Released: Jul 4, 2022

v2.26.5 での更新項目

機能

• The Jars are now signed with SHA256 rather than SHA1.
• Added the ability to allow signatures, that have just been signed, to be queried via various "get" methods.
• Improved the pre-flight process for large embedded fonts, e.g. those used for Chinese, to stop the widths array growing too large for PDF/A.
• Improved the workaround for when 2-byte fonts are used in Form fields, which reduces memory footprint for large forms.
• PDFCanvas.beginTag() will now accept a "ref" attribute which is a space separated list of ids (or uuids) of other nodes, to be converted to the PDF2 "Ref" property on the element. It will now also cleanup name 'atts' (e.g. Table:Scope) which had values set with incorrect leading slashes.
• The PDF Library will now consistently choose the correct rendering intent from ICC ColorSpaces when converting to PNG.
• Minor tweaks to OutputProfile in anticipation of PDF/UA-2.
• Added the ability to set the appropriate feature when private-use range text is added to a PDF.
• Minor bulletproofing to let the PDF Library work with some more types of damaged files.

不具合の修正

• Added a check for the "zero-value EC signature" issue. It is recommended to upgrade the JVM to fix this, this patch ensures that even older JVMs will never validate this type of signature if its found in a PDF.
• Fixed subtle issue converting tiling-patterns with "general" transforms applied to them that could cause them to render to bitmap incorrectly.
• Fixed an issue when merging Structure trees from two documents; iteration over a HashMap was causing random errors in the generated tree.
• Incorrectly identifying some content as being part of a nested marked-content sequence, which is invalid in PDF/UA.
• Fail (rather then loop forever) when attempting to load an invalid collection index from a WOFF font.
• Don't fail when creating an AnnotationShape that has no area.