Question

我有一个python文件，我想从dll调用一个函数。

该功能的原型是：

typedef double real_T;
extern real_T __declspec(dllexport) RectIntersect(const real_T rect1[8], const real_T rect2[8]);

python代码：

import math;
import ctypes;
from ctypes import *
import numpy;


# this module shall always be executed directly
if __name__ == '__main__':
    print "Program started !";
    rect = ctypes.c_double * 8;
    rect1 = rect(1.1, 2.45, 3, 4, 5, 6, 7, 8);
    rect2 = rect(1.6, 3.45, 3.1, 4.1, 5.1, 6.1, 7.1, 8.1);

    # Load DLL into memory.

    hllDll = ctypes.WinDLL ("IntersectDLL.dll");
    hllDll.RectIntersect.argtypes =[ctypes.c_double * 8, ctypes.c_double * 8];
    hllDll.RectIntersect (rect1, rect2);

我收到错误：

Traceback (most recent call last):
  File "C:\Program Files (x86)\Microsoft Visual Studio 10.0\Common7\IDE\Extensions\Microsoft\Python Tools for Visual Studio\2.0\visualstudio_py_util.py", line 70, in exec_file
    exec(code_obj, global_variables)
  File "D:\Sandboxes\SRR2T0\06_Algorithm\05_Testing\05_Test_Environment\algo\smr200_bbt\valf_tests\adma\test.py", line 18, in <module>
    hllDll.RectIntersect (rect1, rect2);
ValueError: Procedure probably called with too many arguments (8 bytes in excess)

请帮助:( .... ....

Answer 1

C声明

real_T RectIntersect(const real_T rect1[8], ...)

完全等同于（并由C编译器替换）：

real_T RectIntersect(const real_T *rect1, ...)

这意味着DLL中的函数实际上并不期望一个8个双精度数组，而是一个指向double的指针，后面可能会再增加7个以生成其中8个数组的数组（但是这部分未被检查）。这意味着你需要其他方法来用ctypes编写它;例如：

hllDll.RectIntersect.argtypes =[ctypes.POINTER(ctypes.c_double * 8),...]

并将其传递给ctypes.byref(rect1)。

或者：您可能没有CFFI（http://cffi.readthedocs.org/）而不是ctypes的问题，而ctypes与C类型的匹配程度更接近：

import cffi
ffi = cffi.FFI()
ffi.cdef("""
    typedef double real_T;
    real_T RectIntersect(const real_T rect1[8], const real_T rect2[8]);
""")
lib = ffi.dlopen("IntersectDLL.dll")
rect1 = ffi.new("real_T[8]", [1.1, 2.45, 3, 4, 5, 6, 7, 8])
rect2 = ffi.new("real_T[8]", [1.6, 3.45, 3.1, 4.1, 5.1, 6.1, 7.1, 8.1])
lib.RectIntersect(rect1, rect2)

来自DLL的python调用函数

1 个答案: