Wait... even better... do this in your swish-e.conf and see if it 
works...


IgnoreMetaTags a

Let us know if that works...


mpeters wrote:
> you could use the <!-- noindex --><!-- index --> tags to comment out 
> specific stuff. Although that would probably be tedious to do that 
> for every link. But if your links are dynamically generated you 
> could have what ever generates them put those tags around.
> 
> Just an idea.
> 
> Or alternatively, you could write a Filter that parses out the <a> 
> tags (using HTML::Parser, etc if you like perl). It shouldn't be 
> that hard. Then you'd just hand the rest off to the libxml parser to 
> do the rest.
> 
> Nic Gibson wrote:
> 
>>Bill, thanks for the help.
>>
>>Having had some sleep (very small babies are not recommended for 
>>sleep), I need to
>>apologise for giving some totally wrong information. On rereading the 
>>docs and looking
>>at what my client is reporting, I realise that a) SWISH-E is performing 
>>exactly as
>>advertised and b) it is not indexing href content, it is indexing the 
>>'label' of the anchor
>>
>>	<a href='....'>Some Text Here</a>
>>
>>The 'Some Text Here' is being indexed. I think this is almost certainly 
>>normal.
>>
>>This leads to another question. Is there is a simple way to stop 
>>SWISH-E indexing this
>>content as it makes the results a little weird.  As an example, 
>>browsing to
>>http://pmr.corbas.co.uk/dynamic/search and entering 'Magnus' as the 
>>search string
>>in the dialog will bring back the appropriate article plus every page 
>>that links to it.
>>
>>Is there a way to modify the indexing to avoid this effect?
>>
>>cheers
>>
>>nic gibson
>>
> 
> 

-- 
Michael Peters
Developer
Plus Three, LP