Il 15/08/2023 16:50, Enrico Maria Giordano ha scritto:

> Please revert make_vc.bat to the SVN version, it is working fine! We
> must be free to change the compiler directory!

I just tried using make_vc.bat from the SVN with xHarbour from GitHub
and all went fine. So there is obviously an error to be fixed in
make_vc.bat from GitHub.

Removing System and Object Files ...
Removing DLL and Object Files ...
Removing Contrib Libraries and Object Files ...
Done ...

****** End of Job *****
Creating System Files (ST) ...
Creating System Files (MT) ...
System Files Succesfully Built ...
Copying System Files to BIN and LIB Folders ...
Done ...

****** End of Job *****
Creating Contrib Libraries ...
Contrib Libraries Succesfully Built ...
Copying Contrib Libraries to LIB Folder ...
Done ...

****** End of Job *****

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

The BCC32 build is failing even with make_vc.bat from SVN:

source\compiler\hbmain.c:
Borland Resource Compiler Version 5.40
Copyright (c) 1990, 1999 Inprise Corporation. All rights reserved.

Error harbour.exe.rc 3 0: Too many digits in a number

** error 20018 ** deleting obj\b32\harbour.exe.res

So there is a bug to fix somewhere else.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Il 15/08/2023 17:01, Ron Pinkas ha scritto:

>> Please revert make_vc.bat to the SVN version, it is working fine! We
>> must be free to change the compiler directory!
>
> make_vc has been cleaned, standardized to provide more consistent support to the many complexities and different tools installed by Microsoft. Support was also added for newer versions of all tools and the O/S itself. So your request to revert seems the wrong approach, instead we should find the exact cause of your issue,

This is not "my issue", this is a bug YOU introduced. Please fix it. If
you need my help then ask for my help, clearly and gently.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Enrico,

> This is not "my issue", this is a bug YOU introduced. Please fix it. If
> you need my help then ask for my help, clearly and gently.

Please let's avoid hostility and getting personal. I apologize, did not mean to say the issue is "you" - I said "your issue" as in the issue you encountered. The correct solution is to find the cause, and only you can trace it and provide the details needed to fix it since I do not have such setup available to me. If you do not care for the the new fixes (and possible "bugs") you can of course keep using the SVN.

Ron

Enrico,

> The BCC32 build is failing even with make_vc.bat from SVN:

I believe you mean make_bc.bat right?
> source\compiler\hbmain.c:
> Borland Resource Compiler Version 5.40
> Copyright (c) 1990, 1999 Inprise Corporation. All rights reserved.
>
> Error harbour.exe.rc 3 0: Too many digits in a number
> ** error 20018 ** deleting obj\b32\harbour.exe.res
>
> So there is a bug to fix somewhere else.

The bug seems to be with the Borland Resource Compiler not able to compile 'hbver.h' (C99 standard long long number). We were forced to change old version numbers relying on SVN version ID because it is not supported by GIT. This is not something I have any control over. We are forced to switch to someway of relying on what is available with GIT if we want to use GIT.

Ron

Il 15/08/2023 17:19, Ron Pinkas ha scritto:

>> This is not "my issue", this is a bug YOU introduced. Please fix it. If
>> you need my help then ask for my help, clearly and gently.
>
> Please let's avoid hostility and getting personal. I apologize, did not mean to say the issue is "you" - I said "your issue" as in the issue you encountered. The correct solution is to find the cause, and only you can trace it and provide the details needed to fix it since I do not have such setup available to me. If you do not care for the the new fixes (and possible "bugs") you can of course keep using the SVN.

Ok, I'll try to find the problem in make_vc.bat and report back here.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Il 15/08/2023 17:33, Ron Pinkas ha scritto:

> Enrico,

>
>> The BCC32 build is failing even with make_vc.bat from SVN:
>
> I believe you mean make_bc.bat right?

Yes, make_bc.bat.

>> source\compiler\hbmain.c:
>> Borland Resource Compiler Version 5.40
>> Copyright (c) 1990, 1999 Inprise Corporation. All rights reserved.
>>
>> Error harbour.exe.rc 3 0: Too many digits in a number
>> ** error 20018 ** deleting obj\b32\harbour.exe.res
>>
>> So there is a bug to fix somewhere else.
>
> The bug seems to be with the Borland Resource Compiler not able to compile 'hbver.h' (C99 standard long long number). We were forced to change old version numbers relying on SVN version ID because it is not supported by GIT. This is not something I have any control over. We are forced to switch to someway of relying on what is available with GIT if we want to use GIT.

hbver.h seems identical (one comment apart) to the old version. What ID
are you referring to?

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

> > The bug seems to be with the Borland Resource Compiler not able to compile 'hbver.h' (C99 standard long long number). We were forced to change old version numbers relying on SVN version ID because it is not supported by GIT. This is not something I have any control over. We are forced to switch to someway of relying on what is available with GIT if we want to use GIT.

> hbver.h seems identical (one comment apart) to the old version. What ID
> are you referring to?

Sorry, meant the AUTO GENERATED hbverbld.h - specifically this:

#define HB_VER_GITFAKEID 202308141804LL
#define HB_VER_CVSID 202308141804LL

Ron

Il 15/08/2023 17:41, Enrico Maria Giordano ha scritto:

> Ok, I'll try to find the problem in make_vc.bat and report back here.

The problem is here:

OLD version

IF EXIST "%CC_DIR%"\vcvarsall.bat CALL "%CC_DIR%"\vcvarsall.bat
set
PATH="%CC_DIR%\bin";%VSCOMMONTOOLS%;"%RC_DIR%";"%BISON_DIR%";%~dp0bin;%PATH%

NEW version

if "%VSINSTALLDIR%"=="" (
IF EXIST "%CC_DIR%"\vcvarsall.bat CALL "%CC_DIR%"\vcvarsall.bat
IF "%VCINSTALLDIR%"=="" (
set
PATH="%CC_DIR%\bin;%VSCOMMONTOOLS%;%RC_DIR%;%BISON_DIR%;%~dp0bin;%PATH%"
)
)

I'm still trying to find the bug...

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Il 15/08/2023 18:07, Enrico Maria Giordano ha scritto:

> I'm still trying to find the bug...

This is working:

IF "%VSINSTALLDIR%"=="" IF EXIST "%CC_DIR%"\vcvarsall.bat CALL
"%CC_DIR%"\vcvarsall.bat
IF "%VSINSTALLDIR%"=="" IF "%VCINSTALLDIR%"=="" set
PATH="%CC_DIR%\bin";%VSCOMMONTOOLS%;"%RC_DIR%";"%BISON_DIR%";%~dp0bin;%PATH%

Can you commit this or change it as you like (and commit)?

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Ron, thank you for your great work in the xHarbour community.

Em segunda-feira, 14 de agosto de 2023 às 15:26:47 UTC-3, Ron Pinkas escreveu:
> As promised the complete xHarbour.com sources of ALL of our products, scripts, documentation, etc. have been uploaded to the xHarbour.org GitHub repository (including all commit history) into the xHarbourBuilder sub folder.
>
> Enjoy!
>
> Ron

> This is working:
>
> IF "%VSINSTALLDIR%"=="" IF EXIST "%CC_DIR%"\vcvarsall.bat CALL
> "%CC_DIR%"\vcvarsall.bat
> IF "%VSINSTALLDIR%"=="" IF "%VCINSTALLDIR%"=="" set
> PATH="%CC_DIR%\bin";%VSCOMMONTOOLS%;"%RC_DIR%";"%BISON_DIR%";%~dp0bin;%PATH%
> Can you commit this or change it as you like (and commit)?

Cool. Maybe () grouping not supported by the Windows version you have. Please test after:

--------
2023-08-15 12:16 UTC-0500 Ron Pinkas <ronpinkas/AT/gmail/com>
* .gitignore
+ *.map

* make_vc.bat
* Removed usage of () to group IF logic
------

Ron

Il 15/08/2023 19:21, Ron Pinkas ha scritto:

>> This is working:
>>
>> IF "%VSINSTALLDIR%"=="" IF EXIST "%CC_DIR%"\vcvarsall.bat CALL
>> "%CC_DIR%"\vcvarsall.bat
>> IF "%VSINSTALLDIR%"=="" IF "%VCINSTALLDIR%"=="" set
>> PATH="%CC_DIR%\bin";%VSCOMMONTOOLS%;"%RC_DIR%";"%BISON_DIR%";%~dp0bin;%PATH%
>> Can you commit this or change it as you like (and commit)?
>
> Cool. Maybe () grouping not supported by the Windows version you have.

I'm using Windows 11.

> Please test after:
>
> --------
> 2023-08-15 12:16 UTC-0500 Ron Pinkas <ronpinkas/AT/gmail/com>
> * .gitignore
> + *.map
>
> * make_vc.bat
> * Removed usage of () to group IF logic
> ------

Now it is working fine, thank you. MSC32 and MSC 64 builds are OK!

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Il 15/08/2023 17:57, Ron Pinkas ha scritto:

>> hbver.h seems identical (one comment apart) to the old version. What ID
>> are you referring to?
>
> Sorry, meant the AUTO GENERATED hbverbld.h - specifically this:
>
> #define HB_VER_GITFAKEID 202308141804LL
> #define HB_VER_CVSID 202308141804LL

Can I see a sample of that .RC file that I can test with BCC resource
compiler?

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

> > Sorry, meant the AUTO GENERATED hbverbld.h - specifically this:
> >
> > #define HB_VER_GITFAKEID 202308141804LL
> > #define HB_VER_CVSID 202308141804LL

> Can I see a sample of that .RC file that I can test with BCC resource
> compiler?

Sure, IIRC just look for harbour.rc

Ron

Il 15/08/2023 19:21, Ron Pinkas ha scritto:

> Cool. Maybe () grouping not supported by the Windows version you have.

Some double-quotes were misplaced too.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

Il 15/08/2023 19:55, Ron Pinkas ha scritto:

>> Can I see a sample of that .RC file that I can test with BCC resource
>> compiler?
>
> Sure, IIRC just look for harbour.rc

There is no harbour.rc in the repository.

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

> >> Can I see a sample of that .RC file that I can test with BCC resource
> >> compiler?
> >
> > Sure, IIRC just look for harbour.rc
>
> There is no harbour.rc in the repository.

Sorry I noticed my mistake and deleted the message. The file is actually:

harbour.exe.rc

It is an automatically generated file by a utility called hbrc (xharbour\utils\misc\hbrc.c)

I suspect this issue can be fixed in that source.

Ron

Il 15/08/2023 20:15, Ron Pinkas ha scritto:

>> There is no harbour.rc in the repository.
>
> Sorry I noticed my mistake and deleted the message. The file is actually:
>
> harbour.exe.rc
>
> It is an automatically generated file by a utility called hbrc (xharbour\utils\misc\hbrc.c)

Ok. Inside it there are:

FILEVERSION 1,2,3,202308151216
PRODUCTVERSION 1,2,3,202308151216

They should be (according to MS docs) four 16bit numbers. So this big
number is wrong.

https://learn.microsoft.com/en-us/windows/win32/menurc/versioninfo-resource

> I suspect this issue can be fixed in that source.

Can you guide me to some tests that I can do?

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

> > harbour.exe.rc
> >
> > It is an automatically generated file by a utility called hbrc (xharbour\utils\misc\hbrc.c)

> Ok. Inside it there are:
>
> FILEVERSION 1,2,3,202308151216
> PRODUCTVERSION 1,2,3,202308151216
>
> They should be (according to MS docs) four 16bit numbers. So this big
> number is wrong.
>
> https://learn.microsoft.com/en-us/windows/win32/menurc/versioninfo-resource
> > I suspect this issue can be fixed in that source.

> Can you guide me to some tests that I can do?

I believe the code can be modified to possibly allocate the FILEVERSION 4 pieces (4x16=64bit) to encode the LL number, and modify the PRODUCTVERSION to omit the LL number (0).

This means the resource will have both the exact Build's Date+Time Stamp as well as the explicit version number specified in the sources.

But truly I never looked into this RC file and have no idea what is even needed for.

Ron

On Tuesday, August 15, 2023 at 3:06:09 PM UTC-5, Ron Pinkas wrote:
> > > harbour.exe.rc
> > >
> > > It is an automatically generated file by a utility called hbrc (xharbour\utils\misc\hbrc.c)
>
> > Ok. Inside it there are:
> >
> > FILEVERSION 1,2,3,202308151216
> > PRODUCTVERSION 1,2,3,202308151216
> >
> > They should be (according to MS docs) four 16bit numbers. So this big
> > number is wrong.
> >
> > https://learn.microsoft.com/en-us/windows/win32/menurc/versioninfo-resource
> > > I suspect this issue can be fixed in that source.
>
> > Can you guide me to some tests that I can do?
> I believe the code can be modified to possibly allocate the FILEVERSION 4 pieces (4x16=64bit) to encode the LL number, and modify the PRODUCTVERSION to omit the LL number (0).
>
> This means the resource will have both the exact Build's Date+Time Stamp as well as the explicit version number specified in the sources.
>
> But truly I never looked into this RC file and have no idea what is even needed for.
>
> Ron

Maybe something like this:

unsigned long long llvalue = HB_VER_CVSID;
unsigned long dw1 = llvalue >> 32; // Get the first DWORD (the higher 32 bits)
unsigned long dw2 = llvalue & 0xFFFFFFFF; // Get the second DWORD (the lower 32 bits)

unsigned short hiword1 = dw1 >> 16; // Get the high word of dw1
unsigned short loword1 = dw1 & 0xFFFF; // Get the low word of dw1
unsigned short hiword2 = dw2 >> 16; // Get the high word of dw2
unsigned short loword2 = dw2 & 0xFFFF; // Get the low word of dw2

fprintf(h, "FILEVERSION %u,%u,%u,%u\n", hiword1, loword1, hiword2, loword2);

rebuild sqlrdd for bcc 7.60 ok :-)
using hbmk2.exe from harbour with -xhb parameters, using bison tools.
sqlrdd parser sql. :-)

On Tuesday, August 15, 2023 at 9:19:08 PM UTC-5, Carlos Vargas wrote:
> rebuild sqlrdd for bcc 7.60 ok :-)
> using hbmk2.exe from harbour with -xhb parameters, using bison tools.
> sqlrdd parser sql. :-)

I hope to add xBuild to the standard build so anyone can then use the included xbp project files and be able to build all of the tools and librarries.

Ron

Il 15/08/2023 22:54, Ron Pinkas ha scritto:

> Maybe something like this:
>
> unsigned long long llvalue = HB_VER_CVSID;
> unsigned long dw1 = llvalue >> 32; // Get the first DWORD (the higher 32 bits)
> unsigned long dw2 = llvalue & 0xFFFFFFFF; // Get the second DWORD (the lower 32 bits)
>
> unsigned short hiword1 = dw1 >> 16; // Get the high word of dw1
> unsigned short loword1 = dw1 & 0xFFFF; // Get the low word of dw1
> unsigned short hiword2 = dw2 >> 16; // Get the high word of dw2
> unsigned short loword2 = dw2 & 0xFFFF; // Get the low word of dw2
>
> fprintf(h, "FILEVERSION %u,%u,%u,%u\n", hiword1, loword1, hiword2, loword2);

Perfect! This is working fine. Please commit. BCC32 and BCC54 builds are OK!

int main( int argc, char * argv[] )
{ if( argc >= 7 )
{
FILE * h = fopen( argv[ 1 ], "wb" );

if( h )
{
SYSTEMTIME t;
char * cCompiler = _hb_verCompiler();
char * aMo[] = { "January", "February", "March",
"April", "May", "June", "July", "August", "September", "October",
"November", "December" };
char szParam[ 265 ];

unsigned long long llvalue = HB_VER_CVSID;
unsigned long dw1 = llvalue >> 32; // Get the first DWORD (the
higher 32 bits)
unsigned long dw2 = llvalue & 0xFFFFFFFF; // Get the second DWORD
(the lower 32 bits)

unsigned short hiword1 = dw1 >> 16; // Get the high word of dw1
unsigned short loword1 = dw1 & 0xFFFF; // Get the low word of dw1
unsigned short hiword2 = dw2 >> 16; // Get the high word of dw2
unsigned short loword2 = dw2 & 0xFFFF; // Get the low word of dw2

GetLocalTime( &t );

fprintf( h, "1 VERSIONINFO\n" );

fprintf(h, "FILEVERSION %u,%u,%u,%u\n", hiword1, loword1, hiword2,
loword2);
fprintf(h, "PRODUCTVERSION %u,%u,%u,%u\n", hiword1, loword1,
hiword2, loword2);

// fprintf( h, "FILEVERSION %d,%d,%d,%lld\n",
HB_VER_MAJOR,HB_VER_MINOR,HB_VER_REVISION,HB_VER_CVSID );
// fprintf( h, "PRODUCTVERSION %d,%d,%d,%lld\n",
HB_VER_MAJOR,HB_VER_MINOR,HB_VER_REVISION,HB_VER_CVSID );

--
Enrico Maria Giordano

http://www.emagsoftware.it
http://www.emagsoftware.it/emgmusic
http://www.emagsoftware.it/spectrum
http://www.emagsoftware.it/tbosg

> Perfect! This is working fine. Please commit. BCC32 and BCC54 builds are OK!
>
>
> int main( int argc, char * argv[] )
> {
> if( argc >= 7 )
> {
> FILE * h = fopen( argv[ 1 ], "wb" );
>
> if( h )
> {
> SYSTEMTIME t;
> char * cCompiler = _hb_verCompiler();
> char * aMo[] = { "January", "February", "March",
> "April", "May", "June", "July", "August", "September", "October",
> "November", "December" };
> char szParam[ 265 ];
> unsigned long long llvalue = HB_VER_CVSID;
> unsigned long dw1 = llvalue >> 32; // Get the first DWORD (the
> higher 32 bits)
> unsigned long dw2 = llvalue & 0xFFFFFFFF; // Get the second DWORD
> (the lower 32 bits)
>
> unsigned short hiword1 = dw1 >> 16; // Get the high word of dw1
> unsigned short loword1 = dw1 & 0xFFFF; // Get the low word of dw1
> unsigned short hiword2 = dw2 >> 16; // Get the high word of dw2
> unsigned short loword2 = dw2 & 0xFFFF; // Get the low word of dw2
> GetLocalTime( &t );
>
> fprintf( h, "1 VERSIONINFO\n" );
> fprintf(h, "FILEVERSION %u,%u,%u,%u\n", hiword1, loword1, hiword2,
> loword2);
> fprintf(h, "PRODUCTVERSION %u,%u,%u,%u\n", hiword1, loword1,
> hiword2, loword2);
>
> // fprintf( h, "FILEVERSION %d,%d,%d,%lld\n",
> HB_VER_MAJOR,HB_VER_MINOR,HB_VER_REVISION,HB_VER_CVSID );
> // fprintf( h, "PRODUCTVERSION %d,%d,%d,%lld\n",
> HB_VER_MAJOR,HB_VER_MINOR,HB_VER_REVISION,HB_VER_CVSID );
> --
Enrico,
I just fixed my code here, and compiled fine!

C:\Users\rhfsy\xharbour.git\trunk>make_bc all
Creating System Files (ST) ...
Creating System Files (MT) ...
System Files Succesfully Built ...
Copying System Files to BIN and LIB Folders ...
Done ...

Creating DLL Files ...
DLL Files Succesfully Built ...
Copying DLL Files to BIN Folder ...
Done ...

Creating Contrib Libraries ...
Contrib Libraries Succesfully Built ...
Copying Contrib Libraries to LIB Folder ...
Done ...

Thank you, and thanks Ron!

Reynaldo Henrique.