TC members,
This update dealt with misordered fields in the CryptographicParameters in each of the testcases.
It did not deal with a less trivial question: Should the odd-numbered testcase appear at all, and if so, should they be slightly refactored?
First, let's deal with the even-numbered tests. Each of them registers an RSA private key, then attempts to decrypt an RSA/OAEP ciphertext with varying types of hash algorithms in both OAEP and in the MGF1 mask generation function. If one accepts the incoming ciphertext as authoritative, then these are good tests to show KMIP server capability to produce expected outcomes (known plaintext). I think they belong in 1.4.
The odd-numbered tests register an RSA public key, then encrypt with OAEP with varying types of hash algorithms in both OAEP and in the MGF1 mask generation function. Because OAEP produces unpredictable ciphertext, there is no authoritative output, so one cannot run the odd tests as known-answer tests. They could however be refactored to both encrypt and decrypt to show that the KMIP server can at least deal with its own output (it could be a bootstrap to get to the even-numbered tests), but it's not worth the effort if the TC is not (in the end) going to add them to the testbucket. So what is the TC's good pleasure?
Bruce Rich
---------- Forwarded message ----------
From:
Bruce Rich <bar@cryptsoft.com>Date: Tue, Sep 13, 2016 at 5:36 PM
Subject: [kmip] Groups - cs-ac-oaep-amended-tests.zip uploaded
To:
kmip@lists.oasis-open.org Submitter's message
Updated CryptographicParameters to deal with field-ordering issues. Thanks to Mark Joseph for noting the problem with the first set.
-- Mr. Bruce Rich