Update dependency com.itextpdf:html2pdf to v6.3.4

[renovate]

This PR contains the following updates:

Package	Change	Age	Confidence
com.itextpdf:html2pdf (source)	6.3.1 → 6.3.4

---

Release Notes

itext/itext-pdfhtml-java (com.itextpdf:html2pdf)

v6.3.4: pdfHTML 6.3.4

Compare Source

pdfHTML is an iText Core add-on for creating PDF from HTML/XML (and associated CSS).

TThis release adds initial support for HTML vertical text rendering in pdfHTML. It supports the CSS vertical writing modes writing-mode: vertical-rl and writing-mode: vertical-lr when they're combined with text-orientation: upright, helping align generated PDFs more closely with browser behavior for East Asian (CJK) documents.

Although the implementation is not complete, we have finalized the vertical text API and improved the way pdfHTML handles vertical layout calculations, word wrapping, text rise, and related CSS property interactions. This includes broader coverage for spacing, overflow, alignment, decorations, floats, inline vertical alignment, images, inline blocks, mixed font sizes, and min-max width behavior in complex layouts.

In addition, this release improves quality and robustness for vertical writing scenarios with expanded CJK font and vertical text validation, plus tests for extreme values and boundary conditions to help guard against stability issues such as infinite loops and null pointer exceptions during conversion.

We've also improved PDF/UA language attribute handling during partial HTML element conversion, so empty language attributes are handled more consistently.

New Feature

• DEVSIX-9749 - DEVSIX-10161 - Improve HTML vertical text drawing

Improvements

• DDEVSIX-10190 - Finalize vertical text API

• DEVSIX-10175 - Improve setting language for partial html element conversion

• DEVSIX-10156 - Create extensive test suite for CJK fonts and vertical text

• DEVSIX-10184 - Test min-max width behavior with vertical text

• DEVSIX-10129 - Create test suite for vertical text drawing properties such as spacing, overflow, alignment, decorations, floats, inline vertical alignment, images, inline blocks, and mixed font sizes.

Bug Fixes

• DEVSIX-10195 - Added extreme-value and boundary-condition coverage for vertical writing to guard against stability failures such as infinite loops and null pointer exceptions during conversion.

• DEVSIX-10133 - Corrected PDF/UA handling for partial HTML conversion when empty language attributes are present.

v6.3.3: pdfHTML 6.3.3

Compare Source

pdfHTML is an iText Core add-on for creating PDF from HTML/XML (and associated CSS).

This release adds support for the CSS :is(), :where(), and :not() pseudo-class selectors, improves tolerance of malformed CSS input, and fixes bugs affecting CSS Grid pagination and list rendering performance.

We've added support for the :is(), :where(), and :not() CSS pseudo-class selectors, contributed via a community pull request. These selectors are now resolved correctly in both iText Core's styled-xml-parser and pdfHTML, including handling of specificity differences between :is() and :where(), nested selectors, and combinators.

We've also improved how pdfHTML handles invalid CSS unit values. Previously, encountering a malformed unit (e.g. %%) could cause a hard failure during conversion. Now pdfHTML handles these cases more tolerantly, logging a warning and continuing processing rather than throwing an exception. This was reported by a customer encountering real-world HTML with non-standard unit patterns.

Two related bugs affecting CSS Grid layouts were resolved. The first caused list items and their symbols to be rendered twice when nested inside grid containers with display:grid and inline-table/list structures, which also led to unexpectedly large output PDFs and very slow conversion times. The second fixed an issue where grid cells split across page breaks would incorrectly use the original cell renderer instead of the split renderer, resulting in duplicated or misplaced content on the subsequent page.

New Feature

• DEVSIX-9749 - Add :is() and :where() selector support

Improvements

• DEVSIX-8800 - Implement a more tolerant error handling for wrong units in pdfHTML

Bug Fixes

• DEVSIX-10029 - pdfHTML conversion is very slow with display:grid and nested inline-table/list structures
• DEVSIX-10033 - Grid uses the original cell renderer instead of the split renderer after partial layout

v6.3.2: pdfHTML 6.3.2

Compare Source

pdfHTML is an iText Core add-on for creating PDF from HTML/XML (and associated CSS).

For this release we took the opportunity to reflect the improved Well-Tagged PDF (WTPDF) creation in iText Core 9.6.0 to enable WTPDF conformance support in pdfHTML. We also added support for one CSS pseudo-class, and improved the implementation of another, along with some more minor refinements and bug fixes.

As noted in the Mozilla Developer Network documentation, the :has() pseudo-class matches an element when the selector inside finds a match relative to that element. In simple terms can be called a “parent selector”, as it allows you to style a parent (or even an element that comes before another) based on what it contains or comes after it.

We actually added this to help with the release-notes-to-pdf generator we recently open-sourced, as we found it would be useful when referencing the table of contributors to a specific release version. Hopefully someone else will find it useful too!

The other pseudo-class is the :not() selector for the @​page at-rule. You can now specify @page :not(:first) to apply styles to all pages except the first page, whereas before it behaved incorrectly.

We’ve refactored how pdfHTML handles dependency injection, related to the changes to make StampingProperties and DocumentPropertiesstateless in iText Core 9.6.0:

• ConverterProperties now stores dependency suppliers instead of dependency instances, and new methods to access dependency classes and retrieve a dependency’s supplier have been introduced.
• ProcessorContext was updated to register dependencies using the new supplier-based API, while the old getDependencies() API is now marked as deprecated and returns fresh instances created via suppliers.

An issue with positioning of elements (such as list items) being lost when the block is split across pages was resolved, along with an issue with HTML tables using collapsed borders with large rowspans that overflow/split onto a new page.

Bug Fixes

• DEVSIX-9700 – Support css :has pseudo class
• DEVSIX-9767 – Add WTPDF creation using pdfHTML

Improvements

• DEVSIX-9613 – Make DocumentProperties/StampingProperties stateless

Bug Fixes

• DEVSIX-2431 – Positioned elements (e.g. absolute positioning) are lost when block is split across pages
• DEVSIX-5036 – Original rowspan is restored while overflowing to the next page
• DEVSIX-9715 – Absolutely positioned elements can be dropped when parent block-element is split across pages.
• DEVSIX-7094 – Page selector is applied even though :not is specified

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • At any time (no schedule defined)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.