Thanks Gabriele - it's good to be precise.
The Red documentation on the differences between READ in Rebol and Red does not highlight the limitations in reading Latin1 - so the issue may be a bug.
Though it does note that the current "simple" functionality may change - so this may be addressed in later iterations.
/binary(so, it could mess up text encoded in something not ASCII-compatible). Red and R3 will try to decode the text, so if it's not UTF-8 you get problems. It might make sense to relax the decoding and just replace anything invalid with a ? or something like that, instead of throwing an error... perhaps with a
/strictrefinement to enforce the current behavior... but in general if you don't know the encoding of some text then it would be best to treat it as binary instead.