Sorry for the inconvenience caused by my frequent PR...
Several days ago, I created a PR to fix the bug of masstree key-compare function, but I found that the bug in scan function may not be fixed correctly. I assumed that memcmp was only used for judging whether two keys are equal or not, but it is also used in scan function to judge "more than", which I haven't noticed before. Therefore, creating a keycmp function is of necessity. I used a keycmp function to do element-wise uint64_t comparison for keys.
However, I can't construct a testcase now for this bug... Maybe a testcase constructor is needed for it.
Sorry for the inconvenience caused by my frequent PR...
Several days ago, I created a PR to fix the bug of masstree key-compare function, but I found that the bug in
scan
function may not be fixed correctly. I assumed thatmemcmp
was only used for judging whether twokeys
are equal or not, but it is also used inscan
function to judge "more than", which I haven't noticed before. Therefore, creating akeycmp
function is of necessity. I used akeycmp
function to do element-wiseuint64_t
comparison for keys.However, I can't construct a testcase now for this bug... Maybe a testcase constructor is needed for it.