When looking at the code-bit processing you will often see :( variable) with (their complement) to do bits or , that is, x | -x
The result of the process is this: get a value of y, y start from the lowest number of consecutive 0, then x is the same.
E.g:
1）
1010 1011 1100 1101 --> x
0101 0100 0011 0011 --> -x
1111 1111 1111 1111 --> x|-x
2)
1010 1011 1100 1000 --> x
0101 0100 0011 1000 --> -x
1111 1111 1111 1000 --> x|-x
3)
1010 1000 0000 0000 --> x
0101 1000 0000 0000 --> -x
1111 1000 0000 0000 --> x|-x
Have this effect due to the definition of complement - adding all 1 bits inverted, and this took from a lowest-order bit first stretch of consecutive 1's all become 0 (since the carry), just the result of this 0 original code 0 in the same, or when the incumbent is still zero.
Reproduced in: https: //www.cnblogs.com/mmars/archive/2013/05/13/3076062.html