I think it is important we maintain some reasoning that an Input script
is actually a script signature, but in the end, these names are more
coherent and understandable when reasoning with our code. So I think
its OK we break tradition with bitcoind.
By removing Address.Error, we remove a code smell.
This part of the code base was also not under any form of test.
Test data and tests have therefore been added verifying its behaviour in
both Wallet and Address tests.
The introduction of these two functions allow for the all the network
related code to be eventually removed from Transaction and Script.
Previously the result for non-standard transactions was undefined
behaviour. This change mandates that an exception is thrown if a
non-standard transaction is input.
These functions are not under test, and are unnecessary bloat due to a
confusing API.
Script.from*(asmStr) were two functions that attempted to parse ASM
codes and produce a script from this.
While useful, an parser can be introduced later under a single function
and under test... removed.
Although Script.extractPublicKeys implementation is likely to be correct,
it is not absolute in that what it returns is even strictly a set of
public keys.
It is a useful function, but can be done in a better way later,
probably checking against the Script templates instead.
Transaction.signWithKeys has some inherent undocumented behaviour, and it is not
clear when you would use it over just Transaction.addOutput and
Transaction.sign individually. Nor does it mimic anything in the
bitcoind API... removed.
