[kepler-dev] [Bug 1335] - finish actor prototype tool

Stephen Andrew Neuendorffer neuendor at eecs.berkeley.edu
Mon Jun 14 15:00:35 PDT 2004


The key thing is that the base classes act as interfaces... they must be 
accessible from the
system classpath.  Derived classes for individual atomic actors must be in 
a separate class
path that is not accessible from the system classpath.  Then it is 
relatively easy to reload
the atomic actors (but not the base classes).

The real problem is not the above.  (which is easy to do: just pass a 
JarClassLoader to the
MoMLParser in the constructor)   The real problem is a good integration 
with the UI for it...
Since this only works for newly constructed objects...  Essentially you 
have to force save
and then throw away all the existing models, and then reload them, which
you may or may not want to do.  This is probably not something you would 
want to
do automatically.  You also have to reload the actor library, which is 
currently only loaded once when vergil is
started and is loaded indirectly by the configuration...  So, the only 
thing that a "reload classes"
button would give you is, perhaps, starting vergil over in the same state, 
as opposed to the initial state.
Is this really worth the work?

Steve

At 02:33 PM 6/14/2004, bugzilla-daemon at ecoinformatics.org wrote:
>http://bugzilla.ecoinformatics.org/show_bug.cgi?id=1335
>
>
>
>
>
>------- Additional Comments From berkley at nceas.ucsb.edu  2004-06-14 14:33 
>-------
>I have looked extensively into trying to recompile and reload a class in java
>and have been unable to figure out how to do it within the framework of 
>kepler.
>  Though there are several schemes to reload classes in java, they all rely on
>not including classes in the main classpath and instead loading interfaces,
>which won't work in kepler because the name of the actor class is hard 
>coded in
>the moml.  I'm pretty much stuck on this right now and I'm not sure where 
>to go
>with it from here.
>_______________________________________________
>kepler-dev mailing list
>kepler-dev at ecoinformatics.org
>http://www.ecoinformatics.org/mailman/listinfo/kepler-dev





More information about the Kepler-dev mailing list