Bug #78101 [ReO->Fbk]: Icorrect behaviour for substr on Windows

Edit report at https://bugs.php.net/bug.php?id=78101&edit=1 ID: 78101 Updated by: cmb@php.net Reported by: mishamosher at gmail dot com Summary: Icorrect behaviour for substr on Windows -Status: Re-Opened +Status: Feedback Type: Bug Package: opcache Operating System: Windows 10 1903 64-bits PHP Version: 7.2.19 -Assigned To: +Assigned To: cmb Block user comment: N Private report: N New Comment: Please run the following variation of your script, and report the results: <?php $in = "/es"; $out = substr($in, 0, 1); var_dump(bin2hex($in), bin2hex($out)); ?> Previous Comments: ------------------------------------------------------------------------ [2019-06-04 01:56:44] requinix@php.net If removing opcache makes your problem go away then this may be a bug after all. ------------------------------------------------------------------------ [2019-06-04 01:00:35] mishamosher at gmail dot com Hello subjective@php.net, Are your results from running the script with PHP as an Apache module? I've thoroughly tried over and over again, and no matter what, I always get the wrong results. var_dump reports <<string(3) "/es">>. I've been able to pin-point the issue to the OPCache extension. If I simply remove "zend_extension=opcache" from php.ini, everything works as it should. Please note that I'm experiencing this issue only with PHP 7.2.19 when it is working as an Apache module. I can use PHP 7.2.11, 5.6.40, 7.3.6, etc., with OPCache enabled and everything works OK in there. As a final note, what can I do to 'ring' attention to an existing filed bug? It is about this one: https://bugs.php.net/bug.php?id=77014 - at least to know if it is a bug or intended behaviour. ------------------------------------------------------------------------ [2019-06-03 21:24:05] subjective@php.net I tested it on Windows using the following script: var_dump( substr("/es", 0, 1) ); var_dump( substr("\xEF\xBB\xBF/es", 0, 1) ); Output: string(1) "/" string(1) "�" In general try using var_dump() on the source string. It will report the byte length, so you will notice invisible characters. ------------------------------------------------------------------------ [2019-06-03 20:55:31] mishamosher at gmail dot com Hello requinix@php.net, I appreciate your attention into the matter. I know for sure that there are no multibyte strings or invisible characters in my testing. Functions strlen, mb_strlen and mb_detect_encoding return 3, 3 and "ASCII" respectively. In the mantime, I've just finished configuring PHP 7.2.11 and PHP 7.3.6, as Apache modules, and they both do not exibit this issue with the very same test file and test environment. The real mistery for me is why specifically PHP 7.2.19 does not want to work as it should. Please note that the issue presents itself only when I run PHP 7.2.19 as an Apache module. The issue does not occur when I run PHP via CMD (php.exe -f script.php). Is there any specific test or debug information that I can provide to further identify the issue? ------------------------------------------------------------------------ [2019-06-03 19:23:35] requinix@php.net Either you typed an invisible character (copy/pasted a BOM?) or that slash is a Unicode slash made up of multiple bytes. If you intend to support multibyte strings and non-ASCII encodings then use mbstring and mb_substr() instead. ------------------------------------------------------------------------ The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at https://bugs.php.net/bug.php?id=78101 -- Edit this bug report at https://bugs.php.net/bug.php?id=78101&edit=1