Pandoc 1.5.x uses the locale information to parse its input files (only
1.5, pre and post version use always UTF-8). Hence we need to enforce a
UTF-8 locale for proper parsing of input files.