Fix signed 16.16 fixed-point conversion.

src/common/version.h

 #define MAJOR_VERSION 1
 #define MINOR_VERSION 0
 #define BUILD_VERSION 0
-#define BUILD_REVISION 1044
+#define BUILD_REVISION 1045
 
 #define STRINGIFY(x) #x
 #define MACRO_STRINGIFY(x) STRINGIFY(x)

src/libGLESv2/VertexDataManager.cpp

@@ -448,8 +448,8 @@ struct ConversionRule : gl::Cast<typename GLToCType<fromType>::type, typename D3
 template <GLenum fromType> struct ConversionRule<fromType, true, D3DVT_FLOAT> : gl::Normalize<typename GLToCType<fromType>::type> { };
 
 // Use a full specialisation for this so that it preferentially matches ahead of the generic normalize-to-float rules.
-template <> struct ConversionRule<GL_FIXED, true, D3DVT_FLOAT> : gl::FixedToFloat<GLuint, 16> { };
-template <> struct ConversionRule<GL_FIXED, false, D3DVT_FLOAT> : gl::FixedToFloat<GLuint, 16> { };
+template <> struct ConversionRule<GL_FIXED, true, D3DVT_FLOAT> : gl::FixedToFloat<GLint, 16> { };
+template <> struct ConversionRule<GL_FIXED, false, D3DVT_FLOAT> : gl::FixedToFloat<GLint, 16> { };
 
 // A 2-stage construction is used for DefaultVertexValues because float must use SimpleDefaultValues (i.e. 0/1)
 // whether it is normalized or not.