[Bug]: Apparent memory issue?

[joelfrederico]

Summary

If you insert an exception here:

matplotlib/lib/matplotlib/backends/backend_macosx.py

Line 68 in 25a54a4

def __init__(self, canvas, num):

the code has a segfault on quitting. This likely means that there's either a memory leak or double-free or some other memory issue someplace.

Proposed fix

Insert an exception here:

matplotlib/lib/matplotlib/backends/backend_macosx.py

Line 68 in 25a54a4

def __init__(self, canvas, num):

Build against a debug version of python and track down the issue using a debugger.

[QuLogic]

The backtrace says it is crashing in [window close] in FigureManager_dealloc. In FigureManager_new, the self->window is allocated, and in FigureManager_init it is initialized using initWithContentRect. If we raise from __init__ immediately, then the super-init is not called and the window is not initialized. Presumably an uninitialized window cannot be closed, but I'm not sure how to check if it has been.

[joelfrederico]

It seems like the real issue is a misuse/disuse of tp_alloc, tp_init, tp_finalize, and tp_dealloc.

The window ought to be allocated in a custom tp_alloc. Currently, this is done in tp_init.

The tp_init should call some init method on the window.

The window should be closed in tp_finalize. This should be decoupled from releasing it. I think that can be done by setting isReleasedWhenClosed to false during tp_init.

The window should be released in tp_dealloc.

Is that all correct?

[tacaswell]

@joelfrederico I milestoned this as 3.6 as you did not label this as a regression, but if there is an easy fix to reduce segfaults, we can definitely backport it.

[tacaswell]

Checking, the state is current the same, however if I am reading this correctly we only have the issue if we raise in the __init__?