On Wed, 21 Dec 2005, Tim Jenness wrote:
> On Wed, 21 Dec 2005, Peter W. Draper wrote:
>
> > On Wed, 21 Dec 2005, Mark Taylor wrote:
> >
> >>> Note, I've also started cleaning up the mk script so that it doesn't use
> >>> Fortran to create the shareable library. I'll regenerate the other
> >>> libraries and make changes for them too.
> >>
> >> good, I'd forgotten about that.
> >
> > Done that. All the libraries are now regenerated, including Windows. Guess
> > we'll need to redo those again, after Tim's latest changes, but that can
> > wait. So that's goodbye to any messing about with Fortran for all the Java
> > JNI code.
> >
>
> Just the SPLAT /NDF interface?
Yes, that still relies on Fortran, but I don't consider it important, like
JNIAST and JNIHDS. SPLAT can use JNIHDS for NDFs and does under Windows
and when a webstart application.
> I've been thinking about the "datGetVC" problem over breakfast. How about
> I change the definition of "maxval" to be the declared size of pntrs[]?
> Then I can check that pntrs[] is large enough to hold the null and set
> status bad if it isn't. Then the caller would know there was a problem and
> be able to fix the code without overwriting random memory.
That's better than the pain I've been through. Wonder if Mark is
persuaded?
Peter.
|