SystemError: ... cellobject.c:22: bad argument to internal ?

[robert]

From the trace of a 2.3.5 software i got:

\'SystemError:
C:\\\\sf\\\\python\\\\dist23\\\\src\\\\Objects\\\\cellobject.c:22: bad
argument to internal
function\\n\']


from the middle of normal function / or its call. What is this?

There is:

http://mail.python.org/pipermail/python-list/2005-January/260273.html

but seems to be something else. There is no dynamic addition of
class-method. But the last function called is a callback function in
3rd level inner function of class method like:

class X:
def a():
def b():
...
def c():... # breaks with systemerror
schedule_callback(c) # -> upon Windows WM_TIMER event
...


Will there be another bug-fix release of Python 2.3 ?

-robert

 

[nnorwitz]

From the trace of a 2.3.5 software i got:

\'SystemError:
C:\\\\sf\\\\python\\\\dist23\\\\src\\\\Objects\\\\cellobject.c:22: bad
argument to internal
function\\n\']
....

Will there be another bug-fix release of Python 2.3 ?

No, is this still a problem in 2.4? 2.4.4 is planned to be released
this summer. Can you post the entire code (that doesn't reference
anything outside the stdlib) that causes this problem?

n

 

[robert]

From the trace of a 2.3.5 software i got:

\'SystemError:
C:\\\\sf\\\\python\\\\dist23\\\\src\\\\Objects\\\\cellobject.c:22: bad
argument to internal
function\\n\']

...


Will there be another bug-fix release of Python 2.3 ?

No, is this still a problem in 2.4? 2.4.4 is planned to be released

( Am bound to py2.3 for this app - as py2.4 more than doubled
distributable app sizes with MSVCRTL71, CJK codecs in core etc... )

this summer. Can you post the entire code (that doesn't reference
anything outside the stdlib) that causes this problem?

Solved - no Python error.

The bug turned out to be a refcount bug in PythonWin's
win32ui.HookMessage internals: the Python callback upon WM_TIMER didn't
INCREF the called function object itself. When the function removed
itself from the hook list (indirectly), the executing functions
closure/cells.. were recycled while the function was still executing.

https://sourceforge.net/tracker/index.php?func=detail&aid=1489690&group_id=78018&atid=551954

-robert