另一种工作解决方案，稍微简单一点。

//Note: Due to a bug with Chrome the 'dragleave' event is fired when hovering the dropzone, then
//      we must check the mouse coordinates to be sure that the event was fired only when 
//      leaving the window.
//Facts:
//  - [Firefox/IE] e.originalEvent.clientX < 0 when the mouse is outside the window
//  - [Firefox/IE] e.originalEvent.clientY < 0 when the mouse is outside the window
//  - [Chrome/Opera] e.originalEvent.clientX == 0 when the mouse is outside the window
//  - [Chrome/Opera] e.originalEvent.clientY == 0 when the mouse is outside the window
//  - [Opera(12.14)] e.originalEvent.clientX and e.originalEvent.clientY never get
//                   zeroed if the mouse leaves the windows too quickly.
if (e.originalEvent.clientX <= 0 || e.originalEvent.clientY <= 0) {

如果你使用的是HTML5，你可以得到父类的clientRect:

let rect = document.getElementById("drag").getBoundingClientRect();

然后在parent.dragleave()中:

dragleave(e) {
    if(e.clientY < rect.top || e.clientY >= rect.bottom || e.clientX < rect.left || e.clientX >= rect.right) {
        //real leave
    }
}

这是一个jsfiddle

接受的答案在某些情况下可能有效，但如果您有多个重叠的子元素，它将很快崩溃。

根据您的用例，本文的顶部答案可能更简单、更清晰

当拖动子元素时，父元素的'dragleave'会触发

   private onDragLeave(e) {
     e.preventDefault();
     e.stopPropagation();

    if (this.isEventFromChild(e)) {
      return;
    }

    this.isDragOver = false;
  }
  
  private isEventFromChild(e): boolean {
    //relatedTarget is null on Safari, so we take it from coordinates
    const relatedTarget = e.relatedTarget || document.elementFromPoint(e.clientX, e.clientY);
    return e.currentTarget.contains(relatedTarget);
  }

下面是Chrome的解决方案:

.bind('dragleave', function(event) {
                    var rect = this.getBoundingClientRect();
                    var getXY = function getCursorPosition(event) {
                        var x, y;

                        if (typeof event.clientX === 'undefined') {
                            // try touch screen
                            x = event.pageX + document.documentElement.scrollLeft;
                            y = event.pageY + document.documentElement.scrollTop;
                        } else {
                            x = event.clientX + document.body.scrollLeft + document.documentElement.scrollLeft;
                            y = event.clientY + document.body.scrollTop + document.documentElement.scrollTop;
                        }

                        return { x: x, y : y };
                    };

                    var e = getXY(event.originalEvent);

                    // Check the mouseEvent coordinates are outside of the rectangle
                    if (e.x > rect.left + rect.width - 1 || e.x < rect.left || e.y > rect.top + rect.height - 1 || e.y < rect.top) {
                        console.log('Drag is really out of area!');
                    }
                })

下面是另一种基于事件时间的方法。

从子元素分派的dragenter事件可以被父元素捕获，并且它总是发生在dragleave之前。这两个事件之间的时间间隔非常短，比任何可能的人类鼠标操作都要短。所以，这个想法是记住dragenter发生的时间，过滤dragleave事件发生“不是太快”之后…

这个简短的例子适用于Chrome和Firefox:

var node = document.getElementById('someNodeId'),
    on   = function(elem, evt, fn) { elem.addEventListener(evt, fn, false) },
    time = 0;

on(node, 'dragenter', function(e) {
    e.preventDefault();
    time = (new Date).getTime();
    // Drag start
})

on(node, 'dragleave', function(e) {
    e.preventDefault();
    if ((new Date).getTime() - time > 5) {
         // Drag end
    }
})