I thought unsigned int meant unsigned long int (32 bit) by default. Likewise, I thought unsigned long meant unsigned long int by default (32 bit). Why the two selectors on NSNumber (unsignedIntValue and unsignedLongValue), or should I be worried about the data type bit lengths?
Thanks!
On LP64 (Mac OS 64 Bit) a long is 64 bit wide while an int has only 32 bit.
I understand OSX and iOS have limited architectures, but in general, you should not make any assumptions about integer sizes, other than the basic relative ordering guaranteed in the standard.
If you really need a specific sized integral value, use the uint8_t, uint16_t, uint32_t, and uint64_t types.
Related
When should I be using NSInteger vs. int when developing for iOS? I see in the Apple sample code they use NSInteger (or NSUInteger) when passing a value as an argument to a function or returning a value from a function.
- (NSInteger)someFunc;...
- (void)someFuncWithInt:(NSInteger)value;...
But within a function they're just using int to track a value
for (int i; i < something; i++)
...
int something;
something += somethingElseThatsAnInt;
...
I've read (been told) that NSInteger is a safe way to reference an integer in either a 64-bit or 32-bit environment so why use int at all?
You usually want to use NSInteger when you don't know what kind of processor architecture your code might run on, so you may for some reason want the largest possible integer type, which on 32 bit systems is just an int, while on a 64-bit system it's a long.
I'd stick with using NSInteger instead of int/long unless you specifically require them.
NSInteger/NSUInteger are defined as *dynamic typedef*s to one of these types, and they are defined like this:
#if __LP64__ || TARGET_OS_EMBEDDED || TARGET_OS_IPHONE || TARGET_OS_WIN32 || NS_BUILD_32_LIKE_64
typedef long NSInteger;
typedef unsigned long NSUInteger;
#else
typedef int NSInteger;
typedef unsigned int NSUInteger;
#endif
With regard to the correct format specifier you should use for each of these types, see the String Programming Guide's section on Platform Dependencies
Why use int at all?
Apple uses int because for a loop control variable (which is only used to control the loop iterations) int datatype is fine, both in datatype size and in the values it can hold for your loop. No need for platform dependent datatype here. For a loop control variable even a 16-bit int will do most of the time.
Apple uses NSInteger for a function return value or for a function argument because in this case datatype [size] matters, because what you are doing with a function is communicating/passing data with other programs or with other pieces of code; see the answer to When should I be using NSInteger vs int? in your question itself...
they [Apple] use NSInteger (or NSUInteger) when passing a value as an
argument to a function or returning a value from a function.
OS X is "LP64". This means that:
int is always 32-bits.
long long is always 64-bits.
NSInteger and long are always pointer-sized. That means they're 32-bits on 32-bit systems, and 64 bits on 64-bit systems.
The reason NSInteger exists is because many legacy APIs incorrectly used int instead of long to hold pointer-sized variables, which meant that the APIs had to change from int to long in their 64-bit versions. In other words, an API would have different function signatures depending on whether you're compiling for 32-bit or 64-bit architectures. NSInteger intends to mask this problem with these legacy APIs.
In your new code, use int if you need a 32-bit variable, long long if you need a 64-bit integer, and long or NSInteger if you need a pointer-sized variable.
If you dig into NSInteger's implementation:
#if __LP64__
typedef long NSInteger;
#else
typedef int NSInteger;
#endif
Simply, the NSInteger typedef does a step for you: if the architecture is 32-bit, it uses int, if it is 64-bit, it uses long. Using NSInteger, you don't need to worry about the architecture that the program is running on.
You should use NSIntegers if you need to compare them against constant values such as NSNotFound or NSIntegerMax, as these values will differ on 32-bit and 64-bit systems, so index values, counts and the like: use NSInteger or NSUInteger.
It doesn't hurt to use NSInteger in most circumstances, excepting that it takes up twice as much memory. The memory impact is very small, but if you have a huge amount of numbers floating around at any one time, it might make a difference to use ints.
If you DO use NSInteger or NSUInteger, you will want to cast them into long integers or unsigned long integers when using format strings, as new Xcode feature returns a warning if you try and log out an NSInteger as if it had a known length. You should similarly be careful when sending them to variables or arguments that are typed as ints, since you may lose some precision in the process.
On the whole, if you're not expecting to have hundreds of thousands of them in memory at once, it's easier to use NSInteger than constantly worry about the difference between the two.
On iOS, it currently does not matter if you use int or NSInteger. It will matter more if/when iOS moves to 64-bits.
Simply put, NSIntegers are ints in 32-bit code (and thus 32-bit long) and longs on 64-bit code (longs in 64-bit code are 64-bit wide, but 32-bit in 32-bit code). The most likely reason for using NSInteger instead of long is to not break existing 32-bit code (which uses ints).
CGFloat has the same issue: on 32-bit (at least on OS X), it's float; on 64-bit, it's double.
Update: With the introduction of the iPhone 5s, iPad Air, iPad Mini with Retina, and iOS 7, you can now build 64-bit code on iOS.
Update 2: Also, using NSIntegers helps with Swift code interoperability.
As of currently (September 2014) I would recommend using NSInteger/CGFloat when interacting with iOS API's etc if you are also building your app for arm64.
This is because you will likely get unexpected results when you use the float, long and int types.
EXAMPLE: FLOAT/DOUBLE vs CGFLOAT
As an example we take the UITableView delegate method tableView:heightForRowAtIndexPath:.
In a 32-bit only application it will work fine if it is written like this:
-(float)tableView:(UITableView *)tableView heightForRowAtIndexPath:(NSIndexPath *)indexPath
{
return 44;
}
float is a 32-bit value and the 44 you are returning is a 32-bit value.
However, if we compile/run this same piece of code in a 64-bit arm64 architecture the 44 will be a 64-bit value. Returning a 64-bit value when a 32-bit value is expected will give an unexpected row height.
You can solve this issue by using the CGFloat type
-(CGFloat)tableView:(UITableView *)tableView heightForRowAtIndexPath:(NSIndexPath *)indexPath
{
return 44;
}
This type represents a 32-bit float in a 32-bit environment and a 64-bit double in a 64-bit environment. Therefore when using this type the method will always receive the expected type regardless of compile/runtime environment.
The same is true for methods that expect integers.
Such methods will expect a 32-bit int value in a 32-bit environment and a 64-bit long in a 64-bit environment. You can solve this case by using the type NSInteger which serves as an int or a long based on the compile/runtime environemnt.
int = 4 byte (fixed irrespective size of the architect)
NSInteger = depend upon size of the architect(e.g. for 4 byte architect = 4 byte NSInteger size)
I am creating a bitmask for the iOS using the data type int32_t. This is then set to a variable that accepts an NSInteger. This throws no compile time errors as expected, but I was wondering - is there was some way that this could cause run-time errors in the future?
In general, it will work. NSInteger will always be at least 32 bits long. If it is 64 bits, your number will be sign extended to match (you might need to cast it). It will not cause run-time errors.
Detailed info about NSInteger
NSInteger is defined using the following code (from NSObjCRuntime.h):
#if __LP64__ || TARGET_OS_EMBEDDED || TARGET_OS_IPHONE || TARGET_OS_WIN32 || NS_BUILD_32_LIKE_64
typedef long NSInteger;
typedef unsigned long NSUInteger;
#else
typedef int NSInteger;
typedef unsigned int NSUInteger;
#endif
This means that NSInteger is equivalent to long when compiling for 64 bit, an embedded OS, iPhone, Windows, or when building 32 bit like 64 bit. Otherwise, it is equivalent to int. Since int and int32_t are equivalent, you will be OK in 32 bit applications on OS X. For the other situations, it depends on the size of long.
In 64 bit OS X applications, long is a 64 bit number. On the iPhone, long is a 32 bit number. I do not know about Windows, and I am not sure whether it uses the 32 bit or 64 bit long when building 32 bit like 64 bit.
Equivalent to 'long' in Java what do we have in Objective C, NSInteger?
In Java, a long is always 64 bits. In C and Objective-C, a long might be 64 bits, or it might be 32 bits, or (in less common cases) it might be something else entirely; the C standard doesn't specify an exact bit width.
On OS X, an NSInteger is 64 bits on 64-bit platforms, and 32 bits on 32-bit platforms. 32-bit Mac platforms are increasingly rare, so you can probably use NSInteger and be fine.
However, if you always want a 64-bit integer, you'll probably want to use the int64_t data type defined in stdint.h.
A Java long is defined as a signed 64-bit value, neither long nor NSInteger guarantee this for Objective-C. For example, on 32-bit systems plaforms, NSInteger and long are 32-bit signed values. If your platform comes with C99 headers (for example when your compiler is gcc based), then you should have stdint.h which has platform independent definitions for integer types with guaranteed sizes. The 64 bit signed type is named int64_t.
#include <stdint.h>
int64_t someVariable; // 64 bit signed integer, like Java's long
You didn't ask, but int32_t is the analogue to Java's int type (a 32-bit integer).
A NSInteger is 32 bits on 32-bit platforms, and 64 bits on 64-bit platforms. Is there a NSLog specifier that always matches the size of NSInteger?
Setup
Xcode 3.2.5
llvm 1.6 compiler (this is important; gcc doesn't do this)
GCC_WARN_TYPECHECK_CALLS_TO_PRINTF turned on
That's causing me some grief here:
#import <Foundation/Foundation.h>
int main (int argc, const char * argv[]) {
#autoreleasepool {
NSInteger i = 0;
NSLog(#"%d", i);
}
return 0;
}
For 32 bit code, I need the %d specifier. But if I use the %d specifier, I get a warning when compiling for 64 bit suggesting I use %ld instead.
If I use %ld to match the 64 bit size, when compiling for 32 bit code I get a warning suggesting I use %d instead.
How do I fix both warnings at once? Is there a specifier I can use that works on either?
This also impacts [NSString stringWithFormat:] and [[NSString alloc] initWithFormat:].
Updated answer:
You can make use of the z and t modifiers to handle NSInteger and NSUInteger without warnings, on all architectures.
You want to use %zd for signed, %tu for unsigned, and %tx for hex.
This information comes courtesy of Greg Parker.
Original answer:
The official recommended approach is to use %ld as your specifier, and to cast the actual argument to a long.
The accepted answer is absolutely reasonable, it is standard conforming, and correct. The only problem is that it doesn't work anymore, which is completely Apple's fault.
The format %zd is the C/C++ standard format for size_t and ssize_t. Like NSInteger and NSUInteger, size_t and ssize_t are 32 bit on a 32 bit system, and 64 bit on a 64 bit system. And that's why printing NSInteger and NSUInteger using %zd worked.
However, NSInteger and NSUInteger are defined as "long" on a 64 bit system, and as "int" on a 32 bit system (which is 64 vs 32 bit). Today, size_t is defined on "long" on all systems, which is the same size as NSInteger (either 64 or 32 bit), but a different type. Either Apple's warnings have changed (so it doesn't allow passing the wrong type to printf, even though it has the right number of bits), or the underlying types for size_t and ssize_t have changed. I don't know which one, but %zd stopped working some time ago. There is no format today that will print NSInteger without warning on both 32 and 64 bit systems.
So the only thing you can do unfortunately: Use %ld, and cast your values from NSInteger to long, or from NSUInteger to unsigned long.
Once you don't build for 32 bit anymore, you can just use %ld, without any cast.
The formatters come from the standard UNIX/POSIX printf function. Use %lu for unsigned long, %ld for long, %lld for long long, and %llu for unsigned long long. Try man printf on the console, but on Mac it is incomplete. The linux manpages are more explicit http://www.manpages.info/linux/sprintf.3.html
Both warnings can only be fixed by NSLog(#"%lu", (unsigned long)arg); combined with a cast as the code will be compiled in 32 AND 64 bit for iOS. Otherwise each compilation creates a separate warning.
NSInteger/NSUInteger are Cocoa-defined replacements for the regular built-in types.
Is there any benefit to using the NS* types over the built-ins? Which do you prefer and why? Are NSInteger and int the same width on 32-bit / 64-bit platforms?
The way I understand it is that NSInteger et al. are architecture safe versions of the corresponding C types. Basically their size vary depending on the architecture, but NSInteger, for example, is guaranteed to hold any valid pointer for the current architecture.
Apple recommends that you use these to work with OS X 10.5 and onwards, and Apple's API:s will use them, so it's definitely a good idea to get into the habit of using them. They require a little more typing, but apart from that it doesn't seem to be any reason not to use them.
Quantisation issues for 64-bit runtime
In some situations there may be good reason to use standard types instead of NSInteger: "unexpected" memory bloat in a 64-bit system.
Clearly if an integer is 8 instead of 4 bytes, the amount of memory taken by values is doubled. Given that not every value is an integer, though, you should typically not expect the memory footprint of your application to double. However, the way that Mac OS X allocates memory changes depending on the amount of memory requested.
Currently, if you ask for 512 bytes or fewer, malloc rounds up to the next multiple of 16 bytes. If you ask for more than 512 bytes, however, malloc rounds up to the next multiple of 512 (at least 1024 bytes). Suppose then that you define a class that -- amongst others -- declares five NSInteger instance variables, and that on a 32-bit system each instance occupies, say, 272 bytes. On a 64-bit system, instances would in theory require 544 bytes. But, because of the memory allocation strategy, each will actually occupy 1024 bytes (an almost fourfold increase). If you use a large number of these objects, the memory footprint of your application may be considerably greater than you might otherwise expect. If you replaced the NSInteger variables with sint_32 variables, you would only use 512 bytes.
When you're choosing what scalar to use, therefore, make sure you choose something sensible. Is there any reason why you need a value greater than you needed in your 32-bit application? Using a 64-bit integer to count a number of seconds is unlikely to be necessary...
64-bit is actually the raison d'être for NSInteger and NSUInteger; before 10.5, those did not exist. The two are simply defined as longs in 64-bit, and as ints in 32-bit:
#if __LP64__ || NS_BUILD_32_LIKE_64
typedef long NSInteger;
typedef unsigned long NSUInteger;
#else
typedef int NSInteger;
typedef unsigned int NSUInteger;
#endif
Thus, using them in place of the more basic C types when you want the 'bit-native' size.
CocoaDev has some more info.
I prefer the standard c style declarations but only because I switch between several languages and I don't have to think too much about it but sounds like I should start looking at nsinteger
For importing and exporting data to files or over the net I use UInt32, SInt64 etc...
These are guaranteed to be of a certain size regardless of the architecture and help in porting code to other platforms and languages which also share those types.