[omniORB] omniORB, solaris10x86 and CORBA::LongDouble

Duncan Grisby duncan at grisby.org
Tue Apr 4 16:37:05 BST 2006


On Tuesday 4 April, "Roloff, Craig A" wrote:

> We use the configure script.  It shows:
> 
> checking for long double... yes
> checking size of long double... 16

So you do indeed have long double...

> Yet the preprocessor shows:
> 
> ../../../bin/omniidl -bcxx -Wba -p../../../../src/lib/omniORB -Wbdebug
> -v -nf -P -WbF -ComniORB4 ../../../../idl/corbaidl.idl
> omniidl: Importing back-end 'cxx'
> omniidl: 'cxx' imported from
> '../../../../src/lib/omniORB/omniidl_be/cxx/__init__.pyc'
> omniidl: Preprocessing '../../../../idl/corbaidl.idl' with
> '/project/aero/swdev/perf/omni407/hp11i/lib/omnicpp -lang-c++ -undef
> -D__OMNIIDL__=0x2420 -DHAS_LongLong -D__OMNIIDL_CXX__
> "../../../../idl/corbaidl.idl"'
> 
> We've tried to trace this problem before without any luck.  Can you
> suggest where to look?

The defines come from idlpython.cc in the IdlPyPlatformDefines function,
where it just sets the defines according to what's set when that file is
compiled. Something somewhere is preventing HAS_LongDouble from being
set, or is undefining it once it's set. I wonder if Python.h is doing
something that prevents SIZEOF_LONGDOUBLE being right. All I can suggest
you do is to add some things to the files being #included to trace why
HAS_LongDouble is not set when idlpython.cc is compiled.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --



More information about the omniORB-list mailing list