Replies: 2 comments
-
|
Discussed a bit here: https://youtube.com/live/NKvU0w3IZQg This seems interesting, but is probably a "not for now" design question! |
Beta Was this translation helpful? Give feedback.
-
|
The more I think about this, the more I wonder if this would leave a door open for larger desktop to desktop comms. One mitigation we could do soonish (or at least before stabilizing) would be to limit net_ids and ports to <= 127, which matches a 1-byte varint limit (this would be annoying for broadcast messages, we might need to move that ID to That being said: I still worry about compatibility: you could have a network with |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
During stream morning stream james mentioned the possible future problem of "what if people wanna have really large networks".
A simple solution could be for the protocol to be generic
These would all be completely incompatible protocols, but since the intended use-case for ergot is not an global public internet protocol but a private closed one where all the parts are probably maintained by the same team. They just have to collectively decide on their network would best be served by having network ids be at a max of 2^8, 2^16 or 2^32.
For a simpler api it might smart to default this generic parameter to u16.
If this has some unforeseen consequence an alternative could also just be to have different features that could be enabled for the different sizes, u8, u16, u32, u64 and u128, Though this approach is less flexible than generics, but would mean technically not using generics if this courses some problem in the implementation
Beta Was this translation helpful? Give feedback.
All reactions