Ryan Leckey

> For instance, multiple threads patching and unpatching the same function can cause race conditions. I'm classing that as a bug for now that will be fixed soon enough (tm)....

As a compromise for test UX, what about if we deprecate the functions, rely _only_ on a `patch!` ( as in #11 ) macro, and have that macro emit `#[cfg(test)]`...

Would `#[inline(never)]` help? Of course if this bug requires optimizations of any kind to manifest I'm going to say it's out of scope of this crate. I wouldn't say no...

I would accept a PR that adds the test under a feature flag like `limitations` or `flaky` or something. It's good to document what we can't do after all. >...

I think this would fall under #109

> no binary output function available for type uint8 that's not an error from SQLx, does that query work in psql?

I'm wary of _implicit_ query batching. In `tokio-postgres`, batching happens automatically if multiple futures from query executions are polled concurrently. Supporting query pipelining is fairly simple at the protocol level,...

> Instead of a new object `Batch`, could we make it work with the existing `Transaction`? We can make pipeline creation generic similar to how transaction creation is, so you...

Postgres uses `$1`, `$2`, etc. for variables. --- > error: unsupported comparison operator: = That's potentially interesting / something we could easily fix if I can pinpoint it. Did you...

I'm working on an end-to-end [asynchronous](https://github.com/mehcode/salt-rs/blob/master/examples/proxy.rs#L15) [web framework](https://github.com/mehcode/salt-rs) and serving files with non-blocking I/O is something I'm very interested in. Likewise, I found this by searching around crates.io to see...