I tried converting http://book.kanunu.org/book4/10406/231373.html to PDF.
I used the following command:
wkhtmltopdf --encoding gbk http://book.kanunu.org/book4/10406/231373.html test.pdf
And I got the following output: https://cloud.githubusercontent.com/assets/1232162/3973254/d67340e6-27e5-11e4-9b30-a7df42e0bc41.png
As you can see the characters slant. I tested out more websites and realized that this happens only for Chinese and Japanese characters. Does anyone know why that is the case?