The C implementation supports both formats "n" and "N". The Python implementation only supports format "n".

>>> from decimal import Decimal
>>> format(Decimal('1e100'), 'n')
'1e+100'
>>> format(Decimal('1e100'), 'N')
'1E+100'
>>> from _pydecimal import Decimal
>>> format(Decimal('1e100'), 'n')
'1e+100'
>>> format(Decimal('1e100'), 'N')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/serhiy/py/cpython/Lib/_pydecimal.py", line 3766, in __format__
    spec = _parse_format_specifier(specifier, _localeconv=_localeconv)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/serhiy/py/cpython/Lib/_pydecimal.py", line 6194, in _parse_format_specifier
    raise ValueError("Invalid format specifier: " + format_spec)
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: Invalid format specifier: N

Interesting. I think the behaviour of the Python implementation behaviour is actually more correct here: neither `int` nor `float` supports 'N', and I'm not seeing any indication in tests or documentation that 'N' should be supported. So is this a bug in libmpdec, or a missing feature in the Python implementation? (Either way, it's definitely a bug that the two aren't aligned.)

>>> format(123, 'n')
'123'
>>> format(123, 'N')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: Unknown format code 'N' for object of type 'int'

I don’t think “N” should be supported. It was never the intention to have it work.

Eric, Serhiy: do you have opinions on the right way forward?

Here are 6 options, on a spectrum of increasing level of acceptance of "N".

-2. Remove "N" support for cdecimal right now (i.e., for Python 3.11), on the basis that there's no need for deprecation warnings, because it was never officially a feature.
-1. Deprecate "N" support for cdecimal, remove it in Python 3.13.
0. Do nothing (the default), leaving _pydecimal and cdecimal inconsistent.
1. Add "N" support to the Python implementation for parity with cdecimal, but don't document it - leave it as an undocumented feature.
2. Officially add "N" support to decimal formatting - add documentation, tests, and fix the Python implementation.
3. Officially add "N" support to all numeric formatting ...

I could be persuaded for any of options -1, 1 and 2. I don't much like option 0.

I support deprecating "N".

I'd support -1.

I have no strong opinion. I found the discrepancy when reviewed one of Mark's PRs. I was surprised because I did not know that "N" is supported.

On other hand, there are reasons for supporting upper-case variant of "n" if we support upper-case variants of "e", "g" and "x" (formats which produce output containing letters).

There is an alternative solution of the problem solved by upper-case variations of formats: add a special converter for converting a string to upper case, so for example f'{x:12.8E}' could be written as f'{x:12.8e!u}'.

I prefer -1 or 2. If we choose deprecation I prefer adding an upper-case converter.

While I'd prefer -1 overall, I would also prefer 3 over 2. If we were adding the feature from scratch, we wouldn't have decimal differ from int, float, and complex.

And I'm not in favor of an uppercase converter, no matter what we do here. The other converters work like:

f'{obj!s}' -> format(str(obj))
f'{obj!r}' -> format(repr(obj))

but the proposed !u would be

f'{obj!u}' -> format(obj).upper()

That is, it operates on the result of __format__, not its input.

You are right about an uppercase converter. I forget this. I withdraw this proposition.

It seems that even if we add support of "N" to all numeric formatting it will not cover all use cases. "n" is a locale specific variant of "g", but other formats could have locale specific variants too. In C all numeric formatting is locale specific. Maybe add a flag which makes any numeric formatting locale specific? Then "n" will became obsolete.