Use CGFloat instead of float? They don't do it

Question

[I might be misunderstanding this entire topic, as I've grown up with languages that allow the dev to almost entirely ignore processor architecture, like Java, except in some particular cases. Please correct me if I have some of the concepts wrong.]

Reading here it seems that the advice is to use CGFloat instead of, say, float, because it future-proofs my code for different processor architectures (64-bit handles float differently). Assuming that is right, then why does UISlider, for instance, use float directly (for the value)? Wouldn't it be wrong (or something) for me to read their float and convert it to a CGFloat, because in any case my code is not right if the architecture changes anyway?

Jason McCreary · Accepted Answer · 2013-08-20T12:32:11.577

6

CGFloat is just a typedef for float. This provides flexibility for CGFloat to be something different down the road. Which is why using it future-proofs your code. Objective-C does this with many types, NSInteger is another example.

Although they can be used interchangeably, I agree that it doesn't appear in the case of UISlider that Apple was dogfooding.

edited Aug 20 '13 at 12:32

answered Jul 09 '10 at 15:38

Jason McCreary

71,546
23
135
174

+1 so what do you think the implications are for code that reads from and writes to a value on a slider? Just use float? – Dan Rosenstark Jul 09 '10 at 15:40
I say if it's a `float` use `float` ;) Consistency is what matters here. Ideally it would be `CGFloat` throughout. – Jason McCreary Jul 09 '10 at 15:44

score 2 · Answer 2 · answered Jun 01 '11 at 18:26

2

CGFloat is part of Core Graphics, and used for things like pixel values, which has changed greatly over the years (anyone here a TRS-80 lover?). SO, using CGFloat for display-related drawing values is advised.

The value of the slider is not graphics-related. It is simply 0.0 to 1.0, which can be handled precisely by a float, and probably a 16-bit float at that.

answered Jun 01 '11 at 18:26

PapaSmurf

2,345
1
20
10

So float is float is float, even if the architecture changes? – Dan Rosenstark Jun 03 '11 at 22:57
what float actually is depends on the architecture (refer to endianess). From a programming perspective, you can always be guaranteed that it will 'work' the same way. Similarly, CGFloat will always give you good result for computer-graphic related computations, no matter what the actual display and architecture. – PapaSmurf Jun 06 '11 at 23:03

Use CGFloat instead of float? They don't do it

2 Answers2

Linked