Both Perl 5 and Python 3 are built with the assumption that a "character" is what the Unicode standard generally refers to as a "character".

This is a serious problem and will I suspect one day be understood to be the biggest long-term mistake made when designing Python 3.

I wrote a comment about this a few months back at https://news.ycombinator.com/item?id=9791211 (see second half of the initial comment and then the followups).