G'Day All,

Dennis: Thanks for your tips on values in the spot filters.
Bill W7OO: Are you ready! This is what you need until Dennis comes up with a better idea.

I have been mucking about today to see if I can write a "private function" to use for alerts and the spot log - to surpress repeated alerts - and try to discover what else might be possible.

I am assuming the spot log will be small, and my PC is fast, and will multi-task with Logic. I need to do more work with UTC time. My goals were to gain experience with a real problem, and have fun.

Who is willing to give it a go - do some testing to see that it works for them? Reply to this topic or email me.

The code returns FALSE if a specific spot has been seen in the last "spot_delay" seconds and is called in an alert filter by, for example ...

.T. AND check_alert( CALL, BAND, MODE, 600, 3000 ) AND <filter expression>

... the leading .T. allows one to "turn the alert off completely", check_alert surpresses repeated alerts by returning .F., and <filter> is the actual alert. The result puts 600 seconds between alerts and flashes a message box for 3 seconds when a spot is surpressed. The message box will go when it is proven to work correctly.

I started with this simple code ...

	 LPARAMETERS spot_call, spot_band, spot_mode, spot_delay, box_delay
	 MESSAGEBOX(spot_call,36,"Alert",1)
         RETURN .T.

... which I compiled in Advanced, Database commands and it worked as expected. Then I added the scan through the spots for the call, to return FALSE if a spot log record is found in the last "spot_delay" seconds. Much reading of manuals and the Internet was included.

I added an index to the spot log for "CALL+DTOC(TIMESTAMP)" to facilitate the scan of the spots.

The following code works for me, surpressing alerts for calls for "spot_delay" seconds. UTC time needs to correctly implemented.

	LPARAMETERS spot_call, spot_band, spot_mode, spot_delay, box_delay

	* Spot time stamps are UTC
	* UTC midnight is 36000 seconds after local midnight
	* need to get utc time from the system and stop using the SECONDS() function.

	utc = 36000

	spot_alias = alias()
	if not used("my_spots") then
		use spots again shared noupdate order call_time alias my_spots in 0
	endif

	select my_spots

	status = .T.

	scan for ;
		my_spots.call = spot_call and ;
		my_spots.band = spot_band and ;
		my_spots.mode = spot_mode and ;
		between((seconds() - utc - ((hour(my_spots.timestamp)*60) + minute(my_spots.timestamp))*60),0,spot_delay)

		status = .F.

	endscan

	IF status = .F.
		MESSAGEBOX(spot_call + " " + transform(status),36,"Alert",box_delay)
	ENDIF

	IF EMPTY( spot_alias )

		SELECT 0
	ELSE
		SELECT ( spot_alias )

	ENDIF

	return status

Dennis: Assistance with getting UTC time from Foxpro would be appreciated.

If you would rather I did not do this - just say so. No worries.
If I am "biting off more than I can chew" - just say so. No worries.

But, if a few hints will get me there - thanks in advanced.

My priority: Low - but others may have a higher priority.

Peter VK4IU

Hi Dennis,

Is there a possibility that you could work something out with Peter to include this code in your next release of LOGic 9? It is a significant addition to the filtering tools we already have.

Thanks,

Bill
W7OO

G'Day,

I have been experimenting with alerts.

A better arrangement with the filter is to put my "alert checker" last. Foxpro will abandon further effort if any one term of a "logical expression" is false.

So like this, the check for duplicates occurs no matter what the call sign is

check_filter(call,band,mode,600,0) and call='K9W'

A better check is

call='K9W' and check_filter(call,band,mode,600,0)

The call to check_filter will only occur if the call is actually K9W.

This is clearly more efficient.