2 Jun 2025 |
emily | (then again, I'd say the same about GCC vs. LLVM) | 12:19:35 |
teo (they/he) | The other thing is GHC supports using a different linker to what it was built with, but I'm not sure if it's worthwhile making use of that. You'd have to split it into a bindist derivation and an installed GHC derivation. It would save building GHC n times though | 12:21:31 |
sterni | emily: it's probably better to use lld over ld.gold longterm, but we'd have to check first how bad ld.bfd is nowadays | 12:43:43 |
sterni | I wonder if GHC's configure script still rejects ld.bfd… | 12:44:12 |
teo (they/he) | I'd be keen to just make it an opt-in thing for now, and then at some point in the future we can make it the default if we want | 12:49:44 |
Alex | In reply to @emilazy:matrix.org OTOH both ld.gold and ld.bfd sort of suck for Linux to be on so maybe someone should make LLD happen system-wide :P I've found mold to be even better than lld. But it doesn't (didn't?) support the weird things GHC does. :( | 18:24:42 |
alexfmpe | wth is going on with git-annex | 20:28:53 |
alexfmpe | every lts bump invalides its hash | 20:29:01 |
maralorn | Well, that’s on purpose. peti left an excellent comment there. 😄 | 20:32:36 |
maralorn |
The Hackage tarball is purposefully broken, because it's not intended to be, like, useful.
| 20:33:06 |
maralorn | Makes me smile everytime I read it. 😄 | 20:33:15 |
alexfmpe | can we, like, shame upstream into doing something different? | 20:54:44 |
sterni | hm did wolfgang talk to anyone before the warnIf change? | 21:28:42 |
sterni | the PR is pretty much what I wanted to do myself, but it's a pretty self merge in any case | 21:29:05 |
sterni | maybe he has telepathic abilities :) | 21:29:13 |
maralorn | I know of nothing. 😄 | 21:41:40 |
maralorn | Sooo … what would an empty queue roughly look like? I guess 50k builds waiting for staging-next is not it? 😄 | 22:44:26 |
sterni | that's a question | 22:45:17 |
sterni | I mean we can probably eval, we share most GHCs with staging-next | 22:45:35 |
sterni | if it's a problem someone can always cancel an eval | 22:45:43 |
4 Jun 2025 |
| la .varik. .VALefor. changed their profile picture. | 00:33:34 |
| Alp Mestanogullari joined the room. | 09:47:39 |
sterni | shenanigans… https://github.com/commercialhaskell/stackage/issues/7768 | 12:07:20 |
sterni | over https://github.com/NixOS/nixpkgs/pull/413885 | 12:19:57 |
maralorn | sterni: Would you prefer all haskell-updates PRs to have that prefix? | 12:21:03 |
sterni | it doesn't really matter, I think it is more useful for other reviewers to immediately see that they can ignore it | 12:21:42 |
sterni | though I think other commiters just ignore Haskell PRs based on the label as a general rule, so probably doesn't matter | 12:22:01 |
sterni | you don't need it for searching since base:haskell-updates works for that | 12:22:16 |
sterni | Alex: https://hydra.nixos.org/build/299142554/nixlog/48 this is the race condition | 12:35:29 |