Hi Alex,
On Thu, Jun 01, 2000 at 09:59:41AM +0100, Alex Hornby wrote:
> I agree that its nicer to put both stubs into a the library, but
> that's not how our project does it :)
Alrighty. :-)
> If no-one was ever going to split the client and server stubs there
> would be no need to generate them separately...
Why do you say that? The stubs and skeletons serve two different
purposes, so it makes perfect sense to generate them separately. What
I was saying was that the stub and skeleton object files can be placed
in the same library without affecting the size of the client binary
that the library is linked into since the linker will only pull in the
stub (client side) objects. Did I misunderstand your original point?
> Incidentally, I think this is what mico does. I ported an old version
> of our trading system to mico from Visibroker (before choosing
> TAO:).
:-)
> In order to use common makefiles and source I had to create a
> script (call it idlcomp) that produced a dummy client header and stub
> that just #included the integrated stub and called the integrated
> stub the same name as the visibroker server stubs.
Ack, that must not have been pleasant. 8)
-Ossama
--
Ossama Othman <[EMAIL PROTECTED]>
Distributed Object Computing Laboratory, Univ. of California at Irvine
1024D/F7A394A8 - 84ED AA0B 1203 99E4 1068 70E6 5EB7 5E71 F7A3 94A8
