Also, the blender systems should depend on the base value computing system, so that the base values are correctly computed by the time they're used for blending.
#jira UE-123542
#rb max.chen
[CL 17405210 by ludovic chabant in ue5-main branch]
- Add support for double precision channels, curves, evaluation, blending, and all other runtime infrastructure.
- Note that, as usual for now, double channels load and save float values.
- Editor side also gains some new track editors for these new types, with some workarounds to correctly recognize between float and double vectors.
- Transform tracks are now operating entirely in doubles.
- Float recomposing APIs for keying tracks in the editor are now using doubles, and have been renamed to "value recomposing".
#rb max.chen
#preflight 6123f6d9e7a3070001ff37ed
[CL 17278174 by ludovic chabant in ue5-main branch]