On 24 Jul 1998, Paul Fisher wrote:
> "Aaron M. Renn" <[EMAIL PROTECTED]> writes:
> > That's not really our target anyway, and Kaffe seems to have it
> > covered to why worry?
>
> I looked at Kaffe's java.lang.Character a few minutes ago -- I
> laughed. :) It works great for the extremely small Basic Latin subset
> of Unicode (0x000 to 0x007F), and that's it. I assume that's why
> it's beta software.
Well, if your target is embedded machines (which TransVirtual seems to be
interested in), you might not want much more. Sun's java.lang.Character
has like 60k of byte code just in the class initializer. That's not too
good if you're trying to fit your whole VM+classes into a couple hundred
K.
