| 13 Oct 2025 |
 raitobezarius | not sure, this seems like a mutable lock to me | 09:53:25 |
 MangoIV | and? | 09:53:33 |
| raitobezarius | if cppnix accepts mutable locks now, it probably needs to explain why? | 09:53:51 |
 K900 | IMO cppnix allowing this was a terrible idea | 09:53:57 |
| raitobezarius | i have zero context behind why it was disallowed and now allowed again | 09:54:17 |
| MangoIV | is there a workaround for lix to allow mutable locks? Shouldn't passing --impure allow them? | 09:54:30 |
| raitobezarius | but for us to consider it as a change that we should do as well (not a bug), we would need the rationale context | 09:54:31 |
| raitobezarius | the node needs to be locked | 09:54:48 |
| raitobezarius | for a path to be locked, it needs to have its content hash serialized in the flake lock | 09:55:23 |
| raitobezarius | if flakes are now "everything is hermetic except for this little exception here, that little exception there and so on", i'm not sure i understand what is the value of it anymore :) | 09:55:55 |
| MangoIV | so it is a bug in upstream nix, in that it doesn't lock it properly? | 09:56:26 |
