output of a page of Persian wikibook
The PDF file that atteched is output of this page:
AS you see the PDF file is ltr(Persian is rtl) and all of character is separated(In Persian and Arabic languages separated and connected character have different form)
I saw this bug on Arabic and Hebrew languages
I have a propose:you can use wkhtmltopdf instead http://www.reportlab.com ,wkhtmltopdf is a open source and linux-based program that Google use it.