[1334] in java-interest
Re: Byte Codes
daemon@ATHENA.MIT.EDU (Smiljan Grmek - KSI ZGB)
Wed Aug 30 13:52:50 1995
Date: Wed, 30 Aug 1995 16:20:10 +0200 (MET DST)
From: Smiljan Grmek - KSI ZGB <sgrmek@public.srce.hr>
To: Philip R Brenan <phil@bga.com>
cc: java-interest <java-interest@java.sun.com>
In-Reply-To: <Pine.3.89.9508291239.A18349-0100000@urchin.bga.com>
On Tue, 29 Aug 1995, Philip R Brenan wrote:
> I was thinking more along the lines of a class for doing byte-coding. It
> would have methods for inserting each byte-code into the current output
> class etc.
>
Of course. This is the way along which I considered a more efficient way
to do dynamic compilation - one would give as parameters a class name and
a method definition which would then be compiled and included in the
definition of the loaded class. The only available environment would be
the classes already loaded. This would give us all sorts of nice features
like iterators, selectors &c.
Alas, no - the price is probably too high. It would include invoking
methods by their symbolic names (bringing Java extremely close to Smalltalk)
and breaching the namespace separation of local and remote code security
feature. It would also require addition to the bytecode repertoire (because
Java programs can modify *only* Java variables).
Smi
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
