[Home] [Groups] - Message: [Prev in Group] [Next in Group]

nu.kanga.list.mud-dev

28017: Re: [MUD-Dev] [Tech] Garbage collection

[Full Header] [Plain Text]
From: Bruce Mitchener <bruce@cubik.org>
Newsgroups: nu.kanga.list.mud-dev
Date: Wed, 03 Sep 2003 01:14:52 -0600
References: [1]
Organization: Kanga.Nu
Brian Hook wrote:

> I've been doing some more research into garbage collection lately
> and I'm trying to get a feel for what the current major MUD
> languages/engines use for garbage collection.

> My gut instinct is that a vanilla mark-and-sweep probably isn't
> acceptable because of hitches, so either most are using explicit
> memory management (doubtful), ref-counting (likely), or an
> incremental collector (complicated).

In Cold, we mainly used ref-counting.  It works pretty well,
especially with a test suite for most of the language and the
ability to run it under Purify and/or Valgrind.

There are some licensable garbage collectors as well.  One nice
looking one has a viral open source license, but can commercially
licensed as well:

   http://www.ravenbrook.com/project/mps/

Since I know that you've been looking at TOM
(http://www.gerbil.org/tom/) some, it has an incremental garbage
collector in its runtime that is pretty nice.

I've been looking at Chicken Scheme
(http://www.call-with-current-continuation.org/chicken.html) some
off and on, and it has a generational collector .. but it uses the C
stack as its nursery.

I think the Boehm GC can do incremental collection.

A tool that is pretty nice when working with a garbage collected
system is GCspy:

   http://www.experimentalstuff.com/Technologies/GCspy/

  - Bruce
_______________________________________________
MUD-Dev mailing list
MUD-Dev@kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev