<stepcut>
whenever I try to use ssh-ng:// substituters I get a lot of "error: unexpected end-of-file". But it is not a hard fail -- it is sometimes able to successfully copy files if I ran the command enough times
clacke_movim has left #nixos [#nixos]
clacke_movim has joined #nixos
<worldofpeace>
jtojnar: I actually didn't get a chance to review properly. I should have a chance now though.
jluttine has joined #nixos
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
polman has quit [Ping timeout: 248 seconds]
polman has joined #nixos
<{^_^}>
[nixpkgs] @jtojnar opened pull request #63493 → meson: enable auto_features by default → https://git.io/fjVYg
<monokrome>
what would we do without samueldr y'all
<samueldr>
you know for at least 8h per day, when I sleep :)
<monokrome>
ya but "wait for samueldr" isn't a good answer =P
<monokrome>
omg {^_^} is the best nickname
<monokrome>
I guess rkhunter would maybe be weird / need custom changes for NixOS
<jtojnar>
worldofpeace: while at it, I added plain buildtype, hopefully the world will not burn
troydm has quit [Quit: What is Hope? That all of your wishes and all of your dreams come true? To turn back time because things were not supposed to happen like that (C) Rau Le Creuset]
troydm has joined #nixos
<{^_^}>
[nixpkgs] @jbaum98 opened pull request #63494 → WIP: framac: 18 -> 19 and update why3 → https://git.io/fjVY1
<samueldr>
isn't it more that rootkits would be weird or need custom change for NixOS (mostly joking)
sindrip has quit [Ping timeout: 256 seconds]
<monokrome>
lol - but also not if they are part a package you install or an overlay
<worldofpeace>
jtojnar: for the auto_features it will probably just be us disabling a lot of features explicitly. so maybe plain buildtype would be the more flamable change
<worldofpeace>
But it's just like change I'd like to see here so I still would like to try
<jtojnar>
we could also reuse thr gnome branch
orivej has quit [Ping timeout: 258 seconds]
<worldofpeace>
👍️ (i forgot)
ottidmes has quit [Quit: WeeChat 2.4]
<jtojnar>
too bad GH does not support renaming branches
drakonis has joined #nixos
selfsymmetric-mu has joined #nixos
jasongrossman has joined #nixos
clacke_movim has left #nixos [#nixos]
xkapastel has quit [Quit: Connection closed for inactivity]
georges has joined #nixos
justanotheruser has joined #nixos
dbmikus has quit [Ping timeout: 248 seconds]
<monokrome>
go
<monokrome>
oops
<monokrome>
sorry type
<monokrome>
tfw you typo the word "typo"
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jgt1 has joined #nixos
jb0nd38372 has joined #nixos
freeman42y has quit [Ping timeout: 252 seconds]
<jb0nd38372>
Hello all
jgt1 has quit [Ping timeout: 252 seconds]
Synthetica has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @tadeokondrak opened pull request #63498 → cum: init at 0.9.1 → https://git.io/fjVOE
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
ym555 has quit [Quit: leaving...]
vk3wtf has joined #nixos
stepcut has quit [Remote host closed the connection]
stepcut has joined #nixos
psque has joined #nixos
<psque>
If I use nix to build a project, the output ends up in a `result` and a symlink is created which prevents this from being garbage collected. Is there any convenient way of specifing that all of the dependencies used in that build process also shouldn't be garbage collected?
<ryantm>
psque: The ones used during build but not used in the output?
<psque>
ryantm: Correct. E.g. the purescript compiler if it's a purescript project
stepcut has quit [Ping timeout: 248 seconds]
<psque>
ryantm: And the build output is the js bundle
<ryantm>
psque: The reason I ask is because the dependencies of the result output won't be garbage collected.
<psque>
ryantm: Right. But as I understand it, in cases like the one I described where the output has no runtime dependencies (because it's just files), all the build tooling gets GCed
<psque>
Oh, interesting. I will look around there and see what they do
<psque>
Thanks
<ryantm>
psque: If you save a gc root of the .drv file, it will ensure the dependencies are not garbage collected. You can find the .drv path with drvpath `pkgs.hello.drvpath`
<ryantm>
psque: You can make a derivation like writeTextFile { name = "deps"; text = ''${pkgs.hello.drvPath}'';} and when you build it, it will make a gc root that protects the dependencies of hello.
mexisme_ has quit [Ping timeout: 245 seconds]
psque has quit [Ping timeout: 244 seconds]
reinhardt has joined #nixos
alex`` has quit [Quit: WeeChat 2.5]
dump has joined #nixos
dump is now known as Guest41369
<Guest41369>
Hi, how should I get nixops to rebuild the machine? Do I have to destroy it first? When I change the configuration files, nixops deploy -d trivial doesn't change the running machine.
jgt1 has joined #nixos
drakonis has quit [Quit: WeeChat 2.5]
nD5Xjz has quit [Ping timeout: 246 seconds]
nD5Xjz has joined #nixos
jgt1 has quit [Ping timeout: 252 seconds]
mexisme_ has joined #nixos
Rusty1 has joined #nixos
hamishmack has joined #nixos
<ivan>
Guest41369: does nixops tell you anything at the end of your deploy?
orivej has joined #nixos
adamantium has joined #nixos
<Guest41369>
Works now. Sorry
Guest41369 has quit [Quit: Page closed]
dbmikus has quit [Ping timeout: 252 seconds]
clacke_movim has joined #nixos
domogled has joined #nixos
MrCCoffee has quit [Quit: Lost terminal]
ajirx has joined #nixos
Rusty1 has quit [Quit: Konversation terminated!]
mexisme_ has quit [Ping timeout: 268 seconds]
dbmikus has joined #nixos
orivej has quit [Ping timeout: 258 seconds]
dansho has joined #nixos
o1lo01ol1o has joined #nixos
wfranzini has quit [Remote host closed the connection]
Netsu has quit [Ping timeout: 256 seconds]
<clacke_movim>
jtojnar: Hm, what would that entail? Announcing the rename to other services?
o1lo01ol1o has quit [Ping timeout: 245 seconds]
dbmikus has quit [Ping timeout: 252 seconds]
mexisme_ has joined #nixos
mexisme_ has quit [Ping timeout: 245 seconds]
jgt1 has joined #nixos
<tdeo>
should python applications be fetched from github or pypi? it's my understanding that libraries should be from pypi and applications should be from the source, no?
dansho has quit [Remote host closed the connection]
dansho has joined #nixos
ericsagnes has joined #nixos
palo1 has joined #nixos
zupo has joined #nixos
palo has quit [Ping timeout: 245 seconds]
palo1 is now known as palo
Profpatsch has joined #nixos
alp has quit [Ping timeout: 258 seconds]
<dminuoso>
tdeo: I'd pick everything from source.
<dminuoso>
It's tons easier to fix mistakes and hardwired assumptions (and python libraries seem to be full of them)
Mateon2 has joined #nixos
Mateon1 has quit [Ping timeout: 258 seconds]
Mateon2 is now known as Mateon1
jgt1 has joined #nixos
jgt1 is now known as jgt
polman has quit [Ping timeout: 248 seconds]
polman has joined #nixos
zupo has quit [Read error: Connection reset by peer]
zupo has joined #nixos
mexisme_ has joined #nixos
_e has joined #nixos
_e has quit [Client Quit]
Yakulu has quit [Quit: Gateway shutdown]
_e has joined #nixos
cfricke has quit [Ping timeout: 246 seconds]
drewr has quit [Ping timeout: 248 seconds]
Yakulu has joined #nixos
Yakulu has left #nixos [#nixos]
Yakulu has joined #nixos
qqlq has quit [Ping timeout: 268 seconds]
qqlq_ has joined #nixos
drewr has joined #nixos
rsa has joined #nixos
wfranzini has joined #nixos
<ambro718>
Why is nix-build for single user install putting build directories into /run/user/<pid>? That's typically a tmpfs and will run out of space quickly.
cfricke has joined #nixos
<tilpner>
ambro718: I'm just guessing, but it could be that without the higher-privileged daemon running it can't properly protect the build directory from other users
<tilpner>
Which doesn't really make sense, so it's probably something else
Anton-Latukha has joined #nixos
mexisme_ has quit [Ping timeout: 245 seconds]
kvaster has joined #nixos
<{^_^}>
[nixpkgs] @BadDecisionsAlex opened pull request #63502 → python3Packages.django-crispy-forms: init at 1.7.2 → https://git.io/fjVsB
<{^_^}>
[nixpkgs] @dmvianna opened pull request #63503 → pythonPackages.pycategories: init at 1.1.0 → https://git.io/fjVs0
asheshambasta_m has joined #nixos
kvaster has quit [Ping timeout: 246 seconds]
ambro718 has quit [Quit: Konversation terminated!]
qqlq_ has quit [Read error: Connection reset by peer]
<clever>
tilpner: i think it just always obeys $TMPDIR, and users have a different value then services
<{^_^}>
[nixpkgs] @flokli merged pull request #63037 → google-cloud-sdk: kubeconfig: don't store absolute path to gcloud binary → https://git.io/fj2eu
<gchristensen>
andi- ^
mexisme_ has joined #nixos
hellrazor has quit [Quit: WeeChat 2.5]
hellrazor has joined #nixos
<mpickering>
Do so libraries get stripped by default during installation?>
Izorkin_ has joined #nixos
pie__ has quit [Ping timeout: 252 seconds]
Izorkin_ has quit [Client Quit]
<karetsu>
if I wanted to combine all of the attributes of a set of .nix files in a directory what would my default.nix look like?
Izorkin_ has joined #nixos
fusion809 has joined #nixos
lassulus has joined #nixos
Izorkin has quit [Ping timeout: 245 seconds]
Izorkin_ is now known as Izorkin
<mpickering>
karetsu: Combine them in what sense?
<mpickering>
you import them all into one file
<mpickering>
and then it depends what you want to do
<mpickering>
for example, all-packages.nix
thc202 has joined #nixos
<karetsu>
suppose I have two files keys.nix and apps.nix and they contain a lot of the same attributes e.g. keys has a foo.bar and apps has a foo.baz declaration, I want to 'compile' them in such a way that I end up using all of those things
<andi->
gchristensen: grr I both hate and love it :P
<gchristensen>
andi-: which? :)
<gchristensen>
I meant the channel bump
<karetsu>
mpickering: i.e. here - https://gitlab.com/karetsu/nix-overlays/tree/master/etc/i3 in my home.nix I want to just import ./etc/i3/common.nix which then deals with the combining of all of the options across all the other files
<andi->
Ahh, I the gce sdk
ajirx has quit [Remote host closed the connection]
<andi->
But it probably isn't what I did fear.. I thought flokli might have packaged all the cpp sdk deps but he didnt
<karetsu>
I don't know what that common.nix should look like because its definitely not doing all of it
<gchristensen>
that does not explain why polybar is trying to *call* systemctl
zupo_ has joined #nixos
<karetsu>
its not, its polybar_start which is trying to
<karetsu>
which is the derivation for the service?
<gchristensen>
right, it doesn't explain that
<gchristensen>
why does the script to execute polybar need to call systemctl?
<karetsu>
are you assuming that I wrote that package?
<gchristensen>
I think I'm asking the question which needs to be asked :P
<elgoosy>
clever: wow, dont get it. I've added nix-channel --add https://nixos.org/channels/nixos-19.03 nixos. Why does it show 18.09? Is it because i had 18.03 and you cant upgrading hopping 2 versions?
pie__ has joined #nixos
<clever>
elgoosy: did you run --add with or without sudo?
<clever>
elgoosy: compare `nix-channel --list` and `sudo nix-channel --list`
<{^_^}>
[nix] @backuitist opened pull request #2958 → Binary cache - improve error message when trying to copy partial closure → https://git.io/fjVnk
<karetsu>
gchristensen: I have almost 0 clue why it does it that way, I think its probably to avoid having a default launch.sh?
zupo has quit [Ping timeout: 248 seconds]
<{^_^}>
[nix] @CSVdB opened pull request #2959 → Check whether flag registry is compeletely used → https://git.io/fjVnI
<clever>
karetsu: what command are you running, to cause the error message?
<elgoosy>
clever: oh, thanks. What is the important difference i should know about channel on root and user?
weedloser has joined #nixos
<clever>
elgoosy: nixos-rebuild always uses the channels on root, and you typically only need the user channels if you lack root or want to avoid sudo more
<clever>
karetsu: that line is why its trying to run systemctl from the service
<clever>
try using /run/current-system/sw/bin/systemctl instead of normal systemctl
<clever>
karetsu: but note, that may just lead to infinite recursion, starting polybar restarts polybar, which restarts polybar....
Makaveli7 has quit [Quit: WeeChat 2.5]
<Unode>
hi all, if I want to create a systemd service in configuration.nix that should be launched at boot once, do I still need to configure a timer for it?
<clever>
karetsu: i would delete the line i linked to
<karetsu>
script is required, I'll just make it my own launch.sh
<clever>
Unode: nope, timers are only needed for scheduling based on time
<Unode>
clever: do I need to set some enable flag or something ? I don't even see the service with 'systemctl list-units'
<clever>
karetsu: then you probably want start, not restart
<karetsu>
clever: even with script set to anything else it still fails, I'll just not use systemctl for it
<clever>
karetsu: you want to replace systemctl with /run/current-system/sw/bin/systemctl
sigmundv has joined #nixos
<karetsu>
I've done that, but the problem comes from the fact that polybar.service is in /nix/store and its that version which is failing rather than my home-manager config
<karetsu>
:(
elgoosy has quit [Remote host closed the connection]
Neo-- has joined #nixos
<clever>
karetsu: you had systemctl in 2 nix files on your repo
pie__ has quit [Read error: Connection reset by peer]
dermetfan has joined #nixos
dermetfan has quit [Client Quit]
<balsoft[m]>
Hello! I've noticed some nixos users (me included) are having some issues with curl
<balsoft[m]>
warning: unable to download 'https://cache.nixos.org/nar/<hash>.nar.xz': HTTP error 200 (curl error: Timeout was reached); retrying in 346 ms
<arianvp>
Heya. Anybody any luck with an apple Thunderbolt 3 to thunderbolt 2 converter on NixOS?
<arianvp>
Linux just gives me "Cannot enable. Maybe USB cable is bad?"
<arianvp>
This is on a ThinkPad t490
<arianvp>
I'm trying to connect to a displayport monitor using it
<exarkun>
I see there are a number of terraform providers packaged for nixos ... but installing them doesn't let terraform use them. are they meant to be used by the nixos packaged terraform? or are they for something else and I should let the terraform cli install plugins its own way?
<hyper_ch2>
<-- just an average home computer user
<exarkun>
there is some overlap
<exarkun>
terraform supports about 100x more kinds of things than nixops
<hyper_ch2>
oh well, I don't know too much about that :)
<hyper_ch2>
good luck with figuring out the issues though :)
<exarkun>
(but of course nixops has more support for things you might want to do with nixos)
jgt has joined #nixos
<hyper_ch2>
(my notebook runs nixos, my home server runs nixos, my office server runs nixos, another office server runs nixos, my vpn server runs nixos, my other vpn server runs nixos, and my mail/webserver runs debian)
<exarkun>
it looks like it's working now ... I think the problem was always just changing around the plugin version without re-initializing terraform
<tilpner>
hyper_ch2: How many servers do you guess the average home computer user has?
<hyper_ch2>
5? :)
<tilpner>
I wish
<hyper_ch2>
they have a nas
<clever>
tilpner: i have 5 machines in the house currently
<hyper_ch2>
they have windows... which is so open that it counts as server ;)
<clever>
hyper_ch2: your guess was dead on!
<tilpner>
clever: My point was that you can't go to #nixos and search for average computer users
<clever>
:D
<hyper_ch2>
tilpner: but I am an averager computer user :)
<tilpner>
:/
<exarkun>
the distribution of computer users is probably so lumpy and smeared out that "average" is a meaningless statistic
<samueldr>
(do I want to be a lumpy nixos user?)
<exarkun>
lumps are alright
sb0 has joined #nixos
<hyper_ch2>
tilpner: btw, wb to #zfs :)
<hyper_ch2>
s/wb/welcome/
xkapastel has joined #nixos
<hyper_ch2>
(didn't notice you in #zfs before)
<tilpner>
Oh, I ask stupid questions from time to time :)
<hyper_ch2>
I do that all the time
<jbgi>
Here is my scenario: nix start building large derivation that is not binary-cached by our remote hydra, I stop the build. Then hydra finish its job and the derivation is available remotely. I retry the nix build locally but it appears nix does retry to download from the binary cache... :( How do I force nix to retry binary caches?
<hyper_ch2>
and sometimes I anger devs so much that they perma-ban me
<tilpner>
hyper_ch2: You should try the special device thing, you can remove them from the pool at any time
<abbec>
hey, I am trying to install vim with python3 support using `nix-env`. However, no matter how I use `--arg` or `--argstr` it does not make a difference
orivej has quit [Ping timeout: 245 seconds]
vmandela has quit [Remote host closed the connection]
jtojnar has quit [Ping timeout: 248 seconds]
<abbec>
for example `nix-env -i vim_configurable --argstr features tiny` still gives me the huge version
vmandela has joined #nixos
<abbec>
am I missing/misunderstanding something?
<clever>
abbec: --arg doesnt work like that
<clever>
abbec: its best to use something like an overlay to do things like this
<clever>
abbec: when nix loads a given file, it will check if the top-level value was a function (the root nixpkgs function in this case), and call it with any args it can accept
<clever>
and then, for every attr in the -A foo.bar.baz, it will repeat that again
<clever>
vim_configurable is not a function, so it wont try to call it with any arguments
<clever>
and unused args are silently ignored
lsyoyom has joined #nixos
<abbec>
ah, it is not a function, got it
<clever>
`nix-env -iA nixos.vim_configurable.customize --argstr features tiny` may work, but then you have another downside, youll forget the magic incantation, and 6 months down the road, you wont remember how to get the new tiny vim when upgrading
<clever>
putting it all into an overlay forces you to keep a hard record of what the trick is, and you can just nix-env -iA nixos.tiny-vim any time you want to upgrade it
domogled has joined #nixos
<abbec>
and I put it in ~/.config/nixpkgs/overlays.nix?
sb0 has joined #nixos
<abbec>
this is on non-nixos
<clever>
abbec: i think so, or the overlays directory
<clever>
abbec: ive not used overlays outside of nixos
<eraserhd>
I'm trying to build something on Mac and it's getting an error that it needs 10.12... I thought that updating stdenv to 10.12 was merged already, though?
sheeldotme has joined #nixos
zachk has joined #nixos
drakonis has joined #nixos
jabranham has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/fjVBM
<{^_^}>
[nixpkgs] @peti pushed to master « ghc-8.8.x: bump version to alpha 2 »: https://git.io/fjVBD
<{^_^}>
[nix] @CSVdB opened pull request #2961 → Added tests for the `nix flake` CLI → https://git.io/fjVBy
hyper_ch2 has quit [Quit: Page closed]
<stepcut>
infinisil: nm. I seriously hacked up the nixos-rebuild script and got things working
Soo_Slow has quit [Remote host closed the connection]
drakonis_ has joined #nixos
xelxebar_ has joined #nixos
johanot has quit [Quit: WeeChat 2.4]
xelxebar has quit [Ping timeout: 256 seconds]
<{^_^}>
[nix-mode] @jabranham opened pull request #82 → Silence byte compiler warnings in nix-shell → https://git.io/fjVBd
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
domogled has quit [Quit: domogled]
<infinisil>
stepcut: What did you do? For future log searchers
ambro718 has joined #nixos
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-stack: clean up deep overrides »: https://git.io/fjVRq
<{^_^}>
[nixpkgs] @peti pushed 5 commits to haskell-updates: https://git.io/fjVRO
ddellacosta has joined #nixos
<pbb>
Hi, how do you guys unlock your luks root partition via ssh in initrd? I can unlock it but how can I continue the boot process from the ssh session?
<pbb>
the init script uses bash read to get the password, and then hands it to cryptsetup, so there is no cryptsetup process I can kill
<stepcut>
infinisil: edited the script so that it did not try to use nix-daemon and set the PATH to favor /root/.nix-profile/bin, and maybe some other things
alex`` has quit [Ping timeout: 245 seconds]
alex`` has joined #nixos
jtojnar_ has joined #nixos
jtojnar has quit [Ping timeout: 252 seconds]
jtojnar_ is now known as jtojnar
ThatDocsLady has quit [Remote host closed the connection]
ThatDocsLady has joined #nixos
<{^_^}>
[nix-mode] @matthewbauer pushed 2 commits to master: https://git.io/fjVR4
<clever>
ambro718: nix will obey TMP or TMPDIR i believe
<clever>
pbb: one minute
domogled has joined #nixos
<clever>
pbb: i think you can just `killall cryptsetup`
orivej has quit [Ping timeout: 268 seconds]
sigmundv has joined #nixos
<clever>
pbb: the stage-1-init.sh will treat that as having gotten the wrong pw, and try again, then i think it will notice its already open and keep going?
Thra11 has joined #nixos
dm9 has joined #nixos
<ambro718>
clever: right, but I can't set it in nix.conf?
<clever>
it calls tempName, with the root set to an empty string
<clever>
which is defined up about 15 lines
<clever>
yeah, it can only use $TMPDIR
sheeldotme has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
when using multi-user nix, nix-daemon is ran by systemd, and then it doesnt point into /run/user/1000/
<ambro718>
clever: Any idea how it could end up using /run/user/<pid> when TMPDIR is *not* set?
<ambro718>
clever: I just installed nix via the install script, does that end up with a multi-user install?
<clever>
ambro718: i think only darwin does multi-user, but gchristensen should be able to confirm
<clever>
ambro718: `ps aux | grep nix-daemon` ?
<ambro718>
clever: Nothing.
<clever>
ambro718: echo $TMPDIR ?
<Shoubit>
Sort of related: what's the best way to get a multi-user Nix up and running outside NixOS? Last time I did it I had to download an old dpkg version of Nix and upgrade through several versions until I got to the latest release.
sigmundv has quit [Quit: Leaving]
<ambro718>
clever: TMPDIR is not set. If I run nix-build then, it ends up using /run/user/pid. If I set TMPDIR=/tmp prior, it uses that.
<clever>
ambro718: if TMPDIR is not set, it will use XDG_RUNTIME_DIR, if that is also not set, it will use /tmp
<clever>
ambro718: it will then overwrite TMPDIR, TEMPDIR, TMP and TEMP, with the choice it made
<ambro718>
clever: Indeed XDG_RUNTIME_DIR is set to /run/user/1000
<clever>
ambro718: clearing that too should force it back to /tmp/
<ambro718>
It's set by the fedora default setup, don't want to mess with it.
<clever>
ambro718: and note, you want `unset`, not `export XDG_RUNTIME_DIR=`
<clever>
ambro718: you may need a PR to nix, to add an option
<clever>
ambro718: libstore/build.cc looks like the best place, change the "" to point to the option, and if the option is "", the old behaviour returns
<ambro718>
It just sucks I can't tell people to install nix with the installed and build my software, because it's going to fail when it ends up using the small tmpfs, and I also can't tell them to edit a config file. Only setting TMPDIR in bashrc or whatever would fix it, which will affect other programs.
<clever>
ambro718: related, one of my old machines had a persistant /tmp, but it cleared it on bootup, and one day after an unplanned reboot, it didnt come up right away
<clever>
ambro718: i went over to the console, to discover there was 20gig of garbage in /tmp, and the boot waits until its all cleaned up before continuing, lol
Ariakenom_ has quit [Ping timeout: 258 seconds]
<clever>
ambro718: if you get them to use a binary cache service (either your own, or cachix), then they can just download a pre-built copy
<clever>
ambro718: and there tmp dir wont be as loaded as heavily
<ambro718>
I just set the option to purge /tmp on startup on my system.
<clever>
it will also make things build much faster for them
<ambro718>
Don't really want to maintain a binary cache, the point of using nix is to reduce maintentance effort.
<ambro718>
And it doesn't help when everyone could have different nixpkgs.
<clever>
ambro718: your default.nix would need to specify which nixpkgs to use
filedesless has quit [Ping timeout: 245 seconds]
ddellacosta has quit [Quit: WeeChat 2.2]
<stepcut>
I am trying to use an ssh-ng:// substituter, but the ssh connection is not finding the right set of keys or connection to ssh-agent or something. Is there anyway to get nix-daemon to run the substituter with ssh -v and log the output somewhere?
<clever>
stepcut: nix-daemon is running under systemd, so it likely wont find your agent, you could use /root/.ssh/ssh_config to force which key to use, or just have /root/.ssh/id_rsa lack a passphrase
hmpffff has joined #nixos
<clever>
stepcut: checking some source...
<stepcut>
if `/root/.ssh/ssh_config` forces a key -- that does not help with the passphrase issue does it ?
<clever>
stepcut: aha, modify the systemd service to set `NIX_SSHOPTS = "-v"`
<stepcut>
in my NIX_PATH I have these bits, ssh-auth-sock=/tmp/ssh-1ymFiHoK6B/agent.9176:ssh-config-file=/home/stepcut/nix-seereason/ssh-config, but I have concerns about ssh-auth-sock working correctly because ssh-agent does not like it when the euid and uid don't match
<clever>
stepcut: similarly, you could just systemd.services.nix-daemon.environment.SSH_AUTH_SOCK = "/whatever";
<clever>
stepcut: ssh-auth-sock in NIX_PATH only works for pkgs.fetchGitPrivate, and wont help with ssh-ng://
<stepcut>
k
<clever>
stepcut: i also have a hack with nc to fix the uid problem you just mentioned
<clever>
stepcut: this is another option, it will just create a directory suitable for http:// cache operations, throw an http-server of your choice at it, and your up-and-running
<clever>
stepcut: you may also want to mix nix sign-paths in also
ubert has quit [Remote host closed the connection]
<stepcut>
I did not see a way to do authentication for http:// caches
<clever>
stepcut: ah, yeah, you would need to use firewall or vpn stuff to restrict access
Soo_Slow has joined #nixos
v88m has joined #nixos
drakonis has joined #nixos
<clever>
stepcut: you may also be able to adjust ssh-serve some....
<clever>
stepcut: this is forcing ssh to run nix-store --serve or nix-daemon --stdio depending on version
gratto has quit [Read error: Connection reset by peer]
<clever>
stepcut: if you modify it, to use nix-store --serve --store local?root=/second/store/
pointfourone has joined #nixos
<clever>
stepcut: then it can only share what is in /second/store/nix/store
<clever>
and youll want nix copy --to local?root=/second/store/ /nix/store/foo to populate it with a subset
gratto has joined #nixos
<samueldr>
hmm, I don't think there's a way to have patchShebangsAuto skip a directory
<stepcut>
not sure that helps me. I do want to share some private stuff, but only with people who have the ssh key and know the passphrase
drakonis_ has quit [Ping timeout: 258 seconds]
pointfourone has quit [Remote host closed the connection]
d10n-work has joined #nixos
vmandela has left #nixos ["Leaving"]
pointfourone has joined #nixos
<clever>
stepcut: this would let you restrict what they can access, when they do have the ssh key
<samueldr>
I'd like to have my cake and eat it too :/ scripts are being patched while they are runtime artifacts for foreign systems (without nix) in vagrant
jbgi has quit [Ping timeout: 246 seconds]
<stepcut>
clever: well, if they have the key and know the passphrase, then I want them to be able to access everything in /nix/store. They already have login access to the machine anyway. Hence the desire to have a passphrase on the key in the first place
o1lo01ol1o has quit [Remote host closed the connection]
pie__ has quit [Ping timeout: 246 seconds]
<clever>
stepcut: ahh, i was thinking of a 2nd key, that can only read the store, via nix.sshServe
<stepcut>
so, I could have developers create a second key that does not have a passphrase and can only access /nix/store, but they'd still need to be able to access everything in there
<clever>
and then since its read-only, its safer to lack a passphrase
<clever>
yeah
<clever>
stepcut: i think your best option is to just run a hydra with an s3 bucket
<stepcut>
it is safer, but less safe than using the agent
<clever>
stepcut: s3 buckets support auth much better
ThatDocsLady_ has joined #nixos
<stepcut>
I've considered the hydra+s3 route
o1lo01ol1o has joined #nixos
<clever>
stepcut: though you can also use `nix copy --to s3://bucket /nix/store/foo` and skip the hydra half
<clever>
stepcut: and just put s3://bucket into the cache list
<ambro718>
How can a program check if nix is using single or multi user mode?
<clever>
ambro718: nix itself checks by seeing if it has write to /nix/var/nix/db/ and also checking $NIX_REMOTE
<ambro718>
thanks
<clever>
NIX_REMOTE=daemon forces it to use the daemon
<clever>
auto is the default when unset, and that will test write access to figure out daemon vs local
<clever>
older versions lack auto, and will actually fail with permission errors if you dont set NIX_REMOTE
<clever>
and newer nixos wont set it if nix is new enough, causing old nix to break
<clever>
stepcut: i dont think thats used on nixos machines
<clever>
stepcut: you probably want systemd.services.nix-daemon.environment.SSH_AUTH_SOCK = "/whatever";
<clever>
stepcut: and then tell ssh-agent to use a given path, rather then making a dynamic one
<stepcut>
clever: well, however nix-daemon gets started. Though in my particular use case, the developers are actually on ubuntu machines with nix installed on top
<{^_^}>
[nixpkgs] @samueldr opened pull request #63529 → vagrant: Don't patch salt provisioner's shebang → https://git.io/fjV0Q
<vaibhavsagar>
nh2: this was my previous job, but yes
<vaibhavsagar>
JSON is a valid terraform input format, and we just used builtins.toJSON
<vaibhavsagar>
and used Nix for templating
<nh2>
vaibhavsagar: OK, that would have been my next question, hcl or json
dm9 has quit [Quit: WeeChat 2.5]
zachk has joined #nixos
<vaibhavsagar>
the escaping is a bit fiddly, but manageable
<{^_^}>
[nixpkgs] @matthewbauer pushed to master « emacsPackages.irony: add clang for Linux builds »: https://git.io/fjV05
<nh2>
vaibhavsagar: makes sense, I think the limitations of HCL are pretty bad, very hard to create multiple `resource`s from a map of stuff, nix does that way better
<{^_^}>
[nixpkgs] @gnidorah opened pull request #63530 → haskellPackages.hakyll-images: unmark as broken → https://git.io/fjVEC
hyper_ch has quit [Read error: Connection reset by peer]
<vaibhavsagar>
palo: just wondering if you got any ideas from tv :)
<palo>
vaibhavsagar: yeah had the inital seed, and used nix-instantiate, I'm using nix-build now (so you can use <nixpkgs>) but tv was on board from the beginning :D
hyper_ch has joined #nixos
skykanin has joined #nixos
<vaibhavsagar>
haha nice
<vaibhavsagar>
I saw that you use krops so I guessed he was involved :)
<palo>
Right now I'm teasing him, that change krops to configure ssh to use a different configuration than ~/.ssh/config
<palo>
I'm trying to use overlays but no luck so far
<palo>
vaibhavsagar: I'm currently fiddeling on plops :D my krops abstraction :D
* palo
love theses names
<manveru>
what's krops? the link in the readme was dead
<skykanin>
So I'm trying to make an overlay for the openjdk11 package to add openjfx support and I've managed to get the openjfx derivation building. I've also modified the openjdk11.nix file and added openjfx as an argument to the function. In the all-packages.nix file I'm supplying that value which is pointing to the openjdk derivation, but when I try to build openjdk11 I get an error "attribute openjfx missing", which
<skykanin>
I don't understand cause I'm supplying it in the all-packages.nix file
<palo>
the article is quite old now, and nixops has some features which are not named in the vs part
<manveru>
yeah
<manveru>
atm i'm using a thin wrapper around nixos-rebuild instead of nixops anyway...
linarcx has quit [Quit: WeeChat 2.5]
<manveru>
it's a mess though, so i'd love something that helps me structure it
thePirateKing has joined #nixos
trevorriles has joined #nixos
<palo>
manveru: I'm very happy with krops
o1lo01ol1o has quit [Remote host closed the connection]
<manveru>
i'll try it combined with flakes and terranix :D
<manveru>
should be fun
o1lo01ol1o has joined #nixos
<vaibhavsagar>
does terranix have any support for the AWS provider?
<palo>
vaibhavsagar: it just generates JSON for terraform, you can use the full AWS api.
<vaibhavsagar>
ah, I see
<vaibhavsagar>
but then why is there specific support for github or cloudflare?
<palo>
vaibhavsagar: but I did not generate any modules for any provider, because of reasons :D
eraserhd has quit [Quit: WeeChat 2.4]
gratto has quit [Read error: Connection reset by peer]
alp has joined #nixos
eraserhd has joined #nixos
<palo>
vaibhavsagar: just a sec
gratto has joined #nixos
<vaibhavsagar>
fair enough :)
o1lo01ol1o has quit [Ping timeout: 246 seconds]
elibrokeit has joined #nixos
linarcx has joined #nixos
<palo>
vaibhavsagar: ah, ... these modules are not optinionated provider modules, I used them before so I put them in there. But if you look at them you can see, they just set providers.github = { ... }; You can do always the same. :D
<vaibhavsagar>
right, I see
<manveru>
i'll use hetzner and vpsfree anyway, plus gandi (and one day i'll get around to write a provider for nearlyfreespeech)
<palo>
vaibhavsagar: I'm not happy with the provider nameing, I would like to call them like api.cloudflare or terranix.cloudflare or terranix.api.cloudflare for the not opinionated modules.
<manveru>
gchristensen: i'm getting timeouts now as well :(
stranger___ has joined #nixos
<manveru>
this is getting way too common
Jackneill has quit [Quit: Leaving]
<stranger___>
after updating to latest nixos-unstable i'm getting "Detected unsafe path transition /var/log → /var/log/journal during canonicalization of /var/log/journal" any idea?
<vaibhavsagar>
I've been meaning to write a blog post about how `nix-copy-closure --to <config> && nix-env --profile /nix/var/nix/profiles/system --set <config> && /nix/var/nix/profiles/system/bin/switch-to-configuration switch` is good enough for 90% of all deployments
<tilpner>
manveru: Copying derivations to the deploy target and then instantiating those remotely works very well here, but that might be exactly what nixos-rebuild does too
<manveru>
exactly
xelxebar_ has quit [Remote host closed the connection]
xelxebar has joined #nixos
<tilpner>
Well, except when the ssh connection dies mid build
<tilpner>
Only way to recover from that is restarting the daemon IME
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
v88m has quit [Read error: Connection reset by peer]
v88m has joined #nixos
mexisme_ has joined #nixos
<psque>
I'm trying to prevent my build dependencies from getting garbage collected. Last night, someone suggested I could do something like `writeTextFile { name = "deps"; text = ''${foo.drvPath}'';}` and this would make a GC root to protect the dependencies. However, when I build a derivation like the above, it looks like it downloads a bunch of stuff beyond what was required to originally build the package `foo`. What's going on?
laalf has joined #nixos
domogled has joined #nixos
Czen has joined #nixos
NinjaTrappeur has quit [Quit: WeeChat 2.4]
wfranzini has quit [Remote host closed the connection]
wfranzini has joined #nixos
mariel__ has quit [Ping timeout: 272 seconds]
trevorriles has joined #nixos
<djahandarie>
Hey folks, I'm trying to install-multi-user on a machine and I get: building '/nix/store/2c4l83wkdfrj7ra9k07l0aqvym1d5z20-user-environment.drv'... error: opening directory '/nix/store/hbhdjn5ik3byg642d1m11k3k3s0kn3py-nix-2.2.2': Permission denied
<djahandarie>
The command does seem to be getting run as sudo, and if I can `sudo ls` that directory just fine, etc.
<djahandarie>
Does anyone know what to look into here?
<djahandarie>
stracing it, it looks like a child process does try to openat the directory and is getting permission denied: [pid 27557] openat(AT_FDCWD, "/nix/store/hbhdjn5ik3byg642d1m11k3k3s0kn3py-nix-2.2.2", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = -1 EACCES (Permission denied)
psyanticy has quit [Quit: Connection closed for inactivity]
<djahandarie>
It looks like at some point in the process forking, there is a setuid(30001), which is one of the nix daemon users IIRC.
<djahandarie>
But /nix/store/2c4l83wkdfrj7ra9k07l0aqvym1d5z20-user-environment.drv is root/root drwxr-x---.
<djahandarie>
So it's not surprising the error is happening, but why are the permissions so messed up?
<manveru>
djahandarie: so it was a single-user install before?
<djahandarie>
Nope, no nix at all, just a standard Ubuntu box. I just rm -rf'd /nix entirely and it ends up in the same state when I try again
tmaekawa has joined #nixos
<manveru>
what commands do you run with sudo?
<manveru>
because you shouldn't use sudo at all with nix in your case
<djahandarie>
Personally? None, I just run ./install --daemon.
<manveru>
hmm
bgamari has joined #nixos
justbeingglad1 has joined #nixos
tmaekawa has quit [Client Quit]
jgt has joined #nixos
freeman42x has joined #nixos
<djahandarie>
I see that the installer sudo rsync's a bunch of files (including the problematic one) into /nix/store. Why would we expect that to result in permissions that would allow for a nixbld user to read it?
justbeingglad1 has left #nixos [#nixos]
<manveru>
what user is nix-daemon running as?
<manveru>
should be root
<djahandarie>
nix-daemon isn't running yet.
<manveru>
can you maybe paste a log of what happens with a clean slate? from the `./install --daemon` to where you see the issue?
<manveru>
i'm having troubles understanding how that's possible
<manveru>
would be nicer if that was just in the script, but i guess it's good to know anyway :)
<djahandarie>
That doesn't include rm'ing /etc/profile.d/nix.sh
Mateon1 has joined #nixos
<djahandarie>
The script does also suggest cleanup instructions, but those *also* don't rm it
<djahandarie>
Anyways, I'll just rm it.
<manveru>
yeah
<djahandarie>
Same output but without the grep error.
<manveru>
as which user do you run this?
<djahandarie>
My personal user ('djahanda')
<djahandarie>
Either way, when I strace, it's clear the thing that is attempting to open the directory is uid 30001, which indeed doesn't have sufficient perms to open that directory (due to the other 0 perms)
<manveru>
yeah
<djahandarie>
Those perms seem to be copied due to the rsync -p
<manveru>
somehow it's not communicating with the daemon
<manveru>
or i got it all wrong :)
<djahandarie>
And the source directory (which came from a extracted tar) does have the other 0 perms in its store
<pie__>
im on nixos
<pie__>
$ ls -alh /nix/store/001s7m9kkiwhw1yr70cyf474hs2iayc1-freeglut-3.0.0.drv
<pie__>
-r--r--r-- 1 root root 2.5K Jan 1 1970 /nix/store/001s7m9kkiwhw1yr70cyf474hs2iayc1-freeglut-3.0.0.drv
<manveru>
yeah, it's all read/only and root:root
<manveru>
except for the /nix/store directory itself
<pie__>
ps aux says nix-daemon runs as root
<djahandarie>
Figured it out.
<djahandarie>
$ getfacl ~
<djahandarie>
group::r-x
<djahandarie>
user::rwx
<djahandarie>
# group: users
<djahandarie>
# owner: djahanda
<djahandarie>
# file: u0/djahanda
<djahandarie>
getfacl: Removing leading '/' from absolute path names
<djahandarie>
other::---
<djahandarie>
I extracted the tar into my home directory, which has that acl.
<manveru>
duh
<djahandarie>
Therefore the perms on the base nix/store get messed up, and the rsync preserves them
<manveru>
and they say the curl instructions are bad :P
orivej has joined #nixos
<djahandarie>
In this case I couldn't curl anyways (due to no outbound network), but it does sound like it would have saved me, maybe
<manveru>
good find though, didn't think of it at all
<djahandarie>
It sounds like the installer should really set its perms explicitly such that this can't even happen though
<manveru>
indeed
<djahandarie>
Well, that was fun. Thanks for the helping thinking it through!
<psque>
I just ran `nix-collect-garbage && nix-build --option keep-outputs true && nix-collect-garbage && nix-build`. The second gc deletes a bunch of newly built stuff that the second nix-build rebuilds. I thought `keep-outputs` was supposed to avoid this?
NinjaTrappeur has joined #nixos
<tilpner>
psque: keep-outputs is only used during gc, but you only pass it during building
<tilpner>
psque: You need to set it in nix.conf for it to have any effect on GC
hyper_ch has quit [Read error: Connection reset by peer]
<psque>
Ahh. So the conf flag is basically about respecting builds that are "tagged" that way?
trevorriles has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<psque>
tilpner: ^
hyper_ch has joined #nixos
<tilpner>
psque: I don't think there's any tagging going on
<psque>
tilpner: What does passing the flag during building actually do then?
<tilpner>
See man nix.conf, I'd only be paraphrasing that
<psque>
tilpner: Okay. I was also hoping to be able to apply `keep-outputs` for just one package but I guess I'll give up on that for now
<manveru>
psque: for nix-shell?
Diagon has joined #nixos
<psque>
manveru: For a purescript project where the build output is a JS bundle but I don't want the purescript compiler GCed
pointfourone has quit [Quit: Leaving]
hyper_ch has quit [Client Quit]
<manveru>
then you could make a gc root for the compiler
<manveru>
gchristensen: something's really up with the cache today :|
genesis has joined #nixos
<psque>
manveru: Yeah, I'm simplifying a bit. There are also the purescript dependencies, etc that get garbage collected
hyper_ch has joined #nixos
<manveru>
well, basically the derivation
<tilpner>
(What ever happened to cowsay.pw?)
<psque>
manveru: Yeah, that's the idea I started with based on someone's idea last night (`writeTextFile { name = "deps"; text = ''${foo.drvPath}'';}`) but it ended up building a whole bunch of irrelevant stuff (10s of GB)
<monokrome>
Yo! Does anyone here know why xrandr on a 3rd monitor would have a crtc error even though it sets up 2nd monitor fine?
<{^_^}>
mrVanDalo/terranix#2 (by nh2, 24 seconds ago, open): Return exit code on failure, use safe bash options
<psque>
manveru: Agreed. The thing I hoped or expected was that it would be a very quick build that would just add the already built dependencies as GC roots. Instead, it ended up downloading e.g. 3 versions of ghc
<manveru>
psque: i'd do something like `mkDerivation rec { ...; passthru = { inherit buildInputs; }; }`, and then `nix-instantiate ./whatever.nix --indirect --add-root ./keep-this-from-gc`
<manveru>
err
<manveru>
`nix-instantiate ./whatever.nix -A buildInputs --indirect --add-root ./keep-this-from-gc`
<manveru>
something like that
<manveru>
haven't really tested it :)
Jackneill has quit [Remote host closed the connection]
<clever>
manveru: you dont need passthru to access the buildInputs
<manveru>
true that
teej has joined #nixos
<rfold>
Where did the Packages button go on the homepage?
<rfold>
Oh I first have to select NixOS from the menu now.
<manveru>
rfold: the new landing page has just been released
<psque>
manveru: Okay, I'll try that second thing. Thanks
<manveru>
rfold: used to be the NixOS was selected by default :)
cjpbirkbeck has joined #nixos
<rfold>
manveru: thanks, I'll figure out how to configure Firefox to redirect me directly to NixOS so my workflow remains
wfranzini has quit [Remote host closed the connection]
<manveru>
i can recommend `nix search` otherwise
PLPD-Bot has joined #nixos
<manveru>
but yeah, depends on what you're looking for
<manveru>
maybe something strange about how nix uses curl?
<gchristensen>
hrm
bgamari_ has joined #nixos
<gchristensen>
that is very ugly
<gchristensen>
manveru: can you send me fastly-debug.com info?
bgamari has quit [Ping timeout: 245 seconds]
<psque>
You can write a default.nix that refers to executables in the `buildPhase` that aren't in `buildInputs` or `nativeBuildInputs`. Even though everything works, this is to be discouraged and semantically suspect?
<gchristensen>
probablyyeah :)
jtojnar has quit [Remote host closed the connection]
nschoe has joined #nixos
<mac10688>
hey guys, sorry if this question is obvious. I've been trying to read up on it and watch youtube videos but people aren't explicitely stating it. Let's say I am developing a haskell app that has the three layers of frontend, server and database. Do I build environments for each layer that will have the tools I need to build and run each one? So I would have a postgresql nix environment that I open a shell
<mac10688>
with, a haskell environment that has its own shell and a purescript environment with its own shell?
MichaelRaskin has joined #nixos
trevorriles has joined #nixos
hyper_ch has quit [Read error: Connection reset by peer]
<psque>
mac10688: You could do either one
fusion809 has quit [Remote host closed the connection]
<ashkitten>
i'm still pretty confused as to why networkmanager stops resolving dns after a rebuild-switch
<ashkitten>
OH
<ashkitten>
activating the configuration overwrites resolv.conf which networkmanager automatically generates on connecting
<nh2>
(how) can I escape the $ of ${interpolation} in normal "nix strings"
<nh2>
I know I can do it with ''$ in ''multiline strings''
<clever>
`nix-env -i htop` will search every single package to see if the name patches htop
<clever>
nix-env -iA nixos.htop will just go directly to htop
<zachk>
wow, whats a good way to kill a running nix-env, ctrl-c doesnt seem to work
<zachk>
I have been using ctrl-z to background it then ps aux | grep nix-env then kill -9 processNumberHere
Yakulu has left #nixos ["Error from remote client"]
<manveru>
zachk: pkill nix-env
alex`` has quit [Ping timeout: 268 seconds]
<zachk>
ty
<manveru>
also ctrl+\ might do it
adamantium has joined #nixos
rfold has joined #nixos
alex`` has joined #nixos
alp has quit [Ping timeout: 258 seconds]
justanotheruser has joined #nixos
<zachk>
pkill doesnt seem to work all the time, and whats ctrl+\ do?
markus1189 has joined #nixos
<clever>
zachk: sends SIGQUIT
<zachk>
ah never knew
<clever>
same
<clever>
i just ran `strace sh` and then whacked ctrl+\ to see what happened
<clever>
in xterm, you can hold control, and hold left-mouse, to get a menu
<clever>
you have the option to send stop, cont, int hup, term, and kill!
<clever>
but kill is send to the shell itself i think, so the whole terminal dies
rfold has quit [Quit: WeeChat 2.4]
civodul has quit [Quit: ERC (IRC client for Emacs 26.2)]
<manveru>
do you know how to add a .nar.xz to /nix/store when i downloaded it manually?
<manveru>
since i already got them, i might as well save myself trying to re-run nixos-rebuild another 100 times :P
<clever>
manveru: i think you could maybe put both the .nar.xz and the .narinfo into a directory, and then use `nix copy --from file:///path/to/dir /nix/store/path`
<clever>
manveru: use --to to make such a dir, and see what the file layout should be
<manveru>
hmm, i don't have narinfo or what the path for them should be :|
<clever>
manveru: and within the narinfo file, is the path that the .nar.xz is at, relative to the narinfo
wfranzini has quit [Remote host closed the connection]
<manveru>
ah, thanks
wfranzini has joined #nixos
<manveru>
let's try that
jabranham has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
jgt has joined #nixos
growpotkin has joined #nixos
erasmas has quit [Quit: leaving]
drakonis_ has joined #nixos
<growpotkin>
Hey could someone help clear up a few question I have about submitting PRs
<growpotkin>
?
<growpotkin>
I have about 6 or 7 python modules I am trying to submit and some of them depend on each other. Is it best to make multiple PRs or submit 1?
<infinisil>
growpotkin: One PR with many commits is better
selfsymmetric-mu has quit [Remote host closed the connection]
ddellacosta has joined #nixos
<growpotkin>
Also I have tried a few times over the last year or so on past PRs to add myself to the maintainers list and I'm always told "to do it on a separate commit". In the past I submitted a separate PR to do this and claim my old packages there; but this was rejected. This time I tried making 2 commits in a single PR. One adds the package with no maintainer and the second adds me to the maintainers list and claims the SINGLE package. I w
<growpotkin>
as just told today not to do this either... So like, how should I add myself to maintainers? When I do how should I claim my old PRs?
jgt has quit [Ping timeout: 252 seconds]
<Church->
Hmm how would I restrict a user from accessing certain dirs?
<growpotkin>
infinisil: When I make a single PR is something like "pythonPackages.{foo,bar,baz}: init at {1,2,3}" an acceptable PR message? Within that PR should I have multiple commits to add each module/package individually or should I just do 1 commit to keep the log relatively clean?
<Church->
Is there a way for me to encode that under users.extraUsers?
<growpotkin>
Church-: I have seen folks do this with an "activation script" before. Home Manager is a good point of reference.
<growpotkin>
Church-: I don't know of an existing option for this off the top of my head though.
<infinisil>
growpotkin: Who told you you shouldn't do that? I've seen it many times, and I encourage this myself, to add oneself as a maintainer in the same PR, as a first commit. A separate PR for claiming maintainership might be good though.
lovesegfault has quit [Ping timeout: 248 seconds]
<infinisil>
growpotkin: PR message doesn't matter much imo, so whatever title that gives a good overview seems fine
<infinisil>
growpotkin: Oh, but commit messages do matter, and you should absolutely use a single commit for every package and module you add
<growpotkin>
infinisil: this time it was "matthiasbeyer"
<{^_^}>
[nixpkgs] @jameysharp opened pull request #63541 → NixOS config option to disable networking entirely → https://git.io/fjVar
<growpotkin>
infinisil: He added that as a revision on two of my most recent PRs. I could look up who it was for the older ones from months ago. I couldn't decifer what the "correct" way was after these revision requests lol.
<infinisil>
You mean he force pushed to keep it a single commit or?
<growpotkin>
infinisil: I think I will close my existing slew of PRs and aggregate them into one.
<Church->
Hey what's the current way to set up steam on nix?
<growpotkin>
infinisil: I originally had 1 commit that both added the modules and added me as maintainer. Then after the request I split them into two commits under 1 PR first adding the module and then a second adding me to the maintainers list and to the new package.
<Church->
Since it seems to no longer work on my laptop
<ivan>
Church-: add steam to systemPackages
<growpotkin>
infinisil: neither seemed to be acceptable to him. He followed up on the second complaining again.
<Church->
That's weird, it no longer appears in rofi now...
<infinisil>
growpotkin: Ah, it seems that he wants to have each commit touch only a single file
<infinisil>
growpotkin: If you add you as a maintainer in a first commit, you can already reference it in the one that adds the package
<infinisil>
It's a bit nicer I guess
<growpotkin>
infinisil: okay so I'm going to close all of those old PRs. I'll make a new one that first adds me to maintainers in a commit. Then a series of commits that add each individual module. Does that sound reasonable?
fadenb has quit [Remote host closed the connection]
<emilsp>
hrm, has anyone had issues running steam with gnome3+wayland ?
seanparsons has joined #nixos
jgt has joined #nixos
polman has quit [Ping timeout: 248 seconds]
polman has joined #nixos
<infinisil>
emilsp: You I assume?
jgt has quit [Ping timeout: 257 seconds]
zachk has quit [Quit: Leaving]
Soo_Slow has quit [Remote host closed the connection]
ddellacosta has quit [Ping timeout: 272 seconds]
jbgi_ has quit [Ping timeout: 268 seconds]
<jackdk>
I am tinkering with some Cabal-3 / GHC 8.8.1alpha2 stuff, and cabal2nix won't build because of deps that are not ready for Cabal-3 (notably, `language-nix`). Is there a way to callCabal2nix that uses cabal2nix built by a different haskell toolchain?
ricekrispie2 has quit [Read error: Connection reset by peer]
polmaan has quit [Read error: Connection reset by peer]
<infinisil>
jackdk: The cabal2nix comes from pkgs.buildPackages.cabal2nix, which in turn comes from pkgs.cabal2nix. So you can add an overlay for that to change it
ricekrispie has joined #nixos
<infinisil>
Probably
<jackdk>
I am filled with renewed hope, and will try this now.
<infinisil>
(I looked at the source of callCabal2nix, which calls callCabal2nixWithOptions, which calls haskellSrc2nix, which uses cabal2nix)