The changes are now in CVS. I have not changed the XMLCh
to the intrinsic wchar_t. I have kept it as unsigned int,
as the parser always assumes that it is UTF-16.

Regards,
Khaled Noaman

[EMAIL PROTECTED] wrote:

Tinny, Dean, etc.,

I didn't get any feedback to my questions below, so I will proceed with
proposing the following changes (I'm just pasting in the DIFF outputs)...
===================================================================
RCS file: /home/cvspublic/xml-xerces/c/src/util/Compilers/BorlandCDefs.hpp,v
retrieving revision 1.8
diff -r1.8 BorlandCDefs.hpp
88,89c88,91
< // typedef wchar_t  XMLCh;
< typedef unsigned short  XMLCh;
---
>  typedef wchar_t  XMLCh;
> // typedef unsigned short  XMLCh;
> // since C++Builder supports the wchar_t standard definition as unsigned short
> // but won't automatically convert between the two

===================================================================
RCS file:
/home/cvspublic/xml-xerces/c/src/util/Platforms/Win32/Win32PlatformUtils.cpp,v
retrieving revision 1.26
diff -r1.26 Win32PlatformUtils.cpp
74c74,78
< #include <crtdbg.h>
---
> #ifdef _MSC_VER
>    #include <crtdbg.h>
> #else
>    #include <assert.h>
> #endif
626,628c630,634
<     //  Windows supports InterlockedCompareExchange only on Windows 98,
<     //  NT 4.0, and newer. Not on Win 95. So we are back to using assembler.
<     //  (But only if building with MSVC.)
---
>     //  InterlockedCompareExchange is only supported on Windows 98,
>     //  Windows NT 4.0, and newer -- not on Windows 95...
>     //  If you are willing to give up Win95 support change this to #if 0
>     //  otherwise we are back to using assembler.
>     //  (But only if building with compilers that support inline assembler.)
630c636
<     #if defined(_MSC_VER)
---
>     #if defined(_MSC_VER) || defined(__BCPLUSPLUS__)
746c752
< #if 0 && _DEBUG
---
> #if 0 && defined(_DEBUG)

I have also re-created the BCB4 project files.  Because the changes were so
numerous -- mostly because all the paths were wrong -- I am simply attaching a
.ZIP with the new versions of all those files. (See attached file:
Projects-Win32-BCB4.zip)

P.S. I probably should add the disclaimer that these contributions are my own
and imply no responsibilities nor liabilities for my employer.
 

            [EMAIL PROTECTED]
            03/28/2001 02:21 AM
 

Please respond to [EMAIL PROTECTED]

To:   [EMAIL PROTECTED]
Subject:  Supporting C++Builder 4 and 5 issues

I am trying to pull together updates for the BCB4 files in CVS and create a
similar BCB5 tree for C++Builder 5 and I would like the community's input on a
few issues...

1.  Am I correct that Visual C++ 6 does not declare the C++ standard type
"wchar_t" as intrinsic like C++Builder does, which allows VC6 to think an
"XMLCh" is the same, while C++Builder thinks they are different?

2.  Given the typedef problem above, is anyone building Xerces under C++Builder
^without^ changing the "XMLCh" typedef in util\Compilers\BorlandCDefs.hpp back
to being "wchar_t"?

3.  MSDN, etc. indicates that the InterlockedCompareExchange API is not
supported on Windows 95, only on later versions.  Secondly, in the
"compareAndSwap" method in util\Platforms\Win32\Win32PlatformUtils.cpp, the
inline assembler section ifdef'ed for VC++ uses instructions that, I think, are
only valid for 80486 and later processors.  Is supporting Windows 95 the reason
we are not using the InterlockedCompareExchange API method?  Additionally, does
Xerces have a stated list of supported processors?

  ------------------------------------------------------------------------
                                 Name: Projects-Win32-BCB4.zip
   Projects-Win32-BCB4.zip       Type: Zip Compressed Data (application/x-zip-compressed)
                             Encoding: base64
                          Description: .ZIP File

  ------------------------------------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to