Sylvain Lafontaine
1/26/2008 8:24:00 AM
No, it's just surprising how fast you can forget some basic things about a
language when it has been some time since the last time you have used it
(and especially but not necessarily when you are doing something else like
waiting for something to finish).
Maybe I should stop trying to make two things at the same time but on the
other hand, a lot of potentially useful posts would be lost; so I suppose
that's probably better to make some basic error from time to time in favor
of the greater good.
--
Sylvain Lafontaine, ing.
MVP - Technologies Virtual-PC
E-mail: sylvain aei ca (fill the blanks, no spam please)
"Carlo Razzeto" <crazzeto@hotmail.com> wrote in message
news:0F759BC2-67BD-49FA-9DEB-6D914AE05379@microsoft.com...
> Ok, lol... After reading what you wrote I was going to have some
> questions, just wondering if there was something I haddn't considered.
> This post tells me that this isn't the case. :)
>
> Carlo
>
> "Sylvain Lafontaine" <sylvain aei ca (fill the blanks, no spam please)>
> wrote in message news:eJxfWD6XIHA.5784@TK2MSFTNGP03.phx.gbl...
>> Sorry, forget about my last comment as I've misread the original post.
>>
>> --
>> Sylvain Lafontaine, ing.
>> MVP - Technologies Virtual-PC
>> E-mail: sylvain aei ca (fill the blanks, no spam please)
>>
>>
>> "Sylvain Lafontaine" <sylvain aei ca (fill the blanks, no spam please)>
>> wrote in message news:OOInqg5XIHA.3940@TK2MSFTNGP05.phx.gbl...
>>> Excerpt for some exception, this is the right thing to do. The finally
>>> portion of a try block has been called; this mean that an error occurred
>>> earlier and that the object is now in an unknown, unstable or unsafe
>>> state and should not be used anymore. Every object who has the IDispose
>>> interface should be disposed of in the finally portion of a try block
>>> and the Dataset object is no exception to this rule.
>>>
>>> Of course, there are always some exceptions to any rule; for example
>>> when you know that some portions of the object are still valid and that
>>> you want to continue to operate on these portions but these are
>>> exceptions and you should take some serious precautions when dealing
>>> with these.
>>>
>>> In your case, the end programmer who might want to call the
>>> DataAdapter.Update() is making an error because he's trying to operate
>>> on an object for which there has been an error and therefore is in an
>>> unknown and unstable state. Even if the Dispose() method would have not
>>> been called on this object; trying to update the database through this
>>> object could lead to some serious damages to the database; because
>>> essentially you are shooting in the dark.
>>>
>>> --
>>> Sylvain Lafontaine, ing.
>>> MVP - Technologies Virtual-PC
>>> E-mail: sylvain aei ca (fill the blanks, no spam please)
>>>
>>>
>>> "Carlo Razzeto" <crazzeto@hotmail.com> wrote in message
>>> news:0A76CA20-B5AF-4330-9EE1-6EC4D14620FC@microsoft.com...
>>>> Hey,
>>>>
>>>> Some one gave a new developer of ours a project to make sure our core
>>>> library funcitons are properly cleaning up resources (that's a good
>>>> thing!), but I had consernse about one thing the developer was doing
>>>> which sparked a debate with my co-worker.
>>>>
>>>> Some of these functions were returning DataSet objects, and the
>>>> developer was calling the Dispose() function on the datasets inside the
>>>> finally portion of a try block. For me, this kind of raised a red flag,
>>>> because it seems to me the funciton is then returning an object which
>>>> it has already "cleaned up". In some cases an "end programmer" might
>>>> even want to take this dataset, modify datarow values and call a
>>>> DataAdapter.Update() on it. It just doesn't seem like a very safe
>>>> practise to me.
>>>>
>>>> Am I being overly cautious here? I guess I just have a problem with
>>>> returning "cleaned up" object in general, to me it seems like no good
>>>> can come from that. Thanks for your thoughts,
>>>>
>>>> Carlo
>>>
>>>
>>
>>
>