Re: [Cobertura-devel] coverage of try/catch

[Elliotte H. <el...@me...>]

Torsten Curdt wrote:
> Hi there,
> 
> I am wondering how you guys think empty try/catch
> should be handled.
> 
> Let's assume I have
> 
>   try {
>     Thread.sleep(1000);
>   } catch(InterruptedException e) {
>     ;
>   }


That block is not empty. It contains an empty statement. Possibly the 
compiler will optimize this away. I'm not sure.

You definitely should test this catch block if you can. That is, you 
should write a test case in which the sleeping thread is interrupted and 
then verify that everything still behaves as it should. If you really 
think it's impossible for the thread to be interrupted, then you should 
rewrite the catch block along these lines:

catch(InterruptedException e) {
     throw new RuntimeException(e, "Something impossible happened") ;
}

-- 
Elliotte Rusty Harold  el...@me...
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim