django.utils.encoding.smart_unicode/smart_str handle strings_only inconsistently

[Aymeric Augustin]

smart_unicode calls is_protected_type, smart_str just checks for None and int.

It seems to me that smart_str should rely on is_protected_type too, for consistency.

If there's a reason for implementing a different check, it should be at least explained in a comment.

[Claude Paroz]

+1 to use is_protected_type in force_bytes

[Aymeric Augustin <aymeric.augustin@…>]

In e492ab8e7eb5fc2c1d4a50669aeaa3d6163dc5f3:

Fixed #18719 -- Made force_bytes more consistent with force_text.