| 17 Sep 2025 |
 sterni (he/him) | cabal doesn't allow declaring dependencies on non-Haskell build tools iirc | 16:08:03 |
| sterni (he/him) | the idea behind configuration-nix was that you would apply this configuration to your package set even if you are building a downstream package set, but I'm not sure if this actually works out in practice. Most likely too many assumptions in there. | 16:09:22 |
| sterni (he/him) | sounds like a nice idea. I think amjoseph tried it before and ran into a lot of problems, but maybe he tried to get powerpc64le-linux to work and ran into some unrelated issues | 16:10:24 |
| sterni (he/him) | an alternative/complementary approach to this is "postprocessing" in cabal2nix which can almost arbitrarily modify the generated package expressions. We are trying to slowly remove that feature though as it is a bit of a nightmare to maintain | 16:12:31 |
| 18 Sep 2025 |
 iqubic (she/her) | So, I'm looking to use XMobar on my NixOS system. I'm already quite familiar with Haskell and the XMobar config style. The main issue I'm running into is that the ALSA module says "to activate this plugin you must pass the with_alsa flag during compilation." How do I do that? | 19:16:21 |
| iqubic (she/her) | I'm enabling XMobar via Home Manager and it gives me an option for "programs.xmobar.package" which defaults to "pkgs.haskellPackages.xmobar" | 19:17:19 |
| iqubic (she/her) | So, my question is... how do I override a Haskell package like this? | 19:17:33 |
| iqubic (she/her) | Nevermind... I'm a moron. Looking at the code in Nixpkgs, with_alsa is a default flag for this package. | 19:20:34 |
| 19 Sep 2025 |
 bglgwyng | I found that plan.josn doesn't contain .cabal's default-extensions information. Doesn't it indicate that it's not possible to reproduce cabal build only with plan.json + ghc? | 08:22:36 |
| sterni (he/him) | iqubic (she/her): otherwise haskell.lib.compose.enableCabalFlag would be the answer. | 10:21:39 |
| sterni (he/him) | bglgwyng: yes and there are also other tasks that Cabal needs to do which is required for some packages to work | 10:23:00 |
| sterni (he/him) | e.g. autogen-modules. I don't know if there are even more basic things Cabal needs to do, not really familiar with the details | 10:23:45 |
| bglgwyng | it seems generic-builder runs Setup.hs to build cabal packages. | 11:00:03 |
| bglgwyng | Do we have modificatino related to ghcWithPacakges between nixpkgs/nixos-25.05 and nixpkgs/HEAD? With nixos-25.05, I found ghcWithPackages builds two random packages with the same version, but with HEAD, it doesnt'. | 11:09:43 |
| bglgwyng | * Do we have modificatino related to ghcWithPacakges between nixpkgs/nixos-25.05 and nixpkgs/HEAD? With nixos-25.05, I found ghcWithPackages builds two random packages with the same version, but with HEAD, it builds only one. | 11:11:42 |
 maralorn | I mean it’s generally much more likely that this is to due to differences in a specific package. But I remember vaguely that we merged something related to ghcWithPackages in the last months. | 11:12:26 |
| bglgwyng | Could you share the commit ref please? I want to compare before and after. | 11:13:43 |
| maralorn | Huh, er maybe we didn’t merge it and it was just a PR. | 11:13:50 |
| bglgwyng | Hm... then now I am going to start binary search over then lol | 11:14:20 |
| maralorn | Huh, I can’t find anything that makes sense. | 11:16:13 |
| maralorn | Could be d3e4f784aa1bb95ad8a13cadc7073e6617658f2f, but seems highly unlikely. | 11:16:32 |
| maralorn | Good luck with that. I find it somehow very satisfying to do that even though it takes much time. It’s so cool when the power of the logarithm works for you. 😄 | 11:17:36 |
| sterni (he/him) | In reply to @bglgwyng:matrix.org
it seems generic-builder runs Setup.hs to build cabal packages. yes, that’s Cabal | 13:12:37 |
 Wolfgang Walther | @sterni @maralorn it looks like the next staging-next cycle will start this weekend. Do you think we're close enough with the current iteration to catch that? I think it would make sense to still give this major update a bit of exposure on unstable, before we hit the freeze. We will certainly have at least one more staging cycle after that, but until that hits master, we're really close to release already. WDYT? | 18:38:49 |
| Wolfgang Walther | * sterni maralorn it looks like the next staging-next cycle will start this weekend. Do you think we're close enough with the current iteration to catch that? I think it would make sense to still give this major update a bit of exposure on unstable, before we hit the freeze. We will certainly have at least one more staging cycle after that, but until that hits master, we're really close to release already. WDYT? | 18:39:16 |
| sterni (he/him) | I feel like it’s possible yeah, especially since we can still fix stuff on staging | 18:42:37 |
| sterni (he/him) | should maybe look at the broken list as of right now | 18:42:57 |
| Wolfgang Walther | Are you going to give the new blazing fast script a run or should I? | 18:43:37 |
| sterni (he/him) | feel free to I’m obly on the phone at the moment | 18:44:27 |
| sterni (he/him) | * feel free to I’m only on the phone at the moment | 18:44:34 |
