On Sep 26, 2011, at 14:00 , Michael Haberler wrote: > The current watchdog has the following issues: > > - it creates a false positive signal until it detects missing transitions. > - if disabled, it can create a false positive signal (ok-out true while no > transitions). > - the manpage is unclear wrt edge- versus level triggered meaning of > enable-in, and the precise meaning of ok-out. > > I propose to change the watchdog component as follows: > > - it should under no circumstances create a positive signal when the > signal(s) are not present. This means there will be a turn-on delay until all > signals are positively identified. > - it should immediately turn off ok-out if disabled. > - clarify the manpage. > > Frankly I think this should be fixed soon. Opinions?
I agree the manage seems wrong. For example, does setting enable-in True really force out-ok to go False? Is the pin named "out-ok" or "ok-out" (the manage uses both names). Stephen, you wrote the watchdog component. Is the current behavior what you intended? I'm always reluctant to change the behavior of existing components, for fear of breaking someone's machine that relies on the old behavior… I haven't replicated Michael's observations, but they sound wrong to me. I'd also love to see a test that demonstrates the expected behavior and use of the watchdog... -- Sebastian Kuzminsky ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
