The site has around 1000 nodes, so Save & Publish on each node is not an option :)

You could republish every node programmatically?

public void RepublishAllNodes()
{
    Node rootNode = new Node(-1);
    foreach (Document document in GetDescendantNodes(rootNode).Select(x => new Document(x.Id)))
    {
        SaveAndPublish(document);
    }
}

/// <summary>
/// Returns the child nodes of this node
/// </summary>
/// <param name="node">The node</param>
/// <returns>The child nodes</returns>
public static IEnumerable<Node> GetChildNodes(INode node)
{
    if (node == null)
        throw new ArgumentNullException("node");

    return node.ChildrenAsList.Cast<Node>();
}

/// <summary>
/// Returns all descendant nodes
/// </summary>
/// <param name="node">The node</param>
/// <param name="includeCurrent">Include the current node in the result</param>
/// <returns>All descendant nodes</returns>
public static IEnumerable<Node> GetDescendantNodes(INode node, bool includeCurrent = false)
{
    if (node == null)
        throw new ArgumentNullException("node");

    if (includeCurrent)
        yield return (Node)node;
    foreach (Node child in GetChildNodes(node))
    {
        yield return child;
        foreach (Node subchild in GetDescendantNodes(child))
        {
            yield return subchild;
        }
    }
}

/// <summary>
/// Saves and publishes a document
/// </summary>
/// <param name="document">The document to save and publish</param>
public static void SaveAndPublish(Document document)
{
    if (document != null)
    {
        document.Save();
        document.Publish(document.User);
        umbraco.library.UpdateDocumentCache(document.Id);
    }
}

Hi Sebastian,

I'm not sure I follow completely - MNTP only stores ID's - so you're very likely to have a macro of some sorts actually generating the actual links, right?

Obviously you haven't got any hardcoded domains in there (since Save & Publish works) but could it maybe be a caching issue?

/Chriztian

-- Just ignore this reply ;-) --

@Chriz:  All the XSLT Macros have 'Cache By Page' enabled.  But that's by default.

Yeah - I'm still wondering - the only thing I can see would be that the MNTP references a Widget type or similar that's using the URLPicker data-type - that one stores a URL along with the id, and if you're using the URL directly (instead of using NiceUrl() with the node-id), this would be the exact scenario you'd encounter.

Alas, please ignore me if that's not the case :-)

/Chriztian

Did you add a  testweb.whatever.dk  as a domain in the back office when you where testing? Thats the only reson i can think of 

Noooooo ... I marked my own post as the solving answer ... not very karma like.  #h5is

Ha - cool!

Use the urlpicker-helper.xslt as well, and you'll never have that problem again :-)

(For the record: _MultiPickerHelper.xslt just delegates to your own templates for the Widgets :-)

/Chriztian

is working on a reply...