Why use "finally"

  • Thread starter Thread starter Guest
  • Start date Start date
G

Guest

Hi,

This is probably a simple question but I'm curious. What is the purpose of
the "finally" statement? There must be a good reason to use it that I'm just
not getting...

How is this:

method1
{
try
{
x;
}
catch
{
y;
}
finally
{
z;
}
}

different than:

method1
{
try
{
x;
}
catch
{
y;
}

z;
}

Thank you,
Joe
 
Often, there will be return statements either in the Try or the Catch, or
both. Putting this code in the finally guarantees that it will be executed
in any case. Where as otherwise, you would have to put a copy of it in the
Try and Catch right before the return.
 
Hi Manohar,

Thanks for the reply. I read the article but I still don't see the
difference. Maybe if the example actually showed a resource being released
instead of a Console.WriteLine it would be easier to understand...

Joe
 
The answer is that in your second example, "z" is not guaranteed to be
executed, whereas in your first example, it is.

One caveat, though. Your specific examples were a bit unrealistic: it's
bad style to simply have a "catch" without saying what to catch, a
"catch-all" if you'll pardon the pun. So the way that your examples are
written, the only way that "z" will not be executed in the second
example is if "y" throws an exception (inside the catch block).

A more realistic case would be one in which the "catch" blocks catch
specific exceptions, in which case you need "finaly" in case some other
kind of exception that you're not catching occurs.

Sometimes people (like me) even use "finally" with no "catch" at all:
we want the exceptions to percolate up to the caller, but we have to be
sure to release some resource, reset some value, or change some setting
before leaving.

In my code, changing the cursor to an hourglass is a good example:

// Change the cursor to an hourglass
try
{
.... do a long, slow, complicated operation...
}
finally
{
// Change the cursor back to the default
}

I change the cursor back to a default in the finally block because I
want that to happen _even if_ the complex operation inside the "try"
blows up. In fact, the operation is complex enough that it has already
blown up on me several times for reasons beyond my control. If my
"Change the cursor back" code weren't within the "finally" then my
cursor would be stuck as an hourglass.
 
Hi Marina,

That actually makes some sense to me. I guess if no returns were coded in
the try or catch blocks then the second way I showed would work the same as
the first then?

Thanks again,
Joe
 
Joe...
try
{
OpenToiletLid();
Flush();
}
catch(OverflowException e)
{
CallPlumber();
}
finally
{
CloseToiletLid();
}

Regards,
Jeff
 
I think my wife would like that if I could find a way to implement the
finally at home :-)
 
Hello!
I think my wife would like that if I could find a way to implement the
finally at home :-)
Click to expand...

LOL!

Another approach - ask your wife to implement ICareless :-)
 
The code in the finally block will execute before the method exits,
even if exceptions are thrown and regardless of if those exceptions
are caught or not.

In your second method, z will not execute if an exception that is not
handled in the catch block is thrown.

Very often, try -- finally is used without a catch block. C# even has
special
shortcut for that, the using statement.

- Magnus
 
Back
Top