From: "Paul Ranson" 

I'm sure you actually know this stuff but a C++ programmer can write (or if
(s)he's one of the incompetent 'really only suitable for Java' types...)
have written a class that manages a database channel and releases it in its
destructor. Now the compiler does the hard work. This has nothing to do
with memory management. AFAIK Java has the concept of scopes but not any
deterministic way for an object to know when it's gone out of scope. The C#
people just seem to be be perverse about this.

So, literally, C++ users don't have to release channels to databases, and
generally they have to spend very little time indeed worrying about
allocating or releasing memory. They can even use GC if they wish, perhaps
under the aegis of .NET...

Paul

"Adam"  wrote in message
news:428afc6a{at}w3.nls.net...
> Paul Ranson wrote:
>> Object lifetime isn't the same thing as memory management. Consider a
>> database transaction, or any resource lock. Why you want your
>> underpressure, tired etc programmers time after time having to do their
>> own rollbacks or releases is beyond me.
>
> A) So people doing C++ don't have to release channels to db'es or to deal
> with transactions? If so then hey there's one more thing they have to do
> as well as allocating & reclaiming memory.
>
> B) Rollbacks etc i.e. transaction handling is most often a business matter
> & not a languge matter esp wrt distributed transactions where 5 different
> systems must all get the message & understand it & then the whole can be
> committed.
>
>
>>
>> OTOH garbage collection isn't a panacea, however you manage memory if
>> your programmers don't know what they're doing you're heading for
>> trouble.
>>
>
> Indeed howeve rnot having to deallocate memory is one less point of
> failure to worry about. Yes you do get hanging references & yes they can
> be annoying to track down but I have yet to see a hanging ref steadily
> chew up memory till there s none left.
>
> Adam

--- BBBS/NT v4.01 Flag-5