Uhg, never thought I'd request this, but here goes.
Can you implement the standard c-lib string functions, for working on both ascii and unicode arrays? strcatA, strcatU, strlenA, strlenU, etc? Obviously, those libs are not complicated and can be implemented in PB, but they are likely to be noticeably faster if compiled in C, and a universal "native pb" version would be handy.
Why?
This will let us convert our unicode strings to the c-style for library calls, and we can still work on the results without extraneous conversions back and forth.
Thanks.