iText 5.5.2 Release: Major Improvements to the CSS Parser

Tags: Release

Two months have passed since iText 5.5.1 and we're ready to release iText 5.5.2. Although the changelog may seem relatively short, there's one line in specific we've put a lot of effort into: Major improvements to the CSS parser. This is actually a complete rewrite of the CSS functionality, resulting in much better interpretation of CSS syntax. You'll find these changes in iText's XML Worker. The HTMLWorker class that was shipped with iText core was finally deprecated.

We noticed that many people are starting to be aware of Tagged PDF, PDF/A and PDF/UA. We received numerous questions on the subject of merging Tagged PDFs, more specifically PDF/A compliant documents. This wasn't supported in previous versions and it's also kind of tricky because you can't just mix the different flavors of PDF/A. For instance: you shouldn't expect that a PDF/A-1B document merged with a PDF/A-2A document will result in a PDF/A-2A document. Nevertheless we've added a class PdfACopy that allows you to make combinations that make sense and that throws an exception when you're trying to make a combination that doesn't.

We made sure that you can't use the PageStamp functionality when using PdfCopy in tagged mode. This way, we avoid that people don't introduce untagged data out of ignorance regarding the nature of Tagged PDF. PdfCopy now also supports merging fields with tags.

When parsing documents with Type 3 fonts, we used to return space characters because Type 3 fonts are user defined fonts where a random character is mapped to a custom made glyph. We've changed this so that iText returns the characters that were stored inside the PDF. This doesn't always make sense because it can result in a set of random characters, but it is more in line with what people expect when parsing a PDF file that contains Type 3 fonts.

Finally, we've introduced functionality to cache objects to improve the PDF writing process, and we've improved the generic testing framework.