rfc:in_operator

This is an old revision of the document!


PHP RFC: In Operator

Introduction

This RFC adds a new in operator which simplifies contains checks for strings and arrays. Currently, we have to use in_array($needle, $haystack, true) or strpos($haystack, $needle) !== false. These functions have a inconsistent parameter order, so it's hard to remember which is the right one for each. The in operator makes these checks way more readable and lowers the cognitive load. Additionally, it also works for Traversable.

Proposal

Add a new operator (expr1) in (expr2). It checks whether expr2 contains expr1.

It uses strict comparison (===) for array values / instances of Traversable and doesn't search recursively.

$contains = "foo" in ["a", "b", "c"]; // false
$contains = "foo" in ["foo", "bar"]; // true
$contains = "foo" in [["foo"], ["bar"]]; // false
$contains = "0e0" in ["0"]; // false, because of strict comparison
$contains = 0 in ["0"]; // false, because of strict comparison
 
$contains = ["foo"] in [["foo"], ["bar"]]; // true
$contains = ["foo"] in ["foo"]; // false

Traversables are only iterated until there's a match.

function gen () {
    yield "foo";
    yield "bar";
    // code below here wouldn't be executed if "bar" matches
    // because it stops if there's a match.
}
 
$contains = "bar" in gen(); // true
$contains = "baz" in gen(); // false

If $haystack is a string, it's a simple contains check.

$contains = "foo" in "foobar"; // true
$contains = "php" in "foobar"; // false

Other expressions than mixed in array|Traversable or string in string throw an EngineException.

Why strict?

It's strict because otherwise something like “foo” in [0] would pass.

Precedence

It should have the same precedence as instanceof, so it's possible to negate it:

if (!$input in $validValues) {
    // ...
}

Backward Incompatible Changes

New reserved keyword in. This affects function, constant, class and method names.

Proposed PHP Version(s)

Next major release, at the time of writing PHP 7.

RFC Impact

New Constants

A T_IN constant for use with ext/tokenizer has been added.

Future Scope

There could be a syntax that allows to check for multiple values at once, e.g.

$contains = ...["foo", "bar"] in ["foo", "baz", "bar"];

Votes

Requires a 2/3 majority. Even if it passes, it will only get merged if the context sensitive lexer gets merged.

Patches and Tests

Rejected Features

Keep this updated with features that were discussed on the mail lists.

Changelog

  • v0.5: Removed integer support, so the strictness is consistent.
  • v0.4: Removed possibility to check multiple values using an array.
rfc/in_operator.1426379832.txt.gz · Last modified: 2017/09/22 13:28 (external edit)