| 6 Aug 2025 |
 dramforever | ... huh | 04:17:39 |
| dramforever | so it's only git+file that eats submodules | 04:18:10 |
 cleverca22 | and the question becomes, which one is "right"? | 04:20:35 |
| cleverca22 | i guess now its time to pop open a ticket on github? | 04:20:57 |
| dramforever | git archive generates empty directory for submodule | 04:21:29 |
| dramforever | so i'm starting to think empty directory is "correct" | 04:21:41 |
| cleverca22 | so fixing nix, would remove the one case where it works, and then pcsx-redux definitely needs a PR to fix it | 04:22:18 |
| dramforever | yes | 04:24:36 |
| dramforever | even though i guess it would be slightly weird now | 04:24:56 |
| dramforever | because until nix gets fixed you kinda need to handle both cases | 04:25:15 |
| cleverca22 | you just need to rmdir, and handle if its already missing | 04:25:33 |
| cleverca22 | and then copy the submodule into place | 04:25:37 |
| cleverca22 | and all of this mess, is only because flakes strip submodules! | 04:25:59 |
| cleverca22 | https://github.com/grumpycoders/pcsx-redux/blob/main/pcsx-redux.nix#L39-L117 | 04:26:45 |
| cleverca22 | this whole mess of code, is there to recreate the submodules | 04:26:55 |
| dramforever | probably add -T to cp? | 04:27:32 |
| dramforever | well you can get flakes to do submodules | 04:28:15 |
| cleverca22 | ive tried before, and it doesnt actually do it for self | 04:28:36 |
| dramforever | but you have to use git+something and then ?submodules=1 | 04:28:36 |
