31 May 2025 |
maralorn | @peterbecich:matrix.org thanks for the fix. However this breakage only happened because there was an accidental merge of staging into haskell-updates. We decided to resume work on the branch once that staging reaches master. Then we can merge master into haskell-updates. | 20:09:47 |
maralorn | * @peterbecich:matrix.org thanks for the fix. However this breakage only happened because there was an accidental merge of staging into haskell-updates. We decided to resume work on the branch once that staging reaches master. Then we can merge master into haskell-updates. Receiving all the fixes. | 20:10:41 |
1 Jun 2025 |
sterni | https://github.com/NixOS/nixpkgs/pull/413046 | 21:38:25 |
sterni | We need to resolve eval issues and can look into starting an eval tomorrow or in the next days, depending on how full the x86_64-linux queue is. | 21:49:12 |
| NixOS Moderation Bot unbanned @mightyiam:matrix.org. | 22:58:02 |
| Sean Thawe joined the room. | 23:39:39 |
2 Jun 2025 |
maralorn | Uh, why is pandoc pinned to 3.6 in stackage? | 11:04:25 |
maralorn | We have a test error for pandoc on haskell-updates and it is fixed in 3.6.4. | 11:04:48 |
maralorn | But apparently we are already backporting 2 patches for that one test … Lets make it 3. 😄 | 11:08:38 |
sterni | maybe newer pandoc no longer has a valid install plan with stackage lts | 11:58:31 |
maralorn | Yes, that is likely it. | 12:07:32 |
Teo (he/him) | Is there an easy way to configure the nixpkgs haskell infra to use lld as the linker? I'm currently working on adding support to haskell.nix for this and once I'm done with that I'd be keen to do the same for nixpkgs if it doesn't already exist | 12:08:39 |
sterni | probably just need to generalize the ld.gold mechanism, the question is of course why you'd do it. | 12:10:22 |
sterni | it's already kind of silly to use gold as a one off in haskellPackages | 12:11:10 |
Teo (he/him) | Basically cause gold is being deprecated and also lld is better apparently | 12:11:23 |
Teo (he/him) | Yeah I think it's kind of a conceptual mess how this all works | 12:11:33 |