Anti-entropy inefficient

[Chat-Wane]

The current mechanism to get the missing elements of the document ( to catch up after being offline ) is inefficient. C.f. ./lib/application => getOperation.

Currently the complexity is k.n^2 where k is the average size of the identifier, n is the number of elements.

Could be improved using the whole pack of operations requested.
Could be improved with a direct access to the tail of identifiers.