- Now all checks are organized to rules.
- Allow to specify read_only rules to avoid problems on learning.
- Use better normalizer for fuzzy module and it now returns values
  from 0 to 1.0 (like bayes does).
- Update configuration accordingly.
- Drop legacy configuration support.
- Detect tanh as well and provide some reasonable (linear) fallback.

Now rspamd can listen on multiply ipv4/ipv6 addresses.
Removed legacy workers (lmtp, kvstorage) as they are never used
in production.
Try to unify workers initialization.

Fix for cmake 2.6.

* Add partial ipv6 support for some rspamd modules.

  - Removed legacy limitation of worker's types;
  - Using GQuarks to identify workers and modules;
  - Remove modules.sh script;
  - Add a common system of workers and modules;
  - Write management and configuration for new architecture.

Rspamd now can detect and work with libevent-2.

Detect and link against bdb
Change version to 0.4.5

* Fix build of rspamd with CMAKE_BINARY_DIR differs from CMAKE_SOURCE_DIR
Rework include style.

* Add initial version of statshow utility for statfiles debugging
* Add debugging for statistics
* Remove unused utilities

* Fix issue with links that are ip addresses in numeric form in surbl
* On Darwin use BSD style sendfile definition
* Reorganize platform specific knobs in CMakeLists
* Use gettimeofday on systems that have not clock_getres
* Use ftime for dns trans id generation on systems without clock_getres

 - remove completely 2tld
 - add option "exception"
 - for domains from exception list check (level + 1) parts of url:
   If we have url mail.some.com.ru and have com.ru in exception list
   then we would check some.com.ru. If we have some.com.ru in exceptions
   list them mail.some.com.ru would be checked and so on.
 - optimized parsing of surbl requests

* Use system mkstemp(3) on systems where it is available as glib implementation
  has poor security and generate rather predictable temporary file names.