[rescue] Mozilla Firefox

[Charles Shannon Hendrix]

Mon, 26 Apr 2004 @ 14:35 -0400, Dave McGuire said:

> On Apr 26, 2004, at 12:08 PM, Charles Shannon Hendrix wrote:
> >>And, at least we know Java's GC works.
> >
> >Do we?  If it works, then it must be improperly written code in the
> >libraries that seems to make Java programs leak/eat memory.
> >
> >No thanks.
> 
>   Ok, now I'm hearing the "twilight zone" theme again.  I'm sorry man, 
> but I've now reached the conclusion that you don't know squat about 
> Java. :-(

That's your option of course, but I can't imagine why.

The problems of the Java garbage collector are widely discussed and
there has been research on fixing it for years, a lot of that done by
IBM.  The early 1.0 and 1.1 Java GCs were particularly bad, fragmenting
severely and frequently failing to deallocate unreferenced objects.

If I observe a Java program leaking memory, it has to be the libraries
(classes) if it isn't the garbage collector.  Well, it could be the JVM
itself I suppose, but I'd think that is less likely than the GC or the
program code.

If you'll spend 15 minutes with Google, you'll find quite a bit of
information on problems with Java garbage collection, both in how the GC
works, and how people (mis)use it.

I run Java 1.4.1 right now, the Sun distribution for Linux x86.

It's possible the default setup is just bad, but I can't usually run
more than one Java application at a time.  

One thing interesting about 1.4.1 is the introduction of parallel
garbage collectors for SMP systems.  I'm wondering if it would help a
really fast uniprocessor system too.  Might make the application run
smoother since the GC would not have to stop the world when doing a GC
sweep.



-- 
shannon "AT" widomaker.com -- ["It's a damn poor mind that can only think
of one way to spell a word." -- Andrew Jackson]