fix: apply the flanking rules to emphasis delimiters

src/Rules/BoldAndItalicRule.php

@@ -14,15 +14,28 @@ final class BoldAndItalicRule implements
     ProvidesFirstChar,
     ProvidesStopChar
 {
+    use ChecksFlanking;
+
     private(set) string $firstChar = '*_';
     private(set) string $stopChar = '_*';
 
     public function shouldParse(Parser $parser): bool
     {
-        return (
-            $parser->comesNext('***', length: 3)
-            || $parser->comesNext('___', length: 3)
-        );
+        if (
+            ! $parser->comesNext('***', length: 3)
+            && ! $parser->comesNext('___', length: 3)
+        ) {
+            return false;
+        }
+
+        $start = $parser->position + 3;
+        $end = strpos($parser->content, $parser->current ?? '', $start);
+
+        return $this->isFlanked(substr(
+            $parser->content,
+            $start,
+            $end === false ? null : $end - $start,
+        ));
     }
 
     public function parse(Parser $parser): Token

src/Rules/BoldRule.php

@@ -11,6 +11,8 @@
 
 final class BoldRule implements Rule, ProvidesFirstChar, ProvidesStopChar
 {
+    use ChecksFlanking;
+
     private(set) string $firstChar = '*_';
     private(set) string $stopChar = '*_';
 
@@ -49,7 +51,7 @@ public function shouldParse(Parser $parser): bool
             return false;
         }
 
-        return true;
+        return $this->isFlanked(substr($content, 0, -1));
     }
 
     public function parse(Parser $parser): Token

src/Rules/ChecksFlanking.php

@@ -0,0 +1,23 @@
+<?php
+
+namespace Tempest\Markdown\Rules;
+
+use Tempest\Markdown\Parser;
+
+trait ChecksFlanking
+{
+    /**
+     * An opening delimiter may not be followed by whitespace, nor a closing
+     * one preceded by it.
+     *
+     * @see https://spec.commonmark.org/0.31.2/#left-flanking-delimiter-run
+     */
+    private function isFlanked(string $content): bool
+    {
+        return (
+            $content !== ''
+            && ! str_contains(Parser::WHITESPACE, $content[0])
+            && ! str_contains(Parser::WHITESPACE, $content[-1])
+        );
+    }
+}

src/Rules/ItalicRule.php

@@ -11,6 +11,8 @@
 
 final class ItalicRule implements Rule, ProvidesFirstChar, ProvidesStopChar
 {
+    use ChecksFlanking;
+
     private(set) string $firstChar = '*_';
     private(set) string $stopChar = '*_';
 
@@ -41,7 +43,7 @@ public function shouldParse(Parser $parser): bool
             return false;
         }
 
-        return true;
+        return $this->isFlanked(substr($end, 0, -1));
     }
 
     public function parse(Parser $parser): Token

tests/EmphasisFlankingTest.php

@@ -0,0 +1,66 @@
+<?php
+
+namespace Tempest\Markdown\Tests;
+
+use PHPUnit\Framework\Attributes\DataProvider;
+use PHPUnit\Framework\Attributes\Test;
+use Tempest\Markdown\Markdown;
+
+final class EmphasisFlankingTest extends ParserTestCase
+{
+    #[Test]
+    #[DataProvider('provideFlanking')]
+    public function delimiters_follow_the_flanking_rules(
+        string $markdown,
+        string $expected,
+    ): void {
+        $this->assertSame(
+            $expected,
+            new Markdown(highlighter: null)->parse($markdown)->html,
+        );
+    }
+
+    public static function provideFlanking(): iterable
+    {
+        yield 'italic' => ['*a*', '<p><em>a</em></p>'];
+        yield 'bold' => ['**a**', '<p><strong>a</strong></p>'];
+        yield 'bold and italic' => [
+            '***a***',
+            '<p><strong><em>a</em></strong></p>',
+        ];
+        yield 'underscore italic' => ['_a_', '<p><em>a</em></p>'];
+        yield 'underscore bold' => ['__a__', '<p><strong>a</strong></p>'];
+        yield 'inside a word' => ['foo*bar*', '<p>foo<em>bar</em></p>'];
+
+        // An opening delimiter may not be followed by whitespace, and a
+        // closing one may not be preceded by it.
+        yield 'spaced asterisks' => ['2 * 3 * 4', '<p>2 * 3 * 4</p>'];
+        yield 'spaced double asterisks' => ['a ** b **', '<p>a ** b **</p>'];
+        yield 'spaced triple asterisks' => [
+            'a *** b ***',
+            '<p>a *** b ***</p>',
+        ];
+        yield 'spaced underscores' => ['x _ y _', '<p>x _ y _</p>'];
+        yield 'space before the closing asterisk' => ['*a *', '<p>*a *</p>'];
+        yield 'space before the closing double asterisk' => [
+            '**a **',
+            '<p>**a **</p>',
+        ];
+        yield 'space after the opening double asterisk' => [
+            '** a**',
+            '<p>** a**</p>',
+        ];
+        yield 'space before the closing double underscore' => [
+            '__a __',
+            '<p>__a __</p>',
+        ];
+        yield 'hotel ratings' => [
+            'Hôtels 3*** et 4****',
+            '<p>Hôtels 3*** et 4****</p>',
+        ];
+        yield 'literal run after emphasis' => [
+            '*a* * b *',
+            '<p><em>a</em> * b *</p>',
+        ];
+    }
+}