Tracking Issue: Combine with Sovereign-Labs/jmt

[preston-evans98]

Overview

Sovereign Labs and Penumbra are combining efforts to make one JMT to rule them all. This is a tracking issue for all of the changes from the Sovereign Labs JMT that we want to port over to this repo:

• Separate values from Leaf Nodes Separate LeafNodes from values #77
• Convert Children from a HashMap to an array Convert Children from a HashMap to an array #75
• Make the JMT generic over hash function Make JMT generic over hash function #79
• Feature gate metrics. See https://github.com/Sovereign-Labs/jellyfish-merkle-generic/blob/master/src/metrics/mod.rs
• Send all updates through the batch_insert API to share intermediate hashes
• Add optional parallelism with rayon
• Add optional borsh serialization #84
• Add a mechanism to share intermediate hashes when reading multiple values
• Consider changing TreeReader::get_value to use Key instead of KeyHash
• Support no_std. Support no_std #87

[hdevalence]

Consider changing TreeReader::get_value to use Key instead of KeyHash

What if we made it take K: Into<KeyHash>?

[preston-evans98]

What if we made it take K: Into<KeyHash>?

This is a good suggestion. I'll open a PR this afternoon.

The underlying motivation for this item is to allow apps to efficiently retrieve a sequential keys using an iterator though, so just switching to Into<KeyHash> won't be sufficient. I'll open a separate issue where we can explore the tradeoffs. #81

[preston-evans98]

Hmm, on further thought I think this change will prevent TreeReader from being object safe

[zbuc]

In our weekly sprint planning, we pulled in all outstanding JMT issues and are wondering what the status of this one is. @preston-evans98 is there anything that needs to be done on our end?

[preston-evans98]

@zbuc Nothing immediate. I'll be trying to add no_std support this weekend, but the diff should be minimal and will just need review. I don't think we have any AIs here for the very short term