1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
|
Note about the "memory leaks" in the Python allocators
======================================================
This concerns functions "PyTuple_SetItem()" and
"PyDict_SetItemString()", whose documentation is the following:
>>>>>
* int PyTuple_SetItem(PyObject *p, Py_ssize_t pos, PyObject *o)
Insert a reference to object o at position pos of the tuple pointed
to by p. Return 0 on success. If pos is out of bounds, return -1 and
set an IndexError exception.
Note: This function "steals" a reference to o and discards a
reference to an item already in the tuple at the affected position.
* int PyDict_SetItemString(PyObject *p, const char *key, PyObject *val)
Insert val into the dictionary p using key as a key. key should be a
const char* UTF-8 encoded bytes string. The key object is created
using PyUnicode_FromString(key). Return 0 on success or -1 on
failure. This function does not steal a reference to val.
<<<<<
As can be seen:
* PyTuple_SetItem() steals the reference.
* PyDict_SetItemString() does *not* steal the reference.
Consequently, when passing an object created by the Orthanc Python
plugin:
* PyTuple_SetItem() must use "PythonObject::Release()".
* PyDict_SetItemString() must use "PythonObject::GetPyObject()".
In releases <= 4.0 of the Python plugin, PyDict_SetItemString() was
using "PythonObject::Release()" instead of
"PythonObject::GetPyObject()", which resulted in memory leaks.
|
