Timeout on Certain Event

[ASTDan]

Hello,

I have an idea/is this possible question.

When using an event structures is it posible to timeout if a certain event didn't happen. I know right now I can time out if no events for the enire structure didn't occured in x amount of time. I was wondering if I can time out only if a specific event didn't occur i.e. user event.

Thanks

Dan

[crelf]

QUOTE (ASTDan @ Apr 19 2008, 09:17 AM)

When using an event structures is it posible to timeout if a certain event didn't happen. I know right now I can time out if no events for the enire structure didn't occured in x amount of time. I was wondering if I can time out only if a specific event didn't occur i.e. user event.

I'm trying to understand the use case here - you want to timeout only if a certain user event doesn't happen, but don't timeout if other user events don't happen, is that right? How's that different from globally timing out? If your event doesn't happen then it'll go to the timeout case. Are you looking for a "IF event occurs THEN do case one ELSE do case two" pattern? The global timeout will fire if your event didn't occur, and the other cases will fire if their events did occur. Maybe I'm missing your point. Can you give us an example?

[Ton Plomp]

QUOTE (crelf @ Apr 19 2008, 05:23 PM)

I'm trying to understand the use case here - you want to timeout only if a certain user event doesn't happen, but don't timeout if other user events don't happen, is that right?

I think Dan wants a timeout not influenced by other events.

This can be done by placing just two cases inside an event structure, one for the timeout and one for the event. Upon a 'event-specific timeout' you can trigger a user event in the main event structure. If the event doesn't timeout (eg. just happens) you could handle the code in the 'single event-event structure' or trigger another user event.

Ton

[crelf]

QUOTE (tcplomp @ Apr 19 2008, 11:39 AM)

I think Dan wants a timeout not influenced by other events.

Ahhh - so basically two parallel event structures?

[ASTDan]

My use case is I have a Custom User event I created. The user event is my device tells me I have data. When my device tells me I have data I fire my Custom User event. What I want is a timeout on this custom user event. Basicaly if I don't see data from my device in say 1 minute, ping the device to see if it is still there running. I don't want my timout reset by somebody clicking away on my user interface.

Does that make sense?

QUOTE (tcplomp @ Apr 19 2008, 11:39 AM)

This can be done by placing just two cases inside an event structure, one for the timeout and one for the event. Upon a 'event-specific timeout' you can trigger a user event in the main event structure. If the event doesn't timeout (eg. just happens) you could handle the code in the 'single event-event structure' or trigger another user event.

Could you post a quick example of this?

Thanks

Dan

[Ton Plomp]

QUOTE (ASTDan @ Apr 19 2008, 11:05 PM)

My use case is I have a Custom User event I created.

Could you post a quick example of this?

Here's the easy route:

And the hard route:

QUOTE (ASTDan @ Apr 19 2008, 11:05 PM)

My use case is I have a Custom User event I created. The user event is my device tells me I have data. When my device tells me I have data I fire my Custom User event. What I want is a timeout on this custom user event. Basicaly if I don't see data from my device in say 1 minute, ping the device to see if it is still there running. I don't want my timout reset by somebody clicking away on my user interface.

I think you should handle this in the place you get your data, not in the consumer of that place.

I would setup a design where my communicator waits for data and than sends data to the user event. If the timeout passes I ping my device. (a little bit like my second design from above).

And yes two event structures in one VI/application is absolutly valid!

Ton

[ASTDan]

Thanks for the idea of the 2 event loops.

I feel so naughty using 2 event loops in one app.

Dan

[Yair]

QUOTE (ASTDan @ Apr 20 2008, 05:45 PM)

I feel so naughty using 2 event loops in one app.

There's absolutely nothing wrong with that. In fact, the two event structures can even handle the same events without race conditions (although in that case you should be aware of the consequences, especially if you use filter events). You might have been thinking of two event structures in the same loop.

The only real thing you should avoid when doing this is wiring the same event registration refnum into two different event structures because that will cause you to lose events. As you can see, Ton used two event registration nodes to avoid this.

[Ton Plomp]

QUOTE (ASTDan @ Apr 20 2008, 04:45 PM)

Thanks for the idea of the 2 event loops.

Looking in retrospect I think the first solution is the easiest (no extra user events).

QUOTE (Yen @ Apr 20 2008, 08:27 PM)

The only real thing you should avoid when doing this is wiring the same event registration refnum into two different event structures because that will cause you to lose events. As you can see, Ton used two event registration nodes to avoid this.

To understand a little bit what Yen is talking about.

Ton

[ASTDan]

QUOTE (tcplomp @ Apr 21 2008, 02:00 AM)

To http://forums.ni.com/ni/board/message?board.id=170&view=by_date_ascending&message.id=247057#M247057' target="_blank">understand a little bit what Yen is talking about.

I read that very good thread and I have an idea. Is it posible to dynamicly register the timeout event? That way I can turn it on and off as needed.

Thanks

Dan

[Yair]

QUOTE (ASTDan @ Apr 21 2008, 02:47 PM)

Is it posible to dynamicly register the timeout event?

Looks like you can't. If you wire an application reference to the register node, the timeout event is disabled.

What you can do instead is dynamically change the timeout value (e.g. using a shift register). When you don't want it to occur, use a value of -1.

[ASTDan]

That's right!! I forgot about the -1 shift register trick.

Thanks for all the ideas!

Dan

[crelf]

QUOTE (ASTDan @ Apr 21 2008, 02:18 PM)

I forgot about the -1 shift register trick.

Ahhhhhhh - sorry, I should have mentioned that in my initial posts - that's why I was confused: I couldn't understand why you'd use the -1 timeout value.