Missing carets when reporting an "object is not callable" error

[PierreQuentel]

Bug report

Bug description:

There is a minor inconsistency between the way "not subscriptable" and "not callable" errors are reported.

When an object is not subscriptable, the exception trace includes the carets ~ and ^ for the subscript value and slice.

>>> t = 0
>>> t[0]
Traceback (most recent call last):
  File "<python-input-1>", line 1, in <module>
    t[0]
    ~^^^
TypeError: 'int' object is not subscriptable

The carets are also displayed if the attempted subscript is assigned

>>> x = t[0]
Traceback (most recent call last):
  File "<python-input-4>", line 1, in <module>
    x = t[0]
        ~^^^
TypeError: 'int' object is not subscriptable

It is the same when an object is not callable

>>> t(0)
Traceback (most recent call last):
  File "<python-input-2>", line 1, in <module>
    t(0)
    ~^^^
TypeError: 'int' object is not callable

but when the attempted call is assigned, we lose the carets

>>> x = t(0)
Traceback (most recent call last):
  File "<python-input-5>", line 1, in <module>
    x = t(0)
TypeError: 'int' object is not callable

instead of the expected

>>> x = t(0)
Traceback (most recent call last):
  File "<python-input-5>", line 1, in <module>
    x = t(0)
        ~^^^
TypeError: 'int' object is not callable

CPython versions tested on:

3.13

Operating systems tested on:

Windows

[Eclips4]

Hello!
Thanks for the report.
After taking a look at Lib/traceback.py:

cpython/Lib/traceback.py

Lines 712 to 723 in bbd3300

	match statement:
	case ast.Return(value=ast.Call()):
	if isinstance(statement.value.func, ast.Name):
	value = statement.value
	case ast.Assign(value=ast.Call()):
	if (
	len(statement.targets) == 1 and
	isinstance(statement.targets[0], ast.Name)
	):
	value = statement.value
	if value is not None and _spawns_full_line(value):
	return False

It looks like intended behavior, cc @pablogsal

[Eclips4]

Ah, yes, this is expected. See #99180.

[PierreQuentel]

Yes, this is how it works now, but it's hard to see why assigning to a failing subscript (x= t[0]) gives a different result from assigning to a failing call (x = t(0)).

[Eclips4]

Yes, this is how it works now, but it's hard to see why assigning to a failing subscript (x= t[0]) gives a different result from assigning to a failing call (x = t(0)).

Please take a look at the #99180.