fix: preserve signed float intdiv operands

[oxura]

Summary

• prefer i32 when std.intdiv must choose between equally ranked signed and unsigned targets
• keep explicitly unsigned operands on the existing u32 path
• cover negative runtime f32 operands across the CPU implementation and generated WGSL

Closes #2892

Verification

• vitest run --config vitest.config.mts from packages/typegpu — 2,370 tests passed
• pnpm --filter typegpu test:types
• oxlint on the changed source and test files
• oxfmt --check on the changed source and test files

[pullfrog]

Run failed. View the logs →

  ^{｜ Rerun failed job ➔ ｜ View workflow run ｜ via Pullfrog ｜ Using DeepSeek Flash (free via Pullfrog for OSS) ｜ 𝕏}

[Copilot]

Pull request overview

This PR fixes a backend mismatch in std.intdiv where two runtime f32 operands could unify to u32, causing negative values to saturate to 0 in generated WGSL and produce incorrect GPU results compared to the CPU implementation.

Changes:

• Adjust std.intdiv’s type unification preference to choose i32 over u32 in signed/unsigned tie cases.
• Add a regression test ensuring negative runtime f32 operands produce signed integer division in both CPU execution and emitted WGSL.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
packages/typegpu/src/std/numeric.ts	Reorders the unify candidate list for intdiv to prefer i32 over u32.
packages/typegpu/tests/std/numeric/intdiv.test.ts	Adds coverage for negative runtime f32 operands and validates the emitted WGSL uses i32 division.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.