Event.data is a global event object

[Nate Cavanaugh]

The behavior I was expecting was that binding data to an even via .bind(type, data, fn) would store the data relevant to that specific item, rather than a global event object.

A quick example:

$(['test1', 'test2', 'test3']).each(function(){ $('#'+this).bind('change', i, selectOnChange) })

Now, when selectOnChange gets executed, event.data will only contain the latest iterator, which leads me to think that event.data is being stored globally for all onchange events, rather than on a per instance basis.

Would it be possible to store this on a per instance basis?

[Arrix]

Cause of the bug: The data is actually stored as a property of the handler function so the latest call to bind(type, data, fn) always overwrites previous data.

[joern]

Added test in [1588].

[dec]

Any update here? This is a major bug causing me major pain. It practically means that all jQuery functions making use of event.data are useless unless you're passing static data (which seems pointless anyway).

[jesse]

Patch to fix Event data from being global

[brandon]

Thanks Jesse for the patch!

Fixed in Rev. 1765.

[arrix]

The fix introduces other problems. We are unable to unbind events bound with data specified. Because when data is given, guid will only be set on the closure, not the original function.

[arrix]

unable to unbind() if data is specified for bind()

[arrix]

Rev 1765 has changed the behavior of bind(). I remember, if we call bind() with the same handler function multiple times for the same element, the handler will only be executed once. Now the handler gets invoked multiple times if data is specified for bind().

[arrix]

patch

Index: E:/zm/jquery/jquery/src/event/event.js
===================================================================
--- E:/zm/jquery/jquery/src/event/event.js	(revision 1771)
+++ E:/zm/jquery/jquery/src/event/event.js	(working copy)
@@ -32,8 +32,11 @@
 		}
 
 		// Make sure that the function being executed has a unique ID
-		if ( !handler.guid )
+		if ( !handler.guid ) {
 			handler.guid = this.guid++;
+			//don't forget to set guid for the original handler function
+			if (fn) fn.guid = handler.guid; 
+		}
 
 		// Init the element's event structure
 		if (!element.$events)
@@ -274,7 +277,7 @@
 	 */
 	bind: function( type, data, fn ) {
 		return this.each(function(){
-			jQuery.event.add( this, type, fn || data, data );
+			jQuery.event.add( this, type, fn || data, fn && data );
 		});
 	},
 	
@@ -309,7 +312,7 @@
 			jQuery.event.add( this, type, function(event) {
 				jQuery(this).unbind(event);
 				return (fn || data).apply( this, arguments);
-			}, data);
+			}, fn && data);
 		});
 	},
 

[brandon]

Fixed in Rev. 1775.

[jesse]

In my patch, I moved the "handler.guid = this.guid++;" code above the code that created wrapper handler function, then had "handler.guid = fn.guid;" which passed the guid on to the wrapper function. This didn't happen when the patch was applied, thus why the bug was introduced.

For cleanliness sake, I suggest either removing the line "handler.guid = fn.guid;", or better, just moving the block above and removing the "if (fn) fn.guid = handler.guid;", as this patch will do:

Index: event.js
===================================================================
--- event.js	(revision 1801)
+++ event.js	(working copy)
@@ -12,6 +12,10 @@
 		// around, causing it to be cloned in the process
 		if ( jQuery.browser.msie && element.setInterval != undefined )
 			element = window;
+
+		// Make sure that the function being executed has a unique ID
+		if ( !handler.guid )
+			handler.guid = this.guid++;
 		
 		// if data is passed, bind to handler 
 		if( data != undefined ) { 
@@ -31,13 +35,6 @@
 			handler.guid = fn.guid;
 		}
 
-		// Make sure that the function being executed has a unique ID
-		if ( !handler.guid ) {
-			handler.guid = this.guid++;
-			// Don't forget to set guid for the original handler function
-			if (fn) fn.guid = handler.guid;
-		}
-
 		// Init the element's event structure
 		if (!element.$events)
 			element.$events = {};

[brandon]

Good call ... thanks Jesse! Fixed in Rev [1962].