Suspected bug in devapi_pipe

Separate issue for code smells reported in #872 (closed).

in cppapi/client/devapi_pipe.cpp line 1051, in method DevicePipeBlob & DevicePipeBlob::operator<<(DevString &datum) we use the index insert_ctr when I would expect insert_ind

			(*insert_elt_array)[insert_ind].value.string_att_value(dvsa);
		(*insert_elt_array)[insert_ctr].inner_blob_name = Tango::string_dup("Scalar");
			insert_ind = -1;

This might be on purpose, but it looks like a bug.