[PATCH] remove warning "call with no type"

[PATCH] remove warning "call with no type"

[Luc Van Oostenryck]

This warning is issued when trying to linearize a call
expressions with a null expr->ctype but such null ctype
is a consequence of an error detected during evaluation
and this error must already have issued a warning or an
error message specific to the problem.

In short, this "call with no type" is non-informative
and redundant, so avoid usless noise and remove this
warning.

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 linearize.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/linearize.c b/linearize.c
index 898acac88..739fc7c07 100644
--- a/linearize.c
+++ b/linearize.c
@@ -1273,10 +1273,8 @@ static pseudo_t linearize_call_expression(struct entrypoint *ep, struct expressi
 	struct symbol *fntype;
 	struct context *context;
 
-	if (!expr->ctype) {
-		warning(expr->pos, "call with no type!");
+	if (!expr->ctype)
 		return VOID;
-	}
 
 	fn = expr->fn;
 	fntype = fn->ctype;
-- 
2.16.0

Re: [PATCH] remove warning "call with no type"

[Christopher Li]

On Wed, Jan 31, 2018 at 2:40 AM, Luc Van Oostenryck
<luc.vanoostenryck@gmail.com> wrote:
> This warning is issued when trying to linearize a call
> expressions with a null expr->ctype but such null ctype
> is a consequence of an error detected during evaluation
> and this error must already have issued a warning or an
> error message specific to the problem.
>
> In short, this "call with no type" is non-informative
> and redundant, so avoid usless noise and remove this
> warning.

The change looks fine to me. This the change assume the
expr->ctype be NULL is result of an error and has been reported
before. I am sure that assumption holds for most of the cases. Not
sure it holds for all corner cases. On the other hand, if the expr->ctype
is NULL, it is already screwed, change it to VOID is not too bad.

Chris

Re: [PATCH] remove warning "call with no type"

[Luc Van Oostenryck]

On Sat, Feb 10, 2018 at 11:00:26AM -0800, Christopher Li wrote:
> On Wed, Jan 31, 2018 at 2:40 AM, Luc Van Oostenryck
> <luc.vanoostenryck@gmail.com> wrote:
> > This warning is issued when trying to linearize a call
> > expressions with a null expr->ctype but such null ctype
> > is a consequence of an error detected during evaluation
> > and this error must already have issued a warning or an
> > error message specific to the problem.
> >
> > In short, this "call with no type" is non-informative
> > and redundant, so avoid usless noise and remove this
> > warning.
> 
> The change looks fine to me. This the change assume the
> expr->ctype be NULL is result of an error and has been reported
> before. I am sure that assumption holds for most of the cases. Not
> sure it holds for all corner cases.

I consider that if it this not the case, it's a bug which
will need to be fixed by adding the proper error reporting.
But I haven't yet seen such case.

-- Luc