Is -D_XOPEN_SOURCE=500 necessary in gcc/config/sparc/sol2.h?

Is -D_XOPEN_SOURCE=500 necessary in gcc/config/sparc/sol2.h?

[Gerhard Franke]

Hello,

since I didn't get any answer to my mail "Compile problems with g++-3.0"
( http://gcc.gnu.org/ml/gcc-help/2001-07/msg00066.html ) from July 5th
I tried to figure out why g++-2.95.3 behaves differently to g++-3.0
in the handling of #include <sys/socket.h>.

In short:
gcc/config/sparc/sol2.h of gcc-3.0 defines CPLUSPLUS_CPP_SPEC to
-D_XOPEN_SOURCE=500 -D_LARGEFILE_SOURCE=1 -D_LARGEFILE64_SOURCE=1 \
-D__EXTENSIONS__
while gcc-2.95.3 does not.

The -D_XOPEN_SOURCE=500 triggers the odd(?) behavior of g++ or the
C preprocessor described in my other mail by defining _XPG4_2 in
sys/feature_tests.h. This activates, for example, the
'#define connect __xnet_conect' line in sys/socket.h and from now on
the preprocessor replaces every occurrence of a connect string with
__xnet_connect.

Can I safely remove the -D_XOPEN_SOURCE=500 or replace it with some
other less offensive defines?


Gerhard

Re: Is -D_XOPEN_SOURCE=500 necessary in gcc/config/sparc/sol2.h?

[Alexandre Oliva]

On Jul 10, 2001, Gerhard Franke <Gerhard.Franke@mni.fh-giessen.de> wrote:

> Can I safely remove the -D_XOPEN_SOURCE=500 or replace it with some
> other less offensive defines?

I'm afraid this would render libstdc++-v3 useless.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Re: Is -D_XOPEN_SOURCE=500 necessary in gcc/config/sparc/sol2.h?

[Gerhard Franke]

On Tue, Jul 10, 2001 at 03:45:12PM -0300, Alexandre Oliva wrote:
> On Jul 10, 2001, Gerhard Franke <Gerhard.Franke@mni.fh-giessen.de> wrote:
 
> > Can I safely remove the -D_XOPEN_SOURCE=500 or replace it with some
> > other less offensive defines?
 
> I'm afraid this would render libstdc++-v3 useless.

Why this?

I just recompiled gcc-3.0 without -D_XOPEN_SOURCE=500 and now it
seems that everything is working...

Why doesn't g++ let the "application writer" decide wether to
define _XOPEN_SOURCE=500 or not?

from sys/feature-tests.h:
* application writers wishing to utilize UNIX 98 functionality
* must define _XOPEN_SOURCE=500.  This turns on UNIX 95 functionality
* which is a subset of UNIX 98, and also turns on POSIX Realtime and
* POSIX Threads functionality.


Gerhard

Re: Is -D_XOPEN_SOURCE=500 necessary in gcc/config/sparc/sol2.h?

[Alexandre Oliva]

On Jul 11, 2001, Gerhard Franke <Gerhard.Franke@mni.fh-giessen.de> wrote:

> On Tue, Jul 10, 2001 at 03:45:12PM -0300, Alexandre Oliva wrote:
>> On Jul 10, 2001, Gerhard Franke <Gerhard.Franke@mni.fh-giessen.de> wrote:
 
>> > Can I safely remove the -D_XOPEN_SOURCE=500 or replace it with some
>> > other less offensive defines?
 
>> I'm afraid this would render libstdc++-v3 useless.

> Why this?

It's a known problem in libstdc++-v3.  Currently, it depends on
features and declarations that are only present with this define, on a
number of platforms.  It's in the wishlist to remove this dependence,
but it couldn't be done in time for GCC 3.0.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me