When passing start_branch on committing from the WebIDE, it's possible
that the branch has changed since editing started, which results in the
change being applied on top of the latest commit in the branch and
overwriting the new changes.

By passing the start_sha instead we can make sure that the change is
applied on top of the commit which the user started editing from.

Whenever we use the rugged implementation, we are going straight to disk
so we want to bypass the disk access check.

Log memory stats after running each spec file and compile the report.

Uses Rails transactional tests instead of DatabaseCleaner
transaction strategy because that doesn't work with JS tests

Loading `ApplicationSetting` from Redis was responsible for at least 50%
of the CPU load of the Redis cluster on GitLab.com. Since these values
generally don't change very much, we can load this from the database and
cache it in memory, skipping Redis altogther. We use
`ActiveSupport::Cache::MemoryStore` as a drop-in replacement for
`RedisCacheStore` even though we probably don't need synchronized access
within `Thread.current`.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/63977

When the SQL query cache is active, the SELECT query for finding
projects to apply service templates returns the same values. This causes
an infinite loop because even though bulk INSERT queries are made, the
cached results never reflect that progress.  To fix this, we call
`Project.uncached` around the query to ensure new data is retrieved.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/63595

This reverts commit 744f1f2e, reversing
changes made to c4d930e5.

This reverts merge request !27514

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

Since we're skipping the example in a before hook, the @_subscribers
instance variable that is normally set in the setup phase of the test by
https://github.com/rails/rails-controller-testing/blob/21014e48be124dd8d0af250b79f682519d445015/lib/rails/controller/testing/template_assertions.rb#L21
isn't set but the teardown phase of the example still happens and tries
to iterate over the array of subcribers: https://github.com/rails/rails-controller-testing/blob/21014e48be124dd8d0af250b79f682519d445015/lib/rails/controller/testing/template_assertions.rb#L54

Using an around hook allows the @_subscribers instance variable to be
set.
Signed-off-by: Rémy Coutable <remy@rymai.me>

Attempt to debug https://gitlab.com/gitlab-org/gitlab-ce/issues/60953

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

It was 2*4 => 8 times before, so 4 times is still half retries

In both e2e QA tests and unit tests, reduce the number of retires
to 2 (i.e., 1 initial and one retry)

Add no-op license helper to CE to make it compatible with EE codebase.
Helper will be overwritten in EE to allow for CE specs to be run under
EE codebase.

https://gitlab.com/gitlab-org/gitlab-ee/issues/6539

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

This brings back some of the changes in
https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/20339.

For users using Gitaly on top of NFS, accessing the Git data directly
via Rugged is more performant than Gitaly. This merge request introduces
the feature flag `rugged_find_commit` to activate Rugged paths.

There are also Rake tasks `gitlab:features:enable_rugged` and
`gitlab:features:disable_rugged` to enable/disable these feature
flags altogether.

Part of four Rugged changes identified in
https://gitlab.com/gitlab-org/gitlab-ce/issues/57317.

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

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

- we now use the hierarchy class also for epics
- also rename supports_nested_groups? into supports_nested_objects?
  - move it to a concern

migration tests.

We want to run CI with rails 4 for a short-term (until we are sure that
we will ship with rails 5). The problem is that rails 4 can not handle
rails 5 schema.rb properly - specifically `t.index` directive can not
handle multiple indexes on the same column.

Because combination of rails 4 + rails 5 schema will be used
only in CI for a short-term, we can just ignore these incompatibility
failures. This patch adds `rails5` helper for specs.

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

Move fixture_path out of support/rspec file which is used by fast_spec_helper, so that fast specs can be run on its own again.

Move `fixture_path` config together with use_transactional_fixtures and
use_instantiated_fixtures. These settings are all added by rspec-rails

https://github.com/rspec/rspec-rails/blob/8c6c9590b94916199950dc8a91a9741d3be30c7c/lib/rspec/rails/configuration.rb#L65-L67

We don't test any specific string generated by it yet, so there is no
point in keeping it enabled when we are disabling it everywhere we test its output.

We need this because `expect_any_instance_of` doesn't work on
prepended models. Now we could use the same code between CE/EE

Adding the :prometheus tag to an rspec test will clear out
memory-mapped files and reset the registry.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/39968