I thought that most of the GC issues were addressed by the generational GC
in the Sparc-Solaris edition of JDK1.2.
The current GC is great for GUI/event-driven apps where it sits idle for
most of the time but completely inadequate for any sort of real-time stuff.
Question: When will the generation GC be available on other platforms?
Someone mentioned that Java3D on Linux was pause-free, does that mean it
uses the new GC?
I think that having an api for the GC would be useful for those real-time
apps but not necessarily having fine-grained control over object garbage
collection
Grant
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Thomas J
> Lukasik
> Sent: Sunday, 7 March 1999 13:53
> To: Brandie Lynn; Alexander Semenov
> Cc: [EMAIL PROTECTED]
> Subject: Re: [java3d] pausing problem
>
>
> I have been thinking lately that perhaps Sun should make Garbage
> Collection
> **optional**, e.g. I may want to take full responsibility for recovering
> memory (in which case I would need a simple mechanism for manually
> destroying individual objects, or else 'marking' them for
> destruction), or I
> may only want automatic Garbage Collection to recover resources
> if and when
> I *forget* to clean up after myself, or maybe I want the Garbage Collector
> to be responsible for 'cleaning up after me' all the time (as it
> works now).
> Of course, there would need to be some settable option to tell
> the VM which
> behavior I want.
>
> This seems like a simple (and workable) approach to recovering memory; let
> me be responsible for myself if I want to, rather than take that
> choice away
> in all cases. Wouldn't such a simple delete mechanism and an 'on/off
> switch' for GC be easy to implement? Am I missing something?
>
> Regards,
> sparky
>
>
> -----Original Message-----
> From: Brandie Lynn <[EMAIL PROTECTED]>
> To: Alexander Semenov <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
> Date: Saturday, March 06, 1999 9:17 PM
> Subject: RE: [java3d] pausing problem
>
>
> >Alexander et al,
> >
> >Thanks for the gc reference. And, it raises a question
> >after thinking over Doug's reply. What are the
> >functional (and philosophical) implications of
> >trying to elimiate GC? How do you do it?
> >It is after all, one of the touted benefits of Java,
> >efficient GC.
> >
> >I'd like to see what Alexander has done and have to
> >admit I've contemplated getting deeper into GC
> >myself. The problems it seems to address were
> >always more in evidence in C++ than in C, at
> >least with the code I worked with. It seemed
> >like the less ccntrol I had over what was happening
> >the more likely there would be problems.
> >
> >Just thoughts...
> >
> >Brandie
> >
> >=====================================================================
> >To subscribe/unsubscribe, send mail to [EMAIL PROTECTED]
> >Java 3D Home Page: http://java.sun.com/products/java-media/3D/
>
> =====================================================================
> To subscribe/unsubscribe, send mail to [EMAIL PROTECTED]
> Java 3D Home Page: http://java.sun.com/products/java-media/3D/
>
=====================================================================
To subscribe/unsubscribe, send mail to [EMAIL PROTECTED]
Java 3D Home Page: http://java.sun.com/products/java-media/3D/
