Has anyone ported globe (chrome experiment) to the latest version three times?
Narration:
The globe code is based on (and contains) threejs v40 and the latest version on github is three times r55 . I was wondering if someone (more knowledgeable then me in this area) could transfer the globe to the new version three times?
Long story:
I fiddled with the googles globe from http://www.chromeexperiments.com/globe . I noticed that it is based on an old version of threejs (on github). Using the latest version (and getting the same results!) Has proven to be tricky. Also see this question .
I changed some of the function names and fumbled around some of the parameters, no problem. Then I turned to shaders. This turned out to be more difficult. The old version seems to have a bug three times when you do Mesh.flipSided = true; : normal vectors appear to be different in shaders between old and new versions. But the shader code in the globe was written for this error, so I had to fix the shader code.
Now I have something similar, but the atmosphere and earth combination doesn't work at all. I suspect this 3D bug to play a role in it, but I'm not sure. Again, this is a flipSided bug that may have been exploited by the globe authors.
Ok, um, I'm stuck here. I can do what I wanted to, stick with the old version, but it's kind of bad.
Can anyone shed some light here?
source to share
As per the change at https://code.google.com/p/webgl-globe/source/browse/globe/globe.js :
"November 5, 2012. Updated to three js r52 and tweenjs r7."
So, on the surface it sounds like it was updated ... which parts of the code specifically give you problems?
Currently, instead of Mesh.flipSided = true
in the mesh material you need to set side:THREE.BackSide
.
Can you submit or submit a link to your code for further investigation? The updated version of the globe project sounds like the most worthy job.
source to share