The problem is with many PDF viewers not rendering those properly and my current software stack not handling this issue. The characters do in fact get filled into the PDF, however, there is an issue with the field’s appearance stream. Simple PDF viewers (such as web browsers) don’t have the code to work around the issue, so, those characters don’t render properly. If you were to open the filled PDF file with a more professional PDF software then you would see that they render the filled data correctly.
You may be able fix the problem by fully embedding the font used in the fields into your PDF, making sure that it contains glyphs for all of your characters. Unfortunately, that my not always work.
I am working on a solution to fix this issue. I plan to have something available for testing soon.