[Objective-C] Primitive Data Types

Overview

A summary of primitive data types in Objective-C.

Note

This post is based on Objective-C 2.0.

Introduction

When first studying Objective-C, I think it is a good idea to strictly distinguish between the data types provided by Foundation (Classes or Type Aliases that redefine types) and the primitive types. In this post, I will cover Primitive Data Types.

Objective-C is a strict superset of ANSI C, so you can use most of the data types provided by ANSI C, as well as structs and unions.

Summary

1. Character Data Type

1.1. char (32bit / 64bit)

TYPE	Size	Range
signed char	1byte	Min : -128 Max : 127
unsigned char	1byte	Min : 0 Max : 255

Note: Although the data stored in char is actually an integer, it is sometimes treated as a character data type, so I separated it from the integer data types. char is both an integer data type and a character data type.

2. Integer Data Types

2.1. short int (32bit / 64bit)

TYPE	Size	Range
signed short int	2byte	Min : -32,768 Max : 32,767
unsigned short int	2byte	Min : 0 Max : 65,535

2.2. int (32bit / 64bit)

TYPE	Size	Range
signed int	4byte	Min : -2,147,483,648 Max : 2,147,483,647
unsigned int	4byte	Min : 0 Max : 4,294,967,295

2.3. long int (32bit)

TYPE	Size	Range
signed long int	4byte	Min : -2,147,483,648 Max : 2,147,483,647
unsigned long int	4byte	Min : 0 Max : 4,294,967,295

2.4. long int (64bit)

TYPE	Size	Range
signed long int	8byte	Min : -9,223,372,036,854,775,808 Max : 9,223,372,036,854,775,807
unsigned long int	8byte	Min : 0 Max : 18,446,744,073,709,551,615

2.5. long long int (32bit / 64bit)

TYPE	Size	Range
signed long long int	8byte	Min : -9,223,372,036,854,775,808 Max : 9,223,372,036,854,775,807
unsigned long long int	8byte	Min : 0 Max : 18,446,744,073,709,551,615

3. Floating-Point Data Types

3.1. float (32bit / 64bit)

TYPE	Size	Range
float	4byte	Min : 1.175494e-38 Max : 3.402823e+38

3.2. double (32bit / 64bit)

TYPE	Size	Range
double	8byte	Min : 2.225074e-308 Max : 1.797693e+308

4. Boolean Data Type

4.1. BOOL

BOOL, used to represent boolean values, is actually a data type redefined with typedef. Depending on the compiler type, version, or the platform environment where the binary runs, it may be a redefined form of int or char. In C, before C99, developers used to define the BOOL data type themselves like this:

typedef int bool

#define false 0
#define true  1

bool bValue;
bValue = true;

Then in C99, a data type for storing boolean values was officially introduced. Since developers had already been redefining the int data type as bool before C99 as shown above, the keyword bool could not be used. Instead, the _BOOL keyword was used to support boolean values.

// stdbool.h

typedef _Bool bool;

#define true   (_Bool)1
#define false  (_Bool)0

However, since _BOOL is a bit longer than bool, as you can see above, stdbool.h separately defines and provides the bool data type. So in C99, if you want to use bool instead of _BOOL, you need to include stdbool.h.

You might wonder why we are talking about C when we are looking at Objective-C. As mentioned at the beginning of this post, Objective-C is a strict superset of C, so in Objective-C as well, BOOL could have been a redefined form of int or char depending on the compiler type, version, or platform environment.

“Could have been”? Not “could be”? What does that mean? That refers to the era when Xcode used GCC before adopting Clang. During the GCC era, the underlying data type could have been either int or char depending on the version.

In the past, there were reportedly many issues caused by such differences in boolean data types. Nowadays, C compilers comply reasonably well with the ANSI C standard, so the underlying data types adopted for boolean types are nearly identical, and true/false logical values are well-normalized to 0 and 1, preventing issues.

During the transitional period, boolean data types could contain data other than 0 and 1 (any data the underlying type could represent), which caused many issues due to logical flaws in true/false comparisons. In older versions of Visual Studio that did not strictly follow the ANSI C standard, bool was char and BOOL was int, which reportedly confused developers even further. Keep this rambling history in mind and exercise extra caution when using boolean data types in older C compilers.

So what about the current Xcode that has adopted Clang? Let us check for ourselves since we are curious.

First, let us look inside objc.h in the macOS SDK and iOS SDK.

// objc.h

/// Type to represent a boolean value.

#if defined(__OBJC_BOOL_IS_BOOL)
    // Honor __OBJC_BOOL_IS_BOOL when available.
#   if __OBJC_BOOL_IS_BOOL
#       define OBJC_BOOL_IS_BOOL 1
#   else
#       define OBJC_BOOL_IS_BOOL 0
#   endif
#else
    // __OBJC_BOOL_IS_BOOL not set.
#   if TARGET_OS_OSX || TARGET_OS_MACCATALYST || ((TARGET_OS_IOS || 0) && !__LP64__ && !__ARM_ARCH_7K)
#      define OBJC_BOOL_IS_BOOL 0
#   else
#      define OBJC_BOOL_IS_BOOL 1
#   endif
#endif

#if OBJC_BOOL_IS_BOOL
    typedef bool BOOL;
#else
#   define OBJC_BOOL_IS_CHAR 1
    typedef signed char BOOL;
    // BOOL is explicitly signed so @encode(BOOL) == "c" rather than "C"
    // even if -funsigned-char is used.
#endif

Wow! Even though I only extracted the BOOL-related content from the header, it is complex. Let us go through it one by one.

4.1.1. Decision Condition 1

#if defined(__OBJC_BOOL_IS_BOOL)
    // Honor __OBJC_BOOL_IS_BOOL when available.
#   if __OBJC_BOOL_IS_BOOL
#       define OBJC_BOOL_IS_BOOL 1
#   else
#       define OBJC_BOOL_IS_BOOL 0
#   endif
#else

When __OBJC_BOOL_IS_BOOL is defined, OBJC_BOOL_IS_BOOL is defined as 0 or 1 based on its value.

4.1.2. Decision Condition 2

#else
    // __OBJC_BOOL_IS_BOOL not set.
#   if TARGET_OS_OSX || TARGET_OS_MACCATALYST || ((TARGET_OS_IOS || 0) && !__LP64__ && !__ARM_ARCH_7K)
#      define OBJC_BOOL_IS_BOOL 0
#   else
#      define OBJC_BOOL_IS_BOOL 1
#   endif
#endif

When __OBJC_BOOL_IS_BOOL is not defined, OBJC_BOOL_IS_BOOL is defined as 0 or 1 based on the target operating system, platform, and architecture type.

4.1.3. Decision Condition 3

#if OBJC_BOOL_IS_BOOL
    typedef bool BOOL;
#else
#   define OBJC_BOOL_IS_CHAR 1
    typedef signed char BOOL;
    // BOOL is explicitly signed so @encode(BOOL) == "c" rather than "C"
    // even if -funsigned-char is used.
#endif

If OBJC_BOOL_IS_BOOL, determined by Decision Condition 1 and Decision Condition 2, is true, then bool is redefined as BOOL. If false, signed char is redefined as BOOL.

4.1.4. Let Us Test Whether the Decision Conditions Are Correct

After looking inside the header, I have a rough idea now. Let us write some code and verify by debugging. First, let us create a macOS Command Line Tool project and write the following code.

NSLog(@"__OBJC_BOOL_IS_BOOL = %@", __OBJC_BOOL_IS_BOOL ? @"defined" : @"undefined");
NSLog(@"TARGET_OS_OSX = %@", TARGET_OS_OSX ? @"YES" : @"NO");
NSLog(@"OBJC_BOOL_IS_BOOL = %@", OBJC_BOOL_IS_BOOL ? @"YES" : @"NO");
NSLog(@"OBJC_BOOL_IS_CHAR = %@", OBJC_BOOL_IS_CHAR ? @"YES" : @"NO");

The results are as follows:

__OBJC_BOOL_IS_BOOL = undefined
TARGET_OS_OSX = YES
OBJC_BOOL_IS_BOOL = NO
OBJC_BOOL_IS_CHAR = YES
Program ended with exit code: 0

Looking at the three decision conditions summarized earlier while understanding the output, we can say the following:

“When running a macOS Command Line Tool built with Xcode 11.5 and macOS 10.15 SDK on x86_64, macOS Catalina 10.15.4, BOOL is a data type that redefines char.”

Let us also test on an iOS device. Create an iOS Single View App project and write the following code.

NSLog(@"__OBJC_BOOL_IS_BOOL = %@", __OBJC_BOOL_IS_BOOL ? @"defined" : @"undefined");
NSLog(@"TARGET_OS_OSX = %@", TARGET_OS_OSX ? @"YES" : @"NO");
NSLog(@"OBJC_BOOL_IS_BOOL = %@", OBJC_BOOL_IS_BOOL ? @"YES" : @"NO");

The results are as follows:

__OBJC_BOOL_IS_BOOL = defined
TARGET_OS_OSX = NO
OBJC_BOOL_IS_BOOL = YES
Program ended with exit code: 0

Oh! Then we could also say the following:

“When running an iOS Single View App built with Xcode 11.5 and iOS 13.5 SDK on iPhone 11 Pro, arm64, iOS 13.5.1, BOOL is a data type that redefines int. Because BOOL is a redefined form of bool, and bool is a redefined form of int.”

Wrong – that was not the case. In the last C development environment I used, bool was int, so I naturally assumed it would also be int in Objective-C, but it turned out to be a char data type. At first I was puzzled when sizeof returned 1, but after checking the Apple developer documentation, the details were well documented.

In the end, BOOL can be summarized in the following table:

TYPE	Typedef	True	False
BOOL	signed char	YES	NO

4.1.5. Tip: How to Check Processor Architecture

I tried to check the processor architecture as I would on Linux by running the following command:

arch
i386

The result was i386. No way. Sir, that cannot be right. My Mac should be 64-bit, so what happened? I was startled and did some googling.

• Stack Exchange Question1
• Stack Exchange Question2

Ah, on Mac, arch always returns i386 for Intel CPUs regardless of whether it is PowerPC or Intel. So if you get i386, it means either 2-way universal or 32-bit only.

Then let us check the machine hardware name. Run the following command:

machine
x86_64h

Or:

uname -m
x86_64

Of course. That is more like it. Phew.

5. Instance Pointer Type

5.1. id (32bit)

TYPE	Size	Range
id	4byte	Instance pointer

5.2. id (64bit)

TYPE	Size	Range
id	8byte	Instance pointer

// objc.h

/// An opaque type that represents an Objective-C class.
typedef struct objc_class *Class;

/// Represents an instance of a class.
struct objc_object {
    Class _Nonnull isa  OBJC_ISA_AVAILABILITY;
};

/// A pointer to an instance of a class.
typedef struct objc_object *id;

Although we said that the data id can represent is an instance pointer, to be more precise, it can hold a pointer to the struct objc_object that has an isa pointer as an instance variable. It serves the same role as a void pointer in C, and polymorphism – that is, dynamic binding and dynamic typing – can be implemented using id pointers. The details of isa will be covered when we discuss the Objective-C runtime.

6. Format Specifiers Type

Data Type	Format Character
char	%c
signed short int	%hd %hi %hx %ho
unsigned short int	%hu %hx %ho
signed int	%d %i %x %o
unsigned int	%u %x %o
signed long int	%ld %li %lx %lo
unsigned long int	%lu %lx %lo
signed long long int	%lld %lli %llx %llo
unsigned long long int	%llu %llx %llo
float	%f %e %g %a
double	%f %e %g %a
id	%p

I tested in Xcode 12 and there were no build warnings when mixing format characters between signed and unsigned. Still, for backward compatibility and issue prevention, it is better to use them correctly.

Let Us Verify with Our Own Eyes

Now let us check the size, minimum, and maximum values of each data type ourselves. I do not have a 32-bit device, so I will only test on a 64-bit device.

1. First, Let Us Look at the Size of Each Data Type

1.1. Character Data Type

1.1.1. char (64bit)

NSLog(@"size of char is %ld", sizeof(char));
NSLog(@"size of unsigned char is %ld", sizeof(unsigned char));

size of char is 1
size of unsigned char is 1
Program ended with exit code: 0

1.2. Integer Data Types

1.2.1. short int (64bit)

NSLog(@"size of short int is %ld", sizeof(short int));
NSLog(@"size of unsigned short int is %ld", sizeof(unsigned short int));

size of short int is 2
size of unsigned short int is 2
Program ended with exit code: 0

1.2.2. int (64bit)

NSLog(@"size of int is %ld", sizeof(int));
NSLog(@"size of unsigned int is %ld", sizeof(unsigned int));

size of int is 4
size of unsigned int is 4
Program ended with exit code: 0

1.2.3. long int (64bit)

NSLog(@"size of long int is %ld", sizeof(long int));
NSLog(@"size of unsigned long int is %ld", sizeof(unsigned long int));

size of long int is 8
size of unsigned long int is 8
Program ended with exit code: 0

1.2.4. long long int (64bit)

NSLog(@"size of long long int is %ld", sizeof(long long int));
NSLog(@"size of unsigned long long int is %ld", sizeof(unsigned long long int));

size of long long int is 8
size of unsigned long long int is 8
Program ended with exit code: 0

1.3. Floating-Point Data Types

1.3.1. float (64bit)

NSLog(@"size of float is %ld", sizeof(float));

size of float is 4
Program ended with exit code: 0

1.3.2. double (64bit)

NSLog(@"size of double is %ld", sizeof(double));

size of double is 8
Program ended with exit code: 0

1.4. Boolean Data Type

1.4.1. BOOL (64bit)

NSLog(@"size of BOOL is %ld", sizeof(BOOL));

size of BOOL is 1
Program ended with exit code: 0

1.5. Instance Pointer Type

1.5.1. id (64bit)

NSLog(@"size of id is %ld", sizeof(id));

size of id is 8
Program ended with exit code: 0

2. Next, Let Us Print the Minimum and Maximum Values of Each Data Type

2.1. Character Data Type

2.1.1. char (64bit)

NSLog(@"CHAR MIN is %d, CHAR MAX is %d", CHAR_MIN, CHAR_MAX);
NSLog(@"UCHAR MAX is %u", UCHAR_MAX);

CHAR MIN is -128, CHAR MAX is 127
UCHAR MAX is 255
Program ended with exit code: 0

2.2. Integer Data Types

2.2.1. short int (64bit)

NSLog(@"SHRT MIN is %hd, SHRT MAX is %hd", SHRT_MIN, SHRT_MAX);
NSLog(@"USHRT MAX is %hu", USHRT_MAX);

SHRT MIN is -32768, SHRT MAX is 32767
USHRT MAX is 65535
Program ended with exit code: 0

2.2.2. int (64bit)

NSLog(@"INT MIN is %d, INT MAX is %d", INT_MIN, INT_MAX);
NSLog(@"UINT MAX is %u", UINT_MAX);

INT MIN is -2147483648, INT MAX is 2147483647
UINT MAX is 4294967295
Program ended with exit code: 0

2.2.3. long int (64bit)

NSLog(@"LONG MIN is %ld, LONG MAX is %ld", LONG_MIN, LONG_MAX);
NSLog(@"ULONG MAX is %lu", ULONG_MAX);

LONG MIN is -9223372036854775808, LONG MAX is 9223372036854775807
ULONG MAX is 18446744073709551615
Program ended with exit code: 0

2.2.4. long long int (64bit)

NSLog(@"LLONG MIN is %lld, LLONG MAX is %lld", LLONG_MIN, LLONG_MAX);
NSLog(@"ULLONG MAX is %llu", ULLONG_MAX);

LLONG MIN is -9223372036854775808, LLONG MAX is 9223372036854775807
ULLONG MAX is 18446744073709551615
Program ended with exit code: 0

2.3. Floating-Point Data Types

2.3.1. float (64bit)

NSLog(@"FLT MIN is %f, FLT MAX is %f", FLT_MIN, FLT_MAX);

FLT MIN is 0.000000, FLT MAX is 340282346638528859811704183484516925440.000000
Program ended with exit code: 0

2.3.2. double (64bit)

NSLog(@"DBL MIN is %f, DBL MAX is %f", DBL_MIN, DBL_MAX);

DBL MIN is 0.000000, DBL MAX is 179769313486231570814527423731704356798070567525844996598917476803157260780028538760589558632766878171540458953514382464234321326889464182768467546703537516986049910576551282076245490090389328944075868508455133942304583236903222948165808559332123348274797826204144723168738177180919299881250404026184124858368.000000
Program ended with exit code: 0

2.4. Boolean Data Type

2.4.1. BOOL (64bit)

BOOL yes = YES;
BOOL no = NO;

NSLog(@"YES is %d", yes);
NSLog(@"NO is %d", no);

YES is 1
NO is 0
Program ended with exit code: 0

2.5. Instance Pointer Type

2.5.1. id (64bit)

id object = [[NSObject alloc] init];
NSLog(@"object address is %p", &object);

object address is 0x7ffeeb41b078
Program ended with exit code: 0

The memory address changes with each execution.

So far, we have explored the primitive data types provided by Objective-C. For those of you who like to dig deep like me, you might have additional questions like “What is the difference between NSInteger and int?” or “What is the difference between CGFloat and float?” You can find out by looking inside the headers yourself. Headers are the specification itself.

Although in real-world development, I believe you will primarily use data types that have been redefined with typedef to match the appropriate architecture and platform. However, having a solid understanding of primitive data types will surely come in handy someday.

[+] If anyone has a 32-bit device, please run the code above and leave the results in the comments, and I will update the post.

References

• https://code.tutsplus.com/tutorials/objective-c-succinctly-data-types--mobile-21986
• https://andybargh.com/primitive-data-types-objective-c/
• https://www.tutorialspoint.com/objective_c/objective_c_data_types.htm