curses.napms() aborts with a SystemError

[devdanzin]

Crash report

What happened?

The abort happens on normal and no-gil debug builds, release builds also hit the SystemError.

import curses
curses.napms(37)

Error is:

Fatal Python error: _Py_CheckFunctionResult: a function returned a result with an exception set
Python runtime state: initialized
_curses.error: must call initscr() first

The above exception was the direct cause of the following exception:

SystemError: <built-in function napms> returned a result with an exception set

Current thread 0x00007fe41a8c3740 (most recent call first):
  File "/home/fusil/python-73/curses-fatal/source.py", line 3 in <module>
Aborted

Found using fusil by @vstinner.

CPython versions tested on:

3.13, 3.14, CPython main branch

Operating systems tested on:

Linux

Output from running 'python -VV' on the command line:

Python 3.14.0a1+ experimental free-threading build (heads/main:d467d9246c, Nov 1 2024, 09:05:56) [GCC 11.4.0]

Linked PRs

• gh-126313: Fix a crash in curse.napms due to incorrect error handling #126351
• [3.13] gh-126313: Fix a crash in curses.napms() due to incorrect error handling (GH-126351) #126493

[picnixz]

Ah I can see what happened. The macro returns 0 so that it can be used as a substitute for NULL in most of the functions it's being used but _curses.napms returns an int. Instead, it should return -1. It's an easy fix (one should simply remove the macro usage for this one and do something like that:

 if (!_PyCursesStatefulCheckFunction(module,
								     curses_initscr_called,
									 "initscr")) {
	return -1;
}

I think it should suffice. I won't make a PR on Saturday (UTC+2) but probably on Sunday.

[vstinner]

Fixed by change 19d9358.

[serhiy-storchaka]

The fix was not backported yet.

[picnixz]

Update: No need for a 3.12 backport because the implementation returns a PyObject * instead of an int. Only a 3.13 backport is needed (which I am taking care of now).

[vstinner]

I merged the 3.13 backport.