Hi Martin, On Thu, Aug 14, 2025 at 02:21:10PM +0200, Martin Uecker wrote: > > > You could also add a warning when _Countof is applied to a > > > non-constant pointer. > > > > Yup, I should add that. I wouldn't enable it in any groups flags, > > though. I don't want people to start writing [const n]. Instead, > > I expect them to use -farray-parameters-are-const in the future. > > Or a warning when array parameters are modified. I think this would > make the most sense at the moment.
That would be effectively implementing -farray-parameters-are-const.
I think it makes sense to actually implement it, which will result in
simpler GCC code, and don't add any special code here.
So, I think I'll leave this without a warning, pending on
-farray-parameters-are-const.
> > > > So, I stand by this patch even with this known limitation.
> > > >
> > > > On the other, hand, it may be a GCC existing bug? I see that
> > > > _Countof(a+0) remembers the array number of elements but _Countof(a+1)
> > > > doesn't. Maybe once you do a++ you should also forget the number of
> > > > elements (or update it)?
> > >
> > > Difficult, as this is then flow-sensitive and not simply a
> > > property of 'a'. That a + 0 works because the expression
> > > is folded before. Maybe you could change it to take the expression
> > > before folding?
> >
> > Hmmm, yeah, that would be good. It would fix the _surprise() test
> > cases. Do you expect any issues by removing the call to c_fully_fold()?
>
> In the code path where we have a PARM_DECL, it is not needed.
> Otherwise it is.
Thanks!
I don't know much about c_fully_fold(). Would this be enough? Or do
I need to add more conditionals?
diff --git i/gcc/c/c-typeck.cc w/gcc/c/c-typeck.cc
index 0080da494a9b..7f24880297de 100644
--- i/gcc/c/c-typeck.cc
+++ w/gcc/c/c-typeck.cc
@@ -3853,40 +3853,43 @@ c_expr_countof_expr (location_t loc, struct
c_expr expr)
}
else
{
bool expr_const_operands = true;
ret.original_code = COUNTOF_EXPR;
ret.original_type = NULL;
ret.m_decimal = 0;
tree folded_expr = c_fully_fold (expr.value,
require_constant_value,
&expr_const_operands);
tree type = TREE_TYPE (folded_expr);
if (POINTER_TYPE_P (type))
{
if (tree rv = c_array_parm_orig_type (folded_expr))
{
pedwarn (input_location, OPT_Wpedantic,
"ISO C does not support array parameters in %qs",
"_Countof");
type = rv;
+
+ /* Don't fold in array parameters. */
+ folded_expr = expr.value;
}
}
ret.value = c_countof_type (loc, type);
c_last_sizeof_arg = expr.value;
c_last_sizeof_loc = loc;
if (is_top_array_vla (type))
{
/* countof is evaluated when given a vla. */
ret.value = build2 (C_MAYBE_CONST_EXPR, TREE_TYPE (ret.value),
folded_expr, ret.value);
C_MAYBE_CONST_EXPR_NON_CONST (ret.value) =
!expr_const_operands;
SET_EXPR_LOCATION (ret.value, loc);
}
pop_maybe_used (is_top_array_vla (type));
}
return ret;
}
Cheers,
Alex
--
<https://www.alejandro-colomar.es/>
signature.asc
Description: PGP signature
