awygle has quit [Quit: No Ping reply in 180 seconds.]
manjaroCinnamon has quit [Remote host closed the connection]
adamantium has quit [Ping timeout: 248 seconds]
awygle has joined #nixos
_e has quit [Ping timeout: 252 seconds]
mtjmullen has quit [Ping timeout: 245 seconds]
rj1 has quit [Ping timeout: 250 seconds]
glowpelt has quit [Ping timeout: 276 seconds]
_e has joined #nixos
mtjmullen has joined #nixos
Okinan has quit [Quit: Konversation terminated!]
atlas_ has joined #nixos
glowpelt has joined #nixos
Okinan has joined #nixos
rj1 has joined #nixos
abf_ has joined #nixos
rj1 has joined #nixos
rj1 has quit [Changing host]
hyper_ch has quit [Ping timeout: 272 seconds]
mexisme has joined #nixos
bsima has joined #nixos
jluttine has quit [Ping timeout: 246 seconds]
jluttine has joined #nixos
monokrome has quit [Ping timeout: 245 seconds]
rapsyd has joined #nixos
___laika has quit [Ping timeout: 244 seconds]
Rusty1 has joined #nixos
rapsyd-halsym has joined #nixos
monokrome has joined #nixos
rapsyd has quit [Ping timeout: 245 seconds]
<rapsyd-halsym>
hi all, I'm trying to get a baremetal nixos boot working, using yubikey for auth.. seeing a segfault after yubikey is found: https://pastebin.com/8bB4gUTv
DariusTheMede has quit [Ping timeout: 246 seconds]
Cale has quit [Ping timeout: 276 seconds]
liberiga has joined #nixos
zeta_0 has quit [Ping timeout: 245 seconds]
andrewzah has quit [Quit: WeeChat 2.5]
zeta_0 has joined #nixos
day|flip has quit [Quit: ERC (IRC client for Emacs 26.1)]
lord| has quit [Quit: WeeChat 2.5]
lord| has joined #nixos
Cale has joined #nixos
<midi[m]>
Tfw package will not update at all but all others will
ToxicFrog has quit [Ping timeout: 245 seconds]
ToxicFrog has joined #nixos
mexisme has quit [Ping timeout: 246 seconds]
porrifolius has joined #nixos
Zer0xp has quit [Ping timeout: 268 seconds]
<porrifolius>
error: Please be informed that this pseudo-package is not the only part of
<porrifolius>
Nixpkgs that fails to evaluate. You should not evaluate entire Nixpkgs
<porrifolius>
by Hydra.
<porrifolius>
without some special measures to handle failing packages, like those taken
<porrifolius>
How do I try to resolve that error?
<porrifolius>
I'm trying to enter a nix-shell for haskell-opencv3 library.
Zer0xp has joined #nixos
<porrifolius>
I cloned the git repo and checked out the latest release, but I get the error after cd'ing into the directory and running nix-shell as instructed by the readme.
cyraxjoe has quit [Read error: Connection reset by peer]
cyraxjoe has joined #nixos
<gchristensen>
while I realize this might add some competition to the spot market which is bad for me, I open sourced the code I use for transient Nix builders on Packet's Spot market: https://github.com/grahamc/packet-nix-builder
<porrifolius>
There was a bad hash error for the nix-pkg version the default.nix was loading... I replaced the hash and got rid of that error. Then I was stuck on the 'please be informed' error.
<porrifolius>
Running nix-shell -v --show-traces does'n give me any more (useful) info as far as I can tell.
<porrifolius>
Ohhh... hang on... maybe I've solved it. It's a multi-repo git repo and there is nix stuff floating around in the root directory so nix-shell does something but now that I'm in the correct sub-dir things are looking a lot more promising!!
vmandela has quit [Remote host closed the connection]
fusion809 has quit [Quit: Leaving]
fusion809 has joined #nixos
Garbanzo has quit [Remote host closed the connection]
wfranzini has quit [Remote host closed the connection]
endformationage has quit [Quit: WeeChat 2.5]
palo1 has joined #nixos
liberiga has quit [Ping timeout: 260 seconds]
palo has quit [Ping timeout: 245 seconds]
palo1 is now known as palo
knupfer has joined #nixos
sb0 has joined #nixos
knupfer has quit [Ping timeout: 250 seconds]
<{^_^}>
[nixpkgs] @peterhoeg merged pull request #65078 → nixos/syncthing: do not use nogroup → https://git.io/fjMJp
<fasd>
how often does nixpkgs stable update emacs package repositories?
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.2)]
<fasd>
and is there a way to search them with nix search
gxt_ has quit [Quit: WeeChat 2.5]
<jackdk>
I ran `nix-shell -p 'nodePackages."@vue/cli"'` and it downloaded eleventy quadrillinon things, so I assume it worked. It dumped me in a shell which tab-completed the `vue` command, anyway
<ldlework>
thanks i just figured that out too
vmandela has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
Drakonis has quit [Quit: WeeChat 2.4]
pmiddend has joined #nixos
hyper_ch2 has joined #nixos
<hyper_ch2>
the last 3 days, my home server rebooted at around 2:30 - 4:30 am and I can't figure out why. journalctl doesn't show anything except it suddenly stops. Any idea how to troubleshoot?
<jackdk>
stay up late with a good book and strong coffee?
bgamari has quit [Ping timeout: 252 seconds]
<hyper_ch2>
jackdk: I gotta get up at 4:30 :(
zupo has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
pmiddend has quit [Quit: WeeChat 2.5]
hyper_ch2 has joined #nixos
sellout- has quit [Ping timeout: 244 seconds]
hyper_ch2 has quit [Remote host closed the connection]
<tilpner>
hyper_ch2: Oh, if it happens every day at 4:30, don't do a memory test. atop would still be good though
smpl[m] has joined #nixos
<hyper_ch2>
tilpner: no, it happens sometime between 2:30 and 4:30 am... not at the same time though
<hyper_ch2>
tilpner: btw, in a nix expression, how do I access the value from the configuration.nix?
tonic1 has joined #nixos
<hyper_ch2>
tilpner: also how would atop help? when I go to sleep, server runs... when I get up, server rebooted (waiting for zfs password)
<tilpner>
hyper_ch2: You *can* use: (import <nixpkgs/nixos> {}).config.*
orivej has quit [Ping timeout: 248 seconds]
<tilpner>
But there might be a better way. Which option do you want to access?
<hyper_ch2>
tilpner: improving my zed PR
<tilpner>
Uhh
<tilpner>
Don't do that for nixpkgs
<hyper_ch2>
by default using system-sendmail if ZED service is enabled
<tilpner>
It's expensive and will not be accepted
<tilpner>
Make it user-configurable, sure, but don't eval a nixos config that might not even exist
<hyper_ch2>
so, I wonder how to write: if ( enableZED = true ); then propagatedBuildInputs = [ "system-sendmail" ]; fi
zeta_0 has quit [Remote host closed the connection]
MatejC has quit [Ping timeout: 264 seconds]
bakakuna1 has joined #nixos
<hyper_ch2>
sendmail is only required if you use zed (and let zed even send emails).
<tilpner>
Take an argument at the top, then do propagatedBuildInputs = lib.optional enableZed system-sendmail;
<tilpner>
Then the user of that package can .override if they disagree with your choice
<hyper_ch2>
argument at the top?
<tilpner>
In the set your expression takes: { stdenv, ..., enableZed ? false }:
<tilpner>
Or true, your decision
wfranzini has joined #nixos
<hyper_ch2>
tilpner: but it's defined in the /etc/nixos/configuration.nix
<hyper_ch2>
so don't I need to access that value somehow?
bakakuna has quit [Ping timeout: 258 seconds]
bakakuna1 is now known as bakakuna
<tilpner>
You should not access that value from a package expression
<tilpner>
If you think it needs to be automatic, write/change a module to pass it down to where your expression is being used
<hyper_ch2>
but how?
<tilpner>
How what?
<hyper_ch2>
the only thing I found in modules files was checking settings in the module file and build system service files
<hyper_ch2>
but if you don't enable ZED, then you don't need sendmail for ZFS
<hyper_ch2>
so sendmail should only be propagated as build input if the user enables ZED in the configuration.nix
<hyper_ch2>
so I have adjusted the module file to enable/disable ZED.... but how do I now reflect that in the package expression to provide propagated build input or not?
Jackneill has joined #nixos
<tilpner>
hyper_ch2: I just skimmed your PR. Can't you just do systemd.services.zfs-sed.path = lib.optional cfgZED.enable pkgs.system-sendmail;?
<hyper_ch2>
to have sendmail available when zed is enabled
<tilpner>
There are other means of providing sendmail when zed is enabled, and not all of them require altering the zfs package expression
<tilpner>
Is sendmail only going to be used in that systemd service?
nwspk has quit [Quit: nwspk]
<hyper_ch2>
maybe
<hyper_ch2>
it's because of the feedback
<hyper_ch2>
I was happy to just provide in the zed.rc the name of the mailer to be used
<hyper_ch2>
but feedback from infinisil said differently
<tilpner>
You don't even need the optional, the entire service is already optional
<matt`>
hey all. i'm creating a custom package for vivado (huge proprietary binary) and i keep getting that the nix daemon is running out of memory. the tar is 21GB, but I have 32GB of system memory. any idea why it might be failing and how i can get around it? i haven't setup up a swap space bc i never imagined needing more than 32G
veske has quit [Ping timeout: 244 seconds]
hyper_ch2 has quit [Remote host closed the connection]
<tilpner>
matt`: Make sure you're running the very latest Nix (perhaps even try unreleased), there have been a fixes that allow hashing in constant space in different situations
<matt`>
thanks tilpner i'll try that
hyper_ch2 has joined #nixos
<tilpner>
Hmm, it's possible that was just NAR download, not local add
<hyper_ch2>
tilpner: so just: path = [ pkgs.system-sendmail ]; or path = [ "pkgs.system-sendmail" ];
<tilpner>
hyper_ch2: no quotes
<hyper_ch2>
(do I even need the pkgs?
<hyper_ch2>
{ config, lib, pkgs, utils, ... }:
<tilpner>
hyper_ch2: Depends on the module, you don't need the pkgs. if you did a (discouraged) with pkgs; or inherit (pkgs) system-sendmail;
<hyper_ch2>
I just leave it
<hyper_ch2>
so, let's rebuild
<hyper_ch2>
tilpner: thx
Tucky has joined #nixos
<hyper_ch2>
zed will be nice to get notification when your zpools have issues
<srhb>
eyJhb: Minor nit, but please alter the humanfriendly commit message to something along the lines of "humanfriendly: add top level attribute" (such that the package name always comes first, like in "foobar: init at ..."
<eyJhb>
srhb: Sooo, I need to change that? :(
<srhb>
eyJhb: Please. :) It should just be a reword and a force push ;)
m0rphism has joined #nixos
<hyper_ch2>
tilpner: now i'm totally confused.. it seems something already does start zfs-zed
<matt`>
tilpner: mind actually pointing out how i can install nix specify a local source dir for nix?
<matt`>
s/install nix//
<tilpner>
matt`: I meant: nix.package = pkgs.nixUnstable;, but I can check if that is worth the risk in a minute
drakonis1 has quit [Read error: Connection reset by peer]
<tilpner>
hyper_ch2: Why is that unexpected? If you enabled zed, your services will be started by zfs.target
<srhb>
eyJhb: Yep, will merge when I have a break from work.
anon has joined #nixos
<eyJhb>
srhb: well currently you are doing great! The weather is amazing ;)
<tilpner>
matt`: Disregard my suggestion, it seems this particular addToStore still reads everything into memory :(
<srhb>
eyJhb: I think it's horrible xD
anon is now known as Guest3254
<srhb>
eyJhb: 28 later, should just take me out and shoot me now.
<matt`>
tilpner: ah ok. so only option would be to create swap space?
<eyJhb>
srhb: haha, so I am guessing waaaay too hot for you :p I have a friend in CPH too, I think she has the same principles currently. SO maybe you can make a deal with each other ;)
<srhb>
eyJhb: That sounds good :P
<tilpner>
matt`: How much memory is actually free of those 32GB?
<matt`>
tilpner: it looks like it only used 60% before stopping...
iMatejC has joined #nixos
<clever>
matt`: i believe nix needs 2 or 3 std::string's, each holding the entire 21gig tar
<tilpner>
D:
<clever>
matt`: worst case, it could need 61gig of ram+swap
nwspk has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<tilpner>
clever: Do you see any reason why we couldn't just replace that StringSink with a streaming alternative?
<tilpner>
I'm surprised it survived this long
<matt`>
clever: ok thanks. i guess i'm setting up some swap then
<clever>
tilpner: yeah, ./foo is always a performance cost
<clever>
tilpner: ive done data processing before with nix, with ~3gig of input files, and ./foo causes the eval to hang for 30+ seconds each time, as it re-hashes it in ram
veske2 has quit [Quit: This computer has gone to sleep]
<tilpner>
:/
<clever>
my solution was to switch it to requireFile, and then add it with nix-store --add-fixed
<clever>
so it only pays that once once
<clever>
but that still has the memory usage problems
<tilpner>
hyper_ch2: Yes, zfs-zed.service is provided by ${zfs}/etc/systemd/system/zfs-zed.service
<clever>
tilpner: one thought i had, is to make a source, that emits ./. as a nar, then run it thru a tee type class, which writes it to a pair of sinks
<siraben>
is it possible to encrypt my ext4 partition after the installation of nixos?
<clever>
tilpner: one sink that hashes, the other that re-unpacks it to /nix/store/temp
<clever>
tilpner: then based on the hash, rename it to /nix/store/hash-name
<clever>
tilpner: but that still has the cost (IO and disk usage) that its going to duplicate the entire thing (21gig in matt`'s case), even if /nix/store already had it
<tilpner>
clever: That looks similar to edolstras commit at a first glance
veske2 has joined #nixos
<tilpner>
Assuming dumpPath creates a NAR dump
<clever>
tilpner: i believe it does
<clever>
tilpner: you can do the same thing with `nix-store --dump /nix/store/foo`
<clever>
(warning, it prints binary to stdout)
hyper_ch2 has quit [Remote host closed the connection]
<hyper_ch2>
tilpner: sorry, the new freenode web irc client just sucks
<tilpner>
hyper_ch2: Why don't you just use hyper_ch? That one seems stable
<hyper_ch2>
tilpner: that's my notebook at home
<hyper_ch2>
which I cannot use here
arahael1 is now known as Arahael
cfricke has joined #nixos
eagle2com has joined #nixos
<tilpner>
Did you try adding it?
nschoe has quit [Client Quit]
bakakuna has quit [Ping timeout: 272 seconds]
<hyper_ch2>
tilpner: I just added systemd.services.zfs-zed.path = [ pkgs.system-sendmail ]; almost at the end and it complains: syntax error, unexpected '=', at /root/nixos-master/nixos/modules/tasks/filesystems/zfs.nix:867:33
bakakuna has joined #nixos
<tilpner>
But it builds without that line?
<{^_^}>
[nixpkgs] @peterhoeg opened pull request #65326 → clang-tools: install clangd as well → https://git.io/fjDus
<hyper_ch2>
yes
<tilpner>
Can you paste the entire file with the line added?
<hyper_ch2>
one option is to remove zfs-zed from: map createZfsService [ "zfs-mount" "zfs-share" "zfs-zed" ]); and create own service file like I did before
<tilpner>
You added it at the wrong location
enuiarte has joined #nixos
<tilpner>
That line is not valid in the context of a list
<tilpner>
You should move it into the first mkIf block
<hyper_ch2>
ah
o1lo01ol1o has quit [Remote host closed the connection]
<hyper_ch2>
I have no idea what I'm doing
o1lo01ol1o has joined #nixos
liberiga has quit [Quit: Leaving]
eagle2com has quit [Read error: Connection reset by peer]
hyper_ch2 has quit [Remote host closed the connection]
psyanticy has joined #nixos
bakakuna has quit [Ping timeout: 245 seconds]
turion has joined #nixos
aexoxea has joined #nixos
hyper_ch2 has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<srhb>
And the answer to your question is in RFC0042
<srhb>
Which was linked :)
<srhb>
hyper_ch2: pending means only you can see it, right?
<tilpner>
hyper_ch2: That "Pending" means GitHub didn't send your comment
<hyper_ch2>
tilpner: no idea
<srhb>
It does.
<hyper_ch2>
why doesn't it send?
<tilpner>
No idea about what? That means you started a review, but didn't submit it
<hyper_ch2>
I didn't do any review.. I just replied
<tilpner>
Yes, and by doing that you started a review
<srhb>
You gotta click more buttons to make it visible :-)
<hyper_ch2>
weird
<tilpner>
There are two buttons when you reply
<hyper_ch2>
but that explains a lot
<tilpner>
"Add single comment" and "Start a review"
<tilpner>
If you just want to send a single comment, click the former
<hyper_ch2>
there's no start a review
<hyper_ch2>
theres' only "[Cance] [Comment]"
<hyper_ch2>
I have no idea what happened now
igor_ has joined #nixos
igor_ has left #nixos [#nixos]
cjpbirkbeck has joined #nixos
<tilpner>
That's because you already started a review. At the very top, there's a "Finish your review" button
<hyper_ch2>
no idea what I did
<hyper_ch2>
I think I undid all comments
<tilpner>
How many did you mean to send?
<hyper_ch2>
anyway, just doing something abstract because you can do something abstract isn't a good argument by it if there's no actual benefit than just having it now abstract
<tilpner>
"why would this abstraction be better except making it more difficult to comprehend/read?" got through
<hyper_ch2>
no idea, I alraedy moved on with that comment
<hyper_ch2>
I did that now again
<tilpner>
If you clicked cancel, then yes, you deleted your comments
<hyper_ch2>
I already moved on
<tilpner>
The benefit is that you can't make typos, and that it's less text and thus easier to read/check
<hyper_ch2>
instead of just making it abstract, it would be much better to loop through the files without having to specify each individual one. But no idea how to accomplish that. At least in the nixpkgs in the phases you can use bash (?) where that is possible but no idea about module files
adamt has joined #nixos
<hyper_ch2>
tilpner: the abstration makes it harder to read/check of what's actually going on
Okinan has quit [Quit: Konversation terminated!]
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
fendor has joined #nixos
<tilpner>
I see how the version in the comment might be harder to read, but I think there's a prettier way to write it
<hyper_ch2>
I'm not against abstraction... just against abstraction without any actual benefit :)
<hyper_ch2>
as said, it would be better if you could just loop through the files in the dir
<tilpner>
Alternative proposal: don't symlink everything individually. Construct new zed.d from old zed.d while adding your zed.rc
<tilpner>
Unless you think a user might want to link in their own files?
<hyper_ch2>
Construct new zed.d from old zed.d while adding your zed.rc --> I have no idea how... and when I attempted, it just didn't let me overwrite the zed.rc
<{^_^}>
[nixos-hardware] @yegortimoshenko pushed 2 commits to master: https://git.io/fjDzm
<{^_^}>
[nixos-hardware] @yegortimoshenko pushed 0 commits to fstrim: https://git.io/fjDzY
<tilpner>
hyper_ch2: Either with buildEnv or runCommand and lndir
<tilpner>
Or cp even
mexisme has joined #nixos
Makaveli7 has joined #nixos
o1lo01ol1o has joined #nixos
<Taneb>
When I try to run "nix-store --import" of something I've just --exported, I get "error: cannot add path '/nix/store/fg4yq8i8wd08xg3fy58l6q73cjy8hjr2-glibc-2.27' because it lacks a valid signature". What's the correct path here?
<tilpner>
But I'm not really convinced of that solution either
<srhb>
Taneb: Either sign it or become trusted enough to not require signatures I guess.
<srhb>
Taneb: (--option require-sigs false should work if you're trusted)
<alexarice[m]>
jackdk: thanks for the suggestion though I don't think this helps that much as I'm trying to write the string "\\.nix'\\" and "\"\\\\.nix'\\\\\"" seems the easiest
<alexarice[m]>
it appears riot also escapes strings so that didn't demo my point very well
o1lo01ol1o has quit [Ping timeout: 246 seconds]
<Taneb>
srhb: thank you
<srhb>
alexarice[m]: Two backslashes, then dot, then nix, then single quote, then two backslashes?
ubert has joined #nixos
<alexarice[m]>
srhb: yes, for use in emacs use-package `:mode` tag
<adamt>
Any chance somebody could backport the cherrypy fix (https://github.com/NixOS/nixpkgs/pull/65217) to 19.03 (the one caused by the upgrade to python 3.7.4)? It doesn't seem to apply cleanly to 19.03, and it's fairly involved, so somebody else should probably take a look at it.
<andi->
gchristensen: that would require a compositor extension that might exist but probably isn't supported by zoom.
<gchristensen>
a compositor extension?
<gchristensen>
fwiw I can screen share under sway with google hangouts/slack/jitsi/appear.in but not zoom -- as long as the other screens are running under Xwayland
<andi->
ahhh
<andi->
you are still trying with legacy foo
<gchristensen>
aye
<gchristensen>
unfortunately I'm switching back to full-on legacy as we speak :P
<andi->
try filing a bug with zoom?
<andi->
Probably not enough market share for them to bother but still
<adamt>
Maybe they could get back some good will if they fixed it, after their screw-up with the mac client.
fragamus has joined #nixos
<infinisil>
andi-: ,locate is using a bit of an outdated nixpkgs version, not sure if that would have an influence
<andi->
I am almost sure that shouldn't but who knows..
selfsymmetric-mu has quit [Remote host closed the connection]
<infinisil>
If I would have to choose a superpower, it would be to immediately know the cause of software bugs/oddities
mexisme has joined #nixos
srid6390803 has joined #nixos
veske has quit [Ping timeout: 245 seconds]
chaoflow has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
enuiarte has joined #nixos
iyzsong has quit [Ping timeout: 250 seconds]
mexisme has quit [Ping timeout: 264 seconds]
<chaoflow>
uname -i says "unknown", same thing in archlinux and alpine docker containers on the same machine, but
o1lo01ol1o has joined #nixos
<chaoflow>
uname -i says "unknown", same thing in archlinux and alpine docker containers on the same machine, but "x86_64" in an ubuntu container. Any ideas?
<chaoflow>
That is on nixos with latest unstable channel.
fragamus has joined #nixos
<infinisil>
That is a bit odd
<infinisil>
Pretty sure uname worked properly at some point
<averell>
uname -m?
<chaoflow>
Need to correct: Wasn't on latest unstable. In the process of updating.
<averell>
cause i get unknown directly on the host
<infinisil>
Oh wait, I was looking at the output of just `uname` and thought I remembered this being longer. But I really remembered `uname -a`, which does include a bunch of stuff
<infinisil>
But the -i flag says it's "print the hardware platform (non-portable)"
<adamt>
Both `uname -i` and `uname -p` returns unknown for my own laptop, and a bunch of servers i just tried, on 19.03. (`-m` works)
<infinisil>
Unknown for all of my machines too
<chaoflow>
infinisil, adamt: Thanks for checking!
domogled has joined #nixos
WeS has joined #nixos
eagle2com has quit [Ping timeout: 248 seconds]
<{^_^}>
[nixpkgs] @Baughn opened pull request #65335 → wifi: Include CRDA regulatory database → https://git.io/fjDrY
chaoflow has quit [Quit: WeeChat 2.5]
mac10688 has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[hydra] @grahamc merged pull request #658 → test.api: use Hydra perl dependencies to run `api-test.pl` → https://git.io/fjwtx
usr has joined #nixos
adamt has quit [Ping timeout: 268 seconds]
<pie_>
elvishjerricco: so im mucking with boot stuff again, i followed your zfs guide but not from an install iso, instead from my running system
<pie_>
and some entries didnt show up in hardware-configuration.nix so i added them manually
<pie_>
but when i try to boot the disk grub says error no such device and dumps some kind of id
<pie_>
and fails to boot
<pie_>
it doesnt ask for a luks password or anything
<elvishjerricco>
pie_: That's odd. You do usually have to manually add luks entries with ZFS, but filesystems themselves have always been properly detected for me. And the grub error is especially weird.
<pie_>
well it kind of makes sense, hold on im still typing
<pie_>
i ran strings on the efi executable and this came up: search.fs_uuid 733f8127ccc7eeea root
<pie_>
i dont have that id anywhere afaict
<pie_>
and thets the id that shows up in the grub error
<pie_>
so i have no idea what is supposed to set that value correctly. this should be what grub uses to search for the luks volume
<elvishjerricco>
That's really weird. I can't imagine why grub would make up a uuid like that
<elvishjerricco>
pie_: You said you're doing this from a full system rather than an iso. Does that mean you're attempting to install this on a system that already has another nixos install on it?
<pie_>
unless nixos-install --root some/path is doing something weird, but even then
<pie_>
yes that is correct
<elvishjerricco>
Worth noting that `installAsRemovable` will overwrite the boot loader of the other system, if they're sharing an ESP.
<pie_>
no tihs is being installed on a completely separate drive
<elvishjerricco>
ah
<elvishjerricco>
Anyway, that UUID has to exist somewhere. There's no way it's randomly generated :P
<pie_>
i dont see anything like it in blkid
<pie_>
unless its some weird encoding
<pie_>
but it cant be because thats just an argument to the grub search command
<pie_>
elvishjerricco: uhh can i maybe tell it to build the system from a different nixpkgs root
<pie_>
though i should still be on stable..
o1lo01ol1o has quit [Remote host closed the connection]
<pie_>
and unstable doesnt allow boot breaking things...
<pie_>
man what the heck is going on here
dingo has joined #nixos
<dingo>
if i create a nixos-container with --config-file i can't specify containers in there right? (like in option 2: http://paste.debian.net/1092921/ )
o1lo01ol1o has joined #nixos
<elvishjerricco>
Yea it's weird. I can't think of anywhere a UUID would come from that wouldn't show up in blkid
<pie_>
elvishjerricco: any idea how to find the code in the module that passes this info to grub
<{^_^}>
[nixpkgs] @vbgl pushed commit from @marsam to master « dune: 1.10.0 -> 1.11.0 »: https://git.io/fjDoI
tonic1 has joined #nixos
ilya-fedin has quit [Quit: Connection closed for inactivity]
<matt``>
hello, how can i specify a custom source repo for nix? I'm trying to test out a pr. I've tried adding a nixSrc attribute in pkgs/tools/package-management/nix/default.nix (and in all-packages) but it's telling me that nix requires boost
matt```` has quit [Remote host closed the connection]
<Thra11>
Is there a built in or common way to deal with the case where the sources of a package are in a subdirectory of a git repo or tarball?
erasmas has joined #nixos
<adisbladis>
Thra11: You can set sourceRoot on the derivation
<pie_>
elvishjerricco: yeah i cant really find the code that sets this. grub docs talk about embedde configuration files but nothing in nixpkgs uses grub-mkimage for this..not sure how grub might be getting that value..
<pie_>
elvishjerricco: can you check whats in your boot64.efi ?
<pie_>
elvishjerricco: just run strings on it
<pie_>
it should be at the very end
<pie_>
Thra11: i think you can set sourceDir but im not 100% sure thats the appropriate solution
<pie_>
Thra11: ok its called something else but..
<Thra11>
adisbladis: Thanks
<pie_>
oh i didnt notice adisbladis xD
<pie_>
adisbladis: hm..sourceRoot doesnt show up in the documentation anywhere? but apparently src is a thing
<elvishjerricco>
pie_: What package is strings in?
<pie_>
oh duh. i dont use nix for a week and this is what happens. i remember src
<pie_>
elvishjerricco: binutils-unwrapped
<infinisil>
Thra11: To be a bit more specific, you need `sourceRoot = "source/thesubdirectory"`
<elvishjerricco>
And what should I be looking for in the output?
<pie_>
sourceroot is in the nixpkgs manual. im just bad at searching apparently.
<pie_>
elvishjerricco: the last three lines or so of the output should be human readable unless youre config is somehow doing something totally different
<pie_>
*your
<pie_>
clever: i have more "fun" grub things if youre interested
<elvishjerricco>
pie_: My last two lines are `cryptomount -u <MY DISK'S UUID>` and `(cryptouuid/<MY DISK'S UUID>)/root/boot@/grub`
* pie_
scratches head
<pie_>
well thats not what i have
<Thra11>
infinisil: With literal "source"?
<elvishjerricco>
That's the output of `strings /efi/EFI/BOOT/BOOTX64.EFI`
Soo_Slow has joined #nixos
<elvishjerricco>
pie_: I'm also.... still on 18.09 :P
<elvishjerricco>
So maybe Grub broke this setup in 19.03?
<Thra11>
infinisil: Or am I supposed to replace the word source with e.g. the name of the root directory in the tarball?
<andi->
elvishjerricco: I am runnnig a very similar setup with 19.03 and unstable
inquisitiv3 has joined #nixos
<elvishjerricco>
andi-: Are you booting straight from ZFS over LUKS with grub? No kernel on a boot partition?
<andi->
I have /boot on a luks device and grub (stage1) decrypts that
<andi->
and afterwards it can boot zfs and whatnot
<pie_>
elvishjerricco: im pretty surei was using 19.03 on the livecd when i did this last time
<elvishjerricco>
Ah. Well I've heard grub's zfs support can be flaky, so maybe it doesn't like ZFS over LUKS anymore
<andi->
I would probably not want to depend on that support... Things like linuxboot might become better alternatives at some point
<elvishjerricco>
Yea. My loooong term goal is to make a little boot loader out of Linux, since I can't seem to figure out linuxboot :P
<andi->
What is actually missing to use nixpkgs to create a "boot loader"? The booable efi payload bits?
<sphalerite>
andi-: what do you mean? nixpkgs already builds grub and systemd-boot for instance..?
<pie_>
why is grub-install a 2000 line c file ;_;
<andi->
sphalerite: the idea is to no longer have systemd-boot or grub as boot loaders since those routinely have to reinvent drivers (think zfs, raid, $hardware) and instead just oint your UEFI/coreboot/u-boot/… at a linux image that acts as bootloader
<sphalerite>
andi-: nothing stopping you from doing that AFAIK. :)
<andi->
sphalerite: exactly that was my question. There is some work on creating standard boot specs so you can have one bootloader that is able to boot dual/multi-boot systems (including windows, …)
<andi->
and that is the harder part of the job
shibboleth has quit [Quit: shibboleth]
__Sander__ has quit [Quit: Konversation terminated!]
<sphalerite>
andi-: you mean… like EFI? \:|
<sphalerite>
s/EFI/UEFI/
adamt has joined #nixos
<andi->
sphalerite: well but not re-implemented by every company producing mainboards
<pie_>
blkid doesnt show anything weird when i use nixos-enter...
<infinisil>
Thra11: Yeah literal source
alex`` has joined #nixos
srid6390803 has quit [Ping timeout: 244 seconds]
<infinisil>
Thra11: By default sourceRoot = "source" I think
<sphalerite>
Has anyone else noticed that opening URLs from another program in firefox has got really slow recently? `firefox https://xkcd.com` takes about 5 seconds for me, used to be practically instant
kraem has quit [Ping timeout: 248 seconds]
kraem has joined #nixos
<pie_>
elvishjerricco: i think it might not be enabling the crypto stuff for some reason but im not sure what the underlying issue is...
<andi->
sphalerite: which firefox version? It is still instant for me even through additional layers of scripts
<literon>
hi! I wonder if it's possible to control the derivation name of sourceByRegex and pals (cleanSourceWith-based).
wfranzini has quit [Remote host closed the connection]
<literon>
I filter ./., which happens to be the top-level directory. The parent dir name seems to be included in the derivation, which changes its hash.
<infinisil>
literon: I think you might be able to do `(sourceByRegex ...).overrideAttrs (old: { name = "foo"; })`
<{^_^}>
[nixpkgs] @ttuegel opened pull request #65339 → pinentry: Use qt5.wrapQtApps as needed → https://git.io/fjDKu
<{^_^}>
[nixpkgs] @doronbehar opened pull request #65340 → mpd-mpris: init at 0.2.3 → https://git.io/fjDKa
wildtrees has joined #nixos
Guest3254 has quit [Quit: WeeChat 2.4]
KingOfKings has joined #nixos
<pie_>
so after a rebuild switch how do i get the iso to load new kernel modules?
KingOfKings has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @WilliButz opened pull request #65341 → highlight: 3.43 -> 3.52, use GitLab source → https://git.io/fjDK1
<slabity>
pie_: I believe the modules are automatically loaded, unless the kernel itself was updated. In which case you will need to reboot
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
KingOfKings has joined #nixos
turion has quit [Ping timeout: 246 seconds]
<slabity>
pie_: Check with `lsmod`, and if it's not loaded then try to manually load it with `modprobe`
KingOfKings has quit [Remote host closed the connection]
<pie_>
hm ok
<pie_>
ill give it a shot
KingOfKings has joined #nixos
KingOfKings has quit [Remote host closed the connection]
KingOfKings has joined #nixos
<{^_^}>
[nixpkgs] @bradleyjensen opened pull request #65342 → Improve ergonomics of clwrapper → https://git.io/fjDKQ
<slabity>
Is there any way to declaratively set the nixpkgs channel without needing to redownload the entire thing each time I run an update? Or need to change the sha256 every time upstream updates
KingOfKings has quit [Remote host closed the connection]
<pie_>
slabity: insmod complains about unknown symbol in module
<vaibhavsagar>
slabity: you want to declaratively set something that changes?
<slabity>
pie_: Not sure what to do about that
domogled has quit [Quit: domogled]
<vaibhavsagar>
you can use `builtins.fetchTarball` to use a single version of nixpkgs
<slabity>
vaibhavsagar: Yea, I know it's impure
<vaibhavsagar>
then you also know that what you're asking for is unreasonable
<slabity>
The problem with fetchTarball is that it redownloads each time it evaluates
<Thra11>
infinisil: Thanks. How do I use it with buildLinux? I tried just setting sourceRoot = "source/subdir". It didn't complain, but it still used "source" as the root.
judson has joined #nixos
<literon>
Is there a difference how nix-store --add hashes inputs between, say, nix 2.1.x and 2.2.x?
hoijui has joined #nixos
hoijui has quit [Remote host closed the connection]
englishm has quit [Excess Flood]
englishm has joined #nixos
<infinisil>
Thra11: What's the full expression?
<tilpner>
literon: I haven't had to update basic output hashes. Which ones are you concerned about?
<Thra11>
infinisil: Well I've got something very similar to linux-hardkernel-4.14.nix. Just using fetchFromGitHub to build linux from a different set of sources. The problem is that the source repo is massive dump of related BSP stuff, with the kernel sources as a subdirectory.
bgamari has joined #nixos
<infinisil>
Thra11: Ah, so the arguments buildLinux takes aren't derivation arguments
<infinisil>
Thra11: So to override sourceRoot you'll have to do `(buildLinux { ... }).overrideAttrs (old: { sourceRoot = "...`
gagbo has quit [Ping timeout: 245 seconds]
<literon>
tilpner: I'm trying to debug hashes over a bunch of files differ on some machines. I'll keep digging then.
<literon>
since it doesn't go through `nix-serve`, I would assume not. Just curious.
<pie_>
elvishjerricco: grub-probe -d /dev/mapper/some_crypt_thingy -t fs_uuid returns the bad uuid
<infinisil>
Thra11: Hm, I'd have said that looks
<elvishjerricco>
Huh...
<infinisil>
correct
<pie_>
elvishjerricco: which doesnt really explain whats happening internally but its irrelevant at least
magnetophon has quit [Ping timeout: 246 seconds]
<pie_>
elvishjerricco: man how does that even work...blkid showd completely different values than what grub says
<elvishjerricco>
Maybe just different formats of the same underlying data?
<{^_^}>
[nixpkgs] @kalbasit merged pull request #65288 → pythonPackages.joblib: fix the build on Darwin → https://git.io/fjDUO
<{^_^}>
[nixpkgs] @kalbasit pushed to master « pythonPackages.joblib: fix the build on Darwin (#65288) »: https://git.io/fjD6X
<infinisil>
Thra11: Does `env | grep sourceRoot` output the correct value when you nix-shell into the build
warbo has joined #nixos
<Thra11>
infinisil: Could you remind me how to do that? I've forgotten which argument I need to pass to nix-shell
<infinisil>
nix-build -A foo -> nix-shell -A foo
<pie_>
elvishjerricco: should it even be looking at that?
<infinisil>
Thra11: So probably just the same
<elvishjerricco>
I have no idea. I really don't know much about grub :P
<pie_>
did i accidentally tell it to use the decrypted disk somehow
* pie_
checks
<infinisil>
Fun fact: nix-build and nix-shell are the same executable
<pie_>
nope i told it to use the right uuid...so why the F does grub decide to use this random /dev/mapper device
<{^_^}>
[nixpkgs] @costrouc opened pull request #65344 → [WIP] Enable testing in core scientific python libraries and move to github as src → https://git.io/fjD69
<Thra11>
infinisil: Ah. I think the problem is that the linux-config dependency is the first failure, but I'm asking it to drop me into a shell ready to build linux itself...
* pie_
scratches head - but i cant do this without having that mounted
Edes has joined #nixos
<infinisil>
Thra11: Ah heh, well then the problem is clear, you've been trying to fix the wrong build
<Edes>
Hi, I'm having an issue when running nixos-rebuild switch
<infinisil>
Thra11: Well, not the one that has been failing at least
<Edes>
also the internet seems to fail sometimes when I run it until reboot, not sure what else breaks
inquisitiv3 has quit [Ping timeout: 258 seconds]
<ldlework>
i don't know much about it; will it be difficult if I want to get a Windows VM running on NixOS?
<infinisil>
Edes: What are the permissions on the two directories?
<niso>
ldlework: getting a windows vm to run should be straight forward
<niso>
ldlework: just use e.g. virt-manager
<Thra11>
infinisil: So I need to work out how to get sourceRoot overridden in the config as well. It looks like it inherits all the other stuff (src, version, etc.) in generic.nix
<infinisil>
Yeah so the problem is at the line I linked
<Edes>
this isn't on 19.03 because it messed up when I switched to unstable
ris has joined #nixos
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
magnetophon has joined #nixos
<ldlework>
niso: I'm guessing once I do `virtualisation.libvirtd.enable = true;` I have to reboot?
<infinisil>
Oh snap, this has been in the couchdb module since it was added in 2014
<Thra11>
infinisil: Well, sticking sourceRoot = ... in the middle of generic.nix seems to have worked. There's probably a way to override it properly, but it's beyond me.
<Thra11>
infinisil: Thanks for your help!
<infinisil>
Thra11: Ah neat, not sure how that works, but glad you got it :)
<niso>
ldlework: probably, and don't forget to add the "libvirtd" group to your current user
<Izorkin>
Please check and merge #61816 #63150 #61722 #61671
<pie_>
if that was enabled id probably have a different boot config...
<pie_>
so what is going wrong..
<Thra11>
infinisil: My (rather limited) understanding is that generic.nix line 94 is constructing the derivation which creates the linux config file. Being an ordinary mkDerivation, you can set sourceRoot there.
<niso>
ldlework: does the error remain if you run it as root?
Olgierd has quit [Quit: WeeChat 1.9.1]
<pie_>
niso: yes. ive gotten it to work before and it was nowhere near this complicated. something seems broken with grub-install or related stuff, i dont know..
zupo has joined #nixos
inquisitiv3 has joined #nixos
<pie_>
i have boot.loader.grub.enableCryptodisk set to true but it doesnt actually seem to be enabled
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pie_>
everyone else's luks isnt broken so i _have_ to be doing something dumb
<niso>
maybe a typo in the uid? :P
<infinisil>
Edes: Actually, you'll probably encounter this error every time you do a nixos-rebuild switch
<infinisil>
Well, not 100% sure on that, but that might happen
<infinisil>
I'm currently writing a PR to fix it though
<elvishjerricco>
pie_: Maybe try going through the steps in a VM first and see if you can get it working there. If so, maybe your hardware is confusing grub somehow?
<yorick>
pie_: works4me
<yorick>
pie_: paste your config?
<niso>
is it possible to use containers.bindMounts with imperativly managed containers?
<{^_^}>
[nixpkgs] @costrouc opened pull request #65346 → pythonPackages.simple-salesforce: init at 0.74.3 → https://git.io/fjDib
<elvishjerricco>
yorick: Fyi, he's trying to get grub to boot straight from LUKS, not just have root on LUKS.
<yorick>
elvishjerricco: my roommate has that I think
<yorick>
pie_: what's the security advantage of this?
<elvishjerricco>
Not much
<elvishjerricco>
But I personally just like having as much as possible in my ZFS pool instead of a FAT partition
<elvishjerricco>
And in theory, with secure boot, it can protect you from your kernel being modified without having to sign every kernel
<elvishjerricco>
Not that signing every kernel is legitimately a problem :P
ixxie has joined #nixos
<yorick>
elvishjerricco: I'd go for TPM to accomplish that
<yorick>
(attestation)
<infinisil>
Edes: Got a nick on Github so I can mention you?
<elvishjerricco>
TPM is harder to manage since you have to re-bootstrap every time your kernel, initrd, parameters, or boot loader changes.
<pie_>
right now im trying to dig out my old config thats supposed to work and see if i can get it to work
<niso>
infinisil: pls link the pr here
<pie_>
but i managed to break the zpool somehow now
<elvishjerricco>
And it's also in theory possible to extract the secret from the tpm.
<infinisil>
niso: Yeah I will, just asking so I can give credit to the reporting of the error
<elvishjerricco>
But the signing key for secure boot can be exposed only when you need it
knupfer has joined #nixos
<elvishjerricco>
pie_: Are you sure it's not a bad drive?
<pie_>
elvishjerricco: no im not sure but i have a filesystem with nix on it so...
<pie_>
?
<{^_^}>
[nixpkgs] @Infinisil opened pull request #65347 → Couchdb: Don't chown /var/log to couchdb → https://git.io/fjDPU
<infinisil>
niso: ^
<infinisil>
Edes: ^
<niso>
infinisil: just realized that you are the all-hies guy...
<infinisil>
Indeed, hello there :)
<niso>
nice work, i'm looking forward on using it at some point :)
___laika has quit [Ping timeout: 245 seconds]
inquisitiv3 has quit [Remote host closed the connection]
alp__ has joined #nixos
inquisitiv3 has joined #nixos
gratto has joined #nixos
alp has quit [Ping timeout: 264 seconds]
<maralorn>
infinisil: With which editor do use hie?
<niso>
maralorn: pretty much all of them (which support lsp)
<clever>
pie_: ?
* niso
gives it another try:
<niso>
is it possible to use containers.bindMounts with imperativly managed containers?
<petercommand>
edef: I am trying to override this package with a developmental version, but wasn't able to build the project properly
<clever>
pie_: one min, something came up
abrar has joined #nixos
<elvishjerricco>
pie_: Have you tried it in a VM yet?
<alienpirate5>
looks impure edef
<edef>
hmm?
<petercommand>
edef: and the issue seems like the cmake project is trying to clone another project from github, and it wasn't able to clone into the source directory
<edef>
petercommand: right, you probably want to use fetchgit with enableSubmodules = true;
<edef>
er, fetchSubmodules = true i think
<edef>
or maybe patch the thing it's trying to download in yourself
<petercommand>
ah, thanks, will try fetchSubmodules = true
<edef>
builds aren't allowed to touch the network because that could introduce impurities
<edef>
note that fetchSubmodules = true will require a new sha256
<pie_>
elvishjerricco: i just broke my vm i think
<pie_>
elvishjerricco: what was i supposed to try again
<petercommand>
edef: I used builtins.fetchGit
<pie_>
(man nothing is going right today)
<edef>
petercommand: that works
xwvvvvwx has joined #nixos
<elvishjerricco>
pie_: Oh wait have you been doing all of this in a vm?
<pie_>
clever: no rush im busy trying to unf*** things
<pie_>
elvishjerricco: ive een doing both
<pie_>
elvishjerricco: i ahve the same problems in a vm with iso
<petercommand>
edef: adding it doesn't work at all..
<edef>
petercommand: sec
<elvishjerricco>
Oh. Huh...
<edef>
petercommand: you're just trying to build taskwarrior master?
<petercommand>
edef: not master
<xwvvvvwx>
Hey I'm trying to use the docker-containers module (
<edef>
petercommand: a branch close to it, at lesat?
<edef>
*least
<petercommand>
edef: a branch
<petercommand>
edef: 2.6.0
<edef>
like, libshared seems to be a submodule
<petercommand>
yes
Edes has quit [Ping timeout: 245 seconds]
<edef>
is that what you're trying to take care of?
<petercommand>
maybe I shouldn't use builtins.fetchGit?
<edef>
i don't think builtins.fetchGit supports fetchSubmodules
<edef>
in fact it gives an error here
<edef>
error: unsupported argument 'fetchSubmodules' to 'fetchGit', at (string):1:32
Guest9944 has joined #nixos
<xwvvvvwx>
but I keep getting an error (The option docker-containers does not exist). I have an exact copy of the mainline tests in my config file. I am using nixops to try and deploy to hetzner. My config is here: https://gist.github.com/xwvvvvwx/fa332632f5d325f2cdb542d6b5efd753
<edef>
so i'm reasonably confident that isn't getting passed in your code
<pie_>
xwvvvvwx: you can run anything from an activation script. there might be more standard ways to do such things though
<xwvvvvwx>
nisowhat is this doing exactly?
<xwvvvvwx>
pie_ what is an activation script?
<ajs124>
sounds like a use for tmpfiles
ambro718 has quit [Quit: Konversation terminated!]
magnetophon has quit [Ping timeout: 272 seconds]
Thra11 has joined #nixos
andrewrk has joined #nixos
<andrewrk>
I just got a Dell XPS 15 7590 and I'm pretty sure I need the latest linux kernel for the wifi to work. is that an option for installing nixos?
<niso>
xwvvvvwx: it allows to set permissions on temporary files
<niso>
xwvvvvwx: other than that you can always stick to bash
<xwvvvvwx>
@niso
<xwvvvvwx>
yeah just got it figure out, worked like a charm
<xwvvvvwx>
thanks!!!
<niso>
yw
vmandela has left #nixos ["Leaving"]
<pie_>
clever: ok managed to unbork zfs
<{^_^}>
[nixpkgs] @matthewbauer opened pull request #65349 → Lookup hunspell dictionaries in XDG_DATA_DIRS → https://git.io/fjDXI
magnetophon has joined #nixos
jabranham has joined #nixos
xwvvvvwx has quit [Remote host closed the connection]
<pie_>
elvishjerricco: ok yeah either its my current system or current nixpkgs thats breaking this because my old config also appears to be broken. (or im doing something wrong)
<andrewrk>
vaibhavsagar, that seems good, do you know the command to do it?
fresheye1 has quit [Quit: WeeChat 2.4]
<elvishjerricco>
pie_: I'll try it out myself when I'm at my desk shortly
<pie_>
elvishjerricco: what i dont understand is wouldnt this break everyone's systems? when does the bootloader get reinstalled?
<buckley310>
nixpkgs/nixos/modules/installer/cd-dvd/installation-cd-minimal-new-kernel.nix this looks like a prefab config for what you need
<pie_>
actually no, it cant be my current system because vm also broke
shibboleth has quit [Quit: shibboleth]
<elvishjerricco>
pie_: Most people don't use grub to boot directly from grub. Most people leave the kernel on an unencrypted partition and let the kernel do the decryption
<elvishjerricco>
Directly from LUKS*
<vaibhavsagar>
andrewrk: actually, look at nixos/modules/installer/cd-dvd/installation-cd-minimal-new-kernel.nix
<pie_>
elvishjerricco: aha.
stepcut has quit [Remote host closed the connection]
<vaibhavsagar>
that looks like it has the latest kernel
<andrewrk>
perfect
<pie_>
time to try to roll back nixpkgs and maybe bisect
<niso>
pie_: just curious: how would you bisect this issue?
SovereignBleak has quit [Quit: Connection closed for inactivity]
<tilpner>
andrewrk: You could also download one from Hydra
<pie_>
niso: instead of rebooting im just checking what gets written to the end of my efi binary thingy
<tilpner>
(Give me a few minutes to get the link, hydra is slow)
orivej has quit [Ping timeout: 258 seconds]
knupfer has joined #nixos
<niso>
ah, good idea :)
<pie_>
niso: im just assuming that and cryptodisk broke at the same time
<pie_>
assuming those are at all related, im rather clueless
<andrewrk>
tilpner, thanks! I just received this new laptop in the mail 10 minutes ago, and leaving for a flight in 25 minutes... I hope I can find out if the wifi will work with the new linux kernel by then!
orivej has joined #nixos
<andrewrk>
what's the best partition table type again for the usb stick? gpt? msdos?
<niso>
same here.... i'm trying to move from docker-compose to nix for project development since a couple of days now, and i'm running in one issue after another (just started reading into dolstras paper...)
* niso
assumed the whole thing was more straight-forward
<pie_>
niso: good luck! nix can be rather difficult sometimes >_>
<pie_>
and when stuff breaks you gotta start digging...
<tilpner>
andrewrk: I found it, but I forgot that it wasn't graphical. You can still use it, it just doesn't give you plasma. https://hydra.nixos.org/build/96749123
<andrewrk>
I don't need graphics. thanks!
<niso>
pie_: thanks, to you too...
o1lo01ol1o has joined #nixos
<niso>
pie_: nothing broke here, guess i'm just lacking one or two litte infos to get stuff running
<pie_>
yeah just saying 'xD
<pie_>
not that its /necessarily/ different for any other linux
<pie_>
i suppose id be kind of screwed if i had to dig around in grub guts in either case, but now id have to figure out how to hook nixos-install or something :I
<andrewrk>
tilpner, is there a way to tell what kernel version this is?
<pie_>
hm...how _would_ one attach a debugger to an activation script?
<lordcirth_>
andrewrk, on a running system? uname -a
<andrewrk>
yeah I'll just boot into it and find out
o1lo01ol1o has quit [Ping timeout: 246 seconds]
fragamus has quit [Ping timeout: 248 seconds]
lovesegfault has joined #nixos
kiloreux has joined #nixos
<pie_>
niso: alternative bisect methods: 1) cry and reboot a lot, 2) use a vm
<kiloreux>
How can I avoid my package being removed by `nix-collect-garbage` and `nix-store --optimise` ?
<pie_>
kiloreux: you give it a root
<niso>
pie_: 3) use aws or hetzner-cloud
<pie_>
i dont know off the top of my head how to do that though, maybe check the nix-store man page for something like --add-root or somesuch
<kiloreux>
Ahh okay, so it's collect-garbage that's guilty here.
<Smith[m]>
Hello ! I'm trying to setup wireguard and I was wondering how can I make private keys available before hand ? Is there some kind of hooks where I can execute a script to fetch my private keys ?
<adamt>
kiloreux: Yup, and you need a link to the pkg that nix knows about, to avoid it being removed. Try googling for gc roots in nix, that can be one solution
Edes has quit [Ping timeout: 268 seconds]
<pie_>
elvishjerricco: unless i did something wrong - im using a checkout of 18.09 and its still broken sooo... :O
<elvishjerricco>
I'm *pretty sure* nixos-config=/etc/nixos/configuration.nix is right, and that you don't need mnt in front because it's relevant after the chroot, but you'll want to double check me on that
<adamt>
andrewrk: If you already have access to nix somehow, you can build an installer containing a newer kernel, although it would have to be based on something newer than stable, unless you want to overlay the kernel yourself.
<pie_>
niso: the point was i think he uses 18.09 and it works for him
<elvishjerricco>
niso: It is but I haven't had time to upgrade and I'm pie_'s reference on how to do what he wants to do :P
orivej has quit [Ping timeout: 245 seconds]
<pie_>
which shoudnt really matter because i had this workng with 19.03 before im _quite_ sure...
<niso>
i see
<pie_>
there has to be something really stupid behind this...
<colemickens>
Hum. XCB is no longer an available QT_QPA_PLATFORM for me. Did something change regarding it possibly?
virus_dave has quit [Quit: virus_dave]
mexisme has joined #nixos
<elvishjerricco>
pie_: Oh crap, I got the same error in my VM
Drakonis has joined #nixos
dl4mfo_ has quit [Ping timeout: 246 seconds]
<pie_>
elvishjerricco: oh crap :D
* pie_
cries
<pie_>
elvishjerricco: this is with your working config yeah?
<elvishjerricco>
There had to be some change in the grub install script or something
<elvishjerricco>
Yea, but I don't know what version of nixos I'm on specifically :P
chloekek has quit [Ping timeout: 244 seconds]
<elvishjerricco>
Let me find out and you can try that revision
<pie_>
so maybe need an older rev than 18.03 becuase you havent reinstalled the bootloader since?
<elvishjerricco>
pie_: 001b34abcb4
<pie_>
i dont know when the bootloader gets reinstalled but im guessing nix somehow keeps track
<clever>
pie_: one sec
<pie_>
elvishjerricco: this is what breaks for you though isnt it?
<clever>
pie_: the state file you pasted, says that the ESP should be mounted to /efi (or /mnt/efi when running install)
<pie_>
clever: yeah it is, thats fine
<elvishjerricco>
clever: So the issue we're having is that attempting to boot directly on zfs over luks from grub isn't working because grub appears to be looking for the zfs pool without opening the luks volume.
<pie_>
clever: ive even removed the file before to check if its generated
<elvishjerricco>
it used to do this correctly
is_null has joined #nixos
<elvishjerricco>
pie_: You should try that rev I posted
Jackneill has quit [Remote host closed the connection]
<elvishjerricco>
It's what my working system uses
<clever>
elvishjerricco: ah, ive never trusted grub with zfs, and i dont bother encrypting /boot either
alp__ has quit [Ping timeout: 252 seconds]
<pie_>
ok 18.03 "still " (???)broken
<pie_>
elvishjerricco: ok will try rev, but you said its borked?
liberiga has joined #nixos
<elvishjerricco>
No, I tried 19.03 in a VM
<elvishjerricco>
I haven't tried that rev in a VM, but it currently works on my system
<pie_>
(im not using vm right now)
<pie_>
(since i figured it doesnt help)
troydm has quit [Ping timeout: 268 seconds]
<pie_>
(and its annoying to un/remount stuff constantly)
<elvishjerricco>
My point being I haven't tried installing from scratch with that rev
<clever>
pie_: why are you encrypting /boot/?
troydm has joined #nixos
<pie_>
clever: dunno elvishjerricco did it and i liked how he did it
<pie_>
:p
porrifolius_ has joined #nixos
<elvishjerricco>
I personally just like that there is only a single file not checksummed by zfs :P
<pie_>
thats a plus. also muh full disk encryption (yeah i know the bootloader isnt verified)
<pie_>
and i have plenty of space on the esp for some "extra" kernels :p
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to master: https://git.io/fjD1s
<alienpirate5>
I'm trying to set up remote builder on a Mac
gratto has quit [Read error: Connection reset by peer]
<elvishjerricco>
I think I'm an idiot
<alienpirate5>
getting a lot of errors
matt` has joined #nixos
gratto has joined #nixos
<elvishjerricco>
I think in my test that failed, I accidentally used the dm-crypt block device in my `boot.luks.devices.test.device` option in `configuration.nix`
<elvishjerricco>
definitely not supposed to do that
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
<clever>
pie_: brb
<pie_>
clever: i forgot, the reason i wasnt really asking about this in #grub is we have all this nix wrappin stuff around it and uhh...
<pie_>
cant really tell if the problem is nix or grub
<clever>
pie_: i'm going to lean towards it being grub
<pie_>
elvishjerricco: not sure what thats supposed to mean but switching the unencrypted and encrypted here sounds like a typical problem :D
<pie_>
clever: i did peek at the file but didnt manage to glean anything informative out of it
<pie_>
clever: grub-install seems to generate the boot config in question internally
<clever>
pie_: so you can just tell #grub what grub-install args are being used, and nix is out of the picture
<pie_>
ah ok
<clever>
pie_: you could also `strace -f -e execve -s 300 .... nixos-install ...` to see the exact commands it ran
<elvishjerricco>
pie_: Can you post your `configuration.nix`?
<clever>
if you dont feel like emulating perl in your brain
<pie_>
thats probably going to be helpful anyway
<pie_>
elvishjerricco: its uhhh complicated but i can paste the "relevant" parts?
<clever>
pie_: you could also then manually run grub-install in a chroot, to see what its doing
<pie_>
elvishjerricco: does your stuff work again?
<elvishjerricco>
pie_: Installing now. Will see soon
<pie_>
clever: i suppose this will also be helpful for figuring out wtf is going on here
<elvishjerricco>
I just need to see your `boot.initrd.luks` section
<pie_>
so im trying to dump the commandline right now
<elvishjerricco>
(protip when nixos-installing a bunch in one day: the iso has max-jobs = 1; use --max-jobs auto to speed up the part where it's just building hundreds of text files)
<elvishjerricco>
Uhhh, this time I didn't even get past the firmware. Grub didn't even start
nixnothing has joined #nixos
<clever>
pie_: reading grub-install.c, i can see a --disk-module arg, that specifies what grub driver to bake into itself, so it can find the next stage, but efi doesnt set one
<elvishjerricco>
Oh, that's what happens when you forget the crucial config line that tells it to install the boot loader :P
<pie_>
clever: well this doesnt seem particularly...specific to anything: /nix/store/533s8q08p0zma4jdnlrxd23zy4aasyv3-grub-2.02/sbin/grub-install --recheck --target=x86_64-efi --boot-directory=/boot --efi-directory=/efi --no-nvram --removableInstalling for x86_64-efi platform.
<clever>
legacy will bake in "biosdisk" which is how you read a hdd via the legacy bios api
<pie_>
clever: actually i bet half of the grub config is in env vars
<clever>
pie_: grub-install doesnt use very many env vars, its likely more inspecting `mount` to see what is mounted where
<elvishjerricco>
(pie_ I bet I know exactly what your problem is if you just show me your luks nix expression :P)
<pie_>
elvishjerricco: sec
<pie_>
elvishjerricco: ok thats actually wrong right now because i switched to testing with my old config...
<nixnothing>
what is pie_ acting as a bouncer for? some rasbpi project?
<pie_>
elvishjerricco: which is funny because it means what im observing is actually independent of the settings there...
<alienpirate5>
is there no way to run a remote builder that cross comiles?
<clever>
alienpirate5: thats entirely possible, but you need to ask for a cross-compile, it wont magically cross with the right build slave
<clever>
alienpirate5: for example, if i try to do a linux->windows cross compile, while on a mac, it will find a linux build slave to do the work
<niso>
the tex files of Dolstras thesis arn't public, right?
<elvishjerricco>
pie_: Which block device does that `/dev/disk/by-uuid/...` point to?
<clever>
pie_: brb
<pie_>
elvishjerricco: sdc2 which is the luks
<alienpirate5>
clever: wjat O ,eman os that it would cross compiule for linux when running on a darwin machine or something
<elvishjerricco>
Oh, ok so maybe I'm wrong...
<alienpirate5>
* @freenode_clever:matrix.org: What I mean is that it would cross compile for linux when running on a darwin machine or something
slyfox has joined #nixos
<elvishjerricco>
Anyway 18.09 worked for me out of the box
<elvishjerricco>
once I fixed my goof
<nixnothing>
wtf, where is /var/lib ..
psyanticy has quit [Quit: Connection closed for inactivity]
<nixnothing>
jk . recent nixos convert
<elvishjerricco>
Trying that de-goofing on unstable now
adamantium has joined #nixos
<alienpirate5>
it does not exist
<alienpirate5>
*
<pie_>
so...how does /nix/store/533s8q08p0zma4jdnlrxd23zy4aasyv3-grub-2.02/sbin/grub-install --recheck --target=x86_64-efi --boot-directory=/boot --efi-directory=/efi --no-nvram --removable decide what the main device is?
<nixnothing>
EVERYTHING IMMUTALATE
<pie_>
or rather what device to boot...
<nixnothing>
immutable*
<pie_>
obviously /efi is where it installs to ...
<elvishjerricco>
pie_: Yea, de-goofing fixed it. I managed to use unstable in a VM just fine
<nixnothing>
i've been considering switching to unstable..
<nixnothing>
one think that has confused me is is configuration.nix verion end line though
is_null has quit [Remote host closed the connection]
<nixnothing>
system.stateVersion = "19.09"; # Did you read the comment?
<pie_>
elvishjerricco: is it possible grub is reusing some state i have somewhere and thats why this is constantly broken
<niso>
nixnothing: something something rfc
<gchristensen>
niso: hm?
<adamt>
nixnothing: I think the comment above it is pretty clear :P
<nixnothing>
pie_: what grub issues are you having
<elvishjerricco>
pie_: Possibly. Maybe there's some in the efi vars in NVRAM?
<ajs124>
pie_: did you remove the contents of /boot/ and /efi/before running this?
<pie_>
elvishjerricco: any idea how to dump that
<nixnothing>
adamt: I figured mostly for dbs, but just unsure what else it effects...
<elvishjerricco>
pie_: Or maybe there's some other disk on the system with grub on it?
<samueldr>
,stateVersion
<{^_^}>
Setting stateVersion to the latest release doesn't upgrade anything and can only break your setup at best. To actually upgrade NixOS see https://nixos.org/nixos/manual/#sec-upgrading. If you want to update the stateVersion option regardless, Ctrl-F for "stateVersion" in https://nixos.org/nixos/manual/release-notes.html to see things that need to be manually migrated with the new value.
<pie_>
ajs124: just started testing that right now
<elvishjerricco>
pie_: efibootmgr, but I have no idea how to use it
<samueldr>
nixnothing: basically, keep it to whatever was set on generation
<pie_>
im 99% sure im looking at the right esp
<samueldr>
though yeah, it's *mostly* databases, but at a moment's notice anything stateful could end up being affected by stateVersion
<samueldr>
like e.g. systemd-timesyncd
<clever>
pie_: run efibootmgr -v
<ajs124>
pie_: how many ESPs do you have?
<clever>
pie_: that will show the partition uuid of whatever the bios is being told to boot
<clever>
Boot0004* UEFI OS HD(1,GPT,27c99b08-455d-4dfe-a44f-6150cbc09ef8,0x800,0x100000)/File(\EFI\BOOT\BOOTX64.EFI)..BO
<pie_>
ajs124: just this one on the external drive
<clever>
pie_: you didnt boot via efi, so you cant tell the bios to use the right file on the right ESP
<pie_>
clever: i was only testing this in a vm anyway :V
<adamt>
nixnothing: If you want to know what actually uses it, you can just grep for "stateVersion" in the nixpkgs repository.
<pie_>
but good to know
<elvishjerricco>
clever: He is getting to grub though, with device = "nodev"
<elvishjerricco>
so efi is working
<gchristensen>
niso: ah yes
<elvishjerricco>
and efi vars are not required
<ajs124>
clever: pie_: it should still detect the well-known™-path, though, shouldn't it?
<clever>
elvishjerricco: it could be an old legacy install he forgot about, or a second ESP
<colemickens>
Is NetworkManager-wait-online.service required for a proper boot with NM? It's supposedly taking the huge vast majority of my boot time according to `systemd-analyze blame`?
<nixnothing>
I just wonder if your tied to that state version though when deploying a new system, or if you can leave blank to use most recent
* niso
thinks the SDF notation of nix should be in the manual. Starting with that would have made things way easier.
<elvishjerricco>
right, and if he followed the blog post he says he did, he does have efiInstallAsRemovable = true;
<ajs124>
clever: " --efi-directory=/efi --no-nvram --removable " was in pie_'s command, so I guess that should be fine?
<clever>
ajs124: ahh, yeah it is
<elvishjerricco>
I don't think the efivars being unsupported thing matters
<adamt>
nixnothing: You can't leave it blank, but it sounds like you're worrying more about it than is strictly necessary :-)
<gchristensen>
yes, just let it be the default and leave it alone
<samueldr>
nixnothing: think of it as "my system is a XX.YY vintage", like a fine old wine
psyanticy has joined #nixos
* gchristensen
wonders if he should keep it as the current version, since his system maintains no state
<samueldr>
you'll have to set it to something, starting soon
<pie_>
elvishjerricco: what was your cryptdevice line again?
<gchristensen>
good
pamplemo1sse has joined #nixos
<{^_^}>
[nixpkgs] @nicknovitski opened pull request #65357 → hyperkit: init at 0.20190201 → https://git.io/fjD11
* samueldr
wonders if the PR was merged
<samueldr>
though, gchristensen, it's likely it wouldn't change much in your case
<pie_>
elvishjerricco: from the binary
<adamt>
It's actually kinda interesting to see how few modules actually rely on it -- 11 modules if I count correctly.
<samueldr>
wondering if the migration thing for timesyncd would have an issue though
<elvishjerricco>
pie_: What binary?
<samueldr>
if you declare yourself as 19.03, but don't have a state to migrate
iqubic has joined #nixos
<adamt>
samueldr: What's changing with timesyncd?
<pie_>
elvishjerricco: the efi binary
<elvishjerricco>
I always though `stateVersion` was kind of bad form. Anything that needs state that may need to migrate should probably manage its own state variable with a file on the system, rather than with this setting that users often mess up
<iqubic>
So, anyone know what the right URI is to use for my new HP printer?
<pie_>
elvishjerricco: i mean when i asked you to take the tail
<elvishjerricco>
pie_: `cryptomount -u 7e7a752a530c431f91625cf801a12910` and `(cryptouuid/7e7a752a530c431f91625cf801a12910)/root/boot@/grub`
<iqubic>
Why is this such a pain in the neck?
o1lo01ol1o has quit [Remote host closed the connection]
<adamt>
elvishjerricco: The postgres module choses a specific postgres pkg based on stateVersion. Having to look in the file system to figure it out would be impure, unless you could live with sticking all possible versions in the store, and just pick one on startup :P
<gchristensen>
iqubic: I can't help you find it, but ipp://THE-IP-OF-YOUR-PRINTER:631 is usually sufficient.
wfranzini has quit [Remote host closed the connection]
Tucky has quit [Quit: WeeChat 2.5]
<elvishjerricco>
pie_: FWIW, something weird just happened. I told you I got unstable working. But then I started a brand new VM, did the exact same steps, and got your error again.
<adamt>
samueldr: Ah, I see. Thanks.
<elvishjerricco>
Whereas the time I succeeded was on the VM that I got working with 18.09, and I had reformatted both partitions.
<elvishjerricco>
So 18.09 worked for me, and wiping that hard drive and installing unstable worked, but installing unstable on a new vm didn't
<elvishjerricco>
wtf
<pie_>
:D
<iqubic>
How can I go about testing this printer? Because I think I just successfully added it.
alp has joined #nixos
<pie_>
iqubic: cups has a "print test page" button somewhere. or you can just print something.
<iqubic>
Where is that print test page option?
o1lo01ol1o has joined #nixos
<iqubic>
I'm using the local webserver as my GUI.
<iqubic>
In other words, I pointed my browser to localhost:631
<gchristensen>
maybe try searchig the internet for "cups print test page"
<adamt>
iqubic: There's also the 'lp' cli tool that can print.
<ajs124>
iqubic: admin -> manage printers -> click on the name of the printer -> select "print test page" from drop down menu
shomodj has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<iqubic>
And now we wait.
<{^_^}>
[nixpkgs] @josephtheengineer opened pull request #65358 → kitty: fix wayland support (EGL: Library not found error) on swaywm → https://git.io/fjD1b
<pie_>
elvishjerricco: i just tried the weird hack of changing the search command with a hex editor to use cryptomount (barely fit) and it failed on boot saying couldt find cryptomount, so grub doesnt even seem to get compiled with luks (?)
<iqubic>
Well, that worked without any hitches.
<elvishjerricco>
pie_: Maybe it has to do with the ZFS version. 18.09 still uses ZFS 0.7, and creating a new pool with 0.8 will make a pool with different feature flags enabled. When I did the wipe + upgrade to unstable, I used the same pool and just deleted the dataset
<clever>
pie_: when you run grub-install, it will figure out which modules it needs at early boot, and merge them into the .efi file
<clever>
pie_: so it wont include luks support if it doesnt think /boot is on luks
<pie_>
clever: makes sense
<clever>
pie_: once it can open /boot, it will load more modules from files
<elvishjerricco>
I have a feeling something about a pool made with ZFS 0.8 causes grub to fail to realize the pool is on LUKS, unlike a pool made with 0.7
<pie_>
elvishjerricco: does that fit with it trying to boot off the zfs partition though?
<clever>
-rwxr-xr-x 1 root root 15776 Oct 29 2017 /boot/grub/x86_64-efi/cryptodisk.mod
<clever>
pie_: such as this
<pie_>
elvishjerricco: ah
<elvishjerricco>
So this is potentially a bug in ZFS, not grub or NixOS!
<pie_>
elvishjerricco: i suppose you could try downgrading to 0.7 on a fresher nixos
<elvishjerricco>
not available in unstable :P
<gchristensen>
zfs can't downgrade
<clever>
evhan: it sounds like maybe its a bug with grub not understanding the output of `zpool status` or such?
<elvishjerricco>
gchristensen: I think he just means downgrading the version you use to create the pool
<clever>
gchristensen: it can, in some limited situations, `man zpool-features`
<pie_>
by downgrade i meant make one with an older zfs
<clever>
gchristensen: you have to know which features 0.7 supports, and check if they can be disabled, or just not enable them when creating/upgrading the pool
<pie_>
elvishjerricco: well you could create the zfs on an old nixos
<gchristensen>
ah
<elvishjerricco>
So to test my hypothesis, I will format the disk with 18.09, then install with unstable
<pie_>
elvishjerricco: and try the installation process on a new one
<elvishjerricco>
exactly
<pie_>
right
<clever>
gchristensen: each feature has different rules, and the man page explains them
<samueldr>
elvishjerricco: grub was updated on unstable, were you wiping the ESP on updates?
<gchristensen>
gotcha
troydm has quit [Ping timeout: 248 seconds]
<elvishjerricco>
samueldr: Yes
* pie_
goes to scream into his pillow for a bit
<samueldr>
hmm, with the ESP wiped then my clue is gone
<elvishjerricco>
samueldr: I think the theory that ZFS 0.8 has a grub regression is likely
<samueldr>
though tbf I haven't followed the discussion closely and shouldn't jump in really :/
<samueldr>
it wouldn't surprise me much, there seems to have been other regressions
<samueldr>
or at least one other :/
<clever>
pie_: try doing grub-install -vvv, with the rest of the args you previously pasted, under a `nix-enter --root /mnt/`
<elvishjerricco>
how can I easily get a file off a virtualbox machine? I don't want to retype this configuration.nix for the fourth time lol
troydm has joined #nixos
<clever>
elvishjerricco: enable ssh?
<samueldr>
elvishjerricco: scp it to your main machine from the VM
<clever>
elvishjerricco: install the gist command?
<samueldr>
(so no need to to enable ssh)
<elvishjerricco>
samueldr: Is it wise that nixos-usntable is already using zfs 0.8 as the stable version?
<samueldr>
couldn't say, I'm not using zfs and know next to nothing
<samueldr>
but if it's deemed stable by the zfs project, I figure it's okay
<gchristensen>
they're typically pretty careful
<iqubic>
Alright... So this printer works.
<iqubic>
This is the first time I've ever been able to print anything at all from a Linux Box.
<iqubic>
Now, time to test the scanner. I'm not sure if this will work or not. I also don't care if it works, but I want to test it.
<gchristensen>
w00t nice, iqubic
<gchristensen>
never have I ever had a scanner work under linux
srid6390803 has quit [Ping timeout: 246 seconds]
adamantium has quit [Remote host closed the connection]
<pie_>
gchristensen: i have had some luck but either i found some driver bugs that almost invisibly affected scan quality or the scanner sucks, idk
<iqubic>
gchristensen: Neither have I. I haven't yet tested this scanner.
<iqubic>
I can use the printer as a means of output, but I'm not yet sure if I can use it as a means of input, if you understand what I'm saying.
<gchristensen>
I do
Soo_Slow has quit [Remote host closed the connection]
Thra11 has quit [Ping timeout: 272 seconds]
gratto has quit [Ping timeout: 245 seconds]
<iqubic>
So I was just trying to connect my bluetooth headset to my computer when I discovered that my new HP Envy 5055 sends out a bluetooth signal that my computer can detect.
Guest47894 has quit [Quit: WeeChat 2.4]
<gchristensen>
:o
<iqubic>
Yeah, it shows up when I run a scan with bluetoothctl.
<adamt>
iqubic: Careful now, you're stretching your luck trying to get both printing, scanning and bluetooth working on the same evening. :P
<iqubic>
Adamt: I have no interest of touch that bluetooth signal.
<inquisitiv3>
Is there any way to connect to a Wireguard server using *.conf files instead of configuring it using the options `networking.wireguard.interfaces.*`? I'm trying to connect to my VPN provider.
<inquisitiv3>
Disclaimer: I have about almost zero knowlege about computer networks.
<iqubic>
I'm treating this a giant jenga tower. Sure everything is standing up and working. But hitting the wrong button, or sending the wrong command might topple the whole system.
<gchristensen>
yeah. erasing my root partition every boot means I can't use my bluetooth headphone anymore.
srid6390803 has joined #nixos
<iqubic>
So, how do I change the default settings of my printer/
<{^_^}>
[nixpkgs] @bjornfor closed pull request #31990 → [WIP] nixos/syncthing: change group from nogroup to syncthing → https://git.io/vFxyq
domogled has quit [Quit: domogled]
<iqubic>
Nevermind... I found it.
<pie_>
iqubic: good job
o1lo01ol1o has quit [Remote host closed the connection]
<pie_>
elvishjerricco: do any flags need to be set on the partitions?
erasmas has quit [Quit: leaving]
<elvishjerricco>
pie_: Not for EFI that I know of
eagle2com has joined #nixos
<iqubic>
And now I'm getting issues where the bottom bit of a page is getting cut off.
liberiga has quit [Ping timeout: 260 seconds]
<iqubic>
When printing that is.
<elvishjerricco>
pie_: Yea, that did the trick. You just have to create the ZFS pool with 0.7
<pie_>
elvishjerricco: jfc
<elvishjerricco>
lol
<ajs124>
inquisitiv3: the .conf file probably contains about 4-5 settings, which can be easily translate into networking.wireguard.interfaces config. the other option would be creating your own systemd service, which just calls wg setconf.
<ajs124>
the settings have the same names and all that (or should, at least)
<elvishjerricco>
pie_: I'm guessing you were enabling zfs unstable on those older NixOS attempts?
webster23 has quit [Ping timeout: 264 seconds]
<elvishjerricco>
Or otherwise avoiding recreating the pool somehow?
<pie_>
elvishjerricco: well so there was something to the hunch of external state given that the various nixpkgs versisns didnt change a damn thing
<pie_>
yeah i wasnt doing the pool
<pie_>
elvishjerricco: now i feel bad about not figuring it out too :P
<elvishjerricco>
Haha you just weren't trying from scratch enough! :P
<pie_>
yeahhh
<elvishjerricco>
Lemme know if it works
<pie_>
trying another workaround right now but i dont think it will change anything
<pie_>
question is if grub is actually broken or just the generation
<pie_>
i found the file where this boot config thingy is stored but im not sure changing it will actually work
<elvishjerricco>
I wager that if I `zpool upgrade` the pool now, grub will probably break on the next boot loader reinstall.
<pie_>
well you seem to be a lot faster than me, so give it a shot :P
<pie_>
and stick something on several github issue trackers ;_;
<pie_>
guess im not used to thinkinf about versioned filesystems either :P
<iqubic>
If I update my printer, and that causes it to reboot, will the ip address change?
<elvishjerricco>
pie_: Odd. `zpool upgrade && nixos-rebuild --install-grub switch && reboot` worked just fine.
<adamt>
iqubic: That's impossible to say, since it depends on how the printer and your network is configured.
<iqubic>
I see.
<elvishjerricco>
I don't in general rely on machines getting their IPs back when they reconnect, but it does seem to happen almost every time on my router
<adamt>
iqubic: Some printers support pointing to using various auto-discovery mechanisms, so that the actual IP doesn't matter.
<pie_>
elvishjerricco: did it actually rebuild grub tho
<pie_>
oh
<elvishjerricco>
pie_: It did.
<pie_>
nvm noticed the flag
<elvishjerricco>
I guess there's still just some lingering difference between an upgraded pool and a straight up new one
<pie_>
elvishjerricco: i think some routers generalyl keep track of mac address assignments and leases
<pie_>
some/most
<elvishjerricco>
I'm gonna ask about the on disk format of a upgrade in #zfsonlinux
<andrewrk>
I'm in the installation cd at root prompt, I don't see anything but the USB stick in parted --list
<pie_>
ok my thing didnt work
eagle2com has quit [Ping timeout: 258 seconds]
owickstrom has quit [Quit: Connection closed for inactivity]
<andrewrk>
What can I try doing to get Linux to see my hard drive?
<mightybyte>
I'm trying to get a nix cache set up on S3. I did a nix copy of my package and it seemed to work, but when I try to build it from another machine I get the following error:
<elvishjerricco>
mightybyte: Does that machine have your aws credentials set up?
<mightybyte>
That item just doesn't exist in that cache but it doesn't seem like it should halt the operation. Any idea how to fix that?
<mightybyte>
elvishjerricco: It doesn't need credentials. The S3 cache is publicly readable.
<elvishjerricco>
hm
<mightybyte>
That hash genuinely doesn't exist.
pamplemo1sse has quit [Ping timeout: 245 seconds]
<iqubic>
So, the printer alignment failed because I can't tell the difference between right and left. The directions said to take the printed page and put it on the front right of the scanner glass. I did front left.
<THFKA4>
andrewrk: kind of a lame tip, but restarting made everything show up for me
<iqubic>
I feel really stupid.
<andrewrk>
THFKA4, I'll try a different .iso
<andrewrk>
You had problems seeing the hdd on a Dell XPS?
<THFKA4>
you can also generate an installer from an existing nixos
<THFKA4>
no, it was some random laptop from 2011
<andrewrk>
Yeah that's what this one is
<THFKA4>
oh, and i had an issue getting to a prompt at all, using the "copy to RAM" option saved me there
<warbo>
hi, I'm running the latest Firefox via buildFHSUserEnv, but I can't get its KeePassXC-Browser plugin (installed via FF, not Nix) to connect to KeePassXC (the Nixpkgs version, overridden to the latest src)
<warbo>
would the fhsuserenv be preventing this somehow?
<warbo>
it seems to use a socket in /run/user/1000 and a "proxy" program in /run/current-system/sw/bin, but both of those paths seem to be accessible from the chroot
<pie_>
elvishjerricco: too bad you cant declaratively roll back your entire disk :P
<pie_>
or jump around rather
nDuff has joined #nixos
<elvishjerricco>
Lol zfs actually added a weak version of that in 0.8 called checkpoint. You can only have one checkpoint on a pool, but it will revert EVERYTHING, included deleted data sets, pool upgrades, and property changes
<andrewrk>
Damn, still no hdd listed in parted with a different iso
<clever>
andrewrk: what does lsblk report?
<andrewrk>
loop0, sda (USB drive)
<andrewrk>
This is the newest Dell XPS
<andrewrk>
I tried with Linux kernel 5.1.19 and 5.2
o1lo01ol1o has joined #nixos
andersk has quit [Remote host closed the connection]
<adamt>
andrewrk: What type of disk does it have? Anything really exotic?
<adamt>
andrewrk: Any chance you have some weird raid-controller enabled in the firmware?
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fjDMQ
m0rphism has quit [Ping timeout: 244 seconds]
Chiliparrot has quit [Read error: Connection reset by peer]
Chiliparrot has joined #nixos
<mightybyte>
When I do "nix copy --to s3://blah-s3-bucket $(nix-build)", it seems that some dependencies are not copied to the cache. Any idea how to get the full set of transitive dependencies cached?
adamt has quit [Quit: WeeChat 2.4]
<Miyu-chan>
What does the flakes RFC mean by "hermetic"? This is not explained in the RFC, nor is this pointed out by any of the PR comments.
wildtrees has quit [Quit: Leaving]
mexisme has joined #nixos
<simpson>
Miyu-chan: Huh, I was expecting it to be defined in the Nix thesis, but it isn't. I presume that it means roughly "isolated", but with some sort of confinement property.
<craige>
Does anyone know how to trigger an update to planet.nixos.org? The sources have been updated but not reflected in the live site.
usr has quit [Read error: Connection reset by peer]
<Miyu-chan>
Right. Thanks. Just my first time seeing that word.
<simpson>
I've seen it before, but it's not common jargon and apparently it's not Nix-specific jargon either.
<simpson>
It roughly means "sealed in both directions". A hermit is one who seals themselves away from the world; a hermetically-sealed subsystem is prohibited from interacting with the rest of the system, except through specified APIs.
<mightybyte>
How can I find why something isn't cached? I'm seeing something get built that I know is in the cache.
warbo has left #nixos ["ERC (IRC client for Emacs 26.1)"]
___laika has joined #nixos
<___laika>
any reasonml/ocaml devs around? i want to prototype a (native, not web) project in reason, and ideally id like to use nix, but it looks like esy is the tool of choice for those languages. esy doesnt seem to be in nixpkgs as far as i can tell, so i was curious if anyone has tried a nix based reason/ocaml workflow and how they felt about it