AutoLoader: Skip tokenizing of irrelevant lines in ClassCollector

This makes AutoLoaderStructureTest in PHPUnit and the
generateLocalAutoload.php maintenance script much faster.

On my machine, it made it 35X faster (or time spent reduced by 97%).

Bug: T225730
Change-Id: Ife959bd17ce9c2ae952dfbd158ddb3d8475e8cb2

1 files changed, 25 insertions, 0 deletions

diff --git a/includes/utils/ClassCollector.php b/includes/utils/ClassCollector.php
index c987354d3737..12b8a707bd6f 100644
--- a/includes/utils/ClassCollector.php
+++ b/includes/utils/ClassCollector.php
@@ -59,6 +59,31 @@ class ClassCollector {
 		$this->alias = null;
 		$this->tokens = [];
 
+		// HACK: The PHP tokenizer is slow (T225730).
+		// Speed it up by reducing the input to the three kinds of statement we care about:
+		// - namespace X;
+		// - [final] [abstract] class X … {}
+		// - class_alias( … );
+		$lines = [];
+		$matches = null;
+		preg_match_all(
+			// phpcs:ignore Generic.Files.LineLength.TooLong
+			'#^\t*(?:namespace |(final )?(abstract )?(class|interface|trait) |class_alias\()[^;{]+[;{]\s*\}?#m',
+			$code,
+			$matches
+		);
+		if ( isset( $matches[0][0] ) ) {
+			foreach ( $matches[0] as $match ) {
+				$match = trim( $match );
+				if ( substr( $match, -1 ) === '{' ) {
+					// Keep it balanced
+					$match .= '}';
+				}
+				$lines[] = $match;
+			}
+		}
+		$code = '<?php ' . implode( "\n", $lines ) . "\n";
+
 		foreach ( token_get_all( $code ) as $token ) {
 			if ( $this->startToken === null ) {
 				$this->tryBeginExpect( $token );