| 2 Jun 2021 |
 hexa | right, I can read backlog in this modern messenger thingy, this is great 😀 | 19:41:42 |
 toonn | hexa: TLDR, hoping the finish the LLVM bump soon so I can get back to the SDK bump. | 19:47:28 |
 LnL | toonn: I'm not sure what's wrong evaluation should either just fail or finish | 20:25:50 |
| toonn | LnL: I think it's because of the failed ofborg-eval? | 20:26:45 |
| LnL | that should only have influence on the tested job not the evaluation overall | 20:27:32 |
| toonn | With this output, https://gist.github.com/GrahamcOfBorg/90d0912d6a8323db383b5b026e63fa5b | 20:27:37 |
| toonn | Oh? | 20:27:45 |
| LnL | but maybe, something to fix either way | 20:27:55 |
| toonn | I'm really confused about it tbh. | 20:28:02 |
| toonn | I didn't touch that attribute. LLVM 11 still seems to define it and it's still in allpackages.nix | 20:28:18 |
| toonn | So how can it be missing? | 20:28:26 |
| LnL | maybe something changed between the llvm7 and llvm11 definitions? | 20:28:48 |
| toonn | Ah, it did disappear from llvm/11/default.nix | 20:32:11 |
| toonn | Missed it because it does appear in llvm/11/llvm/default.nix | 20:32:25 |
| LnL | yeah the llvm expression have a lot of duplication | 20:33:42 |
| LnL | which means inconsistencies, but that's also what makes teaking versions easy since it doesn't impact all of darwin | 20:35:15 |
| LnL | kind of a double edged sword, not sure if the gcc approach is better | 20:35:59 |
| toonn | Hmm, ominous comments suggesting the manpages attributes were commented out on purpose. | 20:44:13 |
| LnL | there where issues with overriding the man output with a different package to avoid rebuilds at some point | 20:49:22 |
| LnL | that's why there's a separate manpages package, if that completely doesn't work anymore it should be removed | 20:50:02 |
| toonn | https://github.com/NixOS/nixpkgs/pull/125184 Actually applies similar fixes as what I just did for llvm-manpages. | 20:57:11 |
| toonn | It should be merged to master soon. | 20:57:24 |
| toonn | I'm not sure whether rebasing on top of that is ok though? Should I just commit the fixes seperately? | 20:58:08 |
| LnL | the more you change compared to that good eval commit the more false positives might show up in the diff | 21:02:58 |
| toonn | LnL: Looks like the eval is progressing now. | 21:57:36 |
| toonn | Not sure what to do with the cherry-pick commit. | 21:57:55 |
| toonn | Do you take notes of commits to remove when merging? | 21:58:26 |
| toonn | Hmm, not sure why the one eval job seems stuck queued, or what neutral status means. | 22:39:20 |
| toonn | Late though, I'll look into it tomorrow or Friday. | 22:39:38 |
| 3 Jun 2021 |
| toonn | LnL: Ok, so two of the checks ended with a status of "neutral." And I think one of them's because we need the intermediate bootstrap-tools. | 10:59:39 |
