Benjamin Saunders

@djc raises some good points that weren't immediately obvious to me. In particular, *not* sending data on a packet that's outgoing anyway and has free space seems pessimal, but sending...

Thought: if your application protocol is not intended to have any data following a certain message, you might want to detect cases where that isn't upheld (i.e. where the peer...

Another related but slightly different case: when opening a new connection and sending 0-RTT data, Quinn will typically send a padded 1200-byte UDP datagram bearing the Initial packet, followed by...

> How do I know if my packet has arrived safely? This depends on what exactly "arrived safely" means to you. Usually what people mean is "has been processed by...

Closing as there don't seem to be any outstanding questions, but feel free to open a new issue if you still have a problem!

Odd, seems like a rounding error, but I don't see why that would round differently. Looks harmless, at least. I don't have a 32-bit system handy to test, unfortunately.

@QUICTester was a consensus reached?

Thanks for the update!

Hmm, wouldn't that compromise sharing with other crates? I guess we need very few definitions so maybe that's fine.

`ExactSizeIterator` indeed already exists for matrix iterators, and iterator traits don't make sense for types that aren't iterators. Not clear which this issue is intended to be about, but in...