Re: ORBit2 ...

[<dahaverk rockwellcollins com>]

I'll let people at the OMG know that ORBit supports the recommendations in
the C RTF since they made sence to the ORB implementers.

The Issues that people are bringing up really need to be captured in a
document or something that I can send to people at the OMG.

Is someone going to volunteer to collect the information into a document?
I can make sure the correct people at the OMG are made
aware of the information.

>        I would recoomend using ORBit2 - not only is the project
>  drasticaly re-structured, but hacking is easier, we can't really touch
>  ORBit-stable currently, it has a reduced memory footprint, abstracted IO
>  layer, is far more generic and powerful etc. etc. :-)

The abstracted IO layer sounds usefull.   As for my progress on getting
ORBit
onto LynxOS.  I'm stuck with the "gettext" stuff.   I haven't quite figured
out
how to get gettext to configure for a target nicely.  (I'll chalk up some
of
the trouble to my inexperience in customizing the Autoconf stuff)

I should look at ORBit2 and see if it is more friendly to embedded systems.

I've had some thoughts about gettext... At least for an embedded system, I
need
to be less dependent on a package that seems to require additional setup
and
configuration just to read command lines.   I'm even considering if I could
"hack"
out the gettext dependencies....  Any thoughts? Especially since when the
embedded
system is deployed, it won't care about having internationalized command
lines like
 a non-embedded system.


Later,

-Dave

>
>        Regards,
>
>                Michael.
>
> --
> mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot




Michael Meeks <michael@ximian.com> on 08/13/2001 12:45:51 PM

To:   <dahaverk@rockwellcollins.com>
cc:

Subject:  ORBit2 ...



Hi David,

On Fri, 10 Aug 2001 dahaverk@rockwellcollins.com wrote:
> The OMG web site is probably only usably fast if you are in the OMG
> office. ;-)

        This needs fixing - it's totaly useless. Surely they can get a
SuperJanet connection or somesuch - it's a charity no ?

> Just so you know.... If the C RTF final report makes sense to you (and
> Others) and ORBit2 can support it...

        Well - the changes are so small that I think we support them
already; and most are just impl. details that we happened to choose the
new standard way anyway it seems :-) any->_value = (char **) for strings
eg.

>  I can let people know at the OMG and it can be moved out of it's
> limbo state.  I think it was the Lockheed Martin OMG rep that told the
> OMG they wouldn't implement it's recommendations... Resulting it it's
> current limbo state.

        Hmm - I think we need to do no work - that's why I'm happy.

> BTW, my main interest in ORBit is to get it running on embedded
> systems. Maybe, adding some of the CORBA Real-Time changes.

        I would recoomend using ORBit2 - not only is the project
drasticaly re-structured, but hacking is easier, we can't really touch
ORBit-stable currently, it has a reduced memory footprint, abstracted IO
layer, is far more generic and powerful etc. etc. :-)

        Regards,

                Michael.

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