[BATIK-1328] No support for unicode characters in U+10000 - U+10FFFF range

• Type: Bug
• Status: Open
• Priority: Trivial
• Resolution: Unresolved
• Affects Version/s: 1.13
• Component/s: SVG DOM
• Assignee: Unassigned
• Reporter: Jasper Krauter
• Created: 03/Jun/22 15:36
• Updated: 03/Jun/22 15:36

Description

The SVG Transcoder checks for valid XML characters but does not take into account characters that, due to the Java String implementation, are represented by two Java chars (UTF-16 Surrogate Pairs). Since neither of those individual chars are a valid XML character on their own, the transcoder fails. But the XML1.0 specification does allow for those characters.

In org.apache.batik.dom.util.DOMUtilities#contentToString, instead of String#charAt, rather String#codePointAt should be used to extract individual characters. Using StringBuffer#appendCodePoint, the code points can properly appended to the output string. The methods that check for character validity already account for code points.

Code example to reproduce the issue:

String svgNS = SVGDOMImplementation.SVG_NAMESPACE_URI;
Document doc = SVGDOMImplementation.getDOMImplementation().createDocument(svgNS, "svg", null);
Element text = doc.createElementNS(svgNS, "text");
text.setTextContent("Hello, world! 👋");
doc.getDocumentElement().appendChild(text);

var transcoder = new SVGTranscoder();
TranscoderOutput out = new TranscoderOutput(new OutputStreamWriter(System.out));
TranscoderInput in = new TranscoderInput(doc);
transcoder.transcode(in, out);

throws

Exception in thread "main" java.lang.RuntimeException: IO:Invalid character
    at batik.transcoder@1.13/org.apache.batik.transcoder.svg2svg.SVGTranscoder.transcode(SVGTranscoder.java:179)