Split crates

[u2]

I consider we could move kvdb in a separate crate. What about removing the avl or moving to another repo firstly, because it is written base on the trie, it's not well designed.

[zeroqn]

For kvdb, should we move hashdb.rs too?

If we don't plan to revisit current avl code in the future then it's ok to remove it.

[zeroqn]

I think our common contains too many crates. If we want to split db into even smaller crates, I think we should move some of crates out of common. Need a big plan here. For example, move all pubsub into pubsub repo. Move hash creates into hash repo.

[u2]

Totally agree with you.

[u2]

About merkel:

• [x] add a TreeBuilder in cita-merklehash
• [x] rename merklehash to static-merkle
• [x] fix the test, merkelhash should build without any features.
• [x] remove null
• [x] "叶子节点怪异折行换掉"

[u2]

https://github.com/cryptape/cita-common/pull/182

[kaikai1024]

Do we consider to publish crates?

[u2]

That depends, we have published some before, like rust-numext and static-merkle-tree.

[kaikai1024]

CITA is almost stable, i think it's time to use the common(split crates) version as other crates.

https://github.com/cryptape/cita/issues/220

[kaikai1024]

What is the progress ?

[u2]

What is the progress ?

Have finished partially, but I think we can do better, it's a long term improvement.

[kaikai1024]

All the independent crates(will be) are maintained in this repo except cita-web3 and logger. Maybe in the future instead of now. The reason is that cita-common needs Cargo.lock