In 2014, @maximpodorov
asked →
BTW, is it proper to use wordwrap(), strlen() and other byte level function for UTF-8 texts?
I took a look and no, I don't think it's proper for Drupal\Core\Mail\MailFormatHelper::wrapMailLine() to call wordwrap() function with $cut_long_words = TRUE, because it could split multi-byte UTF-8 characters.
The MailFormatHelper class could produce invalid UTF-8 if a line is very long (greater than 996 bytes), has multi-byte characters, and has no spaces, thus triggering wrapping at non-space characters. This is especially likely to happen in languages that don't use spaces.
The second call to wordwrap() with $cut_long_words = TRUE should be replaced with something that can safely handle multi-byte characters.
Closed: duplicate
11.0 🔥
This sounded familiar and searching proved my memory correct. There is an earlier issue. #540228: Implement a UTF8-safe wordwrap() → . There is even a patch over there.
This looks like a duplicate therefor, closing as such.