Re: one more stupid patch.

[Michael Meeks <michael helixcode com>]

Hi Mathieu,

On 23 Nov 2000, Mathieu Lacage wrote:
> As maciej said: I have to parse the files separatly so, I need real 
> IDL file, not only bits of IDL declaration.
  
        Sure; I suspected that was the case.

> Also, we should be thinking about the day when someone wants to
> implement a bonobo component with the IDL: this will happen and this
> is the kind of detail which makes it possible: just compile the
> bonobo-control.idl file.
     
        Sure.
     
> > the #pragma inhibit stuff does not matter.
>
> yes, and I am not really willing to do the #pragma stuff: it would
> require quite some testing: I will probably do it later.

        The #pragma inhibit stuff is evil anyway; I don't want it all over
the idl if it is at all possible to avoid, and I see no reason for it.    
  
        Anyway; the patch is nice Mathieu; please commit when you like.

> As a side note, parsing the IDL files to feed data to the IR is not 
> the niocest thing you can dream of. I was toying with the idea of
> reusing mozilla's typelibs but this will wait till I really need it.
  
        I am working on a rather more cunning ( Elliot thinks slow
:-) type based marshaling scheme #2, the first version was too clever by  
half. Anyhow, it should give a nice win for scripting languages,
performance, stub size, cache locality, orb complexity, the price of oil  
etc.
  
> So, may I commit ?
  
        Yes, please do.
  
        Thanks,
  
                Michael.

-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot