Dan Burkert

This gets at the bigger issue of providing exactly-once semantics for clients. It's necessary to have exactly-once client operations for the consensus implementation to provide linearizability (in the case of...

Actually exactly-once might not be necessary for linearizability, it may be a bit stronger than required? CC #106

I don't think it's possible to provide 'at-least once' behavior in a distributed system without allowing retries, but correct me if I'm wrong. The original proposal in this thread handles...

@rbranson ping, I'd like to get your thoughts on this. Thanks!

@rbranson would it be acceptable to add you as a committer on my repo, and have you add me as an [owner](http://doc.crates.io/crates-io.html#cargo-owner) on crates.io? The next step is that I...

It looks like there are [five crates](https://crates.io/crates/mmap/reverse_dependencies) depending on `mmap` right now: - block_allocator - anonymous mmap - context - anonymous stack mmap - coroutine - anonymous stack mmap -...

@per-gron thanks for the PR. Could you explain the motivation behind the unknown fields feature? I've read through the upstream docs linked in the first comment, but I don't really...

Just to set expectations, it's never been a goal of mine to have `prost` officially endorsed by Google or the upstream Protobuf authors or community. Nor has it been a...

Interesting, this is one corner of the protobuf feature set I've never encountered. Could you elaborate a bit on when this comes in handy? If it's a useful feature then...

This also happens when deploying a `dispatch.yaml`.