Change uint_8 to x times binary 1 MSB

Question

I'm writing a Max72xx library for my school project and I have found a way to set columns using the registers that comes with the datasheet. Now, I also found a way to set their height. So I have created a function with two parameters: row, which basically sends the x-axis to the chip, and height which sets the y-axis.

What I did in this code is basically the following:

• The user sets an x-axis between 1 and 8 (parameter const uint8_t& row)
• The user sets an y-axis between 1 and 8 (parameter uint8_t height)
  

I wanted to achieve the following:

• If the height is set to three: the following bit pattern should be created: 1110 000. Or, if the user uses 8, I want the function to create 1111 1111. Or if the user sets 5 as y-axis I want the fucntin to create a bit pattern of 1111 1000, etcetera.

The code I have provided is the following (this piece of code is within the function setColumn):

if (height >= 1 && height <= 8) { // if the height is between the matrix range
    uint8_t pattern = 0x00; // create a pattern to send to the chip
    uint8_t counter = 0; // counter for counting the zeros to shift left later on

    for (uint8_t i = 0; i < height - 1; i++) { // this loop is repeated height amount of times
        pattern |= 0x01; // create 1s in the pattern
        pattern <<= 0x01; // shift it 1 to the left
    }
    pattern |= 0x01; // OR the LSB bit of the pattern

    for (uint8_t i = 7; i > 0; i--) { // count the leading 0 bits to shift them to the MSB position
        if (!(pattern >> i) & 1)
            counter++;
        else break;
    }
    pattern <<= counter;

The MSB is the lowest pixel on the display, the LSB the highest pixel. Now this code works as I described earlier, but I think there is an easier and more efficient way to work this out. I'd like to know some suggestions. Thanks in advance.

Answer 1

Just use this:

uint8_t pattern = (0xff00u >> height);

Or this:

uint8_t pattern = (0xffu << (8 - height));

Don't be afraid of overflows during bit shifting as long as the arguments are unsigned, the behavior is well-defined.