Tolerant Mode

Edit pageLast modified: 02 February 2025

The "tolerant" analysis mode allows parsing a type grammar containing other arbitrary data.

Such a mode will be convenient to use, for example, for analyzing phpdoc docblocks, separating types from their descriptions.

$parser = new TypeLang\Parser\Parser(
    // enable "tolerant" mode
    tolerant: true,
);

As an example, let's try to parse the contents of the " @return " docblock. Similar actions can be implemented for other annotations.

$parser = new TypeLang\Parser\Parser(
    tolerant: true,
);

$content = <<<'PHP'
    @return Example<T> Returns something very interesting!
    PHP;

// There is no need to analyze the "@return" annotation,
// we only get the content:
//   "Example<T> Returns something very interesting!"
$content = substr($content, strlen('@return '));

var_dump($parser->parse($content));
TypeLang\Parser\Node\Stmt\NamedTypeNode
 {...}

The Parser::$lastProcessedTokenOffset contains the offset in bytes (in this case 11) at which unparsable content begins (description of " @return " annotation).

// Next we can get the contents of the docblock description by simply
// getting it by this offset.
$description = substr($content, $parser->lastProcessedTokenOffset);

The $description will contain:

string("Returns something very interesting!")