fix: return contract Addresses and allow passing hex when used as an contract arg

[paulbellamy]

What

fixes #661
fixes #662
fixes #403
replaces #663

Why

#663 missed a few other commands were we take contract IDs as args. Also, we cannot generically accept hex for strval::scaddress_from_json, because with just 32-byte hex, we cannot tell if it is a user address or a contract address.

Incidentally, making this more consistent also fixed a bug in soroban events --id 1, where the given contract id was not padded out to the full 32 bytes. Now, these are all equivalent:

soroban events --id 1
soroban events --id 0000000000000000000000000000000000000000000000000000000000000001
soroban events --id CAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD2KM

Those ids all also work for invoke, deploy, and read

Known limitations

• We should also update soroban-rpc to take a contract strkey in it's params, wherever it currently accepts a contract hex (e.g. getEvents). Then we can update this to pass that in.

There is one case where you cannot use a hex (padded or unpadded) as a contract Address. That is when you are passing an Address argument to a contract. For example:

# Deploy cross_contract contract 'a'
$ soroban contract deploy \
    --wasm ./soroban_cross_contract_a_contract.wasm \
    --source-account default \
    --id a
CAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAVAX5

# Deploy cross-contract contract 'b'
$ soroban contract deploy \
    --wasm ./soroban_cross_contract_b_contract.wasm \
    --source-account default \
    --id b
CAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAXI7N

# Here, we are required to pass the full contract strkey. Passing
# `a`, or `0a`, would not work, because there is no generic way
# to convert a 32-byte hex string to an `ScAddress`. An `ScAddress`
# might be either a user address or a contract address, and the hex
# doesn't have enough information to infer that.
$ soroban contract invoke --source default \
    --id b -- \
    add_with \
    --x 2 --y 3 \
    --contract CAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAVAX5

[tsachiherman]

Please create a fix for that in the system tests first, and then use the updated version of the system test in this PR.