a4de2 Franz Pletz: Revert "dhcpcd: 6.11.5 -> 7.0.0"...
NixOS_GitHub has left #nixos [#nixos]
ryanartecona has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 opened pull request #33617: nixos/udev: fix outdated udev rules for network devices (master...udevd) https://git.io/vNLsX
NixOS_GitHub has left #nixos [#nixos]
<srhb>
gchristensen: tofu. :3
<gchristensen>
!tofu
<{^_^}>
To get a sha256 hash of a new source, you can use the Trust On First Use model: use probably-wrong hash (for example: 0000000000000000000000000000000000000000000000000000) then replace it with the correct hash Nix expected.
fuzzy_id has quit [Remote host closed the connection]
<joepie91>
I like how that's just *probably* wrong :)
coot_ has quit [Quit: coot_]
<gchristensen>
I wrestled with if I should do "probably" or just say "a wrong hash" BUT I couldn't say for sure ;)
<woffs>
👍
amir has quit [Ping timeout: 276 seconds]
amir has joined #nixos
ryantrinkle has joined #nixos
johnsonav has quit [Ping timeout: 255 seconds]
coot_ has joined #nixos
swayf is now known as swayf_away
simendsjo has joined #nixos
<efx[m]>
I'm investigating the ideal way to build a package locally and deploy to a server not running nixos. I have it working using closures, but it looks like I'll need to manage a multi user installation. The docs are not terribly clear, and it seems to require manual configuration and add to the complexity of the deployment. Thus, I'm leaning toward the "one click installation" as it seems to best fit my use. Does anyone uses this
<efx[m]>
for server environments?
<gchristensen>
efx[m]: on some systems like hat, I run a standard single-user install of nix on that system and `nix-copy-closure` the build from the build system to the target system
<gchristensen>
efx[m]: indeed, some servers don't even run nix (but this gets bit weird) and you an just send a tarball and extract it to /, putting all the files you need in to /nix/store, and you don't even need nix installed at all. of course, this is as I said, a bit weird
<efx[m]>
gchristensen: cool, txs for the feedback. IMHO the simplicity of the single user seems clearest. I was wondering about how many people just exported it so you could invoke the binary so it "just works"
spietz has joined #nixos
<gchristensen>
the problem is knowing what you an delete later on, should you ever want to clean up disk space
<efx[m]>
true. In my case, I won't care about what else lives in /nix/ so could probably run nix-collect-garbage -d or something
<efx[m]>
well, rather, just remove the /nix/ directory for option #2
<gchristensen>
right
<gchristensen>
I'd go the single user mode
jrolfs has quit [Ping timeout: 246 seconds]
<efx[m]>
roger, txs
<gchristensen>
the least favorite thing I've done is exported a nix closure with all the /nix/store paths, then used binary patching to replace "/nix/store" with "/mypkgsver" and shoved it in to an RPM. (note len(/nix/store) == len(/mypkgsver) so this generally works ok)
jrolfs__ has quit [Ping timeout: 248 seconds]
johnsonav has joined #nixos
basvandijk has joined #nixos
<manveru>
andi-: so i updated, seems like PTI is active now so meltdown should be fixed, but spectre is still an issue :(
<manveru>
had calls to /bin/echo all over the place
bkchr has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vNLWa
<NixOS_GitHub>
nixpkgs/staging 5837d1a Vladimír Čunát: Merge branch 'master' into staging
NixOS_GitHub has left #nixos [#nixos]
coot_ has quit [Remote host closed the connection]
coot_ has joined #nixos
rogue_koder has joined #nixos
<andi->
manveru: spectre will still be topic for the coming kernel releases... the 4.15-rc7 release mail + follow up from thomas gleixner are actually a good read on that ;-)
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to staging: https://git.io/vNLly
<NixOS_GitHub>
nixpkgs/staging 03dca87 Vladimír Čunát: Revert "gnome3.gnome-screenshot: fix build"...
NixOS_GitHub has left #nixos [#nixos]
hotfuzz_ has joined #nixos
<sirkha>
if I am submitting a PR for a new package, and adding that package requires changes to other packages, is it prefered that they all come in one PR, or a PR that changes the other packages and one that adds the new package?
hotfuzz has quit [Ping timeout: 248 seconds]
Sonarpulse has joined #nixos
<andi->
I'd say it is more important to have proper commits (per change) and then the right order of those. Depending on the size I'd decide if it should be multiple PRs
ryantrinkle has quit [Ping timeout: 265 seconds]
<hyper_ch>
sirkha: make individual commits for each change and then a full PR
nschoe has quit [Quit: Program. Terminated.]
swayf has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<sirkha>
hyper_ch, thanks
asuryawanshi has quit [Remote host closed the connection]
jrolfs has joined #nixos
<sirkha>
andi-thanks, pretty small changes, so one PR, likely
<val_>
I installed nixos on a new machine yesterday and for some reason pip doesnt work at all. I cant "pip install pynacl". It works fine on another Ubuntu machine I own.
<val_>
I am trying in a nix-shell, but always get the error Command "python setup.py egg_info" failed with error code 1 in /run/user/1000/pip-build-k42CRB/pynacl/
<val_>
I remember pip worked absolutely fine on my old nixos machine, does anyone know about pip and nixos?
<aminechikhaoui>
val_: why would you use pip in a nixos machine instead of nix ?
civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
matrium has joined #nixos
sigmundv__ has quit [Ping timeout: 240 seconds]
<val_>
Neo--: huh weird, I put that in my configuration.nix, rebuild, and still cant "import nacl" when I open a python shell
<Neo-->
oh it's not like that :)
<matrium>
hi, how can I get a package I've just built with "nix-build" into scope in my nix-env?
<Neo-->
i just have a weekly with my team, i'll show you an example ofhow to do it when i'm done
<val_>
Neo--: that would be much appreciated! Thanks!
goibhniu has quit [Ping timeout: 240 seconds]
pkill9 has joined #nixos
<tilpner>
matrium - How did you build it?
jrolfs has joined #nixos
jrolfs__ has joined #nixos
<xcmw>
Hi, I have a haskell project that uses nix. My default.nix file was generated using cabal2nix. I need nix to fetch one of the items in executableHaskellDepends from github instead of from where ever it is currently getting it from. I know absolutely nothing about nix and my googling did not turn anything up.
<octe>
why does "go get" find Xlib.h when i use nix-shell -p xorg.libX11 but not iny my normal shell even after i've done nix-env -iA nixos.xorg.libX11 ?
<octe>
is it some environment variable?
<tilpner>
octe - Try installing nixos.xorg.libX11.dev (but don't do that)
<matrium>
tilpner: created a release.nix, containing the derivation and passed it to "nix-build"
<octe>
tilpner, that makes no difference
<matrium>
the build was successfull and nix told me the package and binaries were installed into my package tree
<tilpner>
matrium - Try nix-shell release.nix
Tucky has quit [Quit: WeeChat 2.0.1]
<tilpner>
matrium - If you want to install it permanently, I don't recommend nix-env if you're on NixOS
<matrium>
tilpner: the concrete case is that I want to install haskell-ide-engine and use contained "hie" binary in vscode
<matrium>
*use the
<octe>
tilpner, is this not supposed to work?
<octe>
should i always use nix-shell
<octe>
?
<tilpner>
matrium - Why do you have a release.nix. Is haskellPackages.haskell-ide-engine not okay?
<matrium>
tilpner: as far as I know there is no nix-package for haskell-ide-engine
<tilpner>
matrium - I just told you the name
<octe>
thanks sphalerite
<octe>
it's kind of weird though, if i want to run the compiler from emacs i would then have to wrap that too in nix-shell.. or use different emacs instances started from different nix-shells
<macslayer>
Random question: will installing `mesa_drivers` install `libGL`, as well?
<tilpner>
macslayer - I don't think so
<macslayer>
Or do I perhaps need to set `hardware.opengl = true;`?
<macslayer>
The issue I'm trying to solve is that `-lGL` is not found
<tilpner>
hardware.opengl = true; is invalid
<octe>
how do people usually deal with that when developing?
va1entin has joined #nixos
<macslayer>
Oh, yeah, there's no `enable` in `hardware.opengl`.
<octe>
for example now emacs flycheck runs the go compiler to check for errors, but it won't find the libraries
<tilpner>
octe - I manually invoke nix-build from a terminal c.c
<octe>
and then run an editor from there?
<tilpner>
Maybe someone has an actually good setup, I'm curious too
<clever>
octe: does flycheck have a daemon mode?
<tilpner>
The solution can't be to put your entire DE into nix-shell
<tilpner>
(Yes, it could be)
<matrium>
tilpner: sorry, I'm completely new to nixos. I don't know how to get the haskellPackages.haskell-ide-engine in scope of my nix-env. nix-env -i haskellPackages.haskell-ide-engine does not match a derivation
<octe>
not as far as i know.. but flycheck isn't unique, there's lots of autcompletion, etc, that runs stuff
<octe>
that needs the development libraries
<tilpner>
matrium - Use -iA
<clever>
octe: then the editor has to be ran within the nix-shell
<tilpner>
matrium - So nix-env -iA nixpkgs.haskellPackages.haskell-ide-engine. There is a package, but I can't get it to install properly like that
<matrium>
tilpner: nix-env -f '<nixpkgs>' -iA haskellPackages.haskell-ide-engine’ -> error: attribute ‘haskell-ide-engine’’ in selection path ‘haskellPackages.haskell-ide-engine’’ not found
<macslayer>
I thought the generally recommended way was to do `nix-shell -p [dep packages]`, then run `make` or whatever you need to build the source code.
<val_>
Neo--: and then i'd open that nix expression with nix-shell and end up in an environment with pynacl, python, pip and the other declared dependencies, right?
<Neo-->
val_, pip is not installed
<Neo-->
with this expression
<Neo-->
but pynacl is
<Neo-->
usually you try to avoid using pip, but there are ways
<val_>
Neo--, oh okay, it looked like it should be because its declared as a dependency and also as extraLib
<Neo-->
oooh
<Neo-->
yeah
<Neo-->
i thought i removed it :D
<val_>
Neo--, alright I thought I missed something :D anyway thanks a lot!
<Neo-->
add it to extraLibs
<Neo-->
and in the shellhook you can install it
chreekat_ is now known as chreekat
<Neo-->
but if you can you should use something like pypi2nix
<jsgrant[m]>
Does normal 'unstable' support new 'nix' meta-command yet? Thinking of upgrading on my test/toy box of 17.09 once again.
orivej has quit [Ping timeout: 248 seconds]
<tilpner>
bennofs[m] - I gave up on removing all errors. Now GC_mark_from, GC_header_cache_miss and sha256_block_data_order_avx2 take 25% of CPU. I heard a derivation spawns a lot of GC strings, maybe that can be improved on with pooling or bundling
<tilpner>
jsgrant[m] - It has been available as nixUnstable for a while
<tilpner>
You don't need to use the unstable channel for it
<jsgrant[m]>
Okay, neat ty. :^)
Isorkin has joined #nixos
coot_ has joined #nixos
jrolfs has joined #nixos
jrolfs__ has joined #nixos
<tilpner>
jsgrant[m] - You can make it globally active with nix.package = nixUnstable;, but be aware that nixos-install doesn't work with nixUnstable
<tilpner>
But you can also just play around in a nix-shell
ryanartecona has joined #nixos
szicari has quit [Ping timeout: 265 seconds]
<chreekat>
matrium: haskell-ide-engine isn't on Hackage, it appears.
<jsgrant[m]>
My NixOS box is literally living in the closet right-now until I can more gracefully transition; But 'nix' and the seeming ergonomic and/or community attempts as of recent makes me a bit hopeful in the not-so distant future ... after about 5 failed attempts, going in headfirst.
<jsgrant[m]>
The Rust x Cargo builder integration I've seen a bit of recently, too, re-sparked my interest a bit. :^)
<NixOS_GitHub>
nix/master 7b95836 Eelco Dolstra: Improve error message with --repair for untrusted users
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
[nix] edolstra pushed 1 new commit to master: https://git.io/vNLrf
jrolfs has quit [Ping timeout: 256 seconds]
jrolfs__ has quit [Ping timeout: 260 seconds]
jrolfs__ has joined #nixos
jrolfs has joined #nixos
<chreekat>
tilpner: is that recent? I grepped for haskell-ide-engine in my nixpkgs clone and didn't find it. But I haven't pulled in a few days (doing it now but my network is sloowww right now)
mizu_no__ has joined #nixos
coot_ has quit [Ping timeout: 265 seconds]
<tilpner>
Yes, available on unstable and newer
<tilpner>
(I didn't notice that until you asked)
<chreekat>
matrium: ^ in that case you could upgrade your unstable channel and install from it
<tilpner>
bennofs[m] - A trusted user already should have root rights, no?
<gchristensen>
bennofs[m]: on the build box? no
ixxie has joined #nixos
<gchristensen>
bennofs[m]: if they tricked you in to fetching the compromised sudo, then you build the setuid wrapper with the compromised sudo, yes
<matrium>
chreekat: ok, thanks. but can you please tell me how to get the overlay package into scope of my nix-shell anyway? I think it will come in handy in future
<srhb>
Hmm, don't know enough about Mate to really do anything there..
<bennofs[m]>
oh, because you never run nixos-rebuild on the remote, its safe. the new sudo would only be used when doing a new build...
<matrium>
chreekat: I've created a "hie.nix" with my overlay expression. Would I just add something like "pkgs.haskellPackages.haskell-ide-engine = import ./hie.nix;" to my shell.nix?
randomstrangerb has quit [Ping timeout: 265 seconds]
<pikajude>
cmon hackage2nix, update already
<bennofs[m]>
Are there any ideas on how I could solve https://github.com/bennofs/nix-index/issues/19 ? In short, what I need is a fast way to map nixpkgs attribute -> derivation/output hash (but it'd be ok if it only works for things build on hydra, as I fetch the file listings from there anyway)
<ixxie>
manveru: maybe you should because you don't know if any future packages might use it as well
<manveru>
it's not on spdx.org afaict
jrolfs__ has joined #nixos
jrolfs has joined #nixos
<manveru>
but i'll add it anyway
i-am-the-slime has joined #nixos
jb55 has joined #nixos
<manveru>
i'll add 32bit builds after dinner, but can't really test that
rct has quit [Quit: WeeChat 1.9.1]
jrolfs has quit [Ping timeout: 248 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] jeroendehaas opened pull request #33623: libsigcxx: removed patch that forced -std=c++11 in pkg-config file (master...libsigcxx-remove-cpp11-patch) https://git.io/vNLPA
NixOS_GitHub has left #nixos [#nixos]
jrolfs__ has quit [Ping timeout: 268 seconds]
matrium has quit [Ping timeout: 260 seconds]
jrolfs has joined #nixos
jrolfs__ has joined #nixos
jtojnar has quit [Read error: Connection reset by peer]
jrolfs has quit [Ping timeout: 248 seconds]
jrolfs__ has quit [Ping timeout: 248 seconds]
ryanartecona has quit [Quit: ryanartecona]
simendsj` has joined #nixos
simendsjo has quit [Read error: Connection reset by peer]
jrolfs__ has joined #nixos
simendsj` has quit [Client Quit]
oida has joined #nixos
simendsjo has joined #nixos
jb55 has quit [Ping timeout: 276 seconds]
michaelpj_ has joined #nixos
jrolfs__ has quit [Ping timeout: 256 seconds]
bpa has joined #nixos
phreedom has quit [Ping timeout: 248 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] gnidorah opened pull request #33624: Add secondary X server for Nvidia Optimus as alternative to Bumblebee (master...nvidia) https://git.io/vNL17
NixOS_GitHub has left #nixos [#nixos]
ThatDocsLady has quit [Ping timeout: 252 seconds]
ryanartecona has joined #nixos
jrolfs has joined #nixos
jrolfs__ has joined #nixos
mizu_no__ has quit [Quit: Computer has gone to sleep.]
orivej has joined #nixos
shlevy_ is now known as shlevy
erasmas has quit [Read error: Connection reset by peer]
erasmas has joined #nixos
jrolfs has quit [Ping timeout: 240 seconds]
jrolfs__ has quit [Ping timeout: 240 seconds]
<Baughn>
What's the right way to package a Pythong application that consists of a single .py file, without setuptools but which does have package dependencies?
jrolfs has joined #nixos
jrolfs__ has joined #nixos
<Baughn>
(The "documented in readme file" sort of dependencies)
szicari has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] thanegill opened pull request #33626: pythonPackages.rxpy: init at 1.6.0 (master...rxpy-init) https://git.io/vNLDB
NixOS_GitHub has left #nixos [#nixos]
bpa has quit [Ping timeout: 240 seconds]
jrolfs__ has quit [Ping timeout: 240 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] globin pushed 1 new commit to master: https://git.io/vNLDA
<NixOS_GitHub>
nixpkgs/master 4699d04 Robin Gloster: mongodb: 3.4.2 -> 3.4.10...
NixOS_GitHub has left #nixos [#nixos]
jrolfs has quit [Ping timeout: 260 seconds]
reinzelmann has joined #nixos
m0rphism has quit [Quit: WeeChat 1.9.1]
nithor has quit [Ping timeout: 265 seconds]
jrolfs has joined #nixos
jrolfs__ has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Baughn opened pull request #33627: speedometer: init at 2.8 (master...speedometer) https://git.io/vNLyr
NixOS_GitHub has left #nixos [#nixos]
MP2E has joined #nixos
<dhess>
clever: you around?
MP2E has quit [Remote host closed the connection]
MP2E has joined #nixos
Neo-- has quit [Ping timeout: 276 seconds]
<neonfuz>
anyone know how to get the virtualbox kernel modules?
<neonfuz>
oh, do I have to add it to extraModulePackages
jrolfs__ has quit [Ping timeout: 240 seconds]
jrolfs has quit [Ping timeout: 248 seconds]
randomstrangerb has quit [Ping timeout: 268 seconds]
<tilpner>
neonfuz - You shouldn't need to get them manually, virtualisation.virtualbox.host.enable = true will install them for you
randomstrangerb has joined #nixos
<neonfuz>
oh okay, glad I asked lol, thanks
nithor has joined #nixos
leapingfrogs has joined #nixos
JosW has joined #nixos
<hoverbear>
Yay bear cubs first nix shell
nithor has quit [Ping timeout: 255 seconds]
ssmike has joined #nixos
leapingfrogs has quit [Ping timeout: 260 seconds]
<srhb>
Phew. managed to strip down nixops' evaluator. That's something of a learning experience.
Turion has joined #nixos
<gchristensen>
oh!
mojjo has joined #nixos
<mojjo>
how do you find the name of a package a command comes from..? Is there sth better than `ls -la $(which {command})`
michaelpj_ has quit [Ping timeout: 264 seconds]
<samueldr>
readlink -f would be marginally better than using ls
<tilpner>
nix-index
ssmike1 has joined #nixos
<gchristensen>
samueldr: what for?
<samueldr>
(readlink would be better than using `ls`'s output)
<tilpner>
(Although I get failure caused by: database corrupt, frcode error: incomplete frame)
nithor has joined #nixos
<srhb>
gchristensen: The Simple Stateless Deployer.
<srhb>
gchristensen: (Working on that acronym...)
<gchristensen>
neat :D
<clever>
dhess: yes
<srhb>
Along the way I'm learning a lot about how it works. I was puzzled by how it managed to sort-of-evaluate-but-not-quite in order to get the targetHosts, but it turns out to be quite simple once you get at it :)
<tilpner>
For nix-env -q -a --xml --drv-path --out-path --show-trace
<bennofs[m]>
yeah currently nix-index requires every derivation in nixpkgs to evaluate successfully :/
<bennofs[m]>
which in practice, is often not the case
<gchristensen>
good news, though, every PR is checked for that now, bennofs[m]
<bennofs[m]>
I need to find some way to have a more graceful fallback
reinzelmann has quit [Quit: Leaving]
<dhess>
clever: Do you have any Hydra examples that define a channel? I've been looking around in nixpkgs for it. There's a `channel` function in pkgs/build-support/release/default.nix, but it seems like the channels defined in nixos/release* are just aggregates that also make a tarball
<dhess>
i.e., nothing uses that `channel` function as far as I can tell. Anyway it's a bit confusing.
<LnL>
gchristensen: I think nix-index searches more than what's checked by ofborg
<clever>
dhess: the channel is mostly managed by external scripts, that wait until the tests pass on hydra, then publish the tarball hydra made in a certain job
<gchristensen>
LnL: really? it includes unfree and broken packages. what more is there?
<dhess>
clever: are these scripts run by hydra or are they separate services?
<bennofs[m]>
Perhaps I should lobby more for hydra to provide a job name -> store path mapping :D
<bennofs[m]>
(or actually, the reverse)
<clever>
dhess: cron
<dhess>
clever: huh ok. I'll poke around some more to look for those.
<dhess>
gchristensen: sorry for the really dumb question, but I'm a newb when it comes to the multi-user setup on macOS. (I have only ever run Nix on my Macs as single-user.) What's the equivalent to `nixos-rebuild switch` on the Mac, if I want to update things like nix-daemon?
<bennofs[m]>
gchristensen: do you think it would be possible to expose a "builds.xz" that contains all the json data returned from the "build" hydra api endpoint for each build in the evaluation from the binary cache in addition to the "store-paths.xz"?
<LnL>
gchristensen: hmm, weird I've had issues with haskell stuff not evaluating a few times but that should include those
<bennofs[m]>
LnL: haskellPackages is not recurseIntoAttrs though afaik
jrolfs__ has joined #nixos
jrolfs has joined #nixos
<LnL>
it is in release.nix
<gchristensen>
bennofs[m], LnL, dhess I'm about to go and do something that is going to ruin my night. please ping me again and we can talk about these things tomorrow, or a few days from now. this goes for basically anything you'd need me for.
ryanartecona has quit [Quit: ryanartecona]
<dhess>
oh :(
<dhess>
well good luck
<LnL>
euh, that doesn't sound nice :/
<bennofs[m]>
wow nixpkgs has a handleEvalIssue?
<bennofs[m]>
i should not take month-long pauses from nixpkgs dev, missing too many things :D
<LnL>
oh, that might be the difference
<velovix>
Does anyone know how to change the source directory in a buildRustPackage package? cd-ing in preBuild doesn't seem to have the desired effect
<velovix>
in that the source directory still seems to be the root of the repository
jonge has quit [Quit: quit]
<hoverbear>
velovix: Are you working with a cargo workspace?
va1entin has quit [Quit: WeeChat 1.9.1]
aloiscochard has joined #nixos
<velovix>
This package is built with `cargo build`, if that's what you mean. I'm not familiar with the Rust ecosystem so I'm not 100% sure what you're asking
<velovix>
hoverbear: Above message. Sorry, forgot to put you as a mention
<hoverbear>
velovix: Ah ok. Some rust packages are like 'meta' packages so I thought you might be dealing with one of them
<velovix>
hoverbear: I don't believe it's a meta package. It's xi-editor that I'm making an expression for
<hoverbear>
Ahhh super cool
<hoverbear>
I remember that was a bit strange to build
jrolfs has quit [Ping timeout: 240 seconds]
jrolfs__ has quit [Ping timeout: 248 seconds]
szicari has quit [Quit: szicari]
<velovix>
The only complication I've found so far is that the rust code is in the rust directory instead of in the root of the repository, which is why I'm having problems
szicari has joined #nixos
nithor has joined #nixos
ryanartecona has joined #nixos
<catern>
hey #nixos, is there a way to do something like nix-store --export, but get a tarball out?
<catern>
in other words, get a tarball containing the closure of a store path?
<catern>
nix-store --export creates a NAR
<Sonarpulse>
dhess: I think we're missing the relevant config.lo
<Sonarpulse>
can you do another build with --keep-failed?
vidbina has quit [Ping timeout: 265 seconds]
<Sonarpulse>
it will be way way way more than 5 megs
<Sonarpulse>
so we can just fine the file once it fails
romildo has quit [Quit: Leaving]
vidbina has joined #nixos
MichaelRaskin has joined #nixos
<jabranham>
there's no equivalent of /etc/nixos/configuration.nix for users, right?
nithor has quit [Ping timeout: 276 seconds]
<catern>
oh, heh, I can just tar up the output paths dumped by nix-store -R
<catern>
easy!
<dhess>
Sonarpulse: sure, but oddly I cannot get a remote build to honor --keep-failed
<dhess>
which is why I sent you the partial state
<Sonarpulse>
dhess: oh, can you ssh and do local build?
<dhess>
I will try building directly on the arm box
<dhess>
Sonarpulse: I think so, just have to set up the environment there
<Sonarpulse>
dhess: ok cool, thanks for trying
pie___ has joined #nixos
jonge has joined #nixos
<dhess>
Sonarpulse: which config.log do you need?
pie__ has quit [Ping timeout: 240 seconds]
pie_ has quit [Ping timeout: 240 seconds]
<dhess>
there are like 7 or 8 of them
pie___ is now known as pie_
<dhess>
Sonarpulse: ok I think I've got a local build going from nixpkgs master
nithor has joined #nixos
<dhess>
yeah here it goes
<Sonarpulse>
dhess: one sec
hamishmack has joined #nixos
Fare has quit [Ping timeout: 255 seconds]
<velovix>
hoverbear: Do you know of any resources I could look at that might help me solve this problem?
<dhess>
Sonarpulse: ok I have the full failed build dir now
goibhniu has joined #nixos
vidbina has quit [Ping timeout: 260 seconds]
nithor has quit [Ping timeout: 246 seconds]
szicari has quit [Quit: szicari]
<hoverbear>
velovix: I'm still learning myself :(
randomstrangerb has quit [Ping timeout: 246 seconds]
szicari has joined #nixos
<velovix>
hoverbear: Okay, I'll let you know if I figure anything out :)
mbrgm has quit [Quit: ZNC 1.6.5+deb1 - http://znc.in]
mbrgm has joined #nixos
jrolfs has joined #nixos
nithor has quit [Ping timeout: 255 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nix] shlevy opened pull request #1784: Add setPathName primop to copy paths to the store with a given name. (master...setPathName) https://git.io/vNtvM
NixOS_GitHub has left #nixos [#nixos]
<mpickering>
is the "log diff" button on hydra known to not work?
lexsomething has joined #nixos
jrolfs has quit [Ping timeout: 256 seconds]
Itkovian has joined #nixos
<bennofs[m]>
i believe the log button in general does not work?
<bennofs[m]>
at least it didn't for the mac build that I tried
knupfer has quit [Ping timeout: 265 seconds]
<samueldr>
(for nixos.org hydra, script or content blocking extension may break log viewing, from personal experience)
mbrgm has quit [Quit: ZNC 1.6.5+deb1 - http://znc.in]
mbrgm has joined #nixos
<mpickering>
logs work fine but the log diff button 404s
nithor has joined #nixos
szicari has quit [Quit: szicari]
bennofs has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Luz opened pull request #33630: librepcb-unstable: init at 2017-12-29 (master...librepcb) https://git.io/vNtJ3
NixOS_GitHub has left #nixos [#nixos]
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
pjan_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jabranham has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] andir opened pull request #33631: tomcat{7,8,85,9} fixes for CVE-201712617 & tomcat9 stable bump (master...update-tomcat) https://git.io/vNtTN
NixOS_GitHub has left #nixos [#nixos]
jrolfs__ has joined #nixos
amir has quit [Read error: Connection reset by peer]
joachifm has joined #nixos
amir has joined #nixos
orivej has joined #nixos
jrolfs__ has quit [Ping timeout: 240 seconds]
iqubic has joined #nixos
jb55 has quit [Ping timeout: 260 seconds]
jrolfs__ has joined #nixos
jrolfs has joined #nixos
<iqubic>
So it has come to my attention that I have a ton of Generations on my HardDrive. Is it possible to remove all the generations older than a weel?
<iqubic>
s/weel/week
jtojnar has joined #nixos
Turion has quit [Ping timeout: 246 seconds]
<iqubic>
I'd also like to remove the corresponding entries from my Systemd-boot bootloader list if possible.
<globin>
ikwildrpepper, domenkozar, niksnut: based on PRs like (#33631) I'd like to nominate andir (on GH, andi- on IRC) for committer access, in a discussion a few days ago gchristensen also agreed and proposed this
Fare has quit [Ping timeout: 265 seconds]
bfrog_ has quit [Ping timeout: 268 seconds]
mizu_no_oto has joined #nixos
iqubic` has joined #nixos
avn has quit [Ping timeout: 264 seconds]
ryantrinkle has quit [Ping timeout: 268 seconds]
jrolfs__ has quit [Ping timeout: 256 seconds]
jrolfs has quit [Ping timeout: 256 seconds]
iqubic has quit [Read error: Connection reset by peer]
<tilpner>
There's that long <nix/fetchurl.nix> thing, or you can just use fetchTarball on nixUnstable if you don't care about compatibility
<Ralith>
where can I find it?
* Ralith
is trying to bootstrap a strongly reproducible build
<LnL>
main problem is unpacking it so you can actually import stuff :)
<Ralith>
I still haven't worked out how to actually use nixUnstable
Myrl-saki has quit [Ping timeout: 268 seconds]
<tilpner>
Ralith - nix.package = nixUnstable; or nix-shell -p nixUnstable
* Ralith
wishes 1.12 were out already :(
<Ralith>
when is that happening anyway
<LnL>
don't think you need to upgrade the daemon to use the new fetchTarball
<tilpner>
Ralith - I pinned all my sources instead of using channels, on my laptop. PM/ping if you need help/files
<catern>
Ralith: me too
<tilpner>
LnL - What if using Nix via nixos-rebuild?
<catern>
the worst part is that 1.12 has some ugliness with a 1.11 daemon
<catern>
and I have 1.11 deployed everywhere, so I can't in practice use 1.12
<tilpner>
Using 1.12 on my laptop gave me no problems other than nixos-install not working
<Ralith>
LnL: I just don't want to force all my coworkers to jump through extra hoops to use unstable code
<Ralith>
new tools are always a hard enough sell as it is
<tilpner>
Then use the <nix/fetchurl.nix> thing
<LnL>
the thing I linked works on 1.11 (see the else clause)
<LnL>
just saying that it's a lot nicer with 1.12 :)
Fare has joined #nixos
Itkovian has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
goibhniu has quit [Ping timeout: 260 seconds]
<Ralith>
yeah, that's the plan
<Ralith>
I'm just impatient for 1.12 and all its fancy new features
<Ralith>
especially the first-class S3 support
<samueldr>
is there any trick to debug darwin-specific issues when you... don't have a macOS machine?
<LnL>
yeah http2 cache is also a LOT faster
<Ralith>
I can imagine
<Ralith>
downloads spend practically all their time handshaking as-is
<Ralith>
what's the holdup
<LnL>
samueldr: only evaluation issues
<samueldr>
and, for this instance, I have based off my derivation off of eclipse's, which sets meta.platform to unix, while it seems there's no way for it to work on macOS, is the assert they do right?
<LnL>
you can use nix-instantiate -A hello --argstr system x86_64-darwin to test evaluation
orivej has quit [Ping timeout: 268 seconds]
<samueldr>
okay, and the actual "question" is for this specific case, I have based my derivation off of eclipse's, since it's made the same way, is asserting (like I linked) a right way to resolve the issue?
nithor has joined #nixos
enoeht has joined #nixos
<LnL>
using meta.platforms gives the best error messages