Exception in try block.

[Tejaswini Shirkhedkar]

Hi,
This is the question I came across in one of mock exam sites.

 
class Test{
 
static String raid(){
try
{
throw new Exception();
}
catch (Exception e)
{
int i = 1 / 0 ;
return "Error";
}
finally
{
return "Peace";
}
}
 
public static void main(String args[]){
System.out.println(raid());
}
};
 

Options :

1 Compiler error , return "Error" : statement not reached

2 Compiler error , i = 1 / 0 : Exception must be caught or declared in the throws clause

3 Compiles & prints Error when run

4 Compiles & prints Peace when run

The answer given is option 4. I did compile and run the program and it gives the same output.
I didn't understand the fact that:
In this case, a catch block is throwing an ArithmeticException.This exception is unhandled, then why does it not, either (1) get propogated to the caller?( method main in this case) or (2) give compile time error making the answer option 2 ?
Any explainations?
Thanks

Teju

[ June 27, 2004: Message edited by: Tejaswini Shirkhedkar ]

([C0DE] [/C0DE] tags added - Barry)
[ June 27, 2004: Message edited by: Barry Gaunt ]

[KR Campbell]

ArithmeticException is a runtime exception. It does not need to be declared.

[Gerome Kawa]

Option 4 : prints peace is correct because
- Arithmetic exception is a runtime exception , so won't be caught during compile time.
- finally will always be called !

Hope the explanation helps.

[Tejaswini Shirkhedkar]

Thanks a lot for your replies..
But then, in the output, it should show that java.lang.ArithmeticException etc etc etc..
But the output is just "peace"
Does this mean that the ArithmeticException getting thrown in the catch block is ignored at run time?
Thanks.

[Barry Gaunt]

For me it does throw the ArrayIndexOutOfBoundsException:

 
Now just before execution of main.
execution(void Test.main(String[]))
java.lang.ArrayIndexOutOfBoundsException: 0
at statics.Args.ajc$before$statics_Args$9e(Args.java:12)
at Test.main(Test.java)
  

Notice that it does not print "Error" nor "Peace".

[Tejaswini Shirkhedkar]

Thanks!!
But still not clear
Could you throw a little more light, please?
Thanks

Teju

[Anil Hulikal]

I think your return statement in the finally block is playing a role here. Please try replacing that with a print statement (System.out.println ("peace")). Khalid and Rasmussen's book has the following explanation on the use of transfer statements in finally block:

***
If the finally block executes a control transfer statement such as, a return or a labeled break, then this control transfer statement determines how the execution will proceed�regardless of how the try block or any catch block were executed. In particular, a value returned by a return statement in the finally block will supersede any value returned by a return statement in the try block or a catch block.
***

I wonder if this example is set to demonstrate this point.

Nevertheless, got stumped

[K Anshul]

I got the following warning while compiling the code

 Demo.java:14: warning: finally clause cannot complete normally
                        }
                        ^
1 warning 

(edited: executing -> compiling -Barry)
[ June 28, 2004: Message edited by: Barry Gaunt ]

[Barry Gaunt]

That's correct Anshul, for some reason I did not notice that myself. This problem was discussed a few days ago in this forum. Here it is.

Another is this one.
[ June 28, 2004: Message edited by: Barry Gaunt ]