--- Timothy Wall <email@hidden> wrote:
>
> On Jul 29, 2007, at 8:52 AM, Michael Hall wrote:
> >
> > It would seem to me that a dynamic load should
> search
> > on a symbol and not be tied to a specific absolute
> > path library but I could be wrong. Not that
> familiar
> > with that stuff.
>
> The jnilib is linked against JAWT. The library and
> symbols *could*
> be looked up dynamically via dlopen/dlsym, but it's
> a much bigger
> pain that way.
It would actually be dynamic that way I'd think but
again this is not really the sort of platform where
I've spent most of my time. IBM mainframe you have a
system defined search path for dynamic loads, some
standard libs, some can be varied by the system
programmers, and you find it in those or don't. There
is no reference to any kind of file for the dynamic
load even the ones that are always there.
> >
/Users/twall/Development/ws/jna/build/native/libjnidispatch.jnilib
> That shouldn't be there. Where exactly did it show
> up?
>
The JWS ppc darwin download, extracted and issued the
command.
> > It might work if it's not a symlink (which it
> usually is); you'd want
> the path to be the current JRE, though, not the
> current JDK (unless
> the JDK is always installed by default, which I
> doubt).
>
For OS X the JDK is always installed and I think for
OS X this symlink is always there too. But it is a
symlink so if thats the problem it won't help. Between
the two symlinks there might be differences depending
on whether or not you start tweaking the preferred
version using the java preferences application. I
don't know. But either way I think, a symlink.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Java-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/java-dev/email@hidden
This email sent to email@hidden
