¿ªÔÆÌåÓýI agree, my point is that you should ALWAYS be given the object rather than it belonging to you. Therefore, the whole IDisposable pattern seems flawed from a design perspective. ? All that said, you can still test drive a flawed design¡. J ? From: testdrivendevelopment@... [mailto:testdrivendevelopment@...]
On Behalf Of Brad Wilson
Sent: Monday, February 3, 2014 7:48 AM To: testdrivendevelopment@... Subject: Re: [TDD] How do I unit test a Dispose method ? ? ? ...except, if you're given the object rather than creating it yourself, then its lifetime doesn't belong to you, and you don't need to be concerned with whether it's disposable. ? On Sun, Feb 2, 2014 at 11:36 PM, Amir Kolsky <amir.kolsky@...> wrote: ? If A cares about this specific property it would require B to implement IDisposable. From A¡¯s perspective, it is given an object of type B to perform a specific action. B would be available to A through an interface, not an implementation. Hence A would not know that (say) B1 is the implementation of B, but rather know about B alone ? ? From:
testdrivendevelopment@... [mailto:testdrivendevelopment@...]
On Behalf Of David Burstin ? ? Amir, I feel like I might be missing your point.? ? Why is it unreasonable for A to know that B uses expensive resources and provides a way to release them early? Surely this was one of the considerations when choosing B over X or Y or Z. IMHO it's a characteristic, not an implementation detail. ? On 3 February 2014 15:19, Amir Kolsky <amir.kolsky@...> wrote: ? ? Except that B must already be IDisposable. The IDisposable pattern requires that any object that owns an IDisposable must itself implement IDisposable. ? n? This is inane, as it requires the owning object to know about implementation details of it¡¯s ownee. I disagree. All that A needs to know is that B implements the IDisposable interface. ? ? * A has to know about something pertaining to B that has nothing to do with why A needs it. ? ? ?
|
