Extend RSA-PSS examples.

[duesee]

The examples do not consider the case of broken keys.

[ma-ilsi]

The example in the docs, correct?

[franziskuskiefer]

Yes, the examples in the docs. They are generally correct. But they could be extended to deal with incorrect keys.
Are you interested in working on this @ma-ilsi ?

[ma-ilsi]

Are you interested in working on this @ma-ilsi ?

Yup. I'll take it on.

[ma-ilsi]

Looking at tests/rsapss.cc, it seems a comparison of skey and pkey against NULL (similar to the sign and verify bool checks) will get the message across and make the API reference example clearer.
Apart from that, the load_skey/pkey functions' return values in the docs currently only specify that a pointer to the key is returned. It should also specify that NULL will be returned if parameters are deemed invalid (broken key).

That's my reading of the issue, did I miss something? If my understanding is clear I can PR soon and we'll review.

[franziskuskiefer]

That's my reading of the issue, did I miss something? If my understanding is clear I can PR soon and we'll review.

Sounds about right.
Generally, looking at the function in F* with the pre and post conditions (both in comment and code) should be able to tell whether that's all or if you missed something.