Fixes bad and slow behaviors

src/Rules/ListRule.php

@@ -25,38 +25,41 @@ public function parse(Parser $parser): ?Token
         $newlines = $parser->consumeWhile(Parser::NEW_LINE);
 
         $childContent = '';
-        $indent = strspn($parser->content, ' ', $parser->position);
 
-        while ($indent >= 2 && $parser->current !== null) {
-            if (strspn($parser->content, ' ', $parser->position) < $indent) {
-                break;
-            }
+        while ($parser->current !== null) {
+            $indent = strspn($parser->content, ' ', $parser->position);
 
-            $parser->consume($indent);
-            $childContent .= $parser->consumeUntil(Parser::NEW_LINE) . PHP_EOL;
-            $newlines = $parser->consumeWhile(Parser::NEW_LINE);
-        }
+            if ($indent >= 2) {
+                while ($parser->current !== null && strspn($parser->content, ' ', $parser->position) >= $indent) {
+                    $parser->consume($indent);
+                    $childContent .= $parser->consumeUntil(Parser::NEW_LINE) . PHP_EOL;
+                    $newlines = $parser->consumeWhile(Parser::NEW_LINE);
+                }
 
-        $children = null;
+                if (preg_match('/(?:^|\n)[-*+] /', $childContent)) {
+                    break;
+                }
 
-        if (preg_match('/(?:^|\n)[-*+] /', $childContent)) {
-            $children = $parser->withRules(new ListRule())->lex($childContent)[0];
-        }
+                $content .= ' ' . trim(preg_replace('/\s+/u', ' ', $childContent) ?? '');
+                $childContent = '';
+                continue;
+            }
 
-        if ($children === null && $childContent !== '') {
-            $content .= ' ' . trim(preg_replace('/\s+/u', ' ', $childContent) ?? '');
-        }
+            if ($newlines !== "\n" && $newlines !== "\r\n") {
+                break;
+            }
 
-        while ($children === null && ($newlines === "\n" || $newlines === "\r\n") && $parser->current !== null) {
             $nextLine = substr($parser->content, $parser->position, strcspn($parser->content, Parser::NEW_LINE, $parser->position));
-            if (trim($nextLine) === '' || preg_match('/^(?:[-*+] |[0-9]+[.)] |#{1,6}[ \t]|> |`{3}|~{3})/', $nextLine)) {
+            if (trim($nextLine) === '' || preg_match('/^(?:[-*+] |[0-9]+[.)] |#{1,6}[ \t]|>|`{3}|~{3}|-{3}|={3}|<|:{3})/', $nextLine)) {
                 break;
             }
 
             $content .= ' ' . trim($parser->consumeUntil(Parser::NEW_LINE));
             $newlines = $parser->consumeWhile(Parser::NEW_LINE);
         }
 
+        $children = $childContent === '' ? null : $parser->withRules(new ListRule())->lex($childContent)[0];
+
         $item = new ListItem($content, $children);
 
         if ($parser->lastToken instanceof ListToken) {

src/Rules/ParagraphRule.php

@@ -20,27 +20,31 @@ public function parse(Parser $parser): Token
         $content = '';
 
         while ($parser->current !== null) {
-            $content .= $parser->consumeUntil(Parser::NEW_LINE);
+            $line = $parser->consumeUntil(Parser::NEW_LINE);
 
-            if ($parser->current === null) {
-                break;
-            }
+            // A blank line ends the paragraph and stays available to the newline rule.
+            $endsParagraph =
+                $parser->position >= $parser->length
+                || $parser->comesNext("\n\n", 2)
+                || $parser->comesNext("\r\n\r\n", 4)
+                || $parser->comesNext("\n\r\n", 3)
+                || $parser->comesNext("\r\n\n", 3);
 
-            // A blank line (two consecutive newlines) ends the paragraph
-            if ($parser->comesNext("\n\n", 2) || $parser->comesNext("\r\n\r\n", 4) || $parser->comesNext("\n\r\n", 3) || $parser->comesNext("\r\n\n", 3)) {
-                break;
-            }
+            $newline = $endsParagraph ? '' : $parser->consumeWhile(Parser::NEW_LINE);
 
-            // Single newline — consume it and continue to the next line
-            $content .= $parser->consumeWhile(Parser::NEW_LINE);
-        }
+            $matches = [];
+            if ($content !== '' && preg_match('/\A {0,3}(=+|-+)[ \t]*\z/', $line, $matches)) {
+                $heading = trim($content);
+                $id = mb_strtolower($heading) |> (fn (string $value) => trim(preg_replace('/[^\p{L}\p{N}]+/u', '-', $value) ?? '', '-'));
 
-        $trimmed = rtrim($content, "\r\n");
-        if (preg_match('/\A(.+)\r?\n {0,3}(=+|-+)[ \t]*\z/s', $trimmed, $matches)) {
-            $heading = trim($matches[1]);
-            $id = mb_strtolower($heading) |> (fn (string $value) => trim(preg_replace('/[^\p{L}\p{N}]+/u', '-', $value) ?? '', '-'));
+                return new HeadingToken($heading, $matches[1][0] === '=' ? 1 : 2, $id);
+            }
+
+            $content .= $line . $newline;
 
-            return new HeadingToken($heading, $matches[2][0] === '=' ? 1 : 2, $id);
+            if ($endsParagraph) {
+                break;
+            }
         }
 
         return new ParagraphToken($content);

src/Rules/StrikethroughRule.php

@@ -8,6 +8,7 @@
 use Tempest\Markdown\Rule;
 use Tempest\Markdown\Token;
 use Tempest\Markdown\Tokens\StrikethroughToken;
+use Tempest\Markdown\Tokens\TextToken;
 
 final class StrikethroughRule implements Rule, ProvidesFirstChar, ProvidesStopChar
 {
@@ -16,18 +17,17 @@ final class StrikethroughRule implements Rule, ProvidesFirstChar, ProvidesStopCh
 
     public function shouldParse(Parser $parser): bool
     {
-        if (! $parser->comesNext('~', 1)) {
-            return false;
-        }
-
-        $openingLength = strspn($parser->content, '~', $parser->position);
-
-        return strpos($parser->content, '~', $parser->position + $openingLength) !== false;
+        return $parser->comesNext('~', 1);
     }
 
     public function parse(Parser $parser): Token
     {
-        $parser->consumeWhile('~');
+        $opening = $parser->consumeWhile('~');
+
+        if (strpos($parser->content, '~', $parser->position) === false) {
+            return new TextToken($opening);
+        }
+
         $buffer = $parser->consumeUntil('~');
         $parser->consumeWhile('~');
 

src/Rules/ThinRulerRule.php

@@ -21,6 +21,7 @@ public function shouldParse(Parser $parser): bool
     public function parse(Parser $parser): Token
     {
         $content = $parser->consumeWhile('-');
+        $parser->consumeWhile(" \t");
 
         return new RulerToken(
             $content,

tests/Rules/FrontMatterRuleTest.php

@@ -25,6 +25,8 @@ public function standalone_delimiter_is_a_thematic_break(): void
         $this->assertSame("<hr/>\n", $parser->parse("---\n")->html);
         $this->assertSame('<hr/>', $parser->parse('-----')->html);
         $this->assertSame('<hr/>', new Markdown(highlighter: null)->parse('---')->html);
+        $this->assertSame('<hr/>', new Markdown(highlighter: null)->parse("--- \t")->html);
+        $this->assertSame("<hr/>\n", new Markdown(highlighter: null)->parse("--- \t\n")->html);
     }
 
     #[Test]

tests/Rules/ListRuleTest.php

@@ -44,6 +44,28 @@ public function test_lazy_continuation(): void
         $this->assertSame('<ul><li>one continued</li></ul>', (string) $parser->parse("* one\ncontinued"));
     }
 
+    #[Test]
+    public function test_lazy_continuation_preserves_following_blocks(): void
+    {
+        $parser = new Parser(highlighter: null);
+
+        $this->assertSame("<ul><li>one</li></ul><hr/>\n<p>after</p>", (string) $parser->parse("- one\n---\nafter"));
+        $this->assertSame('<ul><li>one</li></ul><div>block</div>', (string) $parser->parse("- one\n<div>block</div>"));
+    }
+
+    #[Test]
+    public function test_nested_list_after_lazy_continuation(): void
+    {
+        $parser = new Parser(highlighter: null);
+
+        foreach (['-', '*', '+'] as $marker) {
+            $this->assertSame(
+                '<ul><li>one continued<ul><li>child</li></ul></li><li>two</li></ul>',
+                (string) $parser->parse("{$marker} one\ncontinued\n  {$marker} child\n{$marker} two"),
+            );
+        }
+    }
+
     #[Test]
     public function test_lex_multiline_items(): void
     {

tests/Rules/ParagraphRuleTest.php

@@ -58,4 +58,20 @@ public function test_setext_heading_can_span_multiple_lines(): void
 
         $this->assertSame("<h2 id=\"first-second\">First\nSecond</h2>", (string) $parser->parse("First\nSecond\n---"));
     }
+
+    #[Test]
+    public function test_setext_heading_followed_by_paragraph(): void
+    {
+        $parser = new Parser(highlighter: null);
+
+        $this->assertSame('<h2 id="title">Title</h2><p>Body</p>', (string) $parser->parse("Title\n---\nBody"));
+    }
+
+    #[Test]
+    public function test_consecutive_setext_headings(): void
+    {
+        $parser = new Parser(highlighter: null);
+
+        $this->assertSame('<h1 id="one">One</h1><h2 id="two">Two</h2>', (string) $parser->parse("One\n===\nTwo\n---"));
+    }
 }

tests/Rules/StrikethroughRuleTest.php

@@ -8,6 +8,7 @@
 use Tempest\Markdown\Rules\StrikethroughRule;
 use Tempest\Markdown\Rules\TextRule;
 use Tempest\Markdown\Tests\ParserTestCase;
+use Tempest\Markdown\Tokens\TextToken;
 
 class StrikethroughRuleTest extends ParserTestCase
 {
@@ -37,4 +38,26 @@ public function unmatched_tildes_remain_literal(): void
         $this->assertSame('~~~', (string) $parser->parse('~~~'));
         $this->assertSame('<p>Duration: ~2h30.</p>', new Markdown(highlighter: null)->parse('Duration: ~2h30.')->html);
     }
+
+    #[Test]
+    public function test_unmatched_run_is_consumed_as_one_literal_token(): void
+    {
+        $opening = str_repeat('~', 80_000);
+        $parser = new Parser(highlighter: null)->setContent($opening . '**bold**');
+        $rule = new StrikethroughRule();
+
+        $this->assertTrue($rule->shouldParse($parser));
+        $this->assertEquals(new TextToken($opening), $rule->parse($parser));
+        $this->assertSame(strlen($opening), $parser->position);
+        $this->assertTrue($parser->comesNext('**bold**'));
+    }
+
+    #[Test]
+    public function test_inline_formatting_after_an_unmatched_run(): void
+    {
+        $this->assertSame(
+            '<p>About ~<strong>5 km</strong></p>',
+            new Markdown(highlighter: null)->parse('About ~**5 km**')->html,
+        );
+    }
 }