Math.average #1170
Math.average #1170
Conversation
nventuro
added
feature
contracts/math/Math.sol
Outdated
| function average(uint256 _a, uint256 _b) internal pure returns (uint256) { | ||
| // (_a + _b) / 2 can overflow, so we distribute | ||
| uint256 c = (_a / 2) + (_b / 2) + ((_a % 2 + _b % 2) / 2); | ||
| assert((_a <= c && c <= _b) || (_a >= c && c >= _b)); |
There was a problem hiding this comment.
@leonardoalt how do you feel about these sort of assertions? Do you consider them necessary and correct, or do you think they may add too much unnecessary overhead (and gas costs)?
There was a problem hiding this comment.
I find the assertion unnecessary and somewhat arbitrary, as there are other conditions that should be checked in order to verify the correctness of the operation.
The implementation is good. Perhaps we should consider putting the operation in SafeMath, given that it does have overflow protection. What do you think?
|
I've always understood |
|
I've removed the assertion, since we don't yet have a criteria for this kind of stuff. Maybe once we get a FV module? |
The need for an
averagefunction has come up a couple times.I also wanted to try out the new
assertsemantics as described in #1120.