GroupDocs.Merger for .NET 26.10 Release Notes

There are 10 features, improvements, and bug fixes in this release.

This release puts the loading of external resources under your control: images, stylesheets and other files a document references from outside itself are no longer loaded by default, and you can allow trusted locations with a whitelist or decide for each resource with a handler. It also fixes EPUB and TeX documents failing when processed from several threads, EPUB documents that failed to open or reported twice their page count, auto-tagged PDF output that lost its format and title, .NET Framework projects that failed to load an assembly during image and SVG joins, and the output path and joined-document checks of the LowCode products, and moves to a patched System.Security.Cryptography.Pkcs.

Full list of changes in this release

Major Features

• Control of external resources: Remote images, stylesheets, linked pictures and other files a document references are no longer loaded unless you allow them. LoadOptions has new SkipExternalResources, WhitelistedResources and ExternalResourceHandler properties to allow trusted locations, allow everything for documents you trust, or decide, skip or supply each resource in your own code.

• Concurrency fixes: EPUB documents no longer fail intermittently with NullReferenceException or IndexOutOfRangeException, and TeX documents with ArgumentException, when they are processed from several threads.

• EPUB documents: Fixed-layout EPUB documents now have one page per book page, so page N is the book’s page N. EPUB documents whose table of contents file has another extension than .ncx no longer fail with NullReferenceException.

• Auto-tagged PDF output: A PDF document saved with auto-tagging keeps the format and the title of the source. It is no longer marked as PDF/UA-1 unless the source was, and an untitled source no longer gets the title “Tagged PDF”.

• .NET Framework projects: Image joins in packages.config projects no longer fail to load System.ValueTuple, and joins with SVG documents no longer fail to load netstandard on .NET Framework 4.6.2 to 4.7.0 runtimes.

• Patched dependency: The packages depend on System.Security.Cryptography.Pkcs 8.0.2 (.NET 6.0 and .NET 8.0) or 9.0.21 (.NET 10.0), the versions that fix CVE-2026-84008 and CVE-2026-70162.

• LowCode products: An output path given as a bare file name is saved to the current directory, and JoinPdf, JoinDocx and JoinPptx reject a joined document of another format before any output is written.

Public API and backward incompatible changes

1. External resources — denied by default (new API, backward incompatible)

New public types in the GroupDocs.Merger.Domain.Options namespace:

New properties of LoadOptions:

• SkipExternalResources — true by default in every constructor. While true, no external resource is loaded except those matching WhitelistedResources or allowed by the handler.
• WhitelistedResources — absolute URI prefixes of resources that are loaded while SkipExternalResources is true, for example https://cdn.example.com/assets/ or file:///C:/assets/. Empty by default.
• ExternalResourceHandler — null by default.

Backward incompatible changes:

• External resources are no longer loaded by default. Remote resources (http, https, ftp and other schemes), UNC paths, local files, including files in the document’s own folder, and relative references that point outside the document are skipped, and the operation continues. Earlier versions loaded them. Remote images and stylesheets of HTML, Word and presentation documents can therefore be missing from the output, and the page count of HTML and Word documents can change. Content the document carries — data: URIs, cid: references, MHTML parts and packaged media — still loads. To load external resources, whitelist their locations, or set SkipExternalResources to false for documents you trust.
• LoadOptions is now sealed. A class derived from it no longer compiles. Use LoadOptions directly, or implement ILoadOptions; a custom ILoadOptions implementation that does not also implement IResourceLoadingOptions gets the default policy.
• A WhitelistedResources entry that is null, empty, not an absolute URI, carries a query or fragment, or contains %2F, %5C or %2E makes the Merger constructor throw ArgumentException (“Whitelisted resource ‘…’ must be an absolute URI without a query, fragment or encoded ‘/’, ‘' or ‘.’.”).
• If the handler throws, the resource is skipped and the operation throws GroupDocsMergerException (“The external resource handler threw an exception; the resource was skipped. See the inner exception.”) with the handler’s exception as InnerException.
• EPUB documents, and SVG and SVGZ documents in some operations, now have their markup checked for external references. A document whose markup is not well-formed XML fails with GroupDocsMergerException (“The document’s markup could not be read to check it for external resources, so it was not loaded. Set SkipExternalResources to false without an ExternalResourceHandler to load it without the check.”).
• The OleDiagramOptions, OleSpreadsheetOptions and OleWordProcessingOptions constructors that take a file path and a page number render the embedded object’s image with external resources denied, whatever the LoadOptions of the target document allow. To show them, pass your own image data to the constructor that takes it.
• The products of the GroupDocs.Merger.LowCode package always deny external resources.

The options are read when the Merger is created, and documents joined to it follow its options.

using System.Collections.Generic;
using GroupDocs.Merger;
using GroupDocs.Merger.Domain;
using GroupDocs.Merger.Domain.Options;

var loadOptions = new LoadOptions(FileType.HTML)
{
    // Load resources from one trusted folder only; everything else stays denied
    WhitelistedResources = new List<string> { "https://cdn.example.com/assets/" }
};

using (var merger = new Merger("page1.html", loadOptions))
{
    merger.Join("page2.html");
    merger.Save("merged.html");
}

using System;
using GroupDocs.Merger.Domain.Options;

public class TrustedPngHandler : IExternalResourceHandler
{
    public void OnLoading(ExternalResourceLoadingArgs args)
    {
        Uri uri;
        if (Uri.TryCreate(args.Uri, UriKind.Absolute, out uri)
            && uri.Scheme == Uri.UriSchemeHttps
            && uri.Host == "cdn.example.com"
            && uri.AbsolutePath.EndsWith(".png", StringComparison.OrdinalIgnoreCase))
        {
            args.Skipped = false;   // allow; any other resource keeps the default decision
        }
    }
}

See Skip loading of external resources for the matching rules, the handler rules and the supported formats.

2. Auto-tagged PDF documents keep the format and title of the source

No API change. A PDF document saved with PdfSaveOptions.AccesibilitySettings.EnableAutoTagging set to true keeps the PDF format and the title of the source. Previously every auto-tagged document was marked as PDF/UA-1, and an untitled source got the title “Tagged PDF”. Auto-tagging adds a logical structure tree; it does not make a document PDF/UA compliant, and a source that is already marked as PDF/UA keeps its mark.

3. Fixed-layout EPUB documents — one page per book page

No API change. A fixed-layout EPUB document now has one page per book page. Previously each book page was rendered as two pages, so GetDocumentInfo reported twice the page count and page N did not match the book’s page N. Page numbers passed to page operations such as ExtractPages, RemovePages or Split now refer to the book’s pages.

4. GroupDocs.Merger.LowCode — output path and joined documents

• JoinPdf, JoinDocx, JoinPptx and RemovePagesPdf resolve a relative output path against the current working directory. An output path given as a bare file name is now saved to the current directory; previously it was saved to Merger’s default output directory. A null or empty output path throws ArgumentNullException.
• JoinPdf, JoinDocx and JoinPptx throw FileTypeNotSupportedException when a joined document’s extension is not the product’s format, before any output is written, for example “Input format does not match the expected format ‘PDF’; the joined file ‘a.docx’ is ‘Microsoft Word Open XML Document (.docx)’.”. Previously JoinPdf and JoinDocx produced mixed output and JoinPptx failed later with another message. A null or empty joined document path throws ArgumentNullException (“The document at index 0 is null or empty.”).
• All seven products throw FileTypeNotSupportedException (“Input format does not match the expected format ‘PDF’; the supplied file is ‘Unknown’.”) for a source path without an extension, instead of ArgumentException.

5. Dependencies

• The runtime packages depend on System.Security.Cryptography.Pkcs 8.0.2 (GroupDocs.Merger.Net60, GroupDocs.Merger.Net80 and their .Windows packages) or 9.0.21 (GroupDocs.Merger.Net100 and GroupDocs.Merger.Net100.Windows). GroupDocs.Merger.Net60, GroupDocs.Merger.Net60.Windows and GroupDocs.Merger.LowCode also depend on System.Formats.Asn1 8.0.2, and GroupDocs.Merger.LowCode on System.Security.Cryptography.Pkcs 8.0.2. GroupDocs.Merger.Net462 is unchanged.
• In .NET Framework projects, the GroupDocs.Merger package no longer adds a binding redirect for System.ValueTuple to app.config or web.config (packages.config projects), and its .NET Framework build targets set DependsOnNETStandard to true unless the project sets it.

Documentation

New and updated pages for this release:

• Skip loading of external resources — new
• Network access and data privacy — new
• Features overview — control of external resources
• Loading documents — link to the external resources guide
• Installation — the MSI installer includes the -windows builds