False “Invalid mesh topology” warning on the final valid face in `computeSmoothNormals`
Nobody has claimed this yet.
Assessment
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Newbie friendliness
- 74/100
- Issue type
- Bug
- Clarity
- Clearly specified
- Activity status
- Quiet
- Tech stack
- cpp
- Domain
- computer-graphics
Research direction
Start in utils/src/geometry.cpp at computeSmoothNormals and reproduce the warning with the commit-pinned UtahTeapot.usd asset using usdconvert.exe. Check the final-face boundary condition and add regression coverage for a mesh whose face counts end exactly at the index-array length. Done means all faces are processed without an invalid-topology warning.
Written by the indexing model from the issue text.
Description
Hi Everyone, I found an issue while building a cli tool around around what this repo offers.
Description
When exporting a valid USD mesh to STL, glTF, or GLB using Adobe USD File Format Plugins 2026.07, the exporter emits the following warning:
Invalid mesh topology: offset 4944 into indices for face 1235 is larger than total indices 4944
The source mesh topology is valid:
Faces: 1236
Sum of faceVertexCounts: 4944
Number of faceVertexIndices: 4944
Therefore:
sum(faceVertexCounts) == faceVertexIndices.size()
The warning occurs while processing the final face, where the calculated exclusive end offset is exactly equal to the total number of indices.
The export still completes successfully with exit code 0, but the normal-generation loop appears to stop before processing the final face.
Reproduction asset
The issue reproduces using the unmodified Utah Teapot USD asset from the ASWF USD Assets Working Group repository.
The asset is provided under the Public Domain/CC0 license.
- [Commit-pinned UtahTeapot.usd](https://github.com/usd-wg/assets/blob/0cfdafe7ed4b9507a215553f614e63882907074e/full_assets/Teapot/geo/UtahTeapot.usd)
- [Raw UtahTeapot.usd download](https://raw.githubusercontent.com/usd-wg/assets/0cfdafe7ed4b9507a215553f614e63882907074e/full_assets/Teapot/geo/UtahTeapot.usd)
- [Teapot asset directory](https://github.com/usd-wg/assets/tree/0cfdafe7ed4b9507a215553f614e63882907074e/full_assets/Teapot)
- [Asset README and license information](https://github.com/usd-wg/assets/blob/0cfdafe7ed4b9507a215553f614e63882907074e/full_assets/Teapot/README.md)
- [Upstream asset commit](https://github.com/usd-wg/assets/commit/0cfdafe7ed4b9507a215553f614e63882907074e)
No modifications to the upstream USD file are required to reproduce the warning.
Steps to reproduce
Download the commit-pinned UtahTeapot.usd file and export it to STL:
usdconvert.exe UtahTeapot.usd -o UtahTeapot.stl
The same warning can also be reproduced when exporting to glTF or GLB:
usdconvert.exe UtahTeapot.usd -o UtahTeapot.gltf
usdconvert.exe UtahTeapot.usd -o UtahTeapot.glb
Actual result
The export completes with exit code 0, but emits:
Warning: in computeSmoothNormals at line 762 of utils/src/geometry.cpp
-- Invalid mesh topology: offset 4944 into indices for face 1235 is larger than total indices 4944
The warning is produced for STL, glTF, and GLB exports.
FBX and OBJ exports of the same source mesh complete without this warning.
Expected result
- The mesh should not be reported as having invalid topology.
- The final face should be included in smooth-normal generation.
- No topology warning should be emitted when the exclusive end offset is exactly equal to the total number of indices.
Suspected cause
The current implementation in utils/src/geometry.cpp contains the following condition:
if ((size_t)(faceVertexIndex + numFaceVertices) >= totalNumFaceVertices) {
TF_WARN(
"Invalid mesh topology: offset %d into indices for face %zu is larger than "
"total indices %zu",
faceVertexIndex + numFaceVertices,
faceIdx,
totalNumFaceVertices);
break;
}
For the final valid face:
faceVertexIndex + numFaceVertices == totalNumFaceVertices
This equality is valid because the calculated value is the exclusive end offset.
The condition appears to reject a valid final face because it uses >=.
A possible correction would be:
if ((size_t)(faceVertexIndex + numFaceVertices) > totalNumFaceVertices) {
The warning text also says the offset is “larger than” the total number of indices, although in this reproduction the two values are equal.
Potential impact
Because the function calls break, normal generation stops before the final face is processed.
Although the export succeeds, the last face may receive incomplete or incorrect generated normal data.
Environment
OS: Windows 10 x64
Adobe USD File Format Plugins: 2026.07
OpenUSD: 25.11
Affected exports: STL, glTF, GLB
Source mesh faces: 1236
Sum of faceVertexCounts: 4944
faceVertexIndices count: 4944
Suggested regression test
Add a valid mesh where the final face ends exactly at the end of the index array:
sum(mesh.faces) == mesh.indices.size()
Verify that:
computeSmoothNormalsprocesses every face.- No invalid-topology warning is emitted.
- The generated normals include the final face.
Existing-issue search
I searched the repository’s open and closed issues and pull requests for:
- the exact warning text
computeSmoothNormalsfaceVertexIndexfaceVertexIndicestotalNumFaceVertices- mesh topology and normal-generation issues
- related STL and glTF reports
I did not find an existing report for this final-face boundary condition.
- Dominant language
- C++
- Stars
- 396
- Forks
- 42
- PR merge metrics
- No merged PRs in 30d
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from adobe/USD-Fileformat-plugins
-
Use system eigen Open
Difficulty 3/5 1-2 days Newbie friendliness 68/100
-
Difficulty 4/5 3-5 days Newbie friendliness 55/100
adobe/USD-Fileformat-plugins#80 · 1 comment ·
-
Difficulty 4/5 3-5 days Newbie friendliness 35/100
-
Difficulty 1/5 Under an hour Newbie friendliness 50/100
-
Difficulty 5/5 Over a week Newbie friendliness 15/100
All issues in adobe/USD-Fileformat-plugins
Similar issues
-
Difficulty 1/5 Under an hour Newbie friendliness 90/100
AXERA-TECH/ax-llm#77 ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
-
bug-unconfirmed
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
NVIDIA/cuda-samples#453 ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
infiniflow/infinity#3502 ·