Add support for partial rendering.

[X-Ryl669]

This actually allows performing rendering in passes and does not require a width * height * 4 bytes buffer anymore.

On limited memory MCU, this will allow rendering a large animation by trading off memory for CPU consumption instead.

Typically instead of this rendering loop:

   buffer = new uint8[stride * h]; // Huge
   for (i = 0; i < totalFrame; i++) {
        lottie_animation_render(animation, i, buffer, w, h, stride);
   }
   sendBufferToDisplay(buffer);

This allows this kind of loop:

   const int lineCount = 10;
   buffer = new uint8[stride * lineCount]; // Small
   for (i = 0; i < totalFrame; i++) {
       for (l = 0; l < h; l += lineCount) {
          lottie_animation_render_partial(animation, i, buffer, w, h, l, l+lineCount, stride);
          sendBufferToDisplay(buffer, l);
      }
   }

The overhead is higher than a single rendering (mainly, clearing the buffer and computing the clip path n times) but allow to have a lot smaller buffers so it's very useful for limited/embedded system with almost no RAM like ESP32 or STM32.

[X-Ryl669]

CI errors are not related to my changes. You'll need to fix the CI pipeline to include the uint and uchar typedef that's used in the elsewhere in the code.