NMT clarify leaves hashing

[liamsi]

It’s not clear if the namespace is also fed into the hash h as this might or might not be the case depending on what the serialize function looks like. I though I remembered that serialize was narrowed down by specifiying if serialization contains the NID or not but I can't find it currently. In the current form the spec would allow both cases as far as I understand. From a security pov both should be fine as the NID is returned / prefixed in front of the hash anyways (and hence the tree commits to the namespace of each leaf too).

For leaf node node of data d:

node.n_min = d.namespaceID
node.n_max = d.namespaceID
node.v = h(0x00, serialize(d))

https://github.com/lazyledger/lazyledger-specs/blob/master/specs/data_structures.md#namespace-merkle-tree

In the implementation it is not and a leaf hash is computed as:
ns(rawData) || ns(rawData) || hash(leafPrefix || rawData) where rawData does not contain the namespace.
or more precisely:
https://github.com/lazyledger/nmt/blob/4b4c58a9e7c071db650b27a69dccde4821411d07/internal/hasher.go#L63-L67

---

IMHO, this is a perfect candidate to start putting #108 into practice.

[adlerjohn]

Yeah, you're right the spec is unclear what specifically the leaf contains. Either would work, but it needs to be unambiguous.