diff mbox series

C++ PATCH for c++/81676 - bogus -Wunused warnings in constexpr if

Message ID 20190816212900.GV14737@redhat.com
State New
Headers show
Series C++ PATCH for c++/81676 - bogus -Wunused warnings in constexpr if | expand

Commit Message

Marek Polacek Aug. 16, 2019, 9:29 p.m. UTC
This patch is an attempt to fix the annoying -Wunused-but-set-* warnings that
tend to occur with constexpr if.  When we have something like

  template < typename T >
  int f(T v){
      if constexpr(sizeof(T) == sizeof(int)){
	  return v;
      }else{
	  return 0;
      }
  }

and call f('a'), then the condition is false, meaning that we won't instantiate
the then-branch, as per tsubst_expr/IF_STMT:
17284       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
17285         /* Don't instantiate the THEN_CLAUSE. */;
so we'll never get round to mark_exp_read-ing the decls used in the
then-branch, causing finish_function to emit "parameter set but not used"
warnings.

It's unclear how to best deal with this.  Marking the decls DECL_READ_P while
parsing doesn't seem like a viable approach, and since in this testcase the
decl is type-dependent, doing something similiar to the fix for c++/85827 is
out too.

One option is the below, still don't instantiate anything in such a dead
clause, but mark any use of a decl in it as a read.  Which means that "set but
unused" won't be warned about, but that's still better than bogus warnings.
(Clang doesn't seem to handle that, either.)
retrieve_local_specialization is there because we need to set the flag on the
aready instantiated PARM/VAR_DECL, not the template itself.

The good news is that when a param/decl is really unused, we still get the
warning.  And it works with lambdas too.  Better ideas, anyone?
(If anyone's wondering, constexpr function templates don't have this problem.)

Bootstrapped/regtested on x86_64-linux, ok for trunk?

2019-08-16  Marek Polacek  <polacek@redhat.com>

	PR c++/81676 - bogus -Wunused warnings in constexpr if.
	* pt.c (maybe_mark_exp_read_r): New function.
	(tsubst_expr) <case IF_STMT>: Call it for an uninstantiated THEN_CLAUSE
	and ELSE_CLAUSE.

	* g++.dg/cpp1z/constexpr-if30.C: New test.
	* g++.dg/cpp1z/constexpr-if31.C: New test.

Comments

Jason Merrill Aug. 17, 2019, 1:20 a.m. UTC | #1
On 8/16/19 2:29 PM, Marek Polacek wrote:
> This patch is an attempt to fix the annoying -Wunused-but-set-* warnings that
> tend to occur with constexpr if.  When we have something like
> 
>    template < typename T >
>    int f(T v){
>        if constexpr(sizeof(T) == sizeof(int)){
> 	  return v;
>        }else{
> 	  return 0;
>        }
>    }
> 
> and call f('a'), then the condition is false, meaning that we won't instantiate
> the then-branch, as per tsubst_expr/IF_STMT:
> 17284       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
> 17285         /* Don't instantiate the THEN_CLAUSE. */;
> so we'll never get round to mark_exp_read-ing the decls used in the
> then-branch, causing finish_function to emit "parameter set but not used"
> warnings.
> 
> It's unclear how to best deal with this.  Marking the decls DECL_READ_P while
> parsing doesn't seem like a viable approach

Why not?

Jason
Marek Polacek Aug. 19, 2019, 6:28 p.m. UTC | #2
On Fri, Aug 16, 2019 at 06:20:20PM -0700, Jason Merrill wrote:
> On 8/16/19 2:29 PM, Marek Polacek wrote:
> > This patch is an attempt to fix the annoying -Wunused-but-set-* warnings that
> > tend to occur with constexpr if.  When we have something like
> > 
> >    template < typename T >
> >    int f(T v){
> >        if constexpr(sizeof(T) == sizeof(int)){
> > 	  return v;
> >        }else{
> > 	  return 0;
> >        }
> >    }
> > 
> > and call f('a'), then the condition is false, meaning that we won't instantiate
> > the then-branch, as per tsubst_expr/IF_STMT:
> > 17284       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
> > 17285         /* Don't instantiate the THEN_CLAUSE. */;
> > so we'll never get round to mark_exp_read-ing the decls used in the
> > then-branch, causing finish_function to emit "parameter set but not used"
> > warnings.
> > 
> > It's unclear how to best deal with this.  Marking the decls DECL_READ_P while
> > parsing doesn't seem like a viable approach
> 
> Why not?

Well, while parsing, we're in a template and so the condition won't be
evaluated until tsubst_expr.  So we can't tell which branch is dead.

I suppose we could still mark all the _DECLs in then/else branches as read,
in finish_if_stmt if IF_STMT_CONSTEXPR_P, but that seems really really ugly.

--
Marek Polacek • Red Hat, Inc. • 300 A St, Boston, MA
Jason Merrill Aug. 23, 2019, 9:20 p.m. UTC | #3
On 8/19/19 11:28 AM, Marek Polacek wrote:
> On Fri, Aug 16, 2019 at 06:20:20PM -0700, Jason Merrill wrote:
>> On 8/16/19 2:29 PM, Marek Polacek wrote:
>>> This patch is an attempt to fix the annoying -Wunused-but-set-* warnings that
>>> tend to occur with constexpr if.  When we have something like
>>>
>>>     template < typename T >
>>>     int f(T v){
>>>         if constexpr(sizeof(T) == sizeof(int)){
>>> 	  return v;
>>>         }else{
>>> 	  return 0;
>>>         }
>>>     }
>>>
>>> and call f('a'), then the condition is false, meaning that we won't instantiate
>>> the then-branch, as per tsubst_expr/IF_STMT:
>>> 17284       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
>>> 17285         /* Don't instantiate the THEN_CLAUSE. */;
>>> so we'll never get round to mark_exp_read-ing the decls used in the
>>> then-branch, causing finish_function to emit "parameter set but not used"
>>> warnings.
>>>
>>> It's unclear how to best deal with this.  Marking the decls DECL_READ_P while
>>> parsing doesn't seem like a viable approach
>>
>> Why not?
> 
> Well, while parsing, we're in a template and so the condition won't be
> evaluated until tsubst_expr.  So we can't tell which branch is dead.

But if a decl is used on one branch, we shouldn't warn even if it isn't 
used on the selected branch.

Jason
Marek Polacek Aug. 27, 2019, 8:10 p.m. UTC | #4
On Fri, Aug 23, 2019 at 02:20:53PM -0700, Jason Merrill wrote:
> On 8/19/19 11:28 AM, Marek Polacek wrote:
> > On Fri, Aug 16, 2019 at 06:20:20PM -0700, Jason Merrill wrote:
> > > On 8/16/19 2:29 PM, Marek Polacek wrote:
> > > > This patch is an attempt to fix the annoying -Wunused-but-set-* warnings that
> > > > tend to occur with constexpr if.  When we have something like
> > > > 
> > > >     template < typename T >
> > > >     int f(T v){
> > > >         if constexpr(sizeof(T) == sizeof(int)){
> > > > 	  return v;
> > > >         }else{
> > > > 	  return 0;
> > > >         }
> > > >     }
> > > > 
> > > > and call f('a'), then the condition is false, meaning that we won't instantiate
> > > > the then-branch, as per tsubst_expr/IF_STMT:
> > > > 17284       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
> > > > 17285         /* Don't instantiate the THEN_CLAUSE. */;
> > > > so we'll never get round to mark_exp_read-ing the decls used in the
> > > > then-branch, causing finish_function to emit "parameter set but not used"
> > > > warnings.
> > > > 
> > > > It's unclear how to best deal with this.  Marking the decls DECL_READ_P while
> > > > parsing doesn't seem like a viable approach
> > > 
> > > Why not?
> > 
> > Well, while parsing, we're in a template and so the condition won't be
> > evaluated until tsubst_expr.  So we can't tell which branch is dead.
> 
> But if a decl is used on one branch, we shouldn't warn even if it isn't used
> on the selected branch.

I didn't want to mark the decls as read multiple times but we do it anyway
so that's no longer my concern.  So...

Bootstrapped/regtested on x86_64-linux, ok for trunk?

2019-08-27  Marek Polacek  <polacek@redhat.com>

	PR c++/81676 - bogus -Wunused warnings in constexpr if.
	* semantics.c (maybe_mark_exp_read_r): New function.
	(finish_if_stmt): Call it on THEN_CLAUSE and ELSE_CLAUSE.

	* g++.dg/cpp1z/constexpr-if31.C: New test.
	* g++.dg/cpp1z/constexpr-if32.C: New test.

diff --git gcc/cp/semantics.c gcc/cp/semantics.c
index 1f7745933f9..2ba8e635ab7 100644
--- gcc/cp/semantics.c
+++ gcc/cp/semantics.c
@@ -774,6 +774,18 @@ finish_else_clause (tree if_stmt)
   ELSE_CLAUSE (if_stmt) = pop_stmt_list (ELSE_CLAUSE (if_stmt));
 }
 
+/* Callback for cp_walk_tree to mark all {VAR,PARM}_DECLs in a tree as
+   read.  */
+
+static tree
+maybe_mark_exp_read_r (tree *tp, int *, void *)
+{
+  tree t = *tp;
+  if (VAR_P (t) || TREE_CODE (t) == PARM_DECL)
+    mark_exp_read (t);
+  return NULL_TREE;
+}
+
 /* Finish an if-statement.  */
 
 void
@@ -781,6 +793,16 @@ finish_if_stmt (tree if_stmt)
 {
   tree scope = IF_SCOPE (if_stmt);
   IF_SCOPE (if_stmt) = NULL;
+  if (IF_STMT_CONSTEXPR_P (if_stmt))
+    {
+      /* Prevent various -Wunused warnings.  We might not instantiate
+	 either of these branches, so we would not mark the variables
+	 used in that branch as read.  */
+      cp_walk_tree_without_duplicates (&THEN_CLAUSE (if_stmt),
+				       maybe_mark_exp_read_r, NULL);
+      cp_walk_tree_without_duplicates (&ELSE_CLAUSE (if_stmt),
+				       maybe_mark_exp_read_r, NULL);
+    }
   add_stmt (do_poplevel (scope));
 }
 
diff --git gcc/testsuite/g++.dg/cpp1z/constexpr-if31.C gcc/testsuite/g++.dg/cpp1z/constexpr-if31.C
new file mode 100644
index 00000000000..02140cff9fd
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp1z/constexpr-if31.C
@@ -0,0 +1,79 @@
+// PR c++/81676 - bogus -Wunused warnings in constexpr if.
+// { dg-do compile { target c++17 } }
+// { dg-options "-Wall -Wextra" }
+
+template <typename T> int
+f1 (T v)
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return v + x;
+  else
+    return 0;
+}
+
+template <typename T> int
+f2 (T v) // { dg-warning "unused parameter .v." }
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return x;
+  else
+    return 0;
+}
+
+template <typename T> int
+f3 (T v)
+{
+  T x = 0; // { dg-warning "unused variable .x." }
+  if constexpr(sizeof(T) == sizeof(int))
+    return v;
+  else
+    return 0;
+}
+
+template <typename T> int
+f4 (T v)
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return 0;
+  else
+    return v + x;
+}
+
+template <typename T> int
+f5 (T v) // { dg-warning "unused parameter .v." }
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return 0;
+  else
+    return x;
+}
+
+template <typename T> int
+f6 (T v)
+{
+  T x = 0; // { dg-warning "unused variable .x." }
+  if constexpr(sizeof(T) == sizeof(int))
+    return 0;
+  else
+    return v;
+}
+
+int main()
+{
+  f1(0);
+  f1('a');
+  f2(0);
+  f2('a');
+  f3(0);
+  f3('a');
+  f4(0);
+  f4('a');
+  f5(0);
+  f5('a');
+  f6(0);
+  f6('a');
+}
diff --git gcc/testsuite/g++.dg/cpp1z/constexpr-if32.C gcc/testsuite/g++.dg/cpp1z/constexpr-if32.C
new file mode 100644
index 00000000000..13a6039fce6
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp1z/constexpr-if32.C
@@ -0,0 +1,16 @@
+// PR c++/81676 - bogus -Wunused warnings in constexpr if.
+// { dg-do compile { target c++17 } }
+// { dg-options "-Wall -Wextra" }
+
+int main()
+{
+    auto f = [](auto a, auto b) {
+        if constexpr (sizeof(b) == 1) {
+            return a;
+        } else {
+            return b;
+        }
+    };
+
+    return f(1, 1) + f(1, 'a');
+}
Jason Merrill Aug. 28, 2019, 1:18 a.m. UTC | #5
On Tue, Aug 27, 2019 at 4:10 PM Marek Polacek <polacek@redhat.com> wrote:
> On Fri, Aug 23, 2019 at 02:20:53PM -0700, Jason Merrill wrote:
> > On 8/19/19 11:28 AM, Marek Polacek wrote:
> > > On Fri, Aug 16, 2019 at 06:20:20PM -0700, Jason Merrill wrote:
> > > > On 8/16/19 2:29 PM, Marek Polacek wrote:
> > > > > This patch is an attempt to fix the annoying -Wunused-but-set-* warnings that
> > > > > tend to occur with constexpr if.  When we have something like
> > > > >
> > > > >     template < typename T >
> > > > >     int f(T v){
> > > > >         if constexpr(sizeof(T) == sizeof(int)){
> > > > >           return v;
> > > > >         }else{
> > > > >           return 0;
> > > > >         }
> > > > >     }
> > > > >
> > > > > and call f('a'), then the condition is false, meaning that we won't instantiate
> > > > > the then-branch, as per tsubst_expr/IF_STMT:
> > > > > 17284       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
> > > > > 17285         /* Don't instantiate the THEN_CLAUSE. */;
> > > > > so we'll never get round to mark_exp_read-ing the decls used in the
> > > > > then-branch, causing finish_function to emit "parameter set but not used"
> > > > > warnings.
> > > > >
> > > > > It's unclear how to best deal with this.  Marking the decls DECL_READ_P while
> > > > > parsing doesn't seem like a viable approach
> > > >
> > > > Why not?
> > >
> > > Well, while parsing, we're in a template and so the condition won't be
> > > evaluated until tsubst_expr.  So we can't tell which branch is dead.
> >
> > But if a decl is used on one branch, we shouldn't warn even if it isn't used
> > on the selected branch.
>
> I didn't want to mark the decls as read multiple times but we do it anyway
> so that's no longer my concern.  So...
>
> Bootstrapped/regtested on x86_64-linux, ok for trunk?
>
> 2019-08-27  Marek Polacek  <polacek@redhat.com>
>
>         PR c++/81676 - bogus -Wunused warnings in constexpr if.
>         * semantics.c (maybe_mark_exp_read_r): New function.
>         (finish_if_stmt): Call it on THEN_CLAUSE and ELSE_CLAUSE.

I was thinking of adding mark_exp_read to places where we currently
take a shortcut in templates, like check_return_expr or
build_x_binary_op, but this is probably simpler.  The patch is OK.

Jason
diff mbox series

Patch

diff --git gcc/cp/pt.c gcc/cp/pt.c
index 17585119bce..e146ad38443 100644
--- gcc/cp/pt.c
+++ gcc/cp/pt.c
@@ -16995,6 +16995,23 @@  lookup_init_capture_pack (tree decl)
   return r;
 }
 
+/* Callback for cp_walk_tree to mark all instantiations of {VAR,PARM}_DECLs
+   in a tree as read.  The point is to mark the parameters or local variables
+   of a template function as read to prevent various -Wunused warnings.  */
+
+static tree
+maybe_mark_exp_read_r (tree *tp, int *, void *)
+{
+  tree t = *tp;
+  if (VAR_P (t) || TREE_CODE (t) == PARM_DECL)
+    {
+      tree s = retrieve_local_specialization (t);
+      if (s)
+	mark_exp_read (s);
+    }
+  return NULL_TREE;
+}
+
 /* Like tsubst_copy for expressions, etc. but also does semantic
    processing.  */
 
@@ -17282,7 +17299,10 @@  tsubst_expr (tree t, tree args, tsubst_flags_t complain, tree in_decl,
 	  break;
 	}
       if (IF_STMT_CONSTEXPR_P (t) && integer_zerop (tmp))
-	/* Don't instantiate the THEN_CLAUSE. */;
+	/* Don't instantiate the THEN_CLAUSE.  But mark the PARM_DECLs and
+	   VAR_DECLs used in the THEN_CLAUSE as read, so as to suppress
+	   -Wunused-but-set-{variable,parameter} warnings.  */
+	cp_walk_tree (&THEN_CLAUSE (t), maybe_mark_exp_read_r, NULL, NULL);
       else
 	{
 	  tree folded = fold_non_dependent_expr (tmp, complain);
@@ -17296,7 +17316,10 @@  tsubst_expr (tree t, tree args, tsubst_flags_t complain, tree in_decl,
       finish_then_clause (stmt);
 
       if (IF_STMT_CONSTEXPR_P (t) && integer_nonzerop (tmp))
-	/* Don't instantiate the ELSE_CLAUSE. */;
+	/* Don't instantiate the ELSE_CLAUSE.  But mark the PARM_DECLs and
+	   VAR_DECLs used in the ELSE_CLAUSE as read, so as to suppress
+	   -Wunused-but-set-{variable,parameter} warnings.  */
+	cp_walk_tree (&ELSE_CLAUSE (t), maybe_mark_exp_read_r, NULL, NULL);
       else if (ELSE_CLAUSE (t))
 	{
 	  tree folded = fold_non_dependent_expr (tmp, complain);
diff --git gcc/testsuite/g++.dg/cpp1z/constexpr-if30.C gcc/testsuite/g++.dg/cpp1z/constexpr-if30.C
new file mode 100644
index 00000000000..26e0c6f39a6
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp1z/constexpr-if30.C
@@ -0,0 +1,79 @@ 
+// PR c++/81676 - bogus -Wunused warnings in constexpr if.
+// { dg-do compile { target c++17 } }
+// { dg-options "-Wunused-variable -Wunused-parameter" }
+
+template <typename T> int
+f1 (T v)
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return v + x;
+  else
+    return 0;
+}
+
+template <typename T> int
+f2 (T v) // { dg-warning "unused parameter .v." }
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return x;
+  else
+    return 0;
+}
+
+template <typename T> int
+f3 (T v)
+{
+  T x = 0; // { dg-warning "unused variable .x." }
+  if constexpr(sizeof(T) == sizeof(int))
+    return v;
+  else
+    return 0;
+}
+
+template <typename T> int
+f4 (T v)
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return 0;
+  else
+    return v + x;
+}
+
+template <typename T> int
+f5 (T v) // { dg-warning "unused parameter .v." }
+{
+  T x = 0;
+  if constexpr(sizeof(T) == sizeof(int))
+    return 0;
+  else
+    return x;
+}
+
+template <typename T> int
+f6 (T v)
+{
+  T x = 0; // { dg-warning "unused variable .x." }
+  if constexpr(sizeof(T) == sizeof(int))
+    return 0;
+  else
+    return v;
+}
+
+int main()
+{
+  f1(0);
+  f1('a');
+  f2(0);
+  f2('a');
+  f3(0);
+  f3('a');
+  f4(0);
+  f4('a');
+  f5(0);
+  f5('a');
+  f6(0);
+  f6('a');
+}
diff --git gcc/testsuite/g++.dg/cpp1z/constexpr-if31.C gcc/testsuite/g++.dg/cpp1z/constexpr-if31.C
new file mode 100644
index 00000000000..a1db91a5b40
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp1z/constexpr-if31.C
@@ -0,0 +1,16 @@ 
+// PR c++/81676 - bogus -Wunused warnings in constexpr if.
+// { dg-do compile { target c++17 } }
+// { dg-options "-Wunused-variable -Wunused-parameter" }
+
+int main()
+{
+    auto f = [](auto a, auto b) {
+        if constexpr (sizeof(b) == 1) {
+            return a;
+        } else {
+            return b;
+        }
+    };
+
+    return f(1, 1) + f(1, 'a');
+}