[Htmlparser-developer] Table Scanners

[<dha...@po...>]

Hi,

I wrote the following test case

public void testUnClosed () throws ParserException
{
	createParser("<TABLE><TR><TR></TR></TABLE>");
	parseAndAssertNodeCount(1);
=09
assertEquals("Unclosed","<TABLE><TR></TR><TR></TR></TABLE>",node[0].toHt
ml());
}=09

I was expecting one node, since <TABLE> would be the main node and <TR>
would be its children, but I got 5!!!.
Obviously because of that the assert also failed.

Since nesting is allowed by default, if there is code as follows :
...
<TR>
<TD>blah blah</TD>
<TR>
<TD>blah blah</TD>
</TR>
....
Then the second <TR> is incorrectly considered as a child of the first
<TR> whereas in reality, a closing </TR> was missed out should have been
put in. Hence for <TR> nesting should be disallowed through the scanner.
Same holds for <TD> tag. I think this is a bug.

Dhaval