[vector_math] Write out parameters once through setValues

[spydon]

Makes the Vector2, Vector3 and Vector4 members that take an out/result parameter compute their result first and write it with a single setValues call:

• normalizeInto(out) did out..setFrom(this)..normalize(): one pass to copy the components and a second pass to read them back, compute the length and scale them. It now computes the normalized components from this and writes them once; the zero vector is still copied unchanged.
• min, max and mix assigned each component through a separate setter call (result..x = …..y = …). They now make one setValues call.
• copyInto(arg) wrote the target's storage field by field; it is aligned with the same single setValues pattern.

Numeric results are identical (verified by the benchmark checksums and the test suite). vector_math_64 was regenerated with tool/generate_vector_math_64.dart; only the vector files are included, because the generator also produces unrelated diffs in intersection_result.dart and quaternion.dart (those 64-bit files were hand-edited after generation), so they are left untouched here.

A new test, test/out_parameter_test.dart, covers the rewritten members for all three classes: results for normalizeInto (including the zero-length vector and v.normalizeInto(v) aliasing), copyInto, min, max and mix.

Benchmark

Standalone benchmark (not committed) compiled with dart compile exe against main and against this branch, on macOS arm64; 4096 vectors per case, each binary run five times alternately, values are min/median/max of the runs in nanoseconds per call.

Target	Member	main	this PR	reading
Vector2	normalizeInto	2.55 / 2.59 / 2.61	2.37 / 2.39 / 2.55	~8% faster
Vector2	copyInto	1.06 / 1.08 / 1.09	1.05 / 1.08 / 1.14	unchanged
Vector2	min	1.71 / 1.72 / 1.73	1.58 / 1.58 / 1.64	~8% faster, no overlap
Vector2	max	1.69 / 1.73 / 1.73	1.58 / 1.59 / 1.61	~8% faster, no overlap
Vector2	mix	1.78 / 1.82 / 1.82	1.67 / 1.68 / 1.70	~8% faster, no overlap
Vector3	normalizeInto	2.94 / 3.05 / 3.19	2.68 / 2.68 / 2.87	~12% faster
Vector3	copyInto	1.26 / 1.27 / 1.37	1.24 / 1.26 / 1.37	unchanged
Vector3	min	2.14 / 2.18 / 2.22	1.93 / 1.95 / 2.08	~11% faster, no overlap
Vector3	max	2.14 / 2.17 / 2.19	1.93 / 1.94 / 2.08	~11% faster, no overlap
Vector3	mix	3.00 / 3.06 / 3.23	2.90 / 2.91 / 3.11	~5% faster
Vector4	normalizeInto	3.63 / 3.65 / 3.72	3.02 / 3.06 / 3.08	~16% faster, no overlap
Vector4	copyInto	1.35 / 1.37 / 1.41	1.35 / 1.37 / 1.39	unchanged
Vector4	min	3.29 / 3.33 / 3.41	3.10 / 3.16 / 3.33	~5% faster
Vector4	max	3.28 / 3.28 / 3.55	3.09 / 3.14 / 3.58	~4% faster
Vector4	mix	3.50 / 3.52 / 3.75	3.21 / 3.22 / 3.47	~9% faster

The gains are AOT-specific. The same benchmark compiled with dart compile js -O4 and run in node measures neutral: every case is within about 2% with overlapping ranges. Both builds produce identical checksums on both platforms.

Fixes flutter/flutter#192041

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, or I am not using AI tools.
• 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, or I have commented below to indicate which documented exception this PR falls under¹.
• I updated/added any relevant documentation (doc comments with ///).
• I added new tests to check the change I am making, or I have commented below to indicate which test exemption this PR falls under¹.
• All existing and new tests are passing.

https://claude.ai/code/session_01PixXaRTPs8v1P2mQF2QHGG

Footnotes

1. Regular contributors who have demonstrated familiarity with the repository guidelines only need to comment if the PR is not auto-exempted by repo tooling. ↩ ↩²

[gemini-code-assist]

Code Review

This pull request updates Vector2, Vector3, and Vector4 (both 32-bit and 64-bit implementations) so that normalizeInto, copyInto, min, max, and mix write to their output parameters exactly once using setValues. It also adds tests to verify this behavior for subclass observation. The review feedback highlights multiple instances where newly introduced lines exceed the 80-character limit specified in the style guide, suggesting wrapping arguments to improve readability.

[victorsanni]

LGTM with a comment about missing tests.

[victorsanni]

What about tests for vector_math_64?

[spydon]

The convention has always been to test the Float32 library and trust the generator for the 64-bit variant?
But I can add it for this one if you want.

[victorsanni]

Hi @spydon, the PR description checklist row

I read the AI contribution guidelines and understand my responsibilities, or I am not using AI tools.

is unchecked, even though the commit is from a Claude session. Have you read our AI contribution guidelines?

[spydon]

@victorsanni ah yes, both that one and the CLA one is checked now.