| 18 Oct 2024 |
 fgaz | * Apparently pkgsCross.ghcjs.haskell.packages.ghc98.ghcjs-base is null, despite it being defined in hackage-packages.nix. And so is haskellPackages.ghcjs-base. | 12:40:42 |
 alexfmpe | Is ghcjs-base even a thing on new js backend? | 12:43:23 |
| alexfmpe | Wasn't that bundled with old ghcjs? | 12:44:23 |
 maralorn | In reply to @fgaz:matrix.org
Apparently pkgsCross.ghcjs.haskell.packages.ghc98.ghcjs-base is null, despite it being defined in hackage-packages.nix. And so is haskellPackages.ghcjs-base. Packages which are bundled with ghc get set to null in the configuration-ghc-x.y.nix file. Because they usually get compiled within and are provided from within the ghc build. | 12:46:23 |
| maralorn | The fact that we even bother to set it to null makes me think that it still plays a role on the new backend, but I could be wrong. | 12:47:04 |
| maralorn | Maybe that is simply a mistake and ghcjs-base does not exist for the new backend. Maybe you can find the person to ask by a blame on the override which sets it to null. | 12:48:04 |
| fgaz | I don't think ghcjs-base is bundled in ghc. ghcjs-prim is. | 12:48:34 |
| maralorn | huh, I can’t find that override. | 12:48:55 |
| fgaz | same, that's why I'm confused | 12:49:21 |
| maralorn | But still there are overrides for ghcjs-base in the ghcjs-9.x.nix file. | 12:49:27 |
| maralorn | By alexfmpe 😄 | 12:50:00 |
| maralorn | You have to know what’s up, you did it. 😄 | 12:50:12 |
| fgaz | oh but i'm looking in a newer checkout | 12:50:16 |
| fgaz | aha, looks like it was recently fixed. it didn't get to the channels yet | 12:50:44 |
| fgaz | At least I think it was fixed... I'll know after I build ghcjs... again | 12:51:51 |
| fgaz | * oh but i'm looking at a newer checkout, while before I was trying to build with the latest nixpkgs-unstable | 12:54:51 |
| fgaz | * oh but i'm looking at a newer checkout, while before I was trying to build with the latest nixpkgs-unstable edit: nixos-23.05. NIX_PATH mistakes were made ^_^" | 13:42:12 |
| fgaz | * aha, looks like it was recently fixed. it didn't get to the channels yet | 13:42:40 |
| fgaz | * oh but i'm looking at a newer checkout, while before I was trying to build with the latest nixpkgs-unstable edit: nixos-24.05. NIX_PATH mistakes were made ^_^" | 13:42:55 |
| alexfmpe | In reply to @maralorn:maralorn.de
You have to know what’s up, you did it. 😄 That whole PR spree turned into a blur in my memory by now | 14:46:16 |
| maralorn | completely understandable | 15:20:15 |
 sterni | Alex: https://github.com/NixOS/nixpkgs/pull/349570 | 16:19:00 |
| sterni | I still don't get what's the best story for handling extensions | 16:21:33 |
| sterni | providing the latest Cabal version is sure to get you in trouble with inconsistent deps, right? | 16:21:58 |
| 19 Oct 2024 |
| maralorn | Hey everyone in this channel,
as you might now we have a group called the Haskell maintainers team, which keeps the lights on around here. This is on the one hand awesome for organisational purposes on the other hand I don’t like it because it creates an arbitrary distinction to some of you other contributors who do equally important work.
Anyway the main tasks of the haskell team is opening the haskell-update PRs, herding hydra, and merging those PRs, which we do in a bi-weekly rotation. Additionally we also do most of the reviews and generally fix stuff, recently we had a bit of capacity loss and by now basically only sterni and me are part of that bi-weekly rotation. This is kinda bad, because it binds too much of our capacity, the frequency of merges went done significantly because we are nearly saturated just with dealing with PRs. It would help us tremendously if we could again widen the team by one or two people who take on a small bit of responsibility, the core of which is to take responsibility of the PR every few weeks for 2 weeks. It is not actually that much work, on average just one hour per week would be great support. This role is a bit of a thankless job, but is essential work for the Haskell community.
So if anyone with a bit of experience of fixing stuff in haskellPackages would be interested in supporting us, please reach out. Also don’t hesitate to ask if you have any questions.
| 10:15:44 |
| 20 Oct 2024 |
|  fredrikr set a profile picture. | 08:09:11 |
| fredrikr changed their display name from fredrikr79 to fredrikr. | 08:09:23 |
 tfc | maralorn: how much work is that roughly in hours per week? | 09:32:50 |
| maralorn | In reply to @tfc:matrix.org
maralorn: how much work is that roughly in hours per week? It’s really hard to say. As I wrote, on average one per week seems about right. Biggest problem is, that it is a bit unpredictable when you can do the work. There are a lot of roundtriptimes with hydra and complete blockers because of set rebuilds. Thus this is not something which you can concentrate on for 2 consecutive hours every two weeks on Sunday. It’s more something which you do in small gaps distributed over the week. Of course how much work you invest can also be scaled by you. e.g. you could just not review contributed PRs at all and decide that you personally don’t care about any packages and just set the deadline for other people to fix them. i.e. the core responsibility is not that big and would already help us. | 10:02:03 |
| maralorn | The other way around anyone who can do reviews of PRs into haskell-updates would be appreciated. There is actually a fairly simple checklist to go through for a review. And as long as you stay friendly you can’t do much wrong with it. | 10:03:51 |
