Re: mingw-hosted xgcc

["Kai Ruottu" <karuottu@freenet.hut.fi>]

I wrote:

> So the simplest way would to remove these from specs and move them
> to some 'always included' header, or to a header which is included
> in this... Could <sys/types.h> be the proper one?  I moved them
> somewhere, but cannot find where, perhaps I have written them over...

 This kind of 'always included' header is the '_ansi.h', and the
'sys/config.h' is #include'd from it, and already contains all kind
of machine-dependent things...

 So, here is my temporary patch for the 'sys/config.h' for mingw32 host,
where the SIZE_TYPE and PTRDIFF_TYPE definitions must be removed from the 
CPP_SPEC in specs (until they will pass ok through pexecute and the MS
spawn()) :

*** config.h.orig	Wed Oct 07 14:11:52 1998
--- config.h	Sat Feb 20 13:53:09 1999
***************
*** 3,19 ****
  
  /* exceptions first */
  /* ??? Why is much of this stuff duplicated with machine/ieeefp.h?  */
! #if defined(__H8300__) || defined(__H8500__) || defined (__H8300H__) ||  
defined(__W65__) || defined (__H8300S__)
  #define _FLOAT_ARG float
- #define __SMALL_BITFIELDS
  #define _DOUBLE_IS_32BITS
  #define __IEEE_BIG_ENDIAN
  /* ???  This conditional is true for the h8500 and the w65, defining H8300
     in those cases probably isn't the right thing to do.  */
  #define H8300 1
  #endif
  
  #ifdef __W65__
  #define _DOUBLE_IS_32BITS
  #define __SMALL_BITFIELDS
  #define __IEEE_BIG_ENDIAN
--- 3,27 ----
  
  /* exceptions first */
  /* ??? Why is much of this stuff duplicated with machine/ieeefp.h?  */
! #if defined (__H8300__) || defined (__H8300H__) ||  defined (__H8300S__) || 
defined (__H8500__)
  #define _FLOAT_ARG float
  #define _DOUBLE_IS_32BITS
+ #define __SMALL_BITFIELDS
  #define __IEEE_BIG_ENDIAN
+ #if defined (__H8300H__) || defined (__H8300S__)
+ #define __SIZE_TYPE__     long unsigned int
+ #define __PTRDIFF_TYPE__  long int
+ #else
+ #define __SIZE_TYPE__     unsigned int
+ #define __PTRDIFF_TYPE__  int
+ #endif
  /* ???  This conditional is true for the h8500 and the w65, defining H8300
     in those cases probably isn't the right thing to do.  */
  #define H8300 1
  #endif
  
  #ifdef __W65__
+ #define _FLOAT_ARG float
  #define _DOUBLE_IS_32BITS
  #define __SMALL_BITFIELDS
  #define __IEEE_BIG_ENDIAN
***************
*** 21,27 ****
--- 29,44 ----
  #undef UINT_MAX
  #define INT_MAX 32767
  #define UINT_MAX 65535
+ #endif
  
+ #ifdef __mips__
+ #ifdef __LONG_64_BIT__
+ #define __SIZE_TYPE__     long unsigned int
+ #define __PTRDIFF_TYPE__  long int
+ #else
+ #define __SIZE_TYPE__     unsigned int
+ #define __PTRDIFF_TYPE__  int
+ #endif
  #endif
  
  /* 16 bit integer machines */
--------------- clip --------------------------

 For the MIPS case, there aren't anything like '__H8300H__' for '-mh'
etc., just those '-D__SIZE_TYPE__=...' for '-mlong64' or not, so that
'__LONG_64_BIT__' is my addition for the '-mlong64' case to the specs
(my replacement for the '%{mlong64:-D__SIZE_TYPE=....}' is the 
'%{mlong64:-D__LONG_64_BIT__}'). The '__H8500__' case is unchecked, is
it a '__H8300__' or a '__H8300H__' alike ?

 Cheers, Kai
_______________________________________________
New CrossGCC FAQ: http://www.objsw.com/CrossGCC
_______________________________________________
To remove yourself from the crossgcc list, send
mail to crossgcc-request@cygnus.com with the
text 'unsubscribe' (without the quotes) in the
body of the message.