enclave_build: fix argument parsing and image generation

[sstone]

Description of changes:

• fixed inconsistencies in argument names that make it impossible to generate a signed image
• fixed potential image corruption if the output image already exists by truncating it before writing to it
• bump version to 0.2.0

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[raulmldv]

Checked the logs:
Makefile:221: recipe for target 'nitro-format' failed
Make sure to run cargo-fmt --all before commiting.

[sstone]

Checked the logs: Makefile:221: recipe for target 'nitro-format' failed Make sure to run cargo-fmt --all before commiting.

done in dc64679

[raulmldv]

Currently failing on license check make nitro-about. Opened a PR to update our license file. After we merge it, you should rebase.

[lauralt]

Why do we bump the version here? Is the version 0.1.0 released on crates.io?

[sstone]

Since it's a significant change I thought it would make sense to change the version here, but I don't know how this library's maintainers manage versioning.

[lauralt]

If it's not on crates.io, I don't see a point in updating the version here.

[sstone]

I have no idea, I think it's up to the maintainers to choose here.

[lauralt]

Would be nice to follow the 50/72 rule for commit messages. Also please add a Signed-off-by line: git commit -s --amend.

[sstone]

I'll rebase on main and update the commit message.

[petreeftime]

Can you describe a bit more the issue that this patch is fixing? It seems to make the commandline arguments inconsistent, since they follow snake_case rather than kebab-case in enclave_build crate. I think they would all need to be updated to look the same.

[sstone]

Can you describe a bit more the issue that this patch is fixing? It seems to make the commandline arguments inconsistent, since they follow snake_case rather than kebab-case in enclave_build crate. I think they would all need to be updated to look the same.

The issue here is that it is not possible to create a signed image with enclave_build because the private key parameter is not parsed at all (instead the code reuses private_certificate which is imho a bug). I used private-key and not private_key because that's how the command line option is defined as the beginning of main.rs.