`mb_detect_encoding()` does not terminate when called with BOM sequence for UTF-8

[sebastianbergmann]

Description

This call to mb_detect_encoding() does not terminate with PHP 8.3:

var_dump(mb_detect_encoding("\xEF\xBB\xBF", null, true));

With PHP 8.2, the code shown above prints string(5) "UTF-8".

This was originally reported by @nyamsprod for PHPUnit as sebastianbergmann/phpunit#5546.

I did not have success finding the first commit that causes this using git bisect as there were commits between PHP-8.2 and PHP-8.3 for which the build fails (at least on my machine).

PHP Version

PHP 8.3.0-dev (0c22276)

Operating System

Fedora 38

[alexdowad]

Interesting! Let me try it.

[alexdowad]

Excellent catch, sorry for my mistake. I have a fix.

[alexdowad]

I just need to write a good commit message and add a unit test. This should be merged into PHP-8.3 very soon.

[alexdowad]

Just opened #12541

[alexdowad]

The fix has landed on the PHP-8.3 and master branches of php-src. It will be incorporated in the next PHP releases in the 8.3 and 8.4 series.