Hi, We just noticed that the Content-Length header is being incorrectly set (and incorrectly checked if user-supplied) if the POSTed data contains two-byte unicode characters (we are sending utf-8 encoded xml documents via post to an https server)
This is because perl's length command defaults to counting characters and not bytes by placing 'use bytes' at the top of ../5.6.1/LWP/Protocol/http.pm the problem is fixed and since content-length is always in bytes, this should become a permanent part of the length calculation. Cheers, Marcus Taylor Semantico
