public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] [PR85569] skip constexpr target_expr constructor dummy type conversion
@ 2018-11-22 23:39 Alexandre Oliva
  2018-11-27 23:41 ` Jason Merrill
  0 siblings, 1 reply; 4+ messages in thread
From: Alexandre Oliva @ 2018-11-22 23:39 UTC (permalink / raw)
  To: gcc-patches; +Cc: jason, nathan

The testcase is the work-around testcase for the PR; even that had
started failing.  The problem was that, when unqualifying the type of
a TARGET_EXPR, we'd create a variant of the type, then request the
conversion of the TARGET_EXPR_INITIAL to that variant type.  Though
the types are different pointer-wise, they're the same_type_p, so the
resulting modified expr compares cp_tree_equal to the original, which
maybe_constant_value flags as an error.  There's no reason to
construct an alternate TARGET_EXPR or CONSTRUCTOR just because of an
equivalent type, except for another spot that expected pointer
equality that would no longer be satisfied.  Without relaxing the
assert in constexpr_call_hasher::equal, g++.robertl/eb73.C would
trigger an assertion failure.

Regstrapped on i686- and x86_64-linux-gnu.  Ok to install?


for  gcc/cp/ChangeLog

	PR c++/85569
	* constexpr.c (adjust_temp_type): Test for type equality with
	same_type_p.

for  gcc/testsuite

	PR c++/85569
	* g++.dg/cpp1z/pr85569.C: New.
---
 gcc/cp/constexpr.c                   |    4 +
 gcc/testsuite/g++.dg/cpp1z/pr85569.C |   93 ++++++++++++++++++++++++++++++++++
 2 files changed, 95 insertions(+), 2 deletions(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp1z/pr85569.C

diff --git a/gcc/cp/constexpr.c b/gcc/cp/constexpr.c
index 92fd2b2d9d59..bb5d1301b332 100644
--- a/gcc/cp/constexpr.c
+++ b/gcc/cp/constexpr.c
@@ -1060,7 +1060,7 @@ constexpr_call_hasher::equal (constexpr_call *lhs, constexpr_call *rhs)
     {
       tree lhs_arg = TREE_VALUE (lhs_bindings);
       tree rhs_arg = TREE_VALUE (rhs_bindings);
-      gcc_assert (TREE_TYPE (lhs_arg) == TREE_TYPE (rhs_arg));
+      gcc_assert (same_type_p (TREE_TYPE (lhs_arg), TREE_TYPE (rhs_arg)));
       if (!cp_tree_equal (lhs_arg, rhs_arg))
         return false;
       lhs_bindings = TREE_CHAIN (lhs_bindings);
@@ -1276,7 +1276,7 @@ cxx_eval_builtin_function_call (const constexpr_ctx *ctx, tree t, tree fun,
 static tree
 adjust_temp_type (tree type, tree temp)
 {
-  if (TREE_TYPE (temp) == type)
+  if (TREE_TYPE (temp) == type || same_type_p (TREE_TYPE (temp), type))
     return temp;
   /* Avoid wrapping an aggregate value in a NOP_EXPR.  */
   if (TREE_CODE (temp) == CONSTRUCTOR)
diff --git a/gcc/testsuite/g++.dg/cpp1z/pr85569.C b/gcc/testsuite/g++.dg/cpp1z/pr85569.C
new file mode 100644
index 000000000000..aec543041a0f
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/pr85569.C
@@ -0,0 +1,93 @@
+// { dg-do compile { target c++17 } }
+
+#include <utility>
+#include <tuple>
+
+#define LIFT_FWD(x) std::forward<decltype(x)>(x)
+
+template <typename T>
+inline
+constexpr
+auto
+equal(
+  T &&t)
+{
+  return [t = std::forward<T>(t)](const auto& obj)
+    -> decltype(obj == t)
+    {
+      return obj == t;
+    };
+}
+
+template <typename F, typename T>
+struct is_tuple_invocable;
+
+template <typename F, typename ... Ts>
+struct is_tuple_invocable<F, std::tuple<Ts...>>
+{
+  using type = typename std::is_invocable<F, Ts...>::type;
+};
+
+template <typename F>
+inline
+constexpr
+auto
+compose(
+  F&& f
+)
+  noexcept
+-> F
+{
+  return std::forward<F>(f);
+}
+
+namespace detail {
+  template <typename F, typename Tail, typename ... T>
+  inline
+  constexpr
+  auto
+  compose(
+    std::true_type,
+    F&& f,
+    Tail&& tail,
+    T&& ... objs)
+  noexcept(noexcept(f(tail(std::forward<T>(objs)...))))
+  -> decltype(f(tail(std::forward<T>(objs)...)))
+  {
+    return f(tail(std::forward<T>(objs)...));
+  }
+}
+template <typename F, typename ... Fs>
+inline
+constexpr
+auto
+compose(
+  F&& f,
+  Fs&&... fs)
+{
+  return [f = std::forward<F>(f), tail = compose(std::forward<Fs>(fs)...)]
+    (auto&& ... objs)
+    -> decltype(detail::compose(typename std::is_invocable<decltype(compose(std::forward<Fs>(fs)...)), decltype(objs)...>::type{},
+                                f,
+                                compose(std::forward<Fs>(fs)...),
+                                LIFT_FWD(objs)...))
+  {
+    using tail_type = decltype(compose(std::forward<Fs>(fs)...));
+    
+#ifndef NOT_VIA_TUPLE
+    using args_type = std::tuple<decltype(objs)...>;
+    constexpr auto unitail = typename is_tuple_invocable<tail_type, args_type>::type{};
+#else
+    constexpr auto unitail = typename std::is_invocable<tail_type, decltype(objs)...>::type{};
+#endif
+
+    return detail::compose(unitail,  f, tail, LIFT_FWD(objs)...);
+  };
+}
+
+template <auto N>
+constexpr auto eq = equal(N);
+
+static_assert(compose(eq<3>,
+		      std::plus<>{})(1,2),
+              "compose is constexpr");

-- 
Alexandre Oliva, freedom fighter   https://FSFLA.org/blogs/lxo
Be the change, be Free!         FSF Latin America board member
GNU Toolchain Engineer                Free Software Evangelist
Hay que enGNUrecerse, pero sin perder la terGNUra jamás-GNUChe

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] [PR85569] skip constexpr target_expr constructor dummy type conversion
  2018-11-22 23:39 [PATCH] [PR85569] skip constexpr target_expr constructor dummy type conversion Alexandre Oliva
@ 2018-11-27 23:41 ` Jason Merrill
  2018-12-05  6:34   ` Alexandre Oliva
  0 siblings, 1 reply; 4+ messages in thread
From: Jason Merrill @ 2018-11-27 23:41 UTC (permalink / raw)
  To: Alexandre Oliva, gcc-patches; +Cc: nathan

On 11/22/18 6:39 PM, Alexandre Oliva wrote:
> The testcase is the work-around testcase for the PR; even that had
> started failing.  The problem was that, when unqualifying the type of
> a TARGET_EXPR, we'd create a variant of the type, then request the
> conversion of the TARGET_EXPR_INITIAL to that variant type.  Though
> the types are different pointer-wise, they're the same_type_p, so the
> resulting modified expr compares cp_tree_equal to the original, which
> maybe_constant_value flags as an error.  There's no reason to
> construct an alternate TARGET_EXPR or CONSTRUCTOR just because of an
> equivalent type, except for another spot that expected pointer
> equality that would no longer be satisfied.  Without relaxing the
> assert in constexpr_call_hasher::equal, g++.robertl/eb73.C would
> trigger an assertion failure.
> 
> Regstrapped on i686- and x86_64-linux-gnu.  Ok to install?
> 
> 
> for  gcc/cp/ChangeLog
> 
> 	PR c++/85569
> 	* constexpr.c (adjust_temp_type): Test for type equality with
> 	same_type_p.
> 
> for  gcc/testsuite
> 
> 	PR c++/85569
> 	* g++.dg/cpp1z/pr85569.C: New.
> ---
>   gcc/cp/constexpr.c                   |    4 +
>   gcc/testsuite/g++.dg/cpp1z/pr85569.C |   93 ++++++++++++++++++++++++++++++++++
>   2 files changed, 95 insertions(+), 2 deletions(-)
>   create mode 100644 gcc/testsuite/g++.dg/cpp1z/pr85569.C
> 
> diff --git a/gcc/cp/constexpr.c b/gcc/cp/constexpr.c
> index 92fd2b2d9d59..bb5d1301b332 100644
> --- a/gcc/cp/constexpr.c
> +++ b/gcc/cp/constexpr.c
> @@ -1060,7 +1060,7 @@ constexpr_call_hasher::equal (constexpr_call *lhs, constexpr_call *rhs)
>       {
>         tree lhs_arg = TREE_VALUE (lhs_bindings);
>         tree rhs_arg = TREE_VALUE (rhs_bindings);
> -      gcc_assert (TREE_TYPE (lhs_arg) == TREE_TYPE (rhs_arg));
> +      gcc_assert (same_type_p (TREE_TYPE (lhs_arg), TREE_TYPE (rhs_arg)));
>         if (!cp_tree_equal (lhs_arg, rhs_arg))
>           return false;
>         lhs_bindings = TREE_CHAIN (lhs_bindings);
> @@ -1276,7 +1276,7 @@ cxx_eval_builtin_function_call (const constexpr_ctx *ctx, tree t, tree fun,
>   static tree
>   adjust_temp_type (tree type, tree temp)
>   {
> -  if (TREE_TYPE (temp) == type)
> +  if (TREE_TYPE (temp) == type || same_type_p (TREE_TYPE (temp), type))
>       return temp;
>     /* Avoid wrapping an aggregate value in a NOP_EXPR.  */

Hmm, I'm a bit uneasy about this change, but it does make sense to 
follow cp_tree_equal.

Let's replace the == comparison rather than supplement it.  OK with that 
change.

Jason

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] [PR85569] skip constexpr target_expr constructor dummy type conversion
  2018-11-27 23:41 ` Jason Merrill
@ 2018-12-05  6:34   ` Alexandre Oliva
  2018-12-05  8:45     ` [committed] Add testcase for already fixed PR c++/87897 Jakub Jelinek
  0 siblings, 1 reply; 4+ messages in thread
From: Alexandre Oliva @ 2018-12-05  6:34 UTC (permalink / raw)
  To: Jason Merrill; +Cc: gcc-patches, nathan

On Nov 27, 2018, Jason Merrill <jason@redhat.com> wrote:

> Let's replace the == comparison rather than supplement it.  OK with
> that change.

Thanks, here's what I (re)tested and will install eventually.


[PR85569] skip constexpr target_expr constructor dummy type conversion

From: Alexandre Oliva <aoliva@redhat.com>

The testcase is the work-around testcase for the PR; even that had
started failing.  The problem was that, when unqualifying the type of
a TARGET_EXPR, we'd create a variant of the type, then request the
conversion of the TARGET_EXPR_INITIAL to that variant type.  Though
the types are different pointer-wise, they're the same_type_p, so the
resulting modified expr compares cp_tree_equal to the original, which
maybe_constant_value flags as an error.  There's no reason to
construct an alternate TARGET_EXPR or CONSTRUCTOR just because of an
equivalent type, except for another spot that expected pointer
equality that would no longer be satisfied.  Without relaxing the
assert in constexpr_call_hasher::equal, g++.robertl/eb73.C would
trigger an assertion failure.


for  gcc/cp/ChangeLog

	PR c++/85569
	* constexpr.c (adjust_temp_type): Test for type equality with
	same_type_p.
	(constexpr_call_hasher::equal): Likewise.

for  gcc/testsuite

	PR c++/85569
	* g++.dg/cpp1z/pr85569.C: New.
---
 gcc/cp/constexpr.c                   |    4 +
 gcc/testsuite/g++.dg/cpp1z/pr85569.C |   93 ++++++++++++++++++++++++++++++++++
 2 files changed, 95 insertions(+), 2 deletions(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp1z/pr85569.C

diff --git a/gcc/cp/constexpr.c b/gcc/cp/constexpr.c
index 92fd2b2d9d59..a668d14e8bf5 100644
--- a/gcc/cp/constexpr.c
+++ b/gcc/cp/constexpr.c
@@ -1060,7 +1060,7 @@ constexpr_call_hasher::equal (constexpr_call *lhs, constexpr_call *rhs)
     {
       tree lhs_arg = TREE_VALUE (lhs_bindings);
       tree rhs_arg = TREE_VALUE (rhs_bindings);
-      gcc_assert (TREE_TYPE (lhs_arg) == TREE_TYPE (rhs_arg));
+      gcc_assert (same_type_p (TREE_TYPE (lhs_arg), TREE_TYPE (rhs_arg)));
       if (!cp_tree_equal (lhs_arg, rhs_arg))
         return false;
       lhs_bindings = TREE_CHAIN (lhs_bindings);
@@ -1276,7 +1276,7 @@ cxx_eval_builtin_function_call (const constexpr_ctx *ctx, tree t, tree fun,
 static tree
 adjust_temp_type (tree type, tree temp)
 {
-  if (TREE_TYPE (temp) == type)
+  if (same_type_p (TREE_TYPE (temp), type))
     return temp;
   /* Avoid wrapping an aggregate value in a NOP_EXPR.  */
   if (TREE_CODE (temp) == CONSTRUCTOR)
diff --git a/gcc/testsuite/g++.dg/cpp1z/pr85569.C b/gcc/testsuite/g++.dg/cpp1z/pr85569.C
new file mode 100644
index 000000000000..aec543041a0f
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/pr85569.C
@@ -0,0 +1,93 @@
+// { dg-do compile { target c++17 } }
+
+#include <utility>
+#include <tuple>
+
+#define LIFT_FWD(x) std::forward<decltype(x)>(x)
+
+template <typename T>
+inline
+constexpr
+auto
+equal(
+  T &&t)
+{
+  return [t = std::forward<T>(t)](const auto& obj)
+    -> decltype(obj == t)
+    {
+      return obj == t;
+    };
+}
+
+template <typename F, typename T>
+struct is_tuple_invocable;
+
+template <typename F, typename ... Ts>
+struct is_tuple_invocable<F, std::tuple<Ts...>>
+{
+  using type = typename std::is_invocable<F, Ts...>::type;
+};
+
+template <typename F>
+inline
+constexpr
+auto
+compose(
+  F&& f
+)
+  noexcept
+-> F
+{
+  return std::forward<F>(f);
+}
+
+namespace detail {
+  template <typename F, typename Tail, typename ... T>
+  inline
+  constexpr
+  auto
+  compose(
+    std::true_type,
+    F&& f,
+    Tail&& tail,
+    T&& ... objs)
+  noexcept(noexcept(f(tail(std::forward<T>(objs)...))))
+  -> decltype(f(tail(std::forward<T>(objs)...)))
+  {
+    return f(tail(std::forward<T>(objs)...));
+  }
+}
+template <typename F, typename ... Fs>
+inline
+constexpr
+auto
+compose(
+  F&& f,
+  Fs&&... fs)
+{
+  return [f = std::forward<F>(f), tail = compose(std::forward<Fs>(fs)...)]
+    (auto&& ... objs)
+    -> decltype(detail::compose(typename std::is_invocable<decltype(compose(std::forward<Fs>(fs)...)), decltype(objs)...>::type{},
+                                f,
+                                compose(std::forward<Fs>(fs)...),
+                                LIFT_FWD(objs)...))
+  {
+    using tail_type = decltype(compose(std::forward<Fs>(fs)...));
+    
+#ifndef NOT_VIA_TUPLE
+    using args_type = std::tuple<decltype(objs)...>;
+    constexpr auto unitail = typename is_tuple_invocable<tail_type, args_type>::type{};
+#else
+    constexpr auto unitail = typename std::is_invocable<tail_type, decltype(objs)...>::type{};
+#endif
+
+    return detail::compose(unitail,  f, tail, LIFT_FWD(objs)...);
+  };
+}
+
+template <auto N>
+constexpr auto eq = equal(N);
+
+static_assert(compose(eq<3>,
+		      std::plus<>{})(1,2),
+              "compose is constexpr");


-- 
Alexandre Oliva, freedom fighter   https://FSFLA.org/blogs/lxo
Be the change, be Free!         FSF Latin America board member
GNU Toolchain Engineer                Free Software Evangelist
Hay que enGNUrecerse, pero sin perder la terGNUra jamás-GNUChe

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [committed] Add testcase for already fixed PR c++/87897
  2018-12-05  6:34   ` Alexandre Oliva
@ 2018-12-05  8:45     ` Jakub Jelinek
  0 siblings, 0 replies; 4+ messages in thread
From: Jakub Jelinek @ 2018-12-05  8:45 UTC (permalink / raw)
  To: Alexandre Oliva; +Cc: Jason Merrill, gcc-patches, nathan

On Wed, Dec 05, 2018 at 04:34:14AM -0200, Alexandre Oliva wrote:
> for  gcc/cp/ChangeLog
> 
> 	PR c++/85569
> 	* constexpr.c (adjust_temp_type): Test for type equality with
> 	same_type_p.
> 	(constexpr_call_hasher::equal): Likewise.
> 
> for  gcc/testsuite
> 
> 	PR c++/85569
> 	* g++.dg/cpp1z/pr85569.C: New.

This patch fixed also PR87897, I've added the following testcase from it,
regtested on x86_64-linux and committed as obvious:

2018-12-05  Jakub Jelinek  <jakub@redhat.com>

	PR c++/87897
	* g++.dg/init/const13.C: New test.

--- gcc/testsuite/g++.dg/init/const13.C.jj	2018-12-05 09:39:04.604974302 +0100
+++ gcc/testsuite/g++.dg/init/const13.C	2018-12-05 09:38:55.961117120 +0100
@@ -0,0 +1,5 @@
+// PR c++/87897
+// { dg-do compile }
+
+typedef struct A {} B;
+A const a = B ();


	Jakub

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2018-12-05  8:45 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-11-22 23:39 [PATCH] [PR85569] skip constexpr target_expr constructor dummy type conversion Alexandre Oliva
2018-11-27 23:41 ` Jason Merrill
2018-12-05  6:34   ` Alexandre Oliva
2018-12-05  8:45     ` [committed] Add testcase for already fixed PR c++/87897 Jakub Jelinek

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).