I wrote something that I'm sure is absolutely terrible performance-wise.

The idea was to use a regex to find any hints in the RTE editor for dynamic values. An example would be {{name}}. For each match my code would (first) reflect back on the object to see if any properties match this value, and (last) check the content for any properties.

This is just theory and hasn't been tested in a production website yet. Hasn't been thoroughly tested in development either. There are some things I would want to change:

• Obviously the code is pretty incompetent. I'm sure there's a better way to do everything I did.
• I'd prefer if Parse didn't require any arguments (the currently assigned content item) but I couldn't find a way to do dependency injection in my static extension class. The currently assigned content item could instead be the default.
• You could override the default by providing your own IPublishedContent to the Parse method.
• Prioritize the Umbraco content property first, and as a last resort reflect on the C# object. Just for performance reasons.
• Extend other properties (at least strings) so we can parse dynamic values in them as well.

Here is RteParser.cs:

/*
 * Ref: https://github.com/umbraco/Umbraco-CMS/blob/5bfab13dc5a268714aad2426a2b68ab5561a6407/src/Umbraco.Core/Templates/HtmlLocalLinkParser.cs
 * Ref: https://stackoverflow.com/questions/36977442/how-to-to-transform-a-string-of-property-names-into-an-objects-property-values
 */
using System.Text.RegularExpressions;
using Umbraco.Cms.Core.Models.PublishedContent;
using Umbraco.Cms.Core.Strings;
using System.Reflection;
using Umbraco.Cms.Web.Common;

namespace Humble.Umbraco.Parsers
{
public static class IHtmlEncodedStringExtensions
{
    private static readonly Regex ReplacePattern = new Regex(@"{{([a-zA-Z]+)}}", RegexOptions.IgnoreCase | RegexOptions.IgnorePatternWhitespace);

    public static IHtmlEncodedString Parse(this IHtmlEncodedString Html, IPublishedContent Content)
    {
        // Exit: string contents are null or empty
        if (Html == null) return Html;

        // Exit: no current asigned content item
        if (Content == null) return Html;

        // Find matches
        MatchCollection matches = ReplacePattern.Matches(Html.ToString());

        // Exit: no matches
        if (matches.Count.Equals(0)) return Html;

        // Build new string from model data
        var type = Content.GetType();
        string newContents = Html.ToString();

        // Begin replacing values
        foreach (Match match in matches)
        {

            // Check object properties first
            string key = match.Groups[1].Value;
            PropertyInfo typeProperty = type.GetProperty(key, BindingFlags.IgnoreCase | BindingFlags.Instance | BindingFlags.Public);

            if (typeProperty != null)
            {
                newContents = newContents.Replace(match.Value, typeProperty.GetValue(Content, null).ToString());
                continue;
            }

            // Check content properties last
            var contentProperty = Content.GetProperty(key);
            if (contentProperty == null) continue;

            object value = contentProperty.GetValue();
            if (value == null) continue;

            newContents = newContents.Replace(match.Value, value.ToString());
        }

        // Return replaced values
        return new HtmlEncodedString(newContents);
    }
}
}

Feel free to update the namespace. Then add the reference to your _ViewImports.cshtml file. Here's mine:

@using Umbraco.Extensions
@using HumbleUmbraco
@using Umbraco.Cms.Web.Common.PublishedModels
@using Umbraco.Cms.Web.Common.Views
@using Umbraco.Cms.Core.Models.Blocks
@using Humble.Umbraco.Parsers
@addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers

This is how I referenced it via a strongly typed model from a razor partial file:

@inherits UmbracoViewPage<BlockListItem>
@{
if (Model == null)
{
    return;
}

var content = (Humble_Element_RichText)Model.Content;
var settings = Model.Settings;
}

@Html.Raw(content.Copy.Parse(Umbraco.AssignedContentItem))

I am not a programmer so I understand maybe 15 per cent of what you've written : )

I understand the problem, I think.

Can't we take advantage on the engine that's already there?

Because some code takes advantage of the translation engine, looking for the given page I believe this is the

/{localLink:umb://document/ec975190c0314832a16668e02e0916c5}

code. So when you input it, it links to the proper page, in the currently viewed language. So the mechanics is ready, I guess?

Can't we, based on what was found, just "load" this page's meta title and just input it where title=" " is and before the closing of the HTML link?

This, I feel, wouldn't need extensive querying, just take more data based on localLink? Because it definitely takes the "href" part... can't it take what's in the meta title and input it in two places?

is working on a reply...