Skip to content

Clarify default fibre treatments and guard connector rebuilds - #64

Merged
ProjectIdris merged 1 commit into
ProjectIdris:mainfrom
hearthworksai:celyn/material-treatment-default-docs
Nov 8, 2026
Merged

ProjectIdris merged 1 commit into
ProjectIdris:mainfrom
hearthworksai:celyn/material-treatment-default-docs

Conversation

@hearthworksai

Copy link
Copy Markdown
Contributor

Summary

  • Correct the scene-language documentation: omitted/null path treatment uses the named material default, with fibres for muscle, tendon and rope, otherwise plain.
  • Add regression coverage for explicit and implicit fibre treatments during extreme elastic contraction and reverse seeks across low, mid and high quality tiers.
  • No product implementation or API changes.

Regression controls

Both tests pass on current main. Each was also run against a separate, deliberate implementation break with assertions and timeouts unchanged, then rerun after restoring the production file byte-for-byte:

  • treated elastic connectors keep the current bulge and endpoint limits through contraction and reverse seeks: remove the capDepth argument from updateConnectors' tubeGeometry call. The test fails its rounded-end quarter-span assertion (observed bounds approximately -0.02 to 0.04 for a 0.02 span, versus allowed -0.005 to 0.025), then passes after restoration.
  • named fibre materials retain their implicit treatment through contraction and reverse seeks: replace treatmentOf(def) with def.treatment in the connector rebuild's setPathSurfaceCoordinates call. The test fails because muscle/low geometry loses its material coordinates, then passes after restoration.
    These controls demonstrate regression detection, not a runtime defect in current main.

Validation

  • Focused surface/material/elastic suite: 49 passed, 0 failed.
  • Full portable suite, node --test --test-concurrency=2 'test/*.test.ts': 1,970 tests; 1,948 passed, 1 failed, 21 skipped.
  • Sole full-suite failure: a synthetic speech server bound to 127.0.0.2 receives the configured local request, with UND_ERR_SOCKET / local speech server unreachable. The same targeted test fails with the same socket error on an untouched checkout of base main.
  • npm run typecheck, npm run build, and git diff --check: passed.
  • Independent source and red/green evidence review: no blocking findings.

Limits

Native/desktop checks are among the skipped tests; the native fibre shader test requires a display server. This work makes no new GPU rendering, desktop visual-acceptance, or live-model latency claim.

Source

Tested tree: 5e9c8b47bbf1cc3945ef9208aec6960b8c795e15. Published head: 3ae1588d07e2140f5e73cd813fcfec9b07ab5e8c, based directly on main 21cd3c7d5e7d70e252a108da5ba6f8fa393fc9dd. Hosted CI is pending for this new head. Draft for review; no merge requested.

@ProjectIdris
ProjectIdris marked this pull request as ready for review November 8, 2026 05:44
@ProjectIdris
ProjectIdris merged commit b4ce3e2 into ProjectIdris:main Nov 8, 2026
2 checks passed
@ProjectIdris ProjectIdris mentioned this pull request Nov 8, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants