Mailing-List: contact dev-help@geronimo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@geronimo.apache.org
Received-SPF: neutral (asf.osuosl.org: local policy)
Date: Fri, 1 Jul 2005 13:37:19 -0400 (EDT)
From: Aaron Mulder <ammulder@alumni.princeton.edu>
To: dev@geronimo.apache.org
Subject: Re: Questions about CORBA
In-Reply-To: <42C57CB7.9010705@dekasoft.com.ua>
Message-ID: <Pine.LNX.4.58.0507011329260.5464@saturn.opentools.org>
References: <42C4FFA2.6080502@apache.org> <42C50AFA.20904@gmail.com>
 <42C578AE.6000204@toolazydogs.com> <42C57CB7.9010705@dekasoft.com.ua>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE

=09I'm not sure there's a list of problems we had with OpenORB,
though you might poke around the dev list archives.  I haven't really been
involved in the CORBA work, but I've heard several people who were
involved say that OpenORB has issues, and other open source ORBs that
might be candidates don't have Apache-style licenses.  I don't think we're
thrilled with the prospect of coding directly to the JVM ORB, but it got
us over the TCK hump, so it was certainly expedient.

=09As far as I know, we're certainly willing to look at alternatives=20
at this point.  But to go back to your question, someone else may need to=
=20
address the specific problems that were encountered in the past.

=09To Rick and Alan, there's an option 4, Implement Our Own ORB.  I
don't think we're super-excited to start a project of that magnitude from
scratch (we do tend to favor the whole "incorporate best of breed project" =
=20
approach!), but I imagine it would be at least be considered if nothing
else worked out.

Aaron

On Fri, 1 Jul 2005, Viacheslav N tararin wrote:
> Hi.
> Where I can see where OpenORB need to be fixed to be usefull gor Geronimo=
?
>=20
> Alan D. Cabrera =D0=BF=D0=B8=D1=88=D0=B5=D1=82:
>=20
> > There is an option 3, which we did not pursue because of time=20
> > constraints, fix and use OpenORB.
> >
> > Rick, I'm glad to hear that you're working on making our interop code=
=20
> > more portable. It's hard work and it definitely needs to be done. I'm=
=20
> > also glad to see that you're discussing your ideas w/ the community;=20
> > you are an example for us all.
> >
> > I am happy to help out where I can. Can you file a Jira issue so that=
=20
> > people know that you're cleaning up that part of the universe?
> >
> >
> > Regards,
> > Alan
> >
> > On 7/1/2005 2:20 AM, Rick McGuire wrote:
> >
> >> I would state the options "use the ORB in the JVM" or "use some other=
=20
> >> open source implementation". Currently, Geronimo is using the Sun=20
> >> 1.4.2 ORB, and the OpenORB code is going to be removed very shortly.
> >> Unfortunately, the Geronimo code wrapped around the Sun ORB directly=
=20
> >> references Sun internal ORB classes (and in a couple cases,=20
> >> subclasses major ORB classes), so it is highly release dependent. For=
=20
> >> example, the current code won't compile cleanly on Sun's Java 5, or=20
> >> IBM's 1.4.2, because the underlying ORB code is completely different.=
=20
> >> Currently, I see two different solutions to this problem: 1) Write a=
=20
> >> different JVM-specific adaptor for each environment we wish to=20
> >> support, or 2) Try to find a means of implementing the current=20
> >> function using completely portable interceptor code.
> >> Solution 1) presents an interesting build problem, because regardless=
=20
> >> of which JVM you used to build Geronimo, at least one of the adaptors=
=20
> >> will fail to build because the required classes wouldn't be present.=
=20
> >> I still hopeful that solution 2) is possible, but I'm guessing that=20
> >> we might still end up with some small piece of JVM-specific code.
> >>
> >> Rick
> >>
> >> Jacek Laskowski wrote:
> >>
> >>> Hi,
> >>>
> >>> I wonder what options do we have to use as a CORBA implementation in=
=20
> >>> Geronimo (as a ASF project). I'm not familiar with the nuances of=20
> >>> miscellaneous licenses, but unless I'm mistaken we don't have too=20
> >>> much choices - Sun's ORB or OpenORB, right? Again, unless I'm=20
> >>> mistaken if I were to be asked I'd say we ain't too happy with both,=
=20
> >>> but couldn't say why :( Wiki is silent on this, too.
> >>>
> >>> Do we have to build our own CORBA implementation...well, I remember=
=20
> >>> Dain has been working on CORBA stuff in OpenEJB. Is it a complete=20
> >>> rewrite or a wrapper around OpenORB. Hmm, but I also remember we=20
> >>> talked about OpenORB troubles.
> >>>
> >>> Yes, I'm lost and asking for a help - could anyone answer where we=20
> >>> are with the CORBA stuff?
> >>>
> >>> Jacek
> >>>
> >>>
> >
> >
> >
>=20
>=20