Prior to this change, Flowdock used Grit to get the difference between
commits and post that to the remote service. This required direct path
access, which doesn't work with Gitaly.

Fixes gitlab-org/gitaly#1113

Old records have a null value of file_store column.
This causes the problems with current direct_upload implementation,
as this makes it to choose Store::REMOTE instead of Store::LOCAL.

This change moves the store save when change saving the object.

Previously we scheduled a worker to just some this but we were running
into performance issues when the build table was getting too large. So
now we've updated the code such that this column is updated immediately
and incremented/decremented by the correct amount whenever artifacts are
created or deleted.

We've also added the performance optimization that we do not update this
statistic if a project is deleted because it could result in many
updates for a project with many builds.

We saw about 10,000 SQL queries for some commits in the NewNoteWorker,
which stalled the Sidekiq queue for other new notes. The notification
service took up to 8 minutes to process the commits. Avoiding this
N+1 query brings the time down significantly.

Closes #45526

In `arel 7.0` (`7.1.4` version is used for rails5) there were introduced
some changes that break our code in the `User#manageable_groups` method.

The problem is that `arel_table[:id].in(Arel::Nodes::SqlLiteral)` generates
wrong `IN ()` construction. The selection for `IN` is missing:

=> "\"namespaces\".\"id\" IN (0)"

That caused such spec errors for the `rails5` branch:

```
4) User groups with child groups #manageable_groups does not include duplicates if a membership was added for the subgroup
    Failure/Error: expect(user.manageable_groups).to contain_exactly(group, subgroup)

      expected collection contained:  [#<Group id:232 @group29>, #<Group id:234 @group29/group30>]
      actual collection contained:    []
      the missing elements were:      [#<Group id:232 @group29>, #<Group id:234 @group29/group30>]
    # ./spec/models/user_spec.rb:699:in `block (5 levels) in <top (required)>'
    # ./spec/spec_helper.rb:188:in `block (2 levels) in <top (required)>'
    # /var/lib/gems/2.3.0/gems/rspec-retry-0.4.6/lib/rspec/retry.rb:112:in `block in run'
    # /var/lib/gems/2.3.0/gems/rspec-retry-0.4.6/lib/rspec/retry.rb:101:in `loop'
    # /var/lib/gems/2.3.0/gems/rspec-retry-0.4.6/lib/rspec/retry.rb:101:in `run'
    # /var/lib/gems/2.3.0/gems/rspec-retry-0.4.6/lib/rspec_ext/rspec_ext.rb:12:in `run_with_retry'
    # /var/lib/gems/2.3.0/gems/rspec-retry-0.4.6/lib/rspec/retry.rb:30:in `block (2 levels) in setup'
```

This commit changes `User#manageable_groups` in the way to drop the usage of
`Arel::Nodes::SqlLiteral` and adds usage of raw SQL query.
This change should be updated when we're migrated to Rails 5.2 because arel
was fixed in `9.0.0` (which is used in Rails 5.2).

This reverts the addition of the "goldiloader" Gem and all use of it.
While this Gem is very promising it's causing a variety of problems on
GitLab.com due to it eager-loading too much data in places where we
don't expect/can handle this. At least for the time being this means we
have to go back to manually fixing N+1 query problems, but at least
those should not cause a negative impact on availability.

and also make sure that if a field is saved, then the existing cache version is maintained or only upgraded to the version with the same markdown engine.

This is useful for a transition period to migrate away from
`NoninternalAtomicId`. In a situation where both the old and new code to
generate a iid value is run at the same time (for example, during a
deploy different nodes may serve both versions), this will lead to
problems regarding the correct `last_value`. That is, what we track in
`InternalId` may get out of sync with the maximum iid present for
issues.

With this change, we double-check that and correct the `last_value` with
the maximum iid found in issues if necessary.

This is subject to be removed with the 10.8 release and tracked over
here: https://gitlab.com/gitlab-org/gitlab-ce/issues/45389

Closes #45269.

Some specs fail in rails5 branch with errors like

```
1) Group#secret_variables_for when the ref is not protected contains only the secret variables
    Failure/Error: variables = Ci::GroupVariable.where(group: list_of_ids)

    NoMethodError:
      undefined method `arel_table' for Ci::Group:Class
```

This commit fixes it.

In Rails 5.0 the `ActiveRecord::Inheritance::subclass_from_attributes`
method was updated.
Now it calls the `find_sti_class` method [1] which is overriden in the `Event`
model and returns needed class (`Event` vs `PushEvent`).

[1]: https://github.com/rails/rails/blob/5-0-stable/activerecord/lib/active_record/inheritance.rb#L209

This commit fixes the errors like
```
143) User#contributed_projects doesn't include IDs for unrelated projects
      Failure/Error: action = attrs.with_indifferent_access[inheritance_column].to_i

      NoMethodError:
        undefined method `with_indifferent_access' for nil:NilClass
      # ./app/models/event.rb:118:in `subclass_from_attributes'
```
which are raised on the `RAILS5=1 rspec ...` command.

This prevents performing the requests, and disables all emoji reaction buttons

Goldiloader is great, but has several issues with has_many :through relations:

* https://github.com/salsify/goldiloader/issues/12
* https://github.com/salsify/goldiloader/issues/14
* https://github.com/salsify/goldiloader/issues/18

Rather than try to figure out which applies in each case, we should just do the
drudge work of manually disabling autoloading for all relations of this type. We
can always use regular preloading for specific cases, but this way we avoid
generating invalid queries through Goldiloader's magic.

Goldiloader (https://github.com/salsify/goldiloader) can eager load
associations automatically. This removes the need for adding "includes"
calls in a variety of different places. This also comes with the added
benefit of not having to eager load data if it's not used.

Use project_authorizations instead of members to calculate manageable CI projects to speed up query (#41981)

In Rails 5.0 the `ActiveRecord::ConnectionAdapters::Column::TRUE_VALUES`
constant has been removed [1] and the remaining `FALSE_VALUES` constant
has been moved to `ActiveModel::Type::Boolean` [2]

[1]: https://github.com/rails/rails/commit/a502703c3d2151d4d3b421b29fefdac5ad05df61
[2]: https://github.com/rails/rails/commit/9cc8c6f3730df3d94c81a55be9ee1b7b4ffd29f6

The TIMESTAMP data type is used for values that contain both date and
time parts. TIMESTAMP has a range of '1970-01-01 00:00:01' UTC to
'2038-01-19 03:14:07' UTC.

A Forever lib class was included to handle future dates for PostgreSQL
and MySQL, also changes were made to DeployToken to enforce Forever.date

Also removes extra conditional from JwtController

- Adds a default on expires_at datetime
- Modifies deploy tokens views to handle default expires at value
- Use datetime_with_timezone where possible
- Remove unused scopes

Also, fixes broken specs

Also:
- Changes scopes from serializer to use boolean columns
- Fixes broken specs