Provide our own Cython declaration for make_unique

make_unique in Cython's libcpp headers is not annotated with `except
+`. As a consequence, if the constructor throws, we do not catch it in
Python. To work around this (see cython/cython#5560 for details),
provide our own implementation.

Due to the way assignments occur to temporaries, we need to now
explicitly wrap all calls to `make_unique` in `move`, but that is
arguably preferable to not being able to catch exceptions.

- Closes #13743

python/cudf/cudf/_lib/column.pyx

@@ -27,7 +27,7 @@ from cudf.utils.dtypes import _get_base_dtype
 
 from cpython.buffer cimport PyObject_CheckBuffer
 from libc.stdint cimport uintptr_t
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
 from libcpp.utility cimport move
 from libcpp.vector cimport vector
 
@@ -49,6 +49,7 @@ from cudf._lib.cpp.column.column_factories cimport (
     make_numeric_column,
 )
 from cudf._lib.cpp.column.column_view cimport column_view
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.null_mask cimport null_count as cpp_null_count
 from cudf._lib.cpp.scalar.scalar cimport scalar
 from cudf._lib.scalar cimport DeviceScalar
@@ -275,7 +276,7 @@ cdef class Column:
                 self._children = ()
             else:
                 children = Column.from_unique_ptr(
-                    make_unique[column](self.view())
+                    move(make_unique[column](self.view()))
                 ).base_children
                 dtypes = [
                     base_child.dtype for base_child in self.base_children

python/cudf/cudf/_lib/concat.pyx

@@ -1,7 +1,7 @@
-# Copyright (c) 2020-2022, NVIDIA CORPORATION.
+# Copyright (c) 2020-2023, NVIDIA CORPORATION.
 
 from libcpp cimport bool
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
 from libcpp.utility cimport move
 from libcpp.vector cimport vector
 
@@ -12,6 +12,7 @@ from cudf._lib.cpp.concatenate cimport (
     concatenate_masks as libcudf_concatenate_masks,
     concatenate_tables as libcudf_concatenate_tables,
 )
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.table.table cimport table, table_view
 from cudf._lib.utils cimport (
     data_from_unique_ptr,
@@ -30,7 +31,7 @@ cpdef concat_masks(object columns):
     cdef vector[column_view] c_views = make_column_views(columns)
     with nogil:
         c_result = move(libcudf_concatenate_masks(c_views))
-        c_unique_result = make_unique[device_buffer](move(c_result))
+        c_unique_result = move(make_unique[device_buffer](move(c_result)))
     return as_buffer(
         DeviceBuffer.c_from_unique_ptr(move(c_unique_result))
     )

python/cudf/cudf/_lib/copying.pyx

@@ -4,7 +4,7 @@ import pickle
 
 from libc.stdint cimport int32_t, uint8_t, uintptr_t
 from libcpp cimport bool
-from libcpp.memory cimport make_shared, make_unique, shared_ptr, unique_ptr
+from libcpp.memory cimport make_shared, shared_ptr, unique_ptr
 from libcpp.utility cimport move
 from libcpp.vector cimport vector
 
@@ -29,6 +29,7 @@ cimport cudf._lib.cpp.copying as cpp_copying
 from cudf._lib.cpp.column.column cimport column
 from cudf._lib.cpp.column.column_view cimport column_view, mutable_column_view
 from cudf._lib.cpp.libcpp.functional cimport reference_wrapper
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.lists.gather cimport (
     segmented_gather as cpp_segmented_gather,
 )
@@ -80,7 +81,6 @@ def copy_column(Column input_column):
     Deep copied column
     """
     cdef unique_ptr[column] c_result
-
     cdef column_view input_column_view = input_column.view()
     with nogil:
         c_result = move(make_unique[column](input_column_view))

python/cudf/cudf/_lib/cpp/libcpp/memory.pxd

@@ -0,0 +1,12 @@
+# Copyright (c) 2023, NVIDIA CORPORATION.
+
+from libcpp.memory cimport unique_ptr
+
+
+cdef extern from "<memory>" namespace "std" nogil:
+    # The Cython standard header does not have except +, so C++
+    # exceptions from make_unique are not caught and translated to
+    # Python ones. This is not perfectly ergonomic, we always have to
+    # wrap make_unique in move, but at least we can catch exceptions.
+    # See https://github.com/cython/cython/issues/5560
+    unique_ptr[T] make_unique[T](...) except +

python/cudf/cudf/_lib/expressions.pyx

@@ -4,9 +4,11 @@ from enum import Enum
 
 from cython.operator cimport dereference
 from libc.stdint cimport int64_t
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
+from libcpp.utility cimport move
 
 from cudf._lib.cpp cimport expressions as libcudf_exp
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.types cimport size_type
 
 # Necessary for proper casting, see below.
@@ -80,26 +82,26 @@ cdef class Literal(Expression):
     def __cinit__(self, value):
         if isinstance(value, int):
             self.c_scalar.reset(new numeric_scalar[int64_t](value, True))
-            self.c_obj = <expression_ptr> make_unique[libcudf_exp.literal](
+            self.c_obj = <expression_ptr> move(make_unique[libcudf_exp.literal](
                 <numeric_scalar[int64_t] &>dereference(self.c_scalar)
-            )
+            ))
         elif isinstance(value, float):
             self.c_scalar.reset(new numeric_scalar[double](value, True))
-            self.c_obj = <expression_ptr> make_unique[libcudf_exp.literal](
+            self.c_obj = <expression_ptr> move(make_unique[libcudf_exp.literal](
                 <numeric_scalar[double] &>dereference(self.c_scalar)
-            )
+            ))
         elif isinstance(value, str):
             self.c_scalar.reset(new string_scalar(value.encode(), True))
-            self.c_obj = <expression_ptr> make_unique[libcudf_exp.literal](
+            self.c_obj = <expression_ptr> move(make_unique[libcudf_exp.literal](
                 <string_scalar &>dereference(self.c_scalar)
-            )
+            ))
 
 
 cdef class ColumnReference(Expression):
     def __cinit__(self, size_type index):
-        self.c_obj = <expression_ptr>make_unique[libcudf_exp.column_reference](
+        self.c_obj = <expression_ptr>move(make_unique[libcudf_exp.column_reference](
             index
-        )
+        ))
 
 
 cdef class Operation(Expression):
@@ -109,10 +111,10 @@ cdef class Operation(Expression):
         )
 
         if right is None:
-            self.c_obj = <expression_ptr> make_unique[libcudf_exp.operation](
+            self.c_obj = <expression_ptr> move(make_unique[libcudf_exp.operation](
                 op_value, dereference(left.c_obj)
-            )
+            ))
         else:
-            self.c_obj = <expression_ptr> make_unique[libcudf_exp.operation](
+            self.c_obj = <expression_ptr> move(make_unique[libcudf_exp.operation](
                 op_value, dereference(left.c_obj), dereference(right.c_obj)
-            )
+            ))

python/cudf/cudf/_lib/join.pyx

@@ -2,7 +2,7 @@
 
 from cudf.core.buffer import acquire_spill_lock
 
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
 from libcpp.pair cimport pair
 from libcpp.utility cimport move
 
@@ -11,6 +11,7 @@ from rmm._lib.device_buffer cimport device_buffer
 cimport cudf._lib.cpp.join as cpp_join
 from cudf._lib.column cimport Column
 from cudf._lib.cpp.column.column cimport column
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.table.table_view cimport table_view
 from cudf._lib.cpp.types cimport data_type, size_type, type_id
 from cudf._lib.utils cimport table_view_from_columns
@@ -66,8 +67,8 @@ cdef Column _gather_map_as_column(cpp_join.gather_map_type gather_map):
     # help to convert a gather map to a Column
     cdef device_buffer c_empty
     cdef size_type size = gather_map.get()[0].size()
-    cdef unique_ptr[column] c_col = make_unique[column](
+    cdef unique_ptr[column] c_col = move(make_unique[column](
         data_type(type_id.INT32),
         size,
-        gather_map.get()[0].release(), move(c_empty), 0)
+        gather_map.get()[0].release(), move(c_empty), 0))
     return Column.from_unique_ptr(move(c_col))

python/cudf/cudf/_lib/null_mask.pyx

@@ -1,17 +1,18 @@
-# Copyright (c) 2020-2022, NVIDIA CORPORATION.
+# Copyright (c) 2020-2023, NVIDIA CORPORATION.
 
 from enum import Enum
 
 from rmm._lib.device_buffer cimport DeviceBuffer, device_buffer
 
 from cudf.core.buffer import acquire_spill_lock, as_buffer
 
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
 from libcpp.pair cimport pair
 from libcpp.utility cimport move
 
 from cudf._lib.column cimport Column
 from cudf._lib.cpp.column.column_view cimport column_view
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.null_mask cimport (
     bitmask_allocation_size_bytes as cpp_bitmask_allocation_size_bytes,
     bitmask_and as cpp_bitmask_and,
@@ -50,7 +51,7 @@ def copy_bitmask(Column col):
 
     with nogil:
         db = move(cpp_copy_bitmask(col_view))
-        up_db = make_unique[device_buffer](move(db))
+        up_db = move(make_unique[device_buffer](move(db)))
 
     rmm_db = DeviceBuffer.c_from_unique_ptr(move(up_db))
     buf = as_buffer(rmm_db)
@@ -96,7 +97,7 @@ def create_null_mask(size_type size, state=MaskState.UNINITIALIZED):
 
     with nogil:
         db = move(cpp_create_null_mask(size, c_mask_state))
-        up_db = make_unique[device_buffer](move(db))
+        up_db = move(make_unique[device_buffer](move(db)))
 
     rmm_db = DeviceBuffer.c_from_unique_ptr(move(up_db))
     buf = as_buffer(rmm_db)
@@ -110,7 +111,7 @@ def bitmask_and(columns: list):
     cdef unique_ptr[device_buffer] up_db
     with nogil:
         c_result = move(cpp_bitmask_and(c_view))
-        up_db = make_unique[device_buffer](move(c_result.first))
+        up_db = move(make_unique[device_buffer](move(c_result.first)))
     dbuf = DeviceBuffer.c_from_unique_ptr(move(up_db))
     buf = as_buffer(dbuf)
     return buf, c_result.second
@@ -123,7 +124,7 @@ def bitmask_or(columns: list):
     cdef unique_ptr[device_buffer] up_db
     with nogil:
         c_result = move(cpp_bitmask_or(c_view))
-        up_db = make_unique[device_buffer](move(c_result.first))
+        up_db = move(make_unique[device_buffer](move(c_result.first)))
     dbuf = DeviceBuffer.c_from_unique_ptr(move(up_db))
     buf = as_buffer(dbuf)
     return buf, c_result.second

python/cudf/cudf/_lib/parquet.pyx

@@ -32,7 +32,7 @@ from cudf._lib.utils import _index_level_name, generate_pandas_metadata
 from libc.stdint cimport uint8_t
 from libcpp cimport bool
 from libcpp.map cimport map
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
 from libcpp.string cimport string
 from libcpp.unordered_map cimport unordered_map
 from libcpp.utility cimport move
@@ -51,6 +51,7 @@ from cudf._lib.cpp.io.parquet cimport (
     write_parquet as parquet_writer,
 )
 from cudf._lib.cpp.io.types cimport column_in_metadata, table_input_metadata
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 from cudf._lib.cpp.table.table_view cimport table_view
 from cudf._lib.cpp.types cimport data_type, size_type
 from cudf._lib.io.datasource cimport NativeFileDatasource
@@ -644,7 +645,7 @@ cpdef merge_filemetadata(object filemetadata_list):
 
     for blob_py in filemetadata_list:
         blob_c = blob_py
-        list_c.push_back(make_unique[vector[uint8_t]](blob_c))
+        list_c.push_back(move(make_unique[vector[uint8_t]](blob_c)))
 
     with nogil:
         output_c = move(parquet_merge_metadata(list_c))

python/cudf_kafka/cudf_kafka/_lib/kafka.pyx

@@ -1,12 +1,13 @@
-# Copyright (c) 2020-2022, NVIDIA CORPORATION.
+# Copyright (c) 2020-2023, NVIDIA CORPORATION.
 
 from libc.stdint cimport int32_t, int64_t
 from libcpp cimport bool, nullptr
 from libcpp.map cimport map
-from libcpp.memory cimport make_unique, unique_ptr
+from libcpp.memory cimport unique_ptr
 from libcpp.string cimport string
 
 from cudf._lib.cpp.io.types cimport datasource
+from cudf._lib.cpp.libcpp.memory cimport make_unique
 
 from cudf_kafka._lib.kafka cimport kafka_consumer
 
@@ -50,20 +51,20 @@ cdef class KafkaDatasource(Datasource):
 
         if topic != b"" and partition != -1:
             self.c_datasource = <unique_ptr[datasource]> \
-                make_unique[kafka_consumer](configs,
-                                            python_callable,
-                                            python_callable_wrapper,
-                                            topic,
-                                            partition,
-                                            start_offset,
-                                            end_offset,
-                                            batch_timeout,
-                                            delimiter)
+                move(make_unique[kafka_consumer](configs,
+                                                 python_callable,
+                                                 python_callable_wrapper,
+                                                 topic,
+                                                 partition,
+                                                 start_offset,
+                                                 end_offset,
+                                                 batch_timeout,
+                                                 delimiter))
         else:
             self.c_datasource = <unique_ptr[datasource]> \
-                make_unique[kafka_consumer](configs,
-                                            python_callable,
-                                            python_callable_wrapper)
+                move(make_unique[kafka_consumer](configs,
+                                                 python_callable,
+                                                 python_callable_wrapper))
 
     cdef datasource* get_datasource(self) nogil:
         return <datasource *> self.c_datasource.get()