fix: agree one a single default runtime for the whole workspace

[crepererum]

This fixes cargo check --workspace and rust-analyzer.

Also see #1956.

[Kinrany]

Could you explain why changing the runtime to async-std helps?

[crepererum]

I'm not entirely sure what depends on what in this workspace, but it seems the members at least don't agree on the defaults:

https://github.com/launchbadge/sqlx/blob/main/sqlx-macros/Cargo.toml#L19

So the Cargo feature unification will pick both tokio and async-std for the RT crate, which then will fail to compile.

[abonander]

If we have to standardize on a single default runtime for the whole workspace, I'd prefer it to be Tokio.

[crepererum]

If we have to standardize on a single default runtime for the whole workspace, I'd prefer it to be Tokio.

Which would mean adjusting all the examples as well (because there the de facto standard is currently async-std), for which I currently don't have the resources, esp. since compiling them requires quite some manual labor.

[abonander]

Yeah, most of those examples were initially authored when SQLx was exclusively for use with async-std. Since then, the async ecosystem has pretty much centered around Tokio and so it would make more sense to port the examples over as well.

For the most part, it should only require some find-and-replace operations, and you can just let CI build it for you.

[crepererum]

I'll see if I find some time this week. On the first glance, the paw::main macro doesn't play nicely with tokio::main. It's probably a good idea to drop paw altogether (last release was 3y back) and while I'm on it, I would probably replace structopt with an up-to-date clap (clap version 3 incorporated structopt).

[Kinrany]

Clap's API is not completely the same as structopt, removing paw alone may be easier.

[crepererum]

This is now ready to review.

[abonander]

Thanks for going the extra mile. I really appreciate it.

[Kinrany]

As a mere user I appreciate it too! ❤️