[vector_math] Fix Quaternion.rotate() applying the inverse rotation

[AdityaJagtap18]

Quaternion.rotate()/rotated() computed conjugate(this) * [v,0] * this
instead of the standard sandwich product this * [v,0] * conjugate(this),
so every quaternion rotation (in both the 32-bit vector_math.dart and
64-bit vector_math_64.dart variants) rotated vectors by the inverse of
the intended rotation.

Example: rotating (1, 0, 0) by 90° about +Z must give (0, 1, 0) by
the right-hand rule, and matches Matrix4.rotationZ/
Quaternion.asRotationMatrix(), but the old code returned (0, -1, 0).

Fixes flutter/flutter#192350.

What changed

• lib/src/vector_math/quaternion.dart and
  lib/src/vector_math_64/quaternion.dart: corrected rotate() to use
  the standard this * [v,0] * conjugate(this) sandwich product.
• test/fuzz_invariants_test.dart (new): a randomized invariant fuzzer
  covering normalize, normalizeInto consistency, componentwise
  min/max/mix, inverse round-trips for Matrix2/3/4 and Quaternion, cross
  product orthogonality, and — the check that caught this bug —
  Quaternion.rotated() cross-verified against
  Quaternion.asRotationMatrix() on the same vector.
• test/quaternion_test.dart: corrected the 12 hardcoded expected
  vectors in testQuaternionNormalize(), which encoded the old (buggy)
  rotation direction; new values were independently derived from the
  standard rotation-matrix formulas.
• Bumped vector_math to 2.4.3 with a CHANGELOG entry.

Verification

• Derived the correct sandwich-product formula by hand and checked it
  against Matrix4.rotationZ/asRotationMatrix() ground truth for
  several axes/angles.
• Ran the full vector_math test suite after the fix: all 261 tests
  pass (dart test).
• dart format --output=none --set-exit-if-changed and dart analyze
  are clean on all changed files.

Transparency

I used Claude Code to help find this bug (via the fuzz test), derive
and verify the fix, and draft this PR. I reviewed the diff and the math
by hand before submitting, and I'm available to address review feedback.

Pre-Review Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
• I read the AI contribution guidelines and understand my responsibilities.
• I read the Tree Hygiene page, which explains my responsibilities.
• I read and followed the Flutter style guide and ran the auto-formatter.
• I signed the CLA.
• The title of the PR starts with the name of the package surrounded by square brackets, e.g. [vector_math]
• I linked to at least one issue that this PR fixes in the description above.
• I followed the version and CHANGELOG instructions, using semantic versioning and the repository CHANGELOG style.
• I updated/added any relevant documentation (doc comments with ///).
• I added new tests to check the change I am making.
• All existing and new tests are passing.

🤖 Generated with Claude Code

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[gemini-code-assist]

Code Review

This pull request fixes a bug in Quaternion.rotate() and rotated() where the inverse of the intended rotation was being applied in both the 32-bit and 64-bit implementations. It updates the corresponding unit tests, bumps the package version to 2.4.3, and introduces a new fuzz test suite to verify mathematical invariants. The review feedback suggests using the built-in normalized() method instead of cloning and normalizing vectors manually, and removing redundant vector cloning when calling Matrix3.transformed().

[gemini-code-assist]

In vector_math, Vector2, Vector3, and Vector4 all provide a built-in normalized() method that returns a normalized copy of the vector. Using v2.normalized() is more idiomatic and concise than v2.clone()..normalize().\n\nThe same applies to:\n- Line 57: v3.clone()..normalize() -> v3.normalized()\n- Line 62: v4.clone()..normalize() -> v4.normalized()\n- Line 72: v3.clone()..normalize() -> v3.normalized()\n- Line 82: v4.clone()..normalize() -> v4.normalized()

Suggested change

	final Vector2 copy2 = v2.clone()..normalize();
	final Vector2 copy2 = v2.normalized();

[gemini-code-assist]

In vector_math, Matrix3.transformed(Vector3 arg, [Vector3? out]) does not modify the input vector arg. If the out parameter is omitted, it automatically creates a copy of arg internally before transforming it. Therefore, calling v.clone() here is redundant and results in an extra unnecessary allocation. You can pass v directly.

Suggested change

	final Vector3 viaMatrix = m.transformed(v.clone());
	final Vector3 viaMatrix = m.transformed(v);