Leon Matthes

> Using a boxed slice instead of Vec could have negative performance implications in some cases. If you're working with a serialization library such as bincode it's likely that you...

> Updates on this? You might want to take a look at the `moduleverse` branch of my fork. https://github.com/LeonMatthes/enet-rs/tree/moduleverse It contains both the fix from this PR + the endianness...

> Hey @LeonMatthes , thanks for your work on this! While the repo maintainer is unresponsive, would you consider publishing your fork on crates.io under a different name, so that...

@futile Great to see you're back :) Unfortunately this PR has now deviated too far from your work for a rebase to be feasible. (I'm trying right now, but it's...

> While we're at it, Connect is supposed to have data just like Disconnect does. Probably best to keep stuff like that in a separate PR, so this PR doesn't...

Having generations for peers is definitely useful, I was concerned about "dangling" PeerIDs as well. AFAIK enet reuses them. I'll have to figure out a way to count the generations...

And thank you for reviewing this, it's really gotten too big over time. I completely understand that it's a big ask of any maintainer to work with so much code...

Hm, I've made the PeerID generational. I've moved the drop_disconnected call from `process_event` to `service`, as otherwise the stale Peer would not be dropped until a new Event is received....

@futile Sorry that it has taken me so long to get back to this. I've incorporated your feedback wherever I didn't have questions and marked the conversations as resolved. There's...

> @LeonMatthes what exactly is the difference between your `moduleverse` branch and this PR (your `master` branch)? moduleverse contains some older changes, right? @futile is there anything preventing this from...