[Bug c++/77434] warn about suspicious precedence of ternary operator (?:)

[joseph at codesourcery dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77434

--- Comment #2 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
On Wed, 31 Aug 2016, joseph at codesourcery dot com wrote:

> > Code such as the following are suspicious:
> > 
> > int foo(int a, int b)
> > {
> > return (a > 0 && a <= (b == 1) ? 1 : 2);
> 
> Actually I don't think this is suspicious at all; it's just an ordinary 
> conditional expression and this precedence rule doesn't tend to confuse 
> people normally.  Certainly it's less dubious than various existing cases 
> warned about for precedence.

Or, if it's suspicious, it's only because of the boolean expression (b == 
1) in the RHS of <=, where the alternative parse makes more sense.  The 
following is not suspicious and it would seem silly to warn for it:

return (a > 0 && b <= 3 ? 1 : 2);

(because the suggested alternative parse would involve b <= 3 ? 1 : 2 as 
RHS of &&, which is unnatural as a conditional expression in boolean 
context where the halves of the expression aren't boolean).