I think this is what you have in mind:
- (void) logRects
{
CGFloat
x = 10.0,
y = 20.0,
width = 50.0,
height = 60.0;
CGPoint point = {x, y};
CGSize size = {width, height};
CGRect rect1 = {1, 3, size};
CGRect rect2 = {point, size};
CGRect rect3 = {point, size.width, size.height};
//using designated (named) initialisers
CGRect rect4 = {.origin.x=3, .origin.y=5, .size = {100,100}};
//with designated initialisers, order doesn't matter
CGRect rect5 = {.size=size, .origin.x=3, .origin.y=5};
NSLog (@"rect1 %@",NSStringFromCGRect(rect1));
NSLog (@"rect2 %@",NSStringFromCGRect(rect2));
NSLog (@"rect3 %@",NSStringFromCGRect(rect3));
NSLog (@"rect4 %@",NSStringFromCGRect(rect4));
NSLog (@"rect5 %@",NSStringFromCGRect(rect5));
}
But note the discussion here:
Why use functions like CGRectMake?
This kind of compound literal syntax seems to me much easier to read and write, although functions have the edge when it comes to futureproofing ( + you get autocomplete).
update
see also this more recent q&a: