preg_split sometimes returns array of arrays

When PREG_SPLIT_OFFSET_CAPTURE is given, "this changes the return value in an array where every element is an array consisting of the matched string at offset 0 and its string offset into subject at offset 1."

Author: spaze

conf/config.neon

@@ -895,6 +895,11 @@ services:
 		tags:
 			- phpstan.broker.dynamicFunctionReturnTypeExtension
 
+	-
+		class: PHPStan\Type\Php\PregSplitDynamicReturnTypeExtension
+		tags:
+			- phpstan.broker.dynamicFunctionReturnTypeExtension
+
 	-
 		class: PHPStan\Type\Php\ReplaceFunctionsDynamicReturnTypeExtension
 		tags:

src/Type/Php/PregSplitDynamicReturnTypeExtension.php

@@ -0,0 +1,77 @@
+<?php declare(strict_types = 1);
+
+namespace PHPStan\Type\Php;
+
+use PhpParser\Node\Arg;
+use PhpParser\Node\Expr\FuncCall;
+use PhpParser\Node\Name;
+use PHPStan\Analyser\Scope;
+use PHPStan\Reflection\FunctionReflection;
+use PHPStan\Reflection\ParametersAcceptorSelector;
+use PHPStan\Reflection\ReflectionProvider;
+use PHPStan\Type\ArrayType;
+use PHPStan\Type\Constant\ConstantArrayType;
+use PHPStan\Type\Constant\ConstantBooleanType;
+use PHPStan\Type\Constant\ConstantIntegerType;
+use PHPStan\Type\DynamicFunctionReturnTypeExtension;
+use PHPStan\Type\IntegerType;
+use PHPStan\Type\StringType;
+use PHPStan\Type\Type;
+use PHPStan\Type\TypeCombinator;
+
+class PregSplitDynamicReturnTypeExtension implements DynamicFunctionReturnTypeExtension
+{
+
+	private ReflectionProvider $reflectionProvider;
+
+	public function __construct(ReflectionProvider $reflectionProvider)
+	{
+		$this->reflectionProvider = $reflectionProvider;
+	}
+
+
+	public function isFunctionSupported(FunctionReflection $functionReflection): bool
+	{
+		return strtolower($functionReflection->getName()) === 'preg_split';
+	}
+
+
+	public function getTypeFromFunctionCall(FunctionReflection $functionReflection, FuncCall $functionCall, Scope $scope): Type
+	{
+		$flagsArg = $functionCall->args[3] ?? null;
+
+		if ($this->hasFlag($this->getConstant('PREG_SPLIT_OFFSET_CAPTURE'), $flagsArg, $scope)) {
+			$type = new ArrayType(
+				new IntegerType(),
+				new ConstantArrayType([new ConstantIntegerType(0), new ConstantIntegerType(1)], [new StringType(), new IntegerType()])
+			);
+			return TypeCombinator::union($type, new ConstantBooleanType(false));
+		}
+
+		return ParametersAcceptorSelector::selectSingle($functionReflection->getVariants())->getReturnType();
+	}
+
+
+	private function hasFlag(int $flag, ?Arg $expression, Scope $scope): bool
+	{
+		if ($expression === null) {
+			return false;
+		}
+
+		$type = $scope->getType($expression->value);
+		return $type instanceof ConstantIntegerType && ($type->getValue() & $flag) === $flag;
+	}
+
+
+	private function getConstant(string $constantName): int
+	{
+		$constant = $this->reflectionProvider->getConstant(new Name($constantName), null);
+		$valueType = $constant->getValueType();
+		if (!$valueType instanceof ConstantIntegerType) {
+			throw new \PHPStan\ShouldNotHappenException(sprintf('Constant %s does not have integer type.', $constantName));
+		}
+
+		return $valueType->getValue();
+	}
+
+}

tests/PHPStan/Analyser/NodeScopeResolverTest.php

@@ -10048,6 +10048,11 @@ public function dataBug2899(): array
 		return $this->gatherAssertTypes(__DIR__ . '/data/bug-2899.php');
 	}
 
+	public function dataPregSplitReturnType(): array
+	{
+		return $this->gatherAssertTypes(__DIR__ . '/data/preg_split.php');
+	}
+
 	/**
 	 * @dataProvider dataBug2574
 	 * @dataProvider dataBug2577
@@ -10116,6 +10121,7 @@ public function dataBug2899(): array
 	 * @dataProvider dataBug3133
 	 * @dataProvider dataBug2550
 	 * @dataProvider dataBug2899
+	 * @dataProvider dataPregSplitReturnType
 	 * @param string $assertType
 	 * @param string $file
 	 * @param mixed ...$args

tests/PHPStan/Analyser/data/preg_split.php

@@ -0,0 +1,8 @@
+<?php
+
+use function PHPStan\Analyser\assertType;
+
+assertType('array<int, string>|false', preg_split('/-/', '1-2-3'));
+assertType('array<int, string>|false', preg_split('/-/', '1-2-3', -1, PREG_SPLIT_NO_EMPTY));
+assertType('array<int, array(string, int)>|false', preg_split('/-/', '1-2-3', -1, PREG_SPLIT_OFFSET_CAPTURE));
+assertType('array<int, array(string, int)>|false', preg_split('/-/', '1-2-3', -1, PREG_SPLIT_NO_EMPTY | PREG_SPLIT_OFFSET_CAPTURE));