| 5 Apr 2025 |
 hellwolf | I am stuck here atm, building HLS for GHC912 | 17:44:03 |
| hellwolf | something wrong with cabal2nix? | 17:48:23 |
| hellwolf | * could it be something wrong with cabal2nix for this package? | 17:48:29 |
| hellwolf | # A given major version of ghc-exactprint only supports one version of GHC.
ghc-exactprint = doDistribute self.ghc-exactprint_1_9_0_0;
ghc-exactprint_1_9_0_0 = addBuildDepends [
self.Diff
self.extra
self.ghc-paths
self.silently
self.syb
self.HUnit
] super.ghc-exactprint_1_9_0_0;
This is what's written in configuration-ghc-9.10.x
| 17:55:24 |
| hellwolf | alright, I copy this over but for 1_12_0_0 for ghc912, now it compiles. | 18:06:37 |
| hellwolf | but now I am getting a GHC crash when building generic-lens. GREAT. | 18:06:50 |
 maralorn | cabal2nix only creates derivation stubs if there is no valid build plan for our default ghc. | 18:20:10 |
| maralorn | That sounds hairy. | 18:20:26 |
| hellwolf | I got HLS built now, with a few of dontChecks and disabled plugins mostly related to ghc-parser (used by hlint, formolu, ormolu, etc.)
Let me clean it up a bit and send a PR. While I am at it, maybe i should help with 9.8 too.. | 18:34:29 |
| hellwolf | $ ./result/bin/haskell-language-server-wrapper --version
haskell-language-server version: 2.10.0.0 (GHC: 9.12.2) (PATH: /nix/store/21vsy90gw6p5gc3rmzxgpvzg7kcrrghs-haskell-language-server-2.10.0.0/bin/haskell-language-server-wrapper)
| 18:35:29 |
| hellwolf | here you go : https://github.com/NixOS/nixpkgs/pull/396423/files | 20:56:22 |
| 6 Apr 2025 |
| hellwolf | checking https://github.com/cdepillabout/nix-haskell-updates-status
Which of the "Maintained Linux packages with build failure" I could hrlp? | 06:27:30 |
| hellwolf | * checking https://github.com/cdepillabout/nix-haskell-updates-status
Which of the "Maintained Linux packages with build failure" I could help with?
| 06:27:42 |
| maralorn | hellwolf: That actually looks pretty finished. | 09:26:49 |
| maralorn | I will have to fix the nix-output-monitor build. | 09:26:59 |
| maralorn | The upstreaStackHpackVersion needs action. | 09:28:28 |
| maralorn | The others are not really blockers, imo. | 09:28:43 |
| maralorn | We should ping the maintainers again. | 09:28:54 |
| maralorn | Maybe there is still something important in the unfinished builds. | 09:31:21 |
| maralorn | alexfmpe: I get the feeling the cabal-add "fix" broke it on older ghcs. 😄 | 09:32:24 |
| maralorn | https://hydra.nixos.org/build/294449074/nixlog/91/tail | 09:32:37 |
| maralorn | sterni: Are you aware of anything major which needs to happen before we wrap the current PR? (Aside from the regular due diligence.) | 09:33:44 |
 sterni (he/him) | yes needs to be conditional | 13:05:36 |
| sterni (he/him) | I think lower GHC versions are pretty broken I think HLS needs to be redone for a couple of them essentially | 13:05:54 |
| sterni (he/him) | I mean just look what's broken I haven't really checked since the new bump introduced a million regressions | 13:06:26 |
| maralorn | Yeah, no problem. Will do that. I was just wondering whether you have anything else cooking which is blocking a merge. | 13:07:46 |
| sterni (he/him) | no I'm just cooking myself at the moment (cycling) | 13:37:58 |
| maralorn | In reply to @sternenseemann:systemli.org
no I'm just cooking myself at the moment (cycling) Excellent, you do that! | 13:38:54 |
 alexfmpe | oops | 15:02:30 |
| alexfmpe | the first two (coinor-clp, fft) built for me locally yesterday, might be flaky tests I dunno | 15:03:50 |
