Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!gatech!tut.cis.ohio-state.edu!cica!ctrsol!ginosko!usc!hacgate!goddard@aic.hrl.hac.com
From: goddard@aic.hrl.hac.com
Newsgroups: gnu.gcc.bug
Subject: GCC produces strange n_value on SunOS 4.0.3
Message-ID: <5343@hacgate.UUCP>
Date: 4 Oct 89 23:58:48 GMT
Sender: news@hacgate.UUCP
Reply-To: goddard@aic.hrl.hac.com
Distribution: gnu
Organization: Hughes Aircraft Co., El Segundo, CA
Lines: 52

I compile a program with a single global integer, and then run nm on
the .o file.  The value that nm prints for the integer is 8, but the
size of an integer is reported to be 4.  I thought the value printed
by nm was supposed to be the size of the data item.  Am I wrong or is
there a bug in GCC?

Here is the program:

	#include <stdio.h>

	int i;

	main()

	{
	  fprintf(stdout, "int (%d) float (%d) double (%d) long (%d)\n",
		  sizeof(int), sizeof(float), sizeof(double), sizeof(long));
	}

I compile it:

	gcc -c temp.c
	gcc temp.o -o temp

and run it:

	virus% temp
	int (4) float (4) double (8) long (4)
	virus%

and then run nm:

	virus% nm temp.o
	         U __iob
	         U _fprintf
->      00000008 C _i
	0000002c T _main
	00000000 t gcc_compiled.
	virus%

but if I use cc instead of GCC, I get:

	virus% cc -c temp.c
	virus% cc temp.o -o temp
	virus% temp
	int (4) float (4) double (8) long (4)
	virus% nm temp.o
	         U __iob
	         U _fprintf
->      00000004 C _i
	00000000 T _main
	virus%