If i understand your question correctly... one way is to test for color on drawing new shapes. If the color under the new shape is the color of inside of your big shape, draw otherwise not. Alternatively, if the color of the draw is no good for test (same color of other thing in the scene) you may draw a copy of the big shape in a not displayed PGraphics and test against it.

If you don't need the inner shapes to fit perfectly within the bigger shape (i.e. you can have parts of the smaller shapes cut off) there is a much 'cheaper' way; draw all your inner shapes then superimpose a mask of your larger shape over them!

Excuse my poor drawing but here's the idea:

1 Draw inner shapes
2 Load/create mask
3 Superimpose mask
4 Remove unwanted part

If the shapes will tessellate and sizes are known in advance this could work with whole shapes within.

@vk: i think i understand what you're saying. it may have limitations though, i will try that.

I would look at using a mask, my guess is it would be far less computationally expensive - it doesn't have to have a hard edge and could be generated at runtime if you just want something like this T shape.

EDIT: A 'mask' could be a simple array representing black/white pixels (or grey pixels!) test each pixel you want to draw (i.e. in each ellipse) against the array - draw the pixel if array is white, or don't if it's black.

Here's a (very small) T shaped mask in an array.

11111
10001
11011
11011
11011

The problem is you have to check every single pixel you intend to draw (or jump through a lot of tricky calculations). Far easier/quicker to draw all the inner shapes in full then apply a proper mask :D

That is indeed amazing. I didn't know this. Glad to learn. Thanks.