uraimo / bitter Goto Github PK

The last line of the first example of the Specific Byte subscript section reads:

// Let's change the third byte
i[3] = 0xAA   //Error! There is no 3rd byte!

However, there's no actual error here because there is a 3rd byte. I was following along in code and at first didn't catch that there's no error, so I thought maybe that Bitter was just silently ignoring the out-of-bounds index since nothing appeared to change. But of course nothing would change by setting byte 3 to 0xAA since it's already 0xAA.

Anyway, I'd suggest something like this instead:

// Let's change the fourth byte
i[4] = 0xFF   //Error! There is no 4th byte!

Currently, a GYB template is used to generate specific extensions for each integer type.
Even if it's quite easy to figure out how GYB templating works, migrating the template to Sourcery will make the template seem a little less esoteric.
The bash script that generates the classes will have to be updated too.

• Functions to change endianness of the bitpattern content.
• Functions to create patterns in addition to allZeroes and allOnes (e.g. 5 ones) to easily generate masks.
• Migrate the template to Sourcery?

Tagged as 1.2.0, i'm not going to push the pod yet (something else will surely change before the Swift 3.0 is released).

A small method that generates bit mask for a given number of bits could be useful sometimes.
It could look something like mask(bits:msb:), where bits would be the number of bits that need to be set to 1 and msb a boolean(no need for an enum here) that select if those bits need to be set from the MSB(right) or LSB(left).

Note: Since .allZeroes has been deprecated in Swift 4, when this method is available .allOnes should be removed, it makes no sense to keep it around when mask() can do the same thing and its sibling it's not around anymore.