<genesis>
evanjs : seems nix eval doesn't like hexadecimal too
drakonis1 has quit [Read error: Connection reset by peer]
drakonis1 has joined #nixos
<energizer`>
I can do checkConflicts=false; but that seems bad, how do i avoid these? https://bpaste.net/N6DA i dont even understand what it's complaining about with `six`
<{^_^}>
[nixpkgs] @worldofpeace pushed 4 commits to master: https://git.io/JvgGu
<kini>
What precisely do "free" and "unfree" licenses mean in nixpkgs? I get that for specific licenses that exist in SPDX we can just look at the info there, but what about the generic licenses `free` and `unfree`?
drakonis1 has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
<kini>
The text in the nixpkgs manual about `unfreeRedistributable`, `unfree`, and `unfreeRedistributableFirmware` seem to distinguish them based upon what parts are non-redistributable. Does it follow that `free` refers to all licenses where the source and all build artifacts *are* redistributable?
<tbenst[m]>
has anyone encountered `sudo: /run/current-system/sw/bin/sudo must be owned by uid 0 and have the setuid bit set` before and know a fix? No idea how it happened
drakonis1 has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos
<cole-h>
Do you have `security.sudo.enable = true;` in your configuration.nix?
<samueldr>
tbenst[m]: yeah, sounds like your sudo doesn't come from suid wrappers
<samueldr>
it would be /run/wrappers/bin/sudo when it's working properly (assuming nixos)
<cole-h>
And as samueldr touched on, `security.wrappers` doesn't work on non-NixOS. AFAIR, there is no way to get setuid on a non-NixOS system, other than copying it somewhere and setting perms there
drakonis_ has quit [Client Quit]
jgeerds has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « hackage-packages.nix: automatic Haskell package set update »: https://git.io/Jvgna
<flokli>
There should be a comment somewhere, at least in the git history
<{^_^}>
[nixpkgs] @cole-h opened pull request #81344 → qimgv: init at 0.8.9 → https://git.io/Jvgch
<clever>
under normal linux, the nss modules get dlopen()'d into each process doing dns, by glibc
<clever>
but with nix involved, each process might have a different arch (32bit, 64bit), it might have a different glibc, and maybe a different gcc version
<clever>
combined, that makes it a nightmare to keep the nss modules compatible
<flokli>
We run nscd pointed to the configured nss modules. And applications try to resolve through nscd first, if the socket file exists
<clever>
so instead, it loads all the modules into nscd, and proxy's the request thru that
<flokli>
So that way we have a somewhat stable interface
<{^_^}>
[nixpkgs] @jonringer pushed commit from @bhipple to staging « rustPlatform.fetchCargo: handle custom Cargo.lock patchfiles with validation »: https://git.io/JvgCr
h0m1 has quit [Ping timeout: 240 seconds]
<NemesisD>
@lovesegfault i think my problem is i had zsh.oh-my-zsh.enable = true, but not zsh.enable
<NemesisD>
now i just gotta figure out how to reconcile this with my existing nix customization
h0m1 has joined #nixos
<NemesisD>
@lovesegfault also, i'm curious, how do you keep that git repo in sync with the config on your system? maybe you clone the repo and then symlink it all into place on the system?
<clever>
energizer`: the only real advantage is that you dont need root to co-operate with your love of nix :P
<clever>
and more privacy over sources in /nix/store/
<energizer`>
clever: it seems i need root to cooperate enough to enable unprivileged user namespaces
boxscape has quit [Ping timeout: 240 seconds]
<clever>
energizer`: thats an extra requirement that debian tossed in, defeating the main use of namespaces
<energizer`>
clever: everyone can see all my source code if i use a shared store?
<clever>
energizer`: yeah, the source gets copied into /nix/store when built, and the whole store is world-readable
<energizer`>
clever: oh that seems important
<worldofpeace>
DigitalKiwi: lol, that's a secret I'll never tell :D
<energizer`>
i have `grep CONFIG_USER_NS /boot/config-$(uname -r)` -> `CONFIG_USER_NS=y` but https://bpaste.net/HZSQ
<energizer`>
on centOS
<energizer`>
6.10
Supersonic has quit [Ping timeout: 255 seconds]
<cole-h>
I don't know how accurate `/boot/config-...` is since I don't have it. Does it have the same output as `zgrep CONFIG_USER_NS /proc/config.gz`?
<clever>
energizer`: does /proc/sys/kernel/unprivileged_userns_clone exist?
<energizer`>
clever: no
<clever>
energizer`: what error happens if you try to use namespacing?
drakonis has quit [Quit: WeeChat 2.7.1]
<clever>
energizer`: ah, i see it in the paste
<clever>
energizer`: that looks more like it just cant find the directory "nix"
<energizer`>
cole-h: gzip: /proc/config.gz: No such file or directory
<cole-h>
centOS is weird man
<{^_^}>
[nixpkgs] @thoughtpolice pushed 2 commits to master: https://git.io/JvgWi
waleee-cl has quit [Quit: Connection closed for inactivity]
<worldofpeace>
excelsiora: hi 👋
<mlatus_>
Is there a way to override the transitive dependency of a package?
<energizer`>
clever: i unpacked the tarball and now there's a /nix-2.3.3-x86_64-linux/store/ with a bunch of stuff in it, but `$ RUST_BACKTRACE=1 ./nix-user-chroot-bin-1.0.2-x86_64-unknown-linux-musl ~/nix/nix-2.3.3-x86_64-linux/ ` gives the same error as before
<clever>
energizer`: have you read the docs for nix-user-chroot to see how its meant to be used?
<mlatus_>
like if package A depends on B, which depends on C, instead of nested override, is there any more elegent solution? I'm imaging something like A.overlay(self: super: C = ...;)
<clever>
mlatus_: (pkgs.extend (self: super: { C = ...; }).A
<clever>
mlatus_: you can use pkgs.extend to add more overlays to nixpkgs at any time, and then pull any package out, and it will rebuild whatever got stuck between A&C
<cole-h>
clever: That's really cool. Is there a wiki page for that or something I can bookmark?
<clever>
cole-h: not sure
<mlatus_>
clever: never heard of pkgs.extend, will check, thanks
<clever>
energizer`: and what happens when you `unshare --user --pid echo YES` ?
<lovesegfault>
What's the simplest way of adding a `--frobnicate` flag to a systemd unit's ExecStart?
<lovesegfault>
cc. clever worldofpeace
<worldofpeace>
lovesegfault: like how do you reset an ExecStart. You can't add ot it, you have to first clear it then set it. so `ExecStart = [ "" "autahuth -frobnicate" ] `?
<lovesegfault>
worldofpeace: ExecStart is an arr?
<lovesegfault>
Can't I do `systemd....serviceConfig.ExecStart = config.systemd...serviceConfig.ExecStart + " --frobnicate"?
<worldofpeace>
lovesegfault: ¿Qué? I think that would be an infinite recursion.
<lovesegfault>
The good kind :P
<lovesegfault>
I just want to start bluetoothd with `--noplugin=sap`
<clever>
lovesegfault: due to how .serviceConfig is an option and .ExecStart is technically not a nixos option, you cant use things like mkForce on it
<clever>
lovesegfault: if you set it to a list, you will add multiple ExecStart='s to the .service file, and the above would cause it to have 3 entries, the original, an empty, and the new one
<lovesegfault>
Damn I hadn't even thought of that clever
<clever>
and the empty one is a signal to systemd, to ignore the original copy
<clever>
yes, its ugly as heck :P
<lovesegfault>
Wow
<worldofpeace>
yep
<worldofpeace>
I usually comment that wherever I use it
<lovesegfault>
Amazingly ugly
<clever>
fix the module to accept what you want as a proper option!
<worldofpeace>
I do the following to kill udev-settle systemd.services.systemd-udev-settle.serviceConfig.ExecStart = ["" "${pkgs.coreutils}/bin/true"];
<clever>
worldofpeace: you could also just systemd.services.systemd-udev-settle.wantedBy = mkForce []; i believe
<{^_^}>
[nixpkgs] @worldofpeace pushed 3 commits to release-20.03: https://git.io/Jvg83
<worldofpeace>
lovesegfault: you're welcome ✨
<lovesegfault>
Man, bluez is really undocumented
<lovesegfault>
it's scary
<clever>
lovesegfault: i once configured it to emulate a bluetooth speaker, so the phone was then able to connect, and play audio out of the laptop speakers
<lovesegfault>
:O
<{^_^}>
[nixpkgs] @worldofpeace pushed 137 commits to staging-20.03: https://git.io/Jvg8n
<clever>
lovesegfault: its an option i found while digging around in blueman-manager
<clever>
it created an audio source in pulseaudio, and auto-added a loopback, feeding that source back into the primary sink (the alsa device for the laptop)
<{^_^}>
[nixpkgs] @sepiabrown opened pull request #81352 → [nixpkgs/nixos/doc/manual/configuration/modularity] lack environment.systemPackages option in the example → https://git.io/JvgBJ
<{^_^}>
[nixpkgs] @adamse opened pull request #81357 → ghc: use ld.gold on linux for GHC 8.8.3 → https://git.io/JvgRU
vandenoever has quit [Ping timeout: 240 seconds]
o1lo01ol1o has joined #nixos
<koleesch_>
hello
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<koleesch_>
i'm installing nixos on an old laptop from me. I get the message "warning: The option 'networking.enableIntel3945ABGFirmware' defined in '...' has been renamed to 'hardware.enableRedistributableFirmware'".
nerdmaxx has quit [Quit: leaving]
maddo has joined #nixos
<koleesch_>
After the installation do i need to do something based on this message?
o1lo01ol1o has quit [Ping timeout: 265 seconds]
<etu>
koleesch_: Find the location where you have networking.enableIntel3945ABGFirmware set and replace it with hardware.enableRedistributableFirmware
<etu>
koleesch_: For now it's fine, because it's aliased. So it will work. But it won't work forever :)
<{^_^}>
[nixpkgs] @domenkozar closed pull request #81357 → ghc: use ld.gold on linux for GHC 8.8.3 → https://git.io/JvgRU
<koleesch_>
etu: thx
<ar>
/3
iyzsong has quit [Ping timeout: 255 seconds]
iyzsong- has joined #nixos
Baughn has quit [Ping timeout: 240 seconds]
iyzsong has joined #nixos
iyzsong- has quit [Ping timeout: 240 seconds]
knupfer has joined #nixos
<{^_^}>
[nixpkgs] @cdepillabout merged pull request #81356 → spago: unbreak from upgrade to ghc-8.8.2 → https://git.io/JvgB7
<{^_^}>
[nixpkgs] @cdepillabout pushed 6 commits to haskell-updates: https://git.io/JvgRV
knupfer has quit [Ping timeout: 248 seconds]
niso__ has quit [Remote host closed the connection]
niso has quit [Remote host closed the connection]
niso_ has quit [Remote host closed the connection]
<JonReed>
cinimod_: Dunno, then. Nix tends to cache things. It can do it when sha matches something it already fetched. But if that's not it, then it might be something else.
kleisli has quit [Ping timeout: 265 seconds]
kleisli has joined #nixos
kleisli has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @nlewo merged pull request #81352 → [nixpkgs/nixos/doc/manual/configuration/modularity] lack environment.systemPackages option in the example → https://git.io/JvgBJ
<{^_^}>
[nixpkgs] @nlewo pushed commit from @sepiabrown to master « nixos/manual: add environment.systemPackages option in an example »: https://git.io/JvguO
cosimone has joined #nixos
cosimone has quit [Client Quit]
cosimone has joined #nixos
<MasseR>
Don't know if it's related, but on Haskell ecosystem invalid packages are silently ignored. If you for example forget the {} after callHackage
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mmatthieu>
Hi everyone ! Is anybody succeeded in using the `next` browser ? Here it starts, loads pages but UI is blank.
ixxie has joined #nixos
chloekek has quit [Ping timeout: 255 seconds]
ixxie has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @fatpat opened pull request #81363 → new package scaleft-1.41.0 → https://git.io/Jvgu5
Thra11 has joined #nixos
anohigisavay has quit [Quit: Konversation terminated!]
<cinimod_>
I am doing `nix-garbage-collect` to see what happens - I doubt this the optimal nix way of doing things
cosimone has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @petabyteboy pushed commit from @bhipple to master « viu: upgrade cargo fetcher and cargoSha256 (#81240) »: https://git.io/Jvgzv
<Thra11>
I tried to do an automatic `git bisect run` to find out which commit broke a package build. However, I think it's just found me a commit which broke a dependency (fixed by a later commit). Anybody have any tricks for using git bisect with nixpkgs?
<cinimod_>
I am in nix hell :( I think I preferred cabal hell
<Thra11>
Looks like if my test script returns 125, it will be interpretted as `git bisect skip`. So the question is can I tell the difference between a `nix build ` which failed because of a dependency and a `nix build` where all the dependencies built successfully, but the final build failed?
dhess has quit [Remote host closed the connection]
chloekek has joined #nixos
dhess has joined #nixos
<LnL>
yeah you can use that to skip the unrelated cases
<{^_^}>
[nixpkgs] @ggreif opened pull request #81364 → wasmtime: bump to v0.12.0 → https://git.io/Jvgzq
werner291 has joined #nixos
cosimone has quit [Remote host closed the connection]
<Thra11>
Does `nix build`'s exit code tell me anything useful? I can't find any mention of it in the docs
cosimone has joined #nixos
<{^_^}>
[nixpkgs] @yegortimoshenko pushed 3 commits to master: https://git.io/JvgzG
<{^_^}>
[nixpkgs] @yegortimoshenko pushed 2 commits to master: https://git.io/JvgzC
erictapen has quit [Quit: leaving]
<Thra11>
Ah. found it in the realise bit of the nix-store manpage: 100 == generic build failure, 101 == timeout, 102 == hash mismatch, 104 == non deterministic check. Don't think any of these can help me.
<Thra11>
LnL: Do you know of any way to programmatically tell the difference, or am I going to have to do the each build and manually mark it as good, bad or skip?
orivej has quit [Ping timeout: 260 seconds]
<manveru>
cinimod_: you ever tried haskell.nix?
<manveru>
LnL: do you perhaps have an example of using vaultenv with systemd? i've been researching a lot about the nicest approach... but the token based auth doesn't seem very suitable for nixos, maybe using certs might be better?
cosimone_ has joined #nixos
<__monty__>
cinimod_: What's the problem?
cosimone has quit [Ping timeout: 240 seconds]
dhess_ has joined #nixos
vykook has quit [Ping timeout: 258 seconds]
o1lo01ol1o has joined #nixos
<Thra11>
Is there a way to ask nix to build all foo's dependencies, but not foo itself?
<LnL>
I personally don't care as much about fixing the second level, being able to rotate secrets and revoke access on role level is already _way_ better than most other setups
<manveru>
i guess that's very true :)
<manveru>
gotta start somewhere
vykook has quit [Ping timeout: 240 seconds]
<manveru>
anyway, thanks a lot for your time, i'll do that for now
chloekek has joined #nixos
<{^_^}>
[nixpkgs] @mweinelt opened pull request #81369 → nixos/acme: apply chmod and ownership unconditionally → https://git.io/Jvgg1
<LnL>
hm, I just had an idea. I wonder if the vault agent could refresh the secret_id with secret_id_num_uses=1
<LnL>
that way everything rotates nicely but as soon as something fishy happens the cycle breaks
iyzsong- has joined #nixos
iyzsong has quit [Ping timeout: 240 seconds]
<cinimod_>
__monty__: > Does it matter you're jailbreaking everything but random there.
<dhess>
previous good rev was 57ed5db343f7883c90c8a5a6b771216690842723
vykook has joined #nixos
<dhess>
I think... I might have been having some other "expected a string got a set" or similar error with that version, also related to making ISOs/EC2 images
malSet has joined #nixos
ixxie has joined #nixos
<dhess>
ok yeah pretty sure 57ed5db343f7883c90c8a5a6b771216690842723 was the last good build
vykook has quit [Ping timeout: 265 seconds]
<dhess>
gchristensen: ok I've done a couple of deploys now, a reboot, used include/exclude, all looks good to me!
<gchristensen>
! thanks! can you say so on the PR?: )
vykook has joined #nixos
<dhess>
yeah
kleisli has joined #nixos
<excelsiora>
What do I need to know to contribute? I'm already pretty good with core Python.
mexisme_ has joined #nixos
<gchristensen>
hmm if you check out my branch from https://github.com/NixOS/nixops/pull/1232 and run nix-shell, run "mypy --report-html ./report nixops" then open ./report/index.html it'll show untyped areas
cosimone has quit [Remote host closed the connection]
<gchristensen>
add some types where you can make a good guess as to what they are, run mypy to verify, and PR? :)
cosimone has joined #nixos
mexisme_ has quit [Ping timeout: 246 seconds]
<dhess>
gchristensen: I don't see any references to disallowedReferences anywhere that would involve udev or stage-1-init
<gchristensen>
not sure then sorry :/
vykook has quit [Ping timeout: 255 seconds]
<gchristensen>
excelsiora: ^ did you see my reply?
ixxie has quit [Ping timeout: 255 seconds]
<excelsiora>
gchristensen: hm... I'll see what I can do. I suppose it has its own nix defaults, right?
<excelsiora>
nice.
<gchristensen>
excelsiora: this person is working on some of their very strict mypy rule set: https://twitter.com/TheWizardTower/status/1233751761099182080 which you could work on too, but I think just adding types where there are currently absolutely none would be a very good way to go
vykook has joined #nixos
Uniaika has joined #nixos
<Uniaika>
hello there! I'm facing quite a lot of "stack smashing" crashes on nix utilities (nix-env, nix-shell…)
<Uniaika>
Is there a way to know what could cause them?
aveltras has quit [Quit: Connection closed for inactivity]
eyJhb has joined #nixos
eyJhb has joined #nixos
eyJhb has quit [Changing host]
oida has quit [Ping timeout: 240 seconds]
oida has joined #nixos
vykook has joined #nixos
jgeerds has joined #nixos
<tokudan>
assuming I've got a function definition like this, is there an easier way to write stdenv=stdenv? defition: { stdenv, fetchFromGitHub, cc65 ? (import ./cc65.nix {stdenv = stdenv; fetchFromGitHub = fetchFromGitHub;})}:
<vaibhavsagar>
I would prefer 8.8 but I realise the LTS was just released and it might be too new to put in a NixOS stable release
<tokudan>
i tried inherit but couldn't figure out how to inherit both at the same time ^^
dansho has quit [Remote host closed the connection]
dansho has joined #nixos
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nixpkgs] @bhipple opened pull request #81375 → pueue: upgrade cargo fetcher and cargoSha256 → https://git.io/JvgVK
<__monty__>
vaibhavsagar: For more flexibility you should check out haskell.nix, the nixpkgs infra is really focused on shipping software that happens to require building with a GHC, not haskell development per se. Caveat, LTS-15 isn't supported until some changes that should land soon™.
<vaibhavsagar>
__monty__: the default version that's specified for the release also determines what's cached :)
<vaibhavsagar>
__monty__: I've been using the NixOS Haskell infrastructure for a couple of years now, I think it works great even if that isn't the intended purpose
<__monty__>
vaibhavsagar: Sure, caching's something haskell.nix currently kinda lacks. It's the future though : )
dansho has quit [Quit: Leaving]
<{^_^}>
[nixpkgs] @bhipple opened pull request #81376 → webmetro: upgrade cargo fetcher and cargoSha256 → https://git.io/JvgV9
<vaibhavsagar>
__monty__: I'll believe it when it gets bundled with nixpkgs :)
cosimone has quit [Quit: Terminated!]
vykook has quit [Ping timeout: 240 seconds]
<dhess>
Hi vaibhavsagar!
<vaibhavsagar>
Hi dhess!
o1lo01ol1o has quit [Remote host closed the connection]
<{^_^}>
[nixpkgs] @bhipple opened pull request #81379 → cargo-web: upgrade cargo fetcher and cargoSha256 → https://git.io/JvgwG
<{^_^}>
[nixops] @TheWizardTower opened pull request #1233 → Start with types → https://git.io/JvgwZ
Uniaika has left #nixos [#nixos]
phreedom has joined #nixos
o1lo01ol1o has quit [Ping timeout: 255 seconds]
<{^_^}>
[nixpkgs] @bhipple opened pull request #81380 → wasm-pack: upgrade cargo fetcher and cargoSha256 → https://git.io/Jvgwn
vykook has joined #nixos
<symphorien>
_phry: end of life linux versions are removed from nixpkgs. you should switch to 4.19 or 5.4 or latest
<_phry>
symphorien 5_4 doesn't seem to be addressable from my nixpkgs
<_phry>
Currently there's nix linuxPackages 5_whatever available when I query for it in my nix repl. And knowing my kernel version is kind of important for me, because I have to apply a patch
<symphorien>
oh latest is 5.4 but there is no 5.4
<symphorien>
strange
<_phry>
it seems to be in the nixpkgs repo, but not in the current release channel o_O
<{^_^}>
[nixpkgs] @bhipple opened pull request #81381 → rx: upgrade cargo fetcher and cargoSha256 → https://git.io/JvgwR
<{^_^}>
[nixpkgs] @marsam pushed 2 commits to release-19.09: https://git.io/Jvgrg
o1lo01ol1o has joined #nixos
xelxebar has quit [Ping timeout: 240 seconds]
vykook has quit [Ping timeout: 260 seconds]
xelxebar has joined #nixos
o1lo01ol1o has quit [Ping timeout: 258 seconds]
smolck has joined #nixos
<smolck>
Hello! Has anyone built Remacs (https://github.com/remacs/remacs) with a nix derivation? I'm trying to build the necessary derivation, but it's rustup nightly requirement has me stumped.
<smolck>
@cole-h It fails in the build phase: ```Checking Rust toolchain install ...error: could not create home directory: '/homeless-shelter/.rustup'error: caused by: Permission denied (os error 13)Remacs currently requires Rust toolchain version nightly-2020-01-08.Run 'rustup install nightly-2020-01-08'.builder for
<smolck>
'/nix/store/lc6hf8kcgkklrdb1cs9449yjviacjv97-remacs.drv' failed with exit code 1error: build of '/nix/store/lc6hf8kcgkklrdb1cs9449yjviacjv97-remacs.drv' failed(use '--show-trace' to show detailed location information)```
<smolck>
Well, almost. No permissions error but still says it needs rustup nightly
<smolck>
version 2020-01-08
<bhipple>
Ah yeah, I was only looking at fixing the permission error for $HOME :)
<smolck>
;)
<smolck>
It's all good
<smolck>
Hmm, so after putting `rustup install nightly-2020-01-08` in preconfigure with the home dir changed I get this:
<smolck>
```info: syncing channel updates for 'nightly-2020-01-08-x86_64-unknown-linux-gnu'error: could not download file from 'https://static.rust-lang.org/dist/2020-01-08/channel-rust-nightly.toml.sha256' to './.rustup/tmp/yxi80qaeual6b2zt_file'error: caused by: failed to make network requesterror: caused by:
<smolck>
https://static.rust-lang.org/dist/2020-01-08/channel-rust-nightly.toml.sha256: error trying to connect: failed to lookup address information: Name or service not knownerror: caused by: failed to lookup address information: Name or service not knownbuilder for '/nix/store/wq9cllwsck1ik14fh442yj7l779k7dsq-remacs.drv' failed with exit code 1error:
<{^_^}>
[nixpkgs] @bhipple opened pull request #81389 → cargo-generate: upgrade cargo fetcher and cargoSha256 → https://git.io/Jvgoe
<smolck>
build of '/nix/store/wq9cllwsck1ik14fh442yj7l779k7dsq-remacs.drv' failed(use '--show-trace' to show detailed location information)```
<cole-h>
I think you should be able to specify `rustToolchain = "nightly-2020-01-08";`
<bhipple>
In general, you shouldn't be running `rustup` in Nix; you should instead configure it to have the right rust compilers out of the gate
<cole-h>
Where, I have no idea.
mexisme_ has quit [Ping timeout: 272 seconds]
<smolck>
Yeah, I'd prefer not to run rustup for sure.
<bhipple>
so the fact that it's running suggests to me that it thinks it's missing something it should have, and is trying to fill the gap
vykook has joined #nixos
<smolck>
@cole-h I think I know where to do that, saw it while I was searching for how to fix this
<bhipple>
gccemacs has some pretty huge performance improvements. I'd be interested in helping get it packaged in something like the NUR emacs-overlay :D
<smolck>
Really?
<bhipple>
Of course, if you're fiddling with remacs just as a way to learn Rust for fun, gccemacs has nothing to offer
vykook has quit [Ping timeout: 258 seconds]
<smolck>
I was just looking for better performance, and was curious, so I'm willing to try gccemacs
<bhipple>
gccemacs is definitely the thing you should be looking at for performance. Also if you haven't already go to emacs27, which has libjanson support for native JSON serialization/deserialization
<bhipple>
nah as mentioned in that github issue ticket, I haven't gotten past packaging libgccjit yet, but after that the emacs portion should be trivial
phreedom has quit [Ping timeout: 240 seconds]
<{^_^}>
[nixpkgs] @bhipple opened pull request #81391 → cargo-inspect: upgrade cargo fetcher and cargoSha256 → https://git.io/Jvgog
<bhipple>
If you take a shot at it please do post on the issue; I can help push it along
<bhipple>
Need to build a package for libgccjit, yeah
<smolck>
Shouldn't be that hard in theory, right?
<NemesisD>
do you all think it would be possible to keep both my system and home-manager nix configs in a git repo in my home dir and then symlink to /etc/nixos?
<bhipple>
as in package it separately out of gcc, get to the point that we can run `nix-build -A libgccjit` and have it pass, then from there it should be trivial to make an gccemacs overlay :)
<smolck>
Couldn't we just install libgcc and extract it somehow?
<bhipple>
^ Here's how far I've gotten. You need to pass --enable-host-shared to libgcc, which impacts performance, so it's not appropriate to check it into the "main" libgcc package inside nix; it should be a sidecar package called libgccjit specifically
<bhipple>
I haven't gotten it to fully build yet, though
<smolck>
Oh lol that yours, my bad hehe
<bhipple>
:D
<smolck>
What's stopping the buid?
<bhipple>
Give it a try and see, it's been a week or two since I've poked at it
<bhipple>
The closest thing that reads like a from-first-principles bottom-up book walkthrough is the nix-pills ^
waleee-cl has joined #nixos
<excelsiora>
k, I was aware of those, but wasn't sure best place to start, so pills...
<bhipple>
nix pills for sure
LysergicDreams has joined #nixos
<smolck>
Who knew gcc takes forever to compile lol
<smolck>
Well, gccjit
<bhipple>
Assuming you're just trying to learn about Nix in general. If you're trying to install NixOS, the NixOS manual, or if you're trying to do something specific like "how do I build Rust packages in NixPkgs" (the nixpkgs manual) etc.
<bhipple>
yeah it does, hence why it's so painful to get a failure at the last minute :D
<bhipple>
I got past all the really fast up-front failures, at least
<{^_^}>
[nixpkgs] @bhipple opened pull request #81397 → rustup: upgrade cargo fetcher and cargoSha256 → https://git.io/JvgK1
vykook has quit [Ping timeout: 240 seconds]
<jonreeve[m]>
Anyone know how to add a package from GitHub?
<Uniaika>
jonreeve[m]: for GH repos that provide it, you can do something along the lines of
<jonreeve[m]>
I made a shell.nix file for my Haskell project using `cabal2nix --shell . > shell.nix` but now I want to use a version of a haskell package from GitHub which isn't yet on Hackage
<clever>
cinimod_: on line 21, random is getting its deps from that 2nd nixpkgs, and ignoring this overlay
<clever>
cinimod_: you want to override the main pkgs.cabal2nix, to sever the link and make hsuper.callCabal2nix not depend on random
<clever>
cinimod_: you can also do that by just not overriding pkgs.haskellPackages, rename the attr on line 3, and cabal2nix wont notice your overlay entirely
<drew__b>
@cole-h does this use your system cabal?
<drew__b>
or does it install cabal with the project?
<abathur>
has anyone run into needing to pass PS1 through a wrapper script? it looks like by default bash shebangs (at least, perhaps other shells) eat PS1 if they aren't interactive
<cole-h>
drew__b: I don't have anything Haskell installed globally, so what you see there is what I use for Haskell
<abathur>
or, I could rephrase, *since they aren't interactive*, they eat PS1, but if I fabricate a wrapper with a bash shebang using --noprofile --norc -e -i, it'll let PS1 through
<cole-h>
Because you might not have updated your channels recently? If you're using unstable, `haskell-updates` to LTS 15.1 landed yesterday
<drew__b>
they provide shell.nix and other tooling files
<drew__b>
and I can't get it to work with hie
<infinisil>
drew__b: I can tell you that the reason https://pastebin.com/KPnfXf2r doesn't work is because environment.systemPackages is a NixOS attribute. In a random shell.nix it doesn't make any sense
<drew__b>
i see
<drew__b>
any plans on adding instructions to the README similar to what @cole-h has?
<drew__b>
i.e. installation for non nixOS users?
<clever>
cinimod_: line 11, you forgot to callCabal2nix the new random
<clever>
cinimod_: line 25-29, you want ps.random and friends
<clever>
cinimod_: line 36, you want myHaskellPackages
<infinisil>
drew__b: Yeah maybe, I've wanted something similar for a while
<drew__b>
that would be pretty helpful TBH
<infinisil>
drew__b: Does it work with what cole-h suggested?
<drew__b>
well it definitely enters a nix-shell
<drew__b>
and installs stuff
<drew__b>
however I cannot get it to work with that repo
<drew__b>
2. run ./support/copy-tools.sh (something like that, just autocomplete)
<drew__b>
3. Edit the shell.nix to be what cole-h has (but using hie 864)
<drew__b>
then enter a nix-shell
<drew__b>
and open the project in your editor
<drew__b>
i had to add an hie.yaml, but that still didn't work otu
elux has joined #nixos
mexisme_ has joined #nixos
<infinisil>
You haven't mentioned the error yet
<elux>
anyone having install unstable channel at the moment? im getting Error: Failed copying /nix/store/6xqkgnj1hjgbg3jvxvzl74g924vw1lzj-luajit-2.1.0-beta3-luv-1.34.1-1/luv-1.34.1-1-rocks/luv/1.34.1-1/lib/libluv.so.1 to /nix/store/6xqkgnj1hjgbg3jvxvzl74g924vw1lzj-luajit-2.1.0-beta3-luv-1.34.1-1/lib/lua/5.1/libluv.so.1
<infinisil>
(which is the main thing I'm interested in)
<drew__b>
ok let me do it again, i killed that pane
<{^_^}>
Infinisil/all-hies#32 (by Jasagredo, 15 weeks ago, open): Error in emacs: can't load .so/.DLL for libm.so.6 due to GLIBC_2.29
<drew__b>
i get that all the time
<drew__b>
usually if i haven't build the project recently
<infinisil>
I didn't have the time to investigate this error
<infinisil>
Or the motivation hah
PulBhaBa_ has joined #nixos
<drew__b>
this project might be a good chance to test it out quickly
<tokudan>
how do I include the literal _text_ ${abc} in a string in a nix expression? or in other words: how do I prevent nix from interpreting that as "insert variable abc here"?
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos
<srk>
> 'x ''${abc} y'
<{^_^}>
error: syntax error, unexpected $undefined, at (string):284:1
<cole-h>
infinisil: I don't think any non-NixOS-specific installation steps are necessary... Maybe just a `shell.nix` example. The only real difference is s/environment.systemPackages/buildInputs/
<srk>
> 'x \\${abc} y'
<{^_^}>
error: syntax error, unexpected $undefined, at (string):284:1
<tokudan>
srk, ''${ did the trick, thanks :)
<srk>
oo :) good
ddellacosta has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<srk>
> "x ''${abc} y"
<{^_^}>
undefined variable 'abc' at (string):284:8
<samueldr>
when you get deep in quotations, you may want to instead use antiquotation to use another form of escaping
<tokudan>
i usually just add backslashes till it works ;)
<samueldr>
''${"''sample nix string''"}''
<samueldr>
> ''${"''sample nix string''"}''
vykook has quit [Ping timeout: 240 seconds]
<{^_^}>
"''sample nix string''"
<samueldr>
(yes, the example is extremely contrived)
o1lo01ol1o has quit [Ping timeout: 258 seconds]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 260 seconds]
<cinimod_>
clever: I have to cook dinner - thanks - I'll try later
drew__b has quit [Quit: WeeChat 2.7.1]
o1lo01ol1o has joined #nixos
<{^_^}>
[nixops] @grahamc pushed to start-with-types « diff: fixme -> to-done: type hint on nixops.deployment.Deployment »: https://git.io/Jvgib
vykook has joined #nixos
<jonreeve[m]>
<cinimod_ "jonreeve: did you solve your pro"> Not really. I guess I just want a way to say: get me these packages from Hackage, and these from GitHub, and put them all together in a haskell development environment
<cinimod_>
jonreeve[m]: or something like the above
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
niso_ has joined #nixos
vykook has quit [Ping timeout: 260 seconds]
knupfer has joined #nixos
andymandias_ has quit [Ping timeout: 272 seconds]
andymandias has joined #nixos
<Baughn>
I made an attempt at setting -march for my platform.
<Baughn>
Then I tried just recompiling everything *without* that.
<Baughn>
But either my machine is consistently miscompiling Perl, or else there's a bug somewhere which causes a TimeDate test to fail... but not on the compiler farm.
<elux>
who broke nixos unstable channel...? builder for '/nix/store/nv873vikr2nrm02f0lmsicp1lz49yg85-luajit-2.1.0-beta3-luv-1.34.1-1.drv' failed with exit code 1
<samueldr>
elux: if it's pushed to a channel, the channel itself has tested and is valid
<samueldr>
elux: though, it can happen that dependencies that are not part of the tested subset break
<samueldr>
which I think is what is happening here
<teto>
the libluv has some flaws that we are trying to properly fix to clear the neovim issue
<balsoft>
Hi, does anybody know why does grub forcefully copyies kernels and initrd's to `/boot` if it's on another partition? Shouldn't it be able to just mount the root(`/nix/store`) partition and boot from there?
<JonReed>
How can I debug why NixOS doesn't start a graphical session in QEMU? When the machine starts I'm presented with a login console. If I login as root, I can see via `systemctl status display-manager.service` that display manager is running. Yet, I'm in console. It could be I'm missing something very basic, because I'm used to NixOS logging me into a
<JonReed>
desktop on my normal machine and in VM's I used console before.
<JonReed>
When I login as user, `systemctl --user status graphical-session.target` shows "inactive (dead)". I can't figure out how to check GPU memory in QEMU and if perhaps this is a resource issue, but I don't think it is. Running `systemctl start graphical-session.target` still leaves me in console, even though the service starts.
<JonReed>
Of course, just the moment I ask the question I realize what to try. It appears that xorg not running `echo $DISPLAY` is nothing and `ps -aux | grep xorg` shows nothing as well. Hopefully, I'll be able to figure out why.
phreedom_ has quit [Remote host closed the connection]
phreedom has joined #nixos
mexisme_ has joined #nixos
<{^_^}>
[nixops] @grahamc pushed to start-with-types « Revert "diff: fixme -> to-done: type hint on nixops.deployment.Deployment" »: https://git.io/Jvg1v
justanotheruser has quit [Ping timeout: 240 seconds]
`slikts has quit [Quit: No Ping reply in 180 seconds.]
vykook has joined #nixos
<Thra11>
How do I 'fetch' a local directory as `src`?
<cole-h>
Just use `src = insert path here;`?
`slikts has joined #nixos
`slikts has quit [Changing host]
`slikts has joined #nixos
<symphorien>
the path must not be quoted and start with / or ./
ShaRose has quit [Quit: I appear to have left for some reason.]
<Thra11>
cole-h, symphorien: Ok. Thanks. And that will refetch whenever the sources change?
<symphorien>
yes
<symphorien>
it is checked on every evaluation (which may make it slow if the directory is large)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
shibboleth has joined #nixos
vykook has quit [Ping timeout: 255 seconds]
shibboleth has quit [Excess Flood]
shibboleth has joined #nixos
mexisme_ has quit [Ping timeout: 272 seconds]
toppler has quit [Remote host closed the connection]
toppler has joined #nixos
drakonis has quit [Ping timeout: 240 seconds]
<evanjs>
infinisil: ohhh I wonder if some of my paths being gc'ed is due to things under config/new-modules being double imported. Have you had issues with e.g. new-modules/dev not being required by anything, thus being gc'ed on nix-collect-garbage?
iqubic has quit [Remote host closed the connection]
<JonReed>
If I see NixOS showing "Started X11 Server", then it means that it NixOS was on its way to booting me into the graphical session, right? NixOS boots me into console, running either `ps -aux | grep xorg` and `ps -e | grep tty` does not show xorg. It must mean that xorg was started by `display-manager.service` but was then killed for whatever
<JonReed>
reason, right?
<cole-h>
Does NixOS still dump Xorg logs to /var/log/Xorg.*.log? If so, you should check those
Guest2075611 has joined #nixos
jonten has joined #nixos
<Baughn>
balsoft: It can't assume that /nix/store is always readable by the bootloader. A symlink would only work if /boot is the same partition as everything else, and that's unusual. Especially recently, given UEFI boot prevents it.
<balsoft>
Baughn: GRUB can mount ext4/btrfs/zfs, right? Why not just mount the store partition and use that?
<JonReed>
cole-h: I can't find `/var/log/Xorg*`. I would love to find any logs. `journalctl --unit display-manager.service` shows "Started X11 Server" as one of the last entries.
orivej has quit [Ping timeout: 258 seconds]
<Baughn>
balsoft: For one thing, the default bootloader isn't grub.
konobi has quit [Quit: konobi]
<Baughn>
The kernel is tiny. You're right that this optimization could sometimes be made, but it isn't worth the extra complexity.
konobi has joined #nixos
<manveru>
JonReed: check /var/log/X.0.log ?
<Baughn>
Oh, and grub *can't* mount most ZFS filesystems. Its driver isn't up to date, so anything created recently, or upgraded, won't work.
<abathur>
I spelunked in the Nix python builder long enough to figure out how to take some of that work and collapse these down into a single wrapper.
<abathur>
It didn't, as I suspected, fix the issue at hand. Even with a single wrapper, the PS1 still isn't passing through to bin/osh.
<abathur>
I think I did figure out what's going on, though.
<abathur>
Your wrappers use a `/bin/sh` shebang. I'm guessing whatever provides your /bin/sh is provided by a non-bash shell that doesn't block PS1 (IIRC it's probably dash?). Nix patches /bin/sh to nix-installed bash (and, AFAIK, it does so because of /bin/sh portability issues like this).
<abathur>
In a bit of a pickle. I'm setting up nix-based CI for the oil shell project. Debugging a test that breaks in Nix but not existing non-nix CI. The test sets PS1, runs commands in oil's `osh` shell, and checks the output--but non-interactive bash (and thus, any Nix-built/patched wrapper) apparently unsets PS1.
<cole-h>
Wowza
<balsoft>
Baughn: This came up in the context of security. `/nix/store` is assumed to be immutable and also can be checked for correctness pretty easily. This is not the case for the copied kernels. It's simpler for an attacker to change `/boot` than to change athing in store.
<drakonis>
at least you didnt get nuked by sigyn
<cole-h>
That's the most amount of text by one person I've ever seen in IRC
<abathur>
sorry, I just thought I'd copied, and hadn't, and pasting in weechat auto-sent it all
<drakonis>
cole-h: i've seen more
<drakonis>
weechat actually asks if you want to send the batch
<abathur>
I meant to just send the latter
<cole-h>
Weechat prompts me if I paste more than 1 line, yeha
<Baughn>
balsoft: If that's your goal, then I would recommend signing your kernel. Make use of the TPM chip.
<abathur>
it didn't, I must be missing a setting
<JonReed>
weechat loves multilines
<drakonis>
weird
<balsoft>
Baughn: Thanks, I'll pass this onto the person asking :)
<Baughn>
balsoft: Changing /nix/store is pretty easy, by the way.
<evanjs>
yeah glad it does, I would've been kicked so many times if it didn't :D
<Baughn>
It's mounted read-only to prevent accidents. It most certainly isn't a security measure.
<abathur>
cole-h: ty
xeifu has joined #nixos
<balsoft>
Baughn: I know, the person asking has made some measures to prevent anyone from getting RW access to it, presumably.
<balsoft>
Baughn: anyways, thanks for the answers.
hax404 has joined #nixos
monadplus has joined #nixos
<Baughn>
balsoft: It all seems pretty odd. I'm not sure what you'd be defending against -- given root access, there's no difference in difficulty between changing the two.
<Baughn>
And if you lack root, then you can't change either.
<Baughn>
The TPM chip can defend against physical access to a degree, but there are better ways to prevent that. Like guards.
vykook has quit [Ping timeout: 240 seconds]
<balsoft>
I don't know their reasoning, maybe it's an XY problem.
<Baughn>
Maybe. Security is full of snake oil.
zupo has joined #nixos
<JonReed>
Does the default QEMU VM generated by NixOS has enough resources to run an xorg session? I suspect my mysterious xorg problems are related to some resource usage, but I can't figure out how to either check how much VRAM Qemu VM has nor set more vram. I tried giving the QEMU_OPTS `-device VGA,vgamem_mb=256`, but this makes VM not boot at all, it
<JonReed>
greets me with the screen saying "Guest has not initialised the display (yet)".
vykook has joined #nixos
<abathur>
cole-h: odd; i have it set; not sure, unless maybe Terminal.app doesn't support the escape sequences it notes
<manveru>
JonReed: always works for me without any changes... (not very helpful i know)
<JonReed>
manveru: It is very helpful, actually, because now at least I know that it should work.
<manveru>
how do you build the image?
<leons>
Agree with Baughn. The proper way to defend against any modifications, be it on /boot or elsewhere is to store your cryptographic secrets in the TPM and lock them using PCR values
<leons>
However systemd's non-deterministic nature, verifying the userland with PCRs is going to be difficult. Distributions like Alpine do a better job here
<worldofpeace>
JonReed: Use virtualisation.memorySize to control memory in your vm, and virtualisation.cores for the cores. Hope that's useful.
<JonReed>
manveru: `nix-build '<nixpkgs/nixos>' -A vm -I nixos-config=./configuration.nix`
<JonReed>
worldofpeace: Unfortunately, I already set it at 4 gigs and 8 cores :(
<manveru>
can you paste the config?
<manveru>
s/paste/share
<manveru>
i can try reproducing it, if it's not too big
<JonReed>
manveru: default.nix just wraps configuration.nix to give qemu a couple new options
fresheyeball has joined #nixos
<fresheyeball>
so I am setting up a new system with nixops on packet host
<fresheyeball>
and now it wont boot
iqubic has joined #nixos
<fresheyeball>
joining the out of band terminal, it shows it hangs on
<fresheyeball>
setting up /etc...
mehlon has quit [Quit: Connection closed for inactivity]
jonten has quit [Ping timeout: 260 seconds]
vykook has joined #nixos
<JonReed>
Ah, the real command ends up just being `nix-build` and then `/nix/store/<ouput-of-nix-build>/bin/run-nixos-vm`. The previous one I pasted was what I used before wrapping it in default.nix
monadplus has joined #nixos
<fresheyeball>
how can I debug this?
<fresheyeball>
what change could I make the would cause this to occur?
ShaRose has joined #nixos
dashkal has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
<{^_^}>
[nixpkgs] @jonringer pushed commit from @das-g to master « osmpythontools: init at 0.2.6 »: https://git.io/JvgyG
o1lo01ol1o has quit [Ping timeout: 258 seconds]
shibboleth has quit [Quit: shibboleth]
sjuberman has joined #nixos
vykook has quit [Ping timeout: 255 seconds]
<sjuberman>
Hey, I've been using nix-darwin in my Mac for a while and I've liked it a lot. So I decided to try NixOS in a VM to see how it might be better. I've hit a snag where NixOS does not allow running pre-packaged binaries to run without pathcing them first.
<sjuberman>
Is there really now way to run normal binaries in NixOS?
abathur has quit [Read error: Connection reset by peer]
<sjuberman>
I was hoping I could use Nix just to configure my setup (dotfiles/applications/window manger/etc) and the keep my normal workflow
<fionera>
you have to patch them first or compile them yourself since the linker works completely different
<fionera>
i dont know if there is any other way in nix-darwin since I only use nixos directly on my machine. but for me its that case
<symphorien>
you can use steam-run to run many unpatched binaries
<{^_^}>
[nixpkgs] @veprbl pushed 2 commits to release-20.03: https://git.io/Jvgyl
<sjuberman>
Yeah, it's really a problem when you use things like Bazel on daily basis.
<fionera>
downgrading to the kernel from 19.09 didnt help too with the missing gl context
<cole-h>
fionera: If you're comfortable with git, a `git bisect` might be useful (using a local clone of nixpkgs)
vykook has joined #nixos
<sjuberman>
I tried using steam-run and start my zsh session with it and then run Bazel but I hit problems with Git. Seems like a git/libcurl incompatability.
<fionera>
I cant really go away from unstable on the machine since I depend on it with home-manager
<{^_^}>
[nixpkgs] @NinjaTrappeur opened pull request #81415 → linux: add policy routing config flag for aarch64 → https://git.io/Jvgya
<hexa->
fionera: you could very well backport sway to an older home-manager version
shizonic111 has joined #nixos
hiroshi[m]1 has joined #nixos
<fionera>
hexa-: yeah but thats honestly too much work for me as I dont really know nix yet
<cole-h>
To be able to use `<something>` in a Nix file, I just need to have `something=$HOME/something` in my NIX_PATH, right?
<hexa->
fionera: always a first time when you need to get your hands dirty
mexisme_ has joined #nixos
<{^_^}>
[nixpkgs] @jonringer merged pull request #81192 → glances: move from python-module to application → https://git.io/Jvz8J
<{^_^}>
[nixpkgs] @jonringer pushed to master « glances: move from python-module to application »: https://git.io/JvgyM
<clever>
cole-h: you can also have a /foo in NIX_PATH, and then a /foo/something
vykook has quit [Ping timeout: 240 seconds]
<infinisil>
sjuberman: Why can't you use the bazel from nix?
<cole-h>
clever: Sorry, I don't understand. Mind giving an example?
<cole-h>
Oh, do you mean if I have `/foo` in NIX_PATH, I can access `/foo/something` via `<something>`?
<clever>
cole-h: yep
<cole-h>
OK, that's pretty cool. Thanks! clever++
<{^_^}>
clever's karma got increased to 338
<clever>
cole-h: just like #include <something.h> and `-I /foo` would work in gcc
<{^_^}>
#78046 (by symphorien, 5 weeks ago, open): nixos/snapper: automatically create the .snapshots subvolume with correct permissions
nerdmaxx has quit [Quit: Lost terminal]
linarcx has quit [Quit: WeeChat 2.7.1]
wak-work11111 has joined #nixos
<sjuberman>
infinisil: I've been trying to get it to work. I prefer using bazelisk(https://github.com/bazelbuild/bazelisk) to bootstrap bazel. The problem is that bazelisk downloads the bazel binary and that does not fly in NixOS
<clever>
energizer`: the version of nixpkgs you use, is typically set by niv, not $NIX_PATH
<cole-h>
fionera: You have to specify a good commit and a bad commit, and keep providing feedback. `git bisect good release-19.09`, `git bisect bad master`, and then start testing. If you run into a bad commit, `git bisect bad`. If it works fine, `git bisect good`
<fionera>
yeah did that but it always hangs on the first commit
<clever>
fionera: because git bisect is meant to find where it broke, and if its already fixed, it fails
ixxie has joined #nixos
pikajude has joined #nixos
<energizer`>
first off it seems to load everything into memory at once
<clever>
energizer`: git doesnt deal with large files very well, even worse if your modifying the file repeatedly
<gchristensen>
Nix isn't really intended for that
mexisme_ has joined #nixos
<energizer`>
clever: git? did you mean nix?
<clever>
energizer`: for nix, you want to use pkgs.fetchurl to download the large data file, then it wont have to hash it each time, and youll get much better performance
vykook has joined #nixos
fresheyeball has joined #nixos
<energizer`>
clever: and then i transform the file and save it as an $out
<clever>
energizer`: ive done that as well, processing multi-gig files down into a sqlite db and then into output stuff
<energizer`>
gchristensen: in theory it seems like a great fit, no?
<clever>
energizer`: nix deals with compiling the tools if the source ever changes, and re-generating intermediate data if any of the tools ever change
<clever>
energizer`: so its much simpler to manage a complex chain of transformations to the data
<fionera>
cole-h: wrong approach... I am not having a compile error. everything builds fine, but in runtime I dont have a opengl context