PHP RFC: new MyClass()->method() without parentheses
- Date: 2023-12-29
- Author: Valentin Udaltsov (udaltsov.valentin@gmail.com)
- Status: Implemented
- Target Version: PHP 8.4
- Discussion: https://externals.io/message/123031
- First Published at: http://wiki.php.net/rfc/new_without_parentheses
- Implementation: https://github.com/php/php-src/pull/13029
Introduction
The “class member access on instantiation” feature
was introduced in PHP 5.4.0.
Since then constants, properties and methods can be accessed on a newly created instance without an intermediate
variable, but only if the new expression is wrapped in parentheses:
class Request implements Psr\Http\Message\RequestInterface { // ... } // OK $request = (new Request())->withMethod('GET')->withUri('/hello-world'); // PHP Parse error: syntax error, unexpected token "->" $request = new Request()->withMethod('GET')->withUri('/hello-world');
The goal of this RFC is to enable the second syntax to:
- make coding with PHP more convenient and satisfy many users' requests for this feature (see externals#66197, bugs.php#70549, externals#101811, externals#113953)
- decrease the visual debt in all sorts of builders and configurators (more than half a million lines of open source PHP code could be simplified)
- ease switching between other C-like languages that don't require parentheses (Java, C#, TypeScript).
Proposal
This RFC allows to omit parentheses around the new expression with constructor arguments' parentheses. Given class
class MyClass extends ArrayObject { const CONSTANT = 'constant'; public static $staticProperty = 'staticProperty'; public static function staticMethod(): string { return 'staticMethod'; } public $property = 'property'; public function method(): string { return 'method'; } public function __invoke(): string { return '__invoke'; } }
one will be able to write
var_dump( new MyClass()::CONSTANT, // string(8) "constant" new MyClass()::$staticProperty, // string(14) "staticProperty" new MyClass()::staticMethod(), // string(12) "staticMethod" new MyClass()->property, // string(8) "property" new MyClass()->method(), // string(6) "method" new MyClass()(), // string(8) "__invoke" new MyClass(['value'])[0], // string(5) "value" ); $myClass = MyClass::class; var_dump( new $myClass()::CONSTANT, // string(8) "constant" new $myClass()::$staticProperty, // string(14) "staticProperty" new $myClass()::staticMethod(), // string(12) "staticMethod" new $myClass()->property, // string(8) "property" new $myClass()->method(), // string(6) "method" new $myClass()(), // string(8) "__invoke" new $myClass(['value'])[0], // string(5) "value" ); var_dump( new (trim(' MyClass '))()::CONSTANT, // string(8) "constant" new (trim(' MyClass '))()::$staticProperty, // string(14) "staticProperty" new (trim(' MyClass '))()::staticMethod(), // string(12) "staticMethod" new (trim(' MyClass '))()->property, // string(8) "property" new (trim(' MyClass '))()->method(), // string(6) "method" new (trim(' MyClass '))()(), // string(8) "__invoke" new (trim(' MyClass '))(['value'])[0], // string(5) "value" );
This RFC does not change behavior of new expressions without constructor arguments' parentheses:
new MyClass::CONSTANT; // will continue to throw a syntax error new $myClass::CONSTANT; // will continue to throw a syntax error new MyClass::$staticProperty; // will continue to work as `new (MyClass::$staticProperty)` new $myClass::$staticProperty; // will continue to work as `new ($myClass::$staticProperty)` new $myObject->property; // will continue to work as `new ($myObject->property)` new MyArrayConst['class']; // will continue to throw a syntax error new $myArray['class']; // will continue to work as `new ($myArray['class'])`
This RFC allows to omit parentheses around the new anonymous class expression regardless of whether constructor arguments'
parentheses are present or not:
var_dump( // string(8) "constant" new class { const CONSTANT = 'constant'; }::CONSTANT, // string(14) "staticProperty" new class { public static $staticProperty = 'staticProperty'; }::$staticProperty, // string(12) "staticMethod" new class { public static function staticMethod() { return 'staticMethod'; } }::staticMethod(), // string(8) "property" new class { public $property = 'property'; }->property, // string(6) "method" new class { public function method() { return 'method'; } }->method(), // string(8) "__invoke" new class { public function __invoke() { return '__invoke'; } }(), // string(5) "value" new class (['value']) extends ArrayObject {}[0], );
Why the proposed syntax is unambiguous
At first glance new MyClass()->method() expression might seem ambiguous. But according to the same logic new MyClass() is also ambiguous:
is it new (MyClass()) or new MyClass? However, new MyClass() is unambiguously interpreted as an instantiation
of class MyClass with zero constructor arguments, not an instantiation of MyClass() function call result. This
is because PHP interprets the first expression after new as a class name.
Consider also MyClass::new()->method(). It is not interpreted as MyClass::(new()->method()) or MyClass::(new())->method().
It is interpreted as (MyClass::new())->method(). It's natural to assume that MyClass::new() is the first expression to evaluate.
The same is true for new MyClass()->method(): expression new MyClass() is also the first to evaluate.
Here's how it looks at the grammar level.
The formula for the new expression with explicit constructor arguments' parentheses is T_NEW class_name|new_variable|(expr) ctor_arguments,
where new_variable is a variable expression without calls.
new MyClass(); ^ ^ ^ | | | |—T_NEW |—ctor_arguments | |—class_name new $class(); ^ ^ ^ | | | |—T_NEW |—ctor_arguments | |—new_variable (cannot have calls!) new (trim(' MyClass '))(); ^ ^ ^ | | | |—T_NEW |—ctor_arguments | |—(expr)
Once the parser encounters T_NEW, it considers whatever comes next a class name, not a part of another expression.
This guarantees that the new expression with arguments' parentheses can be unambiguously used as a separate expression
without any additional parentheses. It also becomes clear why new_variable cannot have calls and why arguments' parentheses
are crucial for the proposed syntax: parentheses denote the end of the class name and the end of the new expression.
Other syntax ideas
Some of the ideas expressed during the discussion of this RFC are listed below. They are orthogonal to this proposal and require a separate RFC.
Allow to omit the new keyword
Some languages like Kotlin allow to instantiate classes via MyClass() expression. Omitting the “new” keyword in PHP is currently not possible, because classes and functions can have the same name (in Kotlin it is not possible). So, in order to achieve this in PHP, we should first deprecate declaring functions and classes with the same names.
MyClass::new(), MyClass::create()
Introducing a dedicated static constructor like MyClass::new() or MyClass::create() would be a backward compatibility
break for already existing static or object methods named new or create with required parameters and/or return
type different from static. So this idea also requires some deprecations.
Backward Incompatible Changes
None. Any code that is valid before the change will be valid after and interpreted in the same way.
Proposed PHP Version(s)
PHP 8.4
Proposed Voting Choices
This is a simple yes-or-no vote to include this feature. 2/3 majority required to pass.
Voting started on 2024-05-09 and will end on 2024-05-24 00:00 GMT.
Implementation
Pull request contains the final implementation and plenty of tests asserting the expected behavior and backward compatibility: https://github.com/php/php-src/pull/13029