Re: Borland c++ compile problems...

Bruce Momjian wrote:

>OK, patch attached that fixes it.  Should that define just check for GCC
>instead of checking if the compiler isn't MS or Borland?
>
>---------------------------------------------------------------------------
>
>Steve Holdoway wrote:
>  
>
>>Hi folks,
>>
>>It seems that Borland C++ ( I'm using the command line version 5.5 ) 
>>takes exception to #defines like the following...
>>
>>#define open(a, b...) win32_open ( a, b, ##__VA_ARGS__ )
>>  [ to be found in port.h on the current development snapshot ]
>>
>>This causes compilation to fail. I'm after a version of blibpq.dll that 
>>supports ssl connections, and the one I compiled previously ( 7.4.2 I 
>>think ) doesn't seem to.
>>
>>Can anyone suggest a solution. I speak C fluently... it's this pesky 
>>windows stuff that causes me grief!
>>
>>Cheers,
>>
>>Steve
>>
>>---------------------------(end of broadcast)---------------------------
>>TIP 8: explain analyze is your friend
>>
>>    
>>
>
>  
>
>------------------------------------------------------------------------
>
>Index: src/include/port.h
>===================================================================
>RCS file: /cvsroot/pgsql-server/src/include/port.h,v
>retrieving revision 1.43
>diff -c -c -r1.43 port.h
>*** src/include/port.h	11 Jun 2004 16:10:09 -0000	1.43
>--- src/include/port.h	21 Jul 2004 03:01:33 -0000
>***************
>*** 151,157 ****
>  #ifdef WIN32
>  
>  /* open() replacement to allow delete of held files */
>! #ifndef _MSC_VER
>  extern int	win32_open(const char*,int,...);
>  #define 	open(a,b,...)	win32_open(a,b,##__VA_ARGS__)
>  #endif
>--- 151,157 ----
>  #ifdef WIN32
>  
>  /* open() replacement to allow delete of held files */
>! #if !defined(_MSC_VER) && !defined(__BORLANDC__)
>  extern int	win32_open(const char*,int,...);
>  #define 	open(a,b,...)	win32_open(a,b,##__VA_ARGS__)
>  #endif
>  
>
>------------------------------------------------------------------------
>
>
>---------------------------(end of broadcast)---------------------------
>TIP 4: Don't 'kill -9' the postmaster
>  
>
Sorry to be a pest, and all that. However, it looks like make needs to 
be run in src\port, to create pg_config_paths.h, and compile stuff to 
resolve externals _set_noblock, _pg_strcasecmp, _pg_tolower, and _popen 
(? bit worried about that one!)

I had a go at rewriting an existing bcc32.mak, but just seemed to 
generate more errors!

Cheers,

Steve

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
      joining column's datatypes do not match