this was copied from Arch, but the parsing code doesn't understand empty lines and comment lines, so I'm not sure how that is supposed to work. This led to the default papersize being the comment line which makes the lookup fail, and in combination with xpdf not checking any return values results in xpdf crashing.