Re: [Libvirt-cim] [PATCH] Update CS and VSMS for compare_ref changes

KR> I realized I already ack'ed the cu_compare_ref(), but here I KR> realized that you're returning prop from cu_compare_ref(). prop KR> is only meaningful when it's NULL - that is, in the case of an KR> error. Otherwise, you're just returning the last prop in the KR> list. Would a status value be more useful? No, we're returning the first property name that failed the test, NULL if they all passed. This, as in the case of this patch, lets the provider do what it wishes for reporting. Right now, I put that in the status message we send back to the client. I want a rich error message potential without having to pass a broker pointer around, which is why I did it this way. If you use this function to compare a reference to an instance for something other than a GetInstance scenario, you might be interested in just the name of the property that failed, and not just a CMPIStatus you could return to the client. -- Dan Smith IBM Linux Technology Center Open Hypervisor Team email: danms(a)us.ibm.com