Hi^ aLl!
My humble vote for MICO in Postgres CORBA development:
Latest mico (v2.2.3) supports C language mapping. Look into
$MICO_DIST/demo/corba-c.
BTW, mico is good choice for development 'cos it is _stable_ and
_complete_ (excluding Java support). omniORB is fastest but does not
have complete CORBA 2 implementation.
Mico has Any and DynAny support which is critical for PostgreSQL
user-defined types.
OPBit is too incomlete to be used in PostgreSQL in develpoment.
IMHO, the goal is to add CORBA support in Postgres, not to develop some
ORB.
And it is right, user can easy switch to any ORB in compile-time because
CORBA defines strict language mappings and rest of stuff is automaticaly
generaed by tools.
A little work with autoconf/automake needed to provide --with-mico or
--with-omniORB or whatever.
I think that CORBA support is huge advantage for Postgres. I simplifies
a lot of stupid things for developers and eliminates needs for such
things like ODBC, JDBC, etc. Direct type mappings and database access
will seriously improve performance too.
--
With respect,
Alexey Lukin,
JSC "CINET" http://www.ci.net.ua
phone/fax +380-462-101710, 101263