您的位置:首页 > 编程语言 > C语言/C++

C++、Java和C#中的异常处理

2008-06-09 00:53 555 查看
来源:温少的日志

C++中异常:
1、可以抛出任何异常。虽然标准C++定义了std::exception类,但throw的对象可以是任何对象。包括int、char*等。例如:


throw 100;






try

{


// do something





} catch (int num)

{


// do something



}

但是,C++标准库中抛出的所有异常,都是派生自std::exception类。
std::exception定义如下:




class exception

{


public:


exception( );


exception(const char *const&);


exception(const char *const&, int);


exception(const exception&);


exception& operator=(const exception&);


virtual ~exception( );


virtual const char *what( ) const;


};

2、标准C++中,是没有finally子句。熟悉Java和C#的朋友,也许会觉得奇怪!没有finally子句,使用的确不方便。

3、C++中,函数的声明,缺省时可以抛出任何异常的,也可以使用异常规范来约束函数抛出的异常,如:


//可以抛出任何异常




void f()

{


throw 100;


}




//描述该函数不会抛出任何异常




void f() throw()

{


}




//描述该函数会抛出int或者char*类型的异常




void f() throw(int, char*)

{


// do something



}

但是,C++中的异常规范不是很可靠。如下的代码也是可以编译通过的,但是编译器会给出警告。




void f() throw(int, char*)

{


throw "100";


}

4、C++的异常可以派生自多个异常基类。这一个特性使得难以分类的异常容易处理,例如网络文件的异常:




class file_exception

{}




class network_exception

{}




class newwork_file_exception : public file_exception, public network_exception

{}

5、捕捉所有异常的语法。C++中,不是单根继承的,所以你不能,如同Java或者C#那样,通过捕捉某个基类来捕捉所有的异常。




try

{


// do something





} catch (

)

{ //捕捉所有的异常


// do something



}

总结,C++中,命名空间和异常规范是作为针对大规模程序设计而引入的特性。但是由于C++中没有垃圾收集机制,异常造成了资源管理变得非常复杂。在C++中,编写异常安全的代码,是十分困难的。



Java中的异常:

1、可以抛出任何派生自了Throwable的类。Throwable类的定义如下:




public class Throwable implements Serializable

{


public Throwable();


public Throwable(String message);


public Throwable(String message, Throwable cause); //@since 1.4


public Throwable(Throwable cause); //@since 1.4


public String getMessage();


public String getLocalizedMessage(); //@since JDK1.1


public Throwable getCause(); //@since 1.4


public synchronized Throwable initCause(Throwable cause); //@since 1.4


public String toString();


public void printStackTrace();


public void printStackTrace(PrintStream s);


public void printStackTrace(PrintWriter s); //@since JDK1.1


private void printStackTraceAsCause(PrintWriter s,


StackTraceElement[] causedTrace);


public synchronized native Throwable fillInStackTrace();


public StackTraceElement[] getStackTrace(); //@since 1.4


public void setStackTrace(StackTraceElement[] stackTrace); //@since 1.4


}

Throwable下分Exception和Error。




public class Exception extends Throwable

{}




public class Error extends Throwable

{}

而Exception中,有一个特别的派生类,RuntimeException,如下:
public class RuntimeException extends Exception {}
类图如下:



Java divides exceptions in two categories: checked exceptions and unchecked exceptions. Exceptions that are not "checked" by the compiler are called unchecked exceptions; exceptions that are "checked" by the compiler are called checked exceptions. The compiler will not compile a file if it contains a method where a checked exception may occur and the method does not handle the checked exception in a catch block or list it in the method header. The classes
RuntimeException
,
Error
and their subclasses are unchecked exceptions. All other exception classes are checked exceptions.

If a checked exception might occur in a method and the method does not have
catch
block to handle the checked exception, then the checked exception must be declared in the method header using a
throws
clause. For example, if an
IOException
might occur in method
readInteger
and
readInteger
doesn't handle it, then
readInteger
must have the following header:

public int readInteger() throws IOException {

}


If
readInteger
does not list
IOException
in its
throws
clause, compilation will fail with an error message about checked exceptions. Unchecked exceptions may be listed in a
throws
clause, but the compiler does not require it.

In the following code, neither method
readInteger
nor method
main( )
catches
IOException
so they both list
IOException
in their
throws
clauses. When method
main
throws an
IOException
, the exception object is passed to the Java Virtual Machine (JVM). The JVM displays the exception information and terminates the program.

1:
2:
3:
4:
5:
6:
7:
8:
9:
10:
11:
12:
13:
14:
15:
16:
17:
18:
19:
20:
21:
22:
23:
24:
25:
26:
27:
28:
29:
30:
31:
32:
33:
34:
35:
36:
37:
38:
39:
40:
41:
42:
43:
44:
45:
46:
47:
48:
49:
50:
51:
52:
import java.io.*;

/**
* This class provides a method that reads an integer from the
* standard input.
*
* @author author name
* @version 1.0.0
*/
public class IntegerReaderThrowsException  {

private static BufferedReader  stdIn =
new BufferedReader(new  InputStreamReader(System.in));

private static PrintWriter  stdErr =
new PrintWriter(System.err, true);

private static PrintWriter  stdOut =
new PrintWriter(System.err, true);

/**
* Tests method <code>readInteger</code>
*
* @param args  not used
* @throws IOException if error reading from standard input.
*/
public static void main (String[] args) throws IOException {

stdOut.println("The value is: " + readInteger());
}

/**
* Reads an integer from the standard output.
*
* @return the <code>int</code> value.
* @throws IOException if error reading from standard input.
*/
public static int  readInteger() throws IOException {

do  {
try  {
stdErr.print("Enter an integer> ");
stdErr.flush();

return Integer.parseInt(stdIn.readLine());

} catch (NumberFormatException  nfe)  {
stdErr.println("Invalid number format");
}
}  while (true);
}
}

Listing 4 IntegerReaderThrowsException.java
2、 Java中,也使用异常规范,并且在整个基础类库中都使用。Java是基于JVM运行,但是JVM本身就抛出异常,也就是,所有的方法都可能抛出异常。因 此,Java编译器编译不检查Error和RuntimeException。与C++不同,没有异常规范,缺省不允许抛出任何非 RuntimeException和Error的异常。如下:




public void f()

{


throw new RuntimeException(); //正确


}






public void f()

{


throw new Error(); //正确


}

以下的代码编译出错




public void f()

{


throw new Exception();


}

使用异常规范的正确例子:




public WfException extends Exception

{}






public void f() throws WfException

{


throw new WfException();


}

3、Throwable、Exception、RuntimeException中部分成员,是在不同的JDK版本中添加的,编写代码的时候需要注意。如下的代码不能在JDK 1.3中编译通过:




public void f() throws WfException

{




try

{


// do something





} catch (Exception e)

{


throw new WfException ("", e);


}


}

4、Java中的try语法,包括finally子句。如下:




try

{


// do something





} catch (Exception e)

{


// do something





} finally

{


// do something



}

C#中的异常:
1、C#异常的局限性。C#是基于CLR运行的,由于CLR需要支持多种语言,其异常的 实现受到了较大的局限。一个重要的特征就是没有异常规范,很多人对此很失望!Anders Hejlsberg曾专门发表过文章,说明为什么C#不支持异常规范,说得也很有道理。但是,没有异常规范的C#,变得简单,但也更容易犯错误,在这一点 上,我更喜欢Java,而不喜欢C#的方式。没有异常规范,编译器就不会帮我们检查代码,是否做了try ... catch处理。

2、异常的结构。在.NET Framework中,所有的异常派生自System.Exception基类。其下有两个子类:SystemException和 ApplicationException。自行编写异常类,不建议直接派生自Exception类,而是应该派生自 ApplicationException。

3、C#犹如Java,也支持finally的写法。如:


try






{


// do something



}


catch (Exception e)






{


// do something



}


finally






{


// do something
内容来自用户分享和网络整理,不保证内容的准确性,如有侵权内容,可联系管理员处理 点击这里给我发消息
标签: