refactor: separate test logic from html5lib-test parsing

1 files changed, 13 insertions, 3 deletions

diff --git a/integration_tests/tests/test_html5lib.rs b/integration_tests/tests/test_html5lib.rs
index 32dccd4..61e2133 100644
--- a/integration_tests/tests/test_html5lib.rs
+++ b/integration_tests/tests/test_html5lib.rs
@@ -5,6 +5,13 @@ use pretty_assertions::assert_eq;
 use serde::{de::Error as _, Deserialize};
 use std::{collections::BTreeMap, fs::File, io::BufReader, path::Path};
 
+fn parse_tests(
+    reader: impl std::io::Read,
+) -> Result<impl Iterator<Item = Test>, serde_json::Error> {
+    let Tests { tests } = serde_json::from_reader(reader)?;
+    Ok(tests.into_iter().map(undo_double_escaping))
+}
+
 struct ExpectedOutputTokens(Vec<Token<()>>);
 
 impl<'de> Deserialize<'de> for ExpectedOutputTokens {
@@ -250,14 +257,14 @@ fn test_tokenizer_file(path: &Path) {
 
     let f = File::open(path).unwrap();
     let bf = BufReader::new(f);
-    let tests: Tests = serde_json::from_reader(bf).unwrap();
+    let tests = parse_tests(bf).expect(&format!("failed to parse {path:?}"));
 
-    for (i, test) in tests.tests.into_iter().enumerate() {
+    for (i, test) in tests.into_iter().enumerate() {
         run_test(fname, i, test);
     }
 }
 
-fn run_test(fname: &str, test_i: usize, mut test: Test) {
+fn undo_double_escaping(mut test: Test) -> Test {
     test.input = if test.double_escaped {
         unescape(&test.input)
     } else {
@@ -279,7 +286,10 @@ fn run_test(fname: &str, test_i: usize, mut test: Test) {
     } else {
         ExpectedOutputTokens(test.output.0)
     };
+    test
+}
 
+fn run_test(fname: &str, test_i: usize, test: Test) {
     for state in &test.initial_states {
         run_test_inner(
             fname,