feat: JS vector swizzle assignment

[aleksanderkatan]

Claude is a coauthor because I asked it to split my changes into a stack of 3 🤷‍♀️

[github-actions]

pkg.pr.new

packages
Ready to be installed by your favorite package manager ⬇️

https://pkg.pr.new/software-mansion/TypeGPU/typegpu@340b4fe299ba6217f371d7b89d0b2308890ccc6a

benchmark
view benchmark

commit
view commit

[github-actions]

Resolution Time Benchmark

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Random Branching (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [1.15, 2.21, 4.88, 7.34, 8.19, 13.31, 26.45, 27.88]
  line [1.03, 2.08, 4.70, 7.94, 9.31, 12.62, 29.09, 27.36]
  line [1.05, 2.06, 4.65, 6.89, 7.93, 13.42, 26.40, 26.00]

Loading

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Linear Recursion (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [0.33, 0.55, 0.78, 0.90, 1.23, 1.30, 1.52, 1.74]
  line [0.31, 0.57, 0.77, 0.88, 1.20, 1.26, 1.56, 1.68]
  line [0.35, 0.55, 0.78, 0.92, 1.52, 1.23, 1.49, 1.70]

Loading

---
config:
  themeVariables:
    xyChart:
      plotColorPalette: "#E63946, #3B82F6, #059669"
---
xychart
  title "Full Tree (🔴 PR | 🔵 main | 🟢 release)"
  x-axis "max depth" [1, 2, 3, 4, 5, 6, 7, 8]
  y-axis "time (ms)"
  line [1.03, 2.15, 4.25, 7.50, 15.25, 28.11, 61.30, 120.90]
  line [0.94, 2.56, 4.37, 8.02, 14.12, 28.70, 62.77, 125.02]
  line [0.95, 2.34, 4.76, 8.00, 15.54, 29.77, 60.83, 120.67]

Loading

[github-actions]

Bundle size comparison (import * as ... in PR vs import * as ... in target):

🟢 Decreased	➖ Unchanged	🔴 Increased (max 0.37%)	❔ Unknown
0	84	265	0

No notable changes.

If you wish to run a comparison for other, slower bundlers, run the 'Tree-shake test' from the GitHub Actions menu.

[aleksanderkatan]

@pullfrog review

[pullfrog]

✅ No new issues found.

Reviewed changes

Initial review of the JS vector swizzle assignment change.

• Runtime swizzle setters — VecBase now installs a set accessor for every permutation swizzle (2, 3, and 4 components, in both xyzw and rgba spellings); repeated-component swizzles keep getter-only descriptors, so assigning to them throws a TypeError.
• Writable swizzle types — new Perm2/Perm3/Perm4 helpers plus writable index signatures intersected into Swizzle2/Swizzle3/Swizzle4, exposing exactly the permutation swizzles as assignable while repeated ones stay readonly.
• Tests — swizzleAssignment.test.ts covers full, partial, and self assignment, repeated-component rejection, and the WGSL swizzle-assignment evaluation-order example.

( ◦°^°◦) 🏴☠️ Avast! I checked the swizzle setters, ran the full typegpu suite (2902 tests passin'), typechecked the package, and verified the WGSL evaluation-order example — the ship be seaworthy.

ℹ️ Writable swizzle types also reach TGSL, where assignment is not implemented yet

The Swizzle2/3/4 interfaces are shared by the v2f/v3f/v4f types used inside 'use gpu' functions, so TGSL swizzle assignment (e.g. let v = d.vec4f(); v.xy = d.vec2f(1, 2);) now typechecks but still fails at tgpu.resolve time. This looks intentional — the stacked #3205 adds the WGSL codegen (src/tgsl/swizzle.ts) — so no change is needed here; flagging the merge-ordering dependency only.

Technical details

# TGSL swizzle assignment typechecks but is rejected by the generator

## Affected sites
- `packages/typegpu/src/data/wgslTypes.ts:185` — writable `Perm*` index signatures are inherited by the `'use gpu'` vector types.
- `packages/typegpu/src/tgsl/accessProp.ts:199` — existing comment notes a swizzle LHS assignment "is not yet supported in WGSL".

## Observed behavior
- `let v = d.vec4f(); v.xy = d.vec2f(1, 2);` inside a `'use gpu'` fn typechecks, then throws `'v.xy = d.vec2f(1, 2)' is invalid, because the left side is not a reference to an existing value.` during resolution.

## Required outcome
- None for this PR if the stack merges in order (#3204 → #3205).

## Open questions for the human
- Confirm #3204 is expected to land before #3205, accepting the brief type/runtime gap on `main`.

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash (free via Pullfrog for OSS) ｜ 𝕏}

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing. Review the Copilot code review billing settings.

[pullfrog]

ℹ️ Minor suggestions only.

Reviewed changes

• packages/typegpu/src/data/vectorImpl.ts — the static block that defines 2/3/4-length swizzle accessors now builds one descriptor per component tuple and appends a set when the tuple is a permutation. The setter destructures the RHS into locals before writing, so self-assignment (vec.wzyx = vec) and overlapping regions are handled correctly; repeat swizzles stay getter-only and therefore throw. Verified against the tests.
• packages/typegpu/src/data/wgslTypes.ts — adds Perm2/3/4<T> helpers and intersects Swizzle2/3/4 with mutable maps keyed by those permutations, so only permutation swizzles become writable while repeat swizzles stay readonly. Confirmed with tsc --noEmit (valid perms compile, repeats/length mismatch/mixed aliases error). These interfaces are shared with 'use gpu' code, where codegen support arrives in #3205 of the stack.
• packages/typegpu/tests/swizzleAssignment.test.ts — new JS tests covering all permutation counts in both xyzw/rgba variants, partial assignment, repeat rejection, self-assignment, and WGSL evaluation order (#example-95005bab). npx vitest run tests/swizzleAssignment.test.ts passes (6/6).

  ^{｜ Fix all ➔ ｜ Fix 👍s ➔ ｜ View workflow run ｜ Using deepseek-v4.1-flash (free via Pullfrog for OSS) ｜ 𝕏}

[cieplypolar]

Cool PermN types