If the feature flag `ci_dag_limit_needs` is turned on for a GitLab
instance, a job can now depend on up to ten jobs instead of five

Allison Browne authored 5 years ago and Bob Van Landuyt :neckbeard: committed 5 years ago

Add table to store the sentry event and issue ids
that are related to a gitlab issue

This MR extracts a lot of functinality
from ProjectTreeRestorer into RelationTreeRestorer
 in order to introduce a new GroupTreeRestorer
that is going to utilize this logic as well.

This MR Also changes a few other classes
to make them compatible with groups.

Igor Frenkel authored 5 years ago and Nick Thomas committed 5 years ago

- Update mr id in changelog
- Add new dependency scanning var
- Fix position of var

This commit persists and controls auto stop date for environments

Kerri Miller authored 5 years ago and Grzegorz Bizon committed 5 years ago

.

drew authored 5 years ago and Grzegorz Bizon committed 5 years ago

Allows for including or excluding bridge jobs via `rules:` to control
triggering of downstream (multi-project) Pipelines.

Ethan Urie authored 5 years ago and Jan Provaznik committed 5 years ago

Refactored PagesController and pages API to use new service.

This change limits the number of projects displayed
in GET /groups/:id API, to prevent N+1 in Gitaly

This API can be used to gather information on a project's push mirrors
and their status.

Running any Rake task would show:

ARNING: Sidekiq testing API enabled, but this is not the test
environment.  Your jobs will not go to Redis.

To fix this, we move the `require` statement into the import task.
Now this message only shows when running the import job.

Closes https://gitlab.com/gitlab-org/gitlab/issues/38377

We used to return service password for admin users.
This has been removed as part of a security fix
and this commit removes some leftover of this feature

* API::Helpers::Pagination#paginate only exposes offset pagination
* API::Helpers::Pagination#paginate_and_retrieve! supports both types

I wanted to make it explicit that keyset pagination returns an array
instead of a relation. This does not match the behavior of offset
pagination, hence not mixing those in #paginate.

This is in line with what we document:
https://docs.gitlab.com/ee/api/README.html#status-codes

This aligns with offset pagination, where we also allow up to 100
records max.

This is the squashed together change from
https://gitlab.com/gitlab-org/gitlab/merge_requests/19623
that had been reverted previously.

Based on the refactoring in
https://gitlab.com/gitlab-org/gitlab/merge_requests/21130

It works the same way as https://gitlab.com/gitlab-org/gitlab/merge_requests/14846:

It consists of two parts:

1. Redirecting users to the configured external storage
1. Allowing the external storage to request the static object(s)
   on behalf of the user by means of specific tokens

Related to https://gitlab.com/gitlab-com/gl-infra/scalability/issues/4

User can now specify updated_before and updated_after
params to limit pipelines by last update date

Sean Arnold authored 5 years ago and Jan Provaznik committed 5 years ago

- Squashed commit containing GraphQL changes

Gabriel Mazetto authored 5 years ago and Mayra Cabrera committed 5 years ago

An attachment from an unmigrated project will only be using the legacy
storage when stored locally. Any file on object storage already use
hashed storage format.

Matt Kasa authored 5 years ago and Thong Kuah committed 5 years ago

Relates to https://gitlab.com/gitlab-org/gitlab/merge_requests/20244



Co-Authored-By: Thong Kuah <tkuah@gitlab.com>

Moved pushing the FE feature flag

Now, instead of doing ot from Snippets controller, we do it from
gon_helper to make it shared between all types of snippets

This reverts merge request !21088

It uses the existing cache for the
MergeRequestDiffBatch collection class.

Add a layer of CI configuration on top of the .gitlab-ci.yml
which uses `include:` feature to include a file from repository,
from a remote URL or from a template (Auto-DevOps)

Matthias Käppler authored 5 years ago and Grzegorz Bizon committed 5 years ago

We tested this with a large import and it's a pretty safe change

The method does not only prepare the relation (as the `prepare_relation`
suggests), but also triggers batch counting which in turn executes the
query as a side-effect.

This is being signal-boosted with renaming the method to `prepare!`.

We use it in places where we actually want both (preloading/batch
counting) to happen. In other places (projects API), we actually don't
want that and have the ability to split this into multiple steps now.

The idea here is to make it explicit that at this point we are handling
an array and not a relation anymore.

This separates the "relation builder" from the "batch counting" concern.

`gsub(/\/\z/, '')` removes a leading slash (`/`) not a space

```
> 'https://gitlab.com/'.gsub(/\/\z/, '')
=> "https://gitlab.com"

> 'https://gitlab.com/  '.gsub(/\/\z/, '')
=> "https://gitlab.com/  "
```