Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!bbn!gatech!hubcap!rchampe
From: rchampe@hubcap.UUCP (Richard Champeaux)
Newsgroups: comp.sys.amiga
Subject: Re: Aztec C bug
Message-ID: <691@hubcap.UUCP>
Date: Thu, 19-Nov-87 12:53:45 EST
Article-I.D.: hubcap.691
Posted: Thu Nov 19 12:53:45 1987
Date-Received: Sun, 22-Nov-87 02:23:29 EST
References: <532@oscvax.UUCP>
Distribution: na
Organization: Clemson University, Clemson, SC
Lines: 30
Summary: Is that legal to begin with?

In article <532@oscvax.UUCP>, rico@oscvax.UUCP (Rico Mariani) writes:
> 
> main()
> {
> 	struct foo { int x; } bar;
> 
> 	if (bar) ;
> }
> 
> compiles to:
> 
> ;	if (bar) ;
> 	tst.r	-2(a5)		<----- Say what?   what's a .r???
> 	beq	.4

     I'm not an expert in C, but doesn't "if (bar)" mean "if (bar!=0)" and
its illegal to compare a structure to an integer even if the structure is
only an integer, (I could be wrong, like I said, I'm no expert.)
If I remember right, using a structure name alone does not give a pointer
to that structure like using the unsubscripted name of an array would.  To get
a pointer to the structure bar, you would use &bar.
(Once again, I could be wrong, I'm no expert.)
     Anyways, a simmilar thing happened to me a while back, I don't remember
the exact code but it had to do with structures and pointers. I got a 
move.x instead.
     I think it's less of a bug, and more of a very poor way to point out
errors.

Rich Champeaux
Clemson University