Merklization audit #157
Merklization audit #157
Conversation
dr-orlovsky
added
*security*
| _25=25 _26=26 _27=27 _28=28 _29=29 _30=30 _31=31 | ||
| } -- U5 | ||
| cofactor is U16 | ||
| crossSection list len=1..MAX32 |
There was a problem hiding this comment.
Just confirming:
I reread LNPBP-4 because of this change and in fact, we will always have at least one commiment in the MPC. This change reflects the statement above, right?
| } -- U5 | ||
| entropy is U64 | ||
| cofactor is U16 | ||
| messages map len=0..MAX24 |
There was a problem hiding this comment.
Similar to question above, we dont need at least one message in the merkle tree?
There was a problem hiding this comment.
Hm.. On one side no: we may have a merkle tree made of entropy (and its cross-section still will not be an empty set). However this has low utility, so it could be we can prohibit that...
There was a problem hiding this comment.
Giving a second thought: there might be a need to tress with no messages and just an entropy. While it's marginal, even if we require a non-empty message set, people can fake it and still produce trees containing no real messages. Thus, while it is mathematically impossible to have a tree cross-section which is empty, it is mathematically possible to have a tree with no messages, made only of entropy - thus, the difference in these types.
No description provided.