| 17 Aug 2025 |
 Morgan (@numinit) | what combination of preferLocalBuild flags makes this not a thing anymore? (pretty sure I tried, but it didn't work) | 22:24:27 |
 emily | none; it'd need to skip Hydra entirely | 22:43:26 |
| emily | so hydraPlatforms on the root jobs that lead to it | 22:43:31 |
| emily | * so hydraPlatforms = [ ]; on the root jobs that lead to it | 22:43:39 |
| emily | (are those images actually entirely FOSS?) | 22:43:55 |
| emily | (because meta.license would be another route…) | 22:44:03 |
| Morgan (@numinit) | Gotcha, pretty sure I tried applying the toplevel meta, but something may not be taking. Thanks for the heads up | 22:59:15 |
| Morgan (@numinit) | Will get that fixed | 22:59:26 |
| emily | I will be putting up some Android-related PRs soon :) | 23:21:56 |
| emily | (nothing exciting - LLVM clean-up) | 23:22:05 |
| emily | but yeah the weird thing about hydraPlatforms is that anything else can still pull it in | 23:22:39 |
| emily | @numinit:matrix.org it looks like those system images include GMS btw? which I'm pretty sure is non-free | 23:25:20 |
| Morgan (@numinit) | yeah, that is likely | 23:25:37 |
| Morgan (@numinit) | The intent was to mark all of them | 23:25:51 |
| Morgan (@numinit) | they are certainly locked behind unfree for composing a SDK | 23:26:14 |
| emily | I got them through building the emulator | 23:27:29 |
| Morgan (@numinit) | ah, makes sense, yeah we should absolutely propagate that down | 23:28:03 |
| emily | you can just set meta.license on the fetchers | 23:29:34 |
