dotnet/wpf · error · XmlException

SR.Format(SR.InvalidXmlContent…

Error message

SR.Format(SR.InvalidXmlContent, AnnotationXmlConstants.Elements.ContentLocator)

What it means

While reading a ContentLocator element during annotation deserialization, ReadXml expects only Element nodes inside the locator. Any non-element node (text, comments, CDATA, whitespace-only content at a content position) causes XmlException(SR.InvalidXmlContent, "ContentLocator"). The reader is also expected to sit on the ContentLocator EndElement when the loop ends.

Solutions

Fix the annotation XML so ContentLocator contains only well-formed locator-part child elements.
Regenerate the annotation stream from a known-good source instead of hand-editing.
Validate the XML against the annotation schema before passing the stream to the store.

Example fix

// before
<ContentLocator>some text<Part .../></ContentLocator>
// after
<ContentLocator><MyPart xmlns="http://schemas.microsoft.com/annotations"/></ContentLocator>

Defensive patterns

Strategy: try-catch

Validate before calling

// Before loading: ensure ContentLocator elements contain only element children
foreach (var locator in doc.Descendants("ContentLocator"))
    if (locator.Nodes().Any(n => n is XText t && !string.IsNullOrWhiteSpace(t.Value)))
        throw new InvalidDataException("ContentLocator contains non-element content");

Try / catch

try { store.LoadAnnotations(stream); }
catch (XmlException ex) when (ex.Message.Contains("ContentLocator")) { /* discard or repair the malformed annotation file */ }

Prevention

Never hand-edit annotation store files; regenerate them with the app.
Validate annotation XML against the WPF annotation schema before loading.
Keep producer and consumer framework versions aligned.

When it happens

Trigger: Deserializing an annotations stream where a <ContentLocator> contains text or unexpected markup, e.g. <ContentLocator>garbage</ContentLocator> or a child that is not a locator part element.

Common situations: Hand-edited annotation store files; annotation XML generated by other tools that include whitespace/text nodes or extra elements; corrupted streams; version drift where newer annotations embed unknown content.

Related errors

AI-assisted analysis of dotnet/wpf@81131a70a4 (2026-09-14). Data as JSON: /api/errors/c8a85ed371c1a888. Report an issue: GitHub.

Appendix: source

Thrown at src/Microsoft.DotNet.Wpf/src/PresentationFramework/System/Windows/Annotations/LocatorPartList.cs:207

        {
            ArgumentNullException.ThrowIfNull(reader);

            // We expect no attributes on a "ContentLocator", 
            // so throw using the name of one of the unexpected attributes
            Annotation.CheckForNonNamespaceAttribute(reader, AnnotationXmlConstants.Elements.ContentLocator);

            if (!reader.IsEmptyElement)
            {
                reader.Read();  // Reads the start of the "ContentLocator" element

                // ContentLocatorParts cannot write themselves out (see above).  They could read
                // themselves in but instead of having write code in one place and read 
                // code somewhere else - we keep it together in this class.
                while (!(AnnotationXmlConstants.Elements.ContentLocator == reader.LocalName && XmlNodeType.EndElement == reader.NodeType))
                {
                    if (XmlNodeType.Element != reader.NodeType)
                    {
                        throw new XmlException(SR.Format(SR.InvalidXmlContent, AnnotationXmlConstants.Elements.ContentLocator));
                    }

                    ContentLocatorPart part = new ContentLocatorPart(new XmlQualifiedName(reader.LocalName, reader.NamespaceURI));
                    
                    // Read each of the Item elements within the ContentLocatorPart
                    if (!reader.IsEmptyElement)
                    {
                        // We expect no attributes on a locator part tag, 
                        // so throw using the name of one of the unexpected attributes
                        Annotation.CheckForNonNamespaceAttribute(reader, part.PartType.Name);

                        reader.Read(); // Read the start of the locator part tag

                        while (!(XmlNodeType.EndElement == reader.NodeType && part.PartType.Name == reader.LocalName))
                        {
                            if (AnnotationXmlConstants.Elements.Item == reader.LocalName && reader.NamespaceURI == AnnotationXmlConstants.Namespaces.CoreSchemaNamespace)
                            {
                                string name = null;

View on GitHub (pinned to 81131a70a4)