Re: Fwd: [Bayonne-devel]timespec *, timespec *)' throws different exceptions (compiling commoncpp 1.0.7)

[Federico Montesino Pouzols]

        I have been trying to add the AC_TRY_COMPILE, but I have found
that it will not work in FreeBSD due to some funny #defines related to
posix versions. So I have just added the "#ifdef __freeBSD__" you
suggested around the declaration in config.h (that way, those RH and
Debian should not complain too).

        When the AC_TRY_COMPILE check runs under FreeBSD, it reports
that the declaration of nanosleep is there, and any simple program
that #includes time.h or sys/time.h and uses nanosleep will compile
and work well. However, nanosleep is defined in time.h only if
_P1003_1B_VISIBLE_HISTORICALLY is defined (in sys/_posix.h). It seems
that the order in which CC++ includes some system headers implies that
the HISTORICALLY symbol is not defined because some posix version
symbols it depends on are affected.

On Thu, Dec 05, 2002 at 06:56:51PM -0500, David Sugar wrote:
> I think we will have to add a AC_TRY_COMPILE for it...I don't have any
> issue with this on any of the systems I normally test with, but I am
> gathering it is an issue at least on some older rh (pre 7.3, and,
> similarly pre debian r3) systems.  If we added it in initially for
> FreeBSD, maybe we should for the moment just have a "#ifdef __FreeBSD__"  
> around it...
> 
> On Thu, 5 Dec 2002, Federico Montesino Pouzols wrote:
> 
> > 
> >     Oh, I had not had this problem yet, but I have just read a
> > mail posted to bug-commoncpp whose author seems to complain about the
> > same issue. The problem comes from the following lines in config.h:
> > 
> > #ifdef HAVE_PTHREAD_NANOSLEEP
> > #ifndef HAVE_PTHREAD_DELAY
> > #define HAVE_PTHREAD_DELAY
> > extern "C" int nanosleep(const struct timespec *rqtp, struct timespec 
> > *rmtp);
> > #define pthread_delay(x) nanosleep(x, NULL)
> > 
> >     I recall we recently added the declaration of nanosleep
> > because it is missing (or requires the definition of some weird posix
> > version related symbols) in FreeBSD. However, this declaration is
> > conflicting with the one in time.h, which difers because of the 'throw
> > ()'. I guess we will have to make an AC_TRY_COMPILE with nanosleep to
> > see if it is already declared, and declare nanosleep in config.h only
> > when strictly necessary.
> > 
> > On Thu, Dec 05, 2002 at 04:51:01PM -0500, David Sugar wrote:
> > > Here's a curious problem I had not seen before...nanosleep is defined in 
> > > at 
> > > least some glibc releases under time.h...
> > > 
> > > ----------  Forwarded Message  ----------
> > > 
> > > Subject: [Bayonne-devel]timespec *, timespec *)' throws different 
> > > exceptions 
> > > (compiling commoncpp 1.0.7)
> > > Date: 05 Dec 2002 15:57:11 -0500
> > > From: Erik Enge <address@hidden>
> > > To: address@hidden
> > > 
> > > Hi.
> > > 
> > > I'm getting this when trying to compile CommonCPP version 1.0.7:
> > > 
> > > make[1]: Entering directory `/share/pak/commoncpp2/commoncpp2-1.0.7/src'
> > > /bin/sh ../libtool --mode=compile c++ -DHAVE_CONFIG_H -I. -I. 
> > > -I../include   
> > >   -I../src -DCCXX_EXPORT_LIBRARY  -I/usr/local/include -D_GNU_SOURCE 
> > >  -I/usr/include/libxml -I/usr/include -I../include -c thread.cpp rm -f
> > >  .libs/thread.lo
> > > c++ -DHAVE_CONFIG_H -I. -I. -I../include -I../src -DCCXX_EXPORT_LIBRARY
> > >  -I/usr/local/include -D_GNU_SOURCE -I/usr/include/libxml -I/usr/include
> > >  -I../include -c thread.cpp  -fPIC -DPIC -o .libs/thread.lo In file 
> > > included
> > >  from thread.cpp:41:
> > > ../include/cc++/config.h:524: declaration of `int nanosleep (const
> > > timespec *, timespec *)' throws different exceptions
> > > /usr/include/time.h:284: than previous declaration `int nanosleep
> > > (const timespec *, timespec *) throw ()'
> > > make[1]: *** [thread.lo] Error 1
> > > make[1]: Leaving directory `/share/pak/commoncpp2/commoncpp2-1.0.7/src'
> > > make: *** [install-recursive] Error 1
> > > 
> > > Is there anything I can do to make it go away?  I'm getting this error
> > > when I try to compile 1.0.6 also - which previously compiled just fine.
> > > I don't think I changed anything between the compiles.  Confusing.
> > > 
> > > (Couple of unrelated comments; bayonne.sf.net and
> > > gnu.org/software/bayonne differ greatly in content but not in design and
> > > the FAQ page (http://bayonne.myip.org/cgi-bin/bayonne.faq) seems to be
> > > down).
> > > 
> > > Thanks,
> > > 
> > > Erik.
> > > 
> > > 
> > > -------------------------------------------------------
> > > This sf.net email is sponsored by:ThinkGeek
> > > Welcome to geek heaven.
> > > http://thinkgeek.com/sf
> > > _______________________________________________
> > > Bayonne-devel mailing list
> > > address@hidden
> > > https://lists.sourceforge.net/lists/listinfo/bayonne-devel
> > > 
> > > -------------------------------------------------------
> > > 
> > 
>