Re: initialization of final fields

["Jerry Kramskoy" <Jerry dot Kramskoy at insignia dot com>]

Hi Martyn,

I agree that the write*should* be performed just once, but again, the
pragmatic approach taken
by Sun is as I stated to avoid penalising the runtime (interpreter/JIT'd
code) ... so if gcj provides this,
it will be offering something that probably no other Java runtime does.

cheers, Jerry

----- Original Message -----
From: Martyn Pearce <m.pearce@inpharmatica.co.uk>
To: Jerry Kramskoy <Jerry.Kramskoy@insignia.com>
Cc: Martyn Pearce <m.pearce@inpharmatica.co.uk>; Per Bothner
<per@bothner.com>; Godmar Back <gback@cs.utah.edu>;
<java-discuss@sourceware.cygnus.com>; <vgp@dms.at>
Sent: Thursday, November 04, 1999 2:01 PM
Subject: Re: initialization of final fields


>
> | > Pathological example I know, but... I don't think we've covered the
> | > possibility that a given initializer calls such a method more than
> | > once.
> | >
> | > Actually, I don't think it's so patholgical --- the method call might
be
> | > embedded in a loop without thinking.
>
> Jerry Kramskoy writes:
> | Sun's stated semantics for writes to final fields are weaker than your
> | proposal ...
> | the current class (to which the executing method belongs)
> | must be the same as the declaring class for the final field.
>
> I think we're at cross purposes here.  I'm not thinking of which class
> the call is coming from, but how many times:
>
> class MyClass
> {
>   final int widget;
>   int[] bar;
>
>   public
>   MyClass (int count)
>   {
>     bar = new int[count];
>     for (int i = 0; i < count; i++)
>       this.init (i);
>   }
>
>   private void
>   init (int foo)
>   {
>     bar[foo] = 1;
>     widget = foo;
>   }
> }
>
> Which is illegal (not to mention odd), since widget gets assigned to
> more than once, but I think that the scheme proposed by Per doesn't
> catch this.
>
> I hasten to add that I don't suppose it'll occur that often, and that
> Per's scheme is several steps ahead of Sun's --- last I checked, jdk
> javac still can't handle indirect assignments to final fields (i.e.,
> assignments from methods called by the c'tor).  Neither does jikes.
>
> Mx.
>
> PS:  If there's still confusion about which part of the JLS is referred
> to here, the reader is directed to
>
> http://java.sun.com/docs/books/jls/html/1.1Update.html
> D.1.2 New Uses for final
>
> Paragraph 3:
>
> 'You can defer initialization of a final field or variable, as long as
> you initialize it before it is used and assign a value to it exactly
> once. The compiler will check for proper assignment, as will the
> verifier before code is executed.'
>