Add project type for global search from ES

See merge request gitlab-org/gitlab!22821

Enable Code Review Analytics by default

See merge request gitlab-org/gitlab!23285

Fix epic subscriptions not working for subgroups

Closes #197483

See merge request gitlab-org/gitlab!23283

1. Move expectation into the callback we give to `testAction`
2. Uses `jasmine.objectContaining` instead of an off-the-fly custom
matcher.
3. Replaces `stateSubscribed` with simply mutating the shared `state`
object we're creating for each test (it was a little weird to see
`stateSubscribed.subscribes`)

Fix epic dates not updating with subgroups

Closes #197495

See merge request gitlab-org/gitlab!23289

Specify language in fenced code blocks

See merge request gitlab-org/gitlab!23212

Markdownlint expects fenced code blocks to declare a language.

Add language declarations to code blocks

See merge request gitlab-org/gitlab!23217

Markdownlint expects fenced code blocks to declare a language.
This commit adds language declarations to fenced code blocks in
pages in the /administration area.

Added description of issue list limits

Closes #37011

See merge request gitlab-org/gitlab!21807

Describes new method of creating an issue from an epic

See merge request gitlab-org/gitlab!19922

Identify correct Sentry ID for error tracking error detail view operations

See merge request gitlab-org/gitlab!23280

The change in https://gitlab.com/gitlab-org/gitlab/merge_requests/22467
caused the wrong path to be passed to the GraphQL endpoint. The frontend
passed only the immediate path (e.g. charts) instead of the full path
(e.g. gitlab-org/charts).

Closes https://gitlab.com/gitlab-org/gitlab/issues/197495

The change in https://gitlab.com/gitlab-org/gitlab/merge_requests/22872
caused the wrong path to be passed to the GraphQL endpoint. The frontend
passed only the immediate path (e.g. `charts`) instead of the full path
(e.g. `gitlab-org/charts`).

Closes https://gitlab.com/gitlab-org/gitlab/issues/197483

In migrating the error detail view for Sentry errors to GraphQL,
the sentry error id got mis-mapped. This fixes the regression.

Add Code Review Analytics documentation

See merge request gitlab-org/gitlab!23031

Docs: Link to GitLab commit in error detail page

See merge request gitlab-org/gitlab!22395

Support new Azure SCIM pre-setup check

Closes #118558

See merge request gitlab-org/gitlab!22348

12905-c support link reference parsing

Closes #196736 and #12905

See merge request gitlab-org/gitlab!23114

Remove useless attachToDocument usage in ee/spec/frontend

Closes #196750

See merge request gitlab-org/gitlab!23053

Remove useless attachToDocument usage in spec/frontend/environments

Closes #196756

See merge request gitlab-org/gitlab!23048

Remove useless attachToDocument usage in spec/frontend/monitoring

Closes #196758

See merge request gitlab-org/gitlab!23046

Add :sidekiq on migration spec/migrations/20190924152703_migrate_issue_trackers_data_spec.rb

Closes #197482

See merge request gitlab-org/gitlab!23277

API Pagination: Move specific methods out of Base

Closes #37447

See merge request gitlab-org/gitlab!23175

This should allow for workers to be cleared around the spec

Closes https://gitlab.com/gitlab-org/gitlab/issues/197482

Fix analytics tracking for new merge request notes

Closes #197464

See merge request gitlab-org/gitlab!23273

Add retry logic for failures during import

See merge request gitlab-org/gitlab!22265

Rm FF for wrk in progress limits

See merge request gitlab-org/gitlab!23229

Migrate issue trackers data

See merge request gitlab-org/gitlab!18639

When a new note is created for a merge request with no associated
metrics, Sidekiq will error out with `NoMethodError: undefined method
'first_comment_at' for nil:NilClass`.

This commit ensures a merge request metrics exists before attempting to
update it. There is still an outstanding question of why no such metrics
were created, but that can happen for older merge requests.

Closes https://gitlab.com/gitlab-org/gitlab/issues/197464

The only group they share is simple_file_name, so that is now the first
one we check. In the case of a link match, that group is always
populated with a truthy value, so the fall-through is safe.

Link reference_patterns are more specific, so if they overlap we should
prefer the link.