为什么 NSNumber 值相等时指向同一个地址？

发布于 2024-10-04 10:04:01 字数 332 浏览 4 评论 0原文

给出以下代码：

int firstInt, secondInt;

firstInt = 5;
secondInt = 5;

NSNumber *firstNumber = [NSNumber numberWithInt:firstInt];
NSNumber *secondNumber = [NSNumber numberWithInt:secondInt];

到底为什么这两个 NSNumber 实例指向同一个地址？

这让我发疯！

当然，如果您将 secondaryInt 更改为“4”，则一切都会按预期工作。

谢谢，杰里米

原文

Given the following code:

int firstInt, secondInt;

firstInt = 5;
secondInt = 5;

NSNumber *firstNumber = [NSNumber numberWithInt:firstInt];
NSNumber *secondNumber = [NSNumber numberWithInt:secondInt];

Why on Earth do those two NSNumber instances are pointing to the same address?

This drives me crazy!

Of course, if you change secondInt to, say '4', all works as expected.

Thanks,
Jérémy

分享到QQ

分享到微博

如果你对这篇内容有疑问，欢迎到本站社区发帖提问参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。

发布评论

需要登录才能够评论，你可以免费注册一个本站的账号。

时间海 2024-10-11 10:04:01

这可能是编译器优化或实现细节：由于 NSNumber 是不可变的，因此它们不需要是单独的实例。

编辑：可能是考虑到它的实现优化。当随后使用相同的整数调用时， numberWithInt 可能会返回一个单例。

回复收藏 0 原文

风轻花落早 2024-10-11 10:04:01

我的设计直觉告诉我，如果分数的身份与它的单纯值不同很重要，那么您应该对某种分数对象而不是普通的 NSNumber 进行排序。

但是除此之外：在紧要关头，您可以像使用 NSNumber 一样使用普通 NSValue。获取值需要做更多的工作，但 NSValue 本身不具有 NSNumber 对小值所做的实例合并行为。

一些执行所有三种行为的代码：

  // NSValues are always distinct:
  int foo = 5, bar = 5, outfoo, outbar;
  NSValue *one = [NSValue value:&foo withObjCType:@encode(int)];
  NSValue *two = [NSValue value:&bar withObjCType:@encode(int)];

  [one getValue:&outfoo];
  [two getValue:&outbar];
  NSLog(@"one: %@ %x = %d ; two: %@ %x = %d",
        [one class], one, outfoo,
        [two class], two, outbar);

  // by comparison with NSNumber behavior:
  NSNumber *three = [NSNumber numberWithInt:6];
  NSNumber *four = [NSNumber numberWithInt:6];

  NSLog(@"three: %@ %x = %d ; four: %@ %x = %d",
        [three class], three, [three intValue],
        [four class], four, [four intValue]);

  // except when the numbers are big:
  NSNumber *five = [NSNumber numberWithInt:8675309];
  NSNumber *six = [NSNumber numberWithInt:8675309];

  NSLog(@"five: %@ %x = %d ; six: %@ %x = %d",
        [five class], five, [five intValue],
        [six class], six, [six intValue]);

在我的 mac 上，这会产生如下输出：

one: NSConcreteValue 42a8d0 = 5 ; two: NSConcreteValue 42a920 = 5
three: NSCFNumber 404380 = 6 ; four: NSCFNumber 404380 = 6
five: NSCFNumber 1324d0 = 8675309 ; six: NSCFNumber 106e00 = 8675309

My design instinct tells me that, if the score's identity is important as distinct from its mere value, you should be sorting some kind of score object instead of plain NSNumbers.

But that aside: In a pinch, you can use plain NSValue similarly to how you're using NSNumber. It's a little more work to get values out, but NSValue itself doesn't have the instance coalescing behavior NSNumber does for small values.

Some code that exercises all three behaviors:

  // NSValues are always distinct:
  int foo = 5, bar = 5, outfoo, outbar;
  NSValue *one = [NSValue value:&foo withObjCType:@encode(int)];
  NSValue *two = [NSValue value:&bar withObjCType:@encode(int)];

  [one getValue:&outfoo];
  [two getValue:&outbar];
  NSLog(@"one: %@ %x = %d ; two: %@ %x = %d",
        [one class], one, outfoo,
        [two class], two, outbar);

  // by comparison with NSNumber behavior:
  NSNumber *three = [NSNumber numberWithInt:6];
  NSNumber *four = [NSNumber numberWithInt:6];

  NSLog(@"three: %@ %x = %d ; four: %@ %x = %d",
        [three class], three, [three intValue],
        [four class], four, [four intValue]);

  // except when the numbers are big:
  NSNumber *five = [NSNumber numberWithInt:8675309];
  NSNumber *six = [NSNumber numberWithInt:8675309];

  NSLog(@"five: %@ %x = %d ; six: %@ %x = %d",
        [five class], five, [five intValue],
        [six class], six, [six intValue]);

On my mac this yields output like:

one: NSConcreteValue 42a8d0 = 5 ; two: NSConcreteValue 42a920 = 5
three: NSCFNumber 404380 = 6 ; four: NSCFNumber 404380 = 6
five: NSCFNumber 1324d0 = 8675309 ; six: NSCFNumber 106e00 = 8675309

回复收藏 0 原文