CNA-GAP-028: The METAL renderer applies KHR_materials_specular factors only and samples neither specular texture

CNA snapshot 009d40f5  ·  Known Issues › Functional gaps  ·  source links pinned to 009d40f5

Every other PBR renderer samples specularTexture and specularColorTexture; METAL reads the factor and colour but not the two maps, which the renderer-policy test pins as factor-only.

Identifier

CNA-GAP-028

Category

Functional gap

Subsystem

Graphics & renderers

Status

Open

Verified against

CNA 009d40f5 (009d40f5dd085c4e674d3479675fac84b12b3e0a)

Evidence basis

Source-verified: read at TARGET, not executed

Tests touching this area

Yes: see Current tests

Affected contract

PbrEffect rendering of KHR_materials_specular materials on the METAL renderer

Expected behaviour

The imported specularTexture and specularColorTexture should modulate the specular term wherever PBR materials are drawn.

Actual behaviour at TARGET

SpecularTextureInventoryClassifiesEveryPbrRenderer in GltfRendererPbrFallbackPolicyTests.cpp lists metal as the only factor-only renderer (“it genuinely IS factor-only”). gltf-limitations.md names EasyGL, OpenGL4, DirectX9/11/12, SDL GPU and Vulkan as applying the textures.

Source locations

• modules/content/tests/CNA/Internal/GltfImport/GltfRendererPbrFallbackPolicyTests.cpp — SpecularTextureInventoryClassifiesEveryPbrRenderer: factorOnly = {metal}
• docs/gltf-limitations.md — KHR_materials_specular: IMPLEMENTED_WITH_A_NAMED_LIMIT

Evidence

Checked by reading at 009d40f5; not executed; METAL runs only on macOS.

Focused reproduction

No focused reproduction is known. Nothing has been invented here; the evidence above is what exists.

Current tests

The policy test pins the classification so that a change is noticed.

Regression test

When METAL samples the maps, move it to the sampling set and add a pixel witness.

Blast radius

glTF materials with specular maps drawn on METAL: the specular term is uniform across the surface.

Workaround

No workaround is known.