add support for OCSP_copy_nonce

[samuel40791765]

Issues:

Addresses CryptoAlg-2420

Description of changes:

Ruby consumes OCSP_copy_nonce, which copies the nonce from the request to the OCSP response.
OpenSSL's OCSP_copy_nonce directly appends the new OCSP nonce to the list of extensions in the response. This allows multiple OCSP nonces to exist the same response since the existing nonces are not deleted. It's a bit strange to allow multiple nonces to exist in a single response. To make things stranger, OCSP_check_nonce directly takes the first OCSP nonce in the list to compare, so any newer nonces are ignored.
We ultimately decided to follow OpenSSL's faulty implementation and document/test the behavior, there doesn't seem to be much value in diverging from the original and will only introduce more unanticipated behavioral changes.

Call-outs:

N/A

Testing:

Extend upon original OCSP nonce tests

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license and the ISC license.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.31%. Comparing base (b32e641) to head (0632214).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1711   +/-   ##
=======================================
  Coverage   78.31%   78.31%           
=======================================
  Files         580      580           
  Lines       96619    96639   +20     
  Branches    13850    13859    +9     
=======================================
+ Hits        75665    75686   +21     
+ Misses      20340    20338    -2     
- Partials      614      615    +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[WillChilds-Klein]

nit: should we document that the nonce extension is prepended?

[samuel40791765]

Ah yes we should

[samuel40791765]

The more I think about it, I think the correct behavior should be replacing any existing nonce in resp instead of prepending/appending it to the list of extensions. OpenSSL appends the nonce to the list, but it doesn't really make sense for a resp to have two nonces.
My initial concern was that someone may have noticed OpenSSL's double nonce behavior and they're removing the old extra nonce themselves after the call to OCSP_copy_nonce. This might cause our enforced singular nonce to be removed unintentionally. But it's an edge case trying to get around OpenSSL's buggy behavior, enforcing the right behavior might be the better call here.

[samuel40791765]

After discussing with @WillChilds-Klein, ultimately decided to align with OpenSSL's implementation. There's not too much benefit in misaligning here to be fair and creates more unanticipated behavioral changes.