[BATIK-1026] ParsedURL fails using "data:" URI scheme when data contains style sheet including #-character

• Type: Bug
• Status: Open
• Resolution: Unresolved
• Affects Version/s: 1.8
• Component/s: SVG Rasterizer
• Environment: Operating System: Linux; Platform: PC
• Assignee: Batik Developer's Mailing list
• Reporter: Matthias Gerber
• Created: 28/Oct/12 16:26
• Updated: 28/Oct/12 16:26

Description

Transcoder allows to pass a user style sheet uri an a hint.
Passing an data uri, such as "data:text/css;charset=UTF-8, .g

{ display:none;}

" works as expected. But as soon, as the data (style sheet) contains a #-character, the style sheet can not be parsed correctly any more.
The reason for this is in the ParsedULR class (or a class used by it - ParsedURLDataProtocolHandler:parseURL(String urlStr)) because the # character is interpreted as a URL fragment separator, not considering the used URI schema at all. As far as i know is the syntax for data URI's "data:<mediatype>[;base64],<data>". Because of this, the part after the first #-character is taken as URI fragment and is no longer part of the style sheet, resulting in an exception like this:

org.w3c.dom.DOMException: data:text/css;charset=UTF-8,#RestrictedAreas

{ display:none;}

:
Invalid CSS document.
java.lang.NullPointerException
at org.apache.batik.css.engine.CSSEngine.parseStyleSheet(CSSEngine.java:1184)
at org.apache.batik.css.engine.CSSEngine.parseStyleSheet(CSSEngine.java:1125)
at org.apache.batik.bridge.BridgeContext.initializeDocument(BridgeContext.java:386)
at org.apache.batik.bridge.GVTBuilder.build(GVTBuilder.java:55)
at org.apache.batik.transcoder.SVGAbstractTranscoder.transcode(SVGAbstractTranscoder.java:208)
at org.apache.batik.transcoder.image.ImageTranscoder.transcode(ImageTranscoder.java:92)
at org.apache.batik.transcoder.XMLAbstractTranscoder.transcode(XMLAbstractTranscoder.java:142)
at org.apache.batik.transcoder.SVGAbstractTranscoder.transcode(SVGAbstractTranscoder.java:156)

Reproduce:
Passing an URI as "data:text/css;charset=UTF-8,#RestrictedAreas

{ display:none;}

" as a user-stylsheet-uri to the transcoder hints.

Workaround:
a) Use a file instead of encoding data directly in the URL
b) Encode style sheet data using based64

String stylesheet = "#RestrictedAreas

{ display:none;}

"
Charset charset = Charset.forName("UTF-8");
byte[] styleSheetBytes = Base64.encodeBase64(charset.encode(stylesheet).array());
CharBuffer styleSheet = charset.decode(ByteBuffer.wrap(styleSheetBytes));
String dataURL = "data:text/css;base64,"+styleSheet;
hints.put(ImageTranscoder.KEY_USER_STYLESHEET_URI, dataURL);

regards