[SHINDIG-987] NekoParser returns cryptic error messages when parsing bad html

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.1-BETA1
• Fix Version/s: 1.1-BETA1
• Component/s: Java
• Assignee: Unassigned
• Reporter: Paul Lindner
• Created: 19/Mar/09 17:25
• Updated: 10/May/09 06:12
• Resolved: 10/May/09 06:12

Description

startImportantElement can throw exceptions when parsing malformed html:

Given this html:

<div id="div_super" class="div_super" valign:"middle"></div>

You get an exception like this:

org.w3c.dom.DOMException: INVALID_CHARACTER_ERR: An invalid or illegal XML character is specified.
org.apache.xerces.dom.CoreDocumentImpl.createAttribute(Unknown Source)
org.apache.xerces.dom.ElementImpl.setAttribute(Unknown Source)
org.apache.shindig.gadgets.parse.nekohtml.NekoSimplifiedHtmlParser$DocumentHandler.startImportantElement(NekoSimplifiedHtmlParser.java:292)
org.apache.shindig.gadgets.parse.nekohtml.NekoSimplifiedHtmlParser$DocumentHandler.startElement(NekoSimplifiedHtmlParser.java:242)
org.apache.shindig.gadgets.parse.nekohtml.SocialMarkupHtmlParser$SocialMarkupDocumentHandler.startElement(SocialMarkupHtmlParser.java:130)

Which is caused here:

for (int i = 0; i < xmlAttributes.getLength(); i++) {
if (xmlAttributes.getURI != null)

{ element.setAttributeNS(xmlAttributes.getURI(i), xmlAttributes.getQName(i), xmlAttributes.getValue(i)); }

else

{ element.setAttribute(xmlAttributes.getLocalName(i) , xmlAttributes.getValue(i)); }

}

because we're trying to set a tag with a colon in it.

We should probably add some error checking here so that we can more easily identify the offending HTML without using a debugger.