fix: state sync

[miraclx]

Remove type_id from calimero_storage, which unblocks sync.

Address other classes of misuse of the crdt subsystem and questionable design

[MatejVukosav]

Can you elaborate more on misuse and questionable design? What was wrong?

how removing type_id unblocks the sync?

[chefsale]

ship it

[miraclx]

Can you elaborate more on misuse and questionable design? What was wrong?

Misuse: Interface's functions were unintuitive and our attempts to abstract over it to provide easier-to-use APIs through Unordered{Set,Map} & Vector still required calling Interface::save on the collection after all mutations happen. When to call that and by whom was an open question, which led to one class of issues.

Questionable design:

• Many functions in index.rs call get_index multiple times to extract bits of unmutated information, this is still the case, but it's been reduced to an extent.
• Structures contained information like creation, update, merkle hash, etc.. that would be sent over the wire unnecessarily.
• Many functions in interface.rs required a definition of T: Data which led to the type_id requirement.

how removing type_id unblocks the sync?

• Just look at __private.rs from before. Since functions required knowledge of which type had which data updated, we had to exhaustively be aware of all types that constitute the state. That was unwieldy and led us down the path of __private.rs.
  That is no longer the case.