| 23 Jan 2026 |
 alexfmpe | Well actually I'm not sure if shellFor handles the sandwhich situation right | 16:20:41 |
| alexfmpe | If it doesn't, I expect it can be fixed. I've some shitty nix that does that for obelisk projects | 16:21:14 |
| alexfmpe | Context: https://www.well-typed.com/blog/2024/07/hls-multi/#the-need-for-cabal-install-3.12 | 16:21:51 |
 Acid Bong | note that cabal-install must also be in the same Haskell closure to be able to see other libs | 16:52:46 |
| 24 Jan 2026 |
 sterni (he/him) | maralorn: seems to me that we could solve the recent separate bin outputs debacle by moving the haddock docs to devdoc, so they wouldn't be installed on most people's machines. | 16:21:33 |
 maralorn | sterni: What is the recent debacle? | 16:37:32 |
 andromeda | hs-rgfw> Preprocessing executable 'hs-rgfw' for hs-rgfw-0.1.0...
hs-rgfw> Building executable 'hs-rgfw' for hs-rgfw-0.1.0...
hs-rgfw> [1 of 1] Compiling Main ( src/Main.hs, dist/build/hs-rgfw/hs-rgfw-tmp/Main.o )
hs-rgfw> [2 of 2] Linking dist/build/hs-rgfw/hs-rgfw
hs-rgfw> /nix/store/s3w5m3spa1g71hx0yb82lvk6394j3w5j-stdenv-linux/setup: line 1758: date: command not found
I'm getting the above error when running nix build on my project. I suspect it's because I'm overriding preBuild, but I don't know where to look to confirm that or how I could rectify that. | 17:27:48 |
| maralorn | Maybe add core-utils as a dep? | 17:45:44 |
| andromeda | wait I might just be doing PATH wrong, lemme check | 17:48:31 |
| andromeda | yeah shit | 17:48:50 |
| andromeda | I escaped the $ in PATH=${hs-bindgen-cli}/bin:$PATH | 17:49:24 |
| andromeda | in a Rust project recently I had something like PATH=something/bin:\$PATH and I just remembered doing that xD | 18:00:20 |
| andromeda | how do I build the docs for a haskellPackages.callPackage packageDerivation {} sort of thing? | 18:22:38 |
| andromeda | nix build .#default.doc. Why does that work? | 18:24:56 |
| Acid Bong | haskell builder always builds docs by default | 18:29:26 |
| andromeda | my result only has lib bin and nix-support | 18:34:38 |
| andromeda | but when I build default.doc result-doc contains share with the docs inside | 18:35:09 |
| Acid Bong | yes, because doc is a separate input | 20:01:51 |
| Acid Bong | that's why it goes into the separate gcroot | 20:02:14 |
| sterni (he/him) | well not debacle, but I remember people complaining that cachix docs were installed | 20:20:30 |
| andromeda | In reply to @acidbong:envs.net
yes, because doc is a separate input what do you mean 'a seperate input?' To which function? can I add .doc to any flake output to build docs if the package has them? | 21:35:29 |
| andromeda | I don't understand what the .something in #output.something actuaally represents | 21:37:24 |
| Acid Bong | In reply to @andromeda:tchncs.de
what do you mean 'a seperate input?' To which function? can I add .doc to any flake output to build docs if the package has them? i meant output, my bad | 21:38:55 |
| Acid Bong | In reply to @andromeda:tchncs.de
I don't understand what the .something in #output.something actuaally represents https://nixos.org/manual/nixpkgs/unstable/#chap-multiple-output | 21:39:01 |
| 26 Jan 2026 |
 LordMZTE | I have a shellFor with withHoogle = true. When I run hoogle using hoogle serve, I have the issue that all link to haddoc are file:// URLs which my browser won't let me click because "security" and also prevent JS from working on those haddoc pages. Is it possible to make hoogle perhaps serve these files and link to that?
| 12:00:00 |
| LordMZTE | Ah whoops there's --local, must've missed that when I first skimmed --help.
| 12:01:06 |
|  ciznia joined the room. | 12:03:40 |
| maralorn | sterni: I noticed, that nixpkgs knows more SPDX identifiers than cabal. Do you think we could extend cabal2nix to parse identifiers based on availability in nixpkgs? | 12:25:18 |
| sterni (he/him) | We can’t parse SPDX identifiers Cabal doesn’t know, no. At least not without maintaining a downstream SPDX parser | 13:22:45 |
| maralorn | But the parser would simply need to know the list of known SPDX identifiers in nixpkgs, right? | 14:35:25 |
