This is the mail archive of the java@gcc.gnu.org mailing list for the Java project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Calling the garbage collector.



----- Original Message ----- From: "Chris Gray" <mbx03025@scarlet.be>
To: "cr88192" <cr88192@hotmail.com>
Cc: "aph" <aph@redhat.com>; "abhi00" <abhi00@gmail.com>; "java" <java@gcc.gnu.org>
Sent: Monday, July 27, 2009 8:37 AM
Subject: Re: Calling the garbage collector.



-


hmm...

maybe Java needs 'delete'?...

<-- As in "maybe the pope needs a wife"? :-)

Automatic garbage collection means you don't know when garbage will be
collected, it could even be "after the VM exits", i.e. never. Running all
finalizers when the VM exits has been tried, and it was a disaster.

http://www.velocityreviews.com/forums/t145751-finalize-not-guaranteed-to-be-called-ever.html
http://www.hpl.hp.com/personal/Hans_Boehm/popl03/slides.pdf

etc..

PS I'm happy to report that the Mika VM also does not run your finalizer. :-)

-->

I was partly joking, but yeah, GC is has some good points, and manual MM some others.

personally, I prefer a hybrid strategy, where memory can be manually freed/destroyed when it is known that it is not needed anymore, and left for the GC when the usage pattern is "difficult to determine".

in my projects, I use a GC, but much of the time treat it much like good old malloc/free...


so, in cases where it matters, one could essentially force the finalizer to be run and the object to be freed...




Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]