| 27 Jan 2025 |
 Julien | They were substituted from the cache | 16:19:40 |
| Julien | And did not count in the statistics | 16:19:49 |
 Atemu | FODs aren't really interesting from a r13y standpoint; they are bit-reproducible by definition. | 16:20:28 |
| Atemu | Their primary issue is source availability | 16:20:42 |
| Atemu | As long as you can get a FOD from somewhere, it doesn't matter | 16:21:12 |
| Julien | I'm also interested in the question: how much of nixpkgs can we rebuild without the cache (in which cas we would need to rebuild the FODs to check if source is still available), but we did not answer this question in this paper | 16:21:13 |
 raboof | Atemu: that is only true if you trust the original committer and hydra not to collude on you - which is somewhat safe to assume but also good to check, in theory. | 16:22:23 |
| Atemu | That's a valid issue but again separate from r13y IMV | 16:23:13 |
| Atemu | Though I guess one of the reasons to have r13y is to be able to detect such collusions | 16:24:10 |
| 30 Jan 2025 |
|  @a3don:matrix.org left the room. | 03:11:27 |
|  Kira joined the room. | 04:22:41 |
