Refactor RegexGroupParser for more immutability and less pass-by-ref

Author: staabm

src/Type/Regex/RegexAstWalkResult.php

@@ -0,0 +1,105 @@
+<?php declare(strict_types = 1);
+
+namespace PHPStan\Type\Regex;
+
+/** @immutable */
+final class RegexAstWalkResult
+{
+
+	/**
+	 * @param array<int, RegexCapturingGroup> $capturingGroups
+	 * @param list<string> $markVerbs
+	 */
+	public function __construct(
+		private int $alternationId,
+		private int $captureGroupId,
+		private array $capturingGroups,
+		private array $markVerbs,
+	)
+	{
+	}
+
+	public static function createEmpty(): self
+	{
+		return new self(
+			-1,
+			// use different start-index for groups to make it easier to distinguish groupids from other ids
+			100,
+			[],
+			[],
+		);
+	}
+
+	public function nextAlternationId(): self
+	{
+		return new self(
+			$this->alternationId + 1,
+			$this->captureGroupId,
+			$this->capturingGroups,
+			$this->markVerbs,
+		);
+	}
+
+	public function nextCaptureGroupId(): self
+	{
+		return new self(
+			$this->alternationId,
+			$this->captureGroupId + 1,
+			$this->capturingGroups,
+			$this->markVerbs,
+		);
+	}
+
+	public function addCapturingGroup(RegexCapturingGroup $group): self
+	{
+		$capturingGroups = $this->capturingGroups;
+		$capturingGroups[$group->getId()] = $group;
+
+		return new self(
+			$this->alternationId,
+			$this->captureGroupId,
+			$capturingGroups,
+			$this->markVerbs,
+		);
+	}
+
+	public function markVerb(string $markVerb): self
+	{
+		$verbs = $this->markVerbs;
+		$verbs[] = $markVerb;
+
+		return new self(
+			$this->alternationId,
+			$this->captureGroupId,
+			$this->capturingGroups,
+			$verbs,
+		);
+	}
+
+	public function getAlternationId(): int
+	{
+		return $this->alternationId;
+	}
+
+	public function getCaptureGroupId(): int
+	{
+		return $this->captureGroupId;
+	}
+
+	/**
+	 * @return array<int, RegexCapturingGroup>
+	 */
+	public function getCapturingGroups(): array
+	{
+		return $this->capturingGroups;
+	}
+
+	/**
+	 * @return list<string>
+	 */
+	public function getMarkVerbs(): array
+	{
+		return $this->markVerbs;
+	}
+
+}

src/Type/Regex/RegexGroupParser.php

@@ -73,51 +73,39 @@ public function parseGroups(string $regex): ?array
 			$captureOnlyNamed = str_contains($modifiers, 'n');
 		}
 
-		$capturingGroups = [];
-		$alternationId = -1;
-		$captureGroupId = 100;
-		$markVerbs = [];
-		$this->walkRegexAst(
+		$astWalkResult = $this->walkRegexAst(
 			$ast,
 			null,
-			$alternationId,
 			0,
 			false,
 			null,
-			$captureGroupId,
-			$capturingGroups,
-			$markVerbs,
 			$captureOnlyNamed,
 			false,
 			$modifiers,
+			RegexAstWalkResult::createEmpty(),
 		);
 
-		return [$capturingGroups, $markVerbs];
+		return [$astWalkResult->getCapturingGroups(), $astWalkResult->getMarkVerbs()];
 	}
 
-	/**
-	 * @param array<int, RegexCapturingGroup> $capturingGroups
-	 * @param list<string> $markVerbs
-	 */
 	private function walkRegexAst(
 		TreeNode $ast,
 		?RegexAlternation $alternation,
-		int &$alternationId,
 		int $combinationIndex,
 		bool $inOptionalQuantification,
 		RegexCapturingGroup|RegexNonCapturingGroup|null $parentGroup,
-		int &$captureGroupId,
-		array &$capturingGroups,
-		array &$markVerbs,
 		bool $captureOnlyNamed,
 		bool $repeatedMoreThanOnce,
 		string $patternModifiers,
-	): void
+		RegexAstWalkResult $astWalkResult,
+	): RegexAstWalkResult
 	{
 		$group = null;
 		if ($ast->getId() === '#capturing') {
+			$astWalkResult = $astWalkResult->nextCaptureGroupId();
+
 			$group = new RegexCapturingGroup(
-				$captureGroupId++,
+				$astWalkResult->getCaptureGroupId(),
 				null,
 				$alternation,
 				$inOptionalQuantification,
@@ -130,9 +118,11 @@ private function walkRegexAst(
 			);
 			$parentGroup = $group;
 		} elseif ($ast->getId() === '#namedcapturing') {
+			$astWalkResult = $astWalkResult->nextCaptureGroupId();
+
 			$name = $ast->getChild(0)->getValueValue();
 			$group = new RegexCapturingGroup(
-				$captureGroupId++,
+				$astWalkResult->getCaptureGroupId(),
 				$name,
 				$alternation,
 				$inOptionalQuantification,
@@ -176,40 +166,36 @@ private function walkRegexAst(
 		}
 
 		if ($ast->getId() === '#alternation') {
-			$alternationId++;
-			$alternation = new RegexAlternation($alternationId, count($ast->getChildren()));
+			$astWalkResult = $astWalkResult->nextAlternationId();
+			$alternation = new RegexAlternation($astWalkResult->getAlternationId(), count($ast->getChildren()));
 		}
 
 		if ($ast->getId() === '#mark') {
-			$markVerbs[] = $ast->getChild(0)->getValueValue();
-			return;
+			return $astWalkResult->markVerb($ast->getChild(0)->getValueValue());
 		}
 
 		if (
 			$group instanceof RegexCapturingGroup &&
 			(!$captureOnlyNamed || $group->isNamed())
 		) {
-			$capturingGroups[$group->getId()] = $group;
+			$astWalkResult = $astWalkResult->addCapturingGroup($group);
 
 			if ($alternation !== null) {
 				$alternation->pushGroup($combinationIndex, $group);
 			}
 		}
 
 		foreach ($ast->getChildren() as $child) {
-			$this->walkRegexAst(
+			$astWalkResult = $this->walkRegexAst(
 				$child,
 				$alternation,
-				$alternationId,
 				$combinationIndex,
 				$inOptionalQuantification,
 				$parentGroup,
-				$captureGroupId,
-				$capturingGroups,
-				$markVerbs,
 				$captureOnlyNamed,
 				$repeatedMoreThanOnce,
 				$patternModifiers,
+				$astWalkResult,
 			);
 
 			if ($ast->getId() !== '#alternation') {
@@ -218,6 +204,8 @@ private function walkRegexAst(
 
 			$combinationIndex++;
 		}
+
+		return $astWalkResult;
 	}
 
 	private function allowConstantTypes(