[omniORB] Re: omni installation on SCO.

[Allan Clark]

James;

Yes, it was me who suggested the use of FSU Pthreads.

(As I've passed on directly to Roshan) Roshan seems to be experiencing
the same troubles I had porting to UnixWare.  I've seen and solved these
problems.  I have to get my hands on an OpenServer box to resolve it,
since I don't have one with me here (on site with a client on a
2200-system rollout).  No OpenServer in sight.

I'll try to get access and open my worknotes in the next few days.

FYI

Allan

James Riden wrote:
> 
> Roshan Alwares <roshana@mahindrabt.com> writes:
> > Dear James,
> >
> > as suggested by you
> 
> erm, Alan suggested this I think.
> 
> > we have downloaded and installed latest g++ compiler
> > and FSU-Pthreads 3.5 package.
> > after that the following error was encountered,
> >
> > **********
> > g++ -c  -fhandle-exceptions -Wall -Wno-unused -I../include
> > -DIDL_CFE_VERSION='"
> > 1.3.0"' -DCPP_LOCATION=\"gcc\" -I.  -I../../../../include -D__x86__
> > -D__osr5__ -
> > D__OSVERSION__=5 -o drv_fork.o drv_fork.cc
> > cc1plus: warning: -fhandle-exceptions has been renamed to -fexceptions
> > (and is n
> > ow on by default)
> > drv_fork.cc: In function `void DRV_fork()':
> > drv_fork.cc:152: implicit declaration of function `int fork(...)'
> > drv_fork.cc:158: implicit declaration of function `int wait(...)'
> > make[3]: *** [drv_fork.o] Error 1
> > make[3]: Leaving directory
> > `/u02/ra003416/omni/omniORB_280/src/tool/omniidl2/dri
> > ver'
> > make[2]: *** [export] Error 1
> > make[2]: Leaving directory
> > `/u02/ra003416/omni/omniORB_280/src/tool/omniidl2'
> > make[1]: *** [export] Error 1
> > make[1]: Leaving directory `/u02/ra003416/omni/omniORB_280/src/tool'
> > make: *** [export] Error 1
> >
> > *****************
> > we are unable to understand the error. can u please help us and tell us
> > some way out to fix the error.
> > thanking you in anticipation,
> > Roshan.
> 
> Roshan,
>   do a "man fork" and check you're including the appropriate
> file. ("unistd.h" possibly ?)  I'm afraid I don't have a copy of SCO
> to try this on. I guess there are #include "unistd.h" directives
> surrounded by various #ifdef {platform X} / #endif pairs - the SCO one
> may be missing or wrong.
> 
> Oh yes, here we go - eg.
> 
> #if defined(__irix__)
> #include      <unistd.h>              // POSIX standard types
> #include      <wait.h>                // POSIX definition of wait()
> #endif
> 
> Doesn't appear to be an SCO one.
> 
> cheers,
>  James