Add DNC-based implementation, update tests and overrides

[mark-moir]

The main purpose of this PR is to add a second implementation of CryptoInterface based on the DockNetwork crypto (DNC) library, thus evaluating the ability of our abstraction to be supported over another library. It is mostly successful, in that almost all features can be supported by both the cryptography library for AnonCreds v2 (AC2C) and DNC (exceptions noted below).

We also override some tests (more detail below) due to differences between the CryptoInterfaces implemented over AC2C and DNC, respectively:

• AC2C does not yet fully support Verifiable Encryption (in particular, decryption and verification of decryption are not yet implemented). For this reason, tests involving Verifiable Encryption are overridden to Fail for now. Although the tests would fail even if not overridden, overriding them saves testing time, and enables us to filter out expected test failures and provide tailored explanation about the reason for failure in test output.
• AC2C panics in some cases when (deliberately) misused by our tests, such as failing to provide a required accumulator witness. The panic causes the tests to fail, but the main purpose of the test (in this example) is to show that a Prover who lacks an updated witness cannot create a proof and convince a Verifier; the panic might be undesirable from a usability perspective, but it does not represent a failure in terms of the main purpose of the test. Such tests are therefore overridden to Skip.
• DNC has changed to interpret ranges as right-open (see this discussion), so that a range from n to n+1 is now considered to include only n. This caused DNC tests to fail because it refuses to create a range proof for such intervals. This highlights that our original abstraction is not sufficiently flexible to accommodate underlying cryptographic libraries that interpret ranges differently. We plan to address this in future work. In the meantime, the relevant test is overridden to Skip.

This PR also includes:

• some refactoring for readability
• some renaming for clarity
• some improvements to our test framework and its documentation, including about overriding tests
• modified tests, particularly to eliminate tests that include range proofs that specify ranges that include only a single value (ZKPs make no sense in this case because the single-value range reveals the value of any attribute that satisfies it)
• added tests of revocation and subsequent witness update for unrevoked credentials