zack-bitcoin
commented
5 years ago Cool. I like how you used a set to store these values. That is exactly the correct data-structure for maintaining a white list.
The part I don't like is how you re-calculate the same set every time you use it. Every time ?WL happens, the white_list() function is called, and the set is built again. Building the entire set just to look up one thing from the set is not efficient.
It seems to me our choices are: 1) we could maintain a set of white list peers in ram. This algorithm would be O(log(length of white list)), and would require making a new gen_server, around 30 lines of code.
2) we could have a hard-coded list of peers in the config file, and scan through the entire list every time anyone sends our server a message. This algorithm would be O(length of white list), and would require writing like 4 lines of code.
It seems like (1) is more complicated, but (2) is more likely to result in speed issues.
I will merge your changes, and then make some more changes to achieve goal (1).
I just noticed your user name. Now I am considering not merging, and re-writing your idea myself. I don't want "charles manson" in the list of contributors.
Why are you called Charles Manson? I don't want that name associated with Amoveo.