|This of course cannot correct for the reset problem the original poster
|mentioned, however I have never had that concern (my resets tend to cause
|cold boots clearing my RAM and starting scratch). This assumes the increment
|code (BumpMyTmr) is atomic, at least as compared to the ReadMyTmr code. I
|would be interested in seeing the original posters code, testing for the
|various problems (including the reset) can be very tricky.

If I recall, the procedure to normalize a stage was something
like this:

normalize:
        btfss   stagebelow,7
         goto   clearbelow
setbelow:
        movf    thisstage,w
        iorlw   1
        btfsc stagebelow,7
         movwf  thisstage
        goto    done
clearbelow:
        incf    thisstage,w
        andlw   254
        btfss stagebelow,7
         movwf  thisstage
done:

I'd have to spend awhile checking all the different cases to
make sure I got everything right.  Does anyone else know, off
hand, whether the "btfsc movwf" sequence is guaranteed to be
atomic (i.e. that there would be no possibility of an interrupt
changing /stagebelow/ and yet still having the movwf happen)?


Attachment converted: wonderland:WINMAIL.DAT 1 (????/----) (000234CE)