wanadev/gltf-bounding-box

Computes the global bounding box of a gltf model

gltf-bounding-box's Issues

If I replace the gltf file, it won't work

According to the code, I found that when URI was not coded in base64, it could not continue to work.
the err：
gltf-bounding-box\dist\index.js:17335 if (!buffer.uri.startsWith("data:")) {

Cannot read property 'slice' of null when computing bounding box

I am using this to compute the bounding box of a gltf file, but an error happens. Here is the stack trace:

TypeError: Cannot read property 'slice' of null
    at Object._loadBufferView (.../gltf-bounding-box/dist/webpack:/src/gltf-reader.js:92:19)
    at Object._loadAccessor (.../gltf-bounding-box/dist/webpack:/src/gltf-reader.js:48:31)
    at .../gltf-bounding-box/dist/webpack:/src/gltf-reader.js:25:47
    at Array.forEach (<anonymous>)
    at loadPositions (.../gltf-bounding-box/dist/webpack:/src/gltf-reader.js:21:21)
    at .../gltf-bounding-box/dist/webpack:/src/gltf2-bounding-box.js:46:51
    at Array.reduce (<anonymous>)
    at Object.getMeshesTransformMatrices (.../gltf-bounding-box/dist/webpack:/src/gltf2-bounding-box.js:39:8)
    at Object.computeBoundings (.../gltf-bounding-box/dist/webpack:/src/gltf2-bounding-box.js:8:28)
    at Object.computeBoundings (.../gltf-bounding-box/dist/webpack:/src/index.js:23:33)

I am using node 8.15.0 on macOS 10.13.4

Transformations are not applied if they are not in the matrix field

If a gltf file contains transformations in the fields scaling, rotation, and translation, these are not considered in the computation.

This happens because in the function getParentNodesMatrices, only the field matrix is checked.
However, the GLTF specifications says that also three other fields are present (for T, R, and S), which can contain the transformations separately.

NaN when identity matrix not specified

[BUG] Integer division causes center to be rounded even if precision is set

In the files gltf1-bounding-box.js and gltf2-bounding-box.js the center is computed as:

center: {
    x: precise.round((boundings.max[0] + boundings.min[0]) / 2, precision + 1),
    y: precise.round((boundings.max[2] + boundings.min[2]) / 2, precision + 1),
    z: precise.round((boundings.max[1] + boundings.min[1]) / 2, precision + 1),
},

However, the division by 2 is integer division and causes the number argument of precise.round to always be an integer.

To correct this bug, I believe it is enough to divide by 2.0 instead.

Should I make a PR for this?

Support du GLB2

• Supporter le glTF 2 ;
• Supporter le format binaire en extrayant le JSON à l'intérieur.