Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site emory.UUCP
Path: utzoo!linus!vaxine!wjh12!genrad!decvax!mcnc!akgua!emory!km
From: km@emory.UUCP (Ken Mandelberg)
Newsgroups: net.bugs.4bsd,net.lang.pascal
Subject: Re: Bug in Pascal (Pix) on 4.2BSD
Message-ID: <1282@emory.UUCP>
Date: Mon, 13-Feb-84 18:31:02 EST
Article-I.D.: emory.1282
Posted: Mon Feb 13 18:31:02 1984
Date-Received: Fri, 17-Feb-84 05:19:28 EST
References: <1277@emory.UUCP>
Organization: Math & Computer Science, Emory University, Atlanta
Lines: 49


This is a followup to a bug I reported in Pix (really Px)
in reading real numbers.

I just traced the problem to a library routine READ8.c in
/usr/lib/libpc.a . However, when I checked the source (and .o in
the archive) I found that the compiled px contained  a later version
of READ8.c than the library.

Specifically, if I rebuild px and diff the 'what' against the
distributed px I get:

1c1
< px.rebuilt
---
> px.cameoff4.2tape
20c20
< 	EXCEPT.c 1.3 1/10/83
---
> 	EXCEPT.c 1.4 10/1/83
39,40c39,40
< 	READ4.c 1.6 4/9/82
< 	READ8.c 1.5 4/9/82
---
> 	READ4.c 1.8 10/28/83
> 	READ8.c 1.7 10/28/83
42c42
< 	READE.c 1.5 6/10/81
---
> 	READE.c 1.6 10/28/83
75c75
< 	UNSYNC.c 1.3 6/10/81
---
> 	UNSYNC.c 1.4 10/28/83

As it happens the rebuilt version (with the older modules)
does not have my real read bug, so I am happy. However,
I am nervous about using modules that are as much as
two years older than what was imbedded in the px that
came on our 4.2 tape. I don't have much alternative, since
there is no source on our 4.2 tape for the newer versions.

If anyone has the source to the newer modules, or knows what
they fix I would like to hear about it.

Ken Mandelberg
Emory University
Dept of Math and CS
{akgua,sb1,gatech}!emory!km