CHANGES for Follow 2.x

2.04
	  - integrated Cabinet.py
	  - integrated new weblog
	  - other minor tweaks

2.03  - moved data to GDBM shelve for speed, memory efficiency
      - got rid of combocache/  (write perms are not so important now)
      - optional WebLog module lets users specify log format (mail for details)
      - num_show is now keep_days; deletes all old entries, not just sessions
      - onscreen config for limiting sessions by domain name
      - onscreen button in page-by-page to show sessions including current page
      - maxshow removed (now an onscreen config)
      - can now specify multiple names for the site (siteurl)
      - can now ignore directories (directory_ignore)
      - can now limit follow-gather to a maximum number of users (max_clients)
      - can now specify whether page should be popup or text entry (page_popup)
      - can now exclude sessions that include broken hits (exclude_broken)
      - can now exclude sessions that include /robots.txt hits (exclude_robots)
      - sessions of only manual hits longer than one hit are excluded
      - agent_class_ignore is now agent_class_exclude for clarity
      - agent_ignore removed (agent_class_exclude makes redundant)
      - sessions matching agent_class_ignore are now shown and marked EXCLUDED
      - excluded sessions are now shown and marked in session listings
      - file: URL's are now considered manual hits
      - .class files are now considered images (part of a page download)
      - improved news URL handling
      - hostname of incoming referers is normalised to lowercase
      - -v is much more verbose
      - various documentation, performence and interface improvements

2.02  - now finds follow.conf when run from different directory (cron)

2.01  - fixed 'cPickle/copy_reg' problem. Whoops.
      - tweaked ignorable agents in default follow.conf. 

2.0   - first release