In the Key Exchange API, pointers that are not supposed to be NULL are described as Nullable

[Kleidukos]

Hi! I'm writing bindings to the Key Exchange API, and I'm reading the signature of crypto_kx_client_session_keys:

int crypto_kx_client_session_keys(unsigned char rx[crypto_kx_SESSIONKEYBYTES],
                                  unsigned char tx[crypto_kx_SESSIONKEYBYTES],
                                  const unsigned char client_pk[crypto_kx_PUBLICKEYBYTES],
                                  const unsigned char client_sk[crypto_kx_SECRETKEYBYTES],
                                  const unsigned char server_pk[crypto_kx_PUBLICKEYBYTES]);

Now from what I know of the Standard, the array syntax means that these parameters are non-nullable pointers.

Then a little bit down the documentation paragraph, I see this sentence:

rx and tx are both crypto_kx_SESSIONKEYBYTES bytes long. If only one session key is required, either rx or tx can be set to NULL.

My questions is:

Aren't tx and rx supposed to be non-nullable?

[Kleidukos]

So, I was sorely mistaken, it's actually when the keyword static appears in the [ ] that the pointer must be non-NULL.

cf: https://godbolt.org/z/zffd1f7Yn

[jedisct1]

Yes, as documented, these parameters can be NULL.

Functions with parameters that have to be non-NULL are marked with the __attribute__((nonnull)) attribute in their prototype.

[Kleidukos]

Thanks! For now I'm mostly taking the signatures in the docs to document my own bindings to Haskell, but I'll do a sweep in the source code to see if there are any nonnull attributes that I would have missed.