Extra, blank test groups created for KAS-FFC-SSC and KDF vectors

[powersmc]

I noticed when adding some sanity checks in my ACVP driver script recently that KAS-FFC-SSC and KDF (SP 800-108) vectors appear to be creating extra, blank test groups in the response files. For KAS, it seems to create each tgId 3x, while with KDF it was duplicating 2x. For example, if I had a KAS-FFC-SSC vector with 5 tgId within it, I would see this at the end of the response file:

{
        "tgId":1,
        "tests":[
        ]
      },
      {
        "tgId":2,
        "tests":[
        ]
      },
      {
        "tgId":3,
        "tests":[
        ]
      },
      {
        "tgId":4,
        "tests":[
        ]
      },
      {
        "tgId":5,
        "tests":[
        ]
      },
      {
        "tgId":1,
        "tests":[
        ]
      },
      {
        "tgId":2,
        "tests":[
        ]
      },
      {
        "tgId":3,
        "tests":[
        ]
      },
      {
        "tgId":4,
        "tests":[
        ]
      },
      {
        "tgId":5,
        "tests":[
        ]
      }

I've attached a patch that I used to fix this on my side, but just with the caveat that it's not super thoroughly tested. I've only tested it on the specific vectors where it affected me (KAS-FFC-SSC and KDF 1.0). Another note is just that the NIST server does not treat these additional empty test groups as failures or anything - it's more just a formatting issue I noticed.

duplicate_testgroups.patch

[smuellerDD]

Thanks, Mike, for the patch. I have also seen the DH issue a few days ago and fixed it with the changes to the parser_dh.c as attached. Could I ask you to check whether this works for you?

parser_dh.c

Also, I thought I fixed it for the KDFs too, because when I compare your patch with parser_kdf_108.c as found in the current repo, it seems it is covered there. May I ask you to check that?

[powersmc]

For the parser_kdf_108.c that's in master on this repo, that matches exactly with the version I used with the bug manifesting - so I don't think it's fixed there. I will test out your alternate parser_dh.c file today and get back to you.

[powersmc]

Just confirming - that parser_dh.c file attached does address the issue in KAS-FFC-SSC.

[smuellerDD]

Am Freitag, 7. November 2025, 18:28:36 Mitteleuropäische Normalzeit schrieb Mike P: Hi Mike,

powersmc left a comment (smuellerDD/acvpparser#106) Just confirming - that parser_dh.c file attached does address the issue in KAS-FFC-SSC.

Thank you very much, I will apply this shortly. Ciao Stephan

[smuellerDD]

Am Freitag, 7. November 2025, 18:18:34 Mitteleuropäische Normalzeit schrieb Mike P: Hi Mike,

powersmc left a comment (smuellerDD/acvpparser#106) For the parser_kdf_108.c that's in master on this repo, that matches exactly with the version I used with the bug manifesting - so I don't think it's fixed there. I will test out your alternate parser_dh.c file today and get back to you.

Hm, I was asking that, because when applying the patch, the patch tool asked me whether I wanted to revert it (i.e. it thought it was applied). Ok, then I need to manually apply the patch. Ciao Stephan