Hyland.DocumentFilters.Linux-arm64 26.1.0

Learn more about Target Frameworks and .NET Standard.

• • Hyland.DocumentFilters (>= 26.1.0)

# Enhancements

- Introduced HD support for DICOM JPEG SOF3 images with three color channels, enabling successful conversion and rendering of these files. (DF-2501)
- Introduced support for Markdown (text/HD) output when using the feature-limited Document Filters trial without a license. This enables AI and ML workflows to leverage Markdown output formats during evaluation, improving accessibility and utility for trial users. (DF-2725)
- Introduced word-level splitting in the MDAST generation pipeline when the `JSON_INCLUDE_WORD_LOCATIONS` option is enabled. This enhancement ensures text content is divided into individual word nodes, improving accuracy for downstream processing and enabling precise word location tracking for advanced data extraction and AI workflows. (DF-2693)
- Switched JPEG support from libjpeg 9e to libjpeg-turbo 3.1.3. (DF-2495)

# Updates

- HD Mode: Resolved a condition for DOCX files where tables with repeat header rows could be rendered incorrectly, with content missing from rows. This ensures the output matches the original document’s structure and content, improving fidelity for users converting complex Word files. (DF-2579)
- HD Mode: Resolved a condition for DOCX files where text formatting properties could be handled incorrectly. This ensures accurate conversion and extraction of all text and formatting from Claude AI-generated Word documents, supporting increased customer use of AI tools. (DF-2735)
- HD Mode: Resolved a condition for DOCX files where text with a default run-style containing the `vanish` property was omitted, even though Microsoft Word would display it. This ensures all intended text is preserved in output, improving document fidelity for downstream processing. (DF-2431)
- HD Mode: Resolved a condition for DOCX files where textboxes could be rendered with incorrect position and size, and text within textboxes could have incorrect wrapping. (DF-2610)
- HD Mode: Resolved a condition for HTML files containing RSMF content where phone numbers within angle brackets (e.g., <123456789>) were omitted from PDF and HiDef output due to being misinterpreted as HTML tags. This ensures that numeric content inside angle brackets is now correctly included in the output, improving data fidelity for downstream workflows. (DF-2380)
- HD Mode: Resolved a condition for HTML files where content using CSS `unset` and `revert` keywords could be rendered overlapping with other content. (DF-2611)
- HD Mode: Resolved a condition for HTML files where images could be missing from rendered output when the HTML was processed via stream or byte array. (DF-2366)
- HD Mode: Resolved a condition for Microsoft Office files where opening the file for processing could fail with a "SEH: Access violation" error if a numbering level definition was missing in the file model. This ensures affected documents can now be processed without error, improving reliability for HD conversions. (DF-2734)
- HD Mode: Resolved a condition for SVG files with embedded base64-encoded PNG images where the image was not rendered, resulting in a blank output. This ensures that embedded images in SVGs are now correctly displayed in HiDef output, improving fidelity for documents containing complex graphics. (DF-2726)
- Resolved a condition for DOCX files where OCR text was not properly extracted from shapes and drawings when the option `OCR_INLINE_IMAGES` was enabled. (DF-2681)
- Resolved a condition for DOCX files where URLs were incorrectly extracted with the fragment identifier placed at the beginning when the option `TEXT_INLINE_LINKS` was enabled. URL extraction now preserves the correct order, ensuring accurate hyperlink representation for downstream processing and data integrity. (DF-2279)
- Resolved a condition for DOCX files where footnote citations were missing from Markdown and text-mode outputs. Footnotes are now collected and emitted at the end of each section in the output, ensuring important citation content is preserved for downstream workflows and improved document fidelity. (DF-2553)
- Resolved a condition for Excel files where the rightmost column could appear blank in HTML output due to incorrect CSS positioning during conversion. This ensures all columns are retained and visible, preserving data integrity for downstream workflows. (DF-2736)
- Resolved a condition for HTML-based (HTML, MSG, EML) files where malformed or whitespace-padded ending tags would lead to truncated output content. This ensures more complete and accurate extraction of data for downstream processing. (DF-2675)
- Resolved a condition for Linux where running `./isys_doc2text --version` did not display the correct version information. This ensures consistent and accurate version reporting across platforms. (DF-2340)
- Resolved a condition for Microsoft Office 2007+ (DOCX, PPTX, XLSX) files where embedded images and date variables could be missing or incorrect because of incorrect path resolution. (DF-2679)
- Resolved a condition for OLM files where the `x-opfmessagecopysenttime` and `x-opfmessagecopyreceivedtime` header fields were missing from extracted EMLs. These headers are now included when present in the source file, ensuring accurate date values for downstream processing and improved fidelity for email extraction workflows. (DF-2624)
- Resolved a condition for OpenDocument Text (ODT) files where headers and footers were not extracted in text output mode, even when `DOCUMENT_HEADERS=ON` and `DOCUMENT_FOOTERS=ON` (default: ON) were set. This ensures that header and footer content is now included in text exports, improving completeness and fidelity for downstream processing and review. (DF-2626)
- Resolved a condition for PDF to image conversion where high-resolution image output quality was lower in version 25.4 compared to 25.3. This ensures that generated images maintain expected fidelity when using options such as `GRAPHIC_DPI=300` and `GRAPHIC_WIDTH=4000`, benefiting workflows that require high-quality image exports. (DF-2727)
- Resolved a condition for Python bindings where iterating over image objects in document pages failed with `TypeError: 'Iterator' object is not iterable` when using Python 3.13. This update restores compatibility with Python 3.13, ensuring continued reliability and support of the latest Python versions. (DF-2490)
- Resolved a condition for large ZIP files where file identification could take several minutes or longer, especially with files over 10 GB. This enhancement improves processing speed for large archives, ensuring faster identification and better usability for workflows involving sizable ZIP files. (DF-2745)
- Resolved a condition for text input with email-like headers "To:" and "From:" where extracted text could be missing content. (DF-2444)
- Resolved a condition for the Java `RenderPageProperties` API where several properties and functions—such as `SourceRect`, `DestRect`, and `AddRedaction`—were missing compared to the C# and Python implementations. The Java version now achieves feature parity, enabling consistent rendering capabilities and reducing maintenance overhead for teams supporting multiple SDKs. This ensures a more uniform developer experience and improved functionality for Java users. (DF-2765)
- Resolved a condition where EML files starting with ARC headers could be misidentified as ASCII files. (DF-2515)
- Resolved a condition where justified text in input DOCX files was rendered to HTML with extraneous <div> elements containing only spaces and non-breaking spaces. (DF-2646)
- Security: libjpeg-turbo:  CVE-2026-24797: not-affected: This vulnerability is associated with a benchmark tool that is not present in Document Filters. (DF-2746)
- Security: libpng: CVE-2025-28162: patched: Applied security patch to address this issue. (DF-2748)
- Security: libpng: CVE-2025-28164: patched: Applied security patch to address this issue. (DF-2747)
- Security: libpng: CVE-2025-64505: patched: Applied security patch to address this issue. (DF-2720)
- Security: libpng: CVE-2025-64506: patched: Applied security patch to address this issue. (DF-2719)
- Security: libpng: CVE-2025-64720: patched: Applied security patch to address this issue. (DF-2718)
- Security: libpng: CVE-2025-65018: patched: Applied security patch to address this issue. (DF-2717)
- Security: libpng: CVE-2025-66293: patched: Applied security patch to address this issue. (DF-2716)
- Security: libpng: CVE-2026-22801: patched: Applied security patch to address this issue. (DF-2740)
- Security: libpng: CVE-2026-25646: patched: Applied security patch to address this issue. (DF-2773)
- Security: openssl: CVE-2025-66199: not-affected: OpenSSL 1.0.2 is explicitly listed as not affected by this CVE. (DF-2756)
- Security: protobuf: CVE-2026-0994: not-affected: Vulnerability affects the Python package, which is not used by Document Filters. (DF-2749)
- Security: protobuf: CVE-2026-23833: not-affected: Document Filters does not bundle ESPHome or its API component. (DF-2742)
- Security: snappy: CVE-2025-65942: not-affected: This CVE affects VictoriaMetrics. VictoriaMetrics is not present in Document Filters. (DF-2722)
- Security: snappy: CVE-2025-67721: not-affected: Aircompressor is not present in Document Filters. (DF-2730)
- Security: tesseract: CVE-2019-25257: not-affected: The vulnerability is in LogicalDOC's configuration of the `ocr.Tesseract.path` parameter, not in Tesseract itself. Document Filters uses Tesseract, not LogicalDOC Enterprise. (DF-2731)
- Security: tesseract: CVE-2025-60176: not-affected: Document Filters does not use the Tesseract WordPress plugin. (DF-2723)
- Security: unrar: CVE-2026-24857: patched: Applied security patch to address this issue. (DF-2750)
- Security: zlib: CVE-2023-53742: not-affected: The issue is related to the Linux kernel and is not a vulnerability of the zlib library itself. (DF-2732)
- Security: zlib: CVE-2026-22184: not-affected: The vulnerable utility is absent from Document Filters. (DF-2741)
- Security: zlib: CVE-2026-24793: not-affected: Document Filters uses a version of zlib which includes a patch for this issue. (DF-2755)
- Security: zlib: CVE-2026-24799: not-affected: Document Filters uses a version of zlib which includes a patch for this issue. (DF-2754)
- Security: zlib: CVE-2026-24800: not-affected: Document Filters uses a version of zlib which includes a patch for this issue. (DF-2753)
- Security: zlib: CVE-2026-24811: not-affected: Vulnerability affects versions of zlib that did not receive the CVE-2016-9841 upstream fix, which has already been applied to Document Filters. (DF-2752)
- Security: zlib: CVE-2026-24812: patched: Applied security patch to address this issue. (DF-2751)