Re: UTF-8 problems with windows

From:

"Mike Schilling" <mscottschilling@hotmail.com>

Newsgroups:

comp.lang.java.programmer

Date:

Thu, 20 Aug 2009 23:02:47 -0700

Message-ID:

<h6lda9$i34$1@news.eternal-september.org>

Lew wrote:

Roedy Green wrote:

On Fri, 14 Aug 2009 08:51:13 -0400, Lew <noone@lewscanon.com>
wrote,
quoted or indirectly quoted someone who said :

Java references are addresses into memory - they point to objects
to which they refer. They aren't exposed to the programmer as
*machine* addresses, but then Java is not a machine language.

In the original JVMs, references were handles, pointers to
pointers.
That made it easier to shuffle objects around in RAM.

What they are in the JVM is independent of the fact that in the Java
language references function as memory addresses, at a high level.
Due to the restricted nature of references in Java - you can't
assign
arbitrary pointer values such as mid-object addresses, for example -
this high-level abstraction of an address is sufficient.

I think I could make (address,offset) work for lots of different
address implementations, including handles. (You can do it more or
less even in existing Java with a class that holds a reference and a
java.lang.reflect.Field.) What Java does that's really important is
make references completely opaque.

Does a reference fit in an int? None of your beeswax.
What bits are in a null reference? None of your beeswax.
What's the offset of the length field in a String? None of your
beeswax.

Beautiful stuff. One of the big mistakes Java made was to tell us
that chars are 16 bits. Otherwise, a char could hold every Unicode
character and we wouldn't need surrogates.