Test benchmarks in CI

[BD103]

Objective

• This reverts Remove benchmarks from CI test #16833, and completely goes against Do not test benchmarks in CI #16803.
• Turns out running cargo test --benches runs each benchmark once, without timing it, just to ensure nothing panics. This is actually desired because we can use it to verify benchmarks are working correctly without all the time constraints of actual benchmarks.

Solution

• Add the --benches flag to the CI test command.

Testing

• cargo run -p ci -- test

[BenjaminBrienen]

Something happened with the bench

thread 'main' panicked at benches/benches/bevy_tasks/iter.rs:80:25:

[BenjaminBrienen]

Suggested change

	// `--benches` runs each benchmark once, to verify that they behave correctly.
	// `--benches` runs each benchmark once in order to verify that they behave correctly. (Don't panic)

[BenjaminBrienen]

Approving since the failure should be fixed in a separate PR

[BD103]

Approving since the failure should be fixed in a separate PR

Surprisingly not! The panic was an integer overflow that was never caught in our benchmarks because they are usually compiled in release mode. I fixed it in aa2e600, though, to explicitly wrap.