Perturbing/add msm bls

[perturbing]

Description

✨ PR: Add and Test Two MSM Implementations for BLS12-381

Summary

This PR introduces multi-scalar multiplication (MSM) for BLS12-381 elliptic curve points:

---

Changes

⚙️ Internal Module (BLS12_381/Internal.hs)

• Added types for projective/affine points and scalars, including array/block representations
• Added FFI bindings for 6 C functions:
  • blst_{p1s,p2s}_mult_pippenger, blst_{p1s,p2s}_to_affine, and c_blst_{p1s,p2s}_mult_pippenger_scratch_sizeof
• Added helper marshalling functions:
  • withPointArray, withScalarArray, withAffineBlockArrayPtr
• Implemented blsMSM with filtering of points-at-infinity and zero scalars for safety and efficiency

✅ Tests (Test/EllipticCurve.hs)

• Added property tests comparing blsMSM against naive implementations
• Fixed prop_randomFailsFinalVerify to use group-based logic instead of point inequality
• Updated Arbitrary instance for Point
• Increased probability of generating blsZero for better edge-case coverage

Checklist

• Commit sequence broadly makes sense and commits have useful messages
• New tests are added if needed and existing tests are updated
• All visible changes are prepended to the latest section of a CHANGELOG.md for the affected packages.
  New section is never added with the code changes. (See RELEASING.md)
• When applicable, versions are updated in .cabal and CHANGELOG.md files according to the
  versioning process.
• The version bounds in .cabal files for all affected packages are updated.
  If you change the bounds in a cabal file, that package itself must have a version increase. (See RELEASING.md)
• Commits that only contain large amounts of formatting changes were added to .git-blame-ignore-revs
• Self-reviewed the diff

[lehins]

Haven't looked into ffi calls, just at the haskell portion of it.
I'll look into this in more depth once PR is out of draft

[perturbing]

Thank you for taking a first look @lehins, much appreciated!

The FFI is not working yet, I am getting some segmentation faults that I am trying to debug with valgrind.

[perturbing]

Hi, I fixed the bug I encountered with the memory layout (I overlooked how C code wanted the pointer).

I also added a property test, but got some weird behavior where my test's success depends on the running of other tests. See my comments here and here.

[lehins]

I have quickly looked through the PR, but haven't spotted exactly what is wrong with this PR, but there is definitely something seriously wrong with memory management in this PR. All the non-deterministic test failures, which should not be the case with functionality that pretends to be "pure" serve as a good indicator that the functionality has a bug. Moreover, the fact that unrelated tests are affected is another strong indicator that something is seriously wrong!

I'll try to dig deeper into this functionality some time this week. Maybe I can help you get to the bottom of this.

[kwxm]

This looks OK, but there's a problem with inputs involving the zero point which needs to be fixed.

[perturbing]

Hi, I am ready for another round of review (see also the PR description for an overview of the changes).

Thanks in advance, much appreciated.

I think it's best to not yet merge this PR, best to first let @kwxm implement it in plutus via an SRP stanza.

[tdammers]

Looks mostly good to me.

[lehins]

I had some questions and minor suggestions, but other than that it is looking good.

Needs a rebase on master. Also could you please add all visible changes to the Changelog and bump minor version in both cabal file and in the latest section in the changlog file to 2.2.3.0.

[lehins]

This PR looks ready to go. Thank you for the hard work @perturbing!

[lehins]

@perturbing Could I ask you to cleanup the git commit history. Otherwise I am just gonna squash the whole 73 commits into one big one with very little info.

[perturbing]

I think the above is too messy, since the impact of this PR is small, I made a new PR here #536. Closing this one!