Re: preliminary patch for objc_comptypes() fixes

[Stan Shebs <shebs at apple dot com>]

Nicola Pero wrote:

> Hi Stan,
>
> sorry for having been a few days late with this.
No problem, this looks great.  The comprehensive testcases are excellent,
almost as if you used gcov to confirm that all paths were covered... :-)

> Thu Sep 12 12:09:12 2002  Nicola Pero  <n.pero@mi.flashnet.it>
>
> 	Fix PR/7014 and related objc typechecking bugs:
> 	* c-typeck.c (comp_target_types): Added a reflexive argument.
> 	Pass it to ObjC when/if calling objc_comptypes().  Updated all
> 	callers to provide the appropriate reflexive argument.
> 	* objc/objc-act.c (objc_comptypes): Carefully checked and fixed
> 	typechecking for all cases of comparisons and assignments,
> 	particularly the obscure and less common ones involving protocols.
OK for the trunk, with a couple micro-nits:

>       if (flag_objc)
> 	objc_comptypes (type, rhstype, 0);
>       return rhs;
>     }
> -
>   if (coder == VOID_TYPE)
>     {
>       error ("void value not ignored as it ought to be");
Better to leave the blank line in, for spacing.

> 	      for (lproto = lproto_list; lproto; lproto = TREE_CHAIN (lproto))
> 		{
> 		  p = TREE_VALUE (lproto);
> 		  rproto = 0;
> 		  rinter = lookup_interface (rname);
> -
> +		  
> 		  while (rinter && !rproto)
> 		    {
> 		      tree cat;
You're replacing an empty line with a line with extra whitespace;
the "canonical" form is to have the line be totally empty.

Don't forget change entry for testcases!

Stan