Martin Thümmel 178ef391c7 Fix of important memory leak in Python module ...

Dear all,
As described in the issue https://github.com/syoyo/tinyobjloader/issues/188#issue-385341218 there is a memory leak in the function pyLoadObj(). The reference count for all created Python objects must be decreased after they are fed into the lists and dictionaries. The only  exception from this is the function PyTuple_SetItem() in pyTupleFromfloat3(), which decreases the reference counter of the Python object *tuple automatically by calling this function. In all other cases like PyList_Insert(), the references are only borrowed and not decreased. Therefore, each created Python object will remain in the heap, since there is one reference to it left in the counter.

These facts are explained in https://docs.python.org/2/c-api/intro.html#reference-counts in detail.

Best regards
Martin.
P.S. sorry, that i did not put that much effort into a more readable code and just inserted the Py_DECREF() function at every necessary position.

2018-11-28 17:16:12 +01:00

7.3 KiB

Raw Blame History

View Raw