you have to basically bypass or modify mfc to prevent it from making calls
to unicode api's on win9x.
microsoft uses (e.g. in office) shim dll's that have "...U"-style named
entry points that always take WCHAR/WSTR strings. for nt, they pass through
to the native "...W" api calls, and for 9x they convert to and from the
current system codepage and call the "...A" api calls.
have a look at their web site.
a few api's actually exist in the "...W" form on 9x (TextOutW() and such).
you would compile your application with the UNICODE and _UNICODE #define's.
it is a whole lot easier to produce 2 binaries (one with _MBCS, the other
one with both UNICODE's) from a single source tree.
worse news: it sounds in rumours like there will be a consumer version of
nt only around 2003, and we have to live with 9x until then :-( maybe you
can convince your customers to only run nt workstation or a different
beware: should you ever port to win ce - there are only the "...W"
functions, not the "...A" ones.
Markus Scherer IBM RTP +1 919 486 1135 Dept. Fax +1 919 254 6430
Unicode is here! --> http://www.unicode.org/
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:45 EDT