chore!: Update to acvm 0.11.0

[phated]

Related issue(s)

Resolves #1301

Description

Summary of changes

This updates the project to ACVM 0.11.0

This is the version of ACVM that introduced the Result return type to many of the functions on the Backend trait. This allows us to better surface errors from backends—and as such, aztec_backend has been updated to return results for almost all errors (the only unwraps remaining are in the merkle implementation).

This version of ACVM also introduced keccak into the constraint system, so we needed to upgrade barretenberg also. This may cause the same OOM error that the other Keccak PR is exhibiting.

Dependency additions / changes

Test additions / changes

Checklist

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt with default settings.
• I have linked this PR to the issue(s) that it resolves.
• I have reviewed the changes on GitHub, line by line.
• I have ensured all changes are covered in the description.

Documentation needs

• This PR requires documentation updates when merged.

Additional context

[TomAFrench]

I've renamed the type parameters to B as naming them Backend is very confusing as it shadows the Backend trait.

[TomAFrench]

I've updated this to use the commit for the 0.1.0 release of acvm-backend-barretenberg pending noir-lang/acvm-backend-barretenberg#178

[TomAFrench]

LGTM. I've made a couple of tweaks at the end so will let you have last look.

I'm not a fan of how the errors require CliError to be generic as it seems like we're taking on extra complexity (type parameters and casting traits) for not much benefit. I'd like to make a separate proposal which imo simplifies things later.

[phated]

I'm fine with your try_vecmap changes, even though I often don't like to introduce unnecessary lambdas into rust code.

I'm not a fan of how the errors require CliError to be generic as it seems like we're taking on extra complexity (type parameters and casting traits) for not much benefit. I'd like to make a separate proposal which imo simplifies things later.

Happy to take a look, but I experimented with different options and the only two ways I could find to make std::error::Error work on traits is to use Associated Types or Box<dyn std::error::Error + 'lifetime> and I'm not okay with that type leaking into each backend implementations to avoid a generic in nargo-cli.

This still needs an approval from someone that isn't me.