SharePoint Designer Custom Workflow Activity - WTF

I was writing a custom workflow activity for SharePoint Designer today (a pretty cool custom activity, if I do say so myself, which I'll post in a few days) and on my custom activity, one of the settable fields I have happens to be the last word in the sentence:  

So I went to go set the field and this is what I get:

Whenever I went to set the field, the hover action over the activity added a drop down button to the end, thereby making the field impossible to set. Anyways, I'm just going to change the sentence but I thought it was pretty funny and wanted to share it.

Using the jQuery UI Dialog widget for confirmation windows

In our web applications, we often have the need to confirm with the user whether or not they want to proceed with an action they attempted to take. For instance, we might have a delete button on our form that responds to a user click by deleting a record in the database. Before we actually do the delete, we want to double-check with the user first. JavaScript has a confirm(msg) function that will display a standard dialog window that you can use to determine whether or not to continue with an action. The standard dialog window looks like the following (in IE): 

The standard confirm window is not very flexible. First, you only get two buttons (and no more), Ok and Cancel, and you can't change the text of either button. Second, you can't change the title on the title bar. You also can't change the question mark icon inside of the dialog. If you need something more flexible, you're going to have to create your own modal overlay window and recreate the functionality that confirm() gives you. Fortunately, jQuery UI has a dialog widget that will get you most of the way there:

 

As you can see, this dialog window definitely doesn't look like the standard confirm. First, I have more than two buttons and with different text. I also have a different title. It can be resized and repositioned. Also, although you can't see it, this dialog window can have the same visual effects (like applying semi-transparent overlays on everything underneath the window) applied to it that can be applied to other html elements by using jQuery. This would not be possible with the standard confirm. I won't bore you with the details of how to set this up. The documentation on the jQuery site sufficiently explains that. What I want to concentrate on is showing you how to make the jQuery dialog widget behave exactly like the standard JavaScript confirm.

When using the standard confirm, all processing stops until the user clicks on either Ok or Cancel in the window. That means that if the button was supposed to post back to the server, the postback won't occur until a selection is made. Usually, we only go ahead with the postback if the Ok button was clicked. This is typically handled by adding something similar to the OnClientClick event handler on our button:

<asp:Button ID="Button1" runat="server" Text="JavaScript" OnClick="Click" OnClientClick="return confirm('Dude, are you sure?');" />

If the user clicked the Ok button, true would be returned by the confirm(msg) function. If the user selected Cancel, false would be returned. Returning false in OnClientClick will effectively cancel the rest of the event processing (i.e., the event handler for the OnClick event won't be triggered).

Now, for the jQuery dialog, this is how I have the button:

<asp:Button ID="Button2" runat="server" Text="jQuery" OnClick="Click" OnClientClick="showjQueryDialog();return false;" />

I call a client-side function showjQueryDialog() which, amazingly enough, handles showing my jQuery dialog (pre-configured in the document ready event handler). After the call to showjQueryDialog(), I go ahead and just return false. The reason why I have to return false here is because, unlike the confirm() function, the browser doesn't stop processing the rest of the client-side script just because the dialog widget is opened. So we have to manually stop it. Now this raises an issue. So if we always return false, how will our event handler for the OnClick event ever be executed? In order for that event handler to execute, we will need to do the postback ourselves. There are a couple of ways to handle this but this is the approach I take:

1. Create a hidden field on the form (called hdnBtnPostback) who's value will be the exact postback function call I need to make in order to emulate the same postback that would occur had the button processing continued. This value can be set in the Page_Load() of the ASP.NET page. But how do we know what the exact postback function call should be? Fortunately for us, that is one of the methods available from the ClientScriptManager object. So the following code will do the trick:

 

this.hdnBtnPostback.Value = Page.ClientScript.GetPostBackEventReference(Button2, string.Empty);

 

This is will generate the exact same call to the __doPostBack() JavaScript function that is generated by the button. 

2. In the event handler for my dialog's Ok button click, all I have to do then is get the value of this hidden field and pass it to the JavaScript eval() function. This will effectively execute the postback and the event handler for the OnClick event will be processed: 

    1    function showjQueryDialog() {

    2 

    3       $("#dialog").dialog("open");

    4    }

    5 

    6    //document on ready.

    7    $(function(){

    8       $("#dialog").dialog({

    9          autoOpen: false,

   10          modal:true,

   11          buttons : {

   12             "Yes" : function() {              

   13                $(this).dialog("close");

   14                eval($("#<%= hdnBtnPostback.ClientID %>").val());

   15             },

   16             "No" : function() {

   17                $(this).dialog("close");

   18             },

   19             "Maybe": function() {

   20                $(this).dialog("close");

   21                //what should we do when "Maybe" is clicked?

   22             }        

   23          }

   24       });

   25    });

The event handler for the Yes button in the dialog are on lines 12-15. Line 14 is where we actually do the postback manually. 

As you can see, it is quite easy to replace the standard confirm dialog and make it function in the same way. All that is needed is a little elbow grease :).

Creating a SharePoint breadcrumb control with drop down menus

While there have plenty of things that people don't like about Windows Vista, one of the features I do like and I think is underappreciated is the new breadcrumb:

Unfortunately in SharePoint, the breadcrumbs don't work like this. There are no submenus. Lucky for us, it's pretty easy to create a breadcrumb that functions this way. Here's how:

Creating the web control

The first part we'll need is a web control that will render the markup. The web control I created is simple and straightforward. Only two public properties are available: SiteMapProvider (string) and NodeSeparator (string). The SiteMapProvider property is used to specify the named site map provider from the web.config file to use to build the breadcrumb. The NodeSeparator property is used to specify the character(s) to use to separate each item in the breadcrumb (like the sideway triangles in the screenshot above). By default, this is set to the '>' character. Also, this property will only be used if the breadcrumb item doesn't have any sub-items; if it does, then a clickable image like the one above will be displayed instead. Lastly, the breadcrumb's submenus will only go one level deep (like Vista's). So there won't be any "submenus of submenus".

The rendering logic takes place in the RenderContents(HtmlTextWriter) method. First, we get an instance of the SiteMapProvider object that has the same name as the one specified in the control's SiteMapProvider property. Then, using the provider, we traverse our way up the site map from the current page (specified by the SiteMapProvider.CurrentNode property) until we get to the top of the site map. While we traverse up, we will collect the nodes that we passed in a Stack<SiteMapNode> collection:

            SiteMapProvider provider = GetSiteMapProvider();

 

            Stack<SiteMapNode> nodes = new Stack<SiteMapNode>();

 

            SiteMapNode current = provider.CurrentNode;

            while (current != null)

            {

                nodes.Push(current);

                current = current.ParentNode;

            }

 

Once we've reached the top, we then pop the nodes out of the stack. Each node will be represented as a list item (li) in an unordered list (ul). If the node has any subitems, then we will render a clickable image and unordered list that is a child of the list item:

            while (nodes.Count > 0)

            {

                SiteMapNode node = nodes.Pop();

                sb.AppendFormat("<li class='dp-breadcrumbitem'><a href='{0}' title='{1}'>{1}</a>", node.Url, node.Title);

 

                //why not use SiteMapNode.HasChildNodes? see: http://social.msdn.microsoft.com/Forums/en-US/sharepointdevelopment/thread/37d10f92-140f-4ce8-b71c-388163721737/

                if (node.ChildNodes.Count > 0)

                {

                    sb.Append("<img src='/_layouts/images/marr.gif' class='dp-breadcrumbitemimage'/>");

 

                    sb.AppendFormat("<ul id='dp-submenu-{0}' class='ms-topNavFlyOuts dp-breadcrumbsubmenu'>", node.Key);

                    foreach (SiteMapNode subNode in node.ChildNodes)

                    {

                        sb.AppendFormat("<li class='dp-breadcrumbsubmenuitem'><a href='{0}' title='{1}' class='dp-submenulink'>{1}</a></li>", subNode.Url, subNode.Title);

                    }

                    sb.Append("</ul>");

                }

                else

                {

                    if (nodes.Count > 0) sb.AppendFormat("<span class='dp-breadcrumbseperator'>{0}</span>", nodeSeparator);

                }

                sb.Append("</li>");

            }

 

This is the typical html markup that is rendered by the breadcrumb control:

<ul class='dp-breadcrumb'>

    <li class='dp-breadcrumbitem'><a href='link' title-'Item 1'>Item 1</a> <img src='path_to_image' />

        <ul class='ms-topNavFlyOuts dp-breadcrumbsubmenu'>

            <li class='dp-breadcrumbsubmenuitem'><a href='link' title='Sub Item 1' class='dp-submenulink'>Sub Item 1</a></li>

            <li class='dp-breadcrumbsubmenuitem'><a href='link' title='Sub Item 2' class='dp-submenulink'>Sub Item 2</a></li>

        </ul>

    </li>

    <li class='dp-breadcrumbitem'><a href='link' title-'Item 2'>Item 2</a> <img src='path_to_image' />

        <ul class='ms-topNavFlyOuts dp-breadcrumbsubmenu'>

            <li class='dp-breadcrumbsubmenuitem'><a href='link' title='Sub Item 3' class='dp-submenulink'>Sub Item 3</a></li>

            <li class='dp-breadcrumbsubmenuitem'><a href='link' title='Sub Item 4' class='dp-submenulink'>Sub Item 4</a></li>

        </ul>

    </li>

</ul>

 

That is pretty much it for the web control. The next parts we need to build are the CSS and the JavaScript.

CSS and JavaScript

I won't bore you with the details of the CSS since it's available in the zip file below but the most important things with the CSS are: making sure the breadcrumb items are displayed inline and making sure the submenu uses absolute positioning and has a z-index that will place it on top of any other element that is going to be on the page.

Now for the UI magic to happen, we need a little client-side code. Here again, I use my new favorite client-side library, jQuery. The client script primarily consists of three event handlers: an event handler for the click event on the image, an event handler for a submenu item's hover event, and an event handler used to handle clicks anywhere else on the document.

The click event for the image is used to show/hide the corresponding submenu. It will also make sure that no other menu is visible besides the one that corresponds to the image that was clicked. Finally, it will toggle an appropriate image to use, depending on whether or not the submenu is visible or hidden:

    //add the event handler for the click on the image

    $("img.dp-breadcrumbitemimage").click(function(e){   

        e.stopPropagation();

 

        var theImage = $(this);

 

        var left = this.offsetLeft + 12;

        var top = this.offsetTop + 12;

 

        //get the submenu corresponding to the image.           

        var submenu = theImage.next("ul.dp-breadcrumbsubmenu");

 

        //iterate over all the submenus in this breadcrumb and hide any that isn't the target submenu.

        $("ul.dp-breadcrumbsubmenu").each(function(idx){           

            if (this.id != submenu.attr("id")){

                $(this).hide();

                $(this).prev("img.dp-breadcrumbitemimage").attr({src:"/_layouts/images/marr.gif"});

            }

        });

 

        //if the target submenu is visible, hide it. if it's invisible, show it.

        //also change the image that is being displayed.

        if(submenu.is(":visible"))

        {

            theImage.attr({src : "/_layouts/images/marr.gif"});

            submenu.slideUp();

        }

        else

        {

            theImage.attr({src : "/_layouts/images/menu2.gif"});

            submenu.css("position", "absolute").css("top", top).css("left", left).slideDown();

        }       

    });

The hover event handler for each submenu item will take care of making sure the item is highlighted properly. The highlighting we will use is the same highlighting used by the top navigation menu:

    //add a hover event for the items in the submenu so that they are highlighted.

    $("li.dp-breadcrumbsubmenuitem").hover(function(){

        $(this).addClass("ms-topNavFlyOutsHover");

    },

    function(){

        $(this).removeClass("ms-topNavFlyOutsHover");   

    });

Lastly, the click event handler for the document will make sure that if the user clicks outside of the breadcrumb, any visible submenu will be hidden:

    //add a click handler for the entire page so that when the user clicks outside of the breadcrumb, any visible menu will be hidden.

    $(this).click(function(){

        $("ul.dp-breadcrumbsubmenu").slideUp();

        $("img.dp-breadcrumbitemimage").attr({src:"/_layouts/images/marr.gif"});

    });

 

Putting it all together

So now that we've built the components, it's time to put it all together. First, add the breadcrumb.css file and breadcrumb.js file in a document library or folder in the SharePoint site collection. You also need to place the jQuery javascript library in a document library/folder (note: I used jQuery version 1.2.6 for this).

Second, you need to deploy the assembly 'DeviantPoint.SharePoint.Web.UI.dll' to the Global Assembly Cache (GAC) and then reset IIS.

Third, in order to be able to use the web control we developed, you need to add an entry into the SafeControls section of your web.config file:

      <SafeControlAssembly="DeviantPoint.SharePoint.Web.UI, Version=1.0.0.0, Culture=neutral, PublicKeyToken=d767c3b2d2145e42"Namespace="DeviantPoint.SharePoint.Web.UI.WebControls.Navigation"TypeName="*"Safe="True" />

Lastly, you need to modify your site collection's master file by first registering our custom assembly:

<%@ Register Tagprefix="DeviantPoint" Namespace="DeviantPoint.SharePoint.Web.UI.WebControls.Navigation" Assembly="DeviantPoint.SharePoint.Web.UI, Version=1.0.0.0, Culture=neutral, PublicKeyToken=d767c3b2d2145e42" %>

Then you need to add a reference to the css and javascript files we are using:

    <!-- reference to the breadcrumb css -->

    <link href="/Style Library/Breadcrumb.css" type="text/css" rel="stylesheet"/>

 

    <!-- add the reference to the two jquery libraries -->

    <script src="/scripts/jquery-1.2.6.min.js" type="text/javascript" language="javascript"></script>

    <script src="/scripts/Breadcrumb.js" type="text/javascript" language="javascript"></script>   

Note that in this case, I put 'Breadcrumb.css' in the default Style Library document library. I put the javascript files in a folder I created with SharePoint designer called scripts. Also, make sure to place these lines right before the closing 'head' tag in the master file.

Lastly, we need to place the breadcrumb control in an appropriate spot on the master page where we'd like the breadcrumb to appear:

              <asp:ContentPlaceHolder id="PlaceHolderGlobalNavigationSiteMap" runat="server">

                <!-- DEVIANTPOINT breadcrumb -->

                <DeviantPoint:Breadcrumb ID="bcGlobal" runat="server" SiteMapProvider="GlobalNavSiteMapProvider"></DeviantPoint:Breadcrumb>

              </asp:ContentPlaceHolder>

In this case, I actually replaced the default global breadcrumb that comes out of the box with my own breadcrumb. I also used the GlobalNavSiteMapProvider because that site map provider will have all of the nodes for the site collection.

Results

Here are some screenshots of the new breadcrumb in action:

 

 

See? Not too difficult!

Files

In the zip file below, you'll find the assembly that needs to be deployed to the GAC, Breadcrumb.cs, breadcrumb.css, breadcrumb.js, and deviantpoint.master. Deviantpoint.master is just the out of the box master with all the changes to the master file I specified above. You will need to download jQuery v. 1.2.6 yourself.

DeviantPoint_Breadcrumb.zip (11.12 kb)