Re: How to get CMyObject from IMyObject
Matt Houser <MattHouser@discussions.microsoft.com> wrote:
"Igor Tandetnik" wrote:
"Matt Houser" <MattHouser@discussions.microsoft.com> wrote in message
news:A195FF80-1B48-4AF7-87AE-0B8BA4777983@microsoft.com
If I have a IMyObject passed into a COM function, how can I find out
if it's a CMyObject (or one someone else created)?
You can define an internal interface (not documented anywhere) and
implement it on your object alongside IMyObject. Then simply query
for it.
Any hints about creating such a hidden interface? It would need to
be in the IDL correct?
Not necessarily. You can manually define it in a .h file. Or you can
define it in a separate internal IDL file that you don't publish.
You could also use a trick like this:
BEGIN_COM_MAP(CMyObject)
// Usual interface entries here
COM_INTERFACE_ENTRY_IID(CLSID_CMyObject, CMyObject)
END_COM_MAP()
This way, when you QueryInterface for CLSID_CMyObject you get back a raw
CMyObject pointer (AddRef'ed, so don't forget to Release it). Since
CLSID_CMyObject doesn't refer to a real interface, there's definitely no
marshalling support for it, so QI will only succeed from the same
apartment (which is what you want).
However, wouldn't this possibly run into issues if the object gets
marshalled? For example, gets created in one process and passed to
another process? The interface would be satisfied
Only if you build and register marshalling support for it. So don't. To
ensure it doesn't happen by accident, you can mark the interface with
[local] attribute (if you choose to define it in the IDL at all).
Is there any way to know if an object is from the current process or
is marshalled from the outside world?
A query for a non-marshallable interface would fail.
--
With best wishes,
Igor Tandetnik
With sufficient thrust, pigs fly just fine. However, this is not
necessarily a good idea. It is hard to be sure where they are going to
land, and it could be dangerous sitting under them as they fly
overhead. -- RFC 1925