A simple problem, quite irritating: I use Kernow to process data from several languages, most importantly lower Sorbian, a small minority language spoken in Germany, and rely on <xsl:message/> to give warnings and error messages to the user.
However, unicode characters are either rendered as blank boxes or as entities (such as ś). This makes some error messages unintelligible. Wouldn't that perhaps be easy to fix?
I can be reached at my firstname, dot, last name followed by @gmx.net.