regenerate test outputs?

[secona]

#84 highlights the problem of poorly generated test outputs.

In some tests as well, we have JSON keys that are ordered alphabetically and some not. See below for snippets. This can make it hard to see the diff when introducing large changes. Should we regenerate all the tests outputs?

cargo-plumbing/tests/testsuite/lock_dependencies.rs

Lines 762 to 783 in 3a80b98

	assert_e2e().eq(
	latest_lock,
	str![[r#"
	[
	{
	"reason": "lockfile",
	"version": null
	},
	{
	"reason": "locked-package",
	"id": "registry+https://github.com/rust-lang/crates.io-index#a@1.0.0",
	"checksum": "3a351dafbc8a3a9cba7c06dfe8caa11a3a45f800a336bb5b913a8f1e2652d454"
	},
	{
	"reason": "locked-package",
	"id": "lock-dependencies-test@0.1.0",
	"dependencies": [
	"a"
	]
	}
	]
	"#]]

cargo-plumbing/tests/testsuite/lock_dependencies.rs

Lines 648 to 672 in 3a80b98

	assert_e2e().eq(
	latest_lock,
	str![[r#"
	[
	{
	"reason": "lockfile",
	"version": 4
	},
	{
	"checksum": "3a351dafbc8a3a9cba7c06dfe8caa11a3a45f800a336bb5b913a8f1e2652d454",
	"id": "registry+https://github.com/rust-lang/crates.io-index#a@1.0.0",
	"reason": "locked-package"
	},
	{
	"dependencies": [
	"a"
	],
	"id": "lock-dependencies-test@0.1.0",
	"reason": "locked-package"
	}
	]
	"#]]
	.is_json()
	.against_jsonlines(),
	);

[secona]

Hmm, I originally opened this because I did cargo test --workspace --all-features and it outputs very messy test changes. But, doing SNAPSHOTS=overwrite cargo test -p cargo-plumbing actually edits the test in a nice way. So, I guess we don't really need to regenerate the tests.

It does feel wrong, however, to have the tests in different alphabetical orders (probably a nitpick from me). It may also be easier to review when tests are in the same order (?). I'm not sure on that as I don't do reviews.