Add unit test for all HTML 5 entities that CommonMark parsers must su…

…pport
thephpleague · Oct 13, 2019 · 33ddd70 · 33ddd70
1 parent 4cd6bac
commit 33ddd70
Show file tree

Hide file tree

Showing 3 changed files with 2,255 additions and 0 deletions.
diff --git a/composer.json b/composer.json
@@ -27,6 +27,7 @@
         "ext-mbstring": "*"
     },
     "require-dev": {
+        "ext-json": "*",
         "cebe/markdown": "~1.0",
         "commonmark/commonmark.js": "0.29.0",
         "erusev/parsedown": "~1.0",

diff --git a/tests/unit/Util/Html5EntityDecoderTest.php b/tests/unit/Util/Html5EntityDecoderTest.php
@@ -16,4 +16,25 @@ public function testEntityToChar()
         $this->assertEquals('Æ', Html5EntityDecoder::decode('&AElig;'));
         $this->assertEquals('Ď', Html5EntityDecoder::decode('&Dcaron;'));
     }
+
+    /**
+     * @dataProvider htmlEntityDataProvider
+     */
+    public function testAllHtml5EntityReferences(string $entity, string $decoded)
+    {
+        $this->assertEquals($decoded, html_entity_decode($entity, ENT_QUOTES | ENT_HTML5, 'UTF-8'), sprintf('Failed parsing the "%s" entity', $entity));
+    }
+
+    public function htmlEntityDataProvider()
+    {
+        // Test data from https://html.spec.whatwg.org/multipage/entities.json
+        $data = json_decode(file_get_contents(__DIR__ . '/entities.json'), true);
+        foreach ($data as $entity => $info) {
+            // Per the spec, we only care about entities that have a trailing semi-colon.
+            // See https://spec.commonmark.org/0.29/#entity-references
+            if (substr($entity, -1, 1) === ';') {
+                yield [$entity, $info['characters']];
+            }
+        }
+    }
 }