Re: one more stupid patch.

[Maciej Stachowiak <mjs eazel com>]

Michael Meeks <michael helixcode com> writes:

> Hi,
> 
> On 22 Nov 2000, Maciej Stachowiak wrote:
> > Mathieu Lacage <mathieu eazel com> writes:
> > > may I commit ?
> > 
> > You should also add the magic to suppress generation of real stubs and
> > skeletons in generated files (#pragma inhibit and all) so if you
> > include one of these IDL files directly you don't get linker problems.
>   
>         Why then are you compiling the IDL if you are going to link
> against the shared stubs and skels in libbonobo directly anyway ? I can
> only see this patch as being helpful ( perhaps ) for scripting languages,
> where one needs to parse the IDL file; is this what the patch is intended
> for Mathieu ?
>     
>         If it is for a speedup, I think orbit-idl is one extremely fast
> part of the build system;
> 
>         Not that the patch isn't nice, but what are you trying to achieve?
>   

Mathieu is writing a daemon that will feed the Interface Repository
from IDL files found on the system. As such, it needs to be able to
parse individual IDL files, which requires files to have the proper
#includes, so for his particular purposes, the #pragma inhibit stuff
does not matter.

I figure as long as he is making this change he should do the other
necessary steps to make the IDL files properly individually includable
(including only what you need is generally desirable from a code
cleanliness point of view I think).

 - Maciej