[omniORB] omniidl bug?

[Duncan Grisby]

On Tuesday 7 December, "Borges, Mark" wrote:

> $ /opt/omniORB/scripts/omniidl -I. -bpython cdt.idl           
> cdt.idl:6: Integer literal is too large for unsigned long
> omniidl: 1 error.
> --------------------
> 
> Why? Neither omniidl2 nor python-stubber from the XEROX-PARC ILU-2.0beta1
> distribution complain.

[...]

>   if (value_ > 0xffffffff) {

[...]

> $ gcc -dumpversion    
> 2.95.2
> $ gcc -dumpmachine    
> sparc-sun-solaris2.6

I assume the problem is that on Sparc, the 0xffffffff is being
sign-extended to be -1. It's not obvious from the C++ spec what the
behaviour should be. Can you try the following enormous patch to see
if the problem goes away?


--- idlexpr.cc	1999/11/02 17:07:26	1.3
+++ idlexpr.cc	1999/12/08 09:55:13
@@ -144,7 +144,7 @@
 }
 _CORBA_ULong IntegerExpr::evalAsULong() {
 #ifdef HAS_LongLong
-  if (value_ > 0xffffffff) {
+  if (value_ > 0xffffffffU) {
     IdlError(file(), line(), "Integer literal is too large for unsigned long");
     return 1;
   }


Cheers,

Duncan.

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