From: Juan C. R. F. <jc...@ca...> - 2002-04-02 10:34:02
|
Hi my friends The promissed schematics: actual_dialog async_dialog synchro_dialog I send the files in two formats: png and dia (http://www.lysator.liu.se/~alla/dia/) Before seeing the drawings you must consider that the ULT will be an asynchronous thread. Why?, well an example will clarify it: First you will have a thread for each telescope. If you send a "GOTO" command to de telescope, this one will be put in movement, but only the associated XY service will have cpu time until the movement will be completed making it impossible to abort a command sent by mistake while the service does not release the CPU. With a thread per ULT you can abort the previous ULT or you can send commands to other services or instruments in parallel. Of course, will be a contention mechanism to access concurrently to the services. The client apis probably remains without changes, in the service's case I'll modify the handler_output to manage the answer and the handler_input will remain to handle the unsolicited messages, only if the instrument has this capacity, of course!. Well, I'll wait your comments. cheers Juan Carlos P.D. For fausto I have included the corrected spiderd.h file. tonight i'll put it in cvs (See attached file: actual_dialog.dia)(See attached file: actual_dialog.png)(See attached file: async_dialog.dia)(See attached file: async_dialog.png)(See attached file: synchro_dialog.dia)(See attached file: synchro_dialog.png)(See attached file: spider.h) |