some suggested documentation enhancements for range() and noDuplicates about avl HOT 2 OPEN

I do agree that this is not obvious. But I guess the returned nodes in the tree with duplicates are not exactly random, they just depend on the traversal direction and insertion order.
Actually would be interesting to introduce a test for that

from avl.

Yup that's true they're not random in the rand() sense, but for the library user who is not aware of the implementation, it's one of those "it is implementation-defined which node you get" which is effectively the same :)

It's better to have predictable behavior if the user has chosen noDuplicates===false

FYI I just edited the OP above to add noDuplicates-related dox to range() too. I realized the same issue applies to range()

from avl.