Fix tests to remove address_to_object_id_hack hack

[patrickkuo]

• removed unused amount args
• renamed init_state_with_account to init_state_with_object
• added get_object_id function to generate random object id for testing

removing the rest of "address_to_object_id_hack" from client.rs will be more involve and require adding object_id(s?) to ClientState, I will keep this PR short and do it in part 2

[huitseeker]

One quick comment, hope this helps.

I recommend setting a comfortable environment where running cargo fmt is easy for you on any commit, perhaps with a git pre-commit hook or an IDE setup?

[huitseeker]

So, in production we're going to need this to be 100% deterministic, based on a set of inputs (e.g. a hash of the object bytes + the Tx creating it, etc).

The reason for this is that the output of the computation of distinct machines on the network needs to be the same, as the success of a unit of computation will depend on a pool of machines producing a collective cryptographic signature on that output. They need to sign "the same bytes".

What you have would be fine for local unit tests, though.

Here's a set of ideas on how to go about this while we're figuring out how to do this:

• conditional compilation can let this function have different code (hence meaning) when running tests and in production,
• the rand Rust crate contains the RNGs we're looking for,
• therein, the SeedableRng trait captures the kinds of Rng we would pick here, and I'd recommend using an API like it to generate the ObjectID: in production, every ObjectID is generated from runtime seed bytes (TBD) + a PRNG, with the details of the later being set at compilation time.

If you go look into the rand crate, a tip if not to change its version (which is not the latest) for right now (for reasons disconnected to this PR).

[patrickkuo]

Thanks!

This method is intended to be use in testing only, I should move it out of the core package

[gdanezis]

You can also prepend the function with #[cfg(test)] and it will only be included when compiling tests.

[patrickkuo]

removing the rest of "address_to_object_id_hack" from client.rs will be more involve and require adding object_id(s?) to ClientState, I will keep this PR short and do it in part 2

[sblackshear]

LGTM