From owner-freebsd-java Sat Mar 27 10:32: 3 1999 Delivered-To: freebsd-java@freebsd.org Received: from wrath.cs.utah.edu (wrath.cs.utah.edu [155.99.198.100]) by hub.freebsd.org (Postfix) with ESMTP id 9ABC715248 for ; Sat, 27 Mar 1999 10:32:02 -0800 (PST) (envelope-from gback@cs.utah.edu) Received: from sal.cs.utah.edu (sal.cs.utah.edu [155.99.195.64]) by wrath.cs.utah.edu (8.8.8/8.8.8) with ESMTP id LAA10257; Sat, 27 Mar 1999 11:31:39 -0700 (MST) From: Godmar Back Received: (from gback@localhost) by sal.cs.utah.edu (8.8.8/8.8.8) id LAA19292; Sat, 27 Mar 1999 11:31:37 -0700 (MST) Message-Id: <199903271831.LAA19292@sal.cs.utah.edu> Subject: Re: A BSD-licensed JIT (was Re: Development Projects) To: sprice@hiwaay.net (Steve Price) Date: Sat, 27 Mar 1999 11:31:36 -0700 (MST) Cc: freebsd-java@FreeBSD.ORG In-Reply-To: from "Steve Price" at Mar 27, 99 12:14:02 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-java@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > On Sat, 27 Mar 1999, Godmar Back wrote: > > # I'll bite long enough to say this: not all GPLed work is controlled > # by RMS and the FSF. Please, do not equate the GPL and the FSF. > # For instance, Kaffe is GPLed, but the FSF does not hold the copyright and > # does not control it. Kaffe uses the GPL for entirely different motives > # than the FSF does, see www.kaffe.org/FAQ.html. > > My apologies if I drew that analogy. I am talking about JITs > and both the ones that I know of are GPL'd and copyright the > FSF. It is not the GPL per se that bothers me, but the attribution > to RMS and the FSF. You are entirely correct that Kaffe is not > this way. Thanks for pointing this out. > Interesting. Could you point me at the two FSF Jits you are talking about? I only know of TYA, which is copyrighted by ALbrecht Kleine, and the ShuJIT, which is copyrighted by Shudo Kazajuki. > > # On a technical note, I thin it is impossible to develop a high-performance > # JIT independent of the rest of the JVM. I believe that the jit interface > # as defined by Sun and used by such JITs as TYA and Shujo (sp?) makes > # integrating really high-performance JIT hard, if not impossible. > > Please do tell. > The reasons are the lack of an interface to the threading, exception, and gc subsystem which are necessarily to implement high performance write barriers for incremental, synchronous and asynchronous precise exceptions. - Godmar To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-java" in the body of the message