Re: orbit-io-thread ?

[Jules Colding <colding omesc com>]

On Tue, 2005-02-22 at 09:34 +0000, Michael Meeks wrote:
> Hi Jules,
> 
> On Mon, 2005-02-21 at 19:38 +0100, Jules Colding wrote:
> > >       The spec explicitly allows for vendors to add their own "magical
> > > strings", so we're doing nothing wrong here. 
> > 
> > That is correct, we are doing nothing wrong spec-wise. It is just, IMHO,
> > not good taste to attach any special meaning to substrings in the ORB
> > name.
> 
> 	Heh ;-) quite possibly not - but it's the only sensible way to do this.
> Also - AFAIK there is no particular need to set 'link_set_thread_io
> (TRUE)' in the ORB init; ultimately (AFAIR) as soon as you create a
> thread-aware POA then it does this for you anyway [ of course, it's
> prolly not wise to do that from some random thread but ... ].

True: orbit-adaptor.c line 29. Didn't see that one...

Are there any reason to keep the magical string if the POA thread hint
takes care of this?

I would *really* love to shoot it ;-)

Best regards,
  jules