[omniORB] omniORB 3 pre-release 3 is now available

[Gary D. Duzan]

   That does appear to have dealt with the problem. Thanks.

					Gary Duzan
					GTE Laboratories
					Verizon Communications



In Message <Pine.SO4.4.04.10007031554110.1568-100000@onion> ,
   David Scott <djs@uk.research.att.com> wrote:

=>Hi,
=>
=>On Mon, 3 Jul 2000, Gary D. Duzan wrote:
=>
=>>    Under AIX 4.2.1 and xlC 3.1.4, I get this when compiling the
=>> one of our application stubs:
=>>
=>> [undefined typecode error message snipped]
=>
=>This bug has now (hopefully) been fixed in our internal cvs tree and will
=>propagate to the external server overnight. Since the patch is so small,
=>here it is:
=>
=>--- src/lib/omniORB2/omniidl_be/cxx/dynskel/typecode.py	2000/06/26
=>16:23:27	1.14.2.6
=>+++ src/lib/omniORB2/omniidl_be/cxx/dynskel/typecode.py	2000/07/03
=>14:42:34
=>@@ -439,7 +439,19 @@
=>         memberType = child.memberType()
=>         if isinstance(memberType, idltype.Declared):
=>             memberType.decl().accept(self)
=>+        elif isinstance(memberType, idltype.Sequence):
=>+            # anonymous sequence (maybe sequence<sequence<...<T>>>)
=>+            # Find the ultimate base type, and if it's user declared then
=>+            # produce a typecode definition for it.
=>+            base_type = memberType.seqType()
=>+            while isinstance(base_type, idltype.Sequence):
=>+                base_type = base_type.seqType()
=> 
=>+            # careful of recursive structs
=>+            if isinstance(base_type, idltype.Declared) and \
=>+               not(recursive(base_type.decl())):
=>+                base_type.decl().accept(self)
=>+                        
=>
=>Sorry about that!
=>
=>-- 
=>  Dave Scott, Research Engineer, AT&T Laboratories Cambridge UK