PDFreactor 12.5 now available
A release for PDFreactor 12 is now available.
The following issues were fixed:
- In certain documents using multi-column can result in exceptions due to incomplete layouts. (#9717)
- Multi-column spanner boxes with forced column breaks cause exceptions. (#9730)
- Multi-column spanner elements that begin with floated or positioned elements cause exceptions. (#9737)
- Merging documents with a huge amount of bookmarks can cause conversions to terminate with
StackOverflowErrors. (#9519, #9615) - Overlaying a PDF with a shorter PDF can cause conversions to terminate with exceptions. (#9747)
- Using the
insetfunction with an explicitly specified border radius of zero causes a terminating exception. (#9798) - Using the
nonekeyword for color components in certain CSS color functions, likelchorlabcauses terminating exceptions. (#9797) - When PDF tagging is enabled DOM element IDs containing any of the characters
!and^can cause conversions to terminate with exceptions. (#9555) - Creating inspectable PDFs terminates with an exception when exceeding content is detected inside a page margin box. (#9756)
- When passing down styles to iframes inside running elements, an exception occurs due to unfinished layout. (#9772)
hrefattributes in SVG without thexlinknamespace are not supported. (#9693)lengthAdjustin SVGs uses inaccurate width determination and breaks the SVG when the text with a white-space that is not collapsed. (#9773)- Unsupported CSS variables in external SVGs can affect embedded SVGs in the same host document. (#9789)
- Some WebP images cannot be decoded. (#9726, #9729)
- In certain cases when using the CSS functions
leaderandtarget-counter, the determined page number is incorrect. (#9706) - Grid layouts with flexible tracks and direction RTL can lead to overflowing grid items. (#9752)
- Using named strings inside running elements collapses spaces unexpectedly if the named strings can be empty. (#9762)
- JavaScript
window.requestAnimationFrame()callbacks don’t have a timestamp parameter.
This causes missing logos in amCharts 5.15. (#9721) - PDF attachments with the locations
documentspecified via CSS on elements spanning multiple pages are added multiple times. (#9761) - Using
-ro-image-resamplingon an image which is referenced multiple times causes the default compression to change fromjpeg(80%)tolossless(). (#9779) - Recompressing JPEG images in debug mode causes these images to be added to the debug resources attachment twice. (#9780)
- PDFreactor Web Service
- Server log mode
livecauses conversions to hang or stop logging. (#9727) - Specifying complex configuration types in the Python CLI client doesn’t work properly. (#9687)
- Server log mode
The following CVEs were mitigated:
- Mitigated CVEs in packaged Java runtime. (CVE-2026-21925, CVE-2026-21932, CVE-2026-21933, CVE-2026-21945)
- PDFreactor Web Service
- CVE-2025-11143
- CVE-2025-41242
- CVE-2025-48913
- GHSA-72hv-8253-57qq
- CVE-2026-1605
The following functionalities were deprecated:
- Using PDFreactor without a license key is deprecated. Currently, without a license key, PDFreactor runs in evaluation mode. In evaluation mode PDFreactor currently works just like the full version, but the resulting PDF document will include watermarks and additional evaluation pages. This will no longer be possible in future versions.
The following dependencies were updated:
- Packaged Java runtime Eclipse Temurin 21.0.9 to 21.0.10
- com.twelvemonkeys.imageio.* 3.12.0 to 3.13.1
- org.apache.pdfbox.* (repackaged) 3.0.5 to 3.0.7
- PDFreactor Web Service
- com.fasterxml.jackson.* 2.19.0 to 2.21.1
- org.springframework.* 6.2.12 to 6.2.16
- org.eclipse.jetty.* 12.0.30 to 12.0.33
- org.apache.cxf.* 4.1.3 to 4.1.5
For a full list of changes and corrections see the changelog.
Important release notes und upgrading information can be found in the readme.
The PDFreactor 12 installation packages are available for download in the download area.


