[JAVA] The end of catastrophic programming # 01 "Do it with the absolute value of an integer"

An ordinary programmer understands intuitively and reflexively, I would like to write an article about a rudimentary and simple story. Because of its rudimentary nature, heavy programmers like me often do it.

If there are any mistakes or inappropriate content, please comment We will consider corrections and adjustments to the content.

1. Absolute value of an integer

In many programming languages, it returns the absolute value of an integer. Functions and methods are available.

For example When I execute ʻint result = Math.abs (-15); , the argument specifies a negative integer, but The signed 15 is assigned to result`. This doesn't mean much to many programmers.

2. Integer type

Many programming languages such as C, Java, and C # have integer types. It's a type for dealing with integers, but this time it's not the main subject in itself. Let's consider an int type that is easy for experienced people in C, Java, and C # to imagine.

I think the size of the int type depends on the language and environment, For now, consider a 32-bit signed integer type.

The range that a 32-bit signed integer type can represent is, for many programming languages, -2,147,483,648 to 2,147,483,647 It should be.

Just looking at this already has a suspicious atmosphere.

3. Consider the absolute value of the smallest integer of the 32-bit signed integer type

ʻInt result = Math.abs (-2147483648);What happens toresult`? As a positive integer, it is up to ** 2147483647 **, so it seems that it will not be good.

By the way, C # System.Math throws an OverflowException.

    int result = System.Math.Abs(Int32.MinValue);
    Console.WriteLine("OverflowException occurs");

In Java's java.lang.Math, the absolute value is the same, but it is a negative integer -2147483648. (This works as specified.)

int result = Math.abs(Integer.MIN_VALUE);

//It's an absolute value,-Manma of 2147483648.
System.out.printf("result = %d%n", result);

For C / C ++ cmath, in my environment it was a negative integer -2147483648. What happens here is undefined, so it may have different results depending on the environment.

#include <cmath>

int main()
  int result = abs(INT_MIN);

  return 0;

4. The end

If you write code that may pass the minimum value of a signed integer type to a function or method that returns an absolute value, "** Eventually **" There is a possibility that something wrong will happen.

The point is "someday", and in this example, the problem is Of the 4,294,967,296 integer values from -2,147,483,648 to 2,147,483,647 There will be only one (-2,147,483,648 only). The problem may not be obvious. Smaller integer types should be more problematic.

5. Countermeasures

If by design you don't have the possibility of passing a signed integer minimum, that's fine. If not, you need to take action.

Basically, special processing is done for the minimum value of the signed integer type. For example, implement so that the minimum value is excluded, If you also need the absolute value of the minimum value, you may want to use an integer type that can handle larger sizes. (Even larger integer types have problems with minimum values.)

Depending on the requirements, it is possible to use an integer type with arbitrary precision such as BigInteger. For example, in Java and C #, there are BigInteger classes and structs, There is also an abs () method (static method in C #). With BigInteger, you don't have to worry about overflowing. (You may run out of memory, but I would like to pay tribute to anyone who deals with numbers that are so huge that they run out of memory. )

Also, if possible, build unit tests that can evaluate the minimum value of a signed integer type.

6. Acknowledgments

Thank you to everyone who commented.

The content related to @ saka1029's comment will be an independent article in the future. @tsuyoshi_cho's comment has been reflected in this article.

Recommended Posts

The end of catastrophic programming # 01 "Do it with the absolute value of an integer"
The end of catastrophic programming # 03 "Comparison of integers, if" a> b ", assume that it is" a --b> 0 ""
Specify the default value with @Builder of Lombok
Test the contents of an Excel file with JUnit
You can do it with copy! Aspect-oriented programming (Android)
The difference between programming with Ruby classes and programming without it
An example of a small work when you want to divide the definition value according to the environment but do not want to be aware of it
The story of toString () starting with passing an array to System.out.println
[Java] Output the result of ffprobe -show_streams in JSON and map it to an object with Jackson
An error occurred in the free course of RubyOnRails Udemy, solved it, and went through to the end