6 The OLE-Server Library

6.6 Implementing in-process servers

If you want to build your server application as an in-process server -- a server that runs within the client application's process -- there are a few differences from what you would do for a local server, and a few additional considerations.

You must define a subclass of <ole-in-process-server>, page 272, instead of <ole-server-framework>.

: The <ole-in-process-server> class is itself a subclass of <ole-server-framework>, so in-process servers inherit all methods documented for <ole-server-framework>.

You must build your in-process server application as a .DLL file instead of as a .EXE file.

: Since your server application is a dynamic link library instead of a complete program, there will be no main program. The event loop will be provided by the container application.

You can handle persistent storage differently from the way local server storage is handled. Another technique is described in Section 6.7.9 on page 327.
Drawing is handled a little differently. Instead of calling OLE-part-draw-metafile, the function OLE-part-draw, page 331, is called. This function takes one additional parameter, which is a rectangle specifying the size and position within which to draw.
Since there is no main program loop in an in-process server, self-registration and creation of a class factory are handled by code generated by the macro initialize-ole-server, page 326.

Overall, an in-process server application will look something like this:

define constant $my-class-ID = make-GUID(...);

define COM-interface <my-object-class> (<ole-in-process-server>)
  ...
end;

... // methods on <my-object-class>

initialize-ole-server(<my-object-class>, 
                      $my-class-ID, 
                      "my.prog.id",
                      "title of my application");