c++: Ensure !!var is not an lvalue [PR107065]

  Hi!

The TRUTH_NOT_EXPR case in cp_build_unary_op is one of the spots where
we somewhat fold immediately using invert_truthvalue_loc.
I've tried using
  return build1_loc (location, TRUTH_NOT_EXPR, boolean_type_node, arg);
in there instead, but unfortunately that regressed
Wlogical-not-parentheses-*.c pr49706.c pr62199.c pr65120.c sequence-pt-1.C
tests, so at least for backporting that doesn't seem to be a way to go.

So, this patch instead wraps it into NON_LVALUE_EXPR if needed (which also
need a tweak for some tests in the pr47906.c test, but nothing major),
with the intent to make it backportable, and later I'll try to do further
steps to avoid folding here prematurely.  Most of the problems with
build1 TRUTH_NOT_EXPR are that it doesn't even invert comparisons as most
common case and lots of warning code isn't able to deal with ! around
comparisons; so perhaps one way to do this would be fold by hand only
invertable comparisons and for the rest create TRUTH_NOT_EXPR.

Anyway, bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk
(for now) and for release branches (after a while on the trunk)?

2022-12-10  Jakub Jelinek  <jakub@redhat.com>

	PR c++/107065
gcc/cp/
	* typeck.cc (cp_build_unary_op) <case TRUTH_NOT_EXPR>: If
	invert_truthvalue_loc returns obvalue_p, wrap it into NON_LVALUE_EXPR.
	* parser.cc (cp_parser_binary_expression): Don't call
	warn_logical_not_parentheses if current.lhs is a NON_LVALUE_EXPR
	of a decl with boolean type.
gcc/testsuite/
	* g++.dg/cpp0x/pr107065.C: New test.


	Jakub
  

On 12/10/22 04:33, Jakub Jelinek wrote:
> Hi!
> 
> The TRUTH_NOT_EXPR case in cp_build_unary_op is one of the spots where
> we somewhat fold immediately using invert_truthvalue_loc.
> I've tried using
>    return build1_loc (location, TRUTH_NOT_EXPR, boolean_type_node, arg);
> in there instead, but unfortunately that regressed
> Wlogical-not-parentheses-*.c pr49706.c pr62199.c pr65120.c sequence-pt-1.C
> tests, so at least for backporting that doesn't seem to be a way to go.
> 
> So, this patch instead wraps it into NON_LVALUE_EXPR if needed (which also
> need a tweak for some tests in the pr47906.c test, but nothing major),
> with the intent to make it backportable, and later I'll try to do further
> steps to avoid folding here prematurely.  Most of the problems with
> build1 TRUTH_NOT_EXPR are that it doesn't even invert comparisons as most
> common case and lots of warning code isn't able to deal with ! around
> comparisons; so perhaps one way to do this would be fold by hand only
> invertable comparisons and for the rest create TRUTH_NOT_EXPR.
> 
> Anyway, bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk
> (for now) and for release branches (after a while on the trunk)?

OK.

> 2022-12-10  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR c++/107065
> gcc/cp/
> 	* typeck.cc (cp_build_unary_op) <case TRUTH_NOT_EXPR>: If
> 	invert_truthvalue_loc returns obvalue_p, wrap it into NON_LVALUE_EXPR.
> 	* parser.cc (cp_parser_binary_expression): Don't call
> 	warn_logical_not_parentheses if current.lhs is a NON_LVALUE_EXPR
> 	of a decl with boolean type.
> gcc/testsuite/
> 	* g++.dg/cpp0x/pr107065.C: New test.
> 
> --- gcc/cp/typeck.cc.jj	2022-11-30 10:29:42.024701797 +0100
> +++ gcc/cp/typeck.cc	2022-12-09 17:47:54.132856233 +0100
> @@ -7396,9 +7396,13 @@ cp_build_unary_op (enum tree_code code,
>   				   build_zero_cst (TREE_TYPE (arg)), complain);
>         arg = perform_implicit_conversion (boolean_type_node, arg,
>   					 complain);
> -      val = invert_truthvalue_loc (location, arg);
>         if (arg != error_mark_node)
> -	return val;
> +	{
> +	  val = invert_truthvalue_loc (location, arg);
> +	  if (obvalue_p (val))
> +	    val = non_lvalue_loc (location, val);
> +	  return val;
> +	}
>         errstring = _("in argument to unary !");
>         break;
>   
> --- gcc/cp/parser.cc.jj	2022-12-09 11:02:35.871444993 +0100
> +++ gcc/cp/parser.cc	2022-12-09 19:14:26.698847734 +0100
> @@ -10224,7 +10224,10 @@ cp_parser_binary_expression (cp_parser*
>   		      || (TREE_CODE (TREE_TYPE (TREE_OPERAND (current.lhs, 0)))
>   			  != BOOLEAN_TYPE))))
>   	  /* Avoid warning for !!b == y where b is boolean.  */
> -	  && (!DECL_P (tree_strip_any_location_wrapper (current.lhs))
> +	  && (!(DECL_P (tree_strip_any_location_wrapper (current.lhs))
> +		|| (TREE_CODE (current.lhs) == NON_LVALUE_EXPR
> +		    && DECL_P (tree_strip_any_location_wrapper
> +					    (TREE_OPERAND (current.lhs, 0)))))
>   	      || TREE_TYPE (current.lhs) == NULL_TREE
>   	      || TREE_CODE (TREE_TYPE (current.lhs)) != BOOLEAN_TYPE))
>   	warn_logical_not_parentheses (current.loc, current.tree_type,
> --- gcc/testsuite/g++.dg/cpp0x/pr107065.C.jj	2022-12-09 16:22:59.686548071 +0100
> +++ gcc/testsuite/g++.dg/cpp0x/pr107065.C	2022-12-09 16:22:59.686548071 +0100
> @@ -0,0 +1,14 @@
> +// PR c++/107065
> +// { dg-do compile { target c++11 } }
> +
> +template<class, class> struct is_same { static constexpr bool value = false; };
> +template<class T> struct is_same<T, T> { static constexpr bool value = true; };
> +
> +int
> +main ()
> +{
> +  bool b = true;
> +  static_assert (is_same<decltype (!(!b)), bool>::value, "");
> +  auto bb = (!(!b));
> +  static_assert (is_same<decltype (bb), bool>::value, "");
> +}
> 
> 	Jakub
>