


                        Archie 3.5 Beta
                        ---------------



New in this release:
--------------------


- Lock files are now created in ~archie/db/locks

- the ``-t'' switch on arcontrol creates the new files  and work files
  in the specified temp directory and not in ~archie/db/tmp

- New module (webindex)  with the databse in ~archie/db/webindex_db

- New database structure and search engine

  The files Stridx.* in db/anonftp_db or in db/webindex_db 
  correspond to the new search engine.
  
  The site files have now a new internal structure and have their
  filenames in the format ip_addr:port
  For anonftp, the port is always 21 whereas for webindex
  the ports may be different.

  In order to speed the searches, an index of the searchable strings
  is used. This index is not built every time a `arcontrol -u' is run.
  A separate program db_build is used for this.

- Webindex

  Archie still works on a site by site basis. New sites/urls found
  along the way are stored in the ~archie/db/webindex_db/exturls_db.*
  database. The program `extern_urls' can be used to add sites to
  the hostdb database so that Archie can index them.
  
- cgi-bin scripts for accessing the Archie database through
  the Web.

- A much better domain filter  for anonftp than before !!

  The file ~archie/etc/domain.order contains the order of the domains

  domains list on the same line are equivalent.
  domain lines are ordered as they appear .. i.e. domains in line i
  appear before in line j  <==>  i < j

  e.g.
       com:edu
       fr:de

  .com and .edu sites will appear before .fr and .de sites in the
  results.



Currently working on:
---------------------

- Rewrite of the cgi-bin front-end to be more flexible.

- Archie Help page

- Regular Expressions with the new search engine.

- New set of manpages and documentation.


Currently testing:
------------------

- dirsv with the new database technology.


Known Problems:
---------------

- arexchange of webindex will not work 100%
  it will not transfer .excerpt files... we still need to figure out
  how much infromation is gathered first by the indexers...


