feat(service): get rid of cargo dependency

[iamwacko]

Description of change

is_next, is_alpha, ensure_bin, ensure_cdylib, and clean_crate no longer require cargo.

partially closes #756, particularly service part 1.

How Has This Been Tested (if applicable)?

I ran cargo test, cargo fmt, and cargo clippy.

[iamwacko]

I have time, so I can do service part 2 as well.

[iulianbarbu]

I have time, so I can do service part 2 as well.

Sounds good @iamwacko ! Let us know if you need help. We will review this whenever it's ready.

[jonaro00]

It's hard to accurately review Rust code without rust-analyzer at hand, but here are some thoughts I had. Don't worry if I'm wrong about something.

[iulianbarbu]

@iamwacko let me know if you have questions on the changes I proposed and we can set up a call to get through them. The PR is on the right path and we'll just need to polish a bit the code async-wise, while keeping it on par with the old functionality.

We have a codebase walkthrough session next Monday (if that helps) and a mentoring session next Wednesday, but feel free to hit me up with qs in private on Discord if you feel stuck.

[jonaro00]

I found some more small things.

I read some more about tokio's Command.
It sounds like it is the proper replacement for spawn_blocking + std Command, so that should be fine.

[oddgrd]

This looks pretty close @iamwacko! I left a few small comments, but I think the main thing we need to fix is sending the build logs to the build_workspace caller, which Iulian pointed out. Let us know if you are stuck on this and need any help!

[oddgrd]

Nice progress! I took a look at your latest changes and saw what may be an easier way to read the output, let me know if you think it could work!

[iulianbarbu]

Thanks @iamwacko ! This is great 🥳 .

[iulianbarbu]

I think there is a regression on shuttle-next. I am not able to run the next/hello-world because of:

thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: status: Unknown, message: "failed to read input file", details: [], metadata: MetadataMap { headers: {"content-type": "application/grpc", "date": "Mon, 05 Jun 2023 20:43:08 GMT", "content-length": "0"} }', /Users/iulianbarbu/repos/shuttle/cargo-shuttle/src/lib.rs:779:51

We must investigate the failing integration test on cargo-shuttle.

[iamwacko]

I see the problem. - gets turned into _ by cargo, but I didn't account for that so it tries to execute hello-world.wasm instead of hello_world.wasm. I also didn't catch it because for a while cargo-shuttle was failing for unrelated reasons.

[iamwacko]

Now the cargo-shuttle test is failing with thread 'init::non_interactive_basic_init' panicked at 'Could not query the latest version of shuttle-runtime', cargo-shuttle/src/init.rs:994:33. It works locally, so I have no idea what has gone wrong. I haven't even touched the code that would make this happen.

[iamwacko]

Looks like the only test failures I have are #821

[iulianbarbu]

Thanks @iamwacko for fixing the shuttle-next issue. Yes, it seems we're getting the error corresponding to #821 in CI. The tests pass locally for me as well.