fix: don't disallow writing to memory after passing it to brillig

[TomAFrench]

Description

Problem*

Resolves

Summary*

This fixes an issue that @MirandaWood was running into on her batched blobs branch.

@kashbrti and I found the root cause is that if a memory block is passed as an argument to a brillig call then the CircuitSimulator marks it as "used" at which point it will reject any future writes to that memory block. I don't see any reason for why the simulator should reject this so I've replaced this logic with a simple check to ensure that we don't perform any reads/writes from uninitialized memory blocks.

@guipublic I think you originally wrote this logic, can you explain the motivation behind the old behaviour in case I'm missing something?

Additional Context

Documentation*

Check one:

• No documentation needed.
• Documentation included in this PR.
• [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist*

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[michaeljklein]

LGTM w/ a few notes:

1. Should we note this behavior in the BrilligInputs docs?

   noir/acvm-repo/acir/src/circuit/brillig.rs

   Line 38 in a1b74f8

   pub enum BrilligInputs<F> {

2. Is the original example too complex for a Noir-level regression test?

[TomAFrench]

Should we note this behavior in the BrilligInputs docs?

I'm not clear what behaviour you mean, the ability to pass a memory block as input?

Is the original example too complex for a Noir-level regression test?

We could construct one but I think a more focused test is better.

[michaeljklein]

I'm not clear what behaviour you mean, the ability to pass a memory block as input?

Mainly how the initialization works