fix: keep raw text elements out of Markdown parsing

src/Rules/HtmlRule.php

@@ -10,6 +10,13 @@
 
 final class HtmlRule implements Rule, ProvidesFirstChar
 {
+    /**
+     * Elements whose content is raw text rather than Markdown.
+     *
+     * @see https://spec.commonmark.org/0.31.2/#html-blocks
+     */
+    private const array RAW_TEXT_TAGS = ['pre', 'script', 'style', 'textarea'];
+
     public string $firstChar = '<';
 
     public function shouldParse(Parser $parser): bool
@@ -84,6 +91,13 @@ public function parse(Parser $parser): Token
 
         $content .= $parser->consumeWhile(Parser::NEW_LINE);
 
-        return new HtmlToken($content);
+        return new HtmlToken(
+            $content,
+            raw: in_array(
+                strtolower($tagName),
+                self::RAW_TEXT_TAGS,
+                strict: true,
+            ),
+        );
     }
 }

src/Tokens/HtmlToken.php

@@ -18,10 +18,17 @@ final class HtmlToken implements Token
 {
     public function __construct(
         public string $html,
+
+        /** Whether the content is raw text that must never be parsed as Markdown. */
+        public bool $raw = false,
     ) {}
 
     public function parse(Parser $parser): string
     {
+        if ($this->raw) {
+            return $this->html;
+        }
+
         return $parser
             ->forToken($this, [
                 new BoldAndItalicRule(),

tests/Rules/HtmlRuleTest.php

@@ -82,4 +82,64 @@ public function test_void_tags_are_case_insensitive(): void
 
         $this->assertSame("<BR>\n<p>Hello</p>", $html);
     }
+
+    #[Test]
+    public function lex_keeps_script_content_raw(): void
+    {
+        $html =
+            (string) new Parser(highlighter: null, rules: [new HtmlRule()])->parse(
+                '<script>const x = "**a**";</script>',
+            );
+
+        $this->assertSame('<script>const x = "**a**";</script>', $html);
+    }
+
+    #[Test]
+    public function lex_keeps_style_content_raw(): void
+    {
+        $html =
+            (string) new Parser(highlighter: null, rules: [new HtmlRule()])->parse(
+                '<style>a::after{content:"*x*"}</style>',
+            );
+
+        $this->assertSame('<style>a::after{content:"*x*"}</style>', $html);
+    }
+
+    #[Test]
+    public function lex_keeps_pre_and_textarea_content_raw(): void
+    {
+        $parser = new Parser(highlighter: null, rules: [new HtmlRule()]);
+
+        $this->assertSame(
+            '<pre>**a** _b_</pre>',
+            (string) $parser->parse('<pre>**a** _b_</pre>'),
+        );
+
+        $this->assertSame(
+            '<textarea>**a**</textarea>',
+            (string) $parser->parse('<textarea>**a**</textarea>'),
+        );
+    }
+
+    #[Test]
+    public function lex_detects_raw_text_tags_case_insensitively(): void
+    {
+        $html =
+            (string) new Parser(highlighter: null, rules: [new HtmlRule()])->parse(
+                '<SCRIPT>**a**</SCRIPT>',
+            );
+
+        $this->assertSame('<SCRIPT>**a**</SCRIPT>', $html);
+    }
+
+    #[Test]
+    public function lex_still_parses_markdown_in_other_elements(): void
+    {
+        $html =
+            (string) new Parser(highlighter: null, rules: [new HtmlRule()])->parse(
+                '<div>**a**</div>',
+            );
+
+        $this->assertSame('<div><strong>a</strong></div>', $html);
+    }
 }