strip invalid utf-8 pattern contents to keep the error message NEON parsable.

Author: staabm

e2e/bug-12629/phpstan-baseline.neon

@@ -19,13 +19,13 @@ parameters:
 			path: src/bug-12629.php
 
 		-
-			message: '#^Regex pattern is invalid\: Compilation failed\: UTF\-8 error\: byte 2 top bits not 0x80 at offset 0 in pattern\: \!Ã\[€\-Ÿ\]\!u$#'
+			message: '#^Regex pattern is invalid\: Compilation failed\: UTF\-8 error\: byte 2 top bits not 0x80 at offset 0 in pattern$#'
 			identifier: regexp.pattern
 			count: 1
 			path: src/bug-12629.php
 
 		-
-			message: '#^Regex pattern is invalid\: Compilation failed\: UTF\-8 error\: isolated byte with 0x80 bit set at offset 1 in pattern\: \!\[€\-Ÿ\]\!u$#'
+			message: '#^Regex pattern is invalid\: Compilation failed\: UTF\-8 error\: isolated byte with 0x80 bit set at offset 1 in pattern$#'
 			identifier: regexp.pattern
 			count: 1
 			path: src/bug-12629.php

src/Rules/Regexp/RegularExpressionPatternRule.php

@@ -123,6 +123,10 @@ private function validatePattern(string $pattern): ?string
 		try {
 			Strings::match('', $pattern);
 		} catch (RegexpException $e) {
+			if (str_contains($e->getMessage(), 'UTF-8 error')) {
+				// strip invalid utf-8 pattern contents to keep the error message NEON parsable.
+				return substr($e->getMessage(), 0, strrpos($e->getMessage(), ':'));
+			}
 			return $e->getMessage();
 		}