| 31 Aug 2025 |
 Uraraka ~ Ochaco | maybe it's to make things nix-buildable?
either way a full install with prefix will still build manually | 17:45:04 |
 lxsameer | Peri, i never tried any wasm related stuff on haskell. is it ok to use any prelude? | 17:45:13 |
| Uraraka ~ Ochaco | I presume so, everything non-network-related seemed to work for me | 17:45:35 |
| Uraraka ~ Ochaco | I wrote a reflex-dom app with it | 17:46:07 |
| Uraraka ~ Ochaco | wth a load of servant in it | 17:46:22 |
| lxsameer | cool cool. seems like i can handle my workload with it | 17:46:47 |
 maralorn | In reply to @dandart:matrix.org
maybe it's to make things nix-buildable?
either way a full install with prefix will still build manually Yes, it is to make things nix buildable. | 17:46:50 |
| maralorn | But it's still very rough around the reges. | 17:48:16 |
| maralorn | * | 17:48:26 |
| Uraraka ~ Ochaco | I kind of gave up nix-building haskell projects because I wanted to keep up with hackage versions (and especially text 2.1.2) - I had to keep jailbreaking things but the package versions in nixpkgs unstable / haskell-updates were still very behind. I presume they were up to date with various stackage snapshots but I don't use stack as I end up missing updates.
I ended up just doing it by getting only ghc and cabal through nix-shell, and then getting cabal to do everything cabal style. That way I can allow people who don't want to use nix be able to do things just as well with ghcup.
hence haven't tried to nix-build any haskell bits in a good while. | 17:52:03 |
| maralorn | Yes, nixpkgs will lag behind, especially in the last year because we get to little help maintaining it. But very often not having the bleeding edge is fine for most usecases. | 18:22:12 |
| Uraraka ~ Ochaco | that's fair enough, I have contributed fixes to it before | 18:23:22 |
| 1 Sep 2025 |
 Artem | regarding https://github.com/NixOS/nixpkgs/pull/429810#issuecomment-3242423178
it's curious that the test suite hangs on "network client and server" (the last line) Something about network configuration on hydra machines?.. | 13:46:14 |
 sterni (he/him) | either that or related to load on those machines, hard to tell | 13:47:34 |
| sterni (he/him) | I'll ask if they changed anything recently | 13:48:18 |
| sterni (he/him) | maybe someone should try with Nix, I've used Lix | 13:50:14 |
| sterni (he/him) | okay third time's the charm | 14:08:22 |
| 2 Sep 2025 |
 bglgwyng | I've patched glib package and got this message during build
> Running phase: setupCompilerEnvironmentPhase
> Build with /nix/store/j7jc5k1l70axsk8s3zi9m4d0q23w3asp-ghc-9.10.1.
> Running phase: unpackPhase
> unpacking source archive /nix/store/13z7zgx0h4cyjm9ddf7i39d7xr3q72wy-glib-0.13.12.0.tar.gz
> source root is glib-0.13.12.0
> setting SOURCE_DATE_EPOCH to timestamp 1000000000 of file "glib-0.13.12.0/System/Glib/Utils.chs"
> Running phase: patchPhase
> Running phase: compileBuildDriverPhase
> setupCompileFlags: -package-db=/build/tmp.a1u8Em0pDC/setup-package.conf.d -threaded
> [1 of 2] Compiling Main ( Setup.hs, /build/tmp.a1u8Em0pDC/Main.o )
> Setup.hs:5:1: error: [GHC-45102]
> Ambiguous module name ‘Gtk2HsSetup’.
> it was found in multiple packages:
> gtk2hs-buildtools-0.13.12.0 gtk2hs-buildtools-0.13.12.0
I think it's somewhat related to Setup.hs executing during install, but I have no clue how to inspect this more. Does anyone know this issue?
| 11:03:49 |
|  Wolfgang Walther joined the room. | 12:10:07 |
| sterni (he/him) | bglgwyng: what does patched mean here? | 12:11:25 |
| sterni (he/him) | Wolfgang Walther: hi :) | 12:11:38 |
| bglgwyng | I did glib = callCabal2Nix ./glib-x.y.z {...} thing. I'm using haskell-flake so the code is different but the result would be so. | 12:14:43 |
| sterni (he/him) | i guess you should first rule out that you have multiple gtk2hs-buildtools derivations in the build environment | 12:24:38 |
|  @aidalgol:tchncs.de joined the room. | 21:56:50 |
| 3 Sep 2025 |
|  ghpzin joined the room. | 09:31:07 |
|  Gerd Flaig joined the room. | 20:55:18 |
| 4 Sep 2025 |
|  epimonic joined the room. | 00:36:54 |
| @aidalgol:tchncs.de changed their display name from Aidan Gauland to aidalgol[m]. | 11:23:25 |
| @aidalgol:tchncs.de changed their display name from aidalgol[m] to aidalgol. | 11:24:11 |
| maralorn | Ah, yeah, users and their wishes. https://github.com/NixOS/nixpkgs/issues/190542#issuecomment-3253170725
Funny thing is I remember vividly how hard not having the newest version of anything would have annoyed me ten years ago. 😄 | 11:52:17 |
