assigned to @mclow

What do you think this code does?
(and what would you suggest replacing it with?)

Second question first: The minimal change would be to replace the “&” with “&&”, which silenced the warning, obviously has the same effect, and might be what the author originally intended. I hesitated to suggest a change of more than a single character change because I don’t understand why the author wrote it this way; e.g., there may be performance issues I’m not aware of.
I think the intent was just:
else if (__n != 0 && __n != 1)
which is what I would prefer if there’s no reason not to make the code clear.

Second question first: The minimal change would be to replace the “&” with
“&&”, which silenced the warning, obviously has the same effect, and might
be what the author originally intended.

Sorry; that is incorrect.

I think the intent was just:
else if (__n != 0 && __n != 1)
which is what I would prefer if there’s no reason not to make the code clear.

This is incorrect as well.

Consider the following code:

#include

bool foo(unsigned x) { return x & x - 1; }
bool bar(unsigned x) { return x && x - 1; }

int main(int, char*[])
{
for (unsigned i = 1; i < 20; ++i)
if (!foo(i))
std::cout << i << " ";
std::cout << std::endl;
for (unsigned i = 1; i < 20; ++i)
if (!bar(i))
std::cout << i << " ";
std::cout << std::endl;
}

Does this print the same sequence twice? (Answer: No)

The expression (__n & __n - 1) is a check to see if n is not a power of two.
See http://graphics.stanford.edu/~seander/bithacks.html#DetermineIfPowerOf2.

Duh, in that case all I ask is an annotation as in PR 25706.

Annotation added in revision 353448.