如何从double数组初始化元组，避免“存储临时对象的不安全C派生”错误？

尝试从C语言的double数组初始化Python元组，编写了如下Cython代码：

from cpython.ref cimport PyObject
from cpython.tuple cimport PyTuple_New
from cpython.float cimport PyFloat_FromDouble

cdef extern from "Python.h":
    struct PyTupleObject:
        PyObject *ob_item[1]

# ======================================================================
# Tuple utility API
# ======================================================================
cdef tuple from_doubles(unsigned n, const double* buffer):
    """ Create a tuple from an array of floats. """
    cdef tuple t = PyTuple_New(n)
    cdef PyObject** items = (<PyTupleObject*>t).ob_item
    cdef unsigned i
    for i in range(n):
        items[i] = <PyObject*>PyFloat_FromDouble(buffer[i])

    return t

认为PyFloat_FromDouble创建的新对象会被元组“窃取”引用，无需调用Py_INCREF/Py_DECREF，但使用Cython 0.26.1编译时，出现以下错误：

Error compiling Cython file:
------------------------------------------------------------
...
    """
    cdef tuple t = PyTuple_New(n)
    cdef PyObject** items = (<PyTupleObject*>t).ob_item
    cdef unsigned i
    for i in range(n):
        items[i] = <PyObject*>PyFloat_FromDouble(buffer[i])
                  ^
------------------------------------------------------------

fin/tuplex.pyx:19:19: Casting temporary Python object to non-numeric non-Python type

Error compiling Cython file:
------------------------------------------------------------
...
    """
    cdef tuple t = PyTuple_New(n)
    cdef PyObject** items = (<PyTupleObject*>t).ob_item
    cdef unsigned i
    for i in range(n):
        items[i] = <PyObject*>PyFloat_FromDouble(buffer[i])
                  ^
------------------------------------------------------------

fin/tuplex.pyx:19:19: Storing unsafe C derivative of temporary Python reference

需要修正代码，将新创建的PyFloat对象正确存储到元组中。

Cython的类型检查机制不允许直接将临时Python对象强制转换为PyObject*后存入C指针数组，因为它无法识别这种“引用窃取”的行为。正确的做法是使用Python C API提供的PyTuple_SET_ITEM宏，它的行为正是期望的——直接将对象指针存入元组，不增加引用计数，相当于“窃取”引用。

修正步骤 #

1. 在cdef extern from "Python.h"中声明PyTuple_SET_ITEM宏：
2. 替换直接操作ob_item数组的代码，改为调用PyTuple_SET_ITEM。

修正后的完整代码 #

from cpython.ref cimport PyObject
from cpython.tuple cimport PyTuple_New
from cpython.float cimport PyFloat_FromDouble

cdef extern from "Python.h":
    void PyTuple_SET_ITEM(PyObject* t, Py_ssize_t pos, PyObject* o)

# ======================================================================
# Tuple utility API
# ======================================================================
cdef tuple from_doubles(unsigned n, const double* buffer):
    """ Create a tuple from an array of floats. """
    cdef tuple t = PyTuple_New(n)
    cdef unsigned i
    for i in range(n):
        PyTuple_SET_ITEM(<PyObject*>t, i, <PyObject*>PyFloat_FromDouble(buffer[i]))

    return t

关键说明 #

• PyTuple_SET_ITEM是专门为元组初始化设计的宏，直接将对象指针写入元组内部数组，不会触发额外的引用计数操作，完全符合“引用窃取”的需求。
• 使用该宏后，Cython能够正确识别代码语义，避免类型检查错误，同时保证代码符合Python C API的规范。

内容的提问来源于stack exchange，提问作者Sylvain Leroux