* Add full json ledger dumps of ledgers 38129 and 40000 to test/fixtures
* Use `Ledger` to calculate account and transaction hashes and verify
against dumps
This includes supporting files that can sign arbitrary data
with a signature that enables public key recovery. It also
includes the PublicKeyValidator class that can verify whether
a given public key is active for an account by looking in its
AccountRoot.
Amount#parse_quality is made currency-aware. This allows it to adjust for XRP as
the base currency, as well as for interest-bearing or demurring base currencies.
Currency#to_json should output a format that can be successfully converted back
to a Currency object via from_json. So for demurrage currencies it should output
the full hex currency code.
Currency#to_human on the other hand should always print something a human will
understand. For demurrage currencies we'll print the three-letter code for now.
This patch might regress the performance of the Currency class and by extension
the Amount class. Since Amount is on a lot of hot paths in the client we should
make sure this isn't a major problem.
As for compatibility, this patch is a major change, but it should maintain the
public interface very well, which the exception of some strange edge cases (e.g.
Currency.from_json(1337)), which weren't well-defined before anyway. Any code
that accesses _value directly (shame on you!) will need to be fixed. There
aren't any such references in ripple-client or the rippled test suite, so I
think we're looking pretty good.
The UInt160 class used to be hardcoded to be an Account. This commit changes it
so it can be used as an account or a plain hash. It will try to automatically
self-classify based on how it is initialized.
In the future we may want to have some dedicated classes rather than a
single configurable UInt160.
XRP IOUs are no longer valid, however they do exist in the network's historical
transactions, so the serialization and deserialization need to be able to handle
them.
Unfortunately, there are some transactions in the history that use this format,
so we need to support it in order to be able to verify historic ledgers.
