A hole in the serialization spec

[Chris Hegarty]

On 17/02/14 15:50, David M. Lloyd wrote:
> ...
>
> Honestly I'm coming to the conclusion that the spec should just be
> updated to match the existing behavior.  At least this way, it is
> possible to establish and explain exactly what is happening and why, and
> what the effect is on the wire protocol, and how to cope with changes to
> the affected class hierarchies in a compatible way.

Resurrecting an old thread, since I just realized I didn't reply to this.

I have no objection, and would support any effort/attempt to try and 
resolve this issue.

-Chris.