egcs-1.1...

[Sven Radej]

Hi, 
I have problems with the following code:

#include <sys/wait.h>
(...)
int KProcess::exitStatus()  
{
  int _status = status;  //status is global
  return WEXITSTATUS(_status);
}

I tested it on gcc-2.7.2 and __USE_BSD is defined there (and passes). I am
sure that egcs-1.0.3 could  compile this, too.

egcs-1.1 cannot parse this: (in sys/wait.h):
(...)
#ifdef	__USE_BSD

/* Lots of hair to allow traditional BSD use of `union wait'
   as well as POSIX.1 use of `int' for the status word.  */

#ifdef	__GNUC__
#define	__WAIT_INT(status)						      \
  (__extension__ ({ union { __typeof(status) __in; int __i; } __u;	      \
		    __u.__in = (status); __u.__i; }))
#else
(...)

So it cannot use (in same file):
(...)
#define	WEXITSTATUS(status)	__WEXITSTATUS(__WAIT_INT(status))
(...)

Claiming:
kprocess.cpp: In method `int KProcess::exitStatus()':
kprocess.cpp:278: `_status' undeclared (first use this function)
kprocess.cpp:278: warning: ANSI C++ forbids declaration `__in' with no type

While parsing this lines in source:

int KProcess::exitStatus()
{
  int _status = status;
  return WEXITSTATUS(_status);
}

I don't know if libc-5.4.44 does something wrong or is it egcs. 
This is intel, linux 2.0.34, libc5.4.44, egcs 1.1.

egcs was compiled with -O2 flags,  --enable-shared and --prefix=/usr
Everything else runs, and it runs better than with egcs-1.0.3, faster,
smaller, sexier... 

Except of this little problem (and I have found this bug in other non-KDE
apps too), KDE compiles wery fine - and is required for koffice programs due
to corba-stuff and exceptions. So would you please be so kind to remove from
your FAQ  question 26 :-) ?

Thanks, 
--
 Sven Radej

KDE developer
radej@kde.org
Visit http://www.kde.org