[omniORB] issues with solaris and egcs shared libraries with omniOrb 2.7.1

Duncan Grisby dgrisby@uk.research.att.com
Wed, 07 Apr 1999 08:11:02 +0100


On Tuesday 6 April, Guenther Grau wrote:

> Craig Brozefsky wrote:
> [...]
> > Is there a reason omniORB does not use automake/autoconf/libtool?
> > Perhaps the win32 platform and other needs preclude the use of those
> > tools, but they sure would make life easier for those of us trying to
> > port omniORB.  If there is no reason not to use them, perhaps I could
> > try and get omniORB autoconfigscated and libtoolized.
> 
> Sounds good to me. I don't know what the maintainers think about it,
> though :-)

The current omniORB build environment is a cut-down version of the
build environment we use for internal development. If someone wants to
come up with a different build environment, we'll make it part of the
distribution as long as it doesn't cause too much extra work for us.

> > With the omniORB binary package for sunos_5.6 built with SunProC I
> > cannot build the examples using egcs (even after changing the
> > platform/sunos_5.6.mk to use egcs) because when it tries to link there
> > are many unresolvable symbols.  At this point I cannot tell if these
> > are just issues with using egcs, or egcs with SunproC compiled
> > libraries, or just the libraries themselves.
> 
> I think this is a problem of c++ (and the compilers involved).
> In general I'd recommend NOT to mix libraries compiled with different
> compilers. This works mostly for c-libs, but is not standardized for
> c++.

It is indeed not a good idea to mix C++ object files created with
different compilers.

> I suggest you compile omniORB with the compiler you want to use,
> i.e. egcs (Be sure to get the latest egcs-1.1.2, as it containes
> some bug fixes for sparc). Or get the professional C++ compiler
> from sun, which is nice and fast, btw :-)

Watch out that egcs still has problems with exceptions in
multi-threaded code on some platforms.

Cheers,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --