Path: utzoo!attcan!uunet!lll-winken!ames!mailrus!tut.cis.ohio-state.edu!bloom-beacon!chuck-jones.ila-west.dialnet.symbolics.COM!York
From: York@chuck-jones.ila-west.dialnet.symbolics.COM (William M. York)
Newsgroups: comp.windows.x
Subject: Re: ILA's X11 on Symbolics
Message-ID: <19890321232818.5.YORK@Chuck-Jones.ILA-West.Dialnet.Symbolics.COM>
Date: 21 Mar 89 23:28:00 GMT
References: <8903211225.AA02545@EXPIRE.LCS.MIT.EDU>
Sender: daemon@bloom-beacon.MIT.EDU
Organization: The Internet
Lines: 32


    Date: Tue, 21 Mar 89 07:25:22 EST
    From: rws@expo.lcs.mit.edu (Bob Scheifler)

	Has anyone tried out the X server for Symbolics' from ILA?
	I played around with an early version at my Symbolics rep's
	and was disappointed. It seemed pretty buggy.

    Yes, I bought it and have tried it, and I would probably have been less
    charitable than you.  It's pretty hard to believe they call it a product.  E.g.
    causing an xterm to scroll by typing carriage return will crash the server.

The important component of the Symbolics X product is the "remote
screen" client program.  The preliminary release of the X Server (a
direct port of the "sample server" to Symbolics C) was included to fill
a gap in the customers' set of available tools.  It was intended to
assist in the debugging of CLX programs being developed on the Lisp
Machine, not as a serious X Server for Genera, and was documented as
such.  We have gotten substantial QA feedback, and most of the known
bugs are fixed and the fixes will be distributed on the next Symbolics
software ECO tape, which is currently in preparation.

Most of the blowouts are due to the Symbolics C compiler being stricter
than those used in the Unix world.  There are several places where
uninitialized storage is referenced (e.g. the event structure
byte-swapping code), or functions are called with too many arguments
(but via dispatch vectors which defeat compile-time checking) and
Symbolics C generates errors in these cases.

Overall, when you include the bugfixes, I think that the server does
quite well for a C program that has been uprooted and transplanted into
a Lisp environment!