Proposal: Automatic Resource Management

[Stephen Colebourne]

Joshua Bloch wrote:
> On Sat, Mar 7, 2009 at 10:21 AM, Neal Gafter <neal at gafter.com> wrote:
>> I don't see how that solves the problem.  Are you suggesting that
>> everywhere someone handles an exception they should also check for a
>> chain of ignored exceptions and log them?
> 
> I am not suggesting this. Most people have no need to look at these
> exceptions. Those who do can read them by calling the
> getSuppressedExceptions()method.

Most of the time, there are no exceptions in closing. Sometimes there 
are. And of those, just a few we care about. So its a small %age. Thus, 
we shouldn't necessarily worry too much about it.

The proposal does allows the supressed exceptions to be checked immediately:

try (Writer w = open()) {
   // process
} catch (IOException ex) {
   if (ex.getSuppressedExceptions() > 0) {
     // handle exceptions
   }
}

which is pretty neat.

> If people think it's worth it, the stack
> trace printing code could be modified to include suppressed exceptions.   In
> many cases, that would cause suppressed exceptions to be logged
> automatically. But I have some misgivings about this approach: I suspect
> there are programs that parse stack traces.  While I frown on this behavior,
> and the spec does not guarantee that it works, I'd hate to be responsible
> for breaking these programs.

When excepion causes were added, the stack traces changed. This has 
precedent and is a good idea (as it makes stack traces, the standard 
debugging tool, more useful).

Stephen