[Mailman-Users] Setting up indexing on Mailman archives

[Richard Barrett]

On Friday, October 10, 2003, at 02:33  pm, Paul F. Williams wrote:

> I have read through numerous archives on how to setup
> indexing on the Mailman archives but I seem to be missing
> something.
>
> In the past I have set up htdig to make it possible to
> search web sites.  I used the url indexing method.
> In order to accomplish that you need a url starting point
> that will find all the other web documents.
>
> So, my question is
>
> What url do you give htdig (or any other indexer) so it can
> find all the archives?
>
> Or do you need to create a document that contains
> a list of starting urls?
>

One alternative is to use the Mailman-htdig integration patches #444879 
and # 444884 referred to on this FAQ page:

http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq01.011.htp

Even if you go some other route for using htdig you should consider 
using the #444879 patch "Archive indexer control to improve indexing".

> If you need to create a document, what do you put in it?

For public list archives, links to 
http://your.web.domain.tld/pipermail/<listname>/index.html

If you do not use the integration patch you will find it a little 
tricky setting up search for private archives. htdig cannot directly 
index private archives through 
http://your.web.domain.tld/mailman/private/<listname>/ URLs; it does 
not know how to do cookie authentication. For private archives, htidg 
has to be told to index through the file system path to private 
archives but return the matching http URLs in search results. But then 
this is basically insecure because the search results contain 
information from documents which the user may not be authorised to see. 
The integration patch deals with this whole problem.

>
> Thanks,
>
> paulw
>
-----------------------------------------------------------------------
Richard Barrett                               http://www.openinfo.co.uk