Since 2010, I've been maintaining a dbdpg fork with full async support, that is, support for async connect and async prepare in addition to async statement execution. While the design is a bit hacky (read: goto-laden), this is pretty well-tested code as it has been used for backend <-> Postgres communication for somewhat largish VPN servers for mobile devices (up to 9000 users) for years. I'm willing to put some work into this to get it cleaned up and eventually upstreamed if this is possible.
So far, I have a local fork of the git repository with the async changes sitting on a dedicated branch. What would be a sensible way to proceed from there?
