Implement program executor

[Fumuran]

This PR implement new execute_program public procedure for executing arbitrary scripts.

[bobbinth]

Is there a reason not to use the interface I suggested in #1161 (comment)? Basically:

pub fn execute_program(
    &self,
    account_id: AccountId,
    block_ref: BlockNumber,
    tx_script: TransactionScript,
    advice_inputs: AdviceInputs,
) -> Result<[Felt; 16], TransactionExecutorError>

This would also allow specifying inputs for the transaction script (i.e., values that get put on the top of the stack before the script is executed).

The user could then generate TransactionScript using TransactionScript::compile() method which is a bit more general than just using a string.

[Fumuran]

Usage of the TransactionScript will indeed allow user to specify the transaction script inputs, but in that case creation of the final program will be more complicated.

To use this procedure in testing we need to execute the prologue before the provided code. If the code is provided as a string, I can just modify it, adding the execution of the prologue right after the begin. But if we have already compiled TransactionScript, we need to have something like a main script, which will execute the prologue first and then call the code provided by the user. I can implement it, but I thought that for the testing purposes the first approach will be sufficient.

[bobbinth]

But if we have already compiled TransactionScript, we need to have something like a main script, which will execute the prologue first and then call the code provided by the user. I can implement it, but I thought that for the testing purposes the first approach will be sufficient.

This is not just for testing purposes (i.e., it needs to work in the client under regular release). So, we should probably create another main file and use it here.