Does anyone have any pointers on ways to map from commonly generated
IDL stubs to common "extension" mechanism of database applications?
(Sorry for the amorphous summary). I'm investigating the use of
CORBA as a coupling mechanism between two wildly different
1) a database that has a clunky C extension system.
- just packs args and calls function
- has a laborious spec and compile cycle: must use a manual GUI to
say "function X takes three args: a string, an int, etc. etc.)"
- basically just converts strings to C primitives, and calls the C
function(s) you linked against the kernel
- it works, but don't mention OO within 50 feet of it
2) a newer, CORBA-friendly web system (C++ & Perl)
Am I trying to push s*** up hill here? Would ODBC be a better bet?
(We're not using M$ standard at the moment and would need a good
reason to start) :-) Anyone heard of similar work to retro-fit CORBA
systems like this where the programmer is standing on the outside
looking in to an API that also contains main()?
(For the curious, the database system is called Universe).
-- James Preston, waiting for Godot.