#8894 fix explanation

gistfile1.js

var fragment, div, clonedNode, firstP, secondP, clonedP, dbi, clonedBdi,
    node = document.createElement("div");

    // true in all browsers
    console.log( node.parentNode == null );

    // but after
    node.getElementsByTagName();

    // true only in ie < 9
    console.log( node.parentNode != null && node.parentNode.nodeType === 11 );

    // different example –
    node = document.createElement("div");
    fragment = document.createDocumentFragment();

    fragment.appendChild( node );

    // true in all browsers
    console.log( node.parentNode != null );
    console.log( node.parentNode.nodeType === 11 );

    // but after
    clonedNode = node.cloneNode( true );

    // true only in ie < 9
    console.log( clonedNode.parentNode != null );
    console.log( clonedNode.parentNode && clonedNode.parentNode.nodeType === 11 );

    // true in all browsers
    console.log( $( "<div />" ).get( 0 ).parentNode == null );

    // true only in ie < 9
    console.log( $( "<div />" ).clone().get( 0 ).parentNode != null );

    // depending on how many elements (one, or more than one) in inserted collection and
    // what type is parentNode of the element, appendTo and alike methods will use different code path,
    // without jQuery.fn.pushStack, in first case and with jQuery.fn.pushStack, in second

    firstP = $("<p/>"); // created without fragments – parentNode == null
    secondP = $("<p>test</p>"); // created with fragments – parentNode.nodeType === 11

    console.log( firstP.appendTo("<div/>").end().length ); // 1
    console.log( secondP.appendTo("<div/>").end().length ); // 0

    // bad behavior, but consistent in all browsers

    // but now amount of inserted nodes is two, so second code path will be chosen in both cases
    console.log( firstP.appendTo("<div/><div/>").end().length ); // 1
    console.log( secondP.appendTo("<div/><div/>").end().length ); // 1

    // but if
    clonedP = secondP.clone();

    // true – in "normal" browsers, and false – in ie < 9
    console.log( clonedP.get( 0 ).parentNode == null );

    // nodeType === 11 in ie < 9, null in all others
    console.log( clonedP.get( 0 ).parentNode && clonedP.get( 0 ).parentNode.nodeType );

    // a bit more complicated example
    div = $("<div></div>");

    // true in all browser
    console.log( div.get( 0 ).parentNode == null );

    div.find("p");

    // now false in ie < 9
    console.log( div.get( 0 ).parentNode == null );

    // so this is why this will be true in ie < 9 and false in all other browsers
    console.log( clonedP.addClass("test").appendTo("<div/>").end().hasClass("test") === false );
    console.log( $("<div/>").find("p").end().addClass("test").appendTo("<div/>").end().hasClass("test") === false );
    console.log( $("<div/>").text("test").addClass("test").appendTo("<div/>").end().hasClass("test") === false );
    // in "normal" browsers parentNode after jQuery.clone or jQuery.fn.find or jQuery.fn.text, will be null,
    // so second path will apply, in ie < 9 parentNode will be documentFragment, so first code path will apply

    // and there is a html5 elements...
    bdi = $( "<bdi />" );

    // true in all browsers
    console.log( bdi[ 0 ].parentNode == null );

    clonedBdi = bdi.clone();

    // null in all browsers and nodeType === 1, in ie < 9
    console.log( clonedBdi.append("test").get(0).parentNode && clonedBdi.append("test").get(0).parentNode.nodeType );

    // because of it, this statement will be false
    console.log( clonedBdi.addClass("test").appendTo("<div/>").end().hasClass("test") === false );

    // code path should be the same for elements –
    // 1. created with fragments
    // 2. created with disconnected nodes
    // 3. created without fragments or disconnected nodes
    // 4. created with fragments or with disconnected nodes and manipulated after

    // appendTo and alike methods called on disconnected nodes
    // should follow the first path, regardless of type of parentNode