Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!zaphod.mps.ohio-state.edu!math.lsa.umich.edu!math.lsa.umich.edu!hyc
From: hyc@math.lsa.umich.edu (Howard Chu)
Newsgroups: comp.os.minix
Subject: Re: Source Posting Proposal, what about packaging...
Message-ID: <1990Aug4.122641.1868@math.lsa.umich.edu>
Date: 4 Aug 90 12:26:41 GMT
References: <5389@plains.UUCP> <1281@sirius.ucs.adelaide.edu.au>
Sender: usenet@math.lsa.umich.edu
Organization: University of Michigan Math Dept., Ann Arbor
Lines: 47

In article <1281@sirius.ucs.adelaide.edu.au> cagney@chook.ua.oz (Andrew Cagney - aka Noid) writes:
>To me it would be nice to be able to take a posting and enter
>   |magic-command
>and have it unpack it exactly where you want it.

I've got a script called "rnget" that worked pretty well for shar files.
   |rnget <dirname>
No sweat, and it left the header/descriptive text in a READ_ME file.

>Magic-command could be something like:
>   |modified-uud | tee posting.tar.Z | uncompress | ( cd / ; tar -xf - )
>   Extracting /etc/passwd
>   Extracting /etc/groups
>:-)

Actually all you need is "cat part*" in front of the pipeline there...  }-)

>	1. uuencoding is needed. And for more than just bitnet users

atob/btoa is more efficient, why doesn't anyone ever use it?

>	2. Compress -b13 is probably a good idea

Yeah, doesn't seem too bad. (No great loss not getting 16 bits.)

>	3. during the 1.5 upgrade, shar was the cause of several of the
>	   posting problems.

Yeah. Could have been avoided by modifying the shar file, to include the
uuencoded binary and a uudecode command, instead of the simple catenation...

>	4. How portable is tar?

I put together my ST upgrade kit on a NeXT machine, tar'ing files back
and forth with my ST. No compatibility problems there. I regularly swap
tar files across Suns, Vaxes, DECstations, etc. I've never encountered
byte-order problems or any other hassles. 

I think encoded compressed tar files are probably the best way to
distribute major source postings. Even though most bitnet sites aren't
likely to have tar on-site, the destination machine (a Unix or Minix
system) will have all the necessary unpacking tools.
--
  -- Howard Chu @ University of Michigan
  one million data bits stored on a chip, one million bits per chip
	if one of those data bits happens to flip,
		one million data bits stored on the chip...