Hey! I'm currently researching how to simulate ove...

# python-contributors

[SerVB]

Hey! I'm currently researching how to simulate overflowing of signed integer operations in Python. For example, for Byte, how to get -128 from 128.

First attempt:

>>> int.from_bytes((128).to_bytes(2, 'little', signed=True)[:1], 'little', signed=True)
-128

Quite long. Second attempt:

>>> (128).__add__(0x80).__and__(0xFF).__sub__(0x80)  # From Hacker's Delight (Sign Extension)
-128

Or ((128 + 0x80) & 0xFF) - 0x80, but I found the former better – it will show that these are some low level operations, not the code from user's codebase.

The second attempt is better. Looks like this is the way I will use. Any ideas on making it even shorter?

In JS, bit operations are always 32-bit, so | 0 will truncate to 32 bits (to Int), and Byte and Short are made using shifts like 128 << 24 >> 24, so 1 operation for Int and 2 for Byte and Short. We have 3 operations for every integer type for now.

On the other hand, for unsigned integers, it seems it will be easy to just do something like (256) & 0xFF (for UByte) – this is cool.