[3052] in java-interest
Finalize()
daemon@ATHENA.MIT.EDU (Richard Williamson)
Fri Oct 27 19:36:55 1995
From: Richard Williamson <richwill@infoscape.com>
Date: Fri, 27 Oct 95 15:04:50 -0700
To: java-interest@java.sun.com
This comment in the Java documentation was pointed out to me:
>When a user defines the void finalize() method in a class
>definition, finalization is enabled for objects of that class.
>Finalization of an object consists of the system calling the
>object's finalize() method. Finalization normally occurs
>asynchronously at some time after the garbage collection mechanism
>identifies an object as inaccessible. Users can invoke their
>finalize() method explicitly but this doesn't guarantee that the
>system will not call it again at a later time. If a finalized
object >references another finalized object, the objects are
finalized in *>the reverse order of their creation. Java does not
guarantee
*>when or if a given finalized object will have its finalize()
method *>called. Thus, finalization should not be relied on for
program *>correctness. Rather, finalization should be thought of as
an *>optimization.
The final few sentences is this footnote are somewhat alarming.
How can finalize() be used for ANYTHING if it isn't reliably called?
- Richard Williamson
richwill@infoscape.com
-
This message was sent to the java-interest mailing list
Info: send 'help' to java-interest-request@java.sun.com
