🚀 go-pugleaf

For BCB5 I posted QC 1351 which shows that the DFM file does not save a wide
string property properly when a property editor is created by the user to
accept and store in the property a wide character string. From the last
comment by Stefan Hoffmeister, the implication is that this is fixed in
BCB6. The workaround in BCB5 was to do one's own saving and restoring the
wide string property as a binary property rather than let the normal VCL
streaming system do it.

"Paul Kienitz" <pkienitz@prevaluate.com> wrote in message
news:3CFF1810.2F630FF5@prevaluate.com...
> Andrew Woodall wrote:
>
> > Does anybody know how to use Unicode characters in C++ Builder??
> >
> > I want to use the infinity symbol, but when inserting it either into
code in
> > the IDE or into the caption of a label or button, a question mark
appears
> > instead.
> >
> > Is there a way to do this???
>
> Borland fights back pretty hard against any attempt to make it do unicode.
For
> instance, I found that wide-character literal escapes in the form of
L"\x3333"
> do not work, either in the C++ compiler or in the resource compiler.  What
I
> ended up doing was using RCDATA blocks in the resource, which meant that
all my
> unicode strings had to be encoded like this:
>
> SOME_STRING RCDATA { 0x30D5, 0x30A1, 0x30A4, 0x30EB, 0x540D, 0x3A, 0 }
>
> Then I made a generalized string loader which, when given an ID, would
first
> look for an RCDATA and, if that was not found, look for a string resource.
All
> control captions had to be assigned at runtime.
>
> Oddly, though, the .DFM files are able to support unicode, I believe.
They use
> UTF-8 encoding.  The gotcha is that they will probably lose the
information if
> you ever use the design-time editors, which are limited to 8 bit
characters.
>