SnapDown

This library is a little helper to write Markdown-based expectation tests.

Those kind of tests are mostly useful for something like parsers, when you have some texual input and text-serializable output (like an AST tree).

The main convenience of this library is to have both inputs and outputs in the same file.

Works best with datatest.

Example

Test code example:

fn parser(path: &Path) -> datatest_stable::Result<()> {
	run_test::<(), _>(path, |blocks| {
		blocks.chunks_mut(2).for_each(|test| {
			let input = &test[0];
			let output = &test[1];
			
			let syntax = parse(input));
			output.result.set(Some(format!("{:#?}", syntax)));
		});
	})
}

datatest_stable::harness!(parser, "tests/parser", r".*");

Markdown test file example:

```markdown
<Link href="/documents/test/a" />
```

```rust
Root [
    Jsx { closed: true } [
        AngleStart,
        Text ["Link"],
        Space,
        JsxAttr [
            Text ["href"],
            Equals,
            JsxValue [
                String [
                    Quote,
                    ForwardSlash,
                    Text ["documents"],
                    ForwardSlash,
                    Text ["test"],
                    ForwardSlash,
                    Text ["a"],
                    Quote,
                ],
            ],
        ],
        Space,
        ForwardSlash,
        AngleEnd,
    ],
]
```

Configuration

CLI parsing

run_test's first generic argument allows you to provide a Serde deserializable type to control accepted CLI arguments.

Drafts

Test output annotated with --draft will re-generate without diff errors. This is useful for quick iterations while you work on something:

```rust --draft
/* this will re-generate */
```

Use SNAPDOWN_REFRESH=1 environment variable to re-generate all the outputs.