Our development people are using latin1 charsets in Windows, which mostly works when the correct charset it set
however, calculating the changelog doesn't work since jenkins tries to process the strings with utf-8
it would be better if the charset were latin1.