[Date Prev][Date Next][Thread Prev][Thread Next][Thread Index]

Re: [XaraXtreme-dev] UNICODE problems in cxfrec [Fwd: [XaraXtreme-commits] Commit Complete]




These really do all need fixing.  There is also this bit at the top of
cxfrec.cpp:

// This define has different meaning under linux
#if !defined(__WXMSW__)
#undef UNICODE
#endif

...which looks somewhat dodgy...

It's #undef'ing something that won't be defined anyway, isn't
it? (i.e. the author meant '_UNICODE' I think). So whilst
it may be dodgy, it sounds like it's harmlessly dodgy.

Note if you are investigate Unicode wierdness, in some of the
ccfile & filters stuff I converted over people were using "char" as
a holder of 8 bits of arbitrary information, which had since
(sometimes for good reason) been converted to "TCHAR".
These you really do want to do a 1:1 conversion on because the
underlying source was never multibyte UTF-8 in the first place,
but pure ASCII (or should I say, pure 8 bit) and it needs
conversion /back/ to pure 8-bit. I hope I flagged all this
stuff with "do 1:1 BYTE<->TCHAR conversion" or similar.
(i.e. what are were doing is using TCHARs to hold 8 bit
quantities. Wow).

Alex