Version 0.7.0 - icalendar support and quite some fixups

Version 0.7.0.

Updates since 0.6.2:

As for real code changes - actually, not that much. The library is relatively mature now - the remaining changes and issues are mostly either sufficiently significant that they should be procrastinated for a 1.0-release, or I consider them too insignificant to bother with it (except #87 which may come in a 0.8-release). Here is the list of changes:

• Some tweaks to support newer versions of Zimbra and other caldav server implementations with slightly twisted support for the RFC.
• API change: add_event, add_todo and add_journal methods are now deprecated and aliases of save_. New attributes no_create and no_overwrite if one wants to be sure the save_-methods does the right thing.
• Calendars can now be created without a display name
• Quite some minor bugfixes
• Some code refactoring, removing some duplicated code
• library will gracefully create an UID if fed with caldav data missing the UID field
• Earlier, for calendar servers not supporting the digest auth format, every request towards the server would be duplicated (try digest auth, retry basic auth). I'm sure there is a more proper fix for it, but at least now it should only try digest auth once, and then save the authentication information.
• it's now possible to use the caldav library together with icalendar instances (but as of 0.7.0, vobject is still the officially supported icalendar library, one has to use event_object.icalendar_instance rather than event_object.instance, the icalendar library is imported when needed and not included in the requirements).
• the code for calendar searches has been reorganized making it possible for "power users" to do custom calendar searches.

Changes outside the library code itself:

• Makefile
• A python file has been stored in the examples folder, with quite some basic usage examples
• The documentation has been refreshed, manually copied some of the lines from that example file into the documentation. Documentation has been uploaded to ReadTheDocs (but something is buggy there, as the autogenerated parts of the documentation isn't available there)
• Quite some work has been done on the test code - doing my best to uncover weird RFC incompatibilities from the servers tested as well as increasing code coverage.
• Quite some cleanup work in the github issues

ref #97