PARMANOIR

Quick note : if you have a container CALayer containing layers arranged in a grid, have a sublayerTransform setup on it, and want to get its transformed layer coordinates on a mouseDown, you'll need a helper layer :

// We're in mouseDown
CGPoint point = NSPointToCGPoint([self convertPoint:event.locationInWindow fromView:nil]);

// Build a dummy CALayer and insert it
CALayer* helperLayer = [CALayer layer];
[containerLayer addSublayer:helperLayer];

// Get coordinates
CGPoint layerPoint = [[self layer] convertPoint:point toLayer:helperLayer];

// Remove our helper layer
[helperLayer removeFromSuperlayer];

Now why can't we just ask for containerLayer's coordinates ? Because it has a sublayerTransform, so its coordinates are really the same as its parent. We can't ask for one of containerLayer.sublayers either : being laid out (layouted ? :) ) in a grid, they each have their own coordinate system. We therefore use a dummy layer : it has a default position of [0, 0] and will give the point transformed trough sublayerTransform.