Bug#341345: gcc-4.0: Please warn on impossible && conditionals

[Paul Harris <harris.pc@gmail.com>]

Hi Justin,

see lower ...

Justin Pryzby wrote:
> Package: gcc-4.0
> Version: 4.0.2-4
> Severity: wishlist
> Tags: upstream
>
> gcc will warn on something like
>
>   char c;
>   // ...
>   if (c>256) {
>   // warning: comparison is always false due to limited range of data type
>   // ...
>   }
>
> Please consider also warning on things like:


I would imagine this below:

>   if (0==c && 1==c)

isn't a good idea, because it would break symmetry with custom 
operator==() behaviour.

eg, if I did

class MySillyClass { ... };

inline bool operator==(MySillyClass const& a, int b)
{
   return (b == 0 || b == 1);
}


then:

char c;
if (c == 0 && c == 1)  // WARNING

MySillyClass d;
if (d == 0 && d == 1) // should NOT warn


then, if i did:

template <class T>
bool test(T x)
{
   return (x == 0 && x == 1);
}

test() would warn on some types and not others.  That would be annoying!


right?
thanks for the brain snack,
Paul