public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r13-1219] c++: -Waddress and if constexpr [PR94554]
@ 2022-06-23 15:08 Jason Merrill
0 siblings, 0 replies; only message in thread
From: Jason Merrill @ 2022-06-23 15:08 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:124a9e08b7a83795bd4d09001955f0eef68ecd00
commit r13-1219-g124a9e08b7a83795bd4d09001955f0eef68ecd00
Author: Jason Merrill <jason@redhat.com>
Date: Thu Jun 23 00:24:22 2022 -0400
c++: -Waddress and if constexpr [PR94554]
Like we avoid various warnings for seemingly tautological expressions when
substituting a template, we should avoid warning for the implicit conversion
to bool in an if statement. I considered also doing this for the conditions
in loop expressions, but that seems unnecessary, as a loop condition is
unlikely to be a constant.
The change to finish_if_stmt_cond isn't necessary since dependent_operand_p
looks through IMPLICIT_CONV_EXPR, but makes it more constent with
e.g. build_x_binary_op that determines the type of an expression and then
builds it using the original operands.
PR c++/94554
gcc/cp/ChangeLog:
* pt.cc (dependent_operand_p): Split out from...
(tsubst_copy_and_build): ...here.
(tsubst_expr) [IF_STMT]: Use it.
* semantics.cc (finish_if_stmt_cond): Keep the pre-conversion
condition in the template tree.
gcc/testsuite/ChangeLog:
* g++.dg/cpp1z/constexpr-if38.C: New test.
Diff:
---
gcc/cp/pt.cc | 43 +++++++++++++++++++++--------
gcc/cp/semantics.cc | 10 +++++--
gcc/testsuite/g++.dg/cpp1z/constexpr-if38.C | 16 +++++++++++
3 files changed, 55 insertions(+), 14 deletions(-)
diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
index 4535c14c26a..12a2b57d9e3 100644
--- a/gcc/cp/pt.cc
+++ b/gcc/cp/pt.cc
@@ -18547,6 +18547,29 @@ lookup_init_capture_pack (tree decl)
return r;
}
+/* T is an operand of a template tree being substituted. Return whether
+ T is dependent such that we should suppress some warnings that would
+ make sense if the substituted expression were written directly, like
+ template <int I> bool f() { return I == 2; }
+ We don't want to warn when instantiating f that comparing two constants
+ always has the same value.
+
+ This is a more limited concept of dependence than instantiation-dependent;
+ here we don't care whether substitution could fail. */
+
+static bool
+dependent_operand_p (tree t)
+{
+ while (TREE_CODE (t) == IMPLICIT_CONV_EXPR)
+ t = TREE_OPERAND (t, 0);
+ ++processing_template_decl;
+ bool r = (potential_constant_expression (t)
+ ? value_dependent_expression_p (t)
+ : type_dependent_expression_p (t));
+ --processing_template_decl;
+ return r;
+}
+
/* Like tsubst_copy for expressions, etc. but also does semantic
processing. */
@@ -18872,8 +18895,13 @@ tsubst_expr (tree t, tree args, tsubst_flags_t complain, tree in_decl,
IF_STMT_CONSTEVAL_P (stmt) = IF_STMT_CONSTEVAL_P (t);
if (IF_STMT_CONSTEXPR_P (t))
args = add_extra_args (IF_STMT_EXTRA_ARGS (t), args, complain, in_decl);
- tmp = RECUR (IF_COND (t));
- tmp = finish_if_stmt_cond (tmp, stmt);
+ {
+ tree cond = IF_COND (t);
+ bool was_dep = dependent_operand_p (cond);
+ cond = RECUR (cond);
+ warning_sentinel s1(warn_address, was_dep);
+ tmp = finish_if_stmt_cond (cond, stmt);
+ }
if (IF_STMT_CONSTEXPR_P (t)
&& instantiation_dependent_expression_p (tmp))
{
@@ -20422,15 +20450,8 @@ tsubst_copy_and_build (tree t,
warnings that depend on the range of the types involved. */
tree op0 = TREE_OPERAND (t, 0);
tree op1 = TREE_OPERAND (t, 1);
- auto dep_p = [](tree t) {
- ++processing_template_decl;
- bool r = (potential_constant_expression (t)
- ? value_dependent_expression_p (t)
- : type_dependent_expression_p (t));
- --processing_template_decl;
- return r;
- };
- const bool was_dep = dep_p (op0) || dep_p (op1);
+ const bool was_dep = (dependent_operand_p (op0)
+ || dependent_operand_p (op1));
op0 = RECUR (op0);
op1 = RECUR (op1);
diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
index 24cca99e909..2344b5eea00 100644
--- a/gcc/cp/semantics.cc
+++ b/gcc/cp/semantics.cc
@@ -1029,9 +1029,9 @@ maybe_warn_for_constant_evaluated (tree cond, bool constexpr_if)
IF_STMT. */
tree
-finish_if_stmt_cond (tree cond, tree if_stmt)
+finish_if_stmt_cond (tree orig_cond, tree if_stmt)
{
- cond = maybe_convert_cond (cond);
+ tree cond = maybe_convert_cond (orig_cond);
if (IF_STMT_CONSTEXPR_P (if_stmt)
&& !type_dependent_expression_p (cond)
&& require_constant_expression (cond)
@@ -1045,7 +1045,11 @@ finish_if_stmt_cond (tree cond, tree if_stmt)
cond = cxx_constant_value (cond, NULL_TREE);
}
else
- maybe_warn_for_constant_evaluated (cond, /*constexpr_if=*/false);
+ {
+ maybe_warn_for_constant_evaluated (cond, /*constexpr_if=*/false);
+ if (processing_template_decl)
+ cond = orig_cond;
+ }
finish_cond (&IF_COND (if_stmt), cond);
add_stmt (if_stmt);
THEN_CLAUSE (if_stmt) = push_stmt_list ();
diff --git a/gcc/testsuite/g++.dg/cpp1z/constexpr-if38.C b/gcc/testsuite/g++.dg/cpp1z/constexpr-if38.C
new file mode 100644
index 00000000000..3e4fd446dca
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/constexpr-if38.C
@@ -0,0 +1,16 @@
+// PR c++/94554
+// { dg-do compile { target c++17 } }
+// { dg-additional-options -Wall }
+
+int meow() { return 1; }
+void kitty(int);
+template <int (*F)()>
+void test() {
+ if constexpr (F) {
+ kitty(F());
+ } else {
+ kitty(2);
+ }
+}
+template void test<nullptr>();
+template void test<meow>();
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2022-06-23 15:08 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-06-23 15:08 [gcc r13-1219] c++: -Waddress and if constexpr [PR94554] Jason Merrill
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).