| 7 Sep 2025 |
 emily | and yeah, having to keep Hugs working is part of the burden here... | 21:44:03 |
| emily | as C codebases unmaintained for 20 years are no fun to keep going | 21:44:20 |
 Alex | In reply to @emilazy:matrix.org
well AIUI the upstream build system either compiles the C blob or uses GHC right? Not necessarily: https://github.com/augustss/MicroHs?tab=readme-ov-file#bootstrapping-with-hugs
(I also delete the C blobs to be extra sure I don't accidentally use them.) | 21:45:29 |
| emily | ah. but that's a separate branch of MicroHs? | 21:46:21 |
| emily | seems like it'd be ideal for it to be merged with the main one but I guess worst case it's just another hop | 21:46:43 |
| Alex | Yes, it's another hop, but the MicroHs build is under 5min anyway, so I'm not concerned.
It's a good idea to build a stage 2 mhs anyway. | 21:53:37 |
| emily | I do worry a bit about the part where he says the resulting binaries are 10x slower than GHC. | 21:55:36 |
| emily | it's not exactly fast to bootstrap GHC to begin with | 21:55:49 |
| emily | but I guess targeted optimization work could be applied | 21:56:08 |
| 8 Sep 2025 |
|  romildo joined the room. | 21:04:44 |
| romildo | I want to install IHaskell with access to the System.Random package, as Jupyter kernel (on NixOS unstable). I have tried the following, without success:
$ nix-shell -p python313Packages.jupyter-client '(ihaskell.override { packages = (ps: with ps; [ random ]); })'
$ ihaskell install
$ jupyter kernelspec list
Available kernels:
python3 /nix/store/452p7spwcycp694vd791hw2c2a2kl5c5-ihaskell-with-packages/lib/python3.13/site-packages/ipykernel/resources
haskell /home/romildo/.local/share/jupyter/kernels/haskell
$ jupyter console --kernel haskell
Jupyter console 6.6.3
IHaskell 0.11.0.0 GHC 9.8.4
In [1]: import System.Random
...:
<interactive>:1:1: error: [GHC-87110]
Could not find module ‘System.Random’.
Use -v to see a list of the files searched for.
In [2]:
Can anyone help me with this?
| 21:16:33 |
| romildo | * I want to install IHaskell with access to the System.Random package, as a Jupyter kernel (on NixOS unstable). I have tried the following, without success:
$ nix-shell -p python313Packages.jupyter-client '(ihaskell.override { packages = (ps: with ps; [ random ]); })'
$ ihaskell install
$ jupyter kernelspec list
Available kernels:
python3 /nix/store/452p7spwcycp694vd791hw2c2a2kl5c5-ihaskell-with-packages/lib/python3.13/site-packages/ipykernel/resources
haskell /home/romildo/.local/share/jupyter/kernels/haskell
$ jupyter console --kernel haskell
Jupyter console 6.6.3
IHaskell 0.11.0.0 GHC 9.8.4
In [1]: import System.Random
...:
<interactive>:1:1: error: [GHC-87110]
Could not find module ‘System.Random’.
Use -v to see a list of the files searched for.
In [2]:
Can anyone help me with this?
| 21:16:47 |
| 9 Sep 2025 |
 maralorn | Isn't random a core package? | 06:46:21 |
|  @julm:matrix.org joined the room. | 11:06:06 |
| @julm:matrix.org left the room. | 11:06:49 |
 bglgwyng | Is ${pkgs.cabal-install}/bin/cabal wrapped to use the specific ghc binary? I just read the script and I don't think so. However, I have considered it working as expected, using package db configured in pkgs.haskellPackages.ghc, each time I run cabal build things in the shell. | 14:52:33 |
| bglgwyng | hmm... does it just use ghc in PATH? And it has been ok since I always put pkgs.ghc together with pkgs.cabal-install in the shell? | 14:55:19 |
| maralorn | It uses the PATH, pretty sure. | 15:35:10 |
 Artem | not currently; judging by the master table, https://gitlab.haskell.org/ghc/ghc/-/wikis/commentary/libraries/version-history, it was on and off untill GHC 8.4.4 when it was dropped and never picked up since. This sounds funny of course. | 16:37:18 |
| Artem | * not currently; judging by the master table, https://gitlab.haskell.org/ghc/ghc/-/wikis/commentary/libraries/version-history, it was on and off untill GHC 8.4.3 when it was dropped and never picked up since. This sounds funny of course. | 16:38:10 |
| maralorn | Artem: Seems a bit … random. 🤪 | 18:04:21 |
|  Magnolia Mayhem changed their profile picture. | 22:16:18 |
| Magnolia Mayhem changed their profile picture. | 22:17:57 |
| 10 Sep 2025 |
 MangoIV | Heya
Will yall repost the maintainer summary thingy on the Haskell-updates PR? I think that would be nice, to have an update every now and then without having to sift through the jobset
| 12:50:58 |
| maralorn | We sometimes do. But are you aware of https://github.com/cdepillabout/nix-haskell-updates-status ? | 13:15:41 |
| MangoIV | No that’s perfect. Thanks. | 13:16:13 |
| Artem | maybe this link is worth adding to the (ever growing) list of links in the Matrix room description?.. | 13:26:33 |
 chreekat | Why does haskell98 show up under "Top 50 broken packages, sorted by number of reverse dependencies", but can't be found in any of the sections above it? | 13:41:55 |
| Artem | i was asking the same question myself. Worth noting that many of those "top 50" are completely dead packages. So this list isn't very useful I think. It'd be great to implement the "dead" category besides the "broken" category, as discussed on the meeting (https://github.com/NixOS/nixpkgs/pull/429810#issuecomment-3224849043 called "removed" there) | 13:46:02 |
| emily | there's an open PR for that | 13:54:47 |
