Understanding Rust futures by going way too deep

submited by

Style Pass

2021-07-26 06:30:03

So! Rust futures! Easy peasy lemon squeezy. Until it's not. So let's do the easy thing, and then instead of waiting for the hard thing to sneak up on us, we'll go for it intentionally.

Yeah, because it's really convenient. Readers just get confused because subcommands like cargo new, cargo build, cargo test, cargo run etc. are built into cargo, but cargo add isn't.

Then we pick an async runtime, because those futures won't poll themselves... and we'll pick tokio for no reason other than: that's what I've been using a bunch these past few months.

First, for error handling - we're writing an app, we're going to get a bunch of different types from different libraries, it'd be neat if we could have one type to unify them all.

Now we need to install color-eyre as the default panic handler, and I snuck in some environment variable modification so we get backtraces by default.

And finally, because I like my logs to be structured, let's add tracing and to print them with nice colors in the terminal, let's add tracing-subscriber.