lldialog with multiple simlutaneous users

I I got a script where its storing a bunch of strings that are displayed by the dialog box. I'm storing them in a list. Since there's more strings than buttons allowed I have it set up for next and previous buttons to pull up a new dialog box. Something I didn't think about is when two or more people end up using it at once. I tried with two avatars and sometimes it works as its supposed to be, sometimes someone gets a dialog box with bizarre options. I'm 95% sure that it has something to do with the key used in the lldialog function. As it stands right now llDetectedKey(0) is called each time that the object gets touched and then its used from then on. Is this the wrong way to go about it?

I'll post my code when I get home if it'll hellp.

detected key 0 may be fine since the event gets trgged around .1secs

what you will want to do is open independant listens for each user you pop a dialog for... you can use a timer to clear them all after a timeout by changing states to and from an empty state

The script I just posted in

/54/94/235155/1.html

MIGHT help. It is a paged dialog that SHOULD work for multiple simultaneous users, though it hasn't been tested with more than one.

The only way to allow this and to prevent different users from interfering with each other's selections by accident is to track their data separately, whether it means storing their last visited page, separating them by channel, etc.

I'd personally track them in a list of limited length, primarily because opening up a potentially unlimited number of listens can cause errors and slow down your script a good deal.

Right now I have a global integer listener, When the dialog box is created I'm using a line:

listener = llListen(channel,"","","";

So what you're saying is that I change listener into a list and add the avatar's name into the llListen call?

I'll assume the dialog is being called from the touch event? and you call the dialog using something like
llDialog( llDetectedKey( 0 ), message, [buttons], channel );


after that instead of having a single listen declared elswhere call
llListen( channel, "", llDetectedKey, "" );

and a Timeout
llSetTimerEvent( 45.0 );

in your timer call
llSetTimerEvent( 0.0 );
state sListenKill;

in a new state
state sListenKill{
state_entry(){
state default;
}
}

the timeout waits x seconds before triggering, if a new timeout is specified before it's triggered, it waits that amount from the new trigger time... basically delaying it indefinately until it's not in use for at least 45 seconds which will switch states, clearing all listens, then pop back ready to use

Not necessarily. Check out the script in the thread I linked to above. It doesn't record any state to keep track of the page a resident was on. Instead it encodes the numbers of the previous and next pages in the strings used for the buttons. That and the UUID of the speaker can give enough information for multiple simultaneous users in many cases (when the strings themselves are not dynamically changed based on user choice).

The easiest way to do this is to rename your "next" and "previous" buttons with the page number they go to (for example, "Page 2", "Page 3", etc. Then the script can just display the page it's asked for. You can track sessions internally, but it's a pain to do and can create a hole for griefers.

the dtnamic dialog function I wrote for the library does this, more buttons with page numbers, handles multiple users using the above outlined method, seach for my name dynamic, and library