Lexer.php

<?php declare(strict_types = 1);

namespace PHPStan\PhpDocParser\Lexer;

use PHPStan\PhpDocParser\ParserConfig;
use function implode;
use function preg_match_all;

/**
 * Implementation based on Nette Tokenizer (New BSD License; https://github.com/nette/tokenizer)
 */
class Lexer
{

	public const TOKEN_REFERENCE = 0;
	public const TOKEN_UNION = 1;
	public const TOKEN_INTERSECTION = 2;
	public const TOKEN_NULLABLE = 3;
	public const TOKEN_OPEN_PARENTHESES = 4;
	public const TOKEN_CLOSE_PARENTHESES = 5;
	public const TOKEN_OPEN_ANGLE_BRACKET = 6;
	public const TOKEN_CLOSE_ANGLE_BRACKET = 7;
	public const TOKEN_OPEN_SQUARE_BRACKET = 8;
	public const TOKEN_CLOSE_SQUARE_BRACKET = 9;
	public const TOKEN_COMMA = 10;
	public const TOKEN_VARIADIC = 11;
	public const TOKEN_DOUBLE_COLON = 12;
	public const TOKEN_DOUBLE_ARROW = 13;
	public const TOKEN_EQUAL = 14;
	public const TOKEN_OPEN_PHPDOC = 15;
	public const TOKEN_CLOSE_PHPDOC = 16;
	public const TOKEN_PHPDOC_TAG = 17;
	public const TOKEN_DOCTRINE_TAG = 18;
	public const TOKEN_FLOAT = 19;
	public const TOKEN_INTEGER = 20;
	public const TOKEN_SINGLE_QUOTED_STRING = 21;
	public const TOKEN_DOUBLE_QUOTED_STRING = 22;
	public const TOKEN_DOCTRINE_ANNOTATION_STRING = 23;
	public const TOKEN_IDENTIFIER = 24;
	public const TOKEN_THIS_VARIABLE = 25;
	public const TOKEN_VARIABLE = 26;
	public const TOKEN_HORIZONTAL_WS = 27;
	public const TOKEN_PHPDOC_EOL = 28;
	public const TOKEN_OTHER = 29;
	public const TOKEN_END = 30;
	public const TOKEN_COLON = 31;
	public const TOKEN_WILDCARD = 32;
	public const TOKEN_OPEN_CURLY_BRACKET = 33;
	public const TOKEN_CLOSE_CURLY_BRACKET = 34;
	public const TOKEN_NEGATED = 35;
	public const TOKEN_ARROW = 36;

	public const TOKEN_COMMENT = 37;

	public const TOKEN_LABELS = [
		self::TOKEN_REFERENCE => '\'&\'',
		self::TOKEN_UNION => '\'|\'',
		self::TOKEN_INTERSECTION => '\'&\'',
		self::TOKEN_NULLABLE => '\'?\'',
		self::TOKEN_NEGATED => '\'!\'',
		self::TOKEN_OPEN_PARENTHESES => '\'(\'',
		self::TOKEN_CLOSE_PARENTHESES => '\')\'',
		self::TOKEN_OPEN_ANGLE_BRACKET => '\'<\'',
		self::TOKEN_CLOSE_ANGLE_BRACKET => '\'>\'',
		self::TOKEN_OPEN_SQUARE_BRACKET => '\'[\'',
		self::TOKEN_CLOSE_SQUARE_BRACKET => '\']\'',
		self::TOKEN_OPEN_CURLY_BRACKET => '\'{\'',
		self::TOKEN_CLOSE_CURLY_BRACKET => '\'}\'',
		self::TOKEN_COMMA => '\',\'',
		self::TOKEN_COMMENT => '\'//\'',
		self::TOKEN_COLON => '\':\'',
		self::TOKEN_VARIADIC => '\'...\'',
		self::TOKEN_DOUBLE_COLON => '\'::\'',
		self::TOKEN_DOUBLE_ARROW => '\'=>\'',
		self::TOKEN_ARROW => '\'->\'',
		self::TOKEN_EQUAL => '\'=\'',
		self::TOKEN_OPEN_PHPDOC => '\'/**\'',
		self::TOKEN_CLOSE_PHPDOC => '\'*/\'',
		self::TOKEN_PHPDOC_TAG => 'TOKEN_PHPDOC_TAG',
		self::TOKEN_DOCTRINE_TAG => 'TOKEN_DOCTRINE_TAG',
		self::TOKEN_PHPDOC_EOL => 'TOKEN_PHPDOC_EOL',
		self::TOKEN_FLOAT => 'TOKEN_FLOAT',
		self::TOKEN_INTEGER => 'TOKEN_INTEGER',
		self::TOKEN_SINGLE_QUOTED_STRING => 'TOKEN_SINGLE_QUOTED_STRING',
		self::TOKEN_DOUBLE_QUOTED_STRING => 'TOKEN_DOUBLE_QUOTED_STRING',
		self::TOKEN_DOCTRINE_ANNOTATION_STRING => 'TOKEN_DOCTRINE_ANNOTATION_STRING',
		self::TOKEN_IDENTIFIER => 'type',
		self::TOKEN_THIS_VARIABLE => '\'$this\'',
		self::TOKEN_VARIABLE => 'variable',
		self::TOKEN_HORIZONTAL_WS => 'TOKEN_HORIZONTAL_WS',
		self::TOKEN_OTHER => 'TOKEN_OTHER',
		self::TOKEN_END => 'TOKEN_END',
		self::TOKEN_WILDCARD => '*',
	];

	public const VALUE_OFFSET = 0;
	public const TYPE_OFFSET = 1;
	public const LINE_OFFSET = 2;

	private ParserConfig $config; // @phpstan-ignore property.onlyWritten

	private ?string $regexp = null;

	public function __construct(ParserConfig $config)
	{
		$this->config = $config;
	}

	/**
	 * @return list<array{string, int, int}>
	 */
	public function tokenize(string $s): array
	{
		if ($this->regexp === null) {
			$this->regexp = $this->generateRegexp();
		}

		// PREG_PATTERN_ORDER, not PREG_SET_ORDER: it collects the whole PHPDoc
		// into two arrays instead of allocating one array per token. This only
		// pays off while the token patterns have no capturing groups, because
		// every group would get an array of its own, one entry per token.
		preg_match_all($this->regexp, $s, $matches);

		$values = $matches[0];
		if ($values === []) {
			return [['', self::TOKEN_END, 1]];
		}

		$marks = $matches['MARK'];

		$tokens = [];
		$line = 1;
		foreach ($values as $i => $value) {
			$type = (int) $marks[$i];
			$tokens[] = [$value, $type, $line];
			if ($type !== self::TOKEN_PHPDOC_EOL) {
				continue;
			}

			$line++;
		}

		$tokens[] = ['', self::TOKEN_END, $line];

		return $tokens;
	}

	private function generateRegexp(): string
	{
		// every group in here must be non-capturing, see tokenize()
		$patterns = [
			self::TOKEN_HORIZONTAL_WS => '[\\x09\\x20]++',

			self::TOKEN_IDENTIFIER => '(?:[\\\\]?+[a-z_\\x80-\\xFF][0-9a-z_\\x80-\\xFF-]*+)++',
			self::TOKEN_THIS_VARIABLE => '\\$this(?![0-9a-z_\\x80-\\xFF])',
			self::TOKEN_VARIABLE => '\\$[a-z_\\x80-\\xFF][0-9a-z_\\x80-\\xFF]*+',

			// '&' followed by TOKEN_VARIADIC, TOKEN_VARIABLE, TOKEN_EQUAL, TOKEN_EQUAL or TOKEN_CLOSE_PARENTHESES
			self::TOKEN_REFERENCE => '&(?=\\s*+(?:[,=)]|\\.\\.\\.|(?:\\$(?!this(?![0-9a-z_\\x80-\\xFF])))))',
			self::TOKEN_UNION => '\\|',
			self::TOKEN_INTERSECTION => '&',
			self::TOKEN_NULLABLE => '\\?',
			self::TOKEN_NEGATED => '!',

			self::TOKEN_OPEN_PARENTHESES => '\\(',
			self::TOKEN_CLOSE_PARENTHESES => '\\)',
			self::TOKEN_OPEN_ANGLE_BRACKET => '<',
			self::TOKEN_CLOSE_ANGLE_BRACKET => '>',
			self::TOKEN_OPEN_SQUARE_BRACKET => '\\[',
			self::TOKEN_CLOSE_SQUARE_BRACKET => '\\]',
			self::TOKEN_OPEN_CURLY_BRACKET => '\\{',
			self::TOKEN_CLOSE_CURLY_BRACKET => '\\}',

			self::TOKEN_COMMA => ',',
			self::TOKEN_COMMENT => '\/\/[^\\r\\n]*(?=\n|\r|\*/)',
			self::TOKEN_VARIADIC => '\\.\\.\\.',
			self::TOKEN_DOUBLE_COLON => '::',
			self::TOKEN_DOUBLE_ARROW => '=>',
			self::TOKEN_ARROW => '->',
			self::TOKEN_EQUAL => '=',
			self::TOKEN_COLON => ':',

			self::TOKEN_OPEN_PHPDOC => '/\\*\\*(?=\\s)\\x20?+',
			self::TOKEN_CLOSE_PHPDOC => '\\*/',
			self::TOKEN_PHPDOC_TAG => '@(?:[a-z][a-z0-9-\\\\]+:)?[a-z][a-z0-9-\\\\]*+',
			self::TOKEN_DOCTRINE_TAG => '@[a-z_\\\\][a-z0-9_\:\\\\]*[a-z_][a-z0-9_]*',
			self::TOKEN_PHPDOC_EOL => '\\r?+\\n[\\x09\\x20]*+(?:\\*(?!/)\\x20?+)?',

			self::TOKEN_FLOAT => '[+\-]?(?:(?:[0-9]++(?:_[0-9]++)*\\.[0-9]*+(?:_[0-9]++)*(?:e[+\-]?[0-9]++(?:_[0-9]++)*)?)|(?:[0-9]*+(?:_[0-9]++)*\\.[0-9]++(?:_[0-9]++)*(?:e[+\-]?[0-9]++(?:_[0-9]++)*)?)|(?:[0-9]++(?:_[0-9]++)*e[+\-]?[0-9]++(?:_[0-9]++)*))',
			self::TOKEN_INTEGER => '[+\-]?(?:(?:0b[0-1]++(?:_[0-1]++)*)|(?:0o[0-7]++(?:_[0-7]++)*)|(?:0x[0-9a-f]++(?:_[0-9a-f]++)*)|(?:[0-9]++(?:_[0-9]++)*))',
			self::TOKEN_SINGLE_QUOTED_STRING => '\'(?:\\\\[^\\r\\n]|[^\'\\r\\n\\\\])*+\'',
			self::TOKEN_DOUBLE_QUOTED_STRING => '"(?:\\\\[^\\r\\n]|[^"\\r\\n\\\\])*+"',
			self::TOKEN_DOCTRINE_ANNOTATION_STRING => '"(?:""|[^"])*+"',

			self::TOKEN_WILDCARD => '\\*',

			// anything but TOKEN_CLOSE_PHPDOC or TOKEN_HORIZONTAL_WS or TOKEN_EOL
			self::TOKEN_OTHER => '(?:(?!\\*/)[^\\s])++',
		];

		foreach ($patterns as $type => &$pattern) {
			$pattern = '(?:' . $pattern . ')(*MARK:' . $type . ')';
		}

		return '~' . implode('|', $patterns) . '~Asi';
	}

}