<jawr>
im having an issue with nixops where it get's stuck at 'Waiting for SSH' (nixops and nodes are all in virtualbox with HostAdapter, and I can manually ssh in to the boxes). i have tried rm -rf .nixops and recreating the deployment. any suggestions?[A
<veleiro>
what's the linode interface i use? glish or lish or something
<qyliss>
like, can you get somebody in the datacenter to walk over to your server and plug something in, and then give you access through a KVM (as in keyboard/video/mouse, not kernel virtual machine) console or something
<veleiro>
yeah, that's exactly what i'd like to know what to look for
<superbaloo>
depending on what you're trying to do
<danderson>
so, I'm making a dev environment that uses python3 and nmigen, an FPGA toolkit. The problem I seem to be having is that another package pulls in python2 as a dependency, and so my nix shell's PYTHONPATH ends up pointing at python2 things.
<danderson>
but it sounds like python3.withPackages creates a more self-contained python install that doesn't rely on pythonpath, maybe?...
<danderson>
testing.
n-osborne has joined #nixos
n-osborn_ has joined #nixos
<energizer>
danderson: yes that's right
<danderson>
and indeed, it does! Spiffy.
<danderson>
Thanks!
<energizer>
python3Packages relies on PYTHONPATH but ive been convinced that's a bad choice and it shouldn't be exposed
<energizer>
withPackages is cleaner imo
n-osborne has quit [Ping timeout: 272 seconds]
<superbaloo>
meh, it relies on NIX_PYTHONPATH
<danderson>
well, in my case, it relies on something that python2 stomps all over and breaks things :
<danderson>
:)
<danderson>
so withPackages is the way to go.
<superbaloo>
yeah, less likely to interfere for sure
<KarlJoad>
Does Nix have a convenient way to set list environment variables? For example, $PATH uses colons as delimiters in its search list.
<tpw_rules>
yes, i've seen it. what is it? uhhhhhh
kreisys has quit [Client Quit]
<tpw_rules>
there is concatStringsSep
kreisys has joined #nixos
<KarlJoad>
tpw_rules: Yeah... I was thinking there might be a function that takes in a list of strings and the name of the environment variable and constructs PATH=/path/to/new/thing:$PATH for you.
<clever>
jybs_: but i also specially configured my system to sign everything it builds (and downloads, it seems), with its own keys, so i have an extra thing i can trust, for locally built things
simba1 has joined #nixos
<KarlJoad>
Is there a way to reference the directory where the shell.nix file is by a variable that expands to an absolute path?
<clever>
KarlJoad: ./. expands to an absolute path automatically
<KarlJoad>
clever: I'm not using the `src` input though. I'm using a shell.nix and `mkShell.
<clever>
KarlJoad: ./. is a valid value, and can be used anywhere in a nix expression
<jybs_>
And for a binarycache I just have to trust the person who made it right?
<clever>
KarlJoad: the same way strings and numbers can be used as values in an expr
<clever>
jybs_: mostly, your trusting the owner of the private key, who signs the paths
<clever>
jybs_: once signed, anybody can host the binary cache, and the signature stops them from doing nasty things
<jybs_>
But when I add a cahce I don't add a key do I?
<clever>
you usually have to add the public key as well
nodezz has quit [Quit: Connection closed]
<clever>
or nix will just refuse to download anything on the new cache
<KarlJoad>
clever: I have shellHook = ''export TEXINPUTS=${concatStringsSep ":" [ "." "./." "$TEXINPUTS"]}'', but the document isn't compiling (class not found, which is somewhat expected). When I specify the full path manually, it works.
<clever>
KarlJoad: dont quote the path
simba1 has quit [Ping timeout: 272 seconds]
<clever>
KarlJoad: ./. and "./." are very different things
<clever>
./. will automatically expand to the absolute path, "./." is a dumb 3 char string, that remains as-is
<jybs_>
How do you order your caches (sorry lots of questions)?
<etu>
matthewcroughan: `sfdisk --dump $DISK1 | sfdisk $DISK2` will dump the partition scheme from one disk to another, and I'll assume that this would work bad if the target disk is smaller than the source disk
<KarlJoad>
clever: Probably because I tried to recompile my document too quickly.
simba1 has joined #nixos
<KarlJoad>
The shell.nix hadn't been rebuilt by lorri and reloaded by direnv yet.
<clever>
KarlJoad: you may also want to `echo $TEXINPUTS` to confirm its doing what you want
<matthewcroughan>
and I have no idea why it works, it just does
<matthewcroughan>
etu: the /dev/disk/by-uuid changes on every single boot of the system anyway, what is it for?
<matthewcroughan>
boot.loader.grub.mirroredBoots.devices changes every single boot
<matthewcroughan>
last boot it was 4842122891116007044, now it's 18353865175145601115
<etu>
matthewcroughan: Uhm, uuid is embedded in the filesystem and shouldn't change.
<matthewcroughan>
I guess this is because it's randomly booting from one of the disks or the other?
<matthewcroughan>
one of the mirrors
<etu>
That's why one use uuid's
<jybs_>
one more clarification - does the path hash include the output hash?
<etu>
Because then it doesn't matter what you boot from
<KarlJoad>
clever: I was doing that. I think I tried to re-compile before the re-build was done, but ran the `echo` after the build was done and the switch happened.
<matthewcroughan>
etu: multiple? Or just one?
<clever>
jybs_: nope
<matthewcroughan>
do you put more than one disk in your mirrored boots?
<jybs_>
that's annoying
<clever>
KarlJoad: thats why i prefer plain direnv without lorri, so i know when its busy
<jybs_>
So two identical paths could have different contents?
pjt_tmp has quit [Ping timeout: 240 seconds]
<etu>
matthewcroughan: Only one in mirrored, I was very confused by that to begin with. But one mirror means that you have one "main" and one "mirror" which becomes two.
<jybs_>
(on different systems)
<clever>
jybs_: only if the build is impure
<matthewcroughan>
etu: your blog post shows devices = [ "/dev/disk/by-uuid/<FS-ID>" ];
waleee-cl has quit [Quit: Connection closed for inactivity]
<clever>
jybs_: if the paths are identical, the directions on how to build the path are identical
<jybs_>
Sure
<matthewcroughan>
but shouldn't it be devices = [ "/dev/disk/by-uuid/<FS-ID>" "/dev/disk/by-uuid/<FS-ID>" ];
<clever>
jybs_: so its a question of if those directions produce the same result when run on diff machines
<jybs_>
I mean, as soon as something uses a timestamp in something it's not
<clever>
jybs_: `repeat = 1` in `nix.conf`, makes it repeat all builds one extra time (after the normal 1 build), and then automatically compare the outputs
<clever>
jybs_: if the outputs dont match bit-for-bit, the build fails
<jybs_>
Can I see if that was enabled from the metadata?
<matthewcroughan>
you're supposed to give it a list of two disks, not one
<clever>
jybs_: this site is doing comparisons of various things in nixos, and reporting when they differ
<matthewcroughan>
or is this different for efi?
<KarlJoad>
clever: Ehh... I don't exactly change the shell.nix file of this particular repository too frequently. Normally I'm working on writing the lab reports in there.
<etu>
matthewcroughan: Nah, it still uses /boot by default
<etu>
matthewcroughan: So mirrors are "other locations" than /boot
<matthewcroughan>
well the path I have there is not the short path like that
<matthewcroughan>
the path I have is the long path, of which there is only ever one of in the system
<matthewcroughan>
/dev/disk/by-uuid/4842122891116007044 instead of /dev/disk/by-uuid/6258-01A0
<matthewcroughan>
both disks have this same ID
<clever>
jybs_: in the gcc example from that site, it looks like the address of .fini moved by 16 bytes, for no obvious reason
<etu>
matthewcroughan: I think your boot filesystems are on /dev/sda3 and /dev/sdb3
<etu>
matthewcroughan: Because that format of uuid's are used for fat32
<matthewcroughan>
all I did was follow your guide and I have both uuids
<etu>
matthewcroughan: So if you have a zfs mirror on the other two, it's the same filesystem, right?
<matthewcroughan>
yes
<clever>
jybs_: but for asciidoc, it is just a timestamp in a man page
<matthewcroughan>
it's a zfs mirror and nothing more
<etu>
matthewcroughan: Then that output makes sense?
<matthewcroughan>
yes, but which uuid should I be using?
<matthewcroughan>
does it matter?
<etu>
matthewcroughan: the two short ones, one should be mounted on /boot and the other one should be mounted somewhere else to be able to be used as mirrored device
<matthewcroughan>
I've used the long UUID, which means 50% of the time when the system boots up, devices = [a path that doesn't exist]
d3od has joined #nixos
d3od has quit [Excess Flood]
<matthewcroughan>
why does the long one exist though? And why is it the same for both disks?
d3od has joined #nixos
<etu>
matthewcroughan: I think it's your ZFS mirror.
<jawr>
im having some issues with a nixops deployment/nix where after deployment they dont boot (the uuid is correct, but it complains about a missing /mnt-root directory. before i tried to make it fancier it was working, if anyone fancies casting an eye its here https://pastebin.com/GUmC2Xi1
larrym has quit [Quit: Connection closed]
rb2k has joined #nixos
is_null has quit [Remote host closed the connection]
rb2k has quit [Client Quit]
fuzzypixelz has joined #nixos
<fuzzypixelz>
how can I get GNOME 40? is there an available derivation (yet)?
<luxemboye>
If there's an issue with a package (nixpkgs, not upstream), are you supposed to open an issue on GitHub, start a thread on the forums or is there a bug tracker?
dev_mohe has joined #nixos
<cole-h>
GitHub issues are the bug tracker :) I'd go there first.
andreas303 has quit [Remote host closed the connection]
<aforemny>
How can I build my whole nixos-config against pkgsStatic? `nix-build … --arg overlay '[(import pkgs/top-level/static.nix)]'` does seem to be fundamentally wrong because it cannot even build `hello`, while `nix-build pkgsStatic.hello` builds.
<zceejrk>
Hello, can someone help me setup a systemd service in NixOS? I am trying to start a service as a user which has my desired binary in it's nix-env, but when trying to start from systemd I get a not found error. Here is the relevant pastebin: https://paste.debian.net/1184677/
r0bby has quit [Read error: Connection reset by peer]
jfhbrook has joined #nixos
mog- is now known as mog
jlpeters has joined #nixos
gluegadget has joined #nixos
<zceejrk>
adisbladis: Yes I added pkgs.go-ethereum to the path but still the same error...
cz3 has joined #nixos
betawaffle has joined #nixos
CMCDragonkai1 has joined #nixos
SrPx has joined #nixos
<zceejrk>
Does the service need to have the same name as the pkg or something?
r0bby has joined #nixos
<clever>
nope
* clever
checks something
MasseR has joined #nixos
sjlnk has joined #nixos
<j2t>
hi, in dbeaver interface I need to fill a path of "local client". pg_restore is in `/nix/store/anl1m3q95092clq3kyipgbsj8iw3sav2-system-path/bin/pg_restore.`, but this is not the way to do it. How should I do it?
<clever>
zceejrk: aha, ExecStart must be an absolute path, systemd rules
<clever>
zceejrk: but you can instead use .script
<adisbladis>
ExecStart is usually preferable
<clever>
systemd.services.<name>.script
<clever>
Shell commands executed as the service's main process.
<clever>
if using ExecStart, it must be absolute, so you would do "${pkgs.go-ethereum}/bin/geth ..."
<clever>
but .script will auto-generate a shell script for you, and then put the path of the shell-script into ExecStart
CMCDragonkai1 has quit [Client Quit]
Qwerky has quit [Remote host closed the connection]
<infinisil>
clever: zceejrk: I'm fairly certain that if you put the binary into systemd.services.*.path, you can use the binary name directly in ExecStart
<clever>
For each of the specified commands, the first argument must be either an absolute path to an executable or a simple file name without any slashes. Optionally, this filename may be prefixed with a number of special characters:
<clever>
infinisil: from `man systemd.service`
<clever>
sounds like it should allow something in $PATH
<simonpe^^>
I'm not sure how I can add stuff to the debug output of a package that I've build with `mkDerivation { separateDebugInfo = true; ... }`
<simonpe^^>
I want to put a custom symlink in the pointing to something depending on the buildInputs of the original derivation
<pie_>
clever: im using your recovery env (or some older verison of it) , nixos-rebuild fails with some permission denied error when run in nixos-enter, any way to make it work?
<j2t>
can i point the dbeaver gui `/run/current-system/sw/bin/pg_restore` or are there any inconvienences?
<{^_^}>
[rfc39-record] @grahamc pushed commit from rfc39 to main « Automated team sync results. »: https://git.io/Jtwhf
fuiltilt has quit [Quit: WeeChat 2.9]
<clever>
pie_: have you tried nixos-install instead? thats basically a helper to run nixos-rebuild in a chroot for you
<clever>
pie_: the only real difference, is that it uses <nixpkgs> from outside the chroot
<pie_>
ok it actually says "operation not permitted", im guesing it cant write to the store but idk
<pie_>
(guess i could strace it)
<pie_>
clever: guess i could do that and set NIX_PATH
<pie_>
would be nice if i didnt need to :p
<clever>
i dont remember exactly why the problem happens
<pie_>
my bad, google seems to know bout the problem
<clever>
whats google say?
<pie_>
i spoke too soon
CMCDragonkai1 has joined #nixos
ahmedelgabri has joined #nixos
<pie_>
clever: ah might be something about sandboxing. apparently it might be fixed on nix master as of at least oct 28 2020 but im not running that
<clever>
--option sandbox false
<pie_>
clever: can that be passed in an env var or something?
<pie_>
remember im calling nixos-rebuild
<clever>
nixos-rebuild still accepts --option
<pie_>
oh it/extra args are not listed in the man page :I
<clever>
In addition, nixos-rebuild accepts various Nix-related flags, including --max-jobs / -j, --show-trace, --keep-failed, --keep-going and --verbose / -v. See the Nix manual for details.
<pie_>
yeah its in the script but i do- oh ok :I
<pie_>
well they should have put the ... in the summary of whatever :Pű
<pie_>
hm, looking at the script it looks like the support is explicit
<pie_>
so they should have specifically listed the accepted flags
<clever>
PR time!
<pie_>
clever: thanks for the nudge
* pie_
pops yak shave stack
* pie_
adds to todo >.>
civodul has quit [Read error: Connection reset by peer]
<__monty__>
j2t: Sounds like you should add postgresql as a dependency of dbeaver.
oldandwise has quit [Quit: leaving]
oldandwise has joined #nixos
CMCDragonkai1 has quit [Quit: CMCDragonkai1]
tomaw has joined #nixos
<j2t>
__monty__: thanks for your response :). I simple have `dbeaver` in my home.packages in home.nix. How would go about adding it as a dependency? Could you point a similar example or something?
<__monty__>
Hmm, looking at the expression it looks like it should just pick up the right DB tools from the PATH tbh.
<__monty__>
Can't find any documentation on configuring it.
<__monty__>
What errors are you getting?
oldandwise has quit [Quit: leaving]
oldandwise has joined #nixos
<j2t>
__monty__: it does with exception of "Local client" option, which is needed to be able to pg_restore for example. Anyway, what are the inconveniences when i manually set that option in gui to `/run/current-system/sw/bin/pg_restore`?
<{^_^}>
[nix] @edolstra merged pull request #4464 → narinfo: Change NAR URLs to be addressed on the NAR hash instead of the compressed hash → https://github.com/NixOS/nix/pull/4464
<exarkun>
audio playback quality on nixos 20.09 seems significantly degraded from 20.03. pianobar skips constantly. anyone have any suggestions? (plan A is go back to 20.03 I guess)
<exfalso>
hi, we created a module that allows nixops to be used on ubuntu machines to manage systemd services, however the deployment still includes a full nixos tree. Does anyone know whether there is a way to override the "top-level" system derivation?
Qwerky has quit [Remote host closed the connection]
Tv` has joined #nixos
rb2k has joined #nixos
<Tv`>
what's the magic incantation to nix-env -i from a nixpkgs.git clone? i can `nix build -I nixpkgs=. nixpkgs.blah` but nix-env is giving me trouble
jmeredith has joined #nixos
<Tv`>
ahh, `-f .` not `-I`, and then no nixpkgs. prefix on the attr
<KarlJoad>
zceejrk: Unless you C-c it during the actual switching process, I don't think anything can really happen. That's kind of the point of the transactional upgrade.
lsix has quit [Ping timeout: 246 seconds]
Qwerky has quit [Ping timeout: 240 seconds]
lsix has joined #nixos
<{^_^}>
[nix] @grahamc pushed to revert-4464-nar-narhash-addressed « Revert "narinfo: Change NAR URLs to be addressed on the NAR hash instead of the compressed hash" »: https://git.io/JtrBP
<{^_^}>
[nix] @grahamc opened pull request #4535 → Revert "narinfo: Change NAR URLs to be addressed on the NAR hash instead of the compressed hash" → https://github.com/NixOS/nix/pull/4535
<patagonicus>
exfalso: What exactly do you mean by override the top-level system derivation? It's all just modules, so you can disable the modules that set the stuff you want to change and then set it yourself.
<evanjs>
Had to use NixOS options search and noticed that multiline code blocks are now rendered properly. Thanks guys! (wow that was a while back, huh) -- https://github.com/NixOS/nixos-search/pull/239
n-osborne has quit [Remote host closed the connection]
graf_blutwurst has quit [Quit: WeeChat 3.0]
<matthewcroughan>
Just to confirm, you're suggesting that my docs should be made as a PR for this manual source you just linked me to?
<ryantm>
There are three main manuals: Nix, NixOS, and Nixpkgs.
<ryantm>
No, I'm suggesting the opposite of that.
<ryantm>
You are asking where the mdbook source is, but I'm telling you that is only for the Nix tool, and your contribution needs to go in the NixOS manual.
<ryantm>
matthewcroughan: It's currently my main focus for my side project time, but I think it could still be a couple months out, depending on how much time I get.
<ryantm>
If other people help things can move quicker of course.
<matthewcroughan>
ryantm: do you think we should set up a location that people can contribute to that is mdbook based so that they ultimately end up in the docs?
<matthewcroughan>
or do you think this will complicate things?
thelounge52 has quit [Ping timeout: 264 seconds]
thelounge524 is now known as thelounge52
<matthewcroughan>
ryantm: My idea is that we could even move a lot of content off nixos.wiki and transition it to mdbooks form, and clean it up a bit, and add attributions below the pages
<simpson>
matthewcroughan: As somebody who has translated a few pages to Markdown, there's nothing to it but to do it. Putting hands to keyboard is the hardest part.
ahmedelgabri has quit [Ping timeout: 258 seconds]
<prusnak>
matthewcroughan: i do, we use nix in docker for Trezor Firmware reproducible builds
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
cosimone has quit [Remote host closed the connection]
<KarlJoad>
Does anyone have any clues on how can I ignore certain source files in a shell.nix that uses mkShell? I'm working on a LaTeX project, and the shell.nix rebuilds when the temporary files are re-generated.
<elvishjerricco>
KarlJoad: Something I do sometimes is `builtins.fetchGit ./.` because not providing a rev causes it to take the current state of the working tree, capturing only tracked files.
<colemickens>
gchristensen: You'll want to use boot.loader.grub.efiInstallAsRemovable if you want it to be easily bootable on random uefi machines.
<colemickens>
gchristensen: otherwise you'll need a UEFI that allows adding entries, or booting and running nixos-install/efibootmgr in order to boot it.
<gchristensen>
colemickens: what if I want the ESP to be on a USB disk, but it is booting internal drives?
<colemickens>
I haven't run that exact setup, but I'd assume it'd be pretty much the same story. The bios is still going to find the ESP and then look for the default uefi executable and either find or not find it.
<nixy>
tpw_rules: There is a script that is supposed to set up the PATH at `~/.nix-profile/etc/profile.d/nix.sh` you have to make sure whatever type of shell you are in is sourcing it.
<tpw_rules>
how does it get sourced by default? e.g. nix shows up when i connect over ssh but not when i start a terminal program from the deskop
<tpw_rules>
but that script is not sourced in .bashrc
<nixy>
I think it gets added to either .profile or .bashrc. So it might be a login vs non-login shell thing can't fully remember but if it's sourced over ssh it's probably in .profile
<colemickens>
Does anyone know where the discussion(s) of building the nixos docker image with Nix is?
<KarlJoad>
abathur: I saw your message, and I tried that last night. But, I kept ending up with "error: string '/nix/store/yh3z24rd0xchdvv8n0b431sf7jbx12xj-gitignore' cannot refer to other paths"
n-osborne has quit [Remote host closed the connection]
<KarlJoad>
abathur: Nevermind. I was using the builtins.path because of an error I was reading about online. Turns out I didn't need it.
thelounge52 has joined #nixos
Jd007 has joined #nixos
hyper_ch2 has quit [Read error: Connection reset by peer]
srl295 has joined #nixos
hyper_ch2 has joined #nixos
<KarlJoad>
abathur: I still end up with lots of build events like this: "build message, message: BuildEvent(Started { nix_file: NixFile("/home/karl/Repos/ece441-reports/shell.nix"), reason: FilesChanged(["/path/lab_reports/lab_03/report.aux"])", which is a temporary file. Does switching to mkDerivation make more sense here?
<exarkun>
I updated my nixos channel to the 20.09 url and updated, but when I import nixpkgs, the version still starts with 19.03. What else do I need to do?
thelounge52 has quit [Ping timeout: 264 seconds]
eri451 has quit [Ping timeout: 240 seconds]
eri has joined #nixos
<shapr>
exarkun: are you using niv to pin nixpkgs in a repo?
eri is now known as Guest64202
<exarkun>
shapr: no. I have repos that pin nixpkgs but right now I'm just confused about why this is the nixpkgs I get in a `nix repl` interactively
<exarkun>
although that's also an indirection
radvendii has quit [Client Quit]
<exarkun>
what I actually want is an up-to-date version of stack but I'm getting 1.9.3
<shapr>
hm, I only know about niv to pin packages, or updating the channel for NixOS, I don't know the inbetween part of how to update the channel for nix as a package manager inside some other OS
radvendii has joined #nixos
<shapr>
though now that you mention it, I want to know this so I can update nix on my work mac
fuiltilt has joined #nixos
<shapr>
exarkun: if you're using nix on top of some OS, how did you update your channel?
<shapr>
I used nix-channel to set my channel, then nix-channel --update, then nixos-rebuild switch --upgrade
<exarkun>
Hm
<shapr>
is it possible you've installed stack with nix-env ?
<exarkun>
I guess I could be missing the `nixos-rebuild switch` step .... but
<exarkun>
it's not
<shapr>
ok
<stevenroose>
Does anyone know a reference or wiki page describing a bare-minimum sent-only e-mail setup? Basically to (1) get ZFS ZED to send e-mail notifications if a drive fails or (2) have some application send e-mail? I can only find the simple-nixos-mailserver project which is huge and intended as a full mailserver.
<exarkun>
But I am using morph to manage this system
<colemickens>
exarkun: idk what to point you to, but the important thing to notice is that channels are user specific
<colemickens>
exarkun: so if you run `sudo nixos-rebuild` then it's going to use root's channels, so you'll want to make sure root's nix channels are up to ddate.
<exarkun>
is there something other than `sudo` that makes it choose root's channels?
<exarkun>
eg, `wheel` membership?
<colemickens>
well, sudo is `super user do` and effectively runs the command as root
<colemickens>
when you do `sudo true` it runs true as root, so it stands to some reason that `sudo nix...` would interact with root's nix channels
n-osborne has joined #nixos
<exarkun>
yes...
<exarkun>
But my observation is that `nix-channel ...` edited my user's channels and then `nix repl` used root's channel
<exarkun>
with a non-root user and no sudo anywhere
lunik1 has quit [Quit: :x]
<colemickens>
Uh
<colemickens>
`nix repl` likely did something even more different
<colemickens>
which I'd rather not talk about, unless you're really confused about it and want to know
<exarkun>
:(
<colemickens>
What `nix repl` command did you run?
<colemickens>
`nix repl` is a "nix 2.0 cli" command so sometimes it does behave differently
vidbina has joined #nixos
<cole-h>
nix repl is like nix 1.0+2.0 cli
<exarkun>
2.3.7
<colemickens>
ehck thats what I was afraid of
* colemickens
really loathes this conversation
<colemickens>
exarkun: sorry, this is confusing, someday this will be a thing of the past, or well-documented at least
<exarkun>
for what it's worth, I used it to inspect the version of <nixpkgs>, and then I switched to this because it was faster:
<colemickens>
(the `nixpkgs=` component, of NIX_PATH, to be precise)
<exarkun>
Okay, indeed, that does match up
<exarkun>
(I have no idea who set NIX_PATH though)
novafacing has quit [Read error: Connection reset by peer]
<colemickens>
haha, uhh I don't know the default behavior. I dislike all the implicit-ness of it all, so my NixOS config explicitly sets `nix.nixPath = [];`
novafacing has joined #nixos
rb2k has quit [Ping timeout: 264 seconds]
<exarkun>
mmm I like explicit, maybe I should do that
tsrt^ has quit []
alexherbo2 has joined #nixos
ericsagn1 has joined #nixos
lunik1 has joined #nixos
novafacing has quit [Ping timeout: 240 seconds]
Curunr[m] has joined #nixos
rb2k has joined #nixos
n-osborne has quit [Remote host closed the connection]
<remexre>
is there a nix-specific hack necessary for emscripten + ASan/UBSan/etc? it looks like emscripten tries to cache them inside /nix, which ofc fails: https://p.remexre.xyz/XeIswkBPoBg=
<rmcgibbo[m]>
fuzzypixelz: same. the package i install into my home managed config is `vscode-with-extensions.override { vscodeExtensions = list-of-extensions; };`
<FireFly>
is fetchgit behaviour supposed to differ from a `git clone`? I came across a repo with a submodule that fetchgit (with fetchSubmodules = true) fails on, but which `git clone --recurse-submodules` works fine