Separate Compilation

[Gary D. Duzan]

   I'm currently designing a system to work with OmniORB with an IDL
spec that is getting larger all the time. I don't know what a normal
IDL size would be, but mine is currently about 700 lines of IDL, which
is generating 4732 lines of header and 9477 of skeleton code.  And the
skeleton compiles to a 2.5MB object file (with debugging symbols) in 45
seconds without optimization or 45 minutes with optimization using
AIX/xlC. The problem is that portions of the system are only going to
use particular interfaces but still have to carry the whole object
file. I could use separate IDL files, but then I'd have to change the
structure since I have one big module encompasing everything, with
submodules and such inside it.
   Has anyone else run into this sort of problem? How did you deal with
it? Any chance of having flags in omniidl2 which produce separate
skeleton code files for each module/interface/whatever? Or maybe
#pragmas in the IDL to specify which parts go in which compilation
unit.
   Thanks.

					Gary Duzan
					GTE Laboratories