From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII
X-Google-Thread: 103376,aef01dc1d0a3a8bd
X-Google-Attributes: gid103376,public
From: Geoff Bull <geoff@research.canon.com.au>
Subject: Re: Dummy
Date: 2000/02/10
Message-ID: <38A212E8.C5552776@research.canon.com.au>#1/1
X-Deja-AN: 583922384
Content-Transfer-Encoding: 8bit
References: 
 <387b154a.3533365@newsread.albacom.net><3898C380.BC01EC03@earthlink.net>
 <t7d7qe2xab.fsf@calumny.jyacc.com><3899DB45.B481F8E2@averstar.com><87bt5wlt8l.fsf@deneb.cygnus.argh.org>
 <wcc66w4rcy5.fsf@world.std.com> <87jghc$oj5$3@wanadoo.fr>
X-Accept-Language: en
Content-Type: text/plain; charset=iso-8859-1
X-Complaints-To: usenet@research.canon.com.au
X-Trace: cass.research.canon.com.au 950145741 18037 203.12.174.227 (10 Feb
 2000 01:22:21 GMT)
Organization: Canon Information Systems Research Australia
Mime-Version: 1.0
NNTP-Posting-Date: 10 Feb 2000 01:22:21 GMT
Newsgroups: comp.lang.ada
Date: 2000-02-10T01:22:21+00:00
List-Id: <comp.lang.ada>

Jean-Pierre Rosen wrote:
> 
> Robert A Duff <bobduff@world.std.com> a �crit dans le message :
> wcc66w4rcy5.fsf@world.std.com...
> > Florian Weimer <someone@deneb.cygnus.argh.org> writes:
> >
> > > Does this mean that that the finalization of local variables is
> > > non-deterministic in Java?  Ugh.
> >
> > Yes.  Finalization happens in Java whenever the garbage collector gets
> > around to noticing that it's needed.
> >
> It's even worse. The finalization routine is allowed to resurect the object
> (for example, by assigning it to some live variable), however it is
> guaranteed that in that case, finalization will *not* be invoked again, when
> the object becomes inaccessible for the second time.

It's even worse, although finalizers in Java are guaranteed to be never
invoked more than once, the are not guaranteed to be invoked at least
once.
This makes finalizers essentially useless for cleaning up resources etc.
It used to be possible to ask the VM to run finalizers at exit, but that
method has been deprecated.