Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!uunet!aai!leo
From: leo@aai.uu.net (Leo Pinard)
Newsgroups: comp.unix.questions
Subject: Re: Trouble with cxref(1)
Keywords: cxref sun3 3B1
Message-ID: <1990Feb11.053937.7629@aai.uu.net>
Date: 11 Feb 90 05:39:37 GMT
References: <2019@cjsa.WA.COM>
Reply-To: leo@aai.UU.NET (Leo Pinard)
Organization: Amerinex Artificial Intelligence, Amherst MA
Lines: 31

In article <2019@cjsa.WA.COM> jeff@cjsa.WA.COM (Jeffery Small) writes:
>I have been unable to get cxref(1) to work.  I am attempting to run this on
>a set of program files which compile and lint clean.  I have tried all manner
>of invocations with no improvement so I am tossing this out for comments.
>
>Now, on a Sun-3/60 (4.0.3 OS) using either /bin/cxref of /usr/5bin/cxref,
>I get the following message:
>
>    SWITCH ERROR IN TMPSCAN: inafunct = no
>
>and cxref aborts.  Does anyone have any idea what I am doing wrong or what
>problems I should be looking for?

I posted a question about this recently and got an answer for the Sun, at
least.  Jim Litchfield sent me the answer (Thanks, Jim!).  It's a bug in cpp
that can be worked around by running cpp separately, and then cxref.  I use the
following script:

#! /bin/csh
foreach name (*.c)
/usr/lib/cpp -Dunix -Dsun -Dsparc $name >${name:r}.cref.c
end
cxref -cst -o rpe.cref *.cref.c 
foreach name (*.cref.c)
\rm $name
end

To find out what options to pass the compiler, do a cc -dryrun.  The bug is
supposedly fixed in 4.1.

Leo Pinard  leo@aai.uu.net