v0.36.0

0.36.0 (2024-10-22)

⚠ BREAKING CHANGES

• remove pedersen commitment (AztecProtocol/aztec-packages#9107)
• remove pedersen hash opcode (AztecProtocol/aztec-packages#9245)
• Brillig and AVM default all uninitialized memory cells to Field 0 (AztecProtocol/aztec-packages#9057)
• remove keccak256 opcode from ACIR/Brillig (AztecProtocol/aztec-packages#9104)
• Brillig with a stack and conditional inlining (AztecProtocol/aztec-packages#8989)
• avm: remove CMOV opcode (AztecProtocol/aztec-packages#9030)
• Integer division is not the inverse of integer multiplication (#6243)
• kind size checks (#6137)
• Change tag attributes to require a ' prefix (#6235)

Features

• Add checked_transmute (#6262) (2618061)
• Add more Type and UnresolvedType methods (#5994) (8236cbd)
• Allow unconstrained after visibility (#6246) (f6dfbcf)
• Brillig and AVM default all uninitialized memory cells to Field 0 (AztecProtocol/aztec-packages#9057) (70dcf4a)
• Brillig with a stack and conditional inlining (AztecProtocol/aztec-packages#8989) (70dcf4a)
• Don't crash LSP when there are errors resolving the workspace (#6257) (7cc7197)
• Don't suggest private struct fields in LSP (#6256) (2a727b3)
• Handwritten parser (#6180) (c4273a0)
• improve: Remove scan through globals (#6282) (fd91913)
• Inclusive for loop (#6200) (bd861f2)
• Integrate databus in the private kernels (AztecProtocol/aztec-packages#9028) (70dcf4a)
• interpreter: Comptime derive generators (#6303) (d8767b3)
• Kind size checks (#6137) (6e40f62)
• New formatter (#6300) (62404d7)
• Optimize Quoted::as_expr by parsing just once (#6237) (a4fcd00)
• Optimize reading a workspace's files (#6281) (b54ed26)
• perf: Flamegraphs for test program execution benchmarks (#6253) (c186791)
• perf: Follow array sets backwards in array set from get optimization (#6208) (999071b)
• Recover from '=' instead of ':' in struct constructor/pattern (#6236) (9a12f31)
• Remove byte decomposition in compute_decomposition (#6159) (a8bcae2)
• Show LSP diagnostic related information (#6277) (c8a91a5)
• Slightly improve "unexpected token" error message (#6279) (8232bfa)
• Sync from noir (AztecProtocol/aztec-packages#8934) (70dcf4a)
• Sync from noir (AztecProtocol/aztec-packages#9034) (70dcf4a)
• Sync from noir (AztecProtocol/aztec-packages#9099) (70dcf4a)
• Sync from noir (AztecProtocol/aztec-packages#9275) (70dcf4a)
• test: Fuzz poseidon hases against an external library (#6273) (8d8ea89)
• test: Fuzz test poseidon2 hash equivalence (#6265) (f61ba03)
• test: Fuzz test stdlib hash functions (#6233) (1a2ca46)
• test: Include the PoseidonHasher in the fuzzing (#6280) (afb8a7c)
• Trait inheritance (#6252) (d3301a4)
• Visibility for impl functions (#6179) (1b26440)
• Visibility for struct fields (#6221) (fc1c7ab)
• Warn about private types leaking in public functions and struct fields (#6296) (67ac0d6)

Bug Fixes

• Add missing visibility for auto-import names (#6205) (c3cb38a)
• Address inactive public key check in verify_signature_noir (#6270) (e4325aa)
• Allow array map on empty arrays (#6305) (51ae1b3)
• Change tag attributes to require a ' prefix (#6235) (b43dcb2)
• Check for Schnorr null signature (#6226) (2430920)
• Display function name and body when inlining recursion limit hit (#6291) (33a1e7d)
• Do not warn on unused self in traits (#6298) (4d524bf)
• Don't warn on unuse global if it has an abi annotation (#6258) (e13f617)
• Don't warn on unused struct that has an abi annotation (#6254) (8a31632)
• Don't warn twice when referring to private item (#6216) (619c545)
• Enforce correctness of decompositions performed at compile time (#6278) (53252fd)
• frontend: Do not warn when a nes...

nightly-2024-10-24: feat: merge and sort imports (#6322)

# Description

## Problem

Manually arranging imports is time consuming.

## Summary

Added two new configurations, `reorder_imports` (true by default) and
`imports_granularity` (`Preserve` by default), similar to the ones found
in rustfmt. With the default configuration imports are reorganized
alphabetically but not merged.

While implementing this I bumped into one code in the test programs like
this:

```noir
use foo;
use foo::bar;
```

Merging those in Rust ends up like this:

```noir
use foo::{bar, self};
```

We don't have `self` in imports, so I had two choices:
1. Try to leave the above unmodified.
2. Implement `self` in imports.

It turned out that implement `self` in imports is very easy: when we
desugar imports, if the last segment is "self" we just remove it. Then
`use foo::self` is desugared into `use foo`.

## Additional Context

None.

## Documentation

Check one:
- [x] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist

- [x] I have tested the changes locally.
- [x] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.

---------

Co-authored-by: Tom French <[email protected]>

nightly-2024-10-23: chore: remove usage of slices in pedersen hash (#6295)

# Description

## Problem\*

Resolves <!-- Link to GitHub Issue -->

## Summary\*

This PR reworks the pedersen hash implementation to work on arrays as
this removes some brillig overhead.

## Additional Context



## Documentation\*

Check one:
- [x] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [x] I have tested the changes locally.
- [x] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.

nightly-2024-10-22: fix: Allow array map on empty arrays (#6305)

# Description

## Problem\*

Resolves a private issue sent to me on slack

## Summary\*

Array map is a fairly old method written before we had
`std::mem::zeroed`. Now that we have zeroed, we can allow mapping empty
arrays since we can use `zeroed` to get the filler `U` value for the
starting elements.

## Additional Context

While I was at it I included a small clarification to `reduce`'s docs
that it requires a non-empty array

## Documentation\*

Check one:
- [ ] No documentation needed.
- [x] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [x] I have tested the changes locally.
- [x] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.

nightly-2024-10-21

feat: Warn about private types leaking in public functions and struct…

nightly-2024-10-20

feat: Warn about private types leaking in public functions and struct…

nightly-2024-10-19: fix: Do not warn on unused self in traits (#6298)

# Description

## Problem\*

Resolves #6297 

## Summary\*

Sets `warn_if_unused` to `false` for the `self` parameter of a function
if it's for a trait implementation.

## Additional Context



## Documentation\*

Check one:
- [x] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [x] I have tested the changes locally.
- [ ] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.

nightly-2024-10-18: fix: Reject invalid expression with in CLI parser (#6287)

# Description

## Problem\*

Resolves #5560

## Summary\*

Rejects an invalid `--expression-value` in the CLI, which would later
cause a panic in `acvm::compiler::transformers::csat`.

An example error message looks like:

```text
error: invalid value '1' for '--expression-width <EXPRESSION_WIDTH>': minimum value is 3
```

## Additional Context

The issue suggests that
[CSatTransformer::new](https://github.com/noir-lang/noir/blob/ae87d287ab1fae0f999dfd0d1166fbddb927ba97/acvm-repo/acvm/src/compiler/transformers/csat.rs#L24-L27)
should return a `Result` explaining the problem it has with `width`,
rather than doing an `assert!` (without any message) and crashing the
program. I agree, however looking at the chain of calls, none of the
half-dozen functions we go through to reach this use `Result`, and the
`CSatTransformer` is the only option we have. This suggests to me that
this limitation is perhaps supposed to be well-known to the user, ie.
it's not the case that one transformer has X limit and another has Y
limit.

For this reason I added a `pub const MIN_EXPRESSION_WIDTH: usize = 3;`
to the `acvm::compiler` module and using that as a common value for the
assertion as well as the validation in the CLI. Should the assumption of
a single global value change, removing that will force us to update the
validation logic as well.

That said if you prefer going the `Result` route I'm not against it, it
just seemed like an overkill for this single use case.
 
## Documentation\*

Check one:
- [x] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [x] I have tested the changes locally.
- [ ] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.

nightly-2024-10-17: chore: regression test for #5462 (#6286)

# Description

## Problem\*

Regression test for github.com/noir-lang/noir/issues/5462

## Summary\*

- Added the test from [5462](github.com/noir-lang/noir/issues/5462)
- Updated the issue needed to re-enable the `array_regex` execution test

## Additional Context



## Documentation\*

Check one:
- [x] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [x] I have tested the changes locally.
- [x] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.

nightly-2024-10-16: chore: regression test for #5462 (#6286)

# Description

## Problem\*

Regression test for github.com/noir-lang/noir/issues/5462

## Summary\*

- Added the test from [5462](github.com/noir-lang/noir/issues/5462)
- Updated the issue needed to re-enable the `array_regex` execution test

## Additional Context



## Documentation\*

Check one:
- [x] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [x] I have tested the changes locally.
- [x] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.