To use a string binding, a client starts with either an existing string binding or with the components of the binding information. Do not hardcode string bindings into application code. Rather, specify them at runtime using a command argument, environment variable, file, or other means. The simplest way to specify a string binding is for a user to supply a string binding manually to a client. However, this manual approach is awkward for users who must know how to obtain and manipulate the string bindings. Also, if binding information changes, the users are responsible for updating any string bindings used by their clients. Reducing manual intervention in the use of string bindings requires that an application provide its own mechanisms for storing, maintaining, and accessing binding information. In contrast, a directory service such as CDS provides these mechanisms automatically to applications that store binding information in a namespace.
Regardless of how a client obtains a string binding, before establishing a binding, the client must ask the RPC runtime for a binding handle that refers to the server binding information depicted in the string binding. The client converts the string binding into a server binding handle by calling the rpc_binding_from_string_binding( ) routine.
The following pseudocode lists the calls for composing a string binding and for using it to obtain a server binding handle:
rpc_string_binding_compose(...);
rpc_binding_from_string_binding(...);
.
.
.
rpc_string_free(...);