RE: Thoughts on Linux and Objectivity...

[Mark Pollack <markp@bnl.gov>]

Hi Ed,

I don't' think that we can do that since at the meaning he mentioned that
the Java part is just another API for Objectivity.  It is not a 100% java client.
The hierarchy of the client goes something  like this (from my notes)

API - C++/Java
Object Manager
Cache
Storage Manager (Maybe Type Manager too)
OS (Posix)

The only part in Java is the API part.
When he was refering to the port to linux he was saying that the API/Object Manager
Cache and Storage Manager are rather platform neutral and that the OS level was
would get ported. 

I did make an interface to oracle using a 100% java client and then put that inside
of STAF using the Java Nativie Interface methods, but it might be somewhat more complicated to pass objects between the two languages. I managed to do it with basic 
types (string included in that).

- Regards,
Mark



-----Original Message-----
From:	Edward May [SMTP:enm@sgi10.hep.anl.gov]
Sent:	Wednesday, February 04, 1998 3:18 PM
To:	challenge@sseos.lbl.gov
Subject:	Thoughts on Linux and Objectivity...

The following thought occurred to me with regard to the issue
of running Objectivity on Linux PC for the MDC and initial
testing of the GCA architecture... In an initial implementation
the Linux PC are clients to the Objectivity database. A port of
objectivity would mean having the lock server, AMS and the 
client C++ library on the Linux machine. In an initial implementation
we might not need the lock server and AMS to run on Linux (it is
some where else on the network). Java is a "portable" virtual 
machine, so could we not run the Objectivity Java client software
on the Linux pc's to access the data in the Objectivity servers using
NFS or AMS on remote machines? An interface between Java and Staf would
then need to be constructed. Clearly this would not be high speed
way of moving data from Objectivity to STAF but would allow a complete
test of the architecture  and functionality of many of the components!

Comments and/or criticisms ....
Ed
______________________________________________________________
Edward N. May			Phone:		(630)-252-6222
High Energy Physics Division 	FAX:		(630)-252-5076
Argonne National Laboratory	Internet:	may@anl.gov
Argonne, IL 60439 USA		Webpage:	http://sun2.hep.anl.gov