S
smat-consulting
Guest
Your client shouldn't worry about how the server procedure gets the data. Simply ask for the data you want form the server i.e. run only getDataTrip.p on server. The procedure that you're running on the server shouldn't have any DB references in it. It should check if DB is connected, if not connect it. Only once connected, the actual program that's looking at the DB is ran. I am old-school, so I use persistent procedures not classes. I usually run all my programs persistent, starting them off with the activate. The program that is called from the client is a small wrapper - that ensures the DB is connected, then executes an API procedure that is in one of the persistent procedures. That API procedure orchestrates the collection of the data, the preparation (if needed) and so on... That way, the programs don't need to be started over and over again, but are all sitting there waiting. The only program that needs to be restarted with every request is the wrapper that's called from the client... BTW, when using temp-tables always first empty the temp-tables in the server-run procedures, to ensure there's nothing left hanging around from a previous call. And empty them also at the end of the request, so the temp-table space is not used up with stale data...
Continue reading...
Continue reading...