The only way to assert unreachable code is to place the assertion, assert(0), within the suspected unreachable code itself.

However, just because the assertion never triggers it does not follow that the code is unreachable -- only that the condition(s) required to reach that code segment have not thus far arisen. How certain can you be that the condition will never arise?

Unreachable code can sometimes be detected at compile time, depending on the compiler. Simple errors such as placing a return statement before unreachable code can normally be detected this way. However, logic errors may not always be as obvious.

For instance,

if( !condition )

// do something.

else if( condition )

// do something else.

else

assert( false ); // unreachable...

In general, if there is no logical proof to say a code segment is unreachable, it's best to leave the segment in place. An assert( false ) will alert you when the segment is hit, but even if it is never hit, you cannot be certain it will never be hit without a logical proof. If the unreachable code is trivial then it is usually not worth the effort to prove it is unreachable. If it is non-trivial, the only issue is how you test it in the event it is actually reachable. In that case you must force the conditions that would make it reachable.