Updated MotionEvent to use Transform

Updated MotionEvent to use Transform instead of offset and scale. This
allowed a lot of simplification since it no longer needs to manually do
the transformation calculation and instead can rely on functions
provided by the Transform class.

Test: Input works
Test: inputflinger_tests
Test: libinput_tests
Bug: 158476194
Change-Id: Id40ae0184ca778507166e60d0d3eaf9564f3172b
diff --git a/services/inputflinger/tests/InputFlingerService_test.cpp b/services/inputflinger/tests/InputFlingerService_test.cpp
index d6543f2..02342c0 100644
--- a/services/inputflinger/tests/InputFlingerService_test.cpp
+++ b/services/inputflinger/tests/InputFlingerService_test.cpp
@@ -291,8 +291,8 @@
     mInfo.frameBottom = TestInfoFrameBottom;
     mInfo.surfaceInset = TestInfoSurfaceInset;
     mInfo.globalScaleFactor = TestInfoGlobalScaleFactor;
-    mInfo.transform.set(std::array<float, 9>{TestInfoWindowXScale, 0, TestInfoFrameLeft, 0,
-                                             TestInfoWindowYScale, TestInfoFrameTop, 0, 0, 1});
+    mInfo.transform.set({TestInfoWindowXScale, 0, TestInfoFrameLeft, 0, TestInfoWindowYScale,
+                         TestInfoFrameTop, 0, 0, 1});
     mInfo.touchableRegion = TestInfoTouchableRegion;
     mInfo.visible = TestInfoVisible;
     mInfo.canReceiveKeys = TestInfoCanReceiveKeys;