0 + not 0

[Ian Kelly]

On Sat, Jul 11, 2015 at 10:02 AM, Stefan Ram <ram at zedat.fu-berlin.de> wrote:
>   I look at Python 3.4.3:
>
> a_expr ::=  m_expr | a_expr "+" m_expr | a_expr "-" m_expr
>
>   So, »not 0« must be an »m_expr« when used as the right operand of »+«.
>
> m_expr ::=  u_expr | m_expr "*" u_expr | m_expr "//" u_expr | m_expr "/" u_expr | m_expr "%" u_expr
> u_expr ::=  power | "-" u_expr | "+" u_expr | "~" u_expr
> power ::=  primary ["**" u_expr]
> primary ::=  atom | attributeref | subscription | slicing | call
> atom      ::=  identifier | literal | enclosure
> enclosure ::=  parenth_form | list_display | dict_display | set_display | generator_expression | yield_atom
>
>   How can there be a »not«?
>
>   »not« is used in
>
> not_test ::=  comparison | "not" not_test
> and_test ::=  not_test | and_test "and" not_test
> or_test  ::=  and_test | or_test "or" and_test
> conditional_expression ::=  or_test ["if" or_test "else" expression]
> expression_nocond      ::=  or_test | lambda_expr_nocond
> expression             ::=  conditional_expression | lambda_expr
>
>   , but an »expression« is not an »m_expr«.

I must concur. The grammar as written does not actually produce 1 +
not 0. I think it's still worthwhile opening a bug, because the
behavior is surprising and possibly not intentional.