feat: switch in 'use gpu' functions

[aleksanderkatan]

Changes:

• parse/obfuscate/tseynit/generate switch,
• update some examples to use switch.

More than half of this PR is tests.

[github-actions]

pkg.pr.new

packages
Ready to be installed by your favorite package manager ⬇️

https://pkg.pr.new/software-mansion/TypeGPU/eslint-plugin-typegpu@5ab12f79819aeefc651a8621cca16765db8b102e

https://pkg.pr.new/software-mansion/TypeGPU/tinyest@5ab12f79819aeefc651a8621cca16765db8b102e

https://pkg.pr.new/software-mansion/TypeGPU/tinyest-for-wgsl@5ab12f79819aeefc651a8621cca16765db8b102e

https://pkg.pr.new/software-mansion/TypeGPU/typegpu@5ab12f79819aeefc651a8621cca16765db8b102e

https://pkg.pr.new/software-mansion/TypeGPU/@typegpu/color@5ab12f79819aeefc651a8621cca16765db8b102e

https://pkg.pr.new/software-mansion/TypeGPU/@typegpu/gl@5ab12f79819aeefc651a8621cca16765db8b102e

https://pkg.pr.new/software-mansion/TypeGPU/unplugin-typegpu@5ab12f79819aeefc651a8621cca16765db8b102e

benchmark
view benchmark

commit
view commit

[github-actions]

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

🟢 Decreased	➖ Unchanged	🔴 Increased (max 0.78%)	❔ Unknown
0	304	21	0

import * as ... in PR vs import * as ... in target (did bundle size increase?):

Test	tsdown
STATIC_tgpu.ts	274.75 kB ($${\color{red}+0.8\%}$$)
tgpu_fn.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_init.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_lazy.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_slot.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_const.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_unroll.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_resolve.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_accessor.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_comptime.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_vertexFn.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_computeFn.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_fragmentFn.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_privateVar.ts	274.76 kB ($${\color{red}+0.8\%}$$)
tgpu_vertexLayout.ts	274.77 kB ($${\color{red}+0.8\%}$$)
tgpu_workgroupVar.ts	274.77 kB ($${\color{red}+0.8\%}$$)
tgpu_initFromDevice.ts	274.77 kB ($${\color{red}+0.8\%}$$)
tgpu_bindGroupLayout.ts	274.77 kB ($${\color{red}+0.8\%}$$)
tgpu_mutableAccessor.ts	274.77 kB ($${\color{red}+0.8\%}$$)
tgpu_resolveWithContext.ts	274.77 kB ($${\color{red}+0.8\%}$$)
STATIC_allImports.ts	301.01 kB ($${\color{red}+0.7\%}$$)

import { ... } in PR vs import * as ... in PR (is the library tree-Shakeable?):

Test	tsdown
tgpu_init.ts	265.23 kB ($${\color{green}-3.5\%}$$)
tgpu_initFromDevice.ts	264.68 kB ($${\color{green}-3.7\%}$$)
tgpu_resolve.ts	163.87 kB ($${\color{green}-40.4\%}$$)
tgpu_resolveWithContext.ts	163.81 kB ($${\color{green}-40.4\%}$$)
tgpu_bindGroupLayout.ts	62.32 kB ($${\color{green}-77.3\%}$$)
tgpu_mutableAccessor.ts	57.04 kB ($${\color{green}-79.2\%}$$)
tgpu_accessor.ts	57.04 kB ($${\color{green}-79.2\%}$$)
tgpu_privateVar.ts	55.73 kB ($${\color{green}-79.7\%}$$)
tgpu_workgroupVar.ts	55.73 kB ($${\color{green}-79.7\%}$$)
tgpu_const.ts	55.15 kB ($${\color{green}-79.9\%}$$)
tgpu_lazy.ts	54.95 kB ($${\color{green}-80.0\%}$$)
tgpu_fragmentFn.ts	39.68 kB ($${\color{green}-85.6\%}$$)
tgpu_fn.ts	39.62 kB ($${\color{green}-85.6\%}$$)
tgpu_vertexFn.ts	39.50 kB ($${\color{green}-85.6\%}$$)
tgpu_computeFn.ts	39.20 kB ($${\color{green}-85.7\%}$$)
tgpu_vertexLayout.ts	28.33 kB ($${\color{green}-89.7\%}$$)
tgpu_comptime.ts	15.93 kB ($${\color{green}-94.2\%}$$)
tgpu_unroll.ts	1.75 kB ($${\color{green}-99.4\%}$$)
tgpu_slot.ts	1.70 kB ($${\color{green}-99.4\%}$$)

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

[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 [0.84, 1.72, 3.38, 5.76, 6.79, 10.76, 20.28, 21.53]
  line [0.94, 1.81, 3.82, 5.82, 6.69, 11.27, 21.49, 20.84]
  line [0.87, 1.76, 3.42, 5.41, 7.17, 9.45, 22.01, 23.48]

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, 0.44, 0.70, 0.78, 0.99, 1.07, 1.23, 1.40]
  line [0.34, 0.54, 0.67, 0.76, 1.03, 1.06, 1.26, 1.47]
  line [0.32, 0.49, 0.60, 0.73, 1.00, 1.04, 1.28, 1.45]

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 [0.76, 2.11, 3.40, 6.14, 10.65, 23.35, 49.83, 99.10]
  line [0.82, 1.99, 3.05, 6.94, 10.74, 23.68, 50.49, 103.44]
  line [0.80, 1.96, 2.99, 6.09, 10.90, 23.57, 51.32, 102.07]

Loading

[aleksanderkatan]

@pullfrog review

[pullfrog]

✅ No new issues found.

Reviewed changes — one commit (118097f8 Fix break pruning) re-reviewed against feat/null-legal-at-comptime:

• Fixed WGSL switch break pruning in WgslGenerator._emitSwitchStatement: the old filter removed every statement whose resolved code was a bare break; (case endsWithControlFlow === 'break'), which corrupted cases where an intermediate break split reachable from unreachable statements. It now pops only a single trailing bare break — the switch's genuinely redundant terminal break (WGSL switch clauses don't fall through, so it's safe to drop).
• Added the does not prune wrong breaks test, which fails under the old logic and pins the corrected emission (intermediate breaks kept, terminal break pruned, auto-added default).
• Confirmed the pruning is WGSL-only: GlslGenerator._emitSwitchStatement overrides and re-emits without pruning, so C-style fall-through breaks are retained for GLSL.

I verified the change end-to-end: the new test and the full switch.test.ts suite (21 tests) pass.

  ^{｜ View workflow run ｜ Using DeepSeek 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.