[1378] in java-interest
RE: java-interest-digest V1 #138
daemon@ATHENA.MIT.EDU (Robert Stephen Rodgers)
Thu Aug 31 11:44:43 1995
Date: Thu, 31 Aug 1995 08:02:03 -0400 (EDT)
From: Robert Stephen Rodgers <rsrodger@wam.umd.edu>
To: muzo <MuzoK@msn.com>
cc: "'java-interest@java.sun.com'" <java-interest@java.sun.com>
In-Reply-To: <UPMAIL06.199508310337220181@msn.com>
> the resource leak problem under previous versions of windows was caused by OS
> not cleaning up after a program exited. For various moronic reasons, the OS
> doesn't mark all the resources allocated by a process as such and when it
> exists can't GC them.It is not a matter of a single process running
> continuously and leaking resources with no fault of the process.
I'm not sure I follow. Does Java GC specifically clean up
Windows-specific resources? I'm almost 100% sure it doesn't. You can
have a program-internal resource leak that lasts as long as the program
runs. I'm not so much concerned about leaks that exist for the duration
of programs run only briefly, but in terms of programs that the user may
launch at startup and keep open for their entire day. Something that
steadily eats system resources is still a problem.
There are all sorts of examples of Win32 programs doing this. Take a
look at the "Contents" shell add-on from Microsoft, for example.
> Also as HotJava is a 32 process the problem of multiple runs of it leaking
> memory is gone too. Under Win32 (win95 and NT, 2 current implementations) if the
> process doesn't leak any resources,
Well, yes, but what I'm referring to is processes that _do_.
RSR
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
