Am working on a .Net windows application which internally uses a COM dll. Out of the following 4 ways of releasing a COM object, which is the best one?
1.
if (obj != null)
{
Marshal.ReleaseComObject(obj);
}
2.
if (obj != null)
{
Marshal.ReleaseComObject(obj);
obj=null;
}
3.if (obj != null)
{
Marshal.FinalReleaseComObject(obj);
}
4.if (obj != null)
{
Marshal.FinalReleaseComObject(obj);
obj=null;
}
Thanks for reading.
5. Let the GC do it.
Which is fine unless the COM component is holding on to significant resources you need.
If you are basing your approach on checking
obj != null
then settingobj
tonull
is a good idea, unless you are sure this code will not be called multiple times.The best way: need more details about the implementation of your application to give specific advise between use of
ReleaseComObject
andFinalReleaseComObject
, but ideally the answer would be neither.Either just let the GC deal with it as Richard suggested, or your number 2. If you call
FinalReleaseComObject
, you will destroy the runtime callable wrapper, regardless if it's found its way to other parts of your code as well. This could seriously confuse other parts of your code. So you should only callReleaseComObject
once per crossing-into-.NET-land-from-COM transition.