#10 Strange behaviour handing comment and CDATA sections

[Tom Fennelly]

Hi.

I was trying to use Hotsax but I noticed some strange
behaviour when it came to handling of <!-- comments -->
and <![CDATA[]]> sections.

As a test I wrote a very simple ContentHandler &
LexicalHandler (attached) which just output the SAX
event sequence to sysout.

Test 1:

Input:
<x>
Text
<!-- comment -->
</x>

Resulting event sequence:
startDocument
startElement: [, x, ]
characters: [ Text ]
comment: [ comment ]
characters: [ Text ]
endElement: [, x, ]
endDocument

As you can see, the characters " Text " are getting
fired twice.

Test 2:

Input:
<x>
Text
<![CDATA[Cdata]]>
</x>

Resulting event sequence:
startDocument
startElement: [, x, ]
characters: [ Text ]
startCDATA
characters: [ Text ]
endElement: [, x, ]
endDocument

Again, very similar but with the added problem of none
of the CDATA data being fired at the content handler.

[Tom Fennelly]

Parser code - includes the Content/Lexical Handler implementation.

[Simon Massey]

you have a carriage return before your end tag.

if you tried:

<x>
Text
<!-- comment --></x>

<x>
Text
<![CDATA[Cdata]]></x>

then you don't get the problem which is shown by:

@Test
public void testSimpleComment() throws Exception {
String html = "<x>\n" +
"Text\n" +
"<!-- comment -->\n" +
"</x>";

final List<String> text = new ArrayList<String>();

ContentHandler ch = new DefaultContentHandler(){

@Override
public void characters(char[] ch, int start, int length)
throws SAXException {
String t = new String(ch, start, length);
text.add(t);
}

};

parser.setContentHandler(ch);

InputSource input = new InputSource(new StringReader(html));

parser.parse(input);

Assert.assertThat(text.size(), is (2));
Assert.assertThat(text.get(0), is("\nText\n") );
Assert.assertThat(text.get(1), is("\n") );

}

on the version of 0.1.2b which I am running.