Cython函数内存泄漏

时间:2017-09-22 05:35:08

标签: python c memory-leaks cython

我已经将python函数转换为cython函数。现在该功能按预期工作。但是当主程序多次调用此函数时,我会遇到大量内存泄漏。我已经释放了我动态分配的内存,但它似乎没有用。

我在这里做错了什么?

from cpython.mem cimport PyMem_Malloc, PyMem_Free

def longest_common_substring(refWord, stemWord):
cdef:
    int longest, x_longest
    int x, y, k
    Py_ssize_t lengthRefWord
    Py_ssize_t lengthStemWord
    wchar_t *referenceWord = PyUnicode_AsWideCharString(refWord, &lengthRefWord)
    wchar_t *stemmableWord = PyUnicode_AsWideCharString(stemWord, &lengthStemWord)
    int t1 = lengthRefWord+1
    int t2 = lengthStemWord+1
    int **m = <int **> PyMem_Malloc(t1 * sizeof(int *))
    wchar_t tempChar1;
    wchar_t tempChar2;
longest = 0
x_longest = 0

for k in range(t1):
   m[k] = <int *> PyMem_Malloc(t2 * sizeof(int))
for x in range(0, t1):
    for y in range(0, t2):
        m[x][y] = 0

for x in range(1, t1):
    for y in range(1, t2):
       tempChar1 = referenceWord[x - 1]
       tempChar2 = stemmableWord[y - 1]

       if tempChar1 == tempChar2:
           m[x][y] = m[x - 1][y - 1] + 1
           if m[x][y] > longest:
               longest = m[x][y]
               x_longest = x
       else:
           m[x][y] = 0
for k in range(t1):
   PyMem_Free(m[k])
PyMem_Free(m)
return refWord[x_longest - longest: x_longest]

1 个答案:

答案 0 :(得分:2)

PyUnicode_AsWideCharString分配你必须释放的内存。 The documentation

  

成功返回由PyMem_Alloc()分配的缓冲区(使用PyMem_Free()释放它)。

你从这个函数得到两个字符串但是没有任何字符串。