Summary:
This diff addresses the issues raised by kmagiera in https://github.com/facebook/react-native/pull/7884. Transforms should be applied in the order they are defined, just like in `processTransform.js`. A scale applied before a translation, for instance, should give a different result than a translation applied before a scale.
We leverage CATransform3D to do the heavy lifting. A concatenated transform is passed all the way to `RCTViewPropertyMapper`. It is compared with the transform currently applied to the view, and if different, applied. The same approach is used for opacity.
I think it makes the most sense to do this diffing in `RCTViewPropertyMapper`, as opposed to creating and cleaning up an `_updatedPropsDictionary` each frame in `RCTTransformAnimatedNode` and `RCTStyleAnimatedNode`. The node should keep its full value; applying a minimal set of altered props is an optimization. The higher up this optimization is implemented, the more assumptions it makes. e.g. that there will only ever be a sing
Closes https://github.com/facebook/react-native/pull/9050
Differential Revision: D3658139
fbshipit-source-id: ad6286762ef734084cbdf83c9bd9241190302d34
eb96b7fabc