Re: proper way of handling [out,retval] VARIANT * args

Your code is correct. The pointer must not be NULL of course,
and the location it points to is presumed to be uninitialized, thus
VariantInit. Here VariantClear is a clear invitation for a crash.

--
=====================================
Alexander Nickolov
Microsoft MVP [VC], MCSD
email: agnickolov@xxxxxxxx
MVP VC FAQ: http://www.mvps.org/vcfaq
=====================================

"Jason S" <jmsachs@xxxxxxxxx> wrote in message
news:1156349994.825654.193260@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Oh wise masters of COM:

what's the proper way to handle an [out,retval] VARIANT *arg ?
Do you use VariantInit() after checking for a null arg, or
VariantClear?

STDMETHODIMP CBlah::get_something(VARIANT *pVal)
{
if (pVal == NULL)
return E_POINTER;

VariantInit(pVal);

// do some stuff, it may return an error

// everything is great, assign the VARIANT
CComBSTR b = L"something from blah";
V_VT(pVal) = VT_BSTR;
V_BSTR(pVal) = b.Detach();
return S_OK;
}



.