Re: Kava for GCJ?

[Per Bothner <per at bothner dot com>]

Adam Megacz wrote:

> Has anybody heard of a Kava implementation for GCJ?
>
>   http://citeseer.nj.nec.com/bacon02kava.html
>
> Kava basically brings C# structs (unboxed complex types) to Java, but
> does a much better job.  You can even define unboxed scalars that are
> constrained to be less than the length of a given array, thereby
> safely elimnating bounds checks.
>
> Hrm, I wonder if this would be hard to implement... might be worth my
> time.

The problem is that this changes Java in incompatible ways.  I prefer
a solution where you *annotate* a class that can only be uses in
certain ways, and an implementation is *encouraged* to use unboxed
types, but not required to:

public final class Point
{
  // The magic annotation:
  static final boolean gnu$gcj$VALUE_CLASS = true;

  // Fields are final.
  public final float x, y;

  // No public constructors:
  private Point(float x, float y) { this.x = x;  this.y = y; }

  public static Point make (float x, float y)
  { Point p = new Point(x, y);
    // Conceptually:  p = intern(p);
    return p;
  }
  ...
}

I wrote up a more detailed proposal, but it needs some editing,
and I'm leaving on vacation tomorrow morning.  Perhaps you could
remind me to post it if I haven't done so by Thursday next week?

If we come up with and implement a design, we should perhaps
propose it as a JSR.  (I've suggested something similar to some
Sun people last year, and they agree with the concept.)
--
	--Per Bothner
per@bothner.com   http://per.bothner.com/