Wallet test fixes

[dcousens]

As per the suggestion in #247, this PR removes the use of sinon in Transaction and instead tests that the Transaction is exactly as it should be since the hash is deterministic.

It also makes a few fixes thanks to jshint and re-works a test in Wallet which was not testing the desired behaviour.

[coveralls]

Coverage remained the same when pulling 166f2c4 on dcousens:testprefix into f028dff on bitcoinjs:master.

[glorat]

I assume this is just a clean-up and unrelated to the kill sinon specific change?
Edit: Just read your commit comment where you mentioned it :) Cool

[glorat]

+1 like

[dcousens]

We should probably still isolate the issue with sinon before shrugging it off though.

[dcousens]

Removed sinon-removing commit. Turns out the problem was very obscure, but none the less a fault in the test logic.
The lazy evaluation of ecurve.Point._zInv was causing a deep equality to fail when getPrivateKeyForAddress was freshly deriving a new key pair.

[coveralls]

Coverage remained the same when pulling c3f8869 on dcousens:testprefix into f028dff on bitcoinjs:master.