Released: Mar 24, 2025

v2.29 での更新項目

機能

• XMP extension types are now scoped to their schema, so XMP.getType() method returns only global types. Added a new XMP.Schema.getType() method.
• Added support for EA-PDF 1.0 specification to XMP.
• Added support for creating a PDFImage from 16-bit TIFF images.
• Added the Measurement class, for real-world dimensions on images and pages, and support creation of Geospatial PDF from GeoTIFF images (requires GeoTools).
• Added PDFStyle.superscriptClone(size, position) to force a particular size for super/subscript across various fonts.
• Added AcrobatSignatureHandler.setLoginToAuthProvider() for Amazon CloudHSM.
• Added PDFAction.fromPath() to create actions from PDF URL fragments.
• Added PDFAction.getEmbeddedFile() to retrieved EmbeddedFile from actions.
• Improvements to real-number precision when reading/writing numbers.
• Added support for multiple annotations to be specified in beginTag() methods.
• Added support to wrap output underlining in an Artifact to make tagging cleaner.

不具合の修正

• Fixed an issue with computing an OutputProfile from a PDF which was not checking inactive appearances on annotations, meaning some features were potentially missed.
• Fixes to generated RelaxNG schema from XMP objects, and improvements to XMP serialization.
• Manually setting "TextRise" in the PDFStyle or calling PDFStyle.subscriptClone or superscriptClone (which does the same) could result in incorrect layout, and also incorrectly changed the line-height calculations in LayoutBox. This is now fixed, which may result in different (but more regular) spacing between lines that make use of this setting.
• Fixed image rescaling with OutputProfiler.SimpleImageAction.
• The putUserData/getUserData methods on various objects (PDF, PDFPage etc) will no longer add a prefix to the property names that already have one.
• "Separation mode" rendering of overprint-in-stroke-but-not-fill now fixed.
• Fixed a "Separation mode" rendering issue with a particular tiling paint.
• Corrected handling of zero length lines when rendering.
• Fixed some edge-cases relating to Artifacts in PDF/UA-1 and UA-2 validation.
• Fixed exception when profiling a PDF that had been cloned from an original containing invalid annotations.
• Removed DocumentPart.getXMP(), an erratum fixed in the spec has removed this structure from the PDF.
• Viewer: Fixed the JavaScript setting of colors.
• Bugfixes to PDF/A-1, PDF/A-2 and PDF/A-3 OutputProfiles:
  • Zero-length strings with a byte-order-mark were not recognising the BOM.
  • Some TrueType font tests were incorrectly skipped for unembedded fonts.
  • Language codes in PDF/A-1 are now verified against RFC1766, not RFC3066.
  • Feature.NonSymbolicTrueTypeWithoutCMap now matches Acrobat's behaviour.
  • Feature.ICCProfileBadColorComponentCount wasn't tested on output-intent.
  • Correct apply PDF/A-1 rules for subset fonts with zero-length names.
  • Corrected extension schema definition for crs:ToneCurve in PDF/A-1.
  • Filters on non-root XMP streams were not being correctly ignored.
  • XMP dates without timezones disallowed in PDF/A-2 and PDF/A-3.
  • XMP dates without timezones now normalized to UTC not localtime.