Hi Luigi,

The "tounicode" key in the font character tables also supports passing
an array of Unicode codepoints, which is _much_ nicer than trying to
reencode characters into UTF-16BE (especially for non-BMP characters).
The attached patch adds this to the manual.