Re: LuceneIndexContentHandler

[Joerg Heinicke]

On 11.03.2004 20:50, Joerg Heinicke wrote:

Replace the startElements in your sample with endElements and it won't 
work. But adding if(flag) like above additionally to endElement should 
fix this. Did we get it now? :)
I have it done like the following

characters
    flag = true
endElement
   if(flag)
      append(' ');
      flag = false;
startElement
   if(flag)
      append(' ');
      flag = false;
But after having the closer look on the code I really wonder why my 
original proposal doing it in endElement depending on the text test 
should not work. The text only contains child::text() (to say it in 
XPath) and through the stack the text is completely reordered depending 
on the order of the endElement tags in the file:

<element>
    text1
    <element>
        text2
    </element>
    text3
</element>
should end in

"text2 text1 text3", shouldn't it? And this should be doable through the 
text test in endElement.

Joerg