Tracking issue: Improvements to cargo-semver-checks' own CI and test system

[obi1kenobi]

• split up the test suite into separate test crates highlighting each specific semver issue, while running all lints on all test crates: Split each source code in test_crates/ into pair of crates #203
• add a check that the Cargo.toml crate name in new/old matches the name of the directory where the new/old crates are located
• add a check that new/old crates have no cargo fmt issues
• add a script for creating a new pair of test crates given the directory name (Add script for creating everything needed to add a new lint. #236, Make the "new lint" script idempotent for each step separately. #237)
• add a script for creating a new lint file and corresponding test outputs file (because the error message when you forget to add the test outputs file is kind of obscure) (Add script for creating everything needed to add a new lint. #236, Make the "new lint" script idempotent for each step separately. #237)
• witness crates: Add witness crates that demonstrate how each breaking change is breaking #223
• Run cargo build and test built binary to ensure it works #235
• run cargo-semver-checks on the most popular crates on crates.io: Script that runs the lints on most popular crates #207
• generate rustdoc for the current version of the most popular crates on crates.io, then run cargo-semver-checks with that rustdoc as both baseline and current so that any reported issues are definitely false-positive