<samueldr>
zybell_: have you read the FAQ entry they are following? it's using a channel entry that is downloaded beforehand, not by the expression being built
<gchristensen>
justanotheruser: can you paste the output from `sudo cat /root/.nix-channels` then `sudo cat /nix/var/nix/profiles/per-user/root/channels/manifest.nix`
<zybell_>
and that means the file is somewhere local
<zybell_>
I wanted only to try something beginning with /
<gchristensen>
justanotheruser: now the output of `sudo -i nix-channel --update`
<justanotheruser>
unpacking channels...
<gchristensen>
that is all?
rosa has quit [Quit: rosa]
<justanotheruser>
ya
<justanotheruser>
originally it would say it created symlinks in user environment, but since that's already done it doesn't do anything
<dmj`>
I’m on one nixos machine, and I have setup letsencrypt in a nixos module. I now deploy that module to a new nixos machine, and letsencrypt fails to renew the certificate "Failed to start Renew ACME Certificate for mysite.org”
<elvishjerricco>
dmj`: Can is the machine's 443 (or is it 80?) visible publicly at the domain you're trying to renew?
<samueldr>
dmj`: there will be a systemd service with the domain, you can see its log with `journalctl --unit=that-service-name`
<dmj`>
samueldr: I’m using nixops, how do I ssh in?
<elvishjerricco>
dmj`: Yea that config looks like what I use, I think
<elvishjerricco>
just have to make sure the domain is assigned to the machine already.
<dmj`>
elvishjerricco: I pointed my domain to that IP
<elvishjerricco>
By default it uses the challenge auth method, where letsencrypt checks that you control the domain by checking if the answer to a challenge exists at a certain path at the domain
<elvishjerricco>
so yea I'd expect that to work...
ndrei has quit [Remote host closed the connection]
<samueldr>
dmj`: right now this resolves to 174.138.125.187 here
<dmj`>
which I assigned to the droplet
srdqty1 is now known as srdqty
<zybell_>
but what ttl has your domain?
<dmj`>
samueldr: yea, that’s right
kmicklas has quit [Ping timeout: 256 seconds]
pie__ has quit [Ping timeout: 264 seconds]
<dmj`>
samueldr: but the deploy fails because it can’t start the renewal
<dmj`>
samueldr: so I don’t think the service gets started
<dmj`>
elvishjerricco: I disable the nixos module on the old machine, changed the DNS A record to point to the floating IP, assisgned the floating IP to the droplet
<samueldr>
sorry, I haven't used nixops yet
<elvishjerricco>
dmj`: Oh. I guess it's not even beginning to try. You should be able to journalctl the service anyway
<dmj`>
elvishjerricco: I’m not sure how to ssh in w/ nixops
<samueldr>
and in the nginx options there aren't any allowedTCPPorts
rosa has joined #nixos
srdqty has quit [Ping timeout: 276 seconds]
aarvar has quit [Ping timeout: 240 seconds]
<dmj`>
samueldr: added firewall, redeploying...
<dmj`>
removed forceSSL
<samueldr>
AFAIK forceSSL shouldn't be an issue, fuzzy memory, but something about ACME not caring for invalid certs on https, thus a forced 80 → 443 is not an issue
srdqty has joined #nixos
<zybell_>
but 80 *must* answer first
iqubic has joined #nixos
<zybell_>
after that a redirect is ok
rosa has quit [Client Quit]
<samueldr>
only for the webroot challenge, though that's what's used with simp_le iirc
<samueldr>
dmj`: you may want to subscribe to that and/or somehow use that
nur0n0 has joined #nixos
day|flip has quit [Remote host closed the connection]
DwarfMaster has quit [Quit: WeeChat 2.1]
orivej has quit [Ping timeout: 260 seconds]
DwarfMaster has joined #nixos
xcmw has joined #nixos
mounty has quit [Ping timeout: 240 seconds]
spear2 has quit [Quit: Leaving]
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos
exmonader has joined #nixos
mounty has joined #nixos
<dmj`>
samueldr: awesome find
<exmonader>
Don't know if i should ask this here or the xmonad channel.....
<dmj`>
samueldr: thank you
<exmonader>
But i don't have dmenu installed or in my config
<dmj`>
samueldr: I bet I can override my version of nixpkgs to use this patch
<exmonader>
However using the default dmenu key binding brings up dmenu. I'm a bit confused
rosa has quit [Remote host closed the connection]
<exmonader>
I have xmonad with xmobar and rofi installed
<exmonader>
i tried which dmenu, none was found
<exmonader>
At least not in my Path
rosa has joined #nixos
ditadi has quit [Ping timeout: 260 seconds]
<dmj`>
samueldr: actually I’m not sure how to override nixpkgs in a Nixos module
nD5Xjz has quit [Ping timeout: 240 seconds]
<dmj`>
samueldr: oh I think I know how. We can fetch that version of nixpkgs and then just import it and set le in package overrides to the one in this nixpkgs
rosa has quit [Client Quit]
Lears has quit [Remote host closed the connection]
nuncanada has quit [Ping timeout: 264 seconds]
<dmj`>
actually, can you override nixos modules in packageOverrides?
rosa has joined #nixos
silver has quit [Read error: Connection reset by peer]
iyzsong has joined #nixos
ericsagnes has joined #nixos
<samueldr>
it may be possible to, instead, manually create the cert in the right format once logged-in into the machine
<{^_^}>
[nixpkgs] @dezgeg pushed commit from @bkchr to master « kernelPatches: Adds bcm2835_mmal_v4l2_camera_driver »: https://git.io/vpv43
<samueldr>
then simp_le would be fine with using it
<samueldr>
after having read both that issue and the linked upstream issue, once you are using a cert built like it expects, the issue shouldn't come back
<samueldr>
(though, that's a manual step, ugh)
<samueldr>
cc dmj` ^
<dmj`>
samueldr: isn’t the cert in the store?
<dmj`>
If not would I have to manually edit it on each deploy?
mounty has quit [Ping timeout: 264 seconds]
<samueldr>
it shouldn't be in the store [citation needed] since the store is readable by all!
<samueldr>
it's [citation needed] probably under /var/ somewhere with all the stateful things
<taktoa>
I'm getting a weird error with the standard haskell builder ("ghc-pkg: 1: unrecognised field or section: "<hex-gobblydegook>"")
<taktoa>
you can reproduce with `git clone https://github.com/taktoa/eqsat && cd eqsat && nix-build --no-out-link default.nix`
mounty has joined #nixos
fragamus has joined #nixos
<taktoa>
(that repo is pinned to a specific revision of nixpkgs so I'd be very surprised if it doesn't reproduce)
rosa has quit [Quit: rosa]
rosa has joined #nixos
exmonader has left #nixos [#nixos]
taktoa has quit [Remote host closed the connection]
<zybell_>
.crt should be readable by world (after all your server sends it to each client) the paired .key should only be readable by root unless its protected by passwd
<Lears>
Is there some difference in how NixOS treats minor kernel updates versus the initially installed version? On my machine 4.14.24 worked fine, then updated to 4.14.32 which fails to boot so early there are no logs in journalctl. So I upgraded to 4.15.15 which was fine, but now 4.15.17 has exactly the same issue as 4.14.32 ...
<samueldr>
dmj`: \o/ great! even though it's a bit manual, at least there's a known working workaround
<dmj`>
samueldr: hopefully by the time we deploy again, that fix will be present, o/w we should be good
<dmj`>
samueldr: yes! and it seems that script won’t be ran again if a key is already present
jgt has joined #nixos
<samueldr>
as long as you deploy on the same server, where the state files are present, there shouldn't be any issues even if this isn't fix, afaict
<samueldr>
isn't fixed*
* dmj`
redeploys just to be sure
<dmj`>
ocean> deployment finished successfully
<dmj`>
hip hip hooray \o/
ericsagnes has quit [Ping timeout: 240 seconds]
ivanivan has joined #nixos
raynold has joined #nixos
htafdwes has joined #nixos
<htafdwes>
says can't get status of /nix/default.nix
<dmj`>
htafdwes: hmm?
<htafdwes>
do I have to run like nix update or something? this is on a fresh install, I tried to do a nix-build
<htafdwes>
default.nix doesn't exist
<htafdwes>
oh I see it does it wherever I am, I'm trying to run nix-build -A monte
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<htafdwes>
now it says it doesn't exist in /root/default.nix
<htafdwes>
kind of a bummer, I installed it so I could nix-build monte :-(
<ivanivan>
I want to add an fzf subdir under XDG_DATA_HOME ( ~/.local/share/ ) to hold my fzf history.
<ivanivan>
Is there a way to do this programatically in my nixos config, or with a derivation?
<ivanivan>
Or is this a bad idea?
<boomshroom>
htafdwes: where is the nix file for monte?
<htafdwes>
boomshroom: ah, I found out it's actually in their repos
<htafdwes>
I'll have to get it thanks
rosa has joined #nixos
<boomshroom>
htafdwes: The the monte repo has the nix file, you can actually call it straight from the command line with a fancy expression.
<boomshroom>
htafdwes: IF*
<boomshroom>
If nobody minds me asking, can anyone tell why constructors aren't being inserted into a custom crtn and crti?
mounty has quit [Ping timeout: 240 seconds]
ivanivan has quit [Quit: WeeChat 2.0]
gm152 has quit [Quit: Lost terminal]
mounty has joined #nixos
lo_mlatu has quit [Quit: Connection closed for inactivity]
acarrico has quit [Quit: Leaving.]
rosa has quit [Quit: rosa]
schoppenhauer has quit [Ping timeout: 264 seconds]
sehqlr has joined #nixos
schoppenhauer has joined #nixos
ericsagnes has joined #nixos
<tnks>
I just installed nix-repl in NixOS, and am kind of surprised that it's trying to access /nix/var/nix/db
<tnks>
when I run Nix over Debian, I don't seem to have this problem.
Lears has quit [Remote host closed the connection]
<boomshroom>
I've been writing a custom PID1 and have been using NotOS as a base. I currently have 21 different linux installations in squashfses in my store.
<tnks>
I found the bug report about this... involves NIX_REMOTE not being set by the installer.
<adisbladis[m]>
tnks: You are mixing nix1 and nix2. The repl is included in nix 2.
<adisbladis[m]>
tnks: `nix repl`
<tnks>
adisbladis[m]: oh, I see.
<tnks>
good to know.
<boomshroom>
Nix has spoiled me. I was confused why ["-c" "echo hello"] wasn't working in Haskell.
<{^_^}>
[nixpkgs] @matthewbauer pushed 2 commits to master: https://git.io/vpvEO
<{^_^}>
→ 34821556 by @matthewbauer: text_cmds: init at 99
<{^_^}>
→ 9673ac56 by @matthewbauer: unixtools: add col to utillinux
<Lears>
Well, I tried going back to 4.14 and the minor update 4.14.34 boots, which is nice. I wish I knew why so many kernel versions were failing so early though.
exmonader has joined #nixos
drakonis has quit [Remote host closed the connection]
<exmonader>
Hi again.
<exmonader>
I have xterm.enable set to false
<exmonader>
But i still have xterm
<exmonader>
Is there anyway to get rid of xterm? I already have urxvt installed
dcol has joined #nixos
<dcol>
Hello, I am trying to get my emacs daemon on NixOS to work properly. It was working fine until I tried adding some packages and noticed that the first time I added packaged they were present in the emacsclient, however when I tried incrementing the packages with some additional ones these were not present in emacsclient. It's as if the daemon is using an old build of emacs.. If I start a standalone emacs I can find the newly added
<dcol>
packages there. Anyone with an idea as to why this is happening?
rosa has joined #nixos
nuncanada has quit [Read error: Connection reset by peer]
<boomshroom>
Nice, I try to terminate pid1 twice in my test environment. Nothing happens the first time, the vm exits the second time. Neither time did it print out my debug line.
MP2E has quit [Remote host closed the connection]
odi_ has joined #nixos
odi has quit [Ping timeout: 245 seconds]
Guanin_ has quit [Ping timeout: 240 seconds]
odi_ has quit [Remote host closed the connection]
ertes has quit [Read error: Connection reset by peer]
<{^_^}>
→ bf31818f by @thorstenweber83: python: python-digitalocean: fix build
<{^_^}>
→ 908657ab by @FRidh: Merge pull request #38981 from thorstenweber83/tw-fix-python-digitalocean
rosa has quit [Client Quit]
<boomshroom>
It seems as though it does print on the second call. But that still doesn't explain why it doesn't receive the first one.
winem__ has quit [Ping timeout: 240 seconds]
pie__ has quit [Ping timeout: 240 seconds]
rosa has joined #nixos
rauno has joined #nixos
lesnuages has quit [Ping timeout: 264 seconds]
ma27 has quit [Ping timeout: 260 seconds]
odi_ has quit [Ping timeout: 265 seconds]
jacob_ has joined #nixos
rosa has quit [Client Quit]
<vaibhavsagar>
ij: it compiles, but I'm not sure that it works just yet
gspia has quit [Quit: Leaving]
rauno has quit [Ping timeout: 240 seconds]
seafood has quit [Quit: seafood]
seafood has joined #nixos
seafood has quit [Client Quit]
toby1851 has quit [Quit: Konversation terminated!]
toby1851 has joined #nixos
pie__ has joined #nixos
fragamus has joined #nixos
rosa has joined #nixos
rauno has joined #nixos
jgt has quit [Quit: WeeChat 1.9.1]
<monotux>
does anyone use system.autoUpgrade.enable = true; btw? I'm thinking of adding it to a server (private, not production really) but not sure if wise
<{^_^}>
→ 764d03de by @peterhoeg: pythonPackages.envs: init at 1.2.4
<{^_^}>
→ c859867e by @peterhoeg: pythonPackages.warrant: init at 0.6.1
<{^_^}>
→ f6d9eed5 by @peterhoeg: home-assistant: enable cloud integration
<{^_^}>
[nixpkgs] @dotlambda merged pull request #38653 → home-assistant: make hass cloud integration work → https://git.io/vx5ZM
ma27 has quit [Ping timeout: 245 seconds]
rauno has joined #nixos
<boomshroom>
Goodnight guys!
boomshroom has quit [Quit: WeeChat 2.0]
rauno has quit [Remote host closed the connection]
rosa has joined #nixos
rauno has joined #nixos
seville has joined #nixos
rauno has quit [Remote host closed the connection]
rauno has joined #nixos
adetokunbo has joined #nixos
freeman42x]NixOS has joined #nixos
Ariakenom has quit [Read error: Connection reset by peer]
aarvar has joined #nixos
asuryawanshi has joined #nixos
rauno has quit [Remote host closed the connection]
knupfer has joined #nixos
rauno has joined #nixos
hyper_ch2 has joined #nixos
reinzelmann has quit [Quit: Leaving]
rauno has quit [Remote host closed the connection]
rosa has quit [Quit: rosa]
rauno has joined #nixos
benkolera has quit [Disconnected by services]
benkolera has joined #nixos
rosa has joined #nixos
benkolera has quit [Disconnected by services]
winem_ has joined #nixos
benkolera has joined #nixos
benkolera has quit [Disconnected by services]
bkolera is now known as benkolera
benkolera1 has joined #nixos
asuryawanshi has quit [Remote host closed the connection]
jensens has joined #nixos
wangoe has joined #nixos
hph^ has joined #nixos
rosa has quit [Quit: rosa]
rosa has joined #nixos
reinzelmann has joined #nixos
rosa has quit [Quit: rosa]
rosa has joined #nixos
zipnoob has joined #nixos
alex`` has quit [Ping timeout: 240 seconds]
rosa has quit [Client Quit]
odi has joined #nixos
knupfer has quit [Ping timeout: 265 seconds]
rosa has joined #nixos
knupfer has joined #nixos
ThatDocsLady has joined #nixos
thblt has joined #nixos
joehh has quit [Ping timeout: 256 seconds]
selaux has joined #nixos
gspia has joined #nixos
asuryawanshi has joined #nixos
<Akii>
After taking a break I think I can resume. I've two questions. I want to run multiple services on the nixos server. Each of them may require some Nginx config, and probably come from a git repository, need its own user and do some compiling. Should I have a config file for each service or one for nginx, one for users, etc?
roconnor has quit [Ping timeout: 260 seconds]
<Akii>
and I forgot the second question lol
<makefu>
Akii: the splitting is up to you, i normally try to separate these things in functional blocks, for example if i have a webservice which also requires nginx config then i put both in one file. This can be done because of the intelligent merging of the nix `imports = [];`. This makes nix/nixOS fantastic for sharing configuration and services
<ij>
vaibhavsagar, You didn't try running it? :)
<Akii>
Okay, thanks! Now the next question: Previously I'd just create a user, su into that user, go to its home dir, clone the git repo, compiled the thing and started it via supervisord. As far as I know this doesn't work with NixOS anymore because I must express these step as configuration. Is that correct?
<LnL>
yeah, use a hash to make sure those are what you expect
<sphalerite>
vaibhavsagar: well if you're putting in the hash you might as well use the .keys instead of .keyFiles and paste the keys
<vaibhavsagar>
yeah, I guess so
<sphalerite>
I only do the fetchurl one on throwaway machines where I want a quick-n-dirty way
TonyTheLion has joined #nixos
jtojnar has quit [Quit: jtojnar]
paraseba has quit [Ping timeout: 256 seconds]
chisui has joined #nixos
reinhardt has quit [Quit: Leaving]
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @siers to release-18.03 « ruby-modules: Import gemset if it's a path OR a string. (#38959) »: https://git.io/vpvSj
sigmundv has joined #nixos
Guest74202 has quit [Ping timeout: 256 seconds]
<electrocat>
hey i ran this, but the build seems to have succeeded, why am i getting this error? http://termbin.derhaeg.be/rc8g
<clever>
electrocat: you dont need the callPackage there
thblt has quit [Remote host closed the connection]
<Alling>
If I remove ghc-mod, it builds, but then I get no help in Atom.
knupfer has quit [Ping timeout: 240 seconds]
zybell has joined #nixos
<Alling>
If I change ghc802 to ghc822, ghc-mod does not compile, with errors such as "Couldn't match expected type ‘Distribution.Types.UnqualComponentName.UnqualComponentName’ with actual type ‘[Char]’"
odi has joined #nixos
<Alling>
So yeah ... how do I make this build?
sigmundv has quit [Ping timeout: 268 seconds]
<srhb>
Alling: I guess you upgrade ghc-mod to build with GHC 8.2
rosa has quit [Quit: rosa]
simukis has joined #nixos
<Alling>
srhb: It did work fine up until I upgraded to NixOS 18.03 though. So it should be possible to make it build in one way or another.
<srhb>
Alling: In that case you might want to just pull an earlier 802 set that it builds with.
<Alling>
srhb: How would I do that?
<srhb>
Alling: You could have a 17.09 channel for instance (though it's probably not recommendable to keep it around for long)
<pacmeister>
it tries to run the binary /run/wrappers/bin/mount.ecryptfs_private which doesn't exist by default, so i copied it from /nix/store/ecryptfs/bin to /run/wrappers/bin but then i get an fopen error
<clever>
pacmeister: your not supposed to copy into that directory, but instead configure security.wrappers
<pacmeister>
Inserted auth tok with sig [**********] into the user session keyring
<pacmeister>
Inserted auth tok with sig [**********] into the user session keyring
<pacmeister>
fopen: No such file or directory
<pacmeister>
ERROR: Could not mount private ecryptfs directory
<clever>
does ~/.ecryptfs/${alias}.sig exist?, check on both your user and root
<clever>
and try copying it to roots home
patrl has joined #nixos
Myrl-saki has joined #nixos
<Myrl-saki>
So... uh.... I compiled a ghc8.02.
<Myrl-saki>
For ARM.
<clever>
Myrl-saki: ow!
<Myrl-saki>
It took 2 days.
<Myrl-saki>
But.
<Myrl-saki>
It compiled.
<yorick>
shlevy: I see you added nixUnstable back. please keep a nixUnstable (or nixGit or whatever) package in nixpkgs so we can override it when needed :)
<Myrl-saki>
Anyone wants it? :)
<Myrl-saki>
clever: Did you move your target commit?
<pacmeister>
yes there was .ecryptfs/Private.sig, copying .ecryptfs to the root home gets rid of the fopen issue but i still get an error
<clever>
Myrl-saki: my hydra automatically follows one of the nixos channels
<pacmeister>
mount: No such file or directory
<pacmeister>
ERROR: Could not mount private ecryptfs directory
<Myrl-saki>
clever: Oh. How does that work without a bincache?
<clever>
pacmeister: its possible that the setuid root wrappers are breaking the $HOME detection
<clever>
Myrl-saki: i am the binary cache :P
chisui_ has joined #nixos
<clever>
Myrl-saki: hence, why it takes days to build anything
<Myrl-saki>
clever: I mean
<Myrl-saki>
clever: You're compiling everything, don't you have to do an (almost) full system rebuild every update?
<clever>
Myrl-saki: yes
<Myrl-saki>
clever: Especially since you're in unstable.
<Myrl-saki>
clever: Ugh, sorry, I'm so roundabout in saying things. Lmao.
rosa_ has joined #nixos
<Myrl-saki>
clever: Basically, don't the channels update faster than you can do the rebuild?
<clever>
Myrl-saki: probably
<Myrl-saki>
clever: So how do you keep the hydra updated for a moving target?
<{^_^}>
[nixpkgs] @Mic92 opened pull request #39000 → rustc: disable tests for aarch64 → https://git.io/vpfv1
<srhb>
Yikes, hydra is not a release blocker...
<{^_^}>
[nixpkgs] @Mic92 closed pull request #38999 → rustc: fix aarch64 by disabling two doc tests → https://git.io/vpfea
knupfer has joined #nixos
ericsagnes has joined #nixos
<zybell>
clever:its OOM again but in a subproc not equiped to report it. I think I have noticed that you build multilib. Do you need it (as endresult) or do you build a bootstrap? further in constrained env I routinely patched the makefile to remove the .o when the .a was built.
<clever>
zybell: its just normal nixpkgs with defaults right now
<clever>
dmesg says OOM hasnt happened
knupfer has quit [Ping timeout: 264 seconds]
xy2_ has quit [Ping timeout: 255 seconds]
<zybell>
idk if it works with reproducible builds
orivej has joined #nixos
<mkaito>
does anyone actually use nix to build ruby stuff? I can't even get a working dev env up, because it keeps complaining about bundler versions.
simukis has quit [Ping timeout: 256 seconds]
xy2_ has joined #nixos
<zybell>
how about replacing *all* `make ...` with `make ...;df` (even submakes) to see how full your drive gets?
<clever>
zybell: eek, 500mb free on / right now
xcmw has joined #nixos
<yorick>
can I do types.coercedTo list (x: [x]) (listOf string)?
<zybell>
after a failed make removed its 'tainted' produkts? or are the .o files around?
<yorick>
I mean, coercedTo string
<clever>
zybell: make itself will delete any partially made outputs if a given step fails
iyzsong has joined #nixos
<yorick>
I have to turn a string option into a list, I'd like to do it without breaking everything
<clever>
zybell: then nix will just delete the entire directory when the nix-build fails
<Myrl-saki>
clever: I've been looking for a Linux From Scratch thingy, but LFS is too hardcore for me, can NixOS do the same, but less insane?
pie__ has joined #nixos
znc-user has quit [Client Quit]
<Myrl-saki>
clever: Err, NotOS*
<Myrl-saki>
not-OS?
<clever>
Myrl-saki: you can either use not-os to get a heavily stripped down linux os, or you could just rewrite nixpkgs from scratch if your feeling insane :P
<Myrl-saki>
clever: Anyways, use-case is an OS that boots immediately to Wine.
<clever>
Myrl-saki: that could probably be done, just need to get X working under not-os
hyper_ch2_ has joined #nixos
hyper_ch2 has quit [Ping timeout: 260 seconds]
<Myrl-saki>
clever: How about xinit?
<Myrl-saki>
Err, what does systemd do to start up the X server? The X binary?
<clever>
typically, the display manager is ran
<clever>
and that is responsible for running X
<sphalerite>
how do I get gdb to use the debugging symbols created in a derivation with separateDebugInfo = true; ?
<clever>
sphalerite: need to set one of the search path flags, cant remember the right name
<clever>
Myrl-saki: an entire display manager, written in haskell
<sphalerite>
Myrl-saki: you can use plain nixos too, I did that for a talk last summer. Just made a nixos iso that booted straight into my presentation and wrote it to a USB stick
<Myrl-saki>
clever: Also, the xinit problem is only because of systemd, I believe?
<clever>
Myrl-saki: i think so
<clever>
sphalerite: memory footprint is also at plat, a minimal nixos squashfs is about 300mb, but not-os is 40mb
<clever>
at play*
<Myrl-saki>
clever: Thanks, I'll take a look at all of these. LightDM works great, but I like configurability that I understand.
<Myrl-saki>
sphalerite: Just had a bit of friendly argument with friends that Windows sucks. :P
<sphalerite>
lol
<Myrl-saki>
sphalerite: I mean, I could also point to ReactOS, but they have worse compatibility.
jtojnar has joined #nixos
<Myrl-saki>
sphalerite: It's very amazing that Wine is compatible with a lot of things I use.
<sphalerite>
yeah wine is a crazy project
<Myrl-saki>
sphalerite: Some guy ran a ransomware in Wine and it actually ransomewared his VM.
hyper_ch2 has joined #nixos
nckx has quit [Ping timeout: 264 seconds]
Akii has joined #nixos
<clever>
Myrl-saki: i think i saw that during a pentesting defcon talk
Akii has quit [Client Quit]
<sphalerite>
hahaha
lord| has quit [Ping timeout: 240 seconds]
<clever>
the local IT guy wanted to help the pentester, not knowing he was an attacker :P
<clever>
and popped the usb stick into his ubuntu machine
<clever>
and the malware just worked, lol
<srhb>
It's not possible to allow just Hydra to build without using restricted mode, right?
Akii has joined #nixos
<Myrl-saki>
clever: On the topic of malware, does repairing/verifying the nix-store without a bincache rebuild the package then comparing it?
<sphalerite>
what's the quickest way to get debian running in a container of some description on nixos?
<srhb>
sphalerite: docker run debian ?
<clever>
Myrl-saki: for the verify operation, nix keeps the "correct" hash in db.sqlite
<srhb>
... I hate myself for just recommending docker like that. :P
<sphalerite>
Myrl-saki: "repair" checks the package against the hash it was initially registered int he store with
<sphalerite>
srhb: but I don't have docker :p
<Myrl-saki>
Ohh
<srhb>
sphalerite: Well, decide on a container executor and ask again :P
<Myrl-saki>
On another note, why does the /nix/store filename not include the output in the hashing?
<clever>
Myrl-saki: in theory, malware that is aware of nix can just fudge the hashes in db.sqlite, to claim the backdoor'd binaries are not corrupt
Akii has quit [Client Quit]
binarin has joined #nixos
Akii has joined #nixos
<sphalerite>
Myrl-saki: which output?
<sphalerite>
oh you mean the build result?
<Dezgeg>
you could verify the hash against the signature, however
<Myrl-saki>
sphalerite: Yeah.
alexteves_ has joined #nixos
<sphalerite>
because the build result depends on the store path.
<Dezgeg>
I guess the malware would just remove the signature from the database then, though :P
<Myrl-saki>
sphalerite: Oh okay, makes sense.
<clever>
Dezgeg: which is why ive thought of an extra util to audit the store, double-check signatures against cache.nixos.org, and then just anything that was not signed by nixos
<Myrl-saki>
sphalerite: Proof-of-work idea! Finding a fixpoint that allows that. :D
<Myrl-saki>
I wonder how long the runtime would be..
<clever>
Dezgeg: those are either config files built locally, stuff --fallback built, or infested paths
<sphalerite>
Myrl-saki: yes, I'm sure that would be super speedy :p
<Dezgeg>
yup
<sphalerite>
Myrl-saki: in fact there might not even be a fixpoint. It could loop as well
<jtojnar>
btw, regarding wine, is there an example of something like mkDerivation that would run msi installer and wrap the installed executables to run them with wine?
<sphalerite>
Myrl-saki: but ideally the only difference between each version would be the store path itself.
<sphalerite>
Something something intensional store
<clever>
jtojnar: 2 or 3 people tried to make PR's for that, and they where rejected
<sphalerite>
the model where the output hash is included in the path was already mentioned in Eelco's PhD thesis, but it's complicated to actually implement
acarrico has joined #nixos
<sphalerite>
srhb: using docker in the end, thanks for the suggestion
<zybell>
I think the build process *is* buggy. For once reproducible build means: It is reproducible how much diskspace is needed before build starts to be successful. Simply checking with df wont guarantee success (parallel running other build could interfere)but can guaratee failure. If enough space is available it could be comitted to the build by setting up a loop-device.
<jtojnar>
clever: thanks, that is what I was looking for
oida has quit [Remote host closed the connection]
<sphalerite>
Myrl-saki: it is an adjective…
oida has joined #nixos
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Myrl-saki>
sphalerite: I mean, I thought "intensional store" was to be taken as two separate words
<sphalerite>
zybell: it can't guarantee failure either
<sphalerite>
zybell: a build might be running and finish intermediately, and the cleanup of that one frees enough space for the build to succeed
<{^_^}>
[nixpkgs] @clefru opened pull request #39004 → Remove overrides for dead emacs packages. → https://git.io/vpfIz
Akii has joined #nixos
<zybell>
by breaking of the build sure;-)
<sphalerite>
zybell: or there might be filesystem compression at play, which fudges the reported available space
xcmw has joined #nixos
pie__ has quit [Ping timeout: 264 seconds]
<Myrl-saki>
sphalerite: Does the hashing depend on the hash of the dependencies?
xcmw has quit [Client Quit]
<sphalerite>
zybell: no I mean… start with 6GB space available. build A starts, uses up 4GB. While build A is running, build B, which also needs 4GB, starts too. build A finishes and its intermediate files are cleaned up, leaving enough space for build B to succeed even though there was less than 4GB available when it started
<sphalerite>
Myrl-saki: the derivations for the dependencies, yet. The outputs, no.
<Myrl-saki>
sphalerite: Ah.
Akii has quit [Client Quit]
xcmw has joined #nixos
Akii has joined #nixos
<Myrl-saki>
Thanks for the convo. Really enjoy the convo here a lot. :P
<{^_^}>
[nixpkgs] @yegortimoshenko pushed 3 commits to master: https://git.io/vpfIh
<{^_^}>
→ 415093d4 by @yorickvP: rocksdb: package binary tools
<{^_^}>
→ f2e16a64 by @yorickvP: rocksdb: use stdenv.lib instead of lib
<{^_^}>
→ 63a865c4 by @yegortimoshenko: Merge pull request #38998 from yorickvP/rocksdb-full
triston has joined #nixos
triston has quit [Max SendQ exceeded]
<sphalerite>
^^
Akii has quit [Client Quit]
<srhb>
Hmm. I now have restrict-eval = false, pure-eval = false, allowed-uris = file:///nix/store and I still have hydra-evaluator complaining that access to path /nix/store/... is forbidden in restricted mode. What gives?
<zybell>
sphalerite:I did understand that from the beginning. But the failure guarantee was meant another way: In the moment of check 4G can't be *committed* to that build. That leaves two options:Either wait and commit later when 4G are available, or break of and complain directly to the user.Both ways can be combined by Timeout. There is another aspect of checking for space.Assume only minimal space is missing. the last write fails and the binary is
<zybell>
installed with two byte less. The *reproducible* build did *not reproduce*.
<sphalerite>
but yeah the "graphical installer" image doesn't actually have any graphical tools for installing nixos
<htafdwes>
I'm using nixos graphical 18.03
<htafdwes>
sphalerite: ah, so it is a joke
<htafdwes>
lol
<sphalerite>
htafdwes: it's not a joke, it has useful things like a graphical editor and graphical tools for disk partitioning etc
<sphalerite>
which some people may prefer to use :)
<sphalerite>
but there's no installation wizard or anythign
<htafdwes>
oh, so it's not just supposed to be a black screen with a mouse cursor?
<sphalerite>
no
Neo-- has joined #nixos
<sphalerite>
it's supposed to be a proper graphical shell. Plasma, I think.
justanotheruser has quit [Quit: WeeChat 2.0]
<sphalerite>
zybell: so how would you envision determining the amount of space a build takes up without building it? :)
<sphalerite>
zybell: with zfs you could actually set it up to reserve the appropriate amount of space for a build, but that doens't solve the problem of actually knowing how much space you need
justanotheruser has joined #nixos
<sphalerite>
nix isn't a magic spell that makes reproducible builds appear out of thin air. But it is a powerful tool for making more reproducible
<sphalerite>
builds
iyzsong has joined #nixos
nick_l has joined #nixos
<sphalerite>
htafdwes: does the journal say anything relevant maybe? Try journalctl -u display-manager
katona has joined #nixos
<Dezgeg>
you could probably use something like device mapper thin provisioning to be able to dynamically allocate disk space for a build (or kill it when it goes over the limit), but that would be pretty complicated
<nick_l>
How do I upgrade to 18.03? I replaced 17.09 with 18.03 in my deployment script and it justs hangs after building the system configuration... warning: unknown setting 'signed-binary-caches'
<Dezgeg>
try waiting more
<LnL>
that's normal if you don't use nixos-rebuild boot
<htafdwes>
sphalerite: I usually reboot it so didn't check. I'm guessing it might be RAM limitations, cause it's in a virtualbox only like a gig of ram allocated forit.
<nick_l>
Dezgeg: more than 5 minutes?
<nick_l>
With every other upgrade it at least outputs something.
<nick_l>
Or it uses CPU.
<Dezgeg>
hmmh, yeah that should be definitely enough waiting to see some output
<nick_l>
There is nothing to indicate it is doing something.
<sphalerite>
htafdwes: hm. While 1G is a bit tight, I think it should fail more explicitly.
<symphorien>
nick_l: I have experienced that; the solution was to first rebuild 17.09 with nix.package = pkgs.nixUnstable; then restart the nix-daemon then rebuild with 18.03
<sphalerite>
htafdwes: but yeah try giving it more RAM and see if that works
<nick_l>
symphorien: OK, thanks, but this is a horrible upgrade procedure.
<Dezgeg>
well it should really be debugged why it is hanging
<zybell>
sphalerite:By building it ofc;-) Build it in "unlimited" space, monitor df (MAKE="df&&make" CC="df&&gcc"...)while building,calculate the max and patch it into build expression.
<nick_l>
That is also not documented in the release notes, is it?
<LnL>
nick_l: symphorien: that's not neccecary, the warning is normal and just a warning
<symphorien>
LnL: I mean the hang
<symphorien>
I know the warning is harmless
rauno has quit [Ping timeout: 264 seconds]
odi has quit [Ping timeout: 240 seconds]
<sphalerite>
zybell: that seems like a terribly brittle and complex way of doing it.
<sphalerite>
zybell: but feel free to implement it :p
crooksey has joined #nixos
<Dezgeg>
dm-thinp!
<crooksey>
Any oh-my-zsh users here?
Rusty1_ has joined #nixos
<katona>
Hi, how can I print with a Samsung m2020 printer? I have cups enabled, and after I got back some errors I started googling finding out about Samsung unified printer driver and enabled it as well, but still no luck
<zybell>
It will make sure the expression builds at least once
<sphalerite>
!doesanyoneuse
<{^_^}>
Don't ask if anyone uses software xyz, just ask your specific question. This helps solve your problem quicker, and allows people who haven't used xyz but may still be able to help you to do so.
<sphalerite>
crooksey: ^
<LnL>
symphorien: if that's the case then it's a bug
ndrei has quit [Remote host closed the connection]
onlfhykdd has quit [Remote host closed the connection]
aflciqemr has quit [Remote host closed the connection]
ndrei- has quit [Remote host closed the connection]
ndrei_ has quit [Remote host closed the connection]
<Dezgeg>
or I wonder if with cgroups you could get a notification on disk space use
<nick_l>
LnL: so, should I just follow this plan then?
<Dezgeg>
probably not
<nick_l>
LnL: his*
<nick_l>
This is a bug and I don't understand how anyone has been able to upgrade without doing that.
<LnL>
nick_l: I would recommend nixos-rebuild boot && reboot for an upgrade like this
<katona>
I use plasma 5 and with the GUI I tried adding the local printer and set the ppd file in the official driver as well as the generic one and a similar Samsung one
<nick_l>
LnL: after 8 minutes it started to do something.
Akii has joined #nixos
<symphorien>
LnL: I wonder if it is possible to have nixos tests to test whether one can upgrade from $last_stable to master
<LnL>
don't see why not
<crooksey>
Ok, well I am using oh-my-zsh and I cant find a working configuration
<symphorien>
because they would depend on two different revisions of nixpkgs
<adisbladis[m]>
katona: You want to add your printer drivers to services.printing.drivers
<Dezgeg>
you can very well give hydra two different git repositories as source inputs
<crooksey>
But neither seems to work, the first solution via reddit compiles no problem, but no theme settings apply
Akii has joined #nixos
Fare has joined #nixos
<katona>
adisbladis[m]: That's what I tried to do with the Samsung unified driver , but still no luck. Also, where can I find a list of available drivers?
<crooksey>
And the blog.hrwd option failes as oh-my-zsh cant locate the check_for_updates.sh
<katona>
symphorien adisbladis[m] zybell thanks, I got it working! I needed Samsung unified Linux driver indeed just a specific version.
odi has joined #nixos
reinzelmann has quit [Ping timeout: 255 seconds]
<LnL>
crooksey: use program.zsh.enable and programs.zsh.oh-my-zsh.*
reinzelmann has joined #nixos
rosa has joined #nixos
<katona>
One other issue, libreoffice fresh (6.0) as well as normal (5.something) doesn't use any theme on plasma 5 for some reason.
<katona>
I can't currently test it on gnome , but on an another machine I remember it working perfectly.
<katona>
Currently I'm on 18.03
<Aleksejs>
Hi, can I ask the same question again? I haven't got any response. I'm using nixops to deploy to digital ocean. I also want the other guy to be able to deploy. What is the best way to do it? Should we share state file somehow or is there any better way?
<clever>
Aleksejs: generally, you should have a central machine that you run nixops on, where the state file is kept, just left in the default directoryu
<adisbladis[m]>
Aleksejs: Nixops has an export feature to export to json
<symphorien>
here your default.nix evaluates as pkgTest which is a derivation
<symphorien>
so no --attr is needed
jmiven has quit [Quit: co'o]
<rizary>
so about cabal, when i enter nix-shell, cabal doesn't automatically installed. Is it common? or usually when we go into nix-shell, it's already included cabal?
LysergicDreams has quit [Ping timeout: 264 seconds]
jmiven has joined #nixos
<{^_^}>
[nixpkgs] @yorickvP opened pull request #39008 → Youtrack package and service → https://git.io/vpfCm
<{^_^}>
[nix] @NinjaTrappeur opened pull request #2080 → Nix repl does not interrupt partially typed in expressions on Ctrl-C → https://git.io/vpfCY
<symphorien>
sorry I don't use haskell
<symphorien>
but in general you may stack nix-shell so nix-shell -p cabal (or whatever it is) inside the first shell might work
sehqlr has joined #nixos
<infinisil>
Um, what do I do when I'm trying to build a rust package with rustPlatform.buildRustPackage but the source doesn't have a Cargo.lock file?
<zybell>
If you want some default in nix-shell use shell.nix
<TweyII>
infinisil: ‘cargo update’ will give you one, I think
<infinisil>
But where do I run this?
<infinisil>
Because the dependencies want to get fetched before patchPhase even
LysergicDreams has joined #nixos
<TweyII>
Mmm
<zybell>
rizary:If you want some default in nix-shell use shell.nix
<TweyII>
infinisil: It's a binary, right?
<rizary>
zybell: i did
<rizary>
i mean i have shell.nix
<infinisil>
TweyII: The package I'm trying to nixify? Yes
<TweyII>
infinisil: Accepted Cargo practice is to include a Cargo.lock with any binary
<TweyII>
infinisil: So you might want to consider patching upstream
<zybell>
in that dir?
<infinisil>
TweyII: Yeah, but it should be patchable in nix still
<ryantrinkle>
we're doing a lot of installing around the office lately :P
<clever>
ryantrinkle: this nixos module configures a complete router to act as an ipxe server
<infinisil>
sphalerite: Thanks, I'll give it a try
<ryantrinkle>
clever: ah, yes, srhb pointed me at that
<sphalerite>
infinisil: it doesn't work for neovim btw, the settings are different
<clever>
ryantrinkle: it also includes the justdoit script in the netboot image, so you literally type "justdoit" at the root shell, and it installs nixos
<ryantrinkle>
but i'm a little confused: how does this actually get the pxe image?
<ryantrinkle>
oh nvm, i see it now
<clever>
ryantrinkle: line 11 loads the netboot installer profile, which generates an initrd that contains the entire rootfs
<clever>
ryantrinkle: 36 makes a tftp root directory containing the bootloaders
<clever>
ryantrinkle: 42 makes an http dir with the kernel and initrd
<ryantrinkle>
yeah, that's the part i missed the first time around :)
<ryantrinkle>
line 36
<infinisil>
sphalerite: Using normal vim
<sphalerite>
perfect then
<infinisil>
Although I've been wanting to switch to emacs evil mode for more things
<sphalerite>
yep that's what I've been doing too :D
<infinisil>
But vim is just really convenient in the terminal
<sphalerite>
emacs -nw :D
<clever>
ryantrinkle: originally, i had used the netboot_server to spread nixos from my old laptop to my new laptop
<infinisil>
sphalerite: Yeah, but I think there can be problems with key bindings and other stuff
<Dezgeg>
I hear spacemacs is the new hype thing for vim-on-emacs
<sphalerite>
yeah
<infinisil>
Also I really like how emacs still gets that terminal look but isn't as constrained as a terminal
<Dezgeg>
i.e. emacs as the kernel and vim bindings as the distribution
<sphalerite>
yeah I think I need to just invert the relationship, and run my shells in emacs rather than emacs in my shells :D
<infinisil>
(I'll shift this vim/emacs talk to #nixos-chat)
rosa has joined #nixos
<ryantrinkle>
clever: cool :)
rosa has quit [Client Quit]
sehqlr has joined #nixos
<clever>
ryantrinkle: only thing i havent gotten to work yet, is efi based pxe
<ryantrinkle>
ah
nckx has joined #nixos
<clever>
ryantrinkle: but justdoit uses grub in removable efi mode, so the bios thinks the fresh install is on removable media, and boots it without efivars being configured
ryanartecona has quit [Quit: ryanartecona]
kreisys has joined #nixos
<clever>
ryantrinkle: justdoit also has a range of config flags, to support virtio, sata, or nvme, with or without luks, and legacy or uefi
xcmw has quit [Ping timeout: 264 seconds]
rosa has joined #nixos
<kreisys>
hello, I'm trying to setup hydra but it seems like it insists on running all jobs in restricted mode?
<kreisys>
Is there a way to switch it to the nix 2.0 pure mode?
Myrl-saki has quit [Ping timeout: 256 seconds]
<sphalerite>
kreisys: basically not without modifying it
<sphalerite>
kreisys: srhb was just struggling with the same thing earlier
<sphalerite>
hm, debugging stuff is a pain. A lot of packages seem not to have debugging symbols enabled
<kreisys>
clever: I saw that.. I guess I could easily enough comment that line out but then how would I enable pure mode?
<kreisys>
I don't want to just enable impurities because I'm being lazy. that's being lazy 😅
drakonis_ has quit [Ping timeout: 255 seconds]
<clever>
sphalerite: hmmm, i did the cursor change thing in tmux on xfce-term, and its doing weird things
Mateon3 has joined #nixos
<clever>
sphalerite: oh, its screen + xfceterm, after some delay, it prints some garbage to the screen
Mateon1 has quit [Ping timeout: 264 seconds]
Mateon3 is now known as Mateon1
<clever>
sphalerite: every minute or so, tmux seems to try to change the cursor, and the \e[ prints garbage, followed by the 6 q
ma27 has joined #nixos
<sphalerite>
clever: nice
<clever>
also, xfce-term claims TERM=xterm!?
<clever>
still happens without screen in the loop
<clever>
also happens without tmux in the loop, so its xfce-term not supporting the control code, and tmux just blindly repeating it
mounty has quit [Read error: Connection reset by peer]
jensens has quit [Ping timeout: 264 seconds]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mounty has joined #nixos
rosa has quit [Quit: rosa]
drakonis_ has joined #nixos
rosa has joined #nixos
lonokhov has joined #nixos
<iqubic`>
:bd
iqubic` has left #nixos ["ERC (IRC client for Emacs 25.3.1)"]
ashkitten has quit [Quit: brb]
<clever>
havent seen :bd before
alex`` has joined #nixos
Ariakenom has joined #nixos
crooksey has quit [Ping timeout: 276 seconds]
ashkitten has joined #nixos
LysergicDreams has quit [Ping timeout: 240 seconds]
drewr has quit [Remote host closed the connection]
ambro718 has joined #nixos
LysergicDreams has joined #nixos
<zybell>
clever:idk which it is,but there is an old code to ask a term for its version string. How it works:the app prints the code,the term types its string framed in yet another code; so could it be:progr wants to set cursor,tmux repeats faithfully,term interprets as asking for sth,term answers,*that* code crashes tmux(because it contains ^B)?
<clever>
zybell: in the compatability config for xfce-term, the emulation setting is set to xterm
knupfer has joined #nixos
rosa has quit [Quit: rosa]
<sphalerite>
hm. The debugging experience with nix is kind of crap. Most libraries don't have debug info :/
<clever>
zybell: changing it to foobar seriously breaks the terminal, lol
<zybell>
not on your comp;there you only need to turn readline off;I mean sphalerite's
pie_ has quit [Ping timeout: 264 seconds]
xcmw has joined #nixos
<dcol>
I'm trying to make the emacs daemon work.. When I declared the packages for it in my configuration.nix the first time it worked fine - the packages were all present in emacsclient. However when I then tried to add another package to the config it did NOT work. After some digging around I noticed that the daemon is using the same nix-store binary it initially build for emacs and refuses to change to the new one. If I launch emacs
<dcol>
normally then the packages are present. Anyone have an idea as to why this is happening?
<sphalerite>
I'd expect that to install the debugging symbols.
<sphalerite>
Not the library.
<sphalerite>
or tool
<tilpner>
The debug symbols for bin/openssl seem to be inside /lib/debug
chisui has joined #nixos
<sphalerite>
tilpner: but /tmp/debug/lib/debug doesn't exist
<tilpner>
So I don't really see how it could differentiate between lib/bin symbols
<sphalerite>
I don't care about it differentiating between lib and bin symbols, I care about it installing the debug output, not the lib or bin output, to the profile
<johnw>
shlevy: would you have time to fix the lsof build on darwin (again)?
<{^_^}>
→ bd77849b by @dezgeg: nixos/installer/channel: Add some files that the channel also has
<{^_^}>
→ 333a24a6 by @dezgeg: nixos/release.nix: Add SD images for ARMv6 + ARMv7
<tilpner>
No idea, sorry
<kreisys>
ikwildrpepper: thx. that "feels" like the more correct way to do it.. the manual is confusing since it just gives an example using <hello> but no explanation of how inputs are supposed to work
endformationage has joined #nixos
<shlevy>
johnw: what's up with it?
<johnw>
shlevy: it broke again; trying to get the error now to paste into GH
Lears has quit [Remote host closed the connection]
<johnw>
do paths automatically coerce to strings?
<shlevy>
johnw: yes, depends on context
<johnw>
in which contexts?
<shlevy>
e.g. in a derivation { foo = ./foo; } will add it to the store, as will "${./foo}", whereas (toString ./foo) will do "/home/shlevy/foo" or whatever
<johnw>
in context where a string is required, can ./foo be passed?
Synthetica has quit [Quit: Connection closed for inactivity]
<shlevy>
Yeah, it will get added to the store and the string will be that store path
<johnw>
can null be coerced to a path?
lo_mlatu has joined #nixos
<johnw>
I guess I'll find out what nixpkgs actually uses in time
<{^_^}>
→ 9751771c by @eonpatapon: dockerTools.buildImage: add /nix/store with correct permissions
<{^_^}>
→ c4dea092 by @LnL7: Merge pull request #38837 from eonpatapon/fix-38835
MarcWeber has joined #nixos
<shlevy>
Nice :) That looks right
<shlevy>
johnw: I've never seen the #extern directive before :o
sehqlr has quit [Ping timeout: 276 seconds]
<johnw>
what is #extern?
<MarcWeber>
Despite adding reiserfs to availableKernelModules and kernelModules its not includeded in the initrd image so booting fails - is there something trivial I missed?
periklis has quit [Ping timeout: 255 seconds]
<orbekk>
MarcWeber: a guess: boot.initrd.kernelModules
<{^_^}>
→ dfa8a5e9 by @LnL7: docker-tools: add a test for permissions issues with AUFS/overlay
<{^_^}>
→ f3353fff by @eonpatapon: dockerTools.buildImage: add /nix/store with correct permissions
<zybell>
shlevy: the commit from johnw l152 doubles l156 (in the new file) and the following too. I think thats not intended.
<johnw>
zybell: which commit is this?
freeman42x]NixOS has quit [Quit: Leaving]
<zybell>
df9092...
<johnw>
i'm sorry, but which project?
<johnw>
i don't find that commit anywhere in nixpkgs
<zybell>
scrollback
<johnw>
ah
<johnw>
ah, I see whaty ou mean
bkchr has joined #nixos
<johnw>
is there a way to get nix's support for Haskell (I'm using developPackage) to automatically determine dependency versions from a cabal.config freeze file?
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos
<johnw>
when I do overrides with Haskell, how do I get them to be recursively applied everywhere?
<johnw>
I don't want mixed versions of some libraries
<htafdwes>
okay I did the curl |sh thing and nix installed, but I have no idea what to do next.
<htafdwes>
I don't have access to any nix commands.
<tilpner>
It tells you to source another file. Did you do that?
<tilpner>
(Or you could login again)
<htafdwes>
tilpner: uh, which file? I don't know which terminal it was now.
<htafdwes>
okay i'll try relogin
<Lisanna>
hey wait a second, exportReferencesGraph is impure!
<Lisanna>
that's not fair ):
<htafdwes>
nice it worked :-D
<htafdwes>
hey, do you think I could like incrementally move to nixOs from ubuntu by adding more packages from Nix and taking them out of Ubuntu?
<htafdwes>
for instance firefox
katona has quit [Quit: Connection closed for inactivity]
<Lisanna>
My current hypothesis is that if you discard a store path and pass that into exportReferencesGraph that if it exists anyways it will work, even if it's not part of your closure
<LnL>
sure, you can use nix as a second package manager
<Lisanna>
but if it doesn't exist, since it's not part of your closure, exportReferencesGraph complains
<Lisanna>
ugh, replaceDependency is so weird
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos
<htafdwes>
muwahaha I did it! now using nix firefox instead of ubuntu firefox.
acarrico has quit [Ping timeout: 268 seconds]
<Lisanna>
Oh... and IFD + exportReferencesGraph is a bad combination, since you can end up with a string containing /nix/store paths without the appropriate string context
<Lisanna>
that /nix/store path shouldn't be allowed in the sandbox
<{^_^}>
[nixpkgs] @mbode opened pull request #39014 → adr-tools: init at 2.1.0 → https://git.io/vpJvg
<tilpner>
It doesn't for me (with updated path). Are you sure sandboxing is enabled?
<tilpner>
grep 'sandbox ' /etc/nix/nix.conf
<Lisanna>
tilpner yes, I always have sandboxing enabled
<Lisanna>
is that path in your /nix/store?
mounty has joined #nixos
<Lisanna>
wait, with updated path?
<Lisanna>
oh, so you used a glibc that exists on your system
<tilpner>
I nix-built hello, and pasted that path
<Lisanna>
oh wait, glibc is probably already always in the path
<Lisanna>
that was a bad test... yeah, hello works fine
guest__ has joined #nixos
flyx has joined #nixos
<Lisanna>
woo, builtins.storePath can be used to restore a string context to a /nix/store path string
<nh2[m]>
hey, when I change `boot.loader.grub.storePath`, should `nixos-rebuild switch` run `grub-install`? Because it doesn't seem to and my system stops booting when I do that
<sphalerite>
Lisanna: is it just me or do you always seem to be performing various degrees of black magic? :p
<Lisanna>
sphalerite some combination of performing and getting burned by :p
<sphalerite>
:D
<sphalerite>
also, does anyone know a nice tool I could use to make C library calls in an interactive shell?
<Lisanna>
sphalerite CERN or something has something like that I think
<sphalerite>
Lisanna: as in the nuclear research centre..?
<Lisanna>
yeah, it's either CERN or a related group
<sphalerite>
thanks for the pointer, I'll have a look at it!
<Lisanna>
yeah, cling was it
<sphalerite>
not in nixpkgs of course :(
<htafdwes>
lol, I accidentally allocated only 2GB for my root partition, and 8GB to my swap
<htafdwes>
uh, hoow do I fix that?
<tilpner>
Do you have LVM setup?
<sphalerite>
htafdwes: you'll need to boot into the installer again
<sphalerite>
htafdwes: also I hope the swap partition is after the root one?
<htafdwes>
sphalerite: nah, I think it was because the instructions said to fix partition order, but they also said to make the swap first.
<htafdwes>
sphalerite: yes it is after
<Lisanna>
sphalerite so, the awful thing that burned me today was that I was doing an IFD on an exportReferencesGraph result, losing important string context information. I was then passing this result back into exportReferencesGraph, which is not 100% pure, such that if you pass a "/nix/store" string into it without proper string context and it exists, it will still work, but if it doesn't exist, it will throw an error. So when I built my derivation
<Lisanna>
locally, it worked, because all the references *happened* to already exist on my local system, but when I built it remotely, not all of them existed *all the time*, so *sometimes* it would break
<sphalerite>
actually maybe just redoing the whole install would be simpler
hotfuzz_ has joined #nixos
<sphalerite>
unless you've got important stuff in your /home or whatever
<sphalerite>
Lisanna: bleh, the cling repos don't seem to be publicly accessible >_>
sanscoeur has quit [Ping timeout: 256 seconds]
<Lisanna>
):
<sphalerite>
I think they are *supposed* to be, they just aren't
<sphalerite>
oh wait no, it just requires auth via HTTPS, but not via HTTP
<htafdwes>
hmmm but I had to make a new swap partition, is there a way of getting it to recognize that?
<htafdwes>
other than editing /etc/fstab
<sphalerite>
htafdwes: swapon it, then rerun nixos-generate-config
<htafdwes>
sphalerite: awesome
<htafdwes>
wait, but would that destroy what I changed in the config?
astiles has joined #nixos
<htafdwes>
also, does nix work on cygwin?
<htafdwes>
ah I see it doesn't overwrite it nice
<sphalerite>
htafdwes: yeah it only overwrites hardware-configuration.nix
<htafdwes>
so do I just put the packages I want installed by default into my configuration.nix or is there a different way? like a user defined version. cause I got headless servers, and some desktops that use graphical applications.
astiles has quit [Remote host closed the connection]
<zybell>
ssphalerite:python has a FFI and a repl,so you can import any function and call it how you want.
ThatPako has joined #nixos
<zybell>
sphalerite:simpler solution: python has a FFI and a repl,so you can import any function and call it how you want.
chreekat has joined #nixos
<sphalerite>
zybell: yeah, I want to mess with curses though and it seems to use curses for its own REPL
<sphalerite>
stuff went wonky when I tried that
<sphalerite>
grumble grumble libraries with global state
<Lisanna>
sphalerite zybell oh don't try to turn python into a C repl
<Lisanna>
you'll just end up crying in a corner
<sphalerite>
haha
srdqty has quit [Ping timeout: 260 seconds]
<htafdwes>
how do I find a list of what options are available for like xserver.displayManager ?
<sphalerite>
htafdwes: there are display managers like sddm, slim, lightdm (which do the login screen), desktop managers like plasma and gnome (which do a full desktop), and window managers like i3 which just manage windows
<htafdwes>
ThatPako: awesome!
<ThatPako>
yep, that's what the WM is called
<sphalerite>
lol
blankhart has quit [Read error: Connection reset by peer]
blankhart has joined #nixos
chreekat has quit [Quit: quitting]
<infinisil>
Okay I am confused
<zybell>
Lisanna:*I* do not such a thing. but python is packaged for nix already and has a cool tk based IDE, where ncurses cant interfere with your cmdline.
<sphalerite>
htafdwes: it's already included by default, and you configure it via configuration.nix
<htafdwes>
sphalerite: ah, so everything goes into that file? I thought I saw some examples of people importing other files like for nginx or nextcloud
vaninwagen has quit [Ping timeout: 268 seconds]
sanscoeu_ has quit [Remote host closed the connection]
k3k2k3 has left #nixos [#nixos]
<sphalerite>
htafdwes: you can import additional files, but all the modules in nixpkgs/nixos/modules are included automatically
apostolis has joined #nixos
sanscoeur has joined #nixos
<sphalerite>
which is why many of them have enable options, which prevent the stuff from actually being activated
Arcaelyx has quit [Ping timeout: 260 seconds]
<apostolis>
Silly question but how do I assign a name to a derivation. I am doing d = derivation {} but it returns an error.
alexteves_ has quit [Ping timeout: 240 seconds]
ryanartecona has quit [Quit: ryanartecona]
oida has joined #nixos
<sphalerite>
apostolis: derivation { name = "foo"; }
<{^_^}>
→ 2c1f2c0f by @volth: babelstone-han: 11.0.0 -> 11.0.2
<{^_^}>
→ 691cc358 by @Mic92: Merge pull request #39010 from volth/patch-125
<apostolis>
Ok, thanks.
srdqty has joined #nixos
ihar has quit [Quit: WeeChat 2.0]
szicari has quit [Quit: szicari]
vidbina has joined #nixos
joehh has quit [Ping timeout: 264 seconds]
apostolis has quit [Quit: WeeChat 1.6]
nick_l has joined #nixos
<nick_l>
Why can't release notes be complete in NixOS?
<nick_l>
In version 17.09 certain packages were available (e.g. attic), which do not exist in 18.03.
Sonarpulse has joined #nixos
<nick_l>
TeX also lost 5 packages or so.
ma27 has joined #nixos
<johnw>
avn: I'm not interested in stack; there should be a way in Haskell overrides to say "make *all* uses of conduit use this version, recursively)", I just don't seem to know how to say it...
<samueldr>
nick_l: about attic, it is mentioned in the release notes, that it was removed
ouvve has joined #nixos
jtojnar has quit [Read error: Connection reset by peer]
<nick_l>
samueldr: OK, then the TeX packages.
<samueldr>
(though it looks accurate that not everything is mentioned)
jtojnar has joined #nixos
ryanartecona has joined #nixos
<samueldr>
nick_l: could you name one, I want to check something
<ouvve>
Is there a good way to set `srcs` based on a glob?
<nick_l>
samueldr: You are in luck: collection-genericextra
<nick_l>
samueldr: this is another one.
<nick_l>
samueldr: collection-genericrecommended # and yet another one
<nick_l>
samueldr: Uses via texlive.combine { <snip> }
<nick_l>
Used*
<nick_l>
samueldr: I don't really care about those packages, but I do find it annoying that this stuff can't be automatically taken care off.
<nick_l>
Something like "You just switched to 18.03. you were using the following expressions which are now not available anymore, but those are no longer supported". Currently, the setup is that it takes minutes to compute that something is not available and then I can try again.
<nick_l>
Since it is NixOS, one can revert anyway. Not bothering the user with every little detail seems to be more useful than throwing an error on such unimportant things.
<nick_l>
Now, if *all* gcc compilers were to be broken or unavailable, that would be something to be concerned about, but if gcc-bleededge-edition would not be there, who cares? Nobody?
<nick_l>
Documentation is at most a substitute for informing the user of relevant information. The people who don't use certain removed software don't want to hear about it and the ones who do prefer to just see it in their terminals.
<ryantm>
Mic92: Sorry, that wasn't you, and it was a long time ago.
<{^_^}>
[nixpkgs] @pbogdan opened pull request #39020 → cantata: build with vlc_qt5 → https://git.io/vpJcf
alex`` has quit [Ping timeout: 256 seconds]
htafdwes[m] has joined #nixos
<Ralith>
if anyone is having frustrating and mysterious font issues after upgrading to 18.03, rejoice! the problem is that ~/.cache/fontconfig was invalidated and apparently fontconfig doesn't handle this gracefully; nuke it and all will be well
<htafdwes[m]>
how about mimetypes like default applications for xdg open, and when clicking. how to configure?
<nh2[m]>
what web server software runs cache.nixos.org?
jperras has quit [Ping timeout: 240 seconds]
<nh2[m]>
(I mean the upstream of cloudfront)
<jgt>
I have this shell.nix: http://ix.io/17YG It contains a call to mkDerivation for my package 'riskbook'. I don't think I need this, because the same derivation is available in the haskell-packages.nix expression I import. How do I call the derivation in the other file?
<jgt>
I replace `drv = variant (haskellPackages.callPackage f {});` with `drv = variant (haskellPackages.riskbook);`
jackdk has joined #nixos
jtojnar has quit [Ping timeout: 264 seconds]
jperras has quit [Read error: Connection reset by peer]
ericsagnes has quit [Ping timeout: 256 seconds]
<iqubic>
Does anyone know to command line interface to ALSA. I can't seem to find the commands to modify my sound levels.
jperras has joined #nixos
<iqubic>
Specifically I want raise, lower, and mute from the command line. Trying to make my media keys work.
<iqubic>
Forgive me if this is not the right place for this.
<boomshroom>
iqubic: Those are the same commands that the default media keys configuration uses.
<iqubic>
Define default.
<iqubic>
Because I use an obsucre window manager that doesn't support this kind of thing.
timokau has quit [Ping timeout: 240 seconds]
blankhart has quit [Ping timeout: 265 seconds]
<iqubic>
*obscure
<boomshroom>
iqubic: As in, if you include sound.mediaKeys.enable = true; in configuration.nix, it will automatically enable the volume up, down, and mute buttons.
<iqubic>
It will? How does that work?
<iqubic>
How do you bypass the window manager and go straight to commands?
<boomshroom>
iqubic: It's defined in github.com/NixOS/nixpkgs/blob/release-18.03/nixos/modules/services/audio/alsa.nix
<boomshroom>
iqubic: It uses actkbd.
<iqubic>
Do I have to reboot my machine for this to work, or will a simple "nixos-rebuild switch" work?
cheshircat has quit [Ping timeout: 240 seconds]
<boomshroom>
iqubic: nixos-rebuild switch is all that's nessisary. Mind if I ask which window manager you're using?
jperras has quit [Ping timeout: 256 seconds]
<iqubic>
EXWM
jperras has joined #nixos
<iqubic>
And the nixos-rebuild switch didn't work for me.
<boomshroom>
iqubic: Reboot usually makes things work even if it shouldn't be nessisary. If not, the commands are listed in the file I linked.