use PCRE MARK verb to identify token types

Author: thunderer

docker/php-5.x/Dockerfile

@@ -5,7 +5,7 @@ RUN apt update && apt install -y --force-yes libonig-dev libzip-dev
 RUN docker-php-ext-install mbstring zip
 
 RUN php -r "copy('https://getcomposer.org/installer', 'composer-setup.php');" \
- && php -r "if (hash_file('sha384', 'composer-setup.php') === 'c31c1e292ad7be5f49291169c0ac8f683499edddcfd4e42232982d0fd193004208a58ff6f353fde0012d35fdd72bc394') { echo 'Installer verified'; } else { echo 'Installer corrupt'; unlink('composer-setup.php'); } echo PHP_EOL;" \
+ && php -r "if (hash_file('sha384', 'composer-setup.php') === 'dac665fdc30fdd8ec78b38b9800061b4150413ff2e3b6f88543c636f7cd84f6db9189d43a81e5503cda447da73c7e5b6') { echo 'Installer verified'; } else { echo 'Installer corrupt'; unlink('composer-setup.php'); } echo PHP_EOL;" \
  && php composer-setup.php \
  && php -r "unlink('composer-setup.php');" \
  && mv composer.phar /usr/local/bin/composer

docker/php/Dockerfile

@@ -2,7 +2,7 @@ ARG PHP_VERSION=8.0
 FROM php:$PHP_VERSION
 
 RUN php -r "copy('https://getcomposer.org/installer', 'composer-setup.php');" \
- && php -r "if (hash_file('sha384', 'composer-setup.php') === '906a84df04cea2aa72f40b5f787e49f22d4c2f19492ac310e8cba5b96ac8b64115ac402c8cd292b8a03482574915d1a8') { echo 'Installer verified'; } else { echo 'Installer corrupt'; unlink('composer-setup.php'); } echo PHP_EOL;" \
+ && php -r "if (hash_file('sha384', 'composer-setup.php') === 'dac665fdc30fdd8ec78b38b9800061b4150413ff2e3b6f88543c636f7cd84f6db9189d43a81e5503cda447da73c7e5b6') { echo 'Installer verified'; } else { echo 'Installer corrupt'; unlink('composer-setup.php'); } echo PHP_EOL;" \
  && php composer-setup.php \
  && php -r "unlink('composer-setup.php');" \
  && mv composer.phar /usr/local/bin/composer

src/Parser/RegularParser.php

@@ -349,19 +349,9 @@ private function tokenize($text)
 
         $tokens = array();
         $position = 0;
-
         foreach($matches as $match) {
-            switch(true) {
-                case array_key_exists('close', $match): { $token = $match['close']; $type = self::TOKEN_CLOSE; break; }
-                case array_key_exists('open', $match): { $token = $match['open']; $type = self::TOKEN_OPEN; break; }
-                case array_key_exists('separator', $match): { $token = $match['separator']; $type = self::TOKEN_SEPARATOR; break; }
-                case array_key_exists('delimiter', $match): { $token = $match['delimiter']; $type = self::TOKEN_DELIMITER; break; }
-                case array_key_exists('marker', $match): { $token = $match['marker']; $type = self::TOKEN_MARKER; break; }
-                case array_key_exists('ws', $match): { $token = $match['ws']; $type = self::TOKEN_WS; break; }
-                case array_key_exists('string', $match): { $token = $match['string']; $type = self::TOKEN_STRING; break; }
-                default: { throw new \RuntimeException('Invalid token.'); }
-            }
-            $tokens[] = array($type, $token, $position);
+            $token = $match[0];
+            $tokens[] = array((int)$match['MARK'], $token, $position);
             $position += mb_strlen($token, 'utf-8');
         }
 
@@ -373,31 +363,31 @@ private function prepareLexer(SyntaxInterface $syntax)
     {
         // FIXME: for some reason Psalm does not understand the `@psalm-var callable() $var` annotation
         /** @psalm-suppress MissingClosureParamType, MissingClosureReturnType */
-        $group = function($text, $group) {
-            return '(?<'.(string)$group.'>'.preg_replace('/(.)/us', '\\\\$0', (string)$text).')';
+        $group = function($text) {
+            return preg_replace('/(.)/us', '\\\\$0', (string)$text);
         };
         /** @psalm-suppress MissingClosureParamType, MissingClosureReturnType */
         $quote = function($text) {
             return preg_replace('/(.)/us', '\\\\$0', (string)$text);
         };
 
         $rules = array(
-            '(?<string>\\\\.|(?:(?!'.implode('|', array(
+            '\\\\.(*:'.self::TOKEN_STRING.')|(?:(?!'.implode('|', array(
                 $quote($syntax->getOpeningTag()),
                 $quote($syntax->getClosingTag()),
                 $quote($syntax->getClosingTagMarker()),
                 $quote($syntax->getParameterValueSeparator()),
                 $quote($syntax->getParameterValueDelimiter()),
                 '\s+',
-            )).').)+)',
-            '(?<ws>\s+)',
-            $group($syntax->getClosingTagMarker(), 'marker'),
-            $group($syntax->getParameterValueDelimiter(), 'delimiter'),
-            $group($syntax->getParameterValueSeparator(), 'separator'),
-            $group($syntax->getOpeningTag(), 'open'),
-            $group($syntax->getClosingTag(), 'close'),
+            )).').)+(*:'.self::TOKEN_STRING.')',
+            '\s+(*:'.self::TOKEN_WS.')',
+            $group($syntax->getClosingTagMarker()).'(*:'.self::TOKEN_MARKER.')',
+            $group($syntax->getParameterValueDelimiter()).'(*:'.self::TOKEN_DELIMITER.')',
+            $group($syntax->getParameterValueSeparator()).'(*:'.self::TOKEN_SEPARATOR.')',
+            $group($syntax->getOpeningTag()).'(*:'.self::TOKEN_OPEN.')',
+            $group($syntax->getClosingTag()).'(*:'.self::TOKEN_CLOSE.')',
         );
 
-        return '~('.implode('|', $rules).')~us';
+        return '~(?|'.implode('|', $rules).')~us';
     }
 }

tests/ParserTest.php

@@ -318,4 +318,42 @@ public function testInstances()
         static::assertInstanceOf('Thunder\Shortcode\Parser\WordPressParser', new WordpressParser());
         static::assertInstanceOf('Thunder\Shortcode\Parser\RegularParser', new RegularParser());
     }
+
+    /** @dataProvider provideBenchmarks */
+    public function testGeneratedNesting($level, $num)
+    {
+        $parser = new RegularParser();
+        $text = $this->benchmark($level, $num);
+
+        $time = microtime(true);
+        $list = $parser->parse($text);
+        $time = (microtime(true) - $time) * 1000; // ms
+
+        self::assertLessThan(12, $time);
+        self::assertCount(1, $list);
+    }
+
+    private function benchmark($level, $num)
+    {
+        for($i = 1; $i <= $level; $i++) {
+            $text = str_repeat('[s'.$i.']'.$i, $num);
+        }
+        for($i = $level; $i >= 1; $i--) {
+            $text .= str_repeat($i.'[/s'.$i.']', $num);
+        }
+
+        return $text;
+    }
+
+    public static function provideBenchmarks()
+    {
+        $cases = array();
+        foreach(range(10, 100, 10) as $level) {
+            foreach(range(1, 10, 1) as $num) {
+                $cases[] = [$level, $num];
+            }
+        }
+
+        return $cases;
+    }
 }