[omniORB] Is resolve_initial_references() implemented right?

[Jaap de Wolff]

Oeps, I pressed the reply button to soon.

> -----Original Message-----
> From:	Duncan Grisby [SMTP:dgrisby@uk.research.att.com]
> Sent:	Thursday, December 20, 2001 12:40 PM
> To:	Jaap de Wolff
> Cc:	Omniorb-List (E-mail)
> Subject:	Re: [omniORB] Is resolve_initial_references() implemented
> right? 
> 
> On Thursday 20 December, Jaap de Wolff wrote:
> 
> > I am linking to two instances of the omniORB library.
> > (Or in fact I am linking to even five instances of the omniORB library)
> > This is by design:
> > I want to be able to start one executable, which loads depending on a
> > configuration file different DLL's.
> > Each of those DLL's will load his own instance of the omniORB library.
> 
> Why?  Why not make them share a single version of the omniORB library?
> What advantage is there in trying to give each DLL its own version of
> omniORB?
	[Jaap de Wolff] 
	When I use Load Library to load a dll, this dll will first load the
dll's it is depending on. It will share the code, but not the data. (There
are ways to make it possible to let DLL's share data also, but you have to
do this in a ugly way)

> > Also I do not think it is specific to windows, I think that when I port
> this
> > program to linux, I still will have the same problem.
> 
> I don't think so. If you have two Linux shared libraries, both linked
> against the omniORB shared libraries, and you dlopen() them, only one
> copy of the omniORB library will be loaded.
	[Jaap de Wolff]  
	In that case Linux I could better use Linux. But unfortunetly I also
have to use also code on the same PC with only runs on windows .. :(
> Cheers,
> 
> Duncan.
> 
> -- 
>  -- Duncan Grisby  \  Research Engineer  --
>   -- AT&T Laboratories Cambridge          --
>    -- http://www.uk.research.att.com/~dpg1 --