Bug 1811454 [wpt PR 38080] - URL: run IdnaTestV2.txt in WPT, a=testonly

Automatic update from web-platform-tests
URL: run a subset of IdnaTestV2.txt in WPT

This excludes various tests for now due to the open issues mentioned at the top of IdnaTestV2-parser.py.

For whatwg/url#341.
--

wpt-commits: 9216115f5621b04a27e0f2e9bbf1ce44dd7d3b9e
wpt-pr: 38080

testing/web-platform/tests/.gitignore

@@ -46,6 +46,7 @@ scratch
 /css/dist
 /css/dist_last
 /css/tools/cache
+/url/tools/IdnaTestV2.txt
 /webaudio/idl/*
 
 # w3c-test.org PR-branch mirroring

testing/web-platform/tests/url/IdnaTestV2.window.js

@@ -0,0 +1,41 @@
+promise_test(() => fetch("resources/IdnaTestV2.json").then(res => res.json()).then(runTests), "Loading data…");
+
+// Performance impact of this seems negligible (performance.now() diff in WebKit went from 48 to 52)
+// and there was a preference to let more non-ASCII hit the parser.
+function encodeHostEndingCodePoints(input) {
+  let output = "";
+  for (const codePoint of input) {
+    if ([":", "/", "?", "#", "\\"].includes(codePoint)) {
+      output += encodeURIComponent(codePoint);
+    } else {
+      output += codePoint;
+    }
+  }
+  return output;
+}
+
+function runTests(idnaTests) {
+  for (const idnaTest of idnaTests) {
+    if (typeof idnaTest === "string") {
+      continue // skip comments
+    }
+    if (idnaTest.input === "") {
+      continue // cannot test empty string input through new URL()
+    }
+    // Percent-encode the input such that ? and equivalent code points do not end up counting as
+    // part of the URL, but are parsed through the host parser instead.
+    const encodedInput = encodeHostEndingCodePoints(idnaTest.input);
+
+    test(() => {
+      if (idnaTest.output === null) {
+        assert_throws_js(TypeError, () => new URL(`https://${encodedInput}/x`));
+      } else {
+        const url = new URL(`https://${encodedInput}/x`);
+        assert_equals(url.host, idnaTest.output);
+        assert_equals(url.hostname, idnaTest.output);
+        assert_equals(url.pathname, "/x");
+        assert_equals(url.href, `https://${idnaTest.output}/x`);
+      }
+    }, `ToASCII("${idnaTest.input}")${idnaTest.comment ? " " + idnaTest.comment : ""}`);
+  }
+}