Updating an XML File in the 14 Hive Using a Custom Timer Job

Applies To: SharePoint 2010, .NET Framework (C#, VB.NET)

As mentioned in a previous post, I’ve recently put together a solution for automatically configuring your SharePoint servers to use the Adobe PDF icon for PDF files. You can download the solution as well as the source for free from CodePlex here: WireBear PDFdocIcon. I’m going to show some of the code as it currently exists below, but be sure to check out the CodePlex site to ensure you have the latest version.

I’ve also provided the bulk of the code and some explanation for installing/uninstalling a custom job from a SharePoint solution in my last post: Implementing a Custom SharePoint Timer Job. In this post we’ll explore what’s actually happening in the execution of the timer job.

The goal is to update the DOCICON.xml file in the 14\TEMPLATE\XML folder within the SharePoint 2010 Hive to include or remove a mapping entry for a specific file extension. Here is the entire DocIconJob class:

The Code:

Imports Microsoft.SharePoint.Administration
Imports System.IO
Imports Microsoft.SharePoint.Utilities
Imports System.Xml

Public Class DocIconJob
    Inherits SPServiceJobDefinition

#Region "Properties"

    Private _dociconPath As String
    Public ReadOnly Property DocIconPath() As String
        Get
            If String.IsNullOrEmpty(_dociconPath) Then _dociconPath = SPUtility.GetGenericSetupPath("TEMPLATE\XML\DOCICON.XML")
            Return _dociconPath
        End Get
    End Property

    Private Const InstallingKey As String = "DocIconJob_InstallingKey"
    Private Property _installing() As Boolean
        Get
            If Properties.ContainsKey(InstallingKey) Then
                Return Convert.ToBoolean(Properties(InstallingKey))
            Else
                Return True
            End If
        End Get
        Set(ByVal value As Boolean)
            If Properties.ContainsKey(InstallingKey) Then
                Properties(InstallingKey) = value.ToString
            Else
                Properties.Add(InstallingKey, value.ToString)
            End If
        End Set
    End Property

    Private Const FileExtensionKey As String = "DocIconJob_FileExtensionKey"
    Private Property _fileExtension() As String
        Get
            If Properties.ContainsKey(FileExtensionKey) Then
                Return Convert.ToString(Properties(FileExtensionKey))
            Else
                Return String.Empty
            End If
        End Get
        Set(ByVal value As String)
            If Properties.ContainsKey(FileExtensionKey) Then
                Properties(FileExtensionKey) = value
            Else
                Properties.Add(FileExtensionKey, value)
            End If
        End Set
    End Property

    Private Const ImageFilenameKey As String = "DocIconJob_ImageFilenameKey"
    Private Property _imageFilename() As String
        Get
            If Properties.ContainsKey(ImageFilenameKey) Then
                Return Convert.ToString(Properties(ImageFilenameKey))
            Else
                Return String.Empty
            End If
        End Get
        Set(ByVal value As String)
            If Properties.ContainsKey(ImageFilenameKey) Then
                Properties(ImageFilenameKey) = value
            Else
                Properties.Add(ImageFilenameKey, value)
            End If
        End Set
    End Property

#End Region

    Public Sub New()
        MyBase.New()
    End Sub

    Public Sub New(JobName As String, service As SPService, Installing As Boolean, FileExtension As String, ImageFilename As String)
        MyBase.New(JobName, service)
        _installing = Installing
        _fileExtension = FileExtension
        _imageFilename = ImageFilename
    End Sub

    Public Overrides Sub Execute(jobState As Microsoft.SharePoint.Administration.SPJobState)
        UpdateDocIcon()
    End Sub

    Private Sub UpdateDocIcon()
        Dim x As New XmlDocument
        x.Load(DocIconPath)

        Dim mapNode As XmlNode = x.SelectSingleNode(String.Format("DocIcons/ByExtension/Mapping[@Key='{0}']", _fileExtension))

        If _installing Then
            'Create DocIcon entry
            If mapNode Is Nothing Then
                'Create Attributes
                Dim keyAttribute As XmlAttribute = x.CreateAttribute("Key")
                keyAttribute.Value = _fileExtension
                Dim valueAttribute As XmlAttribute = x.CreateAttribute("Value")
                valueAttribute.Value = _imageFilename

                'Create Node
                mapNode = x.CreateElement("Mapping")
                mapNode.Attributes.Append(keyAttribute)
                mapNode.Attributes.Append(valueAttribute)

                Dim byExtensionNode = x.SelectSingleNode("DocIcons/ByExtension")
                Dim NodeAdded As Boolean = False
                If byExtensionNode IsNot Nothing Then
                    'Add in alphabetic order
                    For Each mapping As XmlNode In byExtensionNode.ChildNodes
                        If mapping.Attributes("Key").Value.CompareTo(_fileExtension) > 0 Then
                            byExtensionNode.InsertBefore(mapNode, mapping)
                            NodeAdded = True
                            Exit For
                        End If
                    Next

                    If Not NodeAdded Then byExtensionNode.AppendChild(mapNode)
                    x.Save(DocIconPath)
                End If
            End If
        Else
            'Remove DocIcon entry
            If mapNode IsNot Nothing Then
                Dim byExtensionNode = x.SelectSingleNode("DocIcons/ByExtension")
                If byExtensionNode IsNot Nothing Then
                    byExtensionNode.RemoveChild(mapNode)
                    x.Save(DocIconPath)
                End If
            End If
        End If
    End Sub

End Class

What’s Going On:

Lines 9-73 are just the declaration of and logic needed to persist some properties. Again more information can be found in my last post, but basically I am using the SPJobDefinition’s Properties HashTable to store my own properties as specified in the constructor. Except for in the case of the DocIconPath property which is really just wrapping up some logic to get a reference to the 14 Hive’s TEMPLATE\XML directory using the SPUtility class.

The Execute method beginning in line 86 is what is called when the Timer Job actually runs. I override this method to ensure my custom code gets called instead. My custom code really begins in the UpdateDocIcon method starting at line 90.

In lines 91-94, I load the DOCICON.xml file into and XmlDocument object and attempt to find the mapping node that applies to the appropriate file extension (In this case it’s going to be pdf).

If this job is installing (Running on Solution Activation), then I just check to see if the node was found. If so, all done! If not, then it’s time to add it. I create the node and setup it’s attributes in lines 100-108 using standard objects from the System.Xml namespace.

In order to work, the mapping node needs to be added as a child of the ByExtension element, so we find that in line 110. By default the mapping nodes are listed in alphabetical order by their extension. Since I’m anal, I use a method in lines 114-120 presented by Steve Goodyear to ensure I insert the mapping node in it’s proper position. Failing that, I add it to the end in Line 122 and save the file in line 123.

If this job is uninstalling (Running on Solution Deactivation) and the mapping exists, we delete it and save the file in lines 128-134.

Isn’t that Super Exciting?!?! Hopefully this example will help make the concepts I was talking about in my previous post make some sense. If not, then sadness will fill my soul and flowers will no longer bloom or something.

Automatically Setting Up PDF Icon Mapping in SharePoint 2010

Applies To: SharePoint 2010

Nearly everyone who has ever used SharePoint has had to setup the PDF icon mapping so that PDF documents will have the familiar Adobe logo rather than the blank, unknown icon SharePoint uses by default. This is relatively simple and there are guides to do doing this all over the internet. (Microsoft’s can be found here).

Here is a very brief summay of the steps that must be performed manually on every server:

  1. Copy the PDF icon picture from Adobe and put it in your 14 Hive (TEMPLATE\IMAGES)
  2. Edit the DOCICON.xml file in your 14 Hive (TEMPLATE\XML) to add a Mapping element for pdf documents pointing to your new icon
  3. Reset IIS

These aren’t super complicated steps but there are some pretty big problems (or at least irritations) with using this approach:

  • Manual changes can often be error-prone, especially for those not familiar with XML
  • The change must be performed on every server
  • The change must be performed whenever a new server is added to the farm
  • The change will have to be redone in the event of disaster recovery

So, like many before me, I thought, surely this can all be automated! So I looked and I found some solutions for SharePoint 2007 and several solutions that only worked for Standalone Servers or for only one server in the farm. These were of help, but still no good for my needs. So, I wrote my own.

You can find it over on CodePlex as WireBear PDFdocIcon. There’s some stuff about it’s license over there (Free for personal and commercial use, etc.) and the basic installation instructions. It’s super easy to setup since it’s just a standard SharePoint Solution that you globally deploy.

The full source code is available on CodePlex, but I’ll be going in depth about how it works over the next few posts. But to summarize, here’s what happens:

  • The Adobe PDF icon file is copied to the 14\TEMPLATE\IMAGES folder using standard resource deployment
  • On Activation and Deactivation a one time Service Timer Job is run.
  • On Activation, the Timer job searches for a mapping for PDF documents within the 14\TEMPLATE\XML\DOCICON.xml file. If not found, it adds one (in alphabetic order) and points it to the icon file
  • An IIS Reset is performed to get the changes activated
  • When Deactivating, the Timer job removes the mapping for PDF documents
So why use this thing?
  • The changes will be reapplied in the event of Disaster Recovery
  • The changes will be applied to new servers as they are added to your farm
  • You don’t have to personally edit the 14 Hive on every server in your farm
  • It makes a special place in your heart of hearts that keeps the beast at bay

In making this, I came across several blog entries that were especially helpful, here are most of these (Thanks!):

I’ve found this to be a helpful approach and I hope you do too.

PDF Search Results Direct Link (Eliminating DispForm.aspx Results) Without an iFilter

Applies To: SharePoint

We are utilizing our search functionality much more in SharePoint and one of the more annoying things we found was how PDF files are treated by default. In the search results, the link goes to the DispForm.aspx for the item rather than directly to the item.

The obvious fix is to install an iFilter. Unfortunately, this isn’t always an option. For us, the performance and crawl delay issues didn’t make up for the benefit of having these documents indexed. Fortunately, I came across this answer by daver306 on SharePoint SE that didn’t get a lot of attention but worked perfectly for me.

I wanted to write it up with some added detail and share my experiences. Not only does this allow you to link directly to your PDFs within the search without the use of XSL and allows KnowledgeLake queries to open PDFs directly within the KnowledgeLake Viewer, it’s actually pretty simple to do.

1. Add PDF as a File Type

Within Central Admin, go to your Search Administration (Manage Service Applications > Search Service). From there click on the File Types link under Crawling on the left:

If pdf is not listed, click the New File Type button and type pdf (no period needed) in the File extension box and click OK:

2. Restart the Search Service

This is a very important step. I originally tried to skip it to spare myself some hassle and ended up having to repeat the crawl below. You will need to go to each server running the SharePoint Server Search service and stop it. You can do this through the command line or the Services panel under Administrative Tools:

Once off on all boxes, just go back through and start it again.

3. Reset Your Index

Back on the Search Administration page within Central Administration you will want to click on the Index Reset link under Crawling on the menu on the left:

Press the Reset Now button. Remember that this should be done at a time when your environment is not under heavy use or when search won’t be needed since search results will not be available until after a full crawl completes.

4. Perform a Full Crawl

If you have a pretty standard search setup, then you probably only have one content source. If not, then you already know how to start the full crawls for each of them. If you’ve just got the one, then from the Search Administration page within Central Administration click on the Content Sources link under Crawling on the menu on the left. Hover over your content source and choose Start Full Crawl in the dropdown menu:

After the crawl completes (This could be hours depending on the size of your farm), things should be working as expected. No more DispForm.aspx links in your search results!