Does it support Unicode (UTF-8)?

and if yes, how?

Sure. Internally everything should be handled as UTF-8 unless there's a bug.

The problem is fonts: the default PDF fonts usually have a very small subset of UTF-8 glyphs. So, it's just a matter of embedding other fonts using a stylesheet.

In the source package you can see test_greek.txt and test_greek.style which show this issue.

If you try to render test_greek.txt with the default style, some characters (accented greek letters) will not be visible.

If you use test_greek.style, they will, because it will use the DejaVu font (you need to have this font, of course).

maybe it should use DejaVu by default? (or that's not possible??)

Or try to switch to dejavu if there's more than ascii in the source?

(just some random ideas)

It's possible (and not difficult) to switch it to DejaVu system-wide using a stylesheet and /etc/rst2pdf.conf

I can't ship that by default because not everyone has DejaVu, and also because I prefer that the default produces "standard" PDFs using the default fonts.