Corona provides transitions. According to this post, a transitionTo(...) would suffice:

local rect = display.newRect(50,50,100,100)
rect:setFillColor(0,0,0,0)
transition.to( rect.fill, { r=1, g=1, b=1, a=1,
  time=500,
  transition=easing.inCubic }
)

The user can also cancel the transition if needed (which is waht I would need to do later in the code). In case you're into custom stuff and the easing functions are not enough, you can create your own. A sample can be found here.

The signature of the easing function is (at the moment of writing) function customEasing(t, tMax, start, delta) where:

  • t is the time elapsed since the start
  • tMax is the duration of the transition
  • start is 1 if t in this call of the easing function differs from the t in the last call, otherwise 0
  • delta is 1 if is the last acall of the easign function with the same value of t

An example is:

t tMax start delta
21:49:21.088 948.4 3000 0 0
21:49:21.088 948.4 3000 0 1
21:49:21.108 960.2 3000 1 0
21:49:21.108 960.2 3000 0 0
21:49:21.108 960.2 3000 0 0
21:49:21.108 960.2 3000 0 1
21:49:21.119 975 3000 1 0
21:49:21.119 975 3000 0 0
21:49:21.119 975 3000 0 0
21:49:21.119 975 3000 0 1

The data from above table was created using the following easing function:

local function customEasing(t, tMax, start, delta)
  print (t .. " | " .. tMax .. " | " .. start .. " | " .. delta)
  return start
end

IMPORTANT: The easing function must return relevant data for the value being transitioned!

NOTABLE: A final call is made with the final value of the transition, regardless of the easing function.

Now it's easy to create custom transition functions.

There are advantages and disadvantages of using this approach comared to the previous post (game loop approach). Advantages are:

  • Code is significantly leaner
  • You can concentrate to writing your own easing function instead of folling your source with boilerplate
  • The table above shows this function gets called serveral times for each frame, arguably allowing finer transition calculations

Disadvantages are:

  • A continuous change means we need to restart the transition while in the game loop approach it would mean wwe only need to change the business logic to accomodate the changes.
  • There is a risk to perform needless calculations in the easing function, resulting in pointless use of resources

Which is the best approach? I think it depends... If you only need to e.g. fade in/out an image, then a transition is the way to go. If you need to cycle through colors, then it depends on your requirements (e.g. accuracy, control).

In my case I'll need to continuously cycle through colors so I'm leaning towards the game loop approach.

HTH,