On Wed, 28 Apr 1999 [EMAIL PROTECTED] wrote:
> James Einolf wrote:
> > Call it what you will. Ignoring it (and us) will not make it go away.
> While I certainly understand your frustration with this problem, I think
> its probably safe to say that we already know this. And furthermore, that
> we (the engineers at least) aren't ignoring you. I mean, you were
> responding to Steve's message, right?
>
> As far as I know, no one at Palm has ever seen this. I'm not saying that
> it therefore doesn't exist, but it does make it difficult to debug. I'd be
> happy to put some time into it (my own time even) if I could get a
> reproducible case.
>
> For what its worth (which is probably not very much until I get a device
> that exhibits the problem), I have a couple of ideas:
> 1) The RTC interrupt is disabled when we go to sleep for some reason.
I forget how the '328 handles the real-time-clock. Does it need an
interrupt to keep it going, or will it continue when the processor is
asleep? It occurs to me that if some of the RTC stuff is not being
activated, the unit might show a constantly drifting clock (as it wouldn't
keep time while it was asleep in this particular mode.)
No guarantees, but it might be useful to keep an eye out for this
symptom -- and obviously anything that narrows it down will help.
> 2) Maybe this reason is that the battery is low, or a global or something
> related to the battery subsystem gets smashed.
Hmm. If it really does only happen on the V (is that true? James?) then
battery issues do seem a possibility. Does anyone happen to have any
information about the battery level or usage habits of the folks who have
reported the problem. (I doubt it, just thought I'd ask.)
> So, at least you know we've been thinking about the problem, even if we
> haven't been especially vocal.
I suppose I should point out one of the complaints that various company's
tech support groups have garnered over the years is "a black hole",
meaning that reports go in, and nothing ever comes out except rare tech
info statements or patches, neither of which are released with any
fanfare. (I'm thinking of Borland in particular, to be honest.)
Palm devsup is far from a black hole (_very_ far), but the same problem
still creeps up every now and then, I expect due to the usual problems of
clearing stuff with Legal before making an Official Announcement. Perhaps
some sort of facility (a publically searchable bug-base?) would help folks
to verify that devsupp does know about a particular problem, and perhaps
also help coordinate with techsup.
--
Kenneth Albanowski ([EMAIL PROTECTED], CIS: 70705,126)
