Age | Commit message (Collapse) | Author | |
---|---|---|---|
2023-08-19 | refactor: decouple html5lib_tests from html5tokenizer | Martin Fischer | |
Previously we mapped the test tokens to our own token type. Now we do the reverse, which makes more sense as it enables us to easily add more detailed fields to our own token variants without having to worry about these fields not being present in the html5lib test data. (An alternative would be to normalize the values of these fields to some arbitrary value so that PartialEq still holds but seeing such normalized fields in the diff printed by pretty_assertions on a test failure would be quite confusing). | |||
2023-08-19 | refactor: split off reusable html5lib_tests crate | Martin Fischer | |
2023-08-19 | refactor: separate test logic from html5lib-test parsing | Martin Fischer | |
2023-08-19 | test: enable previously skipped tokenizer test | Martin Fischer | |
2023-08-19 | break!: remove set_last_start_tag from Emitter | Martin Fischer | |
2023-08-19 | refactor: move html5lib test to own crate to fix `cargo test` | Martin Fischer | |
Previously `cargo test` failed because it ran the test_html5lib integration test, which depends on the integration-tests feature (so you always had to run `cargo test` with `--features integration-tests` or `--all-features`, which was annoying). This commit moves the integration tests to another crate, so that the dependency on the feature can be properly defined in a way so that `cargo test` just works and runs the test. |