See also bug 68724.
After listening to Fae's comments about encoding troubles, I think it makes sense to tell GWToolset to automatically change C0 and C1 characters (as encoded in utf-8) into what their code point would be if we were using windows-1252.
The C0 and C1 characters are never what's wanted, invisible PITA. Even if windows-1252 is the wrong interpretation, its a better failure mode than adding invisible characters in the file name.
Version: unspecified
Severity: normal