Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site decvax.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!bellcore!decvax!ccb
From: ccb@decvax.UUCP (Charles C. Bennett)
Newsgroups: net.unix,net.decus
Subject: Re: emulating ".ascii" with Awk
Message-ID: <168@decvax.UUCP>
Date: Thu, 23-Jan-86 19:54:05 EST
Article-I.D.: decvax.168
Posted: Thu Jan 23 19:54:05 1986
Date-Received: Fri, 24-Jan-86 22:21:37 EST
References: <bonnie.684>
Reply-To: ccb@decvax.UUCP (Charles C. Bennett)
Organization: DEC Ultrix Engineering Group - Merrimack, NH
Lines: 25
Xref: watmath net.unix:6912 net.decus:248
Summary: 


Tricky, I don't think awk will do it.
Write a little shell script.

use "od -b" to break the string into bytes in octal
use "sed 's/^.......[ ]*//' to strip addresses from the od output
use tr to convert the spaces in the od output to newlines
prepend the string "ibase=8" to the stuff you have so far
feed the works to the stdin of "bc"
use tr to convert the newlines to commas
use sed to strip the telltale trailing comma

something like:

bc $TMPFILE <<!
ibase=8
`echo $STRING|od -b|sed 's/^.......[ ]*//'|tr ' ' '\012'`
!
echo `cat $TMPFILE|tr '\012' ','`|sed 's/,$//'

the final pipeline leaving the result on stdout.

A 'C' program to do this is trivial.

decvax!ccb