conversion error about obj2gltf HOT 5 CLOSED

@emackey This issue has been popping up and we have an issue open for this CesiumGS/gltf-pipeline#242. I think this should get fixed in the near future.

from obj2gltf.

I believe this is because you have a vertex normal that has a magnitude of 0 on line 13.

from obj2gltf.

Perhaps the converter could do something more useful than throw an uncaught DeveloperError? Can it explain to the user what problem was found, and where?

Better yet, can't the converter just print out a warning and work around the problem with UNIT_Z or something? Real-world data is lumpy and has minor problems sprinkled around. Ordinary users want tools that can tolerate a few bad vectors. You could print a warning, you could put UNIT_Z or UNIT_Y, or you could even copy the zero-length normal through to the output along with the warning. Sure the model won't be perfect but the user will appreciate having it much more than a stack trace. The user would load the bad model into something else, see the misdirected normal vectors, and go fix their model instead of filing bug reports like this one. OK, rant over now, thanks for reading this far 😄

from obj2gltf.

This converts correctly once #49 is in, and mostly in part due to this fix: CesiumGS/cesium#5032

from obj2gltf.

Fixed in #49

from obj2gltf.