You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The registry can lie to clients that there are no entries for a package because there is no way to prove that at a Map Checkpoint there was no value associated with a given key.
In the REST API, requesting to prove that a log was empty can be represented as
This PR will change how the sparse merkle tree is optimized. Currently
it implements certain optimizations that are beneficial in some regard,
but there are trade-offs. [Other
literature](https://github.com/google/trillian/blob/master/docs/papers/RevocationTransparency.pdf)
references this other optimization, which after discussion, we find has
benefits that we would like. Specifically, it will come in handy with
respect to the implementation of [this
issue](#66). [These
notes](https://hackmd.io/6RI9ZyGqR1ykRITggTb-pg) may be helpful for
understanding the refactor.
---------
Signed-off-by: Daniel Macovei <[email protected]>
Co-authored-by: Kyle Brown <[email protected]>
The registry can lie to clients that there are no entries for a package because there is no way to prove that at a Map Checkpoint there was no value associated with a given key.
In the REST API, requesting to prove that a log was empty can be represented as
A proof that a key was empty is a walk from the first non-empty ancestor of the leaf for that key up to the root.
The text was updated successfully, but these errors were encountered: