To answer my own question as far as I understand... There will be a control thread (er, event thread) that will be running perpetually -- but the easy model doesn't have a "worker thread" working all the time -- it only runs periodically when the RUNNING event goes off. I mean, you have to code the event handler to do the work at that point. That seems a bit clumsy if you want your worker to be running all the time, but I suppose you could play IPC or simulated-fork games.
My other problem was that I was trying to use the single-callback routine model, and for whatever reason, I had MUCH better luck with the multi-callback-routine model. I can't prove anything is broken, but if someone else had problems, I would try that change. On Jan 20, 11:14 am, chappaquasu...@yahoo.com (ChapSuite) wrote: > I've read the Win32::Daemon doc once or twice, and started playing > with some of the calls. But there's one thing that troubles me -- > does this whole module expect your service to only run > intermittently? I'd like my program to run continuously with its own > sleeps. > > The first path I've selected is to have one handler for all the > callbacks. It seems that the case for SERVICE_CONTROL_RUNNING should > turn around and fork > the main program, but that seems a bit clunky. > > Thanks!
