On Tue, 23 Apr 2002, Campbell, Garrett wrote:
> Great, a work around.
Excellent. You might want to read in your data objects and re-save them
in a non-version-dependent manner....
p.s. thanks for clarifying the 'save(filename, data...)'
> the older visad.jar at least lets me execute the program as it was
> running yesterday.
> I have no idea why I get the hotspot crash, maybe jre1.3.1 is needed to
> match up with
> visad.jar (april 22).
> Dan Murray suggested the VisADForm.save and it was sure easier that
> trying to figure
> out how to save a area file with all the nav info and then read it back.
> Great I can even read in the old objects.
> What ever changes were made in between your install file and the current
> sure had a big effect.
> this was the save code:
> # select= select.append(1.),select.append(2.)..
> where data=load(adde:pointdata..)
> Someone must have changed the field object.
> If it is not too hard, maybe you should save each version of the install
> for dumb users like me.
> -----Original Message-----
> From: Tom Whittaker [mailto:tomw@xxxxxxxxxxxxx]
> Sent: Tuesday, April 23, 2002 3:48 PM
> To: Campbell, Garrett
> Subject: RE: older verions of visad/jython
> > I even tried java 1.4 but that had yet more problems with java 3d
> > packages.
> There have been previous postings about this. The Java3D 1.3x(beta) or
> something is compatible with Java1.4, previous versions are not.
> I've been looking at your posting to the list. You said:
> > > can not be read in again. (I had an array of floats fselect=
> > > fselect.append(1.)..
> > > which I saved:
> > > VisADForm().save(filename,fselect,1)
> Did you create fslelect as a list in Jython? (Which is really a PyList
> object, probably containing PyFloat objects...)
> I'm only asking because if I take your steps above literally, I cannot
> believe it wrote anything, because the signature for save() takes a
> VisAD Data object as the second parameter, not a Jython "PyList" object.
> Anyway, I hope you can get this squared away. I agree with your other
> statement, though -- people who want long-term persistence shouls use
> the saveBinary. I had so many problems a couple of years ago with
> serializing Java Objects to disk (naively thinking it was a good way to
> not have to decode files every time) that I gave up.
> Tom Whittaker
> University of Wisconsin-Madison
> Space Science and Eng. Center
> ph: 608.262.2759
University of Wisconsin-Madison
Space Science and Eng. Center