trie-db

• Improve TrieCache size by reducing size_of NodeOwned #216

• Fix backward iterator seek #215

• Implements DoubleEndedIterator for trie iterator #208

• Make trie_nodes_recorded_for_key work for inline values #194
• trie-db: Fetch the closest merkle value #199
• fixing triedbmut lookup, added some testing in test. #198

• Fix TrieDBRawIterator::prefix_then_seek #190

• Fix compact proof to skip including value node hashes #187
• Update dependencies. #188

• TrieDBRawIterator and TrieDBNodeIterator now return a node inside of an Arc instead of Rc #185
• TrieDBRawIterator is now Send and Sync #185

• Fix recorder behavior: #184

• Updated hashbrown to 0.13.2: #177
• Iterator refactoring: #181
  • Removed TrieDBKeyIterator::suspend and SuspendedTrieDBKeyIterator
  • Added TrieDBRawIterator

• Do not check for root in TrieDB and TrieDBMut constructors: #155

  To get back the old behavior you have to add the following code:

  if !db.contains(root, EMPTY_PREFIX) {
    return Err(InvalidStateRoot(root))
  }
  

• Introduce trie level cache & recorder: #157

  This pull requests introduced a cache that is directly baked into the trie-db. This cache can be used to speed up accessing data in the trie. Alongside the cache, the recorder was also made a first class citizen of the trie. The pull requests introduces quite a lot of changes to the trie api. TrieDB and TrieDBMut are now for example now using a builder pattern. For more information see the pr.

• Updated hashbrown to 0.12. #150

• Support for value nodes. #142

• Updated hashbrown to 0.11. #131

Added

• Decode with an iterator. #121

• Update hashbrown to 0.9. #118

• Use ahash feature of hashbrown. #103

• Update hashbrown to 0.8. #97

• Added ALLOW_EMPTY to TrieLayout. #92

• Prefix iterator. #39
• Update trie-root to v0.16.0 and memory-db to v0.19.0 and reference-trie to v0.20.0 #78

• Compact proofs support. #45