tim at tjwassoc.co.uk
Thu May 22 15:04:24 PDT 1997
tullmann at facility.cs.utah.edu wrote:
> > > > > Can't this be "fixed" by subclassing them from a common, private
> > > > > subclass?
> > > >
> > > > Possibly. I'll think about that.
> > Now I have thought about it, the superclass would have to be public if it's
> > to be referenced from both java.net and java.io.
> > I was originally planning to put NativeIO in java.io, etc., but it
> > would have to be public so that java.net can reference it.
> Perhaps if each of the packages had its own native interface class,
> java.io will have NativeIO, java.net will have NativeNet, etc. *and*
> duplicate interfaces as necessary. So, both NativeIO and NativeNet
> will have fd_close() functions. Its then up to the implementor of the
> native library to make sure these actually do exactly the same thing.
> Since ther'es no state associated with the Native* classes, they're
> just groupings for the native interfaces, it shouldn't be too hard.
> Some care will have to be taken on the part of the native lib writer,
> but this should solve the security problems (the Native classes can be
> package private), and should prevent adding hacks to Kaffe to
> artifically hide certain class names.
> It is a bit of an ugly hack, though.
To stick my two penneth in, I'm not very much in favour of hacking the
internals of Kaffe around to hide classes and solve unnecssary security
problems - there leadeth the road to ruin me thinks.
Tim Wilkinson Tel/Fax: +44 181 440 0658
T. J. Wilkinson & Associates, Mobile: +44 370 621006
London, UK. Email: tim at tjwassoc.co.uk
More information about the kaffe