2019-03-28
19:46
<
clever >
pie_: just `exit 1`
19:34
<
clever >
but that predates the initrd even having network support
19:34
<
clever >
bgamari_: last time i delt with iscsi, it was my rootfs, so the initrd handled that
19:31
<
clever >
gchristensen: systemd-analyze plot
17:26
<
clever >
nefix: search nixpkgs for wrapProgram
17:25
<
clever >
symphorien: not sure how you would handle that
17:24
<
clever >
nefix: modify all of the binaries to be wrapped with a bash script, to put that package into PATH
17:23
<
clever >
nefix: you usually bake the paths into your package, either direct references, or prefix PATH with a bash wrapper
17:22
<
clever >
nefix: nix doesnt allow that to happen just from installing a package, you must enable a nixos module to do that
17:21
<
clever >
only nixos can manage /etc/, packages cant do it
17:21
<
clever >
nefix: yeah, nixos option
17:18
<
clever >
nefix: environment.etc
17:17
<
clever >
symphorien: only at the bash level
16:54
<
clever >
samueldr: you also passed your own -enable-kvm, which isnt conditional
16:54
<
clever >
samueldr: the qemu-kvm bash script, simply tests for /dev/kvm, and conditionally passes -enable-kvm, so it wont fail like that
16:39
<
clever >
samueldr: on nixos, kvm is 666
07:20
<
clever >
Guest29214: you want to set isNormalUser = true; on the acct
2019-03-27
19:25
<
clever >
ah, thats still os, not arch
19:24
<
clever >
generic-builder.nix:, enableHsc2hsViaAsm ? stdenv.hostPlatform.isWindows && stdenv.lib.versionAtLeast ghc.version "8.4"
18:01
<
clever >
,pr judson__
17:38
<
clever >
then complain to the one that did
17:37
<
clever >
iqubic: include the required fonts inside the flatpak
17:37
<
clever >
iqubic: it likely ignores all system stuff
17:34
<
clever >
__monty__: what if i want to allow older?
15:31
<
clever >
Xyliton: you must add its lib dir to LD_LIBRARY_PATH
15:13
<
clever >
aplainzetakind: did you also ensure to stop nix-daemon?
15:11
<
clever >
blackriversoftwa: it may not obey /etc/group, and then just not put things into kvm
15:10
<
clever >
blackriversoftwa: nix-daemon is doing some container-style logic to change what uid the build thinks it has
15:07
<
clever >
blackriversoftwa: for me, it outputs: uid=1000(nixbld) gid=100(nixbld) groups=100(nixbld)
15:07
<
clever >
blackriversoftwa: nix-build -E 'with import <nixpkgs> {}; runCommand "foo" {} "id"'
15:06
<
clever >
blackriversoftwa: it might be that nix-daemon doesnt obey extra groups, would need to run `id` inside a build
15:01
<
clever >
you could also test that on your ubuntu
15:01
<
clever >
blackriversoftwa: on nixos, its 666, so you dont need a special group
15:01
<
clever >
crw-rw-rw- 1 root root 10, 232 Mar 27 11:58 /dev/kvm
15:00
<
clever >
blackriversoftwa: nothing looks out of place there
14:55
<
clever >
blackriversoftwa: what about the output of `nix show-derivation /nix/store/bvapz42iyv546qfmrwrb7aqskqia7ja6-digital-ocean-image.drv` ?
14:55
<
clever >
blackriversoftwa: thats weird, its showing very clear signs of booting, and already having kvm, before it fails
14:54
<
clever >
tilpner: the first thing the daemon does after picking a user, is to kill every process within the given uid
14:53
<
clever >
tilpner: only when changing his own user, but the nixbld groups are different
14:53
<
clever >
and the other lines nearby, just pastebin the whole output
14:52
<
clever >
can you pastebin the full error msg?
14:51
<
clever >
then it shouldnt need your user to be in the group
14:51
<
clever >
blackriversoftwa: is nix-daemon running?
14:50
<
clever >
blackriversoftwa: and the chmod bits?
14:50
<
clever >
blackriversoftwa: what user/group owns /dev/kvm ?
14:49
<
clever >
blackriversoftwa: your user needs permission to open /dev/kvm
14:40
<
clever >
that changes capslock to shiftlock, so it affects things like numbers
14:40
<
clever >
services.xserver.xkbOptions = "caps:shiftlock";
14:40
<
clever >
betawaffle: no idea, the only keyboard change ive done is shiftlock
14:32
<
clever >
`man vconsole.conf` does exist!
14:32
<
clever >
betawaffle: oh, it even claims to have a man page!
14:32
<
clever >
/nix/store/llvq7j77zkrm8xg93rwr113n3wqynfjm-systemd-239.20190219/example/systemd/system/systemd-localed.service:Documentation=man:systemd-localed.service(8) man:locale.conf(5) man:vconsole.conf(5)
14:31
<
clever >
betawaffle: or the systemd docs
14:31
<
clever >
/nix/store/llvq7j77zkrm8xg93rwr113n3wqynfjm-systemd-239.20190219/share/doc/systemd/NEWS: * If /etc/vconsole.conf is non-existent or empty we will no
14:31
<
clever >
betawaffle: this seems like the most likely hit, so youll want to go to the systemd sources next
14:31
<
clever >
Binary file /nix/store/llvq7j77zkrm8xg93rwr113n3wqynfjm-systemd-239.20190219/lib/systemd/systemd-vconsole-setup matches
14:31
<
clever >
betawaffle: this will search the entire machine for any config or executable that can refer to that file
14:31
<
clever >
[root@amd-nixos:~]$ grep -r vconsole.conf $(nix-store -qR /run/current-system)
13:06
<
clever >
mojjo: not really
13:05
<
clever >
mojjo: you may also not be able to access the data if you change those settings
13:05
<
clever >
mojjo: the bios can be configured to either expose the drives directly, or route them thru the raid controller
12:55
<
clever >
mojjo: line 9, a raid controller, you may need special drivers for it
12:41
<
clever >
mojjo: can you pastebin the output of lspci?
12:39
<
clever >
mojjo: every bios shows it differently
12:37
<
clever >
mojjo: crack the case open, look at it
12:37
<
clever >
mojjo: its more about the physical drive
12:36
<
clever >
mojjo: is it on the sata bus, or directly pcie?
12:34
<
clever >
mojjo: do you know what type of device it is?
12:29
<
clever >
mojjo: does lsblk show it?
12:21
<
clever >
sb0: the attacker needs to know the hash of the build directions to download something
2019-03-26
19:03
<
clever >
eval-time ones are harder to read
19:03
<
clever >
slabity: thats an eval-time failure, rather then a build-time failure
19:02
<
clever >
nefix: does it give a giant spew of "cant build x, 1 dep failed" ?
19:00
<
clever >
nefix: nix already tells you the answer in the error, can you pastebin the entire error output?
17:52
<
clever >
and then it gets messy, you would have to check a lot of things for the propagated build inputs under nix-support
17:52
<
clever >
so you would want to add --all to why-depends
17:51
<
clever >
oh, but a lot of things depend on cups, but dont pull in the .desktop
17:51
<
clever >
christianpoveda: yeah, so gnome-control-center is to blame
17:49
<
clever >
christianpoveda: the output uses colors heavily, so a screenshot may be better
17:48
<
clever >
christianpoveda: nix why-depends /run/current-system /nix/store/b9ypwpa96g1gmxg3n6m04awy4dnss4rw-cups-2.2.6
17:48
<
clever >
christianpoveda: what is the full path in the store?
17:47
<
clever >
christianpoveda: nix-store -qR /run/current-system | grep cups
17:45
<
clever >
christianpoveda: what do you get if you run `nixos-option services.printing.enable` ?
17:42
<
clever >
can you screenshot the thing?
17:41
<
clever >
what if you right click it?
17:40
<
clever >
then i can refer to `ps aux` to find the proc name
17:40
<
clever >
for example, i can see this when i click on my terminal emulator
17:39
<
clever >
_NET_WM_PID(CARDINAL) = 5030
17:39
<
clever >
christianpoveda: run xprop, and click on the app, to find its process name
17:33
<
clever >
which will apply an overlay to gnome3, to create a new gnome3 lacking X, then put that into a nixpkgs overlay, to replace the whole gnome3
17:33
<
clever >
symphorien: you want gnome3 = super.gnome3.gnome3.overrideScope' (_: _: { gnome-documents = null; });
17:32
<
clever >
symphorien: due to how overlays work, that would set gnome3 to a set that contains only gnome-documents = null; and all other gnome things would be missing
16:59
<
clever >
NYXT: and then name your script cc or clang, and put it first in PATH
16:59
<
clever >
NYXT: you could create your own bash script, that runs your clang, and prepends NIX_CFLAGS_COMPILE to the args
16:58
<
clever >
NYXT: 80% of what cc-wrapper does, is force (g)cc to obey $NIX_CFLAGS_COMPILE, because packages often overwrite CFLAGS
16:57
<
clever >
NYXT: `env | grep --color FLAGS`
16:57
<
clever >
NYXT: you need to manually pass it $NIX_CFLAGS_COMPILE and friends
16:46
<
clever >
sphalerite: ah, there is a nix level check, that is somewhat decoupled from the actual config
16:44
<
clever >
sphalerite: an old override i have since disabled
16:44
<
clever >
271 NF_CT_PROTO_DCCP n
16:44
<
clever >
270 extraConfig = ''
16:44
<
clever >
269 XXXlinux_4_9 = pkgs.linux_4_9.override {
16:44
<
clever >
sphalerite:
*digs thru cfg*
16:12
<
clever >
jD91mZM2: correct, you can leave it out
15:21
<
clever >
samrose: what fails if you try to boot the iso on virtualbox?
14:21
<
clever >
JonReed: 27-29 are almost exactly what you want
13:23
<
clever >
slabity: which may also wind up in /run/current-system/sw/etc or sw/share
11:43
<
clever >
sphalerite: i think thats right
11:42
<
clever >
fendor: `cabal install` isnt really supposed to work on nixos, because thats not how nix manages installing things
11:41
<
clever >
sphalerite: i think there is a defconfig entry in the platforms stuff
11:39
<
clever >
fendor: nixpkgs only works fully with the non-new commands
11:39
<
clever >
fendor: all of the new- commands
11:39
<
clever >
so it cant find zlib, even if you do supply it
11:38
<
clever >
fendor: i think the problem is that `cabal new-build` doesnt respect the buildInputs
09:51
<
clever >
qyliss^work: if you add ./vim.nix to your imports section, it will have all of my vim config, including Nix syntax highlighting
09:47
<
clever >
http: is also a function, but http:/ is a url
09:47
<
clever >
a:a colors itself like a function in vim
09:45
<
clever >
have fun! :P
09:43
<
clever >
just like src = "/home/clever/foo.tar.gz";, stop quoting your paths!!
09:42
<
clever >
foo = ../../bar/baz; solved it fully
09:42
<
clever >
Myrl-saki: i recently helped a user that was doing foo = "${../..}/bar/baz"; and then wondered why it was rebuilding uselessly
09:41
<
clever >
> { x, ".", y }: x + y
09:41
<
clever >
qyliss^work: how do you then inherit it, and add it to your function args?
09:41
<
clever >
it will cast to string when you least expect it, with various side-effects
09:41
<
clever >
you can think of paths like numbers in JS :P
09:40
<
clever >
paths arent strings
09:40
<
clever >
> { . = x: y: x+y; }
09:40
<
clever >
qyliss^work: assigning a value to . would be ... difficult
09:39
<
clever >
though you have the limitation that f must be a normal attr
09:39
<
clever >
it would just be a parser level thing, that generates the same AST as `f x y`
09:38
<
clever >
but x `f` y is fairly simple
09:38
<
clever >
infix would be difficult to do with the parser in nix
09:34
<
clever >
Myrl-saki: ive sometimes wanted to use $ in nix
2019-03-25
19:25
<
clever >
neonfuz__: check the go docs
18:54
<
clever >
neonfuz__: i'm thinking, try a newer nixpkgs rev
18:53
<
clever >
neonfuz: which would imply a bug within the derivation, sqlite is enabled, yet silently not included
18:53
<
clever >
neonfuz: yep
18:51
<
clever >
neonfuz: now i can run `nix show-derivation /nix/store/s59vkr1xs9c67nchwwdmpbbappb52wn0-gogs-0.11.86-bin` and see the drv that built it
18:50
<
clever >
neonfuz: i can see that it fetched the exact same gogs build you have, which lacked sqlite
18:50
<
clever >
neonfuz: it is now building...
18:47
<
clever >
and without an fs, i have to switch to .vm
18:47
<
clever >
yeah, deleting the references to them now
18:46
<
clever >
this will actually do what i said
18:46
<
clever >
typo in my cmd
18:45
<
clever >
oh, oops, thats the nixpkgs, not nixos path
18:45
<
clever >
neonfuz: this allows me to build the same cfg, on the same nixpkgs
18:39
<
clever >
neonfuz: can you pastebin your entire configuration.nix file?
18:39
<
clever >
neonfuz: sqlite is also enabled on that version...
18:38
<
clever >
neonfuz: what does `nix eval nixpkgs.lib.version` report?
18:37
<
clever >
neonfuz: sqlite, i mean
18:37
<
clever >
neonfuz: yeah, that one appears to be built without gogs
18:36
<
clever >
neonfuz: what is the full storepath to your gogs binary?
18:33
<
clever >
neonfuz: sqliteSupport defaults to true, and must be set with .override, not config.gogs
14:14
<
clever >
which is rather useless, since it only works as root
14:14
<
clever >
`sudo -u nobody systemctl status ...`
14:14
<
clever >
while yours has all:all, so you can run systemctl status, as any user
14:14
<
clever >
i think that means i can only run zpool as root
14:13
<
clever >
and `sudo zpool status` works without a pw
14:12
<
clever >
cmacrae: from a machine where i do see it working
14:12
<
clever >
(root) NOPASSWD: /run/current-system/sw/bin/zpool status*
14:09
<
clever >
cmacrae: you also missed a / in here
14:09
<
clever >
sudo nix/store/rl4ky8x58ixnfjssliygq7iwyd30l3gn-systemd-239.20190219/bin/systemctl status openvpn-moo
14:09
<
clever >
cmacrae: can you add the -l output to the gist?
14:08
<
clever >
cmacrae: also check `sudo -l` which should list what you can (and cant) do
14:06
<
clever >
cmacrae: and what happens if you run a literal: sudo systemctl '*' openvpn-moo
14:05
<
clever >
clefru: what does `sudo -l` report?
14:02
<
clever >
sphalerite: ive also noticed it, but havent looked into why
13:17
<
clever >
jomik: if condition then (throw "msg") else (buildEnv ...)
13:17
<
clever >
jomik: nope, thats where you want throw instead
13:10
<
clever >
jomik: you want the nix manual, not the nixos manual
13:07
<
clever >
jomik: yeah
13:04
<
clever >
jomik: so you would insert the assert on line 8 maybe
13:03
<
clever >
jomik: the syntax for assertions is weird, it works similar to a let block
13:00
<
clever >
Alling: which says that it has never passed on hydra
13:00
<
clever >
Alling: yep
12:57
<
clever >
clefru: can you run `nix show-derivation /nix/store/foo.drv` on that drv?
12:56
<
clever >
clefru: what the heck did node2nix do to you? lol
12:54
<
clever >
ah, i'm on slim
12:54
<
clever >
/home/clever/.Xauthority
12:54
<
clever >
clefru: XAUTHORITY is in $HOME on nixos, last i checked
12:50
<
clever >
phases="buildPhase" out=/tmp/whatever genericBuild
12:50
<
clever >
you can also set out at the same time
12:48
<
clever >
clefru: what about just running nix-shell, and then `phases="buildPhase" genericBuild` in the directory the src came from?
12:47
<
clever >
clefru: and then re-run nix-build, and nix will run those commands for you, at the right point
12:46
<
clever >
clefru: also, you can just edit the nix file, to add more commands at any point in the derivation
12:42
<
clever >
jomik: you probably want assert or throw
12:38
<
clever >
clefru: it works by adding buildInputs = [ breakpointHook ]; to your derivation
12:37
<
clever >
> breakpointHook
12:33
<
clever >
install is special, and i dont think it can be overrode like that
12:33
<
clever >
only `npm run dist` will run that
12:33
<
clever >
i dont think it does
12:32
<
clever >
clefru: looks like you need to run `babel -d dist src` after npm does most of its work
12:31
<
clever >
clefru: have you tried yarn2nix yet?
12:30
<
clever >
clefru: one mi
12:30
<
clever >
and its likely complaining about /usr/bin/env not existing
12:30
<
clever >
npm ERR! code ENOENT
12:30
<
clever >
npm ERR! path /nix/store/in8240mjl3l857gi1db2ywvwapcqv5hs-node-nanopos-0.1.5/lib/node_modules/nanopos/dist/cli.js
12:29
<
clever >
clefru: /usr/bin/env doesnt exist within the nix sandbox, you cant recreate that issue with su or nix-shell
12:25
<
clever >
clefru: can you pastebin the original error your trying to fix?
12:25
<
clever >
clefru: you may want to `export out=$HOME/foo` first, to make it writable
12:24
<
clever >
clefru: yeah, genericBuild is the usual entry
12:19
<
clever >
clefru: the rest highly depends on what is failing, and why
12:18
<
clever >
clefru: first, you can skip sourcing env-vars, by running `nix-shell /nix/store/foo.drv`, as for $out, you can just re-export out to something else
12:12
<
clever >
fmsbeekmans: you need to have a script that generates the nix, outside of nix
12:10
<
clever >
fmsbeekmans: yeah
12:08
<
clever >
fmsbeekmans: or giving it a wrong hash
12:08
<
clever >
,tofu fmsbeekmans
12:08
<
clever >
fmsbeekmans: by running nix-prefetch-git, in a normal shell
12:05
<
clever >
you will also want to connect the gnd of both, so they have a common reference
12:05
<
clever >
nefix_: so you can then run any serial terminal program on one, to access the serial port of the other
12:05
<
clever >
nefix_: if you connect the tx->rx between both of them, then the serial port on one is linked to the other
12:03
<
clever >
nefix_: yeah
12:03
<
clever >
nefix_: do you happen to have a second matching board?
12:02
<
clever >
nefix_: does the board have a serial port?
11:56
<
clever >
idbloader.img is in that directory
11:56
<
clever >
/nix/store/m7rpgdb47k0qkbw8wq03r47z9gy9w5q5-uboot-rock64-rk3328_defconfig-2017.09/idbloader.img
11:56
<
clever >
nefix: that hydra build is in the cache, you can just: nix-store -r /nix/store/m7rpgdb47k0qkbw8wq03r47z9gy9w5q5-uboot-rock64-rk3328_defconfig-2017.09
11:56
<
clever >
nefix: did you build u-boot for the board?
11:53
<
clever >
fmsbeekmans: but for fetching nixpkgs, you likely want fetchTarball instead
11:53
<
clever >
fmsbeekmans: pkgs.fetchgit is a function for generating such derivations
11:53
<
clever >
fmsbeekmans: all network is disabled during normal derivations, you must use a fixed-output derivation to get network access
11:37
<
clever >
513 type family BoundsOf x
11:37
<
clever >
these 2 things conflict i think
11:36
<
clever >
515 type instance BoundsOf (a->a) = Int
11:36
<
clever >
516 type instance BoundsOf (a->a->a) = (Int,Int)
11:36
<
clever >
fails with the error i gave above
11:34
<
clever >
this is my current patch and nix file
11:34
<
clever >
Conflicting family instance declarations:
11:33
<
clever >
AlexRice[m]: ah, because this is where i need to hide
11:33
<
clever >
54 import GHC.Base
11:32
<
clever >
AlexRice[m]: still fails with the same error with this
11:32
<
clever >
+import Prelude hiding (join)
11:30
<
clever >
testing that...
11:27
<
clever >
but we need a patch file first
11:27
<
clever >
add patches = [ ./foo.patch ]; with overrideCabal
11:26
<
clever >
AlexRice[m]: i see it, join ambiguous
11:24
<
clever >
failing here too, for other reasons
11:24
<
clever >
/nix/store/9hk3204yj7h47d4cfnlfbh0ydl9w8vfc-gnused-4.5/bin/sed: can't read conftest.c: No such file or directory
11:20
<
clever >
so i cant just git clone it
11:20
<
clever >
also, oddly, kmeans-par doesnt have source links in hackage
11:20
<
clever >
binary, in the cabal revision (thanks hackage :P) is the problem
11:20
<
clever >
AlexRice[m]: base can be delt with via jailbreak
11:19
<
clever >
nix-build -E 'with import <nixpkgs>{}; let mypkgs = haskell.packages.ghc844.override { overrides = super: self: { hmatrix = haskell.lib.doJailbreak (self.callHackage "hmatrix" "0.15.0.0" {}); }; }; in mypkgs.kmeans-par'
11:19
<
clever >
0.14.0.0 also has the thing we want...
11:17
<
clever >
we need to go older!
11:17
<
clever >
nix-build -E 'with import <nixpkgs>{}; let mypkgs = haskell.packages.ghc822.override { overrides = super: self: { hmatrix = self.callHackage "hmatrix" "0.13.0.0" {}; }; }; in mypkgs.kmeans-par'
11:17
<
clever >
nope, and binary is also a bootpkg, so you cant override it
11:16
<
clever >
binary <0.6
11:16
<
clever >
but this may have the needed binary?
11:16
<
clever >
nix-build -E 'with import <nixpkgs>{}; let mypkgs = haskell.packages.ghc844.override { overrides = super: self: { hmatrix = haskell.lib.doJailbreak (self.callHackage "hmatrix" "0.13.0.0" {}); }; }; in mypkgs.kmeans-par'
11:15
<
clever >
ghc844 has base 4.11.1
11:15
<
clever >
AlexRice[m]: base is backed into the ghc package, and there is a special override to always ignore base overrides
11:14
<
clever >
844 still fails, checking why...
11:14
<
clever >
AlexRice[m]: nixpkgs doesnt allow base to be overriden, stack might?
11:11
<
clever >
nix-build -E 'with import <nixpkgs>{}; let mypkgs = haskell.packages.ghc844.override { overrides = super: self: { hmatrix = self.callHackage "hmatrix" "0.13.0.0" {}; }; }; in mypkgs.kmeans-par'
11:10
<
clever >
but, i can try another ghc version...
11:10
<
clever >
Alling: i'm on a slightly older nixpkgs
11:09
<
clever >
b: hackage has a cabal revision, adding just that!
11:09
<
clever >
a: jailbreak cant deal with constaints inside an if
11:09
<
clever >
ack, double-whamy
11:08
<
clever >
AlexRice[m]: modified the cabal file, to delete version constraints
11:07
<
clever >
nix-build -E 'with import <nixpkgs>{}; let mypkgs = haskellPackages.override { overrides = super: self: { hmatrix = haskell.lib.doJailbreak (self.callHackage "hmatrix" "0.13.0.0" {}); }; }; in mypkgs.kmeans-par'
11:07
<
clever >
and now ive reproduced the base issue, so i just jailbreak...
11:06
<
clever >
ah, 0.13.0.0
11:05
<
clever >
but, callHackage currently fails, looking into why...
11:04
<
clever >
this will build everything using nix, and override the version of hmatrix as it does so
11:03
<
clever >
nix-build -E 'with import <nixpkgs>{}; let mypkgs = haskellPackages.override { overrides = super: self: { hmatrix = self.callHackage "hmatrix" "0.13.0" {}; }; }; in mypkgs.kmeans-par'
11:00
<
clever >
Alling: you may need to change the version of ghc your using, i believe stack does that via the lts release?