Question

I need to use bit flags with more than 32 bits (33 to be exact right now). I tried and find std::bitset doesn’t handle more than 32 bits (ulong). Do I have to use vector or there’s a way to make bitset to work?

I am limited to c++98 in this project so I can’t use boost.

Thanks.

Edit:

I’d like to do something like this:

const uint64    kBigNumber = 1LL << 33;
std::bitset<33> myBitSet;
...
switch(myBitSet) {
    case kBigNumber:
    // do something
    ...
}

Answer 1

I’ve just retested std::bitset with 65 bits and on my 32-bit Linux it works fine and as expected.

Notable exception is the to_ulong() method which throws exception if any set bit would be truncated during the conversion. Now I think about it and that is rather obvious: there is no other way as to prevent application from getting truncated data. And the behavior is also documented.

---

To the Edit with switch/case. Why do you need std::bitset then? You platform apparently already supports 64 bit numbers – use them. std::bitset is designed to be used as an light-weight bit array with static memory allocation. It is not intended to be used as a replacement for number.