Re: orbit-idl-2 new option request, please comment ...

[Jules Colding <colding omesc com>]

On Thu, 2005-08-11 at 05:26 +0800, KC wrote:
> Also, methods emitted by "orbit-idl-2 --skeleton-impl"
> are "static" function ...  I think this should be re-considered
> too.   For large project, "extern" maybe more useful ... so
> maybe we should have option "--skeleton-impl-no-static".

I agree that splitting the emitted skelimpl-files up in c-files and
headers with "extern" declarations would be good. I missed that option
myself a few months ago, but I solved my needs by using the "--onlytop"
option to orbit-idl-2. 

This make use of IDL files that includes other IDL files almost as easy
as if orbit-idl-2 produced appropriate headers with extern declarations
and such for skelimpl-files.

The drawback of the current approach is that several files including the
same "foo_impl.c" will make the resulting binary slightly larger. Using
appropriately defined headers would, IMHO, be better.


Best regards,
  jules