chore: refactor crypto, attestation into crates

[rjzak]

At present, I moved all the cryptography crates into new cryptography crate, which also includes the Steward-specific extensions. Though the issue #76 asked for the attestation code to be separated, it depends on several of the extensions. While at it, I figured I'd try to put all of the external crates into the new cryptography crate, which re-exports them, so that the versioning management is held at a central spot, since the Rust Crypto crates are versioned separately and break when paired with an incorrect version. (Ex: const-oid is 0.9, elliptic curves are 0.11, der is 0.6, x509 is 0.1, etc).

Signed-off-by: Richard Zak [email protected]

[rjzak]

There's an issue when using Enarx for cargo test --target=wasm32-wasi, in that the Enarx.toml is referenced. But for the newly added sub-crates, there's no Enarx.toml present (nor should there be).

• Cargo config doesn't allow for different run vs. test runner, and
• setting the runner in a config for a sub-crate merely adds to the command which is run, doesn't replace it. (why would they do that?? how is that useful??)

So @rvolosatovs's PR #79 is making more sense, except for the missing arg for opening the port for listening.

[bstrie]

Blocked on me extending enarx run allow specifying config parameters on the CLI.

[rjzak]

The second commit changes the runner to Wasmtime, due to the config issue with Enarx needing Enarx.toml, and there being multiple sub-crates which don't have (don't need) their own Enarx.toml. Enarx can be brought back when @bstrie implements the networking config command line, but I'd like to not wait on that.

[puiterwijk]

Looks good to me