Look at the following C# code:

 try
 {
     //some code where you call another class 

     //or another function and it raises an exception
  }
 catch (SomeException e)
 {
     //Log the exception, or something
     throw e;
 }

Looks OK? In fact, using throw e in the last line will destroy the stack of the exception; the information about where it was originally raised (in the other class or function) will have vanished.

Rather, use the following:

 try
 {
     //some code where you call another class 

     //or another function and it raises an exception
  }
 catch (SomeException e)
 {
     //Log the exception, or something
     throw;
 }

This rethrows the exception but preserves the original context of the exception and the stack.