From owner-freebsd-announce  Tue Feb 13 16:07:25 1996
Return-Path: owner-freebsd-announce
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id QAA19142
          for freebsd-announce-outgoing; Tue, 13 Feb 1996 16:07:25 -0800 (PST)
Date: Tue, 13 Feb 1996 16:07:25 -0800 (PST)
From: owner-freebsd-announce
Message-Id: <199602140007.QAA19142@freefall.freebsd.org>
Sender: owner-freebsd-announce@FreeBSD.ORG
Precedence: bulk
To: undisclosed-recipients:;

	>From owner-majordomo  Tue Feb 13 13:47:33 1996
	Received: (from hsu@localhost)
	          by freefall.freebsd.org (8.7.3/8.7.3) id NAA10512
	          for freebsd-announce; Tue, 13 Feb 1996 13:47:33 -0800 (PST)
	Date: Tue, 13 Feb 1996 13:47:33 -0800 (PST)
	From: Jeffrey Hsu <hsu>
	Message-Id: <199602132147.NAA10512@freefall.freebsd.org>
	To: freebsd-announce
	Subject: JDK 1.0 port to FreeBSD

	Announce
	--------
	I have ports of the JDK 1.0 using Green threads, pthreads, and uthreads.
	This port I'm releasing uses Green threads and is robust enough to
	recompile itself correctly.

	Installation
	------------
	1. Grab ftp.javasoft.com:JDK-1_0-solaris2-sparc.tar.Z.
	2. Grab freefall.FreeBSD.org:/pub/FreeBSD/LOCAL_PORTS/jdk10.tar.gz.
	3. Untar the solaris binary distribution.
	4. Move aside bin/ and untar the FreeBSD distribution.
	5. Install the pdksh port.

	Running the JDK
	---------------
	If your JDK directory is rooted at <jdk>, you need to set
	  1. your path to include the <jdk>/bin directory
	  2. CLASSPATH environment variable to .:<jdk>/lib/classes.zip
	  3. LD_LIBRARY_PATH to include <jdk>/lib/i386
	Then to compile a java program, invoke 'javac program.java'.
	To run it, use 'java <name of main class>'.

	Status
	------
	This port is kind of flacky and I hesitate to release it, but it
	does meet my long-stated criteria of being self-hosting.  The
	problems are related to the kludgy work-around I used to
	get around not having weak symbols and dlsym(RTLD_NEXT).  The
	pthreads port does not have this problem, so the next release may
	use that instead, but I'm leaning towards fixing up the underlying
	operating system rather than working around it.  The next release
	will come when I get appletviewer to work.

	How you can help
	----------------
	We could use a full-featured dlsym() which understands RTLD_NEXT.
	Weak symbols would help a lot too.  There are people already working
	on the linker, so please coordinate with them if you decide to go
	work on this.