diff mbox series

c++: decltype of by-ref capture proxy of ref [PR115504]

Message ID 20240625010015.122204-1-ppalka@redhat.com
State New
Headers show
Series c++: decltype of by-ref capture proxy of ref [PR115504] | expand

Commit Message

Patrick Palka June 25, 2024, 1 a.m. UTC
Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
for trunk/14?

-- >8 --

The capture proxy handling in finish_decltype_type added in r14-5330
was stripping the reference type of a capture proxy's captured variable,
which is desirable for a by-value capture, but not for a by-ref capture
(of a reference).

	PR c++/115504

gcc/cp/ChangeLog:

	* semantics.cc (finish_decltype_type): For a by-reference
	capture proxy, don't strip the reference type (if any) of
	the captured variable.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp1y/decltype-auto8.C: New test.
---
 gcc/cp/semantics.cc                         |  4 +++-
 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C | 11 +++++++++++
 2 files changed, 14 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C

Comments

Jason Merrill June 25, 2024, 1:48 a.m. UTC | #1
On 6/24/24 21:00, Patrick Palka wrote:
> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
> for trunk/14?
> 
> -- >8 --
> 
> The capture proxy handling in finish_decltype_type added in r14-5330
> was stripping the reference type of a capture proxy's captured variable,
> which is desirable for a by-value capture, but not for a by-ref capture
> (of a reference).

I'm not sure why we would want it for by-value, either; regardless of 
the capture kind, decltype(x) is int&.

> 	PR c++/115504
> 
> gcc/cp/ChangeLog:
> 
> 	* semantics.cc (finish_decltype_type): For a by-reference
> 	capture proxy, don't strip the reference type (if any) of
> 	the captured variable.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/cpp1y/decltype-auto8.C: New test.
> ---
>   gcc/cp/semantics.cc                         |  4 +++-
>   gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C | 11 +++++++++++
>   2 files changed, 14 insertions(+), 1 deletion(-)
>   create mode 100644 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> 
> diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
> index 08f5f245e7d..b4f626924af 100644
> --- a/gcc/cp/semantics.cc
> +++ b/gcc/cp/semantics.cc
> @@ -12076,9 +12076,11 @@ finish_decltype_type (tree expr, bool id_expression_or_member_access_p,
>   	    {
>   	      if (is_normal_capture_proxy (expr))
>   		{
> +		  bool by_ref = TYPE_REF_P (TREE_TYPE (expr));
>   		  expr = DECL_CAPTURED_VARIABLE (expr);
>   		  type = TREE_TYPE (expr);
> -		  type = non_reference (type);
> +		  if (!by_ref)
> +		    type = non_reference (type);
>   		}
>   	      else
>   		{
> diff --git a/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> new file mode 100644
> index 00000000000..9a5e435f14f
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> @@ -0,0 +1,11 @@
> +// PR c++/115504
> +// { dg-do compile { target c++14 } }
> +
> +void f(int& x) {
> +  [&x]() {
> +    decltype(auto) a = x;
> +    using type = decltype(x);
> +    using type = decltype(a);
> +    using type = int&; // not 'int'
> +  };
> +}
Patrick Palka June 25, 2024, 3:03 p.m. UTC | #2
On Mon, 24 Jun 2024, Jason Merrill wrote:

> On 6/24/24 21:00, Patrick Palka wrote:
> > Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
> > for trunk/14?
> > 
> > -- >8 --
> > 
> > The capture proxy handling in finish_decltype_type added in r14-5330
> > was stripping the reference type of a capture proxy's captured variable,
> > which is desirable for a by-value capture, but not for a by-ref capture
> > (of a reference).
> 
> I'm not sure why we would want it for by-value, either; regardless of the
> capture kind, decltype(x) is int&.

Ah, makes sense.  But I guess that means

  void f(int& x) {
    [x]() {
      decltype(auto) a = x;
    }
  }

is ill-formed since decltype(x) is int& but the corresponding closure
member is const?  It works if we make the lambda mutable.

Like so?  Bootstrapped and regtested on x86_64-pc-linux-gnu.


-- >8 --

Subject: [PATCH] c++: decltype of capture proxy of ref [PR115504]

The capture proxy handling in finish_decltype_type added in r14-5330
was stripping the reference type of a capture proxy's captured variable.

	PR c++/115504

gcc/cp/ChangeLog:

	* semantics.cc (finish_decltype_type): Don't strip the reference
	type (if any) of a capture proxy's captured variable.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp1y/decltype-auto8.C: New test.
---
 gcc/cp/semantics.cc                         |  1 -
 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C | 18 ++++++++++++++++++
 2 files changed, 18 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C

diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
index 6c1813d37c6..6a383c0f7f9 100644
--- a/gcc/cp/semantics.cc
+++ b/gcc/cp/semantics.cc
@@ -12071,7 +12071,6 @@ finish_decltype_type (tree expr, bool id_expression_or_member_access_p,
 		{
 		  expr = DECL_CAPTURED_VARIABLE (expr);
 		  type = TREE_TYPE (expr);
-		  type = non_reference (type);
 		}
 	      else
 		{
diff --git a/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
new file mode 100644
index 00000000000..1100b94a5b7
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
@@ -0,0 +1,18 @@
+// PR c++/115504
+// { dg-do compile { target c++14 } }
+
+void f(int& x) {
+  [&x]() {
+    decltype(auto) a = x;
+    using type = decltype(x);
+    using type = decltype(a);
+    using type = int&; // not 'int'
+  };
+
+  [x]() mutable {
+    decltype(auto) a = x;
+    using type = decltype(x);
+    using type = decltype(a);
+    using type = int&;
+  };
+}
Jason Merrill June 25, 2024, 3:31 p.m. UTC | #3
On 6/25/24 11:03, Patrick Palka wrote:
> On Mon, 24 Jun 2024, Jason Merrill wrote:
> 
>> On 6/24/24 21:00, Patrick Palka wrote:
>>> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
>>> for trunk/14?
>>>
>>> -- >8 --
>>>
>>> The capture proxy handling in finish_decltype_type added in r14-5330
>>> was stripping the reference type of a capture proxy's captured variable,
>>> which is desirable for a by-value capture, but not for a by-ref capture
>>> (of a reference).
>>
>> I'm not sure why we would want it for by-value, either; regardless of the
>> capture kind, decltype(x) is int&.
> 
> Ah, makes sense.  But I guess that means
> 
>    void f(int& x) {
>      [x]() {
>        decltype(auto) a = x;
>      }
>    }
> 
> is ill-formed since decltype(x) is int& but the corresponding closure
> member is const?  It works if we make the lambda mutable.

Yes, and clang agrees.  Let's also test that case.

> Like so?  Bootstrapped and regtested on x86_64-pc-linux-gnu.
> 
> 
> -- >8 --
> 
> Subject: [PATCH] c++: decltype of capture proxy of ref [PR115504]
> 
> The capture proxy handling in finish_decltype_type added in r14-5330
> was stripping the reference type of a capture proxy's captured variable.
> 
> 	PR c++/115504
> 
> gcc/cp/ChangeLog:
> 
> 	* semantics.cc (finish_decltype_type): Don't strip the reference
> 	type (if any) of a capture proxy's captured variable.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/cpp1y/decltype-auto8.C: New test.
> ---
>   gcc/cp/semantics.cc                         |  1 -
>   gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C | 18 ++++++++++++++++++
>   2 files changed, 18 insertions(+), 1 deletion(-)
>   create mode 100644 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> 
> diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
> index 6c1813d37c6..6a383c0f7f9 100644
> --- a/gcc/cp/semantics.cc
> +++ b/gcc/cp/semantics.cc
> @@ -12071,7 +12071,6 @@ finish_decltype_type (tree expr, bool id_expression_or_member_access_p,
>   		{
>   		  expr = DECL_CAPTURED_VARIABLE (expr);
>   		  type = TREE_TYPE (expr);
> -		  type = non_reference (type);
>   		}
>   	      else
>   		{
> diff --git a/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> new file mode 100644
> index 00000000000..1100b94a5b7
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> @@ -0,0 +1,18 @@
> +// PR c++/115504
> +// { dg-do compile { target c++14 } }
> +
> +void f(int& x) {
> +  [&x]() {
> +    decltype(auto) a = x;
> +    using type = decltype(x);
> +    using type = decltype(a);
> +    using type = int&; // not 'int'
> +  };
> +
> +  [x]() mutable {
> +    decltype(auto) a = x;
> +    using type = decltype(x);
> +    using type = decltype(a);
> +    using type = int&;
> +  };
> +}
Patrick Palka June 25, 2024, 3:45 p.m. UTC | #4
On Tue, 25 Jun 2024, Jason Merrill wrote:

> On 6/25/24 11:03, Patrick Palka wrote:
> > On Mon, 24 Jun 2024, Jason Merrill wrote:
> > 
> > > On 6/24/24 21:00, Patrick Palka wrote:
> > > > Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
> > > > for trunk/14?
> > > > 
> > > > -- >8 --
> > > > 
> > > > The capture proxy handling in finish_decltype_type added in r14-5330
> > > > was stripping the reference type of a capture proxy's captured variable,
> > > > which is desirable for a by-value capture, but not for a by-ref capture
> > > > (of a reference).
> > > 
> > > I'm not sure why we would want it for by-value, either; regardless of the
> > > capture kind, decltype(x) is int&.
> > 
> > Ah, makes sense.  But I guess that means
> > 
> >    void f(int& x) {
> >      [x]() {
> >        decltype(auto) a = x;
> >      }
> >    }
> > 
> > is ill-formed since decltype(x) is int& but the corresponding closure
> > member is const?  It works if we make the lambda mutable.
> 
> Yes, and clang agrees.  Let's also test that case.

Like so?

-- >8 --

Subject: [PATCH] c++: decltype of capture proxy of ref [PR115504]

The capture proxy handling in finish_decltype_type added in r14-5330
was stripping the reference type of a capture proxy's captured variable.

	PR c++/115504

gcc/cp/ChangeLog:

	* semantics.cc (finish_decltype_type): Don't strip the reference
	type (if any) of a capture proxy's captured variable.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp1y/decltype-auto8.C: New test.
---
 gcc/cp/semantics.cc                         |  1 -
 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C | 22 +++++++++++++++++++++
 2 files changed, 22 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C

diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
index 6c1813d37c6..6a383c0f7f9 100644
--- a/gcc/cp/semantics.cc
+++ b/gcc/cp/semantics.cc
@@ -12071,7 +12071,6 @@ finish_decltype_type (tree expr, bool id_expression_or_member_access_p,
 		{
 		  expr = DECL_CAPTURED_VARIABLE (expr);
 		  type = TREE_TYPE (expr);
-		  type = non_reference (type);
 		}
 	      else
 		{
diff --git a/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
new file mode 100644
index 00000000000..55135cecf72
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
@@ -0,0 +1,22 @@
+// PR c++/115504
+// { dg-do compile { target c++14 } }
+
+void f(int& x, const int& y) {
+  [&x]() {
+    decltype(auto) a = x;
+    using type = decltype(x);
+    using type = decltype(a);
+    using type = int&; // not 'int'
+  };
+
+  [x]() {
+    decltype(auto) a = x; // { dg-error "discards qualifiers" }
+  };
+
+  [x]() mutable {
+    decltype(auto) a = x;
+    using type = decltype(x);
+    using type = decltype(a);
+    using type = int&;
+  };
+}
Jason Merrill June 25, 2024, 9:24 p.m. UTC | #5
On 6/25/24 11:45, Patrick Palka wrote:
> On Tue, 25 Jun 2024, Jason Merrill wrote:
> 
>> On 6/25/24 11:03, Patrick Palka wrote:
>>> On Mon, 24 Jun 2024, Jason Merrill wrote:
>>>
>>>> On 6/24/24 21:00, Patrick Palka wrote:
>>>>> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
>>>>> for trunk/14?
>>>>>
>>>>> -- >8 --
>>>>>
>>>>> The capture proxy handling in finish_decltype_type added in r14-5330
>>>>> was stripping the reference type of a capture proxy's captured variable,
>>>>> which is desirable for a by-value capture, but not for a by-ref capture
>>>>> (of a reference).
>>>>
>>>> I'm not sure why we would want it for by-value, either; regardless of the
>>>> capture kind, decltype(x) is int&.
>>>
>>> Ah, makes sense.  But I guess that means
>>>
>>>     void f(int& x) {
>>>       [x]() {
>>>         decltype(auto) a = x;
>>>       }
>>>     }
>>>
>>> is ill-formed since decltype(x) is int& but the corresponding closure
>>> member is const?  It works if we make the lambda mutable.
>>
>> Yes, and clang agrees.  Let's also test that case.
> 
> Like so?

OK.

> -- >8 --
> 
> Subject: [PATCH] c++: decltype of capture proxy of ref [PR115504]
> 
> The capture proxy handling in finish_decltype_type added in r14-5330
> was stripping the reference type of a capture proxy's captured variable.
> 
> 	PR c++/115504
> 
> gcc/cp/ChangeLog:
> 
> 	* semantics.cc (finish_decltype_type): Don't strip the reference
> 	type (if any) of a capture proxy's captured variable.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/cpp1y/decltype-auto8.C: New test.
> ---
>   gcc/cp/semantics.cc                         |  1 -
>   gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C | 22 +++++++++++++++++++++
>   2 files changed, 22 insertions(+), 1 deletion(-)
>   create mode 100644 gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> 
> diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
> index 6c1813d37c6..6a383c0f7f9 100644
> --- a/gcc/cp/semantics.cc
> +++ b/gcc/cp/semantics.cc
> @@ -12071,7 +12071,6 @@ finish_decltype_type (tree expr, bool id_expression_or_member_access_p,
>   		{
>   		  expr = DECL_CAPTURED_VARIABLE (expr);
>   		  type = TREE_TYPE (expr);
> -		  type = non_reference (type);
>   		}
>   	      else
>   		{
> diff --git a/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> new file mode 100644
> index 00000000000..55135cecf72
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
> @@ -0,0 +1,22 @@
> +// PR c++/115504
> +// { dg-do compile { target c++14 } }
> +
> +void f(int& x, const int& y) {
> +  [&x]() {
> +    decltype(auto) a = x;
> +    using type = decltype(x);
> +    using type = decltype(a);
> +    using type = int&; // not 'int'
> +  };
> +
> +  [x]() {
> +    decltype(auto) a = x; // { dg-error "discards qualifiers" }
> +  };
> +
> +  [x]() mutable {
> +    decltype(auto) a = x;
> +    using type = decltype(x);
> +    using type = decltype(a);
> +    using type = int&;
> +  };
> +}
diff mbox series

Patch

diff --git a/gcc/cp/semantics.cc b/gcc/cp/semantics.cc
index 08f5f245e7d..b4f626924af 100644
--- a/gcc/cp/semantics.cc
+++ b/gcc/cp/semantics.cc
@@ -12076,9 +12076,11 @@  finish_decltype_type (tree expr, bool id_expression_or_member_access_p,
 	    {
 	      if (is_normal_capture_proxy (expr))
 		{
+		  bool by_ref = TYPE_REF_P (TREE_TYPE (expr));
 		  expr = DECL_CAPTURED_VARIABLE (expr);
 		  type = TREE_TYPE (expr);
-		  type = non_reference (type);
+		  if (!by_ref)
+		    type = non_reference (type);
 		}
 	      else
 		{
diff --git a/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
new file mode 100644
index 00000000000..9a5e435f14f
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1y/decltype-auto8.C
@@ -0,0 +1,11 @@ 
+// PR c++/115504
+// { dg-do compile { target c++14 } }
+
+void f(int& x) {
+  [&x]() {
+    decltype(auto) a = x;
+    using type = decltype(x);
+    using type = decltype(a);
+    using type = int&; // not 'int'
+  };
+}