We should create a preliminary getting started guide in the readme so that people will understand how to use honu; e.g how to use it if they are already, say, a levelDB user.

We want to increase confidence in the honu code by testing edge cases. A powerful tool for this is fuzz testing

• Implement a Fuzz test in honu_test

Investigate what transactions look for each database and how much work would need to be done to implement in Honu

What other features do these other databases offer vs. level db

what are best practices for database optimization

Research namespaces/indices/buckets with each database (how does it support groupings of objects)

etc.

REF: SC-1275

We want to increase confidence in the versions code by testing edge cases. A powerful tool for this is fuzz testing

• Implement a Fuzz test in versions_test

Reference lines -

Implementation does not copy tombstone data into the parent version. This creates a problem when backtracking through history. A deleted object will not have a stored tombstone transition and can only be identified by looking at nil data.

Storing the tombstone allows for a reconstruction of the history of the object, which is necessary when defining expected behavior for users.

Line 61 just needs:

Tombstone: meta.Version.Tombstone,

Bonus points for adding a complete clone() method for taking care of Version copying. :D

Fix the issue where LevelDBEngine.Iter currently grabs the namespace by splitting on double colons (::). In the case of a namespace called for example "namespace::with::colons", Iter considers the namespace to be "namespace" and will not access the database correctly. As for a possible solution, with PR #12 the namespace is stored in the iterator, meaning we would no longer need to split on double colons. The following tests are currently skipping double colon test cases and need to be amended:

• engines/leveldb/leveldb_test: TestLevelDBIter line 163
• honu_test: TestLevelDBInteractions line 119