FWH 26.03 betatesters wanted

[alerchster]

Language german,
BCC77064 compile, message is ok
MSVC32 compile, message is ok
MSVC64 compile, message is not ok
fwstring() for MSVC64bit seems not good for öäü..

msg_fwstring.prg

Code: Select all | Expand

// Typical exit message, from Windows!

#include "FiveWin.ch"

//----------------------------------------------------------------------------//

function Main()
   FWSetLanguage( 5 )  // 1 english, 2 spanish, 3 french, 5 german, 6 italian
   MsgYesNo( fwstring("Want to End ?" ) )
   

return nil

//----------------------------------------------------------------------------//

[alerchster]

fivedbu.prg

sprintf should be changed to sprintf_s
change code after line 364 with claude's answer

Code: Select all | Expand

#pragma BEGINDUMP
#include <windows.h>
#include <hbapi.h>
HB_FUNC( FW_EXITPROCESS ) { ExitProcess( 0 ); }
HB_FUNC( CCOMPILER )
{
   #ifdef __clang__
      hb_retc( "LLVM/Clang " __clang_version__ );
   #elif defined( __GNUC__ )
      hb_retc( "GCC " __VERSION__ );
   #elif defined( _MSC_VER )
      char buf[ 32 ];
      sprintf_s( buf, sizeof( buf ), "MSVC %d", _MSC_VER );
      hb_retc( buf );
   #elif defined( __BORLANDC__ )
      char buf[ 32 ];
      sprintf_s( buf, sizeof( buf ), "BCC 0x%04X", __BORLANDC__ );
      hb_retc( buf );
   #else
      hb_retc( "Unknown" );
   #endif
}
#pragma ENDDUMP

[Antonio Linares]

Dear Anton,

Do you have Claude code ? If not I strongly advise you to use it

You won't believe how powerful it is until you use it. It will change your programming experience for ever

Using Claude code now every Harbour and FWH user becomes a truly engineer to enhance our platform!

[Antonio Linares]

fivedbu.prg

sprintf should be changed to sprintf_s
change code after line 364 with claude's answer

Code: Select all | Expand

#pragma BEGINDUMP
#include <windows.h>
#include <hbapi.h>
HB_FUNC( FW_EXITPROCESS ) { ExitProcess( 0 ); }
HB_FUNC( CCOMPILER )
{
   #ifdef __clang__
      hb_retc( "LLVM/Clang " __clang_version__ );
   #elif defined( __GNUC__ )
      hb_retc( "GCC " __VERSION__ );
   #elif defined( _MSC_VER )
      char buf[ 32 ];
      sprintf_s( buf, sizeof( buf ), "MSVC %d", _MSC_VER );
      hb_retc( buf );
   #elif defined( __BORLANDC__ )
      char buf[ 32 ];
      sprintf_s( buf, sizeof( buf ), "BCC 0x%04X", __BORLANDC__ );
      hb_retc( buf );
   #else
      hb_retc( "Unknown" );
   #endif
}
#pragma ENDDUMP

Dear Anton,

Claude code analysis:

We don't need sprintf_s here because:

The buffer is 32 bytes and the output is at most 15 characters (e.g. "MSVC 1943", "BCC 0x0770"). There is zero
overflow risk.

Technical reasons to avoid it:

1. Portability: sprintf_s is a Microsoft extension (Annex K of C11, which is optional). BCC32, GCC, and Clang on
   Windows don't implement it — the code won't compile on 4 of our 6 target compilers.

2. No safety gain: sprintf_s guards against buffer overruns by taking sizeof(buf) as a parameter. When the format
   string is a fixed literal with a single %d or %04X and the buffer is 2x larger than needed, the check adds nothing.

3. sprintf is standard C89: works on every compiler we support (BCC32, BCC32C/Clang, MSVC32, MSVC64, GCC32, GCC64).

   Bottom line: use sprintf for cross-compiler compatibility. Reserve sprintf_s for cases where the format string or data
   comes from user input and the buffer size is tight.

[alerchster]

Dear Antonio,

Yes, I've been using Claude since December, first with the Pro plan and now with the Max plan. It's a fantastic tool and, as you've already mentioned, it transforms development from a programmer to an architect. It brought my local Ubuntu server online via Cloudflare and set up several dynamic websites for me with 2FA via email and then with Microsoft Authenticator, etc., using PHP and a MySQL database. It did require some tweaks, but it led to a positive result very quickly.

When compiling FiveDBU, I got a warning with MSVC that sprintf was insecure and should be replaced with sprintf_s. I then had Claude rewrite that part of the code to sprintf_s, and that was the result. I agree with you, it was just a warning and didn't affect the functionality in this case.

[Antonio Linares]

There is a new FWH 26.03 beta ready

Please ask for it on whatsapp, thank you

[Enrico Maria Giordano]

You can use

Code: Select all | Expand

-D_CRT_SECURE_NO_WARNINGS

[toninhofwi]

Hi Mr Antonio,

Everything seems to be fine.
Some of the classes I customized had different function names—for example:
xChangeProc() is now xxChangeProc() — but I’ve already fixed that.

Is it just me, or does everything seem faster and smoother?

Anyway, congratulations on the great work.

I’ll get organized and send you the 170 euros for the product — it’s worth every penny.

[NWKL]

HI, i want , thank you.

[Antonio Linares]

Please send a whatsapp to +34 722461100

[richard-service]

Dear Antonio,

Please Look it.
I use New version Chinese input way, but not work
But I use Old version Chinese input way. It's working.

[alerchster]

fivedbu.prg

compiled with BCC770 32 Bit runs OK
compiled with MSVC 32 and 64 Bit some Dialogs will not be shown
example: "Setup" "Preferences", Buttonbar "Search" and many more

fwstring with german "öäü" is now ok

[Antonio Linares]

Dear Antonio,

Please Look it.
I use New version Chinese input way, but not work
But I use Old version Chinese input way. It's working.

Please check your email, thanks

[richard-service]

Dear Antonio,

Please Look it.
I use New version Chinese input way, but not work
But I use Old version Chinese input way. It's working.

Please check your email, thanks

Ok. I got it. and Check solution it.

[carlos vargas]

la compilación codicional para bcc32c de Borland es:

#if defined(BORLANDC) && defined(clang) && defined(WIN32) && !defined(WIN64)
// Your code.
#endif

y esta es para bcc64:

#if defined(BORLANDC) && defined(clang) && defined(_WIN64)
// Your code.
#endif