tuf-test-vectors

A collection of test vectors that can be used to verify if a TUF or Uptane client conforms to the specification with regards to processing metadata and targets.

Vectors

This repo contains server.py which provides an interactive API that directs clients on how to perform updates. A client should use the following steps to run through all the test vectors.

List Vectors

Client does a GET on / to receive a JSON array of strings naming the rest of the vectors.

$ curl localhost:8080/
["vector_1", "another_vector", ... ]

Initialize the Vector

Because the vectors need to simulate time (via a "step"), each vector needs to be intialized with a POST.

$ curl -X POST localhost:8080/$vector_name/step
{
  "update": {
    "is_success": true
  },
  "targets": {
    "file.txt": {
        "is_success": true
     }
  }
}

The response will tell you what the result of an update cycle should be as well as what targets to download and whether or not they should validate. If there are errors, both the update and target object will have err and err_msg along with is_success. err will be machine parseable and attempts to enumerate common types and err_msg is a plain English sentance.

Update the Local Metadata

The client should attempt to do a full update of all metadata as defined in the spec, and the success of this update should match update.is_success.

Download and Verify Targets

The client should download and verify the target. The success of this should match targets.$target_name.is_success.

Step

The client should POST to the same endpoint /$vector_name/step and repeate. If the call to "step" returns HTTP 204, then the client is done.

Reset

To reset the vector, the client may POST to /$vector_name/reset. The client may run tests in parallel, but the client may not run many tests against the same one vector in parallel.

License

This project is licensed under the MIT license. See LICENSE for more information.