Man, perlmonk dose not support unicode either

Yes and no. The form contents must be encoded using iso-8859-1, but since PerlMonks accepts HTML, you can use entities (& sequences) to represent any unicode character.

You tried to post characters outside of iso-8859-1, so your browser submitted than as entities. The problem is that those entities were inside code tags, so they were presumed to be text, not HTML.

测试

As for perldoc, perldoc -t worked... kinda. It issued a warning and it will only work if you have a UTF-8 terminal.

I don't know if the failure to work without -t is a bug, but it sounds like it from the earlier reply to your post. There's no excuse for -t's behaviour, though. A bug report is in order.