gitlab-org--gitlab-foss/PROCESS.md

22 KiB

GitLab core team & GitLab Inc. contribution process


Table of Contents generated with DocToc


Purpose of describing the contributing process

Below we describe the contributing process to GitLab for two reasons:

  1. Contributors know what to expect from maintainers (possible responses, friendly treatment, etc.)
  2. Maintainers know what to expect from contributors (use the latest version, ensure that the issue is addressed, friendly treatment, etc.).

Common actions

Merge request coaching

Several people from the GitLab team are helping community members to get their contributions accepted by meeting our Definition of done.

What you can expect from them is described at https://about.gitlab.com/job-families/expert/merge-request-coach/.

Milestones on community contribution issues

The milestone of an issue that is currently being worked on by a community contributor should not be set to a named GitLab milestone (e.g. 11.7, 11.8), until the associated merge request is very close to being merged, and we will likely know in which named GitLab milestone the issue will land. There are many factors that influence when a community contributor finishes an issue, or even at all. So we should set this milestone only when we have more certainty.

Note this only applies to issues currently assigned to community contributors. For issues assigned to GitLabbers, we are ambitious in assigning milestones to issues.

Assigning issues

If an issue is complex and needs the attention of a specific person, assignment is a good option but assigning issues might discourage other people from contributing to that issue. We need all the contributions we can get so this should never be discouraged. Also, an assigned person might not have time for a few weeks, so others should feel free to takeover.

Be kind

Be kind to people trying to contribute. Be aware that people may be a non-native English speaker, they might not understand things or they might be very sensitive as to how you word things. Use Emoji to express your feelings (heart, star, smile, etc.). Some good tips about code reviews can be found in our Code Review Guidelines.

Feature freeze on the 7th for the release on the 22nd

After 7th at 23:59 (Pacific Time Zone) of each month, stable branch and RC1 of the upcoming release (to be shipped on the 22nd) is created and deployed to GitLab.com. The stable branch is frozen at the most recent "qualifying commit" on master. A "qualifying commit" is one that is pushed before the feature freeze cutoff time and that passes all CI jobs (green pipeline).

Merge requests may still be merged into master during this period, but they will go into the next release, unless they are manually cherry-picked into the stable branch.

By freezing the stable branches 2 weeks prior to a release, we reduce the risk of a last minute merge request potentially breaking things.

Any release candidate that gets created after this date can become a final release, hence the name release candidate.

Feature flags

Merge requests that make changes hidden behind a feature flag, or remove an existing feature flag because a feature is deemed stable, may be merged (and picked into the stable branches) up to the 19th of the month. Such merge requests should have the ~"feature flag" label assigned, and don't require a corresponding exception request to be created.

A level of common sense should be applied when deciding whether to have a feature behind a feature flag off or on by default.

The following guideliness can be applied to help make this decision:

  • If the feature is not fully ready or functioning, the feature flag should be disabled by default.
  • If the feature is ready but there are concerns about performance or impact, the feature flag should be enabled by default, but disabled via chatops before deployment on GitLab.com environments. If the performance concern is confirmed, the final release should have the feature flag disabled by default.
  • In most other cases, the feature flag can be enabled by default.

For more information on rolling out changes using feature flags, read through the documentation.

In order to build the final package and present the feature for self-hosted customers, the feature flag should be removed. This should happen before the 22nd, ideally at least 2 days before. That means MRs with feature flags being picked at the 19th would have a quite tight schedule, so picking these earlier is preferable.

While rare, release managers may decide to reject picking a change into a stable branch, even when feature flags are used. This might be necessary if the changes are deemed problematic, too invasive, or there simply isn't enough time to properly test how the changes behave on GitLab.com.

Between the 1st and the 7th

These types of merge requests for the upcoming release need special consideration:

  • Large features: a large feature is one that is highlighted in the kick-off and the release blogpost; typically this will have its own channel in Slack and a dedicated team with front-end, back-end, and UX.
  • Small features: any other feature request.

It is strongly recommended that large features be with a maintainer by the 1st. This means that:

  • There is a merge request (even if it's WIP).
  • The person (or people, if it needs a frontend and backend maintainer) who will ultimately be responsible for merging this have been pinged on the MR.

It's OK if merge request isn't completely done, but this allows the maintainer enough time to make the decision about whether this can make it in before the freeze. If the maintainer doesn't think it will make it, they should inform the developers working on it and the Product Manager responsible for the feature.

The maintainer can also choose to assign a reviewer to perform an initial review, but this way the maintainer is unlikely to be surprised by receiving an MR later in the cycle.

It is strongly recommended that small features be with a reviewer (not necessarily a maintainer) by the 3rd.

Most merge requests from the community do not have a specific release target. However, if one does and falls into either of the above categories, it's the reviewer's responsibility to manage the above communication and assignment on behalf of the community member.

Every new feature or change should be shipped with its corresponding documentation in accordance with the documentation process and structure.

What happens if these deadlines are missed?

If a small or large feature is not with a maintainer or reviewer by the recommended date, this does not mean that maintainers or reviewers will refuse to review or merge it, or that the feature will definitely not make it in before the feature freeze.

However, with every day that passes without review, it will become more likely that the feature will slip, because maintainers and reviewers may not have enough time to do a thorough review, and developers may not have enough time to adequately address any feedback that may come back.

A maintainer or reviewer may also determine that it will not be possible to finish the current scope of the feature in time, but that it is possible to reduce the scope so that something can still ship this month, with the remaining scope moving to the next release. The sooner this decision is made, in conversation with the Product Manager and developer, the more time there is to extract that which is now out of scope, and to finish that which remains in scope.

For these reasons, it is strongly recommended to follow the guidelines above, to maximize the chances of your feature making it in before the feature freeze, and to prevent any last minute surprises.

On the 7th

Merge requests should still be complete, following the definition of done.

Feature merge requests

If a merge request is not ready, but the developers and Product Manager responsible for the feature think it is essential that it is in the release, they can ask for an exception in advance. This is preferable to merging something that we are not confident in, but should still be a rare case: most features can be allowed to slip a release.

All Community Edition merge requests from GitLab team members merged on the freeze date (the 7th) should have a corresponding Enterprise Edition merge request, even if there are no conflicts. This is to reduce the size of the subsequent EE merge, as we often merge a lot to CE on the release date. For more information, see Automatic CE->EE merge and Guidelines for implementing Enterprise Edition features.

Documentation merge requests

Documentation is part of the product and must be shipped with the feature.

The single exception for the feature freeze is documentation, and it can be left to be merged up to the 14th if:

  • There is a follow-up issue to add documentation.
  • It is assigned to the developer writing documentation for this feature, and they are aware of it.
  • It is in the correct milestone, with the labels ~Documentation, ~Deliverable, ~missed-deliverable, and "pick into X.Y" applied.
  • It must be reviewed and approved by a technical writer.

For more information read the process for documentation shipped late.

After the 7th

Once the stable branch is frozen, the only MRs that can be cherry-picked into the stable branch are:

  • Fixes for regressions where the affected version xx.x in regression:xx.x is the current release. See Managing bugs section.
  • Fixes for security issues
  • Fixes or improvements to automated QA scenarios
  • Documentation updates for changes in the same release
  • New or updated translations (as long as they do not touch application code)
  • Changes that are behind a feature flag and have the ~"feature flag" label

During the feature freeze all merge requests that are meant to go into the upcoming release should have the correct milestone assigned and the Pick into X.Y label where X.Y is equal to the milestone, so that release managers can find and pick them. Merge requests without this label will not be picked into the stable release.

For example, if the upcoming release is 10.2.0 you will need to set the Pick into 10.2 label.

Fixes marked like this will be shipped in the next RC (before the 22nd), or the next patch release.

If a merge request is to be picked into more than one release it will need one Pick into X.Y label per release where the merge request should be back-ported to. For example:

  • Pick into 10.1
  • Pick into 10.0
  • Pick into 9.5

Asking for an exception

If you think a merge request should go into an RC or patch even though it does not meet these requirements, you can ask for an exception to be made.

Check this guide about how to open an exception request before opening one.

Bugs

A ~bug is a defect, error, failure which causes the system to behave incorrectly or prevents it from fulfilling the product requirements.

The level of impact of a ~bug can vary from blocking a whole functionality or a feature usability bug. A bug should always be linked to a severity level. Refer to our severity levels

Whether the bug is also a regression or not, the triage process should start as soon as possible. Ensure that the Engineering Manager and/or the Product Manager for the relative area is involved to prioritize the work as needed.

Regressions

A ~regression implies that a previously verified working functionality no longer works. Regressions are a subset of bugs. We use the ~regression label to imply that the defect caused the functionality to regress. The label tells us that something worked before and it needs extra attention from Engineering and Product Managers to schedule/reschedule.

The regression label does not apply to ~bugs for new features for which functionality was never verified as working. These, by definition, are not regressions.

A regression should always have the regression:xx.x label on it to designate when it was introduced.

Regressions should be considered high priority issues that should be solved as soon as possible, especially if they have severe impact on users.

Managing bugs

Prioritization: We give higher priority to regressions on features that worked in the last recent monthly release and the current release candidates. The two scenarios below can bypass the exception request in the release process, where the affected regression version matches the current monthly release version.

  • A regression which worked in the Last monthly release
    • Example: In 11.0 we released a new feature X that is verified as working. Then in release 11.1 the feature no longer works, this is regression for 11.1. The issue should have the regression:11.1 label.
    • Note: When we say the last recent monthly release, this can refer to either the version currently running on GitLab.com, or the most recent version available in the package repositories.
  • A regression which worked in the Current release candidates
    • Example: In 11.1-RC3 we shipped a new feature which has been verified as working. Then in 11.1-RC5 the feature no longer works, this is regression for 11.1. The issue should have the regression:11.1 label.
    • Note: Because GitLab.com runs release candidates of new releases, a regression can be reported in a release before its 'official' release date on the 22nd of the month.

When a bug is found:

  1. Create an issue describing the problem in the most detailed way possible.
  2. If possible, provide links to real examples and how to reproduce the problem.
  3. Label the issue properly, using the team label, the subject label and any other label that may apply in the specific case
  4. Notify the respective Engineering Manager to evaluate and apply the Severity label and Priority label. The counterpart Product Manager is included to weigh-in on prioritization as needed.
  5. If the ~bug is NOT a regression:
    1. The Engineering Manager decides which milestone the bug will be fixed. The appropriate milestone is applied.
  6. If the bug is a ~regression:
    1. Determine the release that the regression affects and add the corresponding regression:xx.x label.
      1. If the affected release version can't be determined, add the generic ~regression label for the time being.
    2. If the affected version xx.x in regression:xx.x is the current release, it's recommended to schedule the fix for the current milestone.
      1. This falls under regressions which worked in the last release and the current RCs. More detailed explanations in the Prioritization section above.
    3. If the affected version xx.x in regression:xx.x is older than the current release
      1. If the regression is an ~S1 severity, it's recommended to schedule the fix for the current milestone. We would like to fix the highest severity regression as soon as we can.
      2. If the regression is an ~S2, ~S3 or ~S4 severity, the regression may be scheduled for later milestones at the discretion of the Engineering Manager and Product Manager.

Release retrospective and kickoff

Copy & paste responses

Improperly formatted issue

Thanks for the issue report. Please reformat your issue to conform to the
[contributing guidelines](https://docs.gitlab.com/ee/development/contributing/issue_workflow.html#issue-tracker-guidelines).

Issue report for old version

Thanks for the issue report but we only support issues for the latest stable version of GitLab.
I'm closing this issue but if you still experience this problem in the latest stable version,
please open a new issue (but also reference the old issue(s)).
Make sure to also include the necessary debugging information conforming to the issue tracker
guidelines found in our [contributing guidelines](https://docs.gitlab.com/ee/development/contributing/issue_workflow.html#issue-tracker-guidelines).

Support requests and configuration questions

Thanks for your interest in GitLab. We don't use the issue tracker for support
requests and configuration questions. Please check our
[getting help](https://about.gitlab.com/getting-help/) page to see all of the available
support options. Also, have a look at the [contribution guidelines](https://docs.gitlab.com/ee/development/contributing/index.html)
for more information.

Code format

Please use \`\`\` to format console output, logs, and code as it's very hard to read otherwise.

Issue fixed in newer version

Thanks for the issue report. This issue has already been fixed in newer versions of GitLab.
Due to the size of this project and our limited resources we are only able to support the
latest stable release as outlined in our [contributing guidelines](https://docs.gitlab.com/ee/development/contributing/issue_workflow.html).
In order to get this bug fix and enjoy many new features please
[upgrade](https://gitlab.com/gitlab-org/gitlab-ce/tree/master/doc/update).
If you still experience issues at that time please open a new issue following our issue
tracker guidelines found in the [contributing guidelines](https://docs.gitlab.com/ee/development/contributing/issue_workflow.html#issue-tracker-guidelines).

Improperly formatted merge request

Thanks for your interest in improving the GitLab codebase!
Please update your merge request according to the [contributing guidelines](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/doc/development/contributing/merge_request_workflow.md#merge-request-guidelines).

Accepting merge requests

Is there an issue on the
[issue tracker](https://gitlab.com/gitlab-org/gitlab-ce/issues) that is
similar to this? Could you please link it here?
Please be aware that new functionality that is not marked
[`Accepting merge requests`](https://docs.gitlab.com/ee/development/contributing/issue_workflow.html#label-for-community-contributors)
might not make it into GitLab.

Only accepting merge requests with green tests

We can only accept a merge request if all the tests are green. I've just
restarted the build. When the tests are still not passing after this restart and
you're sure that is does not have anything to do with your code changes, please
rebase with master to see if that solves the issue.