mingw-users Mailing List for MinGW - Minimalist GNU for Windows

Hello Georg,

Georg Fusz <fu...@is...> wrote:

GF> Does somebody know a better documentation of dllwrap as

GF> http://www.xraylith.wisc.edu/~khan/software/gnu-win32/README.dllhelpers-0.2.5.txt

GF> ?
    Dllwrap is deprecated - you should not use it for any new code.
Use ld --shared.


GF> --
GF> Georg Fusz
GF> Technische Universitaet Berlin, Germany


--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135

Thanks, that worked for me. I was missing a little in
the initialization.

Eric


--- Michael Haertl <mic...@gm...> wrote:
> Hello Eric,
> 
> I've included a small test program which run
> successfully 
> on my PC. Means I could see the outgoing data on
> another 
> PC's com port.
> 
> My programming skills are sureley not the best, but
> I 
> think you get the idea anyway. At least this worked
> for me:
> - CreateFile()
> - SetCommState()
> - WriteFile()
> 
> I've built it with Developer Studio 4, but I see no
> real 
> problems compiling it with another c-compiler. Tell
> me if 
> it works for you.
> 
> HTH
> Michael
> 
> 
> Lofstad Eric wrote:
> > Is there a good way to access the serial port
> using
> > mingw since is does not have the termios
> functions?
> > I've tried using some Windows code (CreateFile,
> > BuildCommDBC, and WriteFile).  That compiles and
> links
> > without any warnings but always crashes on
> execution
> > with a invalid page fault in Kernel32.dll.  Listed
> > below are the main lines for I used for opening
> the
> > port.
> > 
> > char *SerPort = "com1";
> > HANDLE Comm;
> > DCB dcb;
> > COMMTIMEOUTS cto = {0};
> > 
> > Comm = CreateFile(SerPort, GENERIC_READ |
> > GENERIC_WRITE, 0, 0, OPEN_EXISTING, 0, 0);
> > 
> > cto.ReadTotalTimeoutConstant = 500;
> > SetCommTimeouts(Comm, &cto);
> > FillMemory(&dcb, sizeof(dcb), 0);
> > dcb.DCBlength = sizeof(dcb);
> > 
> > BuildCommDCB("19200,n,8,1", &dcb);
> > 
> > Eric
> 
> -- 
> /* mic...@gm... */
> 
> 
> 
> 
>
/*****************************************************
>  *
>  * transmit simple commands via rs232 
>  *
> 
>
*****************************************************/
> 
> #include <stdio.h>
> #include <stdlib.h>
> #include <string.h>
> #include <conio.h>
> #include <wtypes.h>
> 
>
/*****************************************************/
> #define TESTSTRING "Test... "
> #define KEYB_Q      'q'
> #define KEYB_PLUS   '+'
> #define KEYB_MINUS  '-'
> #define KEYB_SPACE  ' '
> 
> const char* ComPortStr[] = {
>   "ERR",
>   "COM1",
>   "COM2"
> };
> 
>
/*****************************************************/
> 
> 
> 
> void usage(int argc, char* argv[])
> {
>  
>
printf("********************************************\n");
>   printf("wrong number of arguments. usage:\n\n");
>   printf("%s <port> <vca-nr>\n\n", "txcom");
>   printf("  where <port>    is the number of the
> used com port (1,2)\n");
>   printf("        <vca-nr>  is the number of the
> used VCA (1...24)\n");
> }
> 
> 
> 
> int main(int argc, char* argv[])
> {
>   // com port
>   int    com_no;
>   HANDLE hCom;
>   DCB    dcb;
>   ULONG  nBytesWritten;
> 
>   // vca parameters
>   int vca_no;
>   int vca_val       = 60;
>   int vca_delta     = 10;
>   int vca_mutestate = 0;
>   unsigned char cmd_word[3];
> 
>   // div
>   char c = 0;
> 
> 
>   printf("txcom - send commands to com port
> (test)\n\n");
>   
> 
>   // chk for no of arguments
>   if (argc != 3) {
>     usage(argc, argv);
>     // getchar();
>     exit(-1);
>   }
> 
> 
>   // chk for com port
>   com_no = atoi(argv[1]);
>   if ((com_no != 1) && (com_no !=2)) {
>     printf("cannot open com port %d\n", com_no);
>     exit(-1);
>   } else {
>     // try to open com port
>     hCom = CreateFile(
>       ComPortStr[com_no],
>       GENERIC_READ | GENERIC_WRITE,
>       0,                               /* com devs
> must be opeend w/ exclusive-access */
>       NULL,                            /* no
> security attribs */
>       OPEN_EXISTING,                   /* necessary
> for com ports */
>       0,                               /* no
> overlapped i/o */
>       NULL                             /* must be
> null for com */  
>     );
> 
>     if (!hCom) {
>       printf("could not open com port %d!\n",
> com_no);
>       exit(-1);
>     } else {
>       printf("com port %d opened.\n", com_no);
> 
>       // set dcb struct: 19200baud, 8bits no parity,
> 1 stop bit etc
>       dcb.DCBlength         = sizeof(dcb);          
> // sizeof(DCB)
>       dcb.BaudRate          = CBR_19200;            
> // current baud rate
>       dcb.fParity           = FALSE;                
> // disable parity checking
>       dcb.fOutxCtsFlow      = FALSE;                
> // no CTS output flow control
>       dcb.fOutxDsrFlow      = FALSE;                
> // no DSR output flow control
>       dcb.fDtrControl       = DTR_CONTROL_DISABLE;  
> // DTR flow control type
>       dcb.fDsrSensitivity   = FALSE;                
> // DSR sensitivity
>       dcb.fTXContinueOnXoff = TRUE;                 
> // XOFF continues Tx
>       dcb.fOutX             = FALSE;                
> // XON/XOFF out flow control
>       dcb.fInX              = FALSE;                
> // XON/XOFF in flow control
>       dcb.fNull             = FALSE;                
> // enable null stripping
>       dcb.fRtsControl       = RTS_CONTROL_DISABLE;  
> // RTS flow control
>       dcb.fAbortOnError     = FALSE;                
> // abort reads/writes on error
>       dcb.ByteSize          = 8;                    
> // number of bits/byte, 4-8
>       dcb.Parity            = NOPARITY;             
> // 0-4=no,odd,even,mark,space
>       dcb.StopBits          = ONESTOPBIT;           
> // 
=== message truncated ===


__________________________________________________
Do You Yahoo!?
Yahoo! Messenger - Talk while you surf!  It's FREE.
http://im.yahoo.com/

Is it possible to write the .def-Files by hand for C++-functions?

The names of the functions are expanded, how a look in a .def-file created by
dllwrap shows:

"
.....
dummyN__Fv @ 7 ; dummyN(void)
....
"

"; dummyN(void)" that's only a comment

"@ 7" is the number of the row in the file

"__Fv" is there a rule to construct this suffix?

-- 
Georg Fusz
Technische Universitaet Berlin, Germany


Fon:
Uni.: +49 30 314 26 884
privat: +49 30 815 30 32
Handy: +49 173 20 10 696

Homepage: http://www.cadlab.tu-berlin.de/~fusz/

--- Georg Fusz <fu...@is...> wrote:
> Does somebody know a better documentation of dllwrap as 
> 
>
http://www.xraylith.wisc.edu/~khan/software/gnu-win32/README.dllhelpers-0.2.5.txt
> 
> ?
> 

Uhm, the source code would be the definitive documentation.  What is the
problem?

Cheers,

=====
Earnie Boyd
mailto:ear...@ya...

---         <http://earniesystems.safeshopper.com>         ---
--- Cygwin: POSIX on Windows <http://gw32.freeyellow.com/> ---
---   Minimalist GNU for Windows <http://www.mingw.org/>   ---

__________________________________________________
Do You Yahoo!?
Yahoo! Messenger - Talk while you surf!  It's FREE.
http://im.yahoo.com/

Does somebody know a better documentation of dllwrap as 

http://www.xraylith.wisc.edu/~khan/software/gnu-win32/README.dllhelpers-0.2.5.txt

?


-- 
Georg Fusz
Technische Universitaet Berlin, Germany


Fon:
Uni.: +49 30 314 26 884
privat: +49 30 815 30 32
Handy: +49 173 20 10 696

Home page: http://www.cadlab.tu-berlin.de/~fusz/

We have DLLs built with Microsoft C, and we would like the exported
functions to be called by functions/routines compiled with MingW and
Dec/Compaq Fortran as well as others that we can support.  We have no
problem doing this for Dec/Compaq Fortran, and we found a solution for
MingW.

For MingW calling to Microsoft C, is there a better solution than what
we shown below?

Here is our solution for MingW and Dec/Compaq Fortran calling to
Microsoft C DLL.
__stdcall calling convention is used for 2 reasons:
1) More efficient than the C calling convention.
2) Better match between Microsoft C and Dec/Compaq Fortran: A symbol is
sandwiched between '_' and '@n'.

Is there a switch in gcc, dlltool or dllwrap that would mangle __stdcall
the same way?

Since the default MingW __stdcall does not pre-append '_' to the symbol,
we need to find a work-around.  And our work-around is to
declare/prototype functions with '_' pre-appending to the imported
functions, and our users must use them with '_' in front of all the
functions as shown in a simple example below.

/////////////////////////////

Build a DLL using Microsoft C compiler/linker version 6.x

D:\mixc>type msc.c
#include <stdio.h>
__declspec(dllexport) void __stdcall MSC(void)
{
        printf("MSC\n");
}

D:\mixc> cl msc.c /LD

Look at the symbols in the dll.

D:\mixc> dumpbin /exports msc.dll
...
    ordinal hint RVA      name
          1    0 00001000 _MSC@0
...
/////////////////////////////

Mixing Dec/Compaq Fortran with Microsoft C:

D:\mixc> type df77.f
PROGRAM DF77
CALL MSC()
STOP
END

Build an EXE using Dec/Compaq Fortran version 6.x

D:\mixc> df df77.f msc.lib
D:\mixc>df77
MSC
/////////////////////////////

Mixing MingW with Microsoft C:

D:\mixc>type mingw.c
__declspec(dllimport) void __stdcall _MSC(void);
int main(void)
{
        _MSC();
        return(0);
}

1. Create a module-definition file, for example

D:\mixc> echo EXPORTS > msc.def
D:\mixc> echo _MSC@0 >> msc.def

2. Create an import library

D:\mixc>dlltool --def msc.def --dllname msc.dll --output-lib msc.a

3. Build an EXE using MingW version 2.95.2 and run it.

D:\mixc> gcc mingw.c msc.a
D:\mixc> a.exe
MSC

It would be nice if there is a switch in gcc, dlltool or dllwrap that
would mangle __stdcall the same way as Microsoft C.  Does it exists?

--------------
Yeng Bun, Ph.D.
Senior Software Engineer
ye...@st...
Data Analysis Products Div. of MathSoft, Inc.

Thanks for the report.  I'll look into it.  I plan to release a new w32api
package soon.

Thanks,
Earnie.
--- Jean-Marc Gillet - Axa <jea...@ax...> wrote:
> Are you aware of it ? Here is an extract from
> i386-mingw32msvc\include\winsock2.h :
> 
> [...]
> int WINAPI WSAEnumNetworkEvents(SOCKET s, WSAEVENT hEventObject,
> LPWSANETWORKEVENTS lpNetworkEvents);
> int WINAPI WSAEventSelect(SOCKET s, WSAEVENT hEventObject, long
> lNetworkEvents);
> 
> ifdef __cplusplus   <--- there
> }
> #endif
> #endif
> [EOF]
> 
> Please also CC to mailto:jm...@bi... , as I didn't subscribe to the
> list.
> 
> Long live free software.
> 
> 	Jean-Marc
> _______________________________________________
> MinGW-users mailing list
> Min...@li...
> 
> You may change your MinGW Account Options at:
> http://lists.sourceforge.net/mailman/listinfo/mingw-users


__________________________________________________
Do You Yahoo!?
Yahoo! Messenger - Talk while you surf!  It's FREE.
http://im.yahoo.com/

Hello camilo,

camilo ramos <cr...@co...> wrote:

cr> Hello all,

cr> I was  wondering  about  the  following: Can  one  statically link in *.lib
cr> import library files produced by Visual C++? Note that the  object files
cr> that would be used to produce the *.lib are  not  part of a DLL, neither
cr> there would be a DLL  equivalent  to  such *.lib file. Just the archived
cr> objects the way one can do it in Unix with an *.a archive file for
cr> static linking.

cr>  Is this possible?

     That's exactly what recently released alapha binutils might be
good for. Danny, can you confirm this?

cr>  Any  suggestions  thanked.


cr> Best regards,
cr>  camilo                          mailto:cr...@co...



--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135

Are you aware of it ? Here is an extract from
i386-mingw32msvc\include\winsock2.h :

[...]
int WINAPI WSAEnumNetworkEvents(SOCKET s, WSAEVENT hEventObject,
LPWSANETWORKEVENTS lpNetworkEvents);
int WINAPI WSAEventSelect(SOCKET s, WSAEVENT hEventObject, long
lNetworkEvents);

ifdef __cplusplus   <--- there
}
#endif
#endif
[EOF]

Please also CC to mailto:jm...@bi... , as I didn't subscribe to the
list.

Long live free software.

	Jean-Marc

Is there a good way to access the serial port using
mingw since is does not have the termios functions?
I've tried using some Windows code (CreateFile,
BuildCommDBC, and WriteFile).  That compiles and links
without any warnings but always crashes on execution
with a invalid page fault in Kernel32.dll.  Listed
below are the main lines for I used for opening the
port.

char *SerPort = "com1";
HANDLE Comm;
DCB dcb;
COMMTIMEOUTS cto = {0};

Comm = CreateFile(SerPort, GENERIC_READ |
GENERIC_WRITE, 0, 0, OPEN_EXISTING, 0, 0);

cto.ReadTotalTimeoutConstant = 500;
SetCommTimeouts(Comm, &cto);
FillMemory(&dcb, sizeof(dcb), 0);
dcb.DCBlength = sizeof(dcb);

BuildCommDCB("19200,n,8,1", &dcb);


Eric

__________________________________________________
Do You Yahoo!?
Yahoo! Messenger - Talk while you surf!  It's FREE.
http://im.yahoo.com/

camilo ramos wrote:
> 
> I was  wondering  about  the  following: Can  one  statically link in *.lib
> import library files produced by Visual C++? Note that the  object files
> that would be used to produce the *.lib are  not  part of a DLL, neither
> there would be a DLL  equivalent  to  such *.lib file. Just the archived
> objects the way one can do it in Unix with an *.a archive file for
> static linking.

If the code were in a DLL, then you could use an 'import library'
to tell your program how to call the functions in the DLL. I think
you'd have to create the import library yourself using dlltool in
order to use it with mingw. You should be able to call extern "C"
functions in the DLL no matter what compiler it was built with.

If the code is in the library file itself, then that's a 'static
library'--just a collection of objects. The object file format
is incompatibly different for each compiler I know of, so you
cannot link an msvc static library to a mingw app.

Hello,
here's a sample for the CLASS "bug":

<start test.rc>
#define  STRICT
#define  WIN32_LEAN_AND_MEAN
#include <Windows.h>
#include <WinVer.h>

Test DIALOG 0, 0, 419, 150
STYLE DS_MODALFRAME | DS_3DLOOK | DS_CENTER | WS_POPUP | WS_VISIBLE | WS_CAPTION | WS_SYSMENU
CLASS "Test"
CAPTION "Test"
FONT 8, "MS Sans Serif"
{
 CONTROL "&OK", IDOK, "button", BS_PUSHBUTTON | BS_CENTER | WS_CHILD | WS_VISIBLE | WS_TABSTOP, 214, 116, 50, 14
}
<end test.rc>

windres.exe -i test.rc -o tst.o
=> "windres.exe: test.rc:8: parse error"
CLASS Test (without quotes) works, but rc.exe and Borland brc32.exe
report an error :
- rc.exe: "test.rc (8): error RC2104 : undefined keyword or key name: Test"
- brc32.exe : "Error test.rc 8 7: Expecting quoted string"

The MENUEX bug is more complex, it seems that MENUEX are badly
compiled so when trying to load them, an error is returned, and binary resource
editors can't load the executable.

Herve PARISSI
"You assume too much" (The Menace Phantom)

----- Original Message ----- 
From: "Paul Sokolovsky" <pa...@is...>
To: "Herv? PARISSI" <hpa...@ma...>
Cc: <min...@li...>
Sent: Monday, October 30, 2000 7:56 PM
Subject: Re[2]: [Mingw-users] SourceForge File Release Notice


Hello Hervé,

Hervé PARISSI <hpa...@ma...> wrote:

HP> Alas, windres is still bugged!
HP> - CLASS "class" doesn't work (CLASS class still work but it's not std!)
HP> - MENUEX doesn't work either (LoadMenu fails on it)

    I do not use windres myself, so... patches welcome ;-) You can
also describe precisely syntax and semantics of those construction is
bu report - who knows...

HP> Herve PARISSI
HP> "You assume too much" (The Menace Phantom)


--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135


_______________________________________________
MinGW-users mailing list
Min...@li...

You may change your MinGW Account Options at:
http://lists.sourceforge.net/mailman/listinfo/mingw-users

Hello all,

I was  wondering  about  the  following: Can  one  statically link in *.lib
import library files produced by Visual C++? Note that the  object files
that would be used to produce the *.lib are  not  part of a DLL, neither
there would be a DLL  equivalent  to  such *.lib file. Just the archived
objects the way one can do it in Unix with an *.a archive file for
static linking.

 Is this possible?

 Any  suggestions  thanked.


Best regards,
 camilo                          mailto:cr...@co...

Hello Franco,

Sunday, October 29, 2000, 1:26:51 AM, you wrote:

FB> Am Son, 29 Okt 2000 schrieb camilo ramos:
>>Hello all,
FB> <--snip-->
>>The troublesome part is
>>  const  enum GLenum gl_lights[] = {GL_LIGHT0,GL_LIGHT1,GL_LIGHT2,GL_LIGHT3};

FB> That looks strange to me.
FB> what should gl_lights be ?
FB> a constant array of GLenum values ?
 Yes. I guess so.
FB> then I'd try:
FB>   const  GLenum gl_lights[] = {GL_LIGHT0,GL_LIGHT1,GL_LIGHT2,GL_LIGHT3};
I did. That definition in particular stopped giving trouble, at least :->.
Thanks.


Best regards,
 camilo                            mailto:cr...@co...

Hello Lado,

Lado Brisar <lb...@ma...> wrote:

LB> On 30 Oct 2000, at 0:06, Paul Sokolovsky wrote:

>> 4. Enjoy new releases with cool new features brought to you by Mingw
>>    team.

LB> Great. ;) So, could you please tell us something more about that cool
LB> new features (improvements?, bugfixes?) in latest binutils release?
LB> What's new?

    Hm, it is build with libbfd/libopcodes as dll - it's significant
technological advance, imho. I've fixed all bugs I know in bfd/ld.
Also, there're quite a lot of new features since latest mingw32
release. I hope we'll put up a more specific list to the time of
release, but curious can look right now in ChangeLog, available from
CVSWeb from http://sources.redhat.com/binutils/ . And last, but most
cool feature, imho - ld (yet to be released) will have ability to link
against DLL's data exports without explicit __declspec(dllimport). But
for all this to come, it should be sufficiently tested. So, for
everyone who knows how to do that and have time to, I encourage to do
so. Otherwise, release may happen not so soon ;-)

LB> Regards,
LB> Lado Brisar.


--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135

Hello Hervé,

Hervé PARISSI <hpa...@ma...> wrote:

HP> Alas, windres is still bugged!
HP> - CLASS "class" doesn't work (CLASS class still work but it's not std!)
HP> - MENUEX doesn't work either (LoadMenu fails on it)

    I do not use windres myself, so... patches welcome ;-) You can
also describe precisely syntax and semantics of those construction is
bu report - who knows...

HP> Herve PARISSI
HP> "You assume too much" (The Menace Phantom)


--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135

Alas, windres is still bugged!
- CLASS "class" doesn't work (CLASS class still work but it's not std!)
- MENUEX doesn't work either (LoadMenu fails on it)

Herve PARISSI
"You assume too much" (The Menace Phantom)

----- Original Message ----- 
From: "Paul Sokolovsky" <pa...@is...>
To: <min...@li...>
Sent: Sunday, October 29, 2000 6:49 PM
Subject: Fwd: [Mingw-users] SourceForge File Release Notice


> Hello mingw-users,
> 
>       What you need to know about these releases is that they have NOT
> been tested thoroughly. They are intended as pre-alpha for courageuos
> testers. If in doubt, do not use them. Anyway, install in separate dir
> and prepend it to PATH. Recommended location for installation is
> /mingw (Earnie, have we decided on this?) - that means top-level
> directory "mingw" on your usual development disk.

Christian,

Thanks a lot. It is working right now.

Yoyong

On 30 Oct 2000, at 0:06, Paul Sokolovsky wrote:

> 4. Enjoy new releases with cool new features brought to you by Mingw
>    team.

Great. ;) So, could you please tell us something more about that cool 
new features (improvements?, bugfixes?) in latest binutils release? 
What's new?


Regards,
Lado Brisar.
http://lbrisar.htmlplanet.com

For everybody else wondering why on earth he gets crazy mails with
suggestion to visit some link here's definitive answer: IT IS OFFICIAL
MINGW PACKAGE RELEASE NOTICE. It is common practise to post
new version announcements to the mailing list, to not go too far,
Cygwin mailing list is example. And by the way, aren't for things like
this you did subscribe to mailing list? So:

1. Relax
2. Have a look at subject line
3. Yet better, do your mail client do it for you and sort mail into
   separate folders.
4. Enjoy new releases with cool new features brought to you by Mingw
   team.

This is a forwarded message
From: no...@so... <no...@so...>
To: md...@cy..., Pau...@te... <md...@cy..., Pau...@te...>
Subject: [Bug #119690] I am getting file release notificatios I never signed up for

===8<==============Original message text===============
Bug #119690, was updated on 2000-Oct-29 12:05
Here is a current snapshot of the bug.

Project: SourceForge
Category: (email) Notifications
Status: Closed
Resolution: Invalid
Bug Group: None
Priority: 5
Summary: I am getting file release notificatios I never signed up for

Details: Your system is not sending me file release notifications
that I never wanted or signed up for. Please fix this,
it started just a couple of days ago.

Here is an example of the message I am getting:

A new version of binutils has been released.
You can download it from SourceForge by following this link:

<http://sourceforge.net/project/showfiles.php?group_id=2435&release_id=14867>

You requested to be notified when new versions of this file
were released. If you don't wish to be notified in the
future, please login to SourceForge and click this link:
<http://sourceforge.net/project/filemodule_monitor.php?filemodule_id=11290>


Follow-Ups:

Date: 2000-Oct-29 12:44
By: pfalcon

Comment:
Watch out mailing lists you're subscribed to.

-------------------------------------------------------

Date: 2000-Oct-29 13:19
By: bigdisk

Comment:
Extremely unlikely.

Why not just unsubscribe as the message said to?
-------------------------------------------------------

Date: 2000-Oct-29 13:32
By: pfalcon

Comment:
I created "virtual" user to forward forum posts/release notices to mailing list I co-administer. Bug submitter is on that list. ;-)

Do you consider this dirty? ;-)

-------------------------------------------------------

Date: 2000-Oct-29 13:35
By: mdejong

Comment:
I went to the link the email mentioned to get off
the file notification list. It said I was take off.
That was yesterday, and today I got another notification.

I don't really need more spam. If your system is sending
out mail I never asked for, and continues to send it
even though I told it not too, I consider that spam.

-------------------------------------------------------

For detailed info, follow this link:
http://sourceforge.net/bugs/?func=detailbug&bug_id=119690&group_id=1


===8<===========End of original message text===========



--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135

Hello mingw-users,

      What you need to know about these releases is that they have NOT
been tested thoroughly. They are intended as pre-alpha for courageuos
testers. If in doubt, do not use them. Anyway, install in separate dir
and prepend it to PATH. Recommended location for installation is
/mingw (Earnie, have we decided on this?) - that means top-level
directory "mingw" on your usual development disk.

This is a forwarded message
From: no...@so... <no...@so...>
To: no...@so... <no...@so...>
Subject: [Mingw-users] SourceForge File Release Notice

===8<==============Original message text===============
A new version of binutils has been released.
You can download it from SourceForge by following this link:

<http://sourceforge.net/project/showfiles.php?group_id=2435&release_id=14867>

You requested to be notified when new versions of this file
were released. If you don't wish to be notified in the
future, please login to SourceForge and click this link:
<http://sourceforge.net/project/filemodule_monitor.php?filemodule_id=11290>
_______________________________________________
MinGW-users mailing list
Min...@li...

You may change your MinGW Account Options at:
http://lists.sourceforge.net/mailman/listinfo/mingw-users


===8<===========End of original message text===========



--
Paul Sokolovsky, IT Specialist
http://www.brainbench.com/transcript.jsp?pid=11135

A new version of binutils has been released. 
You can download it from SourceForge by following this link: 

<http://sourceforge.net/project/showfiles.php?group_id=2435&release_id=14867> 

You requested to be notified when new versions of this file 
were released. If you don't wish to be notified in the 
future, please login to SourceForge and click this link: 
<http://sourceforge.net/project/filemodule_monitor.php?filemodule_id=11290> 

> Does anyone knows how to use Playsound in Mingw. I tried Borland,it works
> Fine but with MingW it complains with the following message
>
> e:\files\hello\hellowin.o: In function `WndProc':
> file://e/files/hello/hellowin.c:61: undefined reference to
> `_imp__PlaySoundA@12'

gcc -o hellowin.exe hellowin.c -lwinmm

Christian.

Am Son, 29 Okt 2000 schrieb camilo ramos:
>Hello all,
<--snip-->
>The troublesome part is
>  const  enum GLenum gl_lights[] = {GL_LIGHT0,GL_LIGHT1,GL_LIGHT2,GL_LIGHT3};

That looks strange to me.
what should gl_lights be ? 
a constant array of GLenum values ?

then I'd try:
  const  GLenum gl_lights[] = {GL_LIGHT0,GL_LIGHT1,GL_LIGHT2,GL_LIGHT3};

>
>  while GLenum is defined as:
>
>typedef unsigned int GLenum;
>
>I compiled this on linux gcc-2.8.0.
>
>What am I missing?
>
>Any help thanked in advance.
>
>Best regards,
> camilo                          mailto:cr...@co...

Ciao, 
  Franco

Hello all,

  I got trouble trying to compile a GL app with mingw32 (gcc-2.95.2)
It requires g++ and the error is:

gnuplot_opengl.c: In function `void render()':
gnuplot_opengl.c:347: use of enum `GLenum' without previous declaration
gnuplot_opengl.c:347: elements of array `const enum GLenum gl_lights[]' have inc
omplete type
gnuplot_opengl.c:347: uninitialized const `gl_lights'
gnuplot_opengl.c:347: array size missing in `gl_lights'
gnuplot_opengl.c:347: storage size of `gl_lights' isn't known
gnuplot_opengl.c: In function `float fps()':
gnuplot_opengl.c:541: implicit declaration of function `int gettimeofday(...)'
gnuplot_opengl.c: At top level:
gnuplot_opengl.c:904: parse error before `float'
gnuplot_opengl.c:909: parse error before `float'
gnuplot_opengl.c: In function `void process_input()':
gnuplot_opengl.c:1057: implicit declaration of function `int isinf(...)'
gnuplot_opengl.c:1071: implicit declaration of function `int finite(...)'

The troublesome part is
  const  enum GLenum gl_lights[] = {GL_LIGHT0,GL_LIGHT1,GL_LIGHT2,GL_LIGHT3};

  while GLenum is defined as:

typedef unsigned int GLenum;

I compiled this on linux gcc-2.8.0.

What am I missing?

Any help thanked in advance.

Best regards,
 camilo                          mailto:cr...@co...