www.delorie.com/archives/browse.cgi   search  
Mail Archives: pgcc/1999/05/17/16:55:55

Date: Mon, 17 May 1999 16:48:49 -0400
From: zawrotny AT hhmi DOT umbc DOT edu
To: pgcc AT delorie DOT com
Subject: g77 signal 6 not in egcs-g77
Message-ID: <19990517164849.A2885@gly.hhmi.umbc.edu>
Mime-Version: 1.0
X-Mailer: Mutt 0.95.4i
Reply-To: pgcc AT delorie DOT com

Pgcc people,

I am looking at using pgcc instead of egcs for some of our newer 
workstations and have run across a problem in the newest release
of pgcc-g77 (the rpm from Foyer).  g77 -v gives:

g77 version pgcc-2.91.66 19990314 (egcs-1.1.2 release) (from FSF-g77 version 
0.5.24-19981002)
Driving: g77 -v -c -xf77-version /dev/null -xnone
Reading specs from /opt/pgcc/lib/gcc-lib/i686-pc-linux-gnu/pgcc-2.91.66/specs
gcc version pgcc-2.91.66 19990314 (egcs-1.1.2 release)

The problem is in some legacy code that has mismatched global procedure
declarations and calls.  The error message given is:

Argument #2 of `local' is one type at (2) but is some other type at (1) 
[info -f g77 M GLOBALS]

As indicated in the info page referenced above, I tried using -fno-globals
to prevent the compiler error.  This allows the compile to procede, but
an attempt to execute the program core dumps.  The gdb stack trace is:

(gdb) where
#0  0x4006e811 in __kill ()
#1  0x4006e63f in raise (sig=6) at ../sysdeps/posix/raise.c:27
#2  0x4006f84f in abort () at ../sysdeps/generic/abort.c:83

When compiled with g77 from egcs 1.0.3, the program runs normally, as it
does when compiled on one of our SGIs with f77.  The egcs-g77 version
info is:

/usr/bin/g77 -v
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/egcs-2.90.29/specs
gcc version egcs-2.90.29 980515 (egcs-1.0.3 release)

Although I understand that the proper solution to the problem is to correct
the sources, it really isn't an option for me since the package is not
maintained locally and I don't know any fortran at all.  Also, the proper
behavior of the application under egcs-1.0.3-g77 and f77 seems to suggest
that the problem may be related to some of the optimizations in pgcc-g77,
which is of course why I want to use pgcc.

If I can provide any further information, please feel free to contact me.


Thanks,

Mike

-- 
Dr. Michael Zawrotny
Howard Hughes Medical Institute at UMBC
1000 Hilltop Circle			| email:  zawrotny AT hhmi DOT umbc DOT edu
Baltimore, MD 21250			| phone:  (410) 455-6347

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019