…er (AztecProtocol/aztec-packages#5797)

A follow-up to AztecProtocol/aztec-packages#5737

Separate PR as this touches mainly the AVM transpiler and not Noir
codegen itself. Look in PR #5737 for full details about the switch, but
basically we are moving away from the `Brillig` opcode to a
`BrilligCall` opcode that contains a Brillig call opcode. The AVM needs
to be updated to account for this change.

…ckages#5800)

This is a quick fix after
AztecProtocol/aztec-packages#5737 since that PR
assumes that one generated brillig is valid for all calls to the brillig
function. This is however not true, since the generated brillig can
differ if the arguments contains slices. This is because the entry point
codegen depends on the size of the slice.
We currently cannot copy slices of any length into brillig due to the
limitations of
[CALLDATACOPY](https://yp-aztec.netlify.app/docs/public-vm/instruction-set#calldatacopy)
where the size being copied needs to be known at compile-time. I'm going
to chat with the AVM team to see if we can lift this restriction and
make brillig entry points be able to copy in arguments that contain
slices of any length.

…ckages#5800)

This is a quick fix after
AztecProtocol/aztec-packages#5737 since that PR
assumes that one generated brillig is valid for all calls to the brillig
function. This is however not true, since the generated brillig can
differ if the arguments contains slices. This is because the entry point
codegen depends on the size of the slice.
We currently cannot copy slices of any length into brillig due to the
limitations of
[CALLDATACOPY](https://yp-aztec.netlify.app/docs/public-vm/instruction-set#calldatacopy)
where the size being copied needs to be known at compile-time. I'm going
to chat with the AVM team to see if we can lift this restriction and
make brillig entry points be able to copy in arguments that contain
slices of any length.

…ztec-packages#5687)

Closes AztecProtocol/aztec-packages#5081

This PR introduces autogenerated contract interfaces for easy intra and
inter contract interactions. The `aztec-macro` crate is used to stub
every non-internal private and public function and inject them into a
ghost struct which has the same name as the contract that generated
them. After that, they can be called like this:

```rust

contract ImportTest {

  use dep::my_imported_contract::MyImportedContract;

  #[aztec(private)]
  fn a_private_fn() {
    let deserialized_return = MyImportedContract::at(some_address).another_private_fn(arg1, arg2).call(&mut context);
    MyImportedContract::at(some_address).a_public_fn(arg1).enqueue(&mut context);
  }

  #[aztec(public)]
  fn a_public_fn() {
    let deserialized_return = MyImportedContract::at(some_address).a_public_fn(arg1).call(&mut context);
  }

  #[aztec(private)]
  fn calling_my_own_fns() {
    ImportTest::at(context.this_address).a_private_fn().call(&mut context);
    ImportTest::at(context.this_address).a_public_fn().enqueue(&mut context);
  }

}
```

Return values are `deserialized_into()` automatically, providing "real"
return values thanks to
AztecProtocol/aztec-packages#5633

Also, some general cleanup was required to allow importing contracts in
another contracts. Main changes:

- `HirContext.fully_qualified_struct_path` now uses BFS to avoid
returning the longest path when looking for a struct in a crate. This is
required to avoid pulling structs usually imported in top-level
dependencies (usually notes from our main contract) from other imported
contracts.
- `pack_args_oracle` now has a slice mode in addition to its usual array
mode.

PENDING:

~~AvmContext. The AVM team is discussing supporting args as slices. In
case it's decided not to do that, a workaround could possibly be
implemented using the macro, but it would be fairly complex.~~

Thanks to @fcarreiro and the amazing AVM team, this is now supported for
the AvmContext!

---------

Co-authored-by: esau <152162806+sklppy88@users.noreply.github.com>
Co-authored-by: Álvaro Rodríguez <sirasistant@gmail.com>

…ztec-packages#5687)

Closes AztecProtocol/aztec-packages#5081

This PR introduces autogenerated contract interfaces for easy intra and
inter contract interactions. The `aztec-macro` crate is used to stub
every non-internal private and public function and inject them into a
ghost struct which has the same name as the contract that generated
them. After that, they can be called like this:

```rust

contract ImportTest {

  use dep::my_imported_contract::MyImportedContract;

  #[aztec(private)]
  fn a_private_fn() {
    let deserialized_return = MyImportedContract::at(some_address).another_private_fn(arg1, arg2).call(&mut context);
    MyImportedContract::at(some_address).a_public_fn(arg1).enqueue(&mut context);
  }

  #[aztec(public)]
  fn a_public_fn() {
    let deserialized_return = MyImportedContract::at(some_address).a_public_fn(arg1).call(&mut context);
  }

  #[aztec(private)]
  fn calling_my_own_fns() {
    ImportTest::at(context.this_address).a_private_fn().call(&mut context);
    ImportTest::at(context.this_address).a_public_fn().enqueue(&mut context);
  }

}
```

Return values are `deserialized_into()` automatically, providing "real"
return values thanks to
AztecProtocol/aztec-packages#5633

Also, some general cleanup was required to allow importing contracts in
another contracts. Main changes:

- `HirContext.fully_qualified_struct_path` now uses BFS to avoid
returning the longest path when looking for a struct in a crate. This is
required to avoid pulling structs usually imported in top-level
dependencies (usually notes from our main contract) from other imported
contracts.
- `pack_args_oracle` now has a slice mode in addition to its usual array
mode.

PENDING:

~~AvmContext. The AVM team is discussing supporting args as slices. In
case it's decided not to do that, a workaround could possibly be
implemented using the macro, but it would be fairly complex.~~

Thanks to @fcarreiro and the amazing AVM team, this is now supported for
the AvmContext!

---------

Co-authored-by: esau <152162806+sklppy88@users.noreply.github.com>
Co-authored-by: Álvaro Rodríguez <sirasistant@gmail.com>

Automated pull of development from the
[noir](https://github.com/noir-lang/noir) programming language, a
dependency of Aztec.
BEGIN_COMMIT_OVERRIDE
chore: fix alerts on rust msrv
(#4817)
chore(ci): fix alerts on msrv issues
(#4816)
chore: run clippy (#4810)
chore: optimize poseidon2 implementation
(#4807)
fix: catch panics from EC point creation (e.g. the point is at infinity)
(#4790)
feat: Sync from aztec-packages
(#4792)
feat: lalrpop lexer prototype
(#4656)
feat(nargo): Handle call stacks for multiple Acir calls
(#4711)
fix: proper field inversion for bigints
(#4802)
feat: add `NARGO_FOREIGN_CALL_TIMEOUT` environment variable
(#4780)
chore(debugger): Docs (#4145)
feat: narrow ABI encoding errors down to target problem argument/field
(#4798)
chore: Rename 'global' to 'function' in the monomorphization pass
(#4774)
chore: Add Hir -> Ast conversion
(#4788)
fix: Fix panic when returning a zeroed unit value
(#4797)
END_COMMIT_OVERRIDE

---------

Co-authored-by: vezenovm <mvezenov@gmail.com>
Co-authored-by: Tom French <15848336+TomAFrench@users.noreply.github.com>

Automated pull of development from the
[noir](https://github.com/noir-lang/noir) programming language, a
dependency of Aztec.
BEGIN_COMMIT_OVERRIDE
chore: fix alerts on rust msrv
(#4817)
chore(ci): fix alerts on msrv issues
(#4816)
chore: run clippy (#4810)
chore: optimize poseidon2 implementation
(#4807)
fix: catch panics from EC point creation (e.g. the point is at infinity)
(#4790)
feat: Sync from aztec-packages
(#4792)
feat: lalrpop lexer prototype
(#4656)
feat(nargo): Handle call stacks for multiple Acir calls
(#4711)
fix: proper field inversion for bigints
(#4802)
feat: add `NARGO_FOREIGN_CALL_TIMEOUT` environment variable
(#4780)
chore(debugger): Docs (#4145)
feat: narrow ABI encoding errors down to target problem argument/field
(#4798)
chore: Rename 'global' to 'function' in the monomorphization pass
(#4774)
chore: Add Hir -> Ast conversion
(#4788)
fix: Fix panic when returning a zeroed unit value
(#4797)
END_COMMIT_OVERRIDE

---------

Co-authored-by: vezenovm <mvezenov@gmail.com>
Co-authored-by: Tom French <15848336+TomAFrench@users.noreply.github.com>

* master:
  chore: remove unnecessary casts in `BoundedVec` (#4831)
  fix: issue 4682 and add solver for unconstrained bigintegers (#4729)
  chore(docs): fix wrong Nargo.toml workspace examples (#4822)
  chore: delete unnecessary Prover.toml file (#4829)
  chore: fix alerts on rust msrv (#4817)
  chore(ci): fix alerts on msrv issues (#4816)
  chore: run clippy (#4810)
  chore: optimize poseidon2 implementation (#4807)
  fix: catch panics from EC point creation (e.g. the point is at infinity) (#4790)