Re: orbitcpp idl backend & --onlytop

[Bowie Owens <bowie owens csiro au>]

Hi Robert,

Yes --onlytop (and #pragma inhibit) suppresses the generation of 
definitions for idl elements. The problem is, orbitcpp builds it's own 
representation of the IDL tree from the C data-structures it is given by 
libidl. Because the inhibited elements are never defined they don't 
appear in the C++ representation. There are a few places in the orbitcpp 
backend that when undefined/inhibited elements turn up it is assumed to 
be an error. When, however, it is a legitmate condition. Bug 
http://bugzilla.gnome.org/show_bug.cgi?id=101455 is the same problem but 
with exceptions rather than interfaces. I have attached a comment to the 
above bug so that we don't forget about this. I'll try and fix it as 
soon as possible.

For the moment avoid the problem by not using --onlytop or pragma inhibit.

rm wrote:

> hi all, 
> 	i'm trying to use the onlytop flag of orbit-idl-2. i believe
> the semantics of this flag are to suppress the declarations for
> #include'd idl files and implicitly including the headers which are
> assumed to already exist. this seems to work fine for the c backend,
> but fails for orbit-cpp
>
> for example: 
>
> /*-----foo.idl-----------------------*/
>
> module foo {
>
>  interface bar {
>     
>  };
> };
> /*-----B.idl------------------------*/
> #include "foo.idl"
>
>
> module B {
>  interface C {
>    ::foo::bar get_foo();
>  };
>
> };
> /*----------------------------------*/
> and then trying to compile it (first normally) (orbitcpp CVS):
>
> orbit-idl-2 -l cpp B.idl
> orbit-idl-2 2.6.0 compiling
>  mode, show preprocessor errors, passes: stubs skels common headers skel_impl
> imodule
>
> orbit-idl-2 --onlytop -l cpp B.idl
> orbit-idl-2 2.6.0 compiling
>  mode, show preprocessor errors, passes: stubs skels common headers skel_impl
> imodule
>
> foo.idl:4: unknown identifier ::foo::bar
>
> and finally with the C backend
>
> orbit-idl-2 --onlytop  B.idl
> orbit-idl-2 2.6.0 compiling
>  mode, show preprocessor errors, passes: stubs skels common headers skel_impl
> imodule
> -------------------------------
>
>
> 	it appears that although the backend does not get the parse
> tree of the included file, libIDL does know about it and cross checks
> references to it. it's unclear though how much information libidl
> actually keeps around to it. the docs seem to imply it keeps
> namespace information around for type lookup.
> 	is it the case that the orbitcpp backend keeps it's own type
> information?
>
> 	thanks, 
> 	rob
>
>
>
> ----
> Robert Melby
> Georgia Institute of Technology, Atlanta Georgia, 30332
> uucp:     ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt4255a
> Internet: async cc gatech edu
> _______________________________________________
> orbitcpp-list mailing list
> orbitcpp-list gnome org
> http://mail.gnome.org/mailman/listinfo/orbitcpp-list
>  


--
Bowie Owens

CSIRO Mathematical & Information Sciences
phone  : +61 3 9545 8055
fax    : +61 3 9545 8080
mobile : 0425 729 875
email  : Bowie Owens csiro au