Tests no longer require internet

[UebelAndre]

This makes them more hermetic and makes the tests less flaky and makes porting this project to build with Bazel easier.

The changes in this PR add metadata templates to impl/src/testing/metadata_templates which are the formatted outputs of cargo metadata --format-version 1. Each .template file has the Cargo.toml contents at the top of it that produced it. These templates are then rendered during the relevant tests (test similar to the file name). This avoids having to fetch any metadata from the internet during testing.

This change also adds RazeMetadataFetcher which is the the new class to use for gathering metadata for the planning phase of cargo-raze. Note that this no longer implements MetadataFetcher as MetadataFetcher is intended to be a wrapper around a basic way to generate a cargo_metadata::Metadata object. RazeMetadataFetcher is intended to generate a RazeMetadata object.

This PR is blocked by #298

[UebelAndre]

@acmcarther For this one I wasn't quite sure what to do with the metadata. I opted to store it and try to keep changes to code minimal but it would make sense to change the crates used in some of these tests.

[UebelAndre]

Also worth noting, this only has as many new lines of code because the metadata dumps are pretty printed but I feel that's the correct thing to do as it makes it more readable and makes it easier to track the footprint of that metadata.

[UebelAndre]

@acmcarther This is rebased and ready for review now 🙏

[acmcarther]

I'm not sure how I feel about the templated json, but I think it's clearly better than having the tests depend on fetching stuff from the internet.

Generally LGTM, mostly nits. High level nit as well: Can you update the PR description here to describe the abstractions you're adding and the motivation for adding them? That practice makes it easier to digest medium and larger sized PRs.

[UebelAndre]

@acmcarther how's the description look now? Also, still stuck on the actix-web test

[UebelAndre]

I'm not sure how I feel about the templated json, but I think it's clearly better than having the tests depend on fetching stuff from the internet.

@acmcarther I'm not in love with it either but I do feel it's better than what was before. If there's a simpler solution then I could take another pass. But I do feel this change would be good to get us on the path to something better.

[UebelAndre]

@acmcarther Hey, are you around today? 🙏 😅

[acmcarther]

You're doing that force push thing again which is breaking the diff links.

To be really clear the thing that is broken is clicking file links for prior comments, such as this: https://github.com/google/cargo-raze/pull/300/files/f7c7512837473503fcab89c08d4180968c53e43b#diff-1fb1e871b8f9024bde534a1e34e1b0285f4712c79a01284605eb55cf6e7a3e36

[acmcarther]

I'm ready to merge this once the open question about that test which used actix-web is answered. Sorry for the slowness on the review.

Re: description: The new contents of the description are helpful. I was being a bit more literal with my request to include the abstractions -- i meant to literally include the non-trivial new structs. I don't feel strongly about that, hence I didn't raise that. What we have here looks good to me.

[UebelAndre]

I'm ready to merge this once the open question about that test which used actix-web is answered. Sorry for the slowness on the review.

Re: description: The new contents of the description are helpful. I was being a bit more literal with my request to include the abstractions -- i meant to literally include the non-trivial new structs. I don't feel strongly about that, hence I didn't raise that. What we have here looks good to me.

@acmcarther I can do that from here on. I think this is my last large PR though since I feel all the large missing pieces of functionality are in.

You're doing that force push thing again which is breaking the diff links.

To be really clear the thing that is broken is clicking file links for prior comments, such as this: https://github.com/google/cargo-raze/pull/300/files/f7c7512837473503fcab89c08d4180968c53e43b#diff-1fb1e871b8f9024bde534a1e34e1b0285f4712c79a01284605eb55cf6e7a3e36

Sorry, I forgot that this PR was in review when fixing up my branches. I am making an effort to not force push once a PR is in review. Since all my PRs from here on are smaller, I think this will get easier.

[UebelAndre]

@acmcarther I've also added more to the description.

[UebelAndre]

@acmcarther Hey, are you around today to take another look here? 🙏

[UebelAndre]

Thank you so much! 🙏