PCRE - when to use the 'u' modifier?

Whether or not the u modifier is used depends on the purpose of the regular expression and your skill in composing them.

Here is an example in which I divide a string (UTF-8) into two parts: the first character and all the others:

 <?php $str = 'абвгд'; // модификатор есть preg_match('%^(.)(.+)$%u', $str, $matches); var_dump($matches); // модификатор нет preg_match('%^(.)(.+)$%', $str, $matches); var_dump($matches); 

Result of work:

 array(3) { [0]=> string(10) "абвгд" [1]=> string(2) "а" [2]=> string(8) "бвгд" } array(3) { [0]=> string(10) "абвгд" [1]=> string(1) " " [2]=> string(9) " бвгд" } 

Here a regular season without a modifier worked with an error, since . without the modifier, u corresponds to 1 byte (and not a character) except byte x0D.

And now another example: get a substring between two brackets:

 <?php $str = 'прг[абвгд]ктм'; // модификатор есть preg_match('%\[([^\]]*)\]%u', $str, $matches); var_dump($matches); // модификатор нет preg_match('%\[([^\]]*)\]%', $str, $matches); var_dump($matches); 

Result of work:

 array(2) { [0]=> string(12) "[абвгд]" [1]=> string(10) "абвгд" } array(2) { [0]=> string(12) "[абвгд]" [1]=> string(10) "абвгд" } 

Both options work correctly, since the characters [ and ] uniquely identified in the UTF-8 encoding and their codes are not part of multibyte characters.

UPD

depending on the skill of compiling and in the first case could the result be obtained without a modifier?

 <?php $str = 'abcde'; // модификатор нет preg_match('%^([\x00-\x7F]|[\xC0-\xDF][\x80-\xBF]|[\xE0-\xEF][\x80-\xBF]{2})(.+)$%', $str, $matches); var_dump($matches); $str = 'абвгд'; // модификатор нет preg_match('%^([\x00-\x7F]|[\xC0-\xDF][\x80-\xBF]|[\xE0-\xEF][\x80-\xBF]{2})(.+)$%', $str, $matches); var_dump($matches); $str = 'Ⴀабвгд'; // модификатор нет preg_match('%^([\x00-\x7F]|[\xC0-\xDF][\x80-\xBF]|[\xE0-\xEF][\x80-\xBF]{2})(.+)$%', $str, $matches); var_dump($matches); 

Result of work:

 array(3) { [0]=> string(5) "abcde" [1]=> string(1) "a" [2]=> string(4) "bcde" } array(3) { [0]=> string(10) "абвгд" [1]=> string(2) "а" [2]=> string(8) "бвгд" } array(3) { [0]=> string(13) "Ⴀабвгд" [1]=> string(3) "Ⴀ" [2]=> string(10) "абвгд" }