| 9 May 2026 |
K900 | OK I see https://github.com/NixOS/nixpkgs/pull/513745 | 07:50:47 |
K900 | OK pushed that | 08:17:14 |
leona | i'm not sure if that's actually the correct fix or whether we should integrate our fetcher somehow differently without the scripts feature. but don't really have knowledge about yarn. cc Yureka (she/her) | 08:19:00 |
K900 | OK took some wiggling but I've rebased it and I'm going to merge it because it's better than nothing | 08:20:41 |
K900 | We can refactor the fetcher after branchoff | 08:20:56 |
leona | right | 08:21:16 |
leona | i fear that every new package would be broken with the current state of the fetcher | 08:21:45 |
K900 | I don't think so | 08:22:11 |
K900 | That PR doesn't change every single package that uses yarn-berry_4 | 08:22:29 |
K900 | At least | 08:22:30 |
K900 | I'm not sure why some are affected and some aren't | 08:22:49 |
Yureka (she/her) | I don't have the context | 08:56:26 |
Yureka (she/her) | fetchYarnBerryDeps does not use yarn | 08:56:36 |
Yureka (she/her) | so it is odd that the FOD hashes change | 08:56:43 |
K900 | Yarn 4.14 seems to require lockfile changes for some packages | 08:56:50 |
Yureka (she/her) | Ah | 08:56:55 |
Yureka (she/her) | If the lockfile changes, the FOD hash changes | 08:56:59 |
K900 | Yeah | 08:57:03 |
Yureka (she/her) | because the lockfile is copied to the FOD | 08:57:06 |
K900 | But it's some and not all | 08:57:10 |
K900 | It's weird | 08:57:12 |
Yureka (she/her) | but this does not require any change on the fetchYarnBerryDeps side | 08:57:18 |
Yureka (she/her) | so I'd say it's fine | 08:57:25 |
leona | no, the fod hash didn't change | 08:57:33 |
K900 | Yeah but they patched a bunch of lockfiles and now do we have to do that for everything | 08:57:37 |
leona | the build just failed | 08:57:39 |
Yureka (she/her) | do you have a reproducer I can look at? | 08:57:49 |
K900 | OK wait two different issues | 08:57:54 |
K900 | 4.14 changes FOD hashes | 08:58:00 |
K900 | 4.13 just locks up | 08:58:03 |