omniORB 4.1.1 bugs

The following bugs in omniORB 4.1.1 have been fixed. Update from Subversion to get the fixes.


Summary: Assertion falure in Any unmarshalling (bug number 1)
Date: Mon Oct 15 14:56:37 BST 2007
Reported by: Jian Wu, Clemens Fischer
Link: http://www.omniorb-support.com/pipermail/omniorb-list/2007-October/028948.html
Description: Enum TypeCodes were not marked as complete, meaning that receiving a struct containing an enum would trigger an assertion failure.

Summary: Escaped identifiers in unions cause incorrect errors (bug number 2)
Date: Mon Oct 29 12:26:00 GMT 2007
Reported by: Will Denissen
Link: http://www.omniorb-support.com/pipermail/omniorb-list/2007-October/028956.html
Description: omniidl would incorrectly report errors for union members clashing with keywords, but escaped with a leading underscore.

Summary: localhost server transport rule does not work for IPv4 in IPv6 (bug number 3)
Date: Mon Oct 29 12:28:58 GMT 2007
Reported by: Joe Lee
Description: The localhost server transport rule failed to cope with IPv4 addresses encapsulated in IPv6, meaning connections were incorrectly denied.

Summary: Connections closed during upcalls could be leaked (bug number 4)
Date: Fri Nov 23 14:20:15 GMT 2007
Reported by: Emmanuel Taurel
Description: In thread pool mode, connections closed by a client while the server was in an up-call could be leaked.

Summary: Assertion failure in Any marshalling with nil objref (bug number 5)
Date: Mon Nov 26 16:11:15 GMT 2007
Reported by: Peter S. Housel
Link: http://www.omniorb-support.com/pipermail/omniorb-list/2007-November/029040.html

Summary: Segfault on exit of omniidl (bug number 6)
Date: Wed Dec 5 11:10:28 GMT 2007
Reported by: Juergen Weber
Description: omniidl would segfault on exit if given a forward declaration to a previously fully-declared valuetype.
Link: http://www.omniorb-support.com/pipermail/omniorb-list/2007-December/029062.html

Summary: Incorrect generated code (bug number 7)
Date: Wed Dec 5 11:10:28 GMT 2007
Reported by: Juergen Weber
Description: Incorrect generated code for ::CORBA::AbstractBase.
Link: http://www.omniorb-support.com/pipermail/omniorb-list/2007-December/029067.html

Summary: Race condition between Peek and select thread when data in buffer. (bug number 8)
Date: Sun Dec 9 01:32:28 GMT 2007
Reported by: Dirk Siebnich
Description: A race between the select thread and a thread doing peek when there was data already in the connection buffer could lead to a lock-up.