I'm impressed how fast things are developing right now - WOW!
Martin Evans wrote:
On the point of the Alexander's unicode patch I seem to remember
applying it over a year a go to my copy of DBD::ODBC but it broke
building of DBD::ODBC on UNIX - perhaps my recollection is wrong.
I would expect my patch to break compiling on every platform except
Win2K/WinXP, because I only tested it on those platforms.
I think my changes are quite Window specific, especially the calls to
the Unicode converter routines WideCharToMultiByte() and
MultiByteToWideChar(). wchar.h, wcslen() and wcscpy() are in the Single
UNIX Specification, Version 2, dated 1997. WCHAR should be equivalent to
wchar_t of the Single UNIX Specification, a #define or typedef should be
sufficient, like Microsoft does in wchar.h. See also
<http://www.alexander-foken.de/README.unicode-patch.html#known_problems>.
With a litte bit more knowlegde of the inner workings of Perl's Unicode
support, it should be possible to replace the two converter routines
called by my patch with Perl build-in routines. When that has happened,
the code should work on platforms implementing the Single UNIX
Specification Version 2, assuming the ODBC system wants Unicode in
UTF-16 like Windows does. If the ODBC system expects UTF-8, the calls to
the two converter functions should be called only on Win32.
Hope that helps,
Alexander
--
Alexander Foken
mailto:[EMAIL PROTECTED] http://www.foken.de/alexander/
