<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/fjQ5W
domogled has joined #nixos
kvda has joined #nixos
iqubic has joined #nixos
<iqubic>
adisbladis: etu: can one of you help me with a home-manager and EXWM issue?
<iqubic>
I'm running into errors. I just rebuilt my system with my new home manager config, and stuff isn't working.
drakonis_ has joined #nixos
<iqubic>
It's supposed to be running exwm, but it isn't
jb55 has quit [Ping timeout: 260 seconds]
page has quit [Ping timeout: 244 seconds]
<iqubic>
Like I logged in and emacs was launched, but emacs gave me some errors when loading the config file.
<disasm>
with a `types.nullOr types.path` how can I do something with that option if it's null? currently I'm getting this error: `"The option `services.byron-proxy.topologyFile' is used but not defined."` I'm trying to do: `lib.optionalAttrs (cfg.topologyFile != null) { inherit (cfg) topologyFile; };`
<iqubic>
Warning (initialization): An error occurred while loading '/home/iqubic/.emacs':
<iqubic>
File is missing: Cannot open load file, no such file or directory, exwm
jgt has quit [Ping timeout: 276 seconds]
viric has quit [Ping timeout: 248 seconds]
<infinisil>
disasm: You probably want to set `default = null`
<infinisil>
in the mkOption
nix-build has joined #nixos
{^_^} is now known as Guest77051
<disasm>
infinisil: ah! trying!
<disasm>
infinisil: you finish your classes?
<growpotkin>
Guys maximum priority issue: Spotify and Pulse Audio will not get along and I am forced to listen to spotify in a browser. AND THE BROWSER EXPERIENCE IS SO SUBPAR! lol
nix-build has quit [Remote host closed the connection]
<growpotkin>
We need 100 of our best nerds on the case
<aanderse>
growpotkin++
<Guest77051>
growpotkin's karma got increased to 5
<iqubic>
I think my issue is more important. I literally do not have a WM on my laptop.
<aanderse>
just because of the entertainment value :)
drakonis has quit [Ping timeout: 252 seconds]
<ashkitten>
i can't tell if sarcastic or serious
Chiliparrot has quit [Ping timeout: 264 seconds]
<infinisil>
Guest77051: Is that you {^_^}??? You changed a lot man
<disasm>
iqubic: switch to sway or i3, problem solved! :) just kidding!
<gchristensen>
infinisil: lol!
ddellacosta has joined #nixos
<ashkitten>
omg
Guest77051 has quit [Remote host closed the connection]
<ashkitten>
hmm
<iqubic>
I have no idea how I could go about fixing this, because I don't know what's causing it.
<iqubic>
I.E. the user with the user-name nix-build?
abathur has joined #nixos
viric has joined #nixos
page has joined #nixos
<gchristensen>
did anyone have channels they wanted {^_^} to be in?
jb55 has joined #nixos
<aanderse>
don't forget #nixos-security!
<gchristensen>
already got it :)
<aanderse>
:D
<aanderse>
gchristensen++
<iqubic>
adisbladis: etu: Could one of you help me out here?
<iqubic>
So, I asked home-manager to run emacs for me, and that didn't really work.
mexisme_ has quit [Ping timeout: 258 seconds]
atlas_ has joined #nixos
halfbit has joined #nixos
psy3497 has joined #nixos
gchristensen is now known as {^_^}
{^_^} is now known as gchristensen
judson__ has quit [Ping timeout: 245 seconds]
judson has quit [Ping timeout: 245 seconds]
jb55 has quit [Remote host closed the connection]
<iqubic>
I set as such: `xsession.windowManager.command = "emacs";` And then I also added this overlay: nixpkgs.overlays = [ (import ./exwm-overlay.nix) ];
<colemickens>
If something recently needs to be built again "opencl", what do I need to do the derivation to fix it?
<colemickens>
(meson + pkg-config, if it matters)
<iqubic>
So I'm not sure what I should do here, because my WM is not working.
ddellacosta has quit [Ping timeout: 258 seconds]
{^_^} has joined #nixos
<iqubic>
I did not know that {^_^} was in nixos-emacs.
<iqubic>
Good to know.
<iqubic>
Well, I've screwed up my system, and I'm not quite sure how to fix it.
mmlb3 has joined #nixos
<gchristensen>
woohoo, {^_^} can support many more channels now
<infinisil>
What was limiting it before that is not anymore?
<gchristensen>
freenode was rate-limiting its JOINs and disconnecting it
<gchristensen>
now, it joins a single channel and waits forFreenode to say okay you joined, then it joins the next
<infinisil>
I see, neato
jlv has joined #nixos
<worldofpeace>
anyone remember/know how to test the graphical iso as a vm? trying to add another iso.
<iqubic>
Anyone know the best way to get advice on my EXWM issues?
<simpson>
iqubic: Work on one piece at a time. It looks like you have a big mass of Emacs configuration entangled with EXWM, and I'm guessing that you're changing multiple things at once and not carefully examining what you're changing.
mexisme_ has joined #nixos
<iqubic>
That's right.
<simpson>
Also, maybe it's time to take a break from the computer. Stretch out, get some water, take a short walk.
Ralith_ has joined #nixos
ottidmes has quit [Ping timeout: 246 seconds]
fusion809 has quit [Remote host closed the connection]
gchristensen has quit [Quit: WeeChat 2.4]
<infinisil>
worldofpeace: I think I've just been writing a little configuration.nix which imports the graphical profile and some iso thing, then building the iso with `nix-build nixos --arg configuration ./configuration.nix -A config.system.build.isoImage`
{^_^} is now known as Guest96999
<infinisil>
Actually maybe `nix-build nixos --arg configuration nixos/modules/installer/cd-dvd/installation-cd-graphical-kde.nix -A config.system.build.isoImage` would work
nix-build has joined #nixos
<clever>
infinisil: what about the attrs of release.nix?
<clever>
one of those should already import that for you
* infinisil
checks
<infinisil>
Hehe yeah iso_graphical
gchristensen has joined #nixos
<infinisil>
Oh but as a vm
Guest96999 has quit [Ping timeout: 248 seconds]
<worldofpeace>
yeah I want to build iso_graphical but use the build-vm script
<clever>
ah
<infinisil>
Hm, I'm not sure it makes sense to want to test iso's in a vm?
<clever>
and build-vm boots it very differently
<clever>
so its not testing the iso layer things
gchristensen is now known as {^_^}
{^_^} is now known as gchristensen
<worldofpeace>
so what's the best way to test the nixos iso?
nix-build has quit [Remote host closed the connection]
<clever>
worldofpeace: qemu -cdrom foo.iso i think
{^_^} has joined #nixos
<worldofpeace>
clever: ahh so just regular. it would be nice if there was an interface in nixos. though that's still pretty easy
<jlv>
I'm trying to figure out how buildInputs works, especially with autoPatchelfHook. I made a package and confirmed that it has the required lib in its result, but autoPatchelfHook cannot find the lib when the package is in buildInputs.
<clever>
jlv: have you read the source for autoPatchelfHook ?
<jlv>
clever: I've glanced at it, but don't find it very easy to understand
<clever>
jlv: first, is the fixup phase enabled? did you set phases= ?
<jlv>
clever: All of the other libraries are found by autoPatchelfHook, just not the one in the package I made
<clever>
jlv: is it named the same as what ldd shows?
<jlv>
clever: ldd?
<clever>
jlv: run ldd on the binary your trying to patch
<jlv>
The llib.so file has the same name as what autoPatchelfHook is looking for, if that is what you mean. It would take a minute for me to grab the binary itself and run ldd on it. Everything is being done is a nix file.
<clever>
jlv: what is the output of ldd, what are the contents of the lib dir on the package you made?
jluttine has quit [Ping timeout: 248 seconds]
<jlv>
The result of the package has "lib/x86_64-linux-gnu/libcurl.so.4". (I am trying to fix a version mismatch on libcurl for a debian package, hence why I'm not just using the NixOS curl package)
<clever>
jlv: the library must be in lib, not lib/x86_64-linux-gnu
<jlv>
clever: That fixed it! Thank you :)
Henson has joined #nixos
<jlv>
Unfortunately, I am now getting "no version information available" on a few of the curl dependencies :/ . I guess I can keep adding debian dependencies until all the errors are gone :/ . I'm still not sure why the NixOS curl package doesn't work on this debian package.
<clever>
jlv: try googling the error?
MmeQuignon has quit [Ping timeout: 272 seconds]
<jlv>
clever: My understanding of the "no version information available" error is that the package was built with a higher version of a library than the installing machine has, but debian has curl version 7.64.0 and NixOS has curl a higher curl version (and I've tried with multiple NixOS curl versions), so I assume there is some other difference between
<jlv>
curl built on NixOS and curl built on Debian.
<clever>
jlv: there are special version symbols in a library, which you can see with readelf
<clever>
and thats likely what its checking for
<jlv>
clever: Huh, maybe I can fake that somehow?
xkapastel has quit [Quit: Connection closed for inactivity]
<clever>
jlv: or edit the requirements
jluttine has joined #nixos
<jlv>
Although I'm not sure if that is enough. The error is really more of a runtime warning, but trying to run the application itself results in numerous errors with curl, so I don't think faking it is enough. It seems like there is some more fundamental difference between the curl versions.
<hpfr[m]>
Is there just a plain latest Linux kernel package? I can only find latest-libre, latest_hardened, and a few others but not a simple latest
<hpfr[m]>
And why doesn’t the unstable channel use the latest kernel
<DigitalKiwi>
yes
<iqubic>
hpfr: pkgs.linuxPackages_latest
<hpfr[m]>
That was my guess, it’s weird it doesn’t show up in the web nixpkgs search
<iqubic>
you can add the line "boot.kernelPackages = pkgs.linuxPackages_latest to your configuration.nix to get the lastest kernel
<DigitalKiwi>
doesn't nixpkgs follow a LTS kernel
<DigitalKiwi>
by default
<hpfr[m]>
Uname-r says 4.19.58 but I just thought unstable would be on the latest kernel
<DigitalKiwi>
nixos-unstable is the future stable
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #66313 → Add GNOME3 ISO, Don't run graphical installer ISOs as root → https://git.io/fjQFs
<iqubic>
Is there a bleeding edge channel?
<DigitalKiwi>
arch linux?
fl0_id has quit [Ping timeout: 244 seconds]
<DigitalKiwi>
but really nixos-unstable is still pretty edgy
<iqubic>
LOL.
<DigitalKiwi>
you can use the latest kernel easily enough
<hpfr[m]>
iqubic: you can tie directly into the nixpkgs repo but you’re just missing out on the automated tests before it goes to unstable
<DigitalKiwi>
there's the small channels too
Okinan has quit [Quit: Konversation terminated!]
drakonis has quit [Quit: WeeChat 2.4]
fl0_id has joined #nixos
szicari has quit [Ping timeout: 244 seconds]
szicari has joined #nixos
<hpfr[m]>
Getting an error “failed assertions to use flatpak you must enable XDG desktop portals with xdg.portal.enable” I thought services.flatpak.extraPortals = [ pkgs.xdg-desktop-portal-gtk ]; was all I needed
<elvishjerricco>
pie_: I'd wager you can just luksOpen it a second time.
<pie_>
elvishjerricco: hm
<pie_>
LUKS keyslot 4 is invalid.
<pie_>
Device /dev/sda3 is not a valid LUKS device.
<pie_>
sudo cryptsetup open /dev/sda3 test
<pie_>
not a happy thing
<elvishjerricco>
What did you do to break the keyslot?
<clever>
elvishjerricco: he expanded the partition
<elvishjerricco>
Why would that invalidate the existing luks header?
<clever>
no idea
vmandela has quit [Remote host closed the connection]
vmandela has joined #nixos
<pie_>
elvishjerricco: thats one of the questions
<elvishjerricco>
pie_: You said you backed up the old partition table?
<elvishjerricco>
What happens if your restore that?
<pie_>
hm
<clever>
bbl
<pie_>
still seems unhappy
rsoeldner has joined #nixos
<pie_>
random "You'd need to know the master key (if the luks container is still open, dmsetup table --showkeys) or a valid backup of the LUKS header."
<pie_>
elvishjerricco: any idea how to test a dmsetup entry, i could try reloading that in a new entry
zupo has joined #nixos
jgt has joined #nixos
rsoeldner has quit [Ping timeout: 252 seconds]
palo1 has joined #nixos
<elvishjerricco>
pie_: I think your issue is that cryptsetup doesn't even want to try to do anything with your luks volume as long as its got that invalid keyslot, so I dunno if you can test it at all.
<elvishjerricco>
So it looks like `cryptsetup luksDump --dump-master-key --master-key-file ./master.key` will dump the key in the appropriate format to `./master.key`
<clever>
elvishjerricco: except all luks commands fail due to the header being corrupt
<elvishjerricco>
Oh right
<pie_>
(presumably)
<pie_>
sudo cryptsetup luksOpen /dev/sda3 test --master-key-file msk.bin
<pie_>
fails with the same problem
<elvishjerricco>
pie_: And wait, if your setup is like mine, you should only have 2 key slots. Why would a 4th come into play?
drakonis has quit [Ping timeout: 264 seconds]
<pie_>
assuming i didnt get the argument order wrong
<clever>
the 2nd answer in stackage shows how to convert the hex key to binary
<clever>
pie_: you may need to kill the keyslot first
<pie_>
elvishjerricco: noooo idea, ive been wondering that roo
<clever>
2019-08-08 03:00:25 < clever> and luksKillSlot may be able to nuke slot 4
<pie_>
clever: can it even _see_ keyslots
<pie_>
maybe its just utterly f***ed
<elvishjerricco>
clever: That might not be a good idea though
<pie_>
backups first
<elvishjerricco>
presumably, the header was untouched.
<pie_>
and the redundant copy experiment idea
<elvishjerricco>
So it might be that something is offset wrong or something, and slot 4 contains something important
<pie_>
hm
<elvishjerricco>
pie_: You said you don't have a header backup from before the partition table change, right?
<pie_>
ok hold on let me paste my offesets
<pie_>
elvishjerricco: yeah :/
<pie_>
(well ...now you know)
<elvishjerricco>
And restoring the original partition table didn't fix anything?
<pie_>
didnt seem to
<pie_>
hm, not much point pasting the offsets since its just the original again
<clever>
pie_: id say find the luks channel on irc
<pie_>
googling that error only yields source code >_>
<pie_>
is there even one
rsoeldner has quit [Ping timeout: 252 seconds]
drakonis_ has quit [Ping timeout: 244 seconds]
<pie_>
working on backups real quick
<pie_>
im working on it in #archlinux right now
<pie_>
clever: elvishjerricco^
<pie_>
00000000 4c 55 4b 53 ba *censored* |LUKS....aes.....|
<pie_>
the header still appears mostly intact so wtf lol
<pie_>
(clever: elvishjerricco yall want to do a postmortem on this later? :P)
rsoeldner has joined #nixos
<elvishjerricco>
pie_: Have you tried supplying a `--type` parameter, maybe?
<clever>
pie_: are the censored bytes 01 61 65 73?
<pie_>
clever: ya
<clever>
pie_: that would be a single byte you censored, plus the encoding of aes :P
<pie_>
ya :P
domogled has quit [Quit: domogled]
<clever>
and the remaining bytes in that line are all 0's
<pie_>
except theres two more bytes
<pie_>
not for me
<pie_>
i got bytes after luks
<pie_>
2
<elvishjerricco>
pie_: Did anything happen between the partition resizing and the first time you got that error? Or did it start happening immediately?
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
<pie_>
clever: everythig looks like its intact
<pie_>
on a cursory skim
<clever>
pie_: the error is complaining about the stripe count on slot 4
<clever>
pie_: slot 1 is at 208, slot 2 is at 256, each slot is 48 bytes
<clever>
so slot 4 is at 352 bytes
<clever>
and the stripe width is the 44th byte in the slot
slyfox has joined #nixos
<clever>
so 396 bytes in, youll find a 4 byte int32
<pie_>
clever: someone else also recommended your trick:P
<pie_>
<grawity> 1) back up the current header (first 4096 bytes) to a file; 2) write the master key, unhex'd, to a file; 3) create a new small disk image and luksFormat it with --master-key-file; 4) make sure it uses the same cipher; 5) copy the first 4096 bytes from the image to your partition
<pie_>
clever: can you hex that or can i set hexdump to show int column...
* pie_
checks man page
cfricke has joined #nixos
<clever>
pie_: master key must be binary, read the 2nd answer in the stackexchange link you pasted
<clever>
xxd i think it was
drakonis has quit [Ping timeout: 252 seconds]
<pie_>
clever: they said that for backup reasons
<pie_>
i presume
<clever>
2nd answer
<clever>
`you would use masterKey.bin in your cryptsetup command as the master key file`
boxscape has quit [Remote host closed the connection]
<pie_>
btw wxhexeditor is good for this kind of stuff IIRC
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
linarcx has joined #nixos
<pie_>
uhhhh
<pie_>
if key slot 8 starts at 544
<pie_>
why does 512 say EFI PART
m5n has quit [Quit: -a- IRC for Android 2.1.54]
<clever>
pie_: are you starting 544 from the partition or disk?
<pie_>
partition
jgt has joined #nixos
<pie_>
but i have an idea of at least one thing that happened
<pie_>
clever: are you followingin #archlinux
<clever>
pie_: what is the starting offset of the partition?
<iqubic>
Well, this exwm-overlay that I took from adisbladis isn't working. I added this overlay to my nixpkgs-overlays, via home-manager and told home-manager to run "emacs" as the windowManager command. I have not installed Emacs in any other way than this overlay. Emacs is not found when I log in via LightDM. https://github.com/adisbladis/exwm-overlay/bl
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
Makaveli7 has joined #nixos
logzet has joined #nixos
asymptotically has joined #nixos
joshuagl has joined #nixos
ardumont has quit [Ping timeout: 276 seconds]
jlv has quit [Ping timeout: 260 seconds]
o1lo01ol1o has joined #nixos
<clever>
,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.
sogatori has joined #nixos
shandogs has joined #nixos
<manveru>
iqubic: how did you use the overlay?
<manveru>
iqubic: because it looks like you'll need to add `(emacsWithPackages (epkgs: [ epkgs.exwm ]))` to your systemPackages if you want it globally
<{^_^}>
[nixpkgs] @vbgl pushed commit from @mgttlinger to master « pgsolver: init at 4.1 »: https://git.io/fjQNb
<logzet>
Ahh, thanks johanot, I somehow thought that sudo nix-env would use the system profile
<logzet>
Taneb: No and I see ^^
<clever>
`sudo nix-env` operates on roots ~/.nix-profile, not the system profile
mexisme_ has quit [Ping timeout: 264 seconds]
o1lo01ol1o has joined #nixos
vmandela_ has joined #nixos
camal has quit [Quit: Lost terminal]
<pie_>
elvishjerricco: clever: what would i do without you guys to help me with my filesystem insanities
<pie_>
my god..
<pie_>
its full of TARs!
<logzet>
clever: I may not be smart enough to use NixOS after all … ^^
<pie_>
^totally unrelated joke that just crossed my mind
<elvishjerricco>
pie_: I'm sorry for getting you into these file system insanities :P
vmandela__ has joined #nixos
vmandela has quit [Remote host closed the connection]
<pie_>
this is all my fault
gxt has joined #nixos
<kraem>
gchristensen: i've been searching for what i've done wrong trying to connect to a L2TP vpn for a while. Just found your L2TP github issue. Created /etc/l2tp.secrets as you suggested and it finally works! Thanks!
o1lo01ol1o has quit [Ping timeout: 245 seconds]
<kraem>
correction: /etc/ipsec.secrets
<pie_>
elvishjerricco: ok lets attempt what i originally wanted xD
<pie_>
how do i get these to match
<pie_>
└─sda3 8:3 0 937.8G 0 part
<pie_>
└─root 254:0 0 216.8G 0 crypt
linarcx has quit [Quit: WeeChat 2.5]
<clever>
pie_: the cryptsetup resize that originally alerted you to the fubar
<pie_>
does the encryption have any on disk thing that keeps track of its size or is it just a constant set in memory based on the outer partition info
<pie_>
i wonder if i can update the dm table
<pie_>
the second parameter is size in sectors or somesuch
<elvishjerricco>
I *thought* luks had a size record in the header, but I dunno
<pie_>
though i suppose that what this is supposed to do
<{^_^}>
[nixpkgs] @vbgl pushed commit from @jdanekrh to master « gnome-sharp: mono4 -> mono5 »: https://git.io/fjQxG
linarcx has joined #nixos
<elvishjerricco>
pie_: I just tried setting up a LUKS volume on a 4G file
<elvishjerricco>
Then expanding the file to 8G
<elvishjerricco>
Closing and reopening the LUKS volume (without ever running resize) caused it to show as an 8G device
<ivan>
is there some kind of regression in cryptsetup-askpass in master where it only accepts the password 10% of the time now
<elvishjerricco>
So the size is just the underlying device's size, I guess
<pie_>
yeah i read something like that shoudl work
<elvishjerricco>
pie_: So it sounds like resize just resizes without closing and reopening
<pie_>
elvishjerricco: well im not sure thats a valid conclusion, in theory it *could* be updating something on the device, but yeah i dont thing that would be the case
<pie_>
elvishjerricco: sure but why doesnt it work :P
<elvishjerricco>
I just grew my file to 12G and used resize, and it worked fine
<pie_>
oh you know what
<pie_>
i wonder if zfs being mounted has something to do with it
<pie_>
then again im not closing/opening it...
<pie_>
>:(
<pie_>
[incoherent mumbling]
<clever>
pie_: the root device under zfs should still grow, according to lsblk
<elvishjerricco>
Yea I wouldn't expect ZFS to have anything to do with it
<pie_>
yeah i was grasping at straws
<pie_>
the dmsetup table isnt changing so
<pie_>
hm
mexisme_ has joined #nixos
rauno has quit [Ping timeout: 264 seconds]
<iqubic>
So, I have a few files being imported in my home.nix that I want to debug. Specifically I want to print out the value of a few of these variables during the execution of "home-manager switch" is that possible?
<pie_>
well, im doing dmsetup reload, im assuming thas whats needed
sogatori has quit [Remote host closed the connection]
vmandela has joined #nixos
<pie_>
clever: will yanking the cryptsetup from under zfs hang my entire os
<clever>
pie_: very
<pie_>
i was hoping i could remount doing that but i guess not
<hyper_ch2>
clever: ubuntu goes zfs root in 19.10 :)
<pie_>
like how sometimes a vpn has better timeout behaviour than normal network, and bridge over it :P
<hyper_ch2>
pie_: <3 native zfs encryption :)
<pie_>
hyper_ch2: kind of :( native zfs encryption because it leaks some metadata, but i guess its not the kind of thing that really matters for any practical purposes
<pie_>
for most people anyway
<averell>
it is pretty neat, but the performance is not quite ext4-level IME
<adisbladis>
ext4 - do the wrong thing really really fast
<averell>
at least i don't have to wait for errors then :)
<adisbladis>
^_^
<hyper_ch2>
pie_: what metadata? dataset names? dataset properties?
<clever>
hyper_ch2: i believe all dataset names, snapshot names, and properties
<adisbladis>
hyper_ch2: Iirc it's things like file count, file sizes etc?
<adisbladis>
Or am I wrong?
<clever>
hyper_ch2: because you need to be able to delete a dataset you lack the keys for
<{^_^}>
[nixpkgs] @vcunat merged pull request #66310 → ghc: use ld.gold only on Linux → https://git.io/fjQd9
<hyper_ch2>
it does not leak that to my knowledge
<pie_>
clever: also some structure about data can be deduced but i dont quite remember
<clever>
and how big the dataset is
<pie_>
zfs will not encrypt metadata related to the pool structure, including dataset names, dataset hierarchy, file size, file holes, and dedup tables.
<pie_>
^ quote
<adisbladis>
pie_++
<{^_^}>
pie_'s karma got increased to 4
<adisbladis>
Thanks
gxt has quit [Ping timeout: 260 seconds]
<pie_>
nom nom karma
<hyper_ch2>
pie_: thx... but I fail to see any sensitive information in that - and dedup is only for really experienced users
<adisbladis>
hyper_ch2: Depends on your use case. If I was a chinese dissident I would not use zfs crypto
<pie_>
well i was more worked up about it in the beginning but now that i have more of a clue, i guess most people dont even use dedup
<hyper_ch2>
maybe one shouldn't name a dataset like "mySecretPr0nCollection"
<pie_>
file size and holes might be more interesting though
<elvishjerricco>
hyper_ch2: Leaking file sizes and number of files can be major in some cases. Not normally though
<elvishjerricco>
The dedup table is a massive leak though; but luckily no one uses it
<adisbladis>
pie_: I wonder if we could deduct what packages you have installed by knowing rough file sizes, package layouts etc
<iqubic>
So, how does one debug a nix file?
<elvishjerricco>
I personally don't like that properties in particular are leaked.
<pie_>
adisbladis: project++ lol
<iqubic>
\I have a few files being imported in my home.nix that I want to debug. Specifically I want to print out the value of a few of these variables during the execution of "home-manager switch" is that possible?
<elvishjerricco>
adisbladis: I think it's very likely you could figure some of the packages on a NixOS system through native ZFS encryption, via only file sizes and dnode number if that's exposed. File sizes alone could leak a lot of that information though
<adisbladis>
elvishjerricco: I think so too. If anyone wants a subject for their thesis ^
<pie_>
its booothering me that i cant live resize this
<elvishjerricco>
I need to rewatch that one talk on how zfs encryption is actually implemented. They say file metadata is encrypted, but I forget to what extent
<pie_>
my liffle filesystem, zfs is magic (therefore should be able to do whatever insanity i throw at it, tihs isnt even zfs)
<iqubic>
adisbladis: I can't quite get my EXWM stuff running properly.
<hyper_ch2>
"ZFS will not encrypt metadata related to the pool structure, including dataset and snapshot names, dataset hierarchy, properties, file size, file holes, and deduplication tables" --> but I wonder, if file size even a problem? as it is cow it will be fragmented quickly.
sigmundv has quit [Ping timeout: 246 seconds]
<elvishjerricco>
hyper_ch: It doesn't leak record sizes, it leaks the size of an entire file. You can map out a table of blocks belonging to files in an encrypted dataset without the key.
<elvishjerricco>
As I understand it anyway
<elvishjerricco>
It's pretty easy to tell when a file has the exact same size as GHC 8.2.2 that it is in fact GHC 8.2.2
Guest14863 has joined #nixos
<elvishjerricco>
Also looks like the ZIL can be tampered with without the key
<iqubic>
adisbladis: I've sent a few messages in #nixos-emacs detailing my issues
<iqubic>
But I supposed you've already seen those
orivej has joined #nixos
<hyper_ch2>
I wonder how it would leak that.
<kgz>
what's the latest on rfc 39 (unprivileged maintainer team)?
<eyJhb>
kandinski: yeah you can
<alexarice[m]>
is ZFS only really for servers or are people using it on their personal machines as well, keep hearing everyone going on about it but don't know if its worth looking into
<elvishjerricco>
hyper_ch: I think it basically doesn't encrypt the block pointer hierarchy, so you can follow the pointers from the uberblock all the way down to the user data blocks. Knowing the layout of a ZFS pool, you can identify which blocks are dnodes, which lets you identify which blocks are the parent of a file
<clever>
iqubic: builtins.trace "foo" "bar" will print foo, then return bar
<iqubic>
Nice
<iqubic>
So it works just like trace in haskell
<clever>
yep
<pie_>
elvishjerricco: well...i guess ill reboot, if you dont see me, i f***ed up
<elvishjerricco>
pie_: Good luck.
<alexarice[m]>
has anyone managed to build something using the rust wayland crates and know how you pass C wayland libraries to them?
<hyper_ch2>
elvishjerricco: also interesting to know: what filesize is actually displayed when you have compression enabled?
<elvishjerricco>
hyper_ch2: I suppose the compressed size. Still, if you're just interested in checking for files that you know are vulnerable, you can run such files through all the compression algorithms they're likely to be using to see what you're looking for.
<eyJhb>
*and pie_ was never seen again
ng0 has joined #nixos
<teto>
what's the name of the syntax "assert localServer -> db4 != null;" ? is there any doc on it ?
<hyper_ch2>
elvishjerricco: :)
pie_ has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @domenkozar closed pull request #22357 → [WIP] Use openssl 1.1 by default → https://git.io/vDOGw
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
gagbo has quit [Ping timeout: 245 seconds]
gagbo has joined #nixos
paulandell has joined #nixos
<paulandell>
dear all, please help, I'm configuring a nixos server and creating my configuration.nix, now, I have some scripts I would like to use but they are not fit for a upstream package, how to handle this? how to create a simpke pkgs dir where I will put my default.nix files and can have them "besides" my normal nixpkgs and including them in the config?
<paulandell>
thank you
<Yaniel>
easiest is probably to make them overlays
<elvishjerricco>
hyper_ch2: https://youtu.be/frnLiXclAMo "We only encrypt level zero data." Yea, so that confirms the entire block pointer hierarchy is unencrypted; level zero means leaf node. So like directory listings and vfs metadata are encrypted; but you have a perfect picture of where everything is.
<elvishjerricco>
paulandell: Overlays are good, or you can just do `environment.systemPackages = [(pkgs.callPackage ./mypkg {})];` if you don't want to have to figure out what overlays are :P
<paulandell>
Yaniel: soo the overlays will build my packages and installing them when I run nixos-rebuild switch?
<paulandell>
elvishjerricco: integresting and in root of my /etc/nixos there will be my mypkg/default.nix?
<elvishjerricco>
yea
<{^_^}>
[nixpkgs] @layus opened pull request #66317 → openssl: fix man pages collisions → https://git.io/fjQpB
<sphalerite>
paulandell: I don't think you need an overlay for it, you could for instance put (pkgs.writeScriptBin "do-some-stuff" "#!${stdenv.runtimeShell}\necho Hello World!") in your systemPackages directly
<sphalerite>
paulandell: depending on the size of the script you may want to put it in a separate file
<{^_^}>
[nix] @edolstra pushed to master « Rename file for consistency »: https://git.io/fjQpR
<{^_^}>
[nixpkgs] @zimbatm pushed commit from @bachp to master « gitlab-runner: don't bundle prebuilt docker images (#66225) »: https://git.io/fjQpr
vmandela_ has joined #nixos
gagbo has quit [Ping timeout: 245 seconds]
iqubic has quit [Remote host closed the connection]
schjetne has quit [Ping timeout: 244 seconds]
iqubic has joined #nixos
rauno has joined #nixos
gagbo has joined #nixos
rsoeldner has quit [Remote host closed the connection]
rsoeldner has joined #nixos
vmandela__ has quit [Ping timeout: 244 seconds]
anrgistan has joined #nixos
alter-schjetne has quit [Ping timeout: 248 seconds]
<emilsp>
ah, lol, gradle will not use the build tools from $ANDROID_HOME, but rather it'll use the build tools it downloads on it's own. And those most definitely do not work.
<iqubic>
etu++
<{^_^}>
etu's karma got increased to 7
<iqubic>
adisbladis++
<{^_^}>
adisbladis's karma got increased to 14
<etu>
yay
<adisbladis>
Sweet, internet points
<JonReed>
Nix keeps occasionally but regularly keeps failing because it runs out of space, despite that I have 30 GB of free space for root left. I suspect that it's because tmpfs (showing 3.2 gb) or devtmpfs (1.6gb) running out of space. Are they supposed to have this little space and constantly run out of space or I did something wrong to configure my syst
<JonReed>
em?
pie_ has joined #nixos
JonReed has quit [Remote host closed the connection]
JonReed has joined #nixos
<sphalerite>
johanot: or you're running out of inodes, what does df -i say?
<sphalerite>
er JonReed ^
hyper_ch2 has quit [Remote host closed the connection]
<johanot>
I have plenty of inodes, thanks for asking :)
<sphalerite>
JonReed: ah, you're on a hipster filesystem that doesn't use inodes I guess? :p
<niksnut>
it's pretty common that /run has a limit like 20% of available RAM
nst^ has joined #nixos
zupo has joined #nixos
<JonReed>
niksnut: I have 16GB ram on this system. Only 7% is used currently. If I run `sudo nix-channel --update` it will crash my chrome because the system will run out of space.
pie_ has quit [Ping timeout: 252 seconds]
iqubic has quit [Remote host closed the connection]
<JonReed>
So, if I understand correctly, this is likely the filesystem cutting me with its bleeding edge and not Nix?
orivej has quit [Ping timeout: 248 seconds]
<sphalerite>
JonReed: no, it's more likely the tmpfs running out of space like you suspected. Running out of inodes is a common problem on ext4, but by using btrfs you've avoided that :)
<sphalerite>
JonReed: see if it still happens when you unmount /tmp
<emilsp>
JonReed: do take care though, this does sound like something that btrfs would do.
rsoeldner has quit [Ping timeout: 252 seconds]
<{^_^}>
[nixpkgs] @lightbulbjim opened pull request #66320 → golden-cheetah: use qt5's mkDerivation → https://git.io/fjQpN
iqubic has joined #nixos
rsoeldner has joined #nixos
<iqubic>
I think the version of the emacs package auctex that is in nixpkgs is outdated.
<qyliss>
iqubic: update it! :)
chr15 has joined #nixos
<JonReed>
sphalerite: The `/tmp` exists on a device which still has 29G left. I can't unmount it. This is why I don't understand where or how Nix runs out of space. https://pastebin.com/62E7a9wb
<iqubic>
qyliss: I don't know how.
<qyliss>
iqubic: want to learn? :)
<iqubic>
I guess.
<iqubic>
auctex is used in only 3 places ever in all of nixpkgs.
<qyliss>
(It's much healthier for Nixpkgs if people are empowered to do upgrades and stuff for themselves, rather than relying on a small group of experts)
<qyliss>
If it's not used much, should be an easy upgrade :)
<iqubic>
And one of those times is in a freaking manual xml file.
<iqubic>
So, I'm not sure how I'd go about updating the emacs package auctex.
<sphalerite>
JonReed: whoops, misunderstood.
<iqubic>
Because it's not a normal package, but a package for emacs.
<sphalerite>
JonReed: aah, since you're not on nixos and this looks like a single-user installation — this might be a case of TMPDIR being set to /run/user/$UID and nix trying to use that for builds, which goes horribly wrong because it's generally too small.
<sphalerite>
JonReed: see if running it with TMPDIR=/tmp helps
JonReed has quit [Remote host closed the connection]
<iqubic>
qyliss: Actually, I don't need to update this right now. From the auctex website, I get this bit of info: "Current AUCTeX version: 12.1 (2017-12-10)"
<iqubic>
But I think there are some minor changes to the way emacs and nix interact coming down the pipeline soon. So I'll wait for those to be pulled in first.
Makaveli7 has quit [Quit: WeeChat 2.5]
<JonReed>
sphalerite: I am on NixOS though (i.e. I use `configuration.nix`). Setting TMPDIR didn't solve it. Is there a way to verify somehow which temporary directories Nix uses? I do know that it uses `/tmp` for nix-build because that's where builds end up in if I use `nix-build -K`. Like, when I use nix-channel where does nix unpack those channels in, how
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjQhY
MmeQuignon has joined #nixos
rsoeldner has quit [Read error: Connection reset by peer]
rsoeldner has joined #nixos
<vika_nezrimaya>
any way to generate a kernel config and put it so my kernel won't take hours to compile? probably nix-shell into the linux package, unpackPhase and do the make menuconfig stuff, right?
alex`` has quit [Quit: WeeChat 2.5]
gagbo has quit [Ping timeout: 244 seconds]
rsoeldner has quit [Ping timeout: 252 seconds]
gagbo has joined #nixos
<sphalerite>
JonReed: ooooh sorry, missed that
<JonReed>
If Nix is saying "/nix/store/xnk229qan97prxmwfsgm65py2qpca109-gnutar-1.32/bin/tar: nixos-19.03.173251.56d94c8c69f/pkgs/build-support: Cannot mkdir: No space left on device". How to find that path of "nixos-19.03.173251.56d94c8c69f/pkgs/build-support"?
<ivan>
JonReed: probably a nix-build in your /tmp, maybe --keep-failed if you need to look at it
alex`` has joined #nixos
<JonReed>
ivan: But what if the error is during `sudo nix-channel --update`?
<sphalerite>
JonReed: then it's weird… I'm starting to suspect btrfs weirdness as emilsp said
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjQhX
cfricke has quit [Ping timeout: 258 seconds]
kleisli has quit [Ping timeout: 245 seconds]
nakkle has joined #nixos
angristan has joined #nixos
Guest73 has joined #nixos
anrgistan has quit [Ping timeout: 252 seconds]
<nakkle>
Hey! Yesterday I set up a machine with NixOS, and today I can't do nixos-rebuild anymore. If I do, I get 6 errors of the form "error: file 'nixpkgs' was not found in the Nix search path (add it using $NIX_PATH or -I)". The root user does have the 19.03 channel as 'nixos', and it is updated. Anyone know how to fix this?
<clever>
nakkle: what does `echo $NIX_PATH` say?
arianvp[m] has joined #nixos
<nakkle>
clever: /root/.nix-defexpr/channels
DjDingo has joined #nixos
<clever>
nakkle: thats not what it should be, did you change nix.nixPath or set NIX_PATH in a .bashrc ?
<nakkle>
I've copied a configuration for hydra from a working server, which included this line here: ` extraEnv = { "NIX_PATH" = "ssh-config-file=${sshConfigFile}"; };`, but that only affected the hydra user as far as I can see. Besides that, it's a completely fresh system
<clever>
you can manually export it as a one-time fix, then nixos-rebuild to experiment with a long-term fix
<nakkle>
okay that fixed it for now. root's $NIX_PATH is still what just exported it as. Not sure what happened here
<nakkle>
thanks!
<DjDingo>
Hello #nixos, I am trying to install NixOS on ZFS, as I am reading the wiki I noticed it mentioning odering issue while mounting. Does it mean that the partitions should be ordered?
<DjDingo>
exarkun: https://nixos.wiki/wiki/NixOS_on_ZFS The second entry in "Known issues" it talked bout the ordering issues. In the single-disk installation part I noticed the partition being numbered. Therefore the question.
<gchristensen>
like which ones mount first
<gchristensen>
using legacy, NixOS handles the mounting order and does it correctly
<DjDingo>
So the partions can be numbered diffrently to their actual position on the drive?
<exarkun>
if you're installing NixOS _on_ ZFS then you probably just want to pay attention to that section of the page
byteflam1 has joined #nixos
<gchristensen>
DjDingo: no. the problem is *mount order* not order on disk.
<exarkun>
I haven't installed NixOS _on_ ZFS but I see that the instructions for that direct you to use mountpoint=none for the root filesystem, not mountpoint=legacy
<{^_^}>
[nixpkgs] @marsam merged pull request #66308 → wasm-pack: fix build on darwin → https://git.io/fjQdV
<exarkun>
DjDingo: The issue that refers to is where you have a filesystem you wanted mounted at "/foo" and another you want mounted at "/foo/bar"
<gchristensen>
exarkun: that isn't correct, actually. rpool/root (mountpoint=none) is never mounted. rpool/root/nixos (mountpoint=legacy) is mounted at /
<exarkun>
DjDingo: Chances are things won't go well if the system tries to mount "/foo/bar" before it mounts "/foo", that's all.
<exarkun>
gchristensen: sorry, the "root" filesystem, then?
<exarkun>
Anyway my main point is read the instructions and follow them :)
<gchristensen>
rpool/root never gets mounted, it is like a container for many possible root filesystems
<exarkun>
gchristensen: Maybe "root" was a bad name for it.
<gchristensen>
yeah, "roots" or something
<DjDingo>
ahhh I see. so the numbering of the partition doesn't matter. What does is the order of mounting the partitions?
<gchristensen>
right, DjDingo
<gchristensen>
exarkun: the whole page could stand to be blown up and rewritten from scratch.
<DjDingo>
very nice. Thank you for the help exarkun and gchristensen
<exarkun>
DjDingo: Just as a point of terminology: filesystems get mounted, filesystems can be *on* partitions. But with ZFS, they filesystems are in a "pool" and the "pool" may be on one or more partitions.
<exarkun>
may save you some confusion at some point
<DjDingo>
exarkun: Lesson learned. I am so new to all these terms haha. I am a statistician interested in Functional Programming. Systems are a pretty new territory for me lol.
<gchristensen>
DjDingo: if I may make a suggestion, ext4 is a perfectly fine filesystem :)
<exarkun>
seems like you're on a good path, at least
<exarkun>
ext4 has the advantage of being a lot simpler to deal with, yea
<exarkun>
so does xfs though, and maybe xfs has fewer of ext4's downsides :)
<exarkun>
but zfs definitely has more features ...
<exarkun>
features are great right?
<gchristensen>
nobody ever has to look anything up when it comes to ext4 questions
<DigitalKiwi>
<3 zfs
<DigitalKiwi>
gchristensen: how do you do encryption on ext4
<exarkun>
(I use xfs on / and zfs on everything else, fwiw)
<exarkun>
DigitalKiwi: dmcrypt
<exarkun>
is it still called that
<gchristensen>
DigitalKiwi: luks
<DjDingo>
gchristensen: yeah so I've heard. How NixOS handles the config rollback makes it so safe for simpler usage. I came from FreeBSD tho so think it'll be comfortable to build the system on ZFS
<DjDingo>
gchristensen: such as the video install guide on youtube. I used to follow them strictly and tended to not make any change I wasn't familiar with e.g. filesystems haha
<gchristensen>
yeah, what video install guides? never heard of these
<DjDingo>
oh not for nixos, but other distros. arch, debian, void etc
<hyper_ch2>
DjDingo: investing some time into learning how to use zfs is well worth it. Snapshotting, cloning, backup etc. is such a breeze
<hyper_ch2>
(and not just for NixOS but in general)
<hyper_ch2>
the main thing is: remember to use legacy mounts on nixos
npmccallum has joined #nixos
<DjDingo>
hyper_ch2: yea it is very handy. The 2 reaons I quitted FreeBSD for NixOS are 1, no official VMWARE client in the repo. 2, I'm delving into Haskell and NixOS seems perfect for a better understanding of FP
fendor has joined #nixos
<gchristensen>
NixOS taught me my first functional programming language
<DjDingo>
gchristensen: haha how long have you been using NixOS
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fj7eu
hmemcpy has joined #nixos
orivej has joined #nixos
<hmemcpy>
Hi all. I'm very new to nix, so please bear with me :) I'm trying to use a specific version of a python package, but when I run nix-shell, I still get the old version that's installed on my machine
<{^_^}>
nix#3025 (by zeecde, 4 days ago, open): HTTP error 200 (curl error: Timeout was reached) - unable to update and install
<duairc>
There's one particular large nar that's causing issues for me. I thought I'd be clever and try to nix-prefetch-url it, but that hasn't helped
<MmeQuignon>
Hi, I use moc to listen music. This player can be compiled to play streams but the nix version is not compiled that way. I would like to customise the package and eventually make a pr about this. Is there some documentation to do that ? Thank you !
<ivan>
duairc: what's your nix version?
<duairc>
ivan: 2.2.2
Guest14863 has quit [Ping timeout: 246 seconds]
<ivan>
oh damn "Hydra does not seem to support ranged HTTP requests, which is why the retry functionality is not working for me. It keeps trying to download the entire file, and hits the same timeout over and over."
ericsagnes has quit [Ping timeout: 276 seconds]
<{^_^}>
[nixpkgs] @domenkozar opened pull request #66328 → Extract NixOS options documentation generation to a function → https://git.io/fj7ep
<duairc>
I mean given that I can download the file with wget, is there some I can trick nix into thinking it's downloaded it so that it doesn't try to download it every time?
<duairc>
Maybe after a while it had already downloaded most things, so it wasn't trying to download so many things in parallel, which meant it had more bandwidth for the big file that was failing, and it was able to get it over the line before the timeout?
<duairc>
But this definitely doesn't feel very declarative :(
chris__ has joined #nixos
<chris__>
Anyone know if we have ApacheBench (ab) in nixpkgs?
<Taneb>
chris__: it's in apacheHttpd
<colemickens>
gchristensen: btw, I was hoping to be able to test the wireguard stuff soon, and ack back with something other than "I'll try soon" but I'm using it as motivation to wrap up my azure/nixos work, and of course I'm running into roadblocks again. Hopefully soon TM.
<colemickens>
gchristensen: thank you very much, pre-emptively, too :)
hyper_ch2 has quit [Remote host closed the connection]
gagbo has joined #nixos
acarrico has joined #nixos
<Soo_Slow>
sphalerite, eeeh. Hopefully it will be solved someday :c Nix looks way more interesting than flatpak
<sphalerite>
yeah, it's a painful issue
ixxie has joined #nixos
_kwstas has quit [Quit: _kwstas]
_kwstas has joined #nixos
byteflam1 has quit [Ping timeout: 245 seconds]
Okinan has joined #nixos
_kwstas has quit [Client Quit]
byteflam1 has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
schjetne has quit [Ping timeout: 258 seconds]
gagbo has quit [Ping timeout: 245 seconds]
<Soo_Slow>
is it the same for vulkan applications?
gagbo has joined #nixos
o1lo01ol1o has joined #nixos
byteflam1 has quit [Ping timeout: 268 seconds]
cransom has quit [Quit: WeeChat 2.4]
byteflam1 has joined #nixos
cransom has joined #nixos
byteflam1 has quit [Ping timeout: 245 seconds]
mexisme_ has joined #nixos
Zer000 has joined #nixos
<{^_^}>
[nixpkgs] @veprbl closed pull request #37622 → weechat: Add plugin to config only if it is used → https://git.io/vxlvK
<{^_^}>
[nixpkgs] @veprbl merged pull request #64262 → weechat: prevent impure use of system Python → https://git.io/fj6Yl
<{^_^}>
[nixpkgs] @veprbl pushed commit from @bdesham to master « weechat: prevent impure use of system Python (#64262) »: https://git.io/fj7fK
mexisme_ has quit [Ping timeout: 250 seconds]
byteflam1 has joined #nixos
nixos-usr has joined #nixos
schjetne has joined #nixos
<{^_^}>
[nixpkgs] @vbgl opened pull request #66333 → ocamlPackages.zarith: enable for OCaml 4.08 → https://git.io/fj7fH
nixos-usr has quit [Ping timeout: 260 seconds]
boxscape has joined #nixos
orivej has quit [Ping timeout: 246 seconds]
<chiefgoat>
So I installed Ubuntu to another drive
<chiefgoat>
And it overwrote the EFI stub that kicks off the rest of my NixOS boot
<chiefgoat>
Now I can see a bunch of kernel bzimages and initrds in the nixos folder of my efi partition but none of them will boot. I get a kernel panic every time
<chiefgoat>
Is there someone who can provide a little guidance?
lsix has quit [Quit: Konversation terminated!]
gagbo has quit [Ping timeout: 272 seconds]
gagbo has joined #nixos
<chiefgoat>
Found one that works under systemd/
sigmundv has quit [Ping timeout: 258 seconds]
johnny101m has quit [Remote host closed the connection]
<enteee>
any python people here? I would need a bit of support setting up a python project with a dependency to the cryptography module.
lambda-11235 has joined #nixos
<simpson>
enteee: Sure. I have to catch a bus in a few minutes, but there are others, too. `pythonPackages.cryptography` is probably your starting point.
kfound has joined #nixos
<enteee>
simpson: same bus situation here as well :) but thanks..
<enteee>
simpson: let me share quickly what i have right now.. 1 sec
<enteee>
this is the default.nix that I am using to create my python dev environment, which does include pythonPackages.cryptography
<{^_^}>
[nixpkgs] @bdesham opened pull request #66334 → pinboard-notes-backup: init at 1.0.4.1 → https://git.io/fj7Jq
Abednego has joined #nixos
<enteee>
the problem is now that the pip install -r on line 37 does not find the cryptography module and tires to install it again, which fails for obvious reasons (gcc missing)
shibboleth has quit [Quit: shibboleth]
<simpson>
enteee: I'm not sure about that shellHook. It *does* seem very Pythonic, but not very Nixy. Sadly there's no automatic transformation from requirements.txt to Nix expressions.
<enteee>
simpson: ok.
<simpson>
Best of luck; I gotta grab a bus.
<enteee>
k, thx anyways
veske has quit [Ping timeout: 245 seconds]
<adisbladis>
enteee: Tbh I wouldn't mix nix/python like that. I'd make my nix-shell only provide python/pip.
<adisbladis>
(and the required native deps to install whatever python deps I have)
enuiarte has joined #nixos
<enteee>
adisbladis: hmm.. I am not quite sure if I do understand you. That's kind of what i tried. where do you see the mistake?
gagbo has quit [Ping timeout: 245 seconds]
<enteee>
other that me calling pip install -r ... at the end.
<adisbladis>
enteee: I mean I wouldn't even bother pulling in cryptography with nix
o1lo01ol1o has quit [Remote host closed the connection]
<adisbladis>
If you're going to use pip, use pip all the way
<enteee>
i kind of have to. because pip install cyrptography does not work becuase it requires for example gcc to be present etc..
gagbo has joined #nixos
bdesham has joined #nixos
<enteee>
adisbladis: the main reason for this beeing that cryprography is not a pure python package.
<adisbladis>
Imho `In some cases virtualenv fails to install a library because it requires patching on NixOS (example 1, example 2, general issue). In this cases it is better to replace those libraries with ones from Nix. ` is really bad advice
<enteee>
adisbladis: good to know that this is bad advice. It's what i tried to follow
<enteee>
funny enough my approach worked quite well up until a few days ago where something must have changed in virtualenvwrapper or similar.. that's when things started to go wrong
drakonis has quit [Quit: WeeChat 2.4]
byteflam1 has quit [Quit: Lost terminal]
<enteee>
adisbladis: anyways here comes my bus. thx for your help. I'll try that and do some more digging later.
<emptyflask>
especially the `shells = [...]` bit. shouldn't that use `${pkgs.zsh}/bin/zsh` instead?
<averell>
if you have programs.zsh.enable that link will exist, but yeah you could do that too.
gagbo has quit [Ping timeout: 248 seconds]
gagbo has joined #nixos
johanot has quit [Quit: WeeChat 2.4]
gxt has quit [Ping timeout: 260 seconds]
<averell>
seems rather outdated, rsync and direct ssh with echo works fine for me with zsh. maybe you can reproduce that (like in a small container) and then narrow it down to the relevant config.
WhittlesJr has joined #nixos
linarcx has quit [Read error: Connection reset by peer]
jgt has quit [Ping timeout: 252 seconds]
linarcx has joined #nixos
<averell>
#65283 needs more reviewers, only while supplies last!
<emptyflask>
I wonder if it's just that I need to explicitly install rsync in configuration.nix, I'll have to test when I'm back to my other computer
<emptyflask>
oh. there's a services.rsync.enable
<emptyflask>
that would probably be it
<emptyflask>
*rsyncd
<joepie91>
emptyflask: rsync can operate in multiple modes
<joepie91>
either talking to an rsyncd, or tunneling over another protocol (eg. SSH) to a remote ad-hoc rsync instance
<joepie91>
rsyncd.enable should only be necessary when you intend to talk to rsyncd directly
<joepie91>
if you just want to do eg. rsync-over-SSH, the remote system only needs to have rsync installed and available in the PATH of whatever shell session SSHing to the box produces
<joepie91>
and rsyncd will serve no purpose
<joepie91>
(it will essentially SSH in, start `rsync` in receiving mode, and then blast data over the SSH connection)
<emptyflask>
ah, I see. so yeah I'll have to figure out the rsync-in-the-path issue then
drakonis has joined #nixos
<joepie91>
emptyflask: right. if the target system is NixOS, adding it to environment.systemPackages should be sufficient.
Abednego has quit [Ping timeout: 245 seconds]
drakonis1 has joined #nixos
<joepie91>
emptyflask: I believe it's also possible to give rsync a custom command to execute for starting the remote rsync instance, in which case you could shoehorn in a nix-shell call if you don't want to pollute your system environment with an rsync binary
<joepie91>
but I haven't tested this so I don't know how well it actually works
<chiefgoat>
So it looks like firefox-wayland is a lie?
<chiefgoat>
Running it from Sway without xwayland just fails
<gchristensen>
chiefgoat: perhaps you could reframe that question in to a problem description and a question about how to resolve it?
<gchristensen>
because I assure you, none of the people who work hard to maintain firefox and wayland in nixpkgs wanted to lie to you
<chiefgoat>
gchristensen: it wasn’t intended as insult just a hyperbolic declaration ala the cake is a lie
drakonis has quit [Ping timeout: 250 seconds]
<gchristensen>
please tune down the hyperbole, it hurts to work hard on something and then have people be hyperbolic about it
<chiefgoat>
And there isn’t much in the way of additional detail to furnish since X11 apps will simply halt without further detail when run from s Wayland session
<gchristensen>
well, your second message was way more informative :)
<gchristensen>
what happens if you start firefox with
<gchristensen>
MOZ_ENABLE_WAYLAND=1
<{^_^}>
[nixpkgs] @takeda opened pull request #66336 → pgbouncer: use c-ares for name resolution → https://git.io/fj7JH
<chiefgoat>
gchristensen: arguments to dbus_message_new_method_call() were incorrect
<gchristensen>
not sure
<gchristensen>
I use regular firefox on wayland with MOZ_ENABLE_WAYLAND
<emptyflask>
thanks joepie91 -- I'll try the global install tonight. that sounds like it will work
SovereignBleak has joined #nixos
cfricke has joined #nixos
nexgen has quit [Remote host closed the connection]
<SovereignBleak>
gchristensen: also chiefgoat. are you sure it's running as pure wayland? I've gone months now thinking I was running Firefox as Wayland until I ripped xwayland out and verified for myself. Here's a longer paste of the error: http://dpaste.com/2D72TKY
nexgen has joined #nixos
<gchristensen>
not sure, but if I kill xwayland it works fine
<SovereignBleak>
Hmmm. Qutebrowser too fails to run for me but I'm not sure if that's because the Qt patch bringing in wayland support still hasn't made it down to us.
lordcirth has quit [Remote host closed the connection]
lordcirth has joined #nixos
<SovereignBleak>
Eh, well I need Emacs so I'll just put xwayland back but it is a mystery.
duckonomy has joined #nixos
growpotkin has joined #nixos
selfsymmetric-mu has quit [Remote host closed the connection]
_kwstas has joined #nixos
pietranera has quit [Quit: Leaving.]
zupo has joined #nixos
<kfound>
is there a way to prevent "nix-env -i" doing any local compilation; so only using binary substitutes? "allow-import-from-derivation" and "fallback" don't seem to do what i want
<damesca>
Hi. Looking for help. I'm trying to install a newer version of aws-vault by overriding src/version in the nixpkgs 19.03 version, but getting a ton of build failures (see build config/errors here: https://pastebin.com/Gkh7SVyx ). Can anyone help? This is my first attempt at changing a package, and after reading the manual/googling around I've still
<damesca>
not found a good example that works for my slow brain. Would appreciate if anyone can lay it on me. :)
tewlz has quit [Quit: WeeChat 2.5]
<simpson>
damesca: It would seem that the newer version has different dependencies from the older version. You will have to regenerate the dependencies with go2nix or similar.
<damesca>
Hmmmm.... Ok. I'll look into that. Thanks.
<chiefgoat>
corpix: this was actually fixed in the latest unstable builds but thank you :-) I meant that it doesn’t run under “pure” Wayland
emptyflask has quit [Quit: WeeChat 2.5]
zupo_ has joined #nixos
magneticduck has joined #nixos
wildtrees has joined #nixos
wildtrees has quit [Remote host closed the connection]
zupo has quit [Ping timeout: 245 seconds]
vika_nezrimaya has joined #nixos
wildtrees has joined #nixos
wildtrees has quit [Remote host closed the connection]
jgt has quit [Ping timeout: 250 seconds]
wildtrees has joined #nixos
imj[m] has joined #nixos
chloekek has joined #nixos
dispanser has joined #nixos
gxt has joined #nixos
wildtrees has quit [Remote host closed the connection]
Neo-- has quit [Ping timeout: 272 seconds]
jlv has joined #nixos
<jlv>
Can I override glibc in a package buildInput?
Thra11 has joined #nixos
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
vmandela has joined #nixos
orivej has joined #nixos
_kwstas has quit [Quit: _kwstas]
<etu>
jlv: overrideAttrs is your friend for that
<etu>
or no, override should be enough
<exarkun>
What is the difference between override and overrideAttrs?
<etu>
(package.override { glibc = myGlibc; })
<etu>
exarkun: override overrides the arguments provided to the function, while overrideAttrs overrides the attributes set in a derivation, like buildInputs
<exarkun>
Ah. That seems to make sense. Thanks. What's the difference between overrideAttrs and `... // { more stuff }`?
<jlv>
What if I want to override glibc while making the package derivation itself? It looks like overrideAttrs is for overriding while you are calling a package. To clarify, it looks like glibc is automatically added to every package, and I need a different glibc for one package.
<etu>
> a = { b = 3; }; a.b
<{^_^}>
error: syntax error, unexpected ';', expecting '.' or '=', at (string):26:21
<etu>
> a = { b = 3; }; a.b;
<{^_^}>
error: syntax error, unexpected ';', expecting '.' or '=', at (string):26:21
<etu>
> let a = { b = 3; } in a.b
<{^_^}>
error: syntax error, unexpected IN, expecting ';', at (string):256:20
<etu>
> let a = { b = 3; }; in a.b
<{^_^}>
3
<etu>
> let a = { b = 3; } // { c = 4; }; in a.c
<{^_^}>
4
<etu>
exarkun: as you see above, I make set a that contains b = 3, but in the last line I merge it with another set that contains c = 4
<exarkun>
etu: Yea, I think I understand what // does. It sounds like what it does is a lot like what overrideAttrs does. I guess a difference, though, is that overrideAttrs takes a function and passed the old value as an argument, so the overrides can be based on the original value if you want.
<etu>
exarkun: yeah, on overrideAttrs you can kinda mix and match old and new data
kfound has quit [Ping timeout: 245 seconds]
<jlv>
I get "called with unexpected argument 'glibc'" when I try to override glibc with overrideAttrs. I assume because the package function does not take glibc as an argument, since it is implicitly added to every package.
damesca has quit [Remote host closed the connection]
ris has joined #nixos
dazage has joined #nixos
orivej has quit [Ping timeout: 244 seconds]
<exarkun>
jlv: pastebin exactly what you tried
<jlv>
Whoops, looks like I called .override instead of .overrideAttrs
<jlv>
Noticed as I was copying :)
orivej has joined #nixos
jgt has joined #nixos
<jlv>
Unfortunately, it looks like it doesn't actually override glibc :/
<jlv>
The package still has access to the original glibc :(
gagbo has quit [Ping timeout: 268 seconds]
<jlv>
This is the line with the override `deblibldap.overrideAttrs (oldAttrs: rec { glibc = (callPackage ./deblibldap/deblibc {});})`
gagbo has joined #nixos
johanot has joined #nixos
<{^_^}>
[nixpkgs] @worldofpeace closed pull request #66313 → Add GNOME3 ISO, Don't run graphical installer ISOs as root → https://git.io/fjQFs
_kwstas has joined #nixos
orivej has quit [Ping timeout: 272 seconds]
mthst has joined #nixos
_kwstas has quit [Quit: _kwstas]
Jackneill has quit [Remote host closed the connection]
<jlv>
symphorien: Is it possible to use a different glibc then? Like make a different stdenv or something? I'm trying to add a debian package and am running into a lot of issues with incompatible libraries, which is now at the point where the NixOS glibc appears incompatible.
<linarcx>
aanderse: No, flutter + qt development. Most of my times i'm working with these two things. all of other parts of nixos are awesome.
<symphorien>
You mean a stdenv with musl
<symphorien>
?
<{^_^}>
[nixpkgs] @sikmir opened pull request #66337 → qlandkartegt: init at 1.8.1 → https://git.io/fj7T0
<jlv>
I would really like to switch to NixOS as my main computer, but really need this one unfree debian package :(
<aanderse>
linarcx: it has been too many years since i've done qt development... and i've never done it on nixos, so i'm not sure what to say about that
<jlv>
symphorien: I'm not sure what that is. Maybe?
<aanderse>
but yeah, flutter was painful
<aanderse>
hopefully one day i'll get back at it and try to make it work
<linarcx>
aanderse: Yeah.. qt creator has some bad issues on nixos. Also i reported them on github. But.. I hope in future these thinkgs fix.
<jlv>
The specific error is "/nix/store/iykxb0bmfjmi7s53kfg6pjbfpd8jmza6-glibc-2.27/lib/libc.so.6: version `GLIBC_2.28' not found (required by /nix/store/jcjfkl4n1w6bffha01np9aqm0jn16wyv-deblibldap-2.4.2/lib/libldap_r-2.4.so.2)"
<symphorien>
^ do this with an old nixos release instead of unstable
<symphorien>
> stdenv.libc.version
<{^_^}>
attribute 'libc' missing, at (string):256:1
<symphorien>
> glibc.version
<{^_^}>
"2.27"
<symphorien>
So you need a more recent libc
Guest28197 has quit [Changing host]
atlas_ has joined #nixos
<jlv>
symphorien: That sounds a little harder on my end
<symphorien>
Is your unfree debian package released also for rhel ?
<symphorien>
This would lift the "recent libc" requirement
<karetsu>
suppose I want to release my default sink when I use lightdm's `dm-tool switch-to-greeter` so that when my other users log on that source is also picked up as the default for them - I have pacmd but not what to do to allow my user to release the source so that the other user (who had it as their last source) can have it automatically allocated for them
<jlv>
I've also tried Docker, but there appears to be some issues with binding the drive directory. Probably because it is a virtual filesystem.
SovereignBleak has quit [Quit: Connection closed for inactivity]
xantoz has quit [Quit: WeeChat 2.5]
xantoz has joined #nixos
atlas_ is now known as hacker
linarcx has quit [Quit: WeeChat 2.5]
hacker is now known as atlas_
<bdesham>
just tell me if this is already covered in the manual, but... if a Python program has "foo[bar]" in its requirements.txt, how would I package that?
<bdesham>
the "foo" here is coming from pypi but doesn't exist in nixpkgs yet, afaict
<bdesham>
(to be clear, it's the [bar] option I'm not sure how to handle)
<ottidmes>
Umm, I am trying to install NixOS on a new NVMe SSD and am getting Buffer I/O error on dev nvme0n1 when e.g. trying to write the partition table after making one with gdisk, does this mean I am screwed and need to get it replaced? Seems like a hardware issue :(
alex`` has quit [Ping timeout: 246 seconds]
<simpson>
bdesham: Usually by reading the source of `foo` and figuring out what the `[bar]` feature installs.
<joepie91>
ottidmes: have a look at your smart data
zeta_0 has joined #nixos
<bdesham>
simpson: ok, thanks
alex`` has joined #nixos
<ottidmes>
joepie91: smartctl right?
<jlv>
It looks like `fetchurl` has issues with the ~ character
<ottidmes>
joepie: smartctl -H /dev/nvme0, and smartctl -H /dev/nvme0n1 both give me result: PASSED
<ottidmes>
havent even written 25MB to it, but it does have ErrCount 56, just not sure how bad it is
<jlv>
How am I supposed to fetchurl a url with a ~? It throws an error that ~ is an invalid character, but if I try to url encode it with %7E, it throws an error that % is an invalid character...
<srhb>
jlv: "it" ?
<bdesham>
jlv: are you quoting the URL?
<jlv>
bdesham: yes
<jlv>
srhb: fetchurl
<jlv>
Same error with nix-prefetch-url
<srhb>
jlv: Hm, yes. builtins.fetchurl does not appear to have that restriction.
<srhb>
Oh
<srhb>
It's because of the store name of course
<srhb>
You just need to rename that.
<schmittlauch[m]>
Does the nixpkgs manual also have a top-level name in the pkgs set and thus can be installed/ built with nix build?
<srhb>
jlv: So just feed name to pkgs.fetchurl to rename the output.
Abednego has joined #nixos
<schmittlauch[m]>
The manual itself only describes the build process in a nix-shell in a directory
<jlv>
Maybe the problem is with the downloaded file. Looks like the file has a ~ in it name.
<srhb>
jlv: Yep, that's what I said above :)
<jlv>
Got ya :)
<ottidmes>
joepie91 : I am going to check my sedutil settings, it might just cause those error to occur due to them, fingers crossed
<joepie91>
ottidmes: may also just be an unseated cable :)
<srhb>
schmittlauch[m]: I'm not sure it is, but it's manual in <nixpkgs/pkgs/top-level/release.nix>
Lokawn has quit [Remote host closed the connection]
<jlv>
Whoo! Looks like I got seadrive-daemon working on NixOS! Thanks for everyone's help :)
<jlv>
I'll be testing seadrive-gui next for the full experience, but that is open source at least, and I already got it to build correctly, just need to test
<jlv>
seadrive-daemon is a bit of a mess with several libraries pulled in from Debian, but it works!
<kellyjac06>
I was going to ask, is home-manager is the most common nixos way to deal with dotfiles, configs, and installing programs that would maybe be more user focused rather than system(?)
<ashkitten>
oof, i'm having an issue where i built a binary with gcc but it's not in the nix store, and the interpreter doesn't exist anymore because i've upgraded glibc since and deleted old generations
<ashkitten>
how do i deal with that?
<ashkitten>
if i rebuild the binary or patch it, the interpreter will get out of date again
<zeta_0>
kellyjac06: i don't know about the dotfiles, but the home-manager is user specific, so the local(user) and global(system) packages are seperate
<srhb>
zeta_0: Where's the problem in that expression exactly?
<mdash>
ashkitten: the preferred solution is to build it with nix instead so it stays up to date
<mdash>
ashkitten: what program is it?
<ashkitten>
oh hang on
<ashkitten>
i'm gonna package the program
<ashkitten>
that'll take care of it
<mdash>
that's usually easiest in the long run
<mdash>
(though often really frustrating in the short term... I have been trying to package pleroma, it's not been fun)
<ashkitten>
good luck
<kellyjac06>
zeta_0 I think I saw ages ago someones dotfiles that were something like this where most of their config files had been rewritten as .nix: https://gist.github.com/LnL7/f1266f840871d471256008c70c463fbc IDK if they were using home-manager though, I can't seem to find the repo anymore
<vaibhavsagar>
zeta_0: this is day 3 of you pestering people
<zeta_0>
srhb: well i don't think there is a problem, i am just trying to figure out the correct way to paste it into home.nix, i am not sure where i made the syntax errors
zupo has joined #nixos
<zeta_0>
kellyjac06: i've never worked with dotfiles before, so i don't know
<zeta_0>
kellyjac06: but i am really enjoying home.nix, it has saved me a lot of frustration
<srhb>
zeta_0: It looks like you took some quotes that were meant to just highlight code and kept them in your setup.
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fj7II
<blackriversoftwa>
can someone help me with dnsmasq on nixos a bit?
<blackriversoftwa>
I want to set up dnsmasq to provide dhcp on a secondary network interface
andrewrk has joined #nixos
<blackriversoftwa>
and have the dhcp clients addressable by hostname from my nixos machine (the one running dnsmasq)
<blackriversoftwa>
the idea is to be able to plug in my raspberry pi to a second ethernet port and be able to ssh to it by hostname
<andrewrk>
I'm getting `qt.qpa.plugin: Could not find the Qt platform plugin "xcb" in ""` from obs-studio, anyone know how to fix it? this is 19.09.git.b5f5c97 (Loris)
<blackriversoftwa>
but I don't know how to set up dns on my NixOS machine so it uses dnsmasq as a secondary nameserver
<blackriversoftwa>
I tried and lost the ability to resolve internetn names
<blackriversoftwa>
(I tried putting the address of the interface that dnsmasq was listening on in `networking.networkmanager.appendNameservers`
<vaibhavsagar>
andrewrk: are you using `nix-env` for OBS?
<andrewrk>
vaibhavsagar, I was just gonna use `nix-shell -p` but I'm happy to make a env file for it
<vaibhavsagar>
I've had a bunch of issues with `nix-env` and Qt, so much so that I avoid `nix-env` and use `environment.systemPackages` wherever possible
<vaibhavsagar>
`nix-shell -p` should work fine
<vaibhavsagar>
basically Qt shits itself whenever there is a version mismatch
<samueldr>
with the new wrapQtAppsHook, this should slowly go away
<adisbladis>
For a Qt app to run the stars need to align and you need to make a religious sacrifice :/
<samueldr>
derivations not using it still will need to be fixed as they are found
<samueldr>
in fact, it should be breaking even more decisively quickly, so the leftovers should be picked up quickly I hope
<blackriversoftwa>
srhb ok but I'm already using systemd-resolved for my main local network
<blackriversoftwa>
so that's what's listening on 127.0.0.1
<srhb>
blackriversoftwa: So you want the other interface that dnsmasq is listening on in your resolv.conf?
prakritti has left #nixos [#nixos]
<blackriversoftwa>
I want my resolv.conf to point both to dnsmasq (just for the Raspberry pis on a little subnet) and systemd-resolved (for my main network/internet)
<{^_^}>
[nixpkgs] @jonringer opened pull request #66343 → ocamlPackages_latest.campl4: init at 4.08+1 → https://git.io/fj7IB
<blackriversoftwa>
the only reason I'm using dnsmasq at all is to serve dhcp to the raspberry pis and to let me resolve their hostnames
<blackriversoftwa>
srhb: cc
__red__ has quit [Remote host closed the connection]
<srhb>
blackriversoftwa: You can manually point networking.nameservers at the address dnsmasq is listening at then.
<Thra11>
If I'm writing a package where the buildPhase outputs a bunch of files, should I output the files directly into $out and skip the installPhase, or is there a reason to build the files first and have an install phase which copies them to $out?
<srhb>
blackriversoftwa: Or you could add it to resolved. Whichever makes more sense.
<blackriversoftwa>
srhb: but then I wouldn't get my dhcp configured nameservers for internet, would I?
<srhb>
blackriversoftwa: Depends how they're being configured.
<andrewrk>
thanks samueldr after updating my unstable channel it's fixed
<samueldr>
if you see others, and can't fix them + send PR, reply on that tracking issue with the package that needs fixing
<andrewrk>
I was a bit reluctant to do that because I have this crazy patch against linux-firmware: https://clbin.com/L9ZUI
<andrewrk>
ok thanks I will
<samueldr>
it'll eventually be sorted out
Guest73 has quit [Remote host closed the connection]
spoutnik3567[m] has joined #nixos
<{^_^}>
[nixpkgs] @domenkozar reopened pull request #22357 → [WIP] Use openssl 1.1 by default → https://git.io/vDOGw
shibboleth has joined #nixos
polman has quit [Ping timeout: 245 seconds]
polman has joined #nixos
<ashkitten>
how can i unset a value from a set?
orbekk has quit [Quit: WeeChat 2.4]
orbekk has joined #nixos
magneticduck has quit [Quit: Leaving]
chloekek has quit [Ping timeout: 248 seconds]
<{^_^}>
[nixpkgs] @jlesquembre opened pull request #66344 → clj-kondo: init at 2019.07.31-alpha → https://git.io/fj7I1
<blackriversoftwa>
srhb: I figured it out. Turning on dnsmasq defaults resolved to being disabled, so I had to add `services.resolved.enable = true;` to my module for this to work
<blackriversoftwa>
I also had to add `bind-interfaces` to my dnsmasq options so it didn't try to bind port 53 on the wildcard address
domogled has joined #nixos
mexisme_ has joined #nixos
npmccallum has quit [Quit: npmccallum]
<ivegotasthma>
hello
<ivegotasthma>
is there a place for the most recent discussions of flakes?
zeta_0 has quit [Quit: doing a rebuild, i'll be back soon]
<sugi[m]>
i'm kind of wondering what's going on with the postgresql tests on grahamc ofborg in this PR: https://github.com/NixOS/nixpkgs/pull/66347. the tests finished in half a minute? did the tests really run or did someone else make a similar PR and thus no tests needed to be run?
<{^_^}>
[nixpkgs] @pierothebear opened pull request #66348 → [python3][dask] update dask to 2.2.0 → https://git.io/fj7LW
ixxie has quit [Ping timeout: 244 seconds]
info_supplicant[ has joined #nixos
mexisme_ has quit [Ping timeout: 268 seconds]
ddellacosta has joined #nixos
<ivegotasthma>
I'm packaging a python library and I'm wondering, what should go into the propagated build inputs?
zeta_0 has joined #nixos
<ivegotasthma>
should all my buildInputs be in the propagatedBuildInputs too?
Thra11 has quit [Ping timeout: 245 seconds]
enuiarte has joined #nixos
<infinisil>
ivegotasthma: All python deps should go in there
karetsu has quit [Quit: WeeChat 2.4]
<infinisil>
If you have dependencies only needed during the build itself, those should go in nativeBuildInputs
<vaibhavsagar>
ivegotasthma: if it builds just fine but fails on runtime complaining about a missing dependency, that dependency should go in propagatedBuildInputs
mexisme_ has joined #nixos
bakakuna has quit [Ping timeout: 272 seconds]
<ivegotasthma>
thanks guys
<kellyjac06>
iouhAOUHIW1
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<kellyjac06>
iwfKKJWW221
kellyjac06 has left #nixos [#nixos]
<ivegotasthma>
is this a riddle?
<infinisil>
Haha I'm wondering too, it sure feels like one
<averell>
someones gonna be changing passwords...
wayne_ has joined #nixos
<{^_^}>
[nixpkgs] @NilsIrl opened pull request #66349 → Remove useless dont_clobber_environment.patch for Godot → https://git.io/fj7L1
<ivegotasthma>
averell: lol
WhittlesJr has quit [Remote host closed the connection]
andrewrk has left #nixos ["Leaving"]
ng0 has quit [Quit: Alexa, when is the end of world?]
<AWizzArd>
If NixOS wants to compile software it means that no binary version is available, right? But who adds binary versions of software? If I compile it, is it getting uploaded somewhere and stored for all future NixOS users?
<infinisil>
With Nix it would be rather secure because of its purity, people could still upload bad build results though. So we should only allow it for deterministic builds, such that a build result only gets approved if e.g. 3 different people upload the exact same result