Bump version to 0.2.0-beta and fixed section culling

gradle.properties

@@ -11,7 +11,7 @@ loader_version=0.14.21
 fabric_version=0.84.0+1.20.1
 
 # Mod Properties
-mod_version=0.1.17-alpha
+mod_version=0.2.0-beta
 maven_group=me.cortex
 archives_base_name=nvidium
 # Dependencies

src/main/java/me/cortex/nvidium/managers/SectionManager.java

@@ -81,8 +81,11 @@ public void uploadChunkBuildResult(ChunkBuildOutput result) {
         Vector3i min  = output.min();
         Vector3i size = output.size();
 
+        //NOTE:TODO: The y encoded height position only has a range of like 6 bits max, that gives 18 bits free/spare for something
+        // realistically it would only be 16 free bits cause ee but still thats 2 bytes free
+
         int px = section.getChunkX()<<8  | size.x<<4 | min.x;
-        int py = section.getChunkY()<<24 | size.y<<4 | min.y;
+        int py = section.getChunkY()<<24 | size.y<<4 | min.y;//TODO: figure out how to make this << 8 just like the others
         int pz = section.getChunkZ()<<8  | size.z<<4 | min.z;
         int pw = addr;
         new Vector4i(px, py, pz, pw).getToAddress(segment);

src/main/resources/assets/nvidium/shaders/occlusion/section_raster/mesh.glsl

@@ -42,6 +42,7 @@ void emitParital(int visIndex) {
     gl_MeshPrimitivesNV[gl_LocalInvocationID.x+8].gl_PrimitiveID = visIndex;
 }
 
+//TODO: Check if the section can be culled via fog
 void main() {
     int visibilityIndex = (int)(_visOutBase|gl_WorkGroupID.x);
 
@@ -53,9 +54,10 @@ void main() {
     vec3 mins = (header.xyz&0xF)-ADD_SIZE;
     vec3 maxs = mins+((header.xyz>>4)&0xF)+1+(ADD_SIZE*2);
     ivec3 chunk = ivec3(header.xyz)>>8;
-    chunk.y >>= 16;
+    chunk.y >>= 16;//TODO: figure out how to remove this from here and in the SectionManager, when i remove it everything dies
     ivec3 relativeChunkPos = (chunk - chunkPosition.xyz);
     vec3 corner = vec3(relativeChunkPos<<4);
+    vec3 cornerCopy = corner;
 
     //TODO: try mix instead or something other than just ternaries, i think they get compiled to a cmov type instruction but not sure
     corner += vec3(((gl_LocalInvocationID.x&1)==0)?mins.x:maxs.x, ((gl_LocalInvocationID.x&4)==0)?mins.y:maxs.y, ((gl_LocalInvocationID.x&2)==0)?mins.z:maxs.z);
@@ -68,12 +70,14 @@ void main() {
         emitParital(prim_payload);
     }
     if (gl_LocalInvocationID.x == 0) {
-        //ivec3 absRelPos = abs(relativeChunkPos);
-        //int maxDist = min(absRelPos.x, min(absRelPos.y, absRelPos.z));
+        cornerCopy += subchunkOffset.xyz;
+        vec3 minPos = mins + cornerCopy;
+        vec3 maxPos = maxs + cornerCopy;
+        bool isInSection = all(lessThan(minPos, vec3(ADD_SIZE))) && all(lessThan(vec3(-ADD_SIZE), maxPos));
 
         //Shift and set, this gives us a bonus of having the last 8 frames as visibility history
-        //sectionVisibility[visibilityIndex] = uint8_t(lastData<<1) | uint8_t(maxDist<=1?1:0);//Inject visibility aswell
-        sectionVisibility[visibilityIndex] = uint8_t(lastData<<1) | uint8_t(0);
+        sectionVisibility[visibilityIndex] = uint8_t(lastData<<1) | uint8_t(isInSection?1:0);//Inject visibility aswell
+        //sectionVisibility[visibilityIndex] = uint8_t(lastData<<1) | uint8_t(0);
 
         gl_PrimitiveCountNV = 12;
     }

src/main/resources/assets/nvidium/shaders/terrain/translucent/mesh.glsl

@@ -74,8 +74,8 @@ void main() {
     vec4 addiAO;
     vec4 tintBO;
     vec4 addiBO;
-    computeFog(isSphericalFog, posA+subchunkOffset.xyz, tintA, fogColour, fogStart, fogEnd, tintAO, addiAO);
-    computeFog(isSphericalFog, posB+subchunkOffset.xyz, tintB, fogColour, fogStart, fogEnd, tintBO, addiBO);
+    computeFog(isCylindricalFog, posA+subchunkOffset.xyz, tintA, fogColour, fogStart, fogEnd, tintAO, addiAO);
+    computeFog(isCylindricalFog, posB+subchunkOffset.xyz, tintB, fogColour, fogStart, fogEnd, tintBO, addiBO);
     OUT[(gl_LocalInvocationID.x<<1)|0].tint = tintAO;
     OUT[(gl_LocalInvocationID.x<<1)|0].addin = addiAO;
     OUT[(gl_LocalInvocationID.x<<1)|1].tint = tintBO;