search

parallaxsecond / rust-tss-esapi

TSS 2.0 Enhanced System API (ESAPI) Rust wrapper
https://docs.rs/tss-esapi/
Apache License 2.0
89 stars 52 forks source link

Add testing against tpm2-tss v3.0.0 #269

Closed ionut-arm closed 3 years ago

ionut-arm commented 3 years ago

Our current CI workflows test against version 2.3.3 of the tpm2-tss library. However, we have functionality intended for compatibility with 3.0.0 which we do not automatically test (see https://github.com/parallaxsecond/rust-tss-esapi/pull/267). Ideally, all such functionality should be tested to make sure we are indeed compatible.

A new workflow should be created to run all our tests against tpm2-tss v3.0.0.

Superhepper commented 3 years ago

Well on fedora we run against 3.0 version. Is it some kind of special workflow that is needed?

ionut-arm commented 3 years ago

Hmm, it may run against 3.0.0 "incidentally", but I'd rather have a workflow that runs against that specific version of tpm2-tss. For example, there's already a 3.1.0 tpm2-tss version out, we can't have guarantees about which one we're actually testing against. I think the Fedora workflow is more about checking that the crate works well with Fedora in particular, not with a specific version of the library.

Superhepper commented 3 years ago

True it inly takes the latest package available for the Fedora distribution.

I wonder if we should make it so that for each platform we make it build and run against several different version of the libarary.

Example:
  Ubuntu:
    install tpm2-tss 2.3.3
      build (without generating bindings)
      run tests

      build (generate bindings)
      run tests

      uninstall tpm2-tss 2.3.3

  install tpm2-tss 2.4.0
    build (generate bindings)
    run tests

    uninstall tpm2-tss 2.4.0

  install tpm2-tss 3.0.0
    build (generate bindings)
    run tests

    uninstall tpm2-tss 3.0.0

Needs to make the script that runs take a version and an option to set if the bindings should be generated or not. And then we could set up all of the containers to run in parallel.