mbed

if you #include <vector>, you can use std::vector<bool> which uses bits to represent elements. But why do you need it? mbed should have enough memory for a char array of a reasonable size, or even int array, and both will likely be faster than the bit array.

because in my PCA9532 program, I wan't an int (16 bit), to represent each LED (1 as on, 0 as off).
For the expander to do this, it needs at least four char's to be sent to it. A char may look like this for turning every second LED:
0b01-00-01-00 <- This is where my single bit array comes to life:

0b01-01-01-01-01-01-01-01 <- Converts to -> 0b00-01-00-01, 0b00-01-00-01, 0b00-01-00-01, 0b00-01-00-01

So I load the int into and array using for loop, but, I only wan't to use the LSB of a char, or a single bit.

This is where I need the single bit.

/ Lerche

There's no (AFAIK) easy way of converting a bool vector to an integer, so you could make a simple wrapper for setting bits in an integer:

void PCA9532_set_leds(uint16_t bits)
{
  // send 'bits' value to PCA9532
}

//16-bit bit array
class Bits16
{
  uint16_t m_bits;
public:
  Bits16() : m_bits(0) {};
  // set a bit 'bitno' to 1
  void set_bit(int bitno)   { m_bits |= (1<<bitno); };
  // set a bit 'bitno' to 0
  void clear_bit(int bitno) { m_bits &= ~(1<<bitno); };  
  // return the combined integer value
  operator uint16_t() { return m_bits; };
};

Bits16 leds;
int main( void )
{
    for ( int i=0; i < 16; i++ )
    {
      leds.set_bit(i);
      PCA9532_set_leds(leds); // 'leds' is automatically converted to uint16_t
    }
}