Db namespaces added to newKvt API

[web3-developer]

This change adds a namespace parameter to the newKvt API. The namespace parameter is a required field and so must be provided. For convenience I've created another proc called newDefaultKvt which will use the default namespace. The older kvt legacy kvt API doesn't have this namespace parameter so we don't need to update all the existing code that uses the older API. Inside the legacy backend the kvtMethods proc checks the namespace and if it is not equal to the default namespace then it adds the prefix byte to the key. This is done to keep the functionality the same while allowing us to change the way we use the newKvt API. When using the new API and setting a non default namespace we should not add the prefix byte before passing in the key.

Currently the legacy database and many of the tests rely on the use of the prefix byte existing in the keys and so it is very difficult to introduce the namespace and remove the byte prefix all at once so I've opted to do the change in smaller steps. This change doesn't yet use the column families in RocksDb, it simply introduces changes to the newKvt API.

Next steps:

• Add support for the RocksDb column families in the Aristo backend
• Over time update all the existing code to use the newKvt API and cleanup the code that is setting the prefix byte before passing in the keys.
• Once all existing code is using the newKvt API then Nimbus could switch over to the Aristo backend.