在平常项目开发中,异常抛出和捕获是再日常不过的事情。经过try-catch咱们能够方便的捕获异常,同时经过查看异常堆栈咱们能发现抛出异常代码的位置。ide
例以下面这段代码:spa
1 using System; 2 using System.IO; 3 using System.Runtime.CompilerServices; 4 using System.Runtime.InteropServices; 5 6 namespace ExamplesProject 7 { 8 class Program 9 { 10 static int m = 0; 11 static void Main(string[] args) 12 { 13 try 14 { 15 int a = 10, b = 0; 16 Console.WriteLine(a / b); 17 } 18 catch (Exception ex) 19 { 20 Console.WriteLine(ex.ToString()); 21 } 22 23 Console.ReadLine(); 24 } 25 } 26 }
这段代码很是简单,运行后他抛出了以下异常:code
System.DivideByZeroException: Attempted to divide by zero. at ExamplesProject.Program.Main(String[] args) in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 16
没有问题,堆栈信息明确指出了抛出异常的位置,也正是除零异常发生的位置。对象
假如由于种种缘由,须要把main方法的代码逻辑抽取到另外的方法里,像下面这样:blog
1 using System; 2 using System.IO; 3 using System.Runtime.CompilerServices; 4 using System.Runtime.InteropServices; 5 6 namespace ExamplesProject 7 { 8 class Program 9 { 10 static int m = 0; 11 static void Main(string[] args) 12 { 13 try 14 { 15 Divide(); 16 } 17 catch (Exception ex) 18 { 19 Console.WriteLine(ex.ToString()); 20 } 21 22 Console.ReadLine(); 23 } 24 25 public static void Divide() 26 { 27 try 28 { 29 int a = 10, b = 0; 30 Console.WriteLine(a / b); 31 } 32 catch (Exception ex) 33 { 34 throw ex; 35 } 36 } 37 } 38 }
这段代码乍一看,好像也没有什么问题,divide的方法本身捕获并从新抛出了异常。开发
再来看一下异常信息:string
System.DivideByZeroException: Attempted to divide by zero. at ExamplesProject.Program.Divide() in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 34 at ExamplesProject.Program.Main(String[] args) in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 15
异常堆栈显示,抛出异常的方法是Divide(),异常代码位置是34行。可这里并非异常真正的发生位置,真正的异常位置是第30行。it
发生了什么??为何不是第30行。io
再来看一个例子。class
假如业务变复杂了,代码又嵌套了一层。调用关系变复杂了,main()->divide()->divide1()。
1 using System; 2 using System.IO; 3 using System.Runtime.CompilerServices; 4 using System.Runtime.InteropServices; 5 6 namespace ExamplesProject 7 { 8 class Program 9 { 10 static int m = 0; 11 static void Main(string[] args) 12 { 13 try 14 { 15 Divide(); 16 } 17 catch (Exception ex) 18 { 19 Console.WriteLine(ex.ToString()); 20 } 21 22 Console.ReadLine(); 23 } 24 25 public static void Divide() 26 { 27 try 28 { 29 Divide1(); 30 } 31 catch (Exception ex) 32 { 33 throw ex; 34 } 35 } 36 37 public static void Divide1() 38 { 39 try 40 { 41 int a = 10, b = 0; 42 Console.WriteLine(a / b); 43 } 44 catch (Exception ex) 45 { 46 throw ex; 47 } 48 } 49 } 50 }
运行代码,抛出以下异常:
System.DivideByZeroException: Attempted to divide by zero. at ExamplesProject.Program.Divide() in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 33 at ExamplesProject.Program.Main(String[] args) in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 15
为何异常位置是33行,而不是42行呢。代码嵌套的越深,错误隐藏的越深。若是是在大型系统中,将很难发现错误的真正位置。
做者就犯过这样的错误啊。问题到底在哪里呢?
问题的根源是,以上代码抛出的异常并无体现层次关系,或者说异常的嵌套关系。理论上讲,divide1()方法中catch捕获到的异常为最内层异常,catch捕获后处理完,若是要向上层抛出,应该从新实例化一个新的异常对象,并将当前异常做为其innerException, 再向上抛出。以此类推,divide()方法捕获处理后若是要抛出异常也应该这样作。这样最外层的main方法catch到的异常才是完整的异常,天然包含完整的堆栈信息,错误定位就是精准的。
改造后的例子:
1 using System; 2 using System.IO; 3 using System.Runtime.CompilerServices; 4 using System.Runtime.InteropServices; 5 6 namespace ExamplesProject 7 { 8 class Program 9 { 10 static int m = 0; 11 static void Main(string[] args) 12 { 13 try 14 { 15 Divide(); 16 } 17 catch (Exception ex) 18 { 19 Console.WriteLine(ex.ToString()); 20 } 21 22 Console.ReadLine(); 23 } 24 25 public static void Divide() 26 { 27 try 28 { 29 Divide1(); 30 } 31 catch (Exception ex) 32 { 33 throw new Exception(ex.Message, ex); 34 } 35 } 36 37 public static void Divide1() 38 { 39 try 40 { 41 int a = 10, b = 0; 42 Console.WriteLine(a / b); 43 } 44 catch (Exception ex) 45 { 46 throw new Exception(ex.Message, ex); 47 } 48 } 49 } 50 }
运行后,抛出以下异常:
System.Exception: Attempted to divide by zero. ---> System.Exception: Attempted to divide by zero. ---> System.DivideByZeroException: Attempted to divide by zero. at ExamplesProject.Program.Divide1() in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 42 --- End of inner exception stack trace --- at ExamplesProject.Program.Divide1() in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 46 at ExamplesProject.Program.Divide() in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 29 --- End of inner exception stack trace --- at ExamplesProject.Program.Divide() in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 33 at ExamplesProject.Program.Main(String[] args) in D:\Projects\ExamplesProject\ExamplesProject\Program.cs:line 15
这一次,异常堆栈精准的定位了异常代码行第42行。
总结,以上例子很是简单,但是在实际开发中,咱们老是会不经意忽略一些细节,最终致使上了生产后异常格外的难以追踪。