Path: utzoo!attcan!uunet!lll-winken!elroy.jpl.nasa.gov!jpl-devvax!lwall
From: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Newsgroups: comp.lang.perl
Subject: Re: Behavior of a2p
Message-ID: <8587@jpl-devvax.JPL.NASA.GOV>
Date: 3 Jul 90 21:41:23 GMT
References: <182@sun13.scri.fsu.edu>
Reply-To: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 45

In article <182@sun13.scri.fsu.edu> nall@sun8.scri.fsu.edu (John Nall) writes:
: In converting a few awk scripts using a2p, it seems that the
: behavior of a2p is not exactly what I think it should be.
: (This is pretty nit-picky, I realize...:-)  )
: 
: As an example:  Consider the following awk script:
: 
: 	{print NF}
: 
: a2p would convert this into the following (leaving out some of
: the header material):
: 
: 	$[ = 1;
: 	$, = ' ';
: 	$\ = "n";
: 
: 	while (<>) {
: 		$Fld = split(' ', $_, 999);
: 		print $#Fld;
: 	}
: 
: Now consider one line of data, consisting of the following:
: 
: 	1 2 3 4 5
: 
: The awk version would print 5, saying there are 5 fields.
: 
: The perl version, produced by a2p, would print 6, saying
: there are 6 fields.  One can "fix" the perl version in either
: one of two ways: (1) by adding chop; after the while(<>) {
: or (2) by changing the split to "split(' ',$_).  The behavior
: of split without the LIMIT being specified is, of course,
: documented.
: 
: So it would seem that the scripts produced by a2p, in order
: to give the same result as the original awk script, should
: do one of these two things, since it is fairly common for awk
: scripts to assume that NF is going to be the actual number
: of fields.

After the next patch, a2p will throw in a chop if it sees NF.
Way 2 would run into the problem that NF could be too small
if trailing fields were null.

Larry