[[email protected]: Re: GC]

------- Start of forwarded message -------
X-Authentication-Warning: gerd.dnsq.org: gerd set sender to [email protected] using -f
Sender: [email protected]
To: [email protected]
Subject: Re: GC
Reply-To: [email protected]
From: [email protected] (Gerd Moellmann)
Date: 26 Apr 2002 20:39:22 +0200
In-Reply-To: <[email protected]>
X-Sender: [email protected]

Richard Stallman <[email protected]> writes:

> Could you tell me what igc is?

I had some code for a simple non-generational and non-incremental
mostly-copying collector lying around that I wrote in the early 90's
for use in software project of mine, after having read the first paper
about Bartlett's mostly-copying algorithm.  The code I wrote was never
used, and I forgot about it until 2001, when I found it on a floppy
disk again.

I beefed the code up a bit using today's techniques, and made the
collector generational and incremenal using VM page protection, with
the long term goal of maybe using it in Emacs.  The collector is
currently a standalone test program.  More testing would have been
needed, and some Emacs-specific optimizations and changes would have
been necessary before using it in Emacs, but the difficult parts were
all done.

(When I thought Stefan Monnier were ready to take over Emacs
maintenance, I sent him a copy of the code.  Andrew Innes also has a
copy because he expressed interest in the techniques I used in the
collector.)

> This is another example of free software attacked by a patent;

Very true.  And, in the case of igc, hadn't Paul Wilson mentioned the
patent on a mailing list which someone else read who knows I'm
interested in GC algorithms, and hadn't she mentioned it to me, I
wouldn't even have noticed.
------- End of forwarded message -------