When converting HTML to PDF, font size specified in % is interpreted as pixels

[RadekWikturna]

Describe the bug
When converting HTML to PDF, using HTMLWorker.parseToList, font size specified in % is interpreted as pixels

To Reproduce

@Test	
public void testFontSize() throws Exception {
	StringReader reader = new StringReader(
			"<span style=\"font-size:8.0pt\">Text 8.0pt</span><br/>"
			+ "<span style=\"font-size:20px\">Text 20px</span><br/>"
			+ "<span style=\"font-size:85%\">Text 85%</span><br/>"
			+ "<span style=\"font-size:1.5em\">Text 1.5em</span><br/>");
	StyleSheet styleSheet = new StyleSheet();
	Map<String, Object> interfaceProps = new HashMap<>();
	List<Element> elements = HTMLWorker.parseToList(reader, styleSheet, interfaceProps);
		
        Document document = new Document();
        PdfWriter instance = PdfWriter.getInstance(document, new FileOutputStream("Font Size.pdf"));
        document.open();
        instance.getInfo().put(PdfName.CREATOR, new PdfString(Document.getVersion()));
        for(Element e: elements) {
        	document.add(e);
        }
        document.close();
}

Expected behavior
All font size variants, including %, should be properly interpreted and shown approximately as the HTML in a browser.

Screenshots
HTML in a browser:

Generarted PDF:

System:

• OS: Windows 11
• Used Font: default

[asturio]

Pull requests are welcome!

[asturio]

Although I really don't know how much of HTML and CSS parsing should be integrated in OpenPDF as a project with the goal of generating PDF from Java...

[RadekWikturna]

I understand it's not trivial. However, there's API that is supposed to do that (packages text.html, text.html.simpleparser - e.g. , HTMLParser, HTMLWorker) so some basic stuff should work. I've been using iText 2.1.7 for years in our project and we've used it also for converting HTML to PDF and neither in iText this worked correctly, so I had to modify some original iText classes. But now we are about to switch to OpenPDF (mainly to avoid vulnerabilities in old iText) and so I'm restesting stuff that we needed to somehow handle and some problems seem be be fixed in OpenPDF, but some remain - like this one. If there's noone to fix it, I may try it myself - but I've never participated on an opens source project...

[asturio]

That is a great opportunity for starting :-). If you already have some fixes for iText in your own fork, than you could contribute with those changes.

I really recommend using GitHub and contribute to the project, instead of having an old fork of an unmaintained itext.

[RadekWikturna]

OK, I'll try to fix this myself - at lest I'll learn how to create pull request etc.