Swap width and height in TouchVideoFrame

Currently in TouchVideoFrame and related code, the field 'width'
precedes the field 'height'.
But these fields should be  interpreted as follows:
width -> number  of columns
height -> number of rows
In most notations in mathematics, the matrix sizes are designated as
'mxn', where m is the number of rows and n is the number of columns.

So we make TouchVideoFrame consistent with this, and swap these 2
fields.

Test: atest libinput_tests
The actual test is added in the next commit.
Bug: 123241238

Change-Id: I808e7f354bd7b62d5599324eef205bf4450a91c1
diff --git a/include/input/TouchVideoFrame.h b/include/input/TouchVideoFrame.h
index a0f4f27..566c334 100644
--- a/include/input/TouchVideoFrame.h
+++ b/include/input/TouchVideoFrame.h
@@ -30,20 +30,20 @@
  */
 class TouchVideoFrame {
 public:
-    TouchVideoFrame(uint32_t width, uint32_t height, std::vector<int16_t> data,
+    TouchVideoFrame(uint32_t height, uint32_t width, std::vector<int16_t> data,
             const struct timeval& timestamp);
 
     bool operator==(const TouchVideoFrame& rhs) const;
 
     /**
-     * Width of the frame
-     */
-    uint32_t getWidth() const;
-    /**
      * Height of the frame
      */
     uint32_t getHeight() const;
     /**
+     * Width of the frame
+     */
+    uint32_t getWidth() const;
+    /**
      * The touch strength data.
      * The array is a 2-D row-major matrix, with dimensions (height, width).
      * Total size of the array should equal getHeight() * getWidth().
@@ -56,8 +56,8 @@
     const struct timeval& getTimestamp() const;
 
 private:
-    uint32_t mWidth;
     uint32_t mHeight;
+    uint32_t mWidth;
     std::vector<int16_t> mData;
     struct timeval mTimestamp;
 };