Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!mailrus!umix!hyc
From: hyc@umix.cc.umich.edu (Howard Chu)
Newsgroups: comp.sources.d
Subject: Re: bsd version of arc
Message-ID: <4008@umix.cc.umich.edu>
Date: 15 Apr 88 07:08:54 GMT
References: <4395@b-tech.UUCP> <28083.1988Apr6.21:19:52@mill.me.uucp> <4358@hoptoad.uucp>
Reply-To: hyc@umix.cc.umich.edu (Howard Chu)
Organization: University of Michigan Computing Center, Ann Arbor
Lines: 77
Summary: I've been away from Usenet for awhile, sorry...
UUCP-Path: {uunet,rutgers}!umix!hyc


In article <4358@hoptoad.uucp> gnu@hoptoad.uucp (John Gilmore) writes:
>I have a version of "arc" with numerous bugs removed.  It seems to
>create, extract, list, and test files in all archives I have run across
>(except damaged archives, which it complains about now rather than core
>dumping on).  Arc files it has created have been moved to msdos without
>trouble, and vice verse.  It's still really slow and stupid, but I
>didn't fix design problems, just bugs.

I have just posted a version to alt.sources. This program had been tested
only on BSD machines at the time I posted it. I have since sent a copy to
Jon Zeeff (b-tech!zeeff) who tells me he was able to get it running on his
Sys V system, with the help of a BSD compatibility library previously posted
to comp.sources.unix. I'll be getting things together for an "official"
release to comp.sources.misc within a day or two.

The code I just released is based on the PC sources for ARC 5.20, which
is significantly faster than all the previous Unix ports released so far.
(The previous releases were based on version 5.12. The primary difference
is that ARC 5.20 will no longer use Huffman Squeezing to compress files,
and saves a great deal of time by not going thru those calculations.) The
code implements every PC option correctly. It also munges text files so that
CRs are stripped for use on Unix boxes, and re-inserted when compressing
files, such that they will look correct on MSDOS systems upon extraction,
and vice versa. The Squashing code written by Dan Lanciani for version 5.12
has been merged in. I'm quite confident that this is the end-all be-all of
ARC implementations for Unix. The "official" release is waiting for me to
finish wading thru the remaining unresolved lint output. I have already
received notes from people who received the alt.sources posting - they
literally just unshar'ed the posting and typed "make" and it worked.
There's very little left to do before I will send this to Brandon Allbery,
who will hopefully have it out on comp.sources.misc Very Soon.
>
>Many months ago, I sent my version to somebody at umix who claimed to
>be willing to maintain a master "Unix ARC" source.  I no longer have a
>record of who that was, and they don't seem to have distributed it or
>answered any of these net queries.

As the summary said, I haven't been paying much attention to the net
these days. Sorry for the negligence on my part. I was able to use a
little of what you sent me, but basically started over again when I
got hold of the 5.20 sources.
>
>I am *not* willing to send out copies to everybody who asks.  I am *not*
>willing to act as a clearing house for arc bugfixes.  I *am* willing to send
>my version to a person who *is* willing to do the above two things.

I will act as the central clearinghouse. I don't expect much in the way
of bug reports or fixes, because I've done a lot of beating on this code
in the months that have gone by. There's bound to be a bug in there somewhere,
but the ones that I've encountered are no longer present.
>
>In other words, I don't mind sharing my bug fixes, I just have no
>interest being stuck maintaining the damn thing, or having to handle a
>lot of email about it.  How about a serious volunteer responding by
>posting in comp.sources.d?  (Then everybody can email their versions to
>this person, who had better be willing to merge a lot of bugfixes from
>a lot of wierd versions of arc.)
>-- 
>{pyramid,pacbell,amdahl,sun,ihnp4}!hoptoad!gnu			  gnu@toad.com
>  I forsee a day when there are two kinds of C compilers: standard ones and 
>  useful ones ... just like Pascal and Fortran.  Are we making progress yet?
>	-- ASC:GUTHERY%slb-test.csnet

While I'm eager to field bug reports and hear about bug fixes, I'd prefer
that people held off just a bit longer until I get my current release out
the door. I suspect that you'll want to discard anything you've been using
up to this point after you start using this.

Again, sorry for the delay in responding. [I was switching jobs, in a sense,
which is why I wasn't keeping in touch with the world. I still have a small
fraction here at the CC, but I've been busy enough moving around that only
now have things settled down reasonably.]

  /
 /_ , ,_.                      Howard Chu
/ /(_/(__                University of Michigan
    /           Computing Center          College of LS&A
   '              Unix Project          Information Systems