This ensures the right caches are flushed when adding a branch via the
UI, instead of only flushing this one random cache.

This ensures Repository#add_tag calls Repository#before_push_tag instead
of just 1 random cache expiration method.

This ensures that Repository#rm_branch calls
before_remove_branch/after_remove_branch instead of just 1 random cache
expiration method.

The methods used for this are Repository#tag_count and
Repository#branch_count which cache their output in Redis as well as
memoizing it in an instance variable. Both methods have a corresponding
methods/hooks to flush the caches at the right time.

This is only needed when calling the "expire_cache" method as the other
expiration methods don't depend on Git data.

This keeps all the cache expiration code in a single file/class instead
of spreading it all across the codebase.

This ensures that the emptiness cache (used for Repository#empty? and
Repository#has_visible_content?) is flushed after comitting changes
(using the web editor, API or Git) for new repositories. Once a
repository is no longer empty there's no need to explicitly flush the
cache for Repository#empty?. The cache for
Repository#has_visible_content? in turn is already flushed whenever
needed.

Fixes gitlab-org/gitlab-ce#13387

* Now checking if a commit is already reverted or there is a conflict is
  much more faster.
* No longer need to create a new branch.

* Not required to run hooks since it's an internal commit

If path_with_namespace is nil Repository#raw_repository will also return
nil. Apparently code out there creates a Repository instance without a
namespace path. Right.

Just checking raw_repository is no longer accurate to determine if a
repository exists or not.

Now that Repository#raw_repository no longer sets the autocrlf option it
will also no longer raise any NoRepository errors since it doesn't
access Rugged any more. This also means that Repository#exists? can't
simply return the raw repository as this is no indication of whether or
not the repository actually exists (besides returning a non boolean is
weird in the first place).

To solve this problem Repository#exists? now properly checks if the
repository exists and returns true/false instead of a
Gitlab::Git::Repository or nil object.

Setting the "autocrlf" Git option is an overkill since it's rarely
actually needed. More importantly, it has quite the impact on
performance (see gitlab-org/gitlab-ce#13457 for more information).

By setting "autocrlf" when creating or updating files we guarantee the
option is always set properly when we actually need it _without_
introducing overhead for requests that have nothing to do with this
option.

Fixes gitlab-org/gitlab-ce#13457