This prevents DoS attacks caused by request loops.

Renamed optional.arsnova to ui.arsnova-slogan and made comments start
with an uppercase letter.

WE Offshore Sprint: Session Info Feature

We had the task to implement a Session-Info Form, where the speaker
could add more information to a session, which is already created or to
add more information to a session which is created at that moment.

Therefore we needed to use the update Method and change some code in the
backend as well:

1. Changed **updateSession** Method in SessionService
  * First: we get the session that should be updated from the DB with
    the getSessionFromKeyword-Method
  * Second: we update the values of the session with the new values that
    were entered
  * Third: we check the logo just like in the saving Method of a session
  * we return the updated session
2. Changed **getSessionFromKeyword** Method in CouchDBDao
  * We added these fields, so that they are also fetched:
     * ppAuthorName
     * ppAuthorMail
     * ppUniversity
     * ppLogo
     * ppSubject
     * ppLicense
     * ppDescription
     * ppFaculty
     * ppLevel

See merge request !7

refactored handling of logo during the saving of a new session and the update of an existing session out into a new method handleLogo

Retrieved @Cacheable and changed updateSession method in SessionService, now getting the session by keyword and set all fields that may have changed to a new value

Removed @Chacheable from getSession function, because the Update of a session wont work if only the cache is used for loading the content of a session

This merge introduces pagination for lists of sessions (user specific:
owned and visited), interposed questions and free text answers and skill
questions. While the first three are paginated through the database,
skill questions are paginated in memory so caching is still possible.

Frontend usage

To enable pagination, new HTTP headers based on RFC 7233 have been
introduced:

Request: Range: items=<start>-<end>, e.g. Range: items=0-9
Response: Content-Range: items <start>-<end>/<total>, e.g. Content-Range: items 0-9/42

Since it is not possible to determine the total count of items via
CouchDB w/o performance losses, it is currently only set accurately for
skill questions. In other cases total is set to -1.

Backend usage

Pagination support can be added to controllers by extending the newly
introduced PaginationController. Additionally, applicable methods have
to be marked with the @Pagination annotation so the Range header is
parsed and the Content-Range header is set for responses. Pagination
parameters are accessible in the controller through the offset and limit
fields. If pagination on the DBAL is not possible, lists can be wrapped
with PaginationListDecorator<T> to enable in memory pagination.

See merge request !3

Requesting partial lists would interfere with Spring caching so
pagination is handled by the newly introduced PaginationListDecorator
in the controllers. This additionally allows to determine the total
count of items (independent from pagination parameters) which is not
possible with CouchDB.

The mentioned method is also changed to only retrieve http header
informations instead of whole URI content.

Caching fixes

This PR fixes some caching bugs:

* Cached methods have been called locally inside of `CouchDBDao`. The caching proxy was bypassed in these cases. All public methods (even those not involved in caching) are now called on the proxy instance. The here provided implementation is not optimal since it goes against SoC but can serve as a temporary fix. A better solution might involve AspectJ's weaving.

* A wrong key for cache eviction has been used for `NewAnswerEvent` in `CacheBuster`.

See merge request !2

CacheEvict optimization

Currently, in a lot of places the whole cache is cleared when only a single entry changes. Even completely unrelated caches are sometimes unnecessarily cleared.

This PR adds `key` and `condition` elements to the annotations and replaces the `allEntries` element where possible.

There is still room for additional optimizations but these require some refactoring. For now, the related code locations have been commented on to keep track of them.

The last commit of this PR might be obsoleted later by [GH16](https://github.com/thm-projects/arsnova-backend/pull/16). But for now it should massively increase the effectiveness of answer statistics caching.

See merge request !1

Added TODOs to keep track of methods still evicting the whole answers
cache for future improvements. Additional refactoring is necessary to
implement these.