| Message ID | CADzB+2kbdcbJ65BZsSKPw1bdmBOVwX--4Vdsznh62jtB+Sspaw@mail.gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | C++ PATCHes to xvalue handling | expand |
On 23/05/18 18:21, Jason Merrill wrote: > The first patch implements the adjustments from core issues 616 and > 1213 to the value category of subobjects of class prvalues: they were > considered prvalues themselves, but that was kind of nonsensical. Now > they are considered xvalues. Along with this, I've removed the > diagnostic distinction between xvalues and prvalues when trying to use > one or the other as an lvalue; the important thing is that they are > rvalues. > > The second patch corrects various issues with casts and xvalues/rvalue > references: we were treating an xvalue operand to dynamic_cast as an > lvalue, and we were objecting to casts from prvalue to rvalue > reference type. > With the second patch: commit f7d2790049fd1e59af4b69ee12f7c101cfe4cdab Author: jason <jason@138bc75d-0d04-0410-961f-82ee72b054a4> Date: Wed May 23 17:21:39 2018 +0000 Fix cast to rvalue reference from prvalue. * cvt.c (diagnose_ref_binding): Handle rvalue reference. * rtti.c (build_dynamic_cast_1): Don't try to build a reference to non-class type. Handle xvalue argument. * typeck.c (build_reinterpret_cast_1): Allow cast from prvalue to rvalue reference. * semantics.c (finish_compound_literal): Do direct-initialization, not cast, to initialize a reference. git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@260622 138bc75d-0d04-0410-961f-82ee72b054a4 I have observed the following failure in Spec2017 while building 510.parest_r on aarch64-none-linux-gnu aarch64-none-linux-gnu-g++ -c -o source/numerics/matrices.all_dimensions.o -DSPEC -DNDEBUG -Iinclude -I. -DSPEC_AUTO_SUPPRESS_OPENMP -mcpu=cortex-a57+crypto -Ofast -fomit-frame-pointer -fpermissive -DSPEC_LP64 source/numerics/matrices.all_dimensions.cc source/numerics/matrices.all_dimensions.cc: In static member function 'static void dealii::MatrixTools::apply_boundary_values(const std::map<unsigned int, double>&, dealii::BlockSparseMatrix<number>&, dealii::BlockVector<number>&, dealii::BlockVector<number>&, bool)': source/numerics/matrices.all_dimensions.cc:469:50: error: lvalue required as unary '&' operand [this_sparsity.get_rowstart_indices()[row]]; ^ source/numerics/matrices.all_dimensions.cc:472:55: error: lvalue required as unary '&' operand [this_sparsity.get_rowstart_indices()[row]+1], ^ source/numerics/matrices.all_dimensions.cc:474:55: error: lvalue required as unary '&' operand [this_sparsity.get_rowstart_indices()[row+1]], ^ source/numerics/matrices.all_dimensions.cc:479:49: error: lvalue required as unary '&' operand [this_sparsity.get_rowstart_indices()[row]], ^ source/numerics/matrices.all_dimensions.cc:481:51: error: lvalue required as unary '&' operand [this_sparsity.get_rowstart_indices()[row+1]], ^ source/numerics/matrices.all_dimensions.cc:510:50: error: lvalue required as unary '&' operand [this_sparsity.get_rowstart_indices()[0]]); Sudi > Tested x86_64-pc-linux-gnu, applying to trunk. >
On Fri, May 25, 2018 at 12:40 PM, Sudakshina Das <sudi.das@arm.com> wrote: > On 23/05/18 18:21, Jason Merrill wrote: >> >> The first patch implements the adjustments from core issues 616 and >> 1213 to the value category of subobjects of class prvalues: they were >> considered prvalues themselves, but that was kind of nonsensical. Now >> they are considered xvalues. Along with this, I've removed the >> diagnostic distinction between xvalues and prvalues when trying to use >> one or the other as an lvalue; the important thing is that they are >> rvalues. >> >> The second patch corrects various issues with casts and xvalues/rvalue >> references: we were treating an xvalue operand to dynamic_cast as an >> lvalue, and we were objecting to casts from prvalue to rvalue >> reference type. >> > > With the second patch: > commit f7d2790049fd1e59af4b69ee12f7c101cfe4cdab > Author: jason <jason@138bc75d-0d04-0410-961f-82ee72b054a4> > Date: Wed May 23 17:21:39 2018 +0000 > > Fix cast to rvalue reference from prvalue. > > * cvt.c (diagnose_ref_binding): Handle rvalue reference. > * rtti.c (build_dynamic_cast_1): Don't try to build a reference to > non-class type. Handle xvalue argument. > * typeck.c (build_reinterpret_cast_1): Allow cast from prvalue to > rvalue reference. > * semantics.c (finish_compound_literal): Do direct-initialization, > not cast, to initialize a reference. > > git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@260622 > 138bc75d-0d04-0410-961f-82ee72b054a4 > > I have observed the following failure in Spec2017 while building > 510.parest_r on aarch64-none-linux-gnu > > aarch64-none-linux-gnu-g++ -c -o source/numerics/matrices.all_dimensions.o > -DSPEC -DNDEBUG -Iinclude -I. -DSPEC_AUTO_SUPPRESS_OPENMP > -mcpu=cortex-a57+crypto -Ofast -fomit-frame-pointer -fpermissive > -DSPEC_LP64 source/numerics/matrices.all_dimensions.cc > > source/numerics/matrices.all_dimensions.cc: In static member function > 'static void dealii::MatrixTools::apply_boundary_values(const > std::map<unsigned int, double>&, dealii::BlockSparseMatrix<number>&, > dealii::BlockVector<number>&, dealii::BlockVector<number>&, bool)': > > source/numerics/matrices.all_dimensions.cc:469:50: error: lvalue required as > unary '&' operand > > [this_sparsity.get_rowstart_indices()[row]]; > > ^ > > source/numerics/matrices.all_dimensions.cc:472:55: error: lvalue required as > unary '&' operand > > [this_sparsity.get_rowstart_indices()[row]+1], > > ^ > > source/numerics/matrices.all_dimensions.cc:474:55: error: lvalue required as > unary '&' operand > > [this_sparsity.get_rowstart_indices()[row+1]], > > ^ > > source/numerics/matrices.all_dimensions.cc:479:49: error: lvalue required as > unary '&' operand > > [this_sparsity.get_rowstart_indices()[row]], > > ^ > > source/numerics/matrices.all_dimensions.cc:481:51: error: lvalue required as > unary '&' operand > > [this_sparsity.get_rowstart_indices()[row+1]], > > ^ > > source/numerics/matrices.all_dimensions.cc:510:50: error: lvalue required as > unary '&' operand > > [this_sparsity.get_rowstart_indices()[0]]); > > Sudi Thanks, investigating. Jason
On Fri, May 25, 2018 at 4:08 PM, Jason Merrill <jason@redhat.com> wrote: > On Fri, May 25, 2018 at 12:40 PM, Sudakshina Das <sudi.das@arm.com> wrote: >> On 23/05/18 18:21, Jason Merrill wrote: >>> >>> The first patch implements the adjustments from core issues 616 and >>> 1213 to the value category of subobjects of class prvalues: they were >>> considered prvalues themselves, but that was kind of nonsensical. Now >>> they are considered xvalues. Along with this, I've removed the >>> diagnostic distinction between xvalues and prvalues when trying to use >>> one or the other as an lvalue; the important thing is that they are >>> rvalues. >>> >>> The second patch corrects various issues with casts and xvalues/rvalue >>> references: we were treating an xvalue operand to dynamic_cast as an >>> lvalue, and we were objecting to casts from prvalue to rvalue >>> reference type. >>> >> >> With the second patch: >> commit f7d2790049fd1e59af4b69ee12f7c101cfe4cdab >> Author: jason <jason@138bc75d-0d04-0410-961f-82ee72b054a4> >> Date: Wed May 23 17:21:39 2018 +0000 >> >> Fix cast to rvalue reference from prvalue. >> >> * cvt.c (diagnose_ref_binding): Handle rvalue reference. >> * rtti.c (build_dynamic_cast_1): Don't try to build a reference to >> non-class type. Handle xvalue argument. >> * typeck.c (build_reinterpret_cast_1): Allow cast from prvalue to >> rvalue reference. >> * semantics.c (finish_compound_literal): Do direct-initialization, >> not cast, to initialize a reference. >> >> git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@260622 >> 138bc75d-0d04-0410-961f-82ee72b054a4 >> >> I have observed the following failure in Spec2017 while building >> 510.parest_r on aarch64-none-linux-gnu >> >> aarch64-none-linux-gnu-g++ -c -o source/numerics/matrices.all_dimensions.o >> -DSPEC -DNDEBUG -Iinclude -I. -DSPEC_AUTO_SUPPRESS_OPENMP >> -mcpu=cortex-a57+crypto -Ofast -fomit-frame-pointer -fpermissive >> -DSPEC_LP64 source/numerics/matrices.all_dimensions.cc >> >> source/numerics/matrices.all_dimensions.cc: In static member function >> 'static void dealii::MatrixTools::apply_boundary_values(const >> std::map<unsigned int, double>&, dealii::BlockSparseMatrix<number>&, >> dealii::BlockVector<number>&, dealii::BlockVector<number>&, bool)': >> >> source/numerics/matrices.all_dimensions.cc:469:50: error: lvalue required as >> unary '&' operand >> >> [this_sparsity.get_rowstart_indices()[row]]; >> >> ^ >> >> source/numerics/matrices.all_dimensions.cc:472:55: error: lvalue required as >> unary '&' operand >> >> [this_sparsity.get_rowstart_indices()[row]+1], >> >> ^ >> >> source/numerics/matrices.all_dimensions.cc:474:55: error: lvalue required as >> unary '&' operand >> >> [this_sparsity.get_rowstart_indices()[row+1]], >> >> ^ >> >> source/numerics/matrices.all_dimensions.cc:479:49: error: lvalue required as >> unary '&' operand >> >> [this_sparsity.get_rowstart_indices()[row]], >> >> ^ >> >> source/numerics/matrices.all_dimensions.cc:481:51: error: lvalue required as >> unary '&' operand >> >> [this_sparsity.get_rowstart_indices()[row+1]], >> >> ^ >> >> source/numerics/matrices.all_dimensions.cc:510:50: error: lvalue required as >> unary '&' operand >> >> [this_sparsity.get_rowstart_indices()[0]]); >> >> Sudi > > Thanks, investigating. Fixed thus. commit 2c71bc15d22c4b831ff2e27ef7759de4dc78ef6c Author: Jason Merrill <jason@redhat.com> Date: Fri May 25 16:25:10 2018 -0400 CWG 616, 1213 - value category of subobject references. * tree.c (lvalue_kind): Fix handling of ARRAY_REF of pointer. diff --git a/gcc/cp/tree.c b/gcc/cp/tree.c index 9d978160292..f21daaca1d0 100644 --- a/gcc/cp/tree.c +++ b/gcc/cp/tree.c @@ -95,14 +95,24 @@ lvalue_kind (const_tree ref) case TRY_CATCH_EXPR: case REALPART_EXPR: case IMAGPART_EXPR: - case ARRAY_REF: case VIEW_CONVERT_EXPR: - op1_lvalue_kind = lvalue_kind (TREE_OPERAND (ref, 0)); - if (op1_lvalue_kind == clk_class) - /* in the case of an array operand, the result is an lvalue if that - operand is an lvalue and an xvalue otherwise */ - op1_lvalue_kind = clk_rvalueref; - return op1_lvalue_kind; + return lvalue_kind (TREE_OPERAND (ref, 0)); + + case ARRAY_REF: + { + tree op1 = TREE_OPERAND (ref, 0); + if (TREE_CODE (TREE_TYPE (op1)) == ARRAY_TYPE) + { + op1_lvalue_kind = lvalue_kind (op1); + if (op1_lvalue_kind == clk_class) + /* in the case of an array operand, the result is an lvalue if + that operand is an lvalue and an xvalue otherwise */ + op1_lvalue_kind = clk_rvalueref; + return op1_lvalue_kind; + } + else + return clk_ordinary; + } case MEMBER_REF: case DOTSTAR_EXPR: diff --git a/gcc/testsuite/g++.dg/template/array31.C b/gcc/testsuite/g++.dg/template/array31.C new file mode 100644 index 00000000000..007d0dccf89 --- /dev/null +++ b/gcc/testsuite/g++.dg/template/array31.C @@ -0,0 +1,7 @@ +int *g(); + +template <class T> +void f(int i) +{ + int *p = &g()[3]; +}
Hi Jason On 25/05/18 21:55, Jason Merrill wrote: > On Fri, May 25, 2018 at 4:08 PM, Jason Merrill <jason@redhat.com> wrote: >> On Fri, May 25, 2018 at 12:40 PM, Sudakshina Das <sudi.das@arm.com> wrote: >>> On 23/05/18 18:21, Jason Merrill wrote: >>>> >>>> The first patch implements the adjustments from core issues 616 and >>>> 1213 to the value category of subobjects of class prvalues: they were >>>> considered prvalues themselves, but that was kind of nonsensical. Now >>>> they are considered xvalues. Along with this, I've removed the >>>> diagnostic distinction between xvalues and prvalues when trying to use >>>> one or the other as an lvalue; the important thing is that they are >>>> rvalues. >>>> >>>> The second patch corrects various issues with casts and xvalues/rvalue >>>> references: we were treating an xvalue operand to dynamic_cast as an >>>> lvalue, and we were objecting to casts from prvalue to rvalue >>>> reference type. >>>> >>> >>> With the second patch: >>> commit f7d2790049fd1e59af4b69ee12f7c101cfe4cdab >>> Author: jason <jason@138bc75d-0d04-0410-961f-82ee72b054a4> >>> Date: Wed May 23 17:21:39 2018 +0000 >>> >>> Fix cast to rvalue reference from prvalue. >>> >>> * cvt.c (diagnose_ref_binding): Handle rvalue reference. >>> * rtti.c (build_dynamic_cast_1): Don't try to build a reference to >>> non-class type. Handle xvalue argument. >>> * typeck.c (build_reinterpret_cast_1): Allow cast from prvalue to >>> rvalue reference. >>> * semantics.c (finish_compound_literal): Do direct-initialization, >>> not cast, to initialize a reference. >>> >>> git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@260622 >>> 138bc75d-0d04-0410-961f-82ee72b054a4 >>> >>> I have observed the following failure in Spec2017 while building >>> 510.parest_r on aarch64-none-linux-gnu >>> >>> aarch64-none-linux-gnu-g++ -c -o source/numerics/matrices.all_dimensions.o >>> -DSPEC -DNDEBUG -Iinclude -I. -DSPEC_AUTO_SUPPRESS_OPENMP >>> -mcpu=cortex-a57+crypto -Ofast -fomit-frame-pointer -fpermissive >>> -DSPEC_LP64 source/numerics/matrices.all_dimensions.cc >>> >>> source/numerics/matrices.all_dimensions.cc: In static member function >>> 'static void dealii::MatrixTools::apply_boundary_values(const >>> std::map<unsigned int, double>&, dealii::BlockSparseMatrix<number>&, >>> dealii::BlockVector<number>&, dealii::BlockVector<number>&, bool)': >>> >>> source/numerics/matrices.all_dimensions.cc:469:50: error: lvalue required as >>> unary '&' operand >>> >>> [this_sparsity.get_rowstart_indices()[row]]; >>> >>> ^ >>> >>> source/numerics/matrices.all_dimensions.cc:472:55: error: lvalue required as >>> unary '&' operand >>> >>> [this_sparsity.get_rowstart_indices()[row]+1], >>> >>> ^ >>> >>> source/numerics/matrices.all_dimensions.cc:474:55: error: lvalue required as >>> unary '&' operand >>> >>> [this_sparsity.get_rowstart_indices()[row+1]], >>> >>> ^ >>> >>> source/numerics/matrices.all_dimensions.cc:479:49: error: lvalue required as >>> unary '&' operand >>> >>> [this_sparsity.get_rowstart_indices()[row]], >>> >>> ^ >>> >>> source/numerics/matrices.all_dimensions.cc:481:51: error: lvalue required as >>> unary '&' operand >>> >>> [this_sparsity.get_rowstart_indices()[row+1]], >>> >>> ^ >>> >>> source/numerics/matrices.all_dimensions.cc:510:50: error: lvalue required as >>> unary '&' operand >>> >>> [this_sparsity.get_rowstart_indices()[0]]); >>> >>> Sudi >> >> Thanks, investigating. > > Fixed thus. Thanks for fixing this! Sudi >
commit 23a88cb6641f097dd0a9faebfafbe2f1e876714b
Author: Jason Merrill <jason@redhat.com>
Date: Thu May 3 08:39:26 2018 -0400
Fix cast to rvalue reference from prvalue.
* cvt.c (diagnose_ref_binding): Handle rvalue reference.
* rtti.c (build_dynamic_cast_1): Don't try to build a reference to
non-class type. Handle xvalue argument.
* typeck.c (build_reinterpret_cast_1): Allow cast from prvalue to
rvalue reference.
* semantics.c (finish_compound_literal): Do direct-initialization,
not cast, to initialize a reference.
diff --git a/gcc/cp/cvt.c b/gcc/cp/cvt.c
index 30b44b7d7ea..3f87317a47d 100644
--- a/gcc/cp/cvt.c
+++ b/gcc/cp/cvt.c
@@ -381,7 +381,8 @@ diagnose_ref_binding (location_t loc, tree reftype, tree intype, tree decl)
{
tree ttl = TREE_TYPE (reftype);
- if (!CP_TYPE_CONST_NON_VOLATILE_P (ttl))
+ if (!TYPE_REF_IS_RVALUE (reftype)
+ && !CP_TYPE_CONST_NON_VOLATILE_P (ttl))
{
const char *msg;
diff --git a/gcc/cp/rtti.c b/gcc/cp/rtti.c
index 426a23276e0..6692fb7ff86 100644
--- a/gcc/cp/rtti.c
+++ b/gcc/cp/rtti.c
@@ -616,22 +616,22 @@ build_dynamic_cast_1 (tree type, tree expr, tsubst_flags_t complain)
else
{
expr = mark_lvalue_use (expr);
-
- exprtype = build_reference_type (TREE_TYPE (expr));
+ exprtype = TREE_TYPE (expr);
/* T is a reference type, v shall be an lvalue of a complete class
type, and the result is an lvalue of the type referred to by T. */
-
- if (! MAYBE_CLASS_TYPE_P (TREE_TYPE (exprtype)))
+ if (! MAYBE_CLASS_TYPE_P (exprtype))
{
errstr = _("source is not of class type");
goto fail;
}
- if (!COMPLETE_TYPE_P (complete_type (TREE_TYPE (exprtype))))
+ if (!COMPLETE_TYPE_P (complete_type (exprtype)))
{
errstr = _("source is of incomplete class type");
goto fail;
}
+
+ exprtype = cp_build_reference_type (exprtype, !lvalue_p (expr));
}
/* The dynamic_cast operator shall not cast away constness. */
diff --git a/gcc/cp/semantics.c b/gcc/cp/semantics.c
index 94e8f54254d..46251deaa6c 100644
--- a/gcc/cp/semantics.c
+++ b/gcc/cp/semantics.c
@@ -2734,7 +2734,10 @@ finish_compound_literal (tree type, tree compound_literal,
compound_literal
= finish_compound_literal (TREE_TYPE (type), compound_literal,
complain, fcl_context);
- return cp_build_c_cast (type, compound_literal, complain);
+ /* The prvalue is then used to direct-initialize the reference. */
+ tree r = (perform_implicit_conversion_flags
+ (type, compound_literal, complain, LOOKUP_NORMAL));
+ return convert_from_reference (r);
}
if (!TYPE_OBJ_P (type))
diff --git a/gcc/cp/typeck.c b/gcc/cp/typeck.c
index 82089c45105..a694499190f 100644
--- a/gcc/cp/typeck.c
+++ b/gcc/cp/typeck.c
@@ -7315,13 +7315,19 @@ build_reinterpret_cast_1 (tree type, tree expr, bool c_cast_p,
type = cv_unqualified (type);
/* [expr.reinterpret.cast]
- An lvalue expression of type T1 can be cast to the type
+ A glvalue expression of type T1 can be cast to the type
"reference to T2" if an expression of type "pointer to T1" can be
explicitly converted to the type "pointer to T2" using a
reinterpret_cast. */
if (TYPE_REF_P (type))
{
- if (! lvalue_p (expr))
+ if (TYPE_REF_IS_RVALUE (type))
+ {
+ if (!obvalue_p (expr))
+ /* Perform the temporary materialization conversion. */
+ expr = get_target_expr_sfinae (expr, complain);
+ }
+ else if (!lvalue_p (expr))
{
if (complain & tf_error)
error ("invalid cast of an rvalue expression of type "
diff --git a/gcc/testsuite/g++.dg/cpp0x/rv-cast6.C b/gcc/testsuite/g++.dg/cpp0x/rv-cast6.C
new file mode 100644
index 00000000000..3ae5691c5fd
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp0x/rv-cast6.C
@@ -0,0 +1,11 @@
+// Test that a prvalue can be used where a glvalue is expected.
+// { dg-do compile { target c++11 } }
+
+struct A { virtual void f(); };
+struct B : A {};
+
+auto && a = static_cast<A&&>(B());
+auto && b = reinterpret_cast<A&&>(B());
+auto && c = dynamic_cast<A&&>(B());
+auto && d = dynamic_cast<B&&>(static_cast<A&&>(B()));
+auto && e = const_cast<B&&>(B());