When I try to apply the patch, I found too many change on eth type defines, for example you change so functions to take explicit types float and double.
Like these functions
- Code: Select all
void TransformTriplex (float* const dst, dgInt32 dstStrideInBytes, const float* const src, dgInt32 srcStrideInBytes, dgInt32 count) const;
void TransformTriplex (double* const dst, dgInt32 dstStrideInBytes, const dgFloat64* const src, dgInt32 srcStrideInBytes, dgInt32 count) const;
void TransformTriplex (double* const dst, dgInt32 dstStrideInBytes, const float* const src, dgInt32 srcStrideInBytes, dgInt32 count) const;
That’s not what I had in mind, in my mind the change be a single define in command lines, and the code should compile correctly.
I though a making the build and make the nesseray change, but I was busy with the destruction and automatic uvUnWrap code for the Appling the texture to the chunks.
Let us do this, give a day or two to bring the destruction code to stable point, make the project with double in a separate branch.
Basically all that is needed to do is to make the
-full implementation of the dgVector using dgFloat64 and
-Define dgFloat32 to be double,
-compile and put if def out to all of the function that will conflict because they take explicit dgFloa32 and dgFloat64 and parameters
That about a weekend of work, I will do that and you do the test.
Bu I cannot take that patch that explicitly declares function to take float and double, because that make the code no cross plalform.
Is that OK?
