Re: [tz] Non-ASCII chars in southamerica file

Thanks for catching these problems. Two proposed patches are attached. The first fixes the problems by going back to ASCII. The second puts in a check for this problem, so that non-ASCII bytes like that don't slip into future releases. At some point we may well want to add non-ASCII characters, but they should be UTF-8 I expect. I've pushed these proposed patches into the unofficial experimental repository at github.