Try ... Catch ... Finally confusion

[Sean Kirkpatrick]

I got caught with my pants down the other day when trying to explain
Try...Catch...F inally and things didn't work as I had assumed. Perhaps
someone can explain to me the purpose of Finally. I've looked at several
texts that I have and none of them address this specific point.

If I call some method that throws an exception in my routine Foo,

sub foo
call bar <- throws an exception
do something else <- never get here
end sub

control is passed back up the call stack to the first valid catch block.
So far so good.

Now if I do this

sub foo
try
call bar (which throws an exception)
catch
handle the exception <- only if exception thrown
end try

do something else <- since exception handled, do this

end sub

execution goes to the catch block and then continues to do something
else. So far so good.

If I do this

sub foo
try
call bar (which throws an exception)
catch
handle the exception <- only if exception thrown
finally
always do this
end try

do something else <- since exception handled, do this

end sub

the Finally block ALWAYS gets executed regardless of an exception being
thrown. Ok.

I submit that the second case is identical to the third case if I
rewrite number 2 as follows:


sub foo
try
call bar (which throws an exception)
catch
handle the exception <- only if exception thrown
end try

always do this
do something else <- since exception handled, do this

end sub

This begs the question - what is the use of Finally? I had assumed,
wrongly as it turns out, that in the event of an exception, Catch and
Finally would execute and control would return to the caller. If no
exception, Finally would execute and control continues to the next
statement.

I can see no use for Finally!

Sean

[Mythran]

Re: Try ... Catch ... Finally confusion

[color=blue]
> sub foo
> try
> call bar (which throws an exception)
> catch
> handle the exception <- only if exception thrown
> finally
> always do this
> end try
>
> do something else <- since exception handled, do this
>
> end sub
>
> the Finally block ALWAYS gets executed regardless of an exception being
> thrown. Ok.
>
> I submit that the second case is identical to the third case if I
> rewrite number 2 as follows:
>
>
> sub foo
> try
> call bar (which throws an exception)
> catch
> handle the exception <- only if exception thrown
> end try
>
> always do this
> do something else <- since exception handled, do this
>
> end sub
>
> This begs the question - what is the use of Finally? I had assumed,
> wrongly as it turns out, that in the event of an exception, Catch and
> Finally would execute and control would return to the caller. If no
> exception, Finally would execute and control continues to the next
> statement.
>
> I can see no use for Finally!
>
> Sean[/color]

The finally always runs. After the finally, it's iffy. For example:

Dim conn As SqlConnection = New SqlConnection(. ..)
Dim ds As DataSet
Try
ds = GetDataFromData base(conn)
Catch ex As SqlException
' Log the error message or handle it some other way.
Return
Finally
' Perform cleanup, whether or not an exception was raised as well
' as whether or not we are exiting the function because of a
' raised exception, we still need to close the instance of the
' SqlConnection object...we can do that here. If we try to
' Return anywhere in a Try..Catch block, Finally will still be
' executed.
conn.Close()
End Try

' We only get this far if an exception was not thrown. The Finally
' block above was executed and the connection was closed.
DoSomethingWith TheDataSet(ds)

HTH! :)

Mythran

[m.posseth]

Re: Try ... Catch ... Finally confusion

Finally always runs even if you are in a function and perform a hard return
statement finally is run ,, so it is the perfect place to put in code that
should run nomather what ( for example object cleanup code )

regrds

Michel Posseth

"Sean Kirkpatrick" <nannernanner@c ommunity.nospam > wrote in message
news:%23KfWeNtv FHA.1256@TK2MSF TNGP09.phx.gbl. ..[color=blue]
>I got caught with my pants down the other day when trying to explain
>Try...Catch... Finally and things didn't work as I had assumed. Perhaps
>someone can explain to me the purpose of Finally. I've looked at several
>texts that I have and none of them address this specific point.
>
> If I call some method that throws an exception in my routine Foo,
>
> sub foo
> call bar <- throws an exception
> do something else <- never get here
> end sub
>
> control is passed back up the call stack to the first valid catch block.
> So far so good.
>
> Now if I do this
>
> sub foo
> try
> call bar (which throws an exception)
> catch
> handle the exception <- only if exception thrown
> end try
>
> do something else <- since exception handled, do this
>
> end sub
>
> execution goes to the catch block and then continues to do something else.
> So far so good.
>
> If I do this
>
> sub foo
> try
> call bar (which throws an exception)
> catch
> handle the exception <- only if exception thrown
> finally
> always do this
> end try
>
> do something else <- since exception handled, do this
>
> end sub
>
> the Finally block ALWAYS gets executed regardless of an exception being
> thrown. Ok.
>
> I submit that the second case is identical to the third case if I rewrite
> number 2 as follows:
>
>
> sub foo
> try
> call bar (which throws an exception)
> catch
> handle the exception <- only if exception thrown
> end try
>
> always do this
> do something else <- since exception handled, do this
>
> end sub
>
> This begs the question - what is the use of Finally? I had assumed,
> wrongly as it turns out, that in the event of an exception, Catch and
> Finally would execute and control would return to the caller. If no
> exception, Finally would execute and control continues to the next
> statement.
>
> I can see no use for Finally!
>
> Sean[/color]

[topdawg147@hotmail.com]

Re: Try ... Catch ... Finally confusion

Sometimes you don't want to handle the error at this function. That is,
you'll handle it somewhere else up the chain. If this happens, you
still need to clean up everything. Also, you don't have to use Catch in
a Try statement.

Try

Cursors.Current = Cursor.WaitCurs or
ListView1.Begin Update()

' Do stuff.

Finally
Cursors.Current = Cursor.Default
ListView1.EndUp date()
End Try

[Paul Clement]

Re: Try ... Catch ... Finally confusion

On Wed, 21 Sep 2005 10:39:05 -0700, Sean Kirkpatrick <nannernanner@c ommunity.nospam > wrote:


¤ If I do this
¤
¤ sub foo
¤ try
¤ call bar (which throws an exception)
¤ catch
¤ handle the exception <- only if exception thrown
¤ finally
¤ always do this
¤ end try
¤
¤ do something else <- since exception handled, do this
¤
¤ end sub
¤
¤ the Finally block ALWAYS gets executed regardless of an exception being
¤ thrown. Ok.
¤
¤ I submit that the second case is identical to the third case if I
¤ rewrite number 2 as follows:
¤
¤
¤ sub foo
¤ try
¤ call bar (which throws an exception)
¤ catch
¤ handle the exception <- only if exception thrown
¤ end try
¤
¤ always do this
¤ do something else <- since exception handled, do this
¤
¤ end sub
¤
¤ This begs the question - what is the use of Finally? I had assumed,
¤ wrongly as it turns out, that in the event of an exception, Catch and
¤ Finally would execute and control would return to the caller. If no
¤ exception, Finally would execute and control continues to the next
¤ statement.
¤
¤ I can see no use for Finally!

Yes, Finally always executes when the flow of control leaves the Try statement. It is an optional
keyword.

It guarantees the execution of code in the Finally block if an exception is generated in the Catch
block or when you execute an Exit Try statement.


Paul
~~~~
Microsoft MVP (Visual Basic)

[Sean Kirkpatrick]

Re: Try ... Catch ... Finally confusion

Paul Clement wrote:[color=blue]
> ¤ I can see no use for Finally!
>
> It guarantees the execution of code in the Finally block if an exception is generated in the Catch
> block or when you execute an Exit Try statement.
>[/color]
Ok, this answer makes some sense - if I'm in a Catch block and *another*
exception occurs, Finally will ensure that the original cleanup code
gets called, though I submit that in the case of Exit Try, there is no
difference between my original examples 2 & 3.

I'm going to have to play with it a bit to grock it completely.

Thanks Paul!

Sean

[david]

Re: Try ... Catch ... Finally confusion

On 2005-09-21, Sean Kirkpatrick <nannernanner@c ommunity.nospam > wrote:[color=blue]
> Paul Clement wrote:[color=green]
>> ¤ I can see no use for Finally!
>>
>> It guarantees the execution of code in the Finally block if an exception is generated in the Catch
>> block or when you execute an Exit Try statement.
>>[/color]
> Ok, this answer makes some sense - if I'm in a Catch block and *another*
> exception occurs, Finally will ensure that the original cleanup code
> gets called,[/color]

Keep in mind that often *another* exception is your own...

Try
DoSomething()
Catch(ex As Exception)
LogSomeInformat ion(ex.Message)
Throw
Finally
CleanUp()
End Try

....or, something I commonly do...

Try
DoSomething()
Catch(ex as Exception)
Throw New SpecificExcepti onType(ex)
Finally
....

[color=blue]
> though I submit that in the case of Exit Try, there is no
> difference between my original examples 2 & 3.[/color]

They are, but that's because you're eating the exception, which should
be a very, very rare thing. Usually, Try-Finally blocks are much more
common than Try-Catch blocks..

Try
DoSomething()
Finally
' do your cleanup, and let the exception propagate up
....

[Cor Ligthert [MVP]]

Re: Try ... Catch ... Finally confusion

Sean,

Know that you can set a try block in a try block

\\\
Try
connection.open
Try
do the reading
Catch
reader errohandling
return
End Try
Catch
connection open error handling
Finally
connection close 'this will always be done whatever error
End Try
///

I hope this helps,

Cor