[API] Fix URLs in the 'Link' header for 'GET /projects/:id/repository/contributors' when no value is passed for 'order_by' or 'sort'

Signed-off-by: Rémy Coutable <remy@rymai.me>

When a request contains an ETag value in its If-None-Match header, the backend
may send a request ID (from Rack) that does not correspond to a value in Peek's
Redis cache (because we aborted the Rails processing in our ETag middleware).

Because a cached response (304) has to replace the headers with those from the
previous 200 - https://tools.ietf.org/html/rfc7234#section-4.3.4 - we add a
custom header that will only be present in cache hits, that can tell the
frontend to ignore these.

This adds a method to track errors that can be recovered from in
sentry.

It is useful when debugging performance issues, or exceptions that are
hard to reproduce.

This is called repeatedly when viewing a merge request, and this should
improve performance significantly by avoiding shelling out to git every time.

This should help https://gitlab.com/gitlab-com/infrastructure/issues/4027.

Closes #45247.

Direct disk access is impossible on environments where components run in
their own container. There for constructing the path should not be done
except for exceptional cases.

One of the considerations, instead of logging, was to use Sentry. For
now I've chosen not to use this as the impact of this change is hard to
determine. Getting this in now will allow us to checkout the impact on
dev.gitlab.org and staging.

Since !18150 abstract reference filter supports also `group`
as a resource parent (not only `project`). Some reference filters,
which inherit from abstract filter, depend on parent being `project`
when checking a reference.

This patch adds explicit check of parent class to the given filters.

Fixes gitaly#1126

Clients can now request the attributes from `$GIT_DIR/info/attributes`
through Gitaly. The Gitaly migration is described in gitlab-org/gitaly#1082.

The parser algorithm was implemented in a way it could handle both file
contents or a File handle, and both were already tested.

Other than that, using the boy scout rule, I've removed a class,
InfoAttributes, as it was delegating everything to the parser and
therefor wasn't really needed in my opinion.

This refactors the Markdown pipeline so it supports the rendering of
multiple documents that may belong to different projects. An example of
where this happens is when displaying the event feed of a group. In this
case we retrieve events for all projects in the group. Previously we
would group events per project and render these chunks separately, but
this would result in many SQL queries being executed. By extending the
Markdown pipeline to support this out of the box we can drastically
reduce the number of SQL queries.

To achieve this we introduce a new object to the pipeline:
Banzai::RenderContext. This object simply wraps two other objects: an
optional Project instance, and an optional User instance. On its own
this wouldn't be very helpful, but a RenderContext can also be used to
associate HTML documents with specific Project instances. This work is
done in Banzai::ObjectRenderer and allows us to reuse as many queries
(and results) as possible.

It also adds a fallback that covers other edge cases that might surface
later.

This could only be possible for users that can create merge requests
within a project.

So they need to be a allowed to create a branch and create a merge request.

By default this was an OPT_IN RPC, and now graduates to OPT_OUT as the
acceptance testing was successful in:
gitaly#1009

This follows the conventions of our [migration
process](https://gitlab.com/gitlab-org/gitaly/blob/master/doc/MIGRATION_PROCESS.md)

By default this was an OPT_IN RPC, and now graduates to OPT_OUT as the
acceptance testing was successful in:
https://gitlab.com/gitlab-org/gitaly/issues/1008

This follows the conventions of our [migration
process](https://gitlab.com/gitlab-org/gitaly/blob/master/doc/MIGRATION_PROCESS.md)

Rename delete_protected_branch ability to push_to_delete_protected_branch to prevent confusion with destroy_protected_branch

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

* Use the correct key prefix
* Clear old cache keys

TODO:

At some point we could remove clearing old cache keys.

This is useful for annotated tags, where the deferenced target is not the same as the tag object. At the moment there is no way to differentiate the two through the tag API.

This change adds a "target" property and leaves the existing "commit" property alone so that existing behaviour is not altered.

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

Also, fixes broken specs

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

- Remove extra method for authorize_admin_project
- Ensure project presence
- Rename 'read_repo' to 'read_repository' to be more verbose