| | |
| | The precise rules are as follows: suppose that the |
| | result formatted with presentation type ``'e'`` and |
- | | precision ``p-1`` would have exponent ``exp``. Then |
- | | if ``-4 <= exp < p``, the number is formatted |
- | | with presentation type ``'f'`` and precision |
+ | | precision ``p-1`` would have exponent ``exp``. Then, |
+ | | if ``m <= exp < p``, where ``m`` is -4 for floats and -6 |
+ | | for :class:`Decimals <decimal.Decimal>`, the number is |
+ | | formatted with presentation type ``'f'`` and precision |
| | ``p-1-exp``. Otherwise, the number is formatted |
| | with presentation type ``'e'`` and precision ``p-1``. |
| | In both cases insignificant trailing zeros are removed |