Austin Bonander

@Ralith if you read the description, I had a question of exactly _how_ those tests should work.

I'll try to when I get some breathing room. Things are a bit hectic at work right now. I'm having to squeeze in SQLx work during my weekend.

@Ralith @djc rebased and added test cases. > I can think of a couple interesting cases off-hand: > > 1. Construct with send window I > > 2. Shrink send...

Supporting a completion-based runtime would be a significant refactor as all the buffered reads and writes are based around polling. I can imagine how it might _look_, but that's more...

> I like the simplified approach with only with_default_attributes and attribute(...) but maybe it would also need clear_attributes because the user can also define the connection attribute via the connection...

If I had to guess, I would say that Postgres is probably expecting the partial index expression to (more or less) exactly match the one on the index itself; it's...

> only the first migration will be applied correctly. More specifically, all migrations up through the one that creates the schema will succeed, but at that point the migration machinery...

There isn't one single fix in the works because the error doesn't have one and only one single cause. If it was caused by a single obviously fixable issue, it'd...

> Adding something on PgConnectionOptions seems pretty doable This is reasonable. I'd accept a PR for this.

I don't think a trait for the _user-facing_ part of the API is the right direction, precisely because it'll have to be imported everywhere which is going to be a...