Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!lll-winken!uunet!crdgw1!steinmetz!davidsen
From: davidsen@steinmetz.ge.com (Wm. E. Davidsen Jr)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: lzhuf is much better than compress -b16
Message-ID: <13591@steinmetz.ge.com>
Date: 12 Apr 89 19:15:40 GMT
References: <56621@yale-celray.yale.UUCP> <56625@yale-celray.yale.UUCP>
Reply-To: davidsen@crdos1.UUCP (bill davidsen)
Distribution: usa
Organization: General Electric CRD, Schenectady, NY
Lines: 26

In article <56625@yale-celray.yale.UUCP> spolsky-joel@CS.YALE.EDU (Joel Spolsky) writes:

| If anybody wants a copy of the source code suitable for SunOS 4.0 (and
| probably many other unices) send me eMail. The only major change is
| replacing ints with shorts. 

  I'm sending a request by mail, but I'll make a few comments here.
First, converting int to short will only put off the problem, since many
machines don't have 16 bit short. There are some logic errors which
assume "silent anding" to 16 bits, and I admit I haven't found them all.

  Also, I have a version which works on over 99% of the files I've
tried. It hangs on at least two, which work with the 16 bit version.
Again I suspect a trunctation assumption.

  I also have another (just one) file which works on 386 and VAX but not
on Sun or Convex. I believe it's related to using LSB byte order. I hope
we can have a single version which will run on everything from Z80 to
Cray. The original author has done a good job, let's try to get a really
fixed version before propigating it.

  I have sent a test file to Rahul to help validate the new zoo versions.
-- 
	bill davidsen		(wedu@crd.GE.COM)
  {uunet | philabs}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me