Relay-Version: version B 2.10.2 9/18/84; site lsuc.UUCP
Posting-Version: version B 2.10.1.1 9/4/83; site lpi3230.UUCP
Path: lsuc!pesnta!lpi3230!steve
From: steve@lpi3230.UUCP (Steve Burbeck)
Newsgroups: pe.cust.bugs
Subject: Re: floating point doesn't work on our awk
Message-ID: <147@lpi3230.UUCP>
Date: 27 Feb 85 15:21:41 GMT
Date-Received: 27 Feb 85 19:11:20 GMT
Expires: 15 Mar 85 08:00:00 GMT
Organization: Linus-Pauling Institute, Palo Alto, Calif.
Lines: 31

>Floating-point doesn't work on our awk. (We have a 3220 running
>Edition VII, no floating-point hardware, floating point is trapped
>by the kernel and works in C.)
>Is there a fix?

1) this has been reported on the net some months ago by someone.
   The only fix I know of is to use an older version of awk.  You
   are using the v2.4 version I would guess.  Our version (v.2.2.1.1
   or v.2.3 dated Aug 23 1980) works with floats just fine.

2) But this brings up again the question of why PE doesn't make itself
   responsible for maintaining and distributing a bug list so that each
   site doesn't have to find out for itself about each bug.  I'll bet
   that you wasted quite a bit of time before you gave up and decided
   it wasn't something you were doing, it was a bug in awk.  At the
   last Interchange meetings, both Axel and Carl made rather strong
   arguments to PE about just this problem.  There was no response.

3) Another related point is that it would be helpful if PE would at least
   include a list of all commands that have been changed in each new
   distribution.  There are two benefits.  First, if a command such as
   awk seems to have a problem, you could at least see whether it has
   been changed (and hence may be suspected of harboring a new bug).
   Second, if a new distribution has few changes it would be easier to
   install just the changes into your currently working system than to
   have to install the whole thing, and then add back in your own site
   dependent system changes (e.g., uucp patches, permission changes,
   etc.).

					Steve Burbeck
					Linus Pauling Institute