Issue1256669
This issue tracker has been migrated to GitHub,
and is currently read-only.
For more information,
see the GitHub FAQs in the Python's Developer Guide.
Created on 2005-08-11 12:49 by bheld, last changed 2022-04-11 14:56 by admin. This issue is now closed.
Messages (8) | |||
---|---|---|---|
msg26015 - (view) | Author: Ben Held (bheld) | Date: 2005-08-11 12:49 | |
Having recently upgraded to Python 2.4, I am having a large memory leak with the following code built with VC++ 6.0: PyObject *pName, *pModule; Py_Initialize(); pName = PyString_FromString(argv[1]); pModule = PyImport_Import(pName); Py_DECREF(pName); PyObject* pModule2 = PyImport_ReloadModule(pModule); Py_DECREF(pModule2); Py_DECREF(pModule); Py_Finalize(); return 0; I get leaks of over 500 kb. I have another program which is much more complex, in which every call to PyImport_ReloadModule is leaking 200+ kb, even though I am calling Py_DECREF correctly. |
|||
msg26016 - (view) | Author: Martin v. Löwis (loewis) * | Date: 2005-08-13 13:34 | |
Logged In: YES user_id=21627 How do you know there is a memory leak? |
|||
msg26017 - (view) | Author: Ben Held (bheld) | Date: 2005-08-16 13:56 | |
Logged In: YES user_id=1327580 Boundschecker shows the leak and I have verified this by watching the process memory increase via the task manager. |
|||
msg26018 - (view) | Author: Collin Winter (collinwinter) * | Date: 2005-09-14 17:53 | |
Logged In: YES user_id=1344176 I've been unable to verify this on Linux. I've tested python versions 2.2.3, 2.3.5 and 2.4.1, all compiled with gcc 3.3.5 on Debian 3.1 under kernel 2.6.8. I used the sample program provided by Ben, modified with an infinite loop over the PyImport_ReloadModule/PyDECREF(pModule2) lines, sleeping for 1 second after every 25 iterations. I tested reloading the modules distutils, os.path, distutils.command.sdist for 300+ iterations each under each python version. No memory leak was observed. |
|||
msg26019 - (view) | Author: Ben Held (bheld) | Date: 2005-09-14 18:09 | |
Logged In: YES user_id=1327580 This behavior is evident with Python 2.3.5 built on Windows with VC++ 6.0. |
|||
msg26020 - (view) | Author: Neal Norwitz (nnorwitz) * | Date: 2005-09-20 06:02 | |
Logged In: YES user_id=33168 I just tested with Python 2.4.2 on Linux with valgrind. It doesn't report any leaks. So this could be specific to Python 2.3.5, Windows, VC++ 6.0 or some other variation. Can you provide more info? For example, what module are you passing on the command line? Can you provide that code? What objects specifically do you think are leaking? Where were they allocated? |
|||
msg26021 - (view) | Author: Ben Held (bheld) | Date: 2005-11-11 16:13 | |
Logged In: YES user_id=1327580 Turns out this is not a Python issue, but a SWIG issue. It only leaks for modules that were created with SWIG. Please close this item |
|||
msg26022 - (view) | Author: Georg Brandl (georg.brandl) * | Date: 2005-11-11 19:46 | |
Logged In: YES user_id=1188172 Closing as requested. |
History | |||
---|---|---|---|
Date | User | Action | Args |
2022-04-11 14:56:12 | admin | set | github: 42267 |
2005-08-11 12:49:35 | bheld | create |