I'm using both Latin1 and UTF8 encoded files in my projects. The browser showing diffs seems always to assume the UTF8 encoding, resulting in incorrectly showing non-ascii characters. Is there a way to tell SourceTree (or Git) about the encoding of a file?
Addendum: picking an encoding on a per-project basis would be sufficient rather than on a per-file basis. The reason I'm using Latin1 in some projects is that these are legal documents that I share with non-developers on Windows machines that seem to prefer it.