The MUST be normalized to have a linear sum of 1.0 requirement was morphed into a SHOULD be as close as reasonably possible to '1.0' for a given vertex. requirement.
I'm not sure if this is a problem with the specification or validator. However, as the specification is no longer in pre-release, I'm assuming this is the way that it shall stand.
Hi, in the following commits:
https://github.com/KhronosGroup/glTF/commit/0a61cf8d029ece4e1d895852a8b61371217c7677 https://github.com/KhronosGroup/glTF/commit/94ed6813112aa2286eb342fd9d493ba872975bc8#diff-ca628c11679a8c4c0c4031440c7c1cda27c8896ebc87c3bd1e8640efc8942d99R1451
The
MUST be normalized to have a linear sum of 1.0requirement was morphed into aSHOULD be as close as reasonably possible to '1.0' for a given vertex.requirement.However, glTF-Validator still marks this as an error rather than a warning.
I'm not sure if this is a problem with the specification or validator. However, as the specification is no longer in pre-release, I'm assuming this is the way that it shall stand.
To explain, I believe this might be an error with the specification as it appears that https://github.com/KhronosGroup/glTF/commit/94ed6813112aa2286eb342fd9d493ba872975bc8 may have seen the intermediate explanation with "should" in the paragraph and upgraded it to BCP 14, whereas https://github.com/KhronosGroup/glTF/commit/0a61cf8d029ece4e1d895852a8b61371217c7677 stated it was not a breaking change and was only meant to elaborate.