Generate warning on possible confusing use of operators with same precedence?

[zevv]

On #nim the following came up: consider evaluating the following expression:

not 3 == 4

The above evaluates left-to-right as (not 3) == 4 since not and == have same precedence, although the user actually hoped for not (3 == 4) to happen.

Easily explained by pointing to the manual, but still a potential pitfall.

Modern C compilers tend to notice this and advice the user to use parenthesis. For example, clang says:

ico@platdoos:/tmp$ clang -c t.c -Wall
t.c:4:9: warning: logical not is only applied to the left hand side of this comparison
      [-Wlogical-not-parentheses]
        return !3 == 4;
               ^  ~~
t.c:4:9: note: add parentheses after the '!' to evaluate the comparison first
        return !3 == 4;
               ^
                (     )
t.c:4:9: note: add parentheses around left hand side expression to silence this warning
        return !3 == 4;

Is this something Nim should signal and warn about as well?

[Araq]

Instead we should have had bitnot, bitand and bitor...

[mratsim]

But the rule is that unary always binds stronger than binary so they do not have the same precedence since now not is always unary.