Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Defining a versioning and branching strategy #895

Merged
merged 1 commit into from
May 4, 2021
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
141 changes: 141 additions & 0 deletions versioning-and-branching.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,141 @@

# OSCAL Versioning

The OSCAL project uses [semantic versioning](https://semver.org/spec/v2.0.0.html) to version releases.

Semantic versions are in the form of `MAJOR.MINOR.PATCH`. Given a version number, increment the:

1. MAJOR version when an incompatible model change is made,
1. MINOR version when a new optional feature is added in a backwards compatible manner, and
1. PATCH version when backwards compatible bug fixes are made.

An incompatible model change will occur when all content produced under a given MAJOR version is no longer schema valid with the new model.

A backwards compatible feature change can entail an addition to a model that adds a new construct, perhaps even deprecating an existing construct, but all previously created content is still schema valid with the new model.

# OSCAL Branching

The main branches of this repository are:
- `main` is the current supported, production-ready release.
- `develop` is the current set of development changes for the next release.
- This branch can be considered an integration branch where development code can be tested
- This branch will be merged into `main` when the developed code is ready for release.
- `nist-pages` is the currently deployed website content, which is managed by the CI/CD process.
- `release-*` where `*` matches a MAJOR.MINOR version number.

## Git Setup

To use this strategy, the following Git configuration is needed:

You must do all work in a personal fork of the OSCAL Git repository.

```
git remote add upstream git@github.com:usnistgov/OSCAL.git
```

## Release Branches

A *release branch* is used to stage code in preparation of a new release. Refinements to code can be made in a release branch in preparation for a new release.

Release branches represent production-ready code that is near-ready for release.
- Branched off of `develop`.
- Must be merged back into `develop` as releases are made.
- Must be merged back into `main` to reflect the latest release
- Release branches will be named `release-MAJOR.MINOR`, e.g. release 1.0, release 1.1, release 2.0

### Creating a Release Branch

A release branch can be created by issuing the following Git commands:

```
git checkout -b release-1.2 develop
# TODO: need a method to bump version numbers in metaschemas and content
git commit -a -m "Bumped version number to 1.2"
git push --set-upstream upstream release-1.2
```

### Releasing a Release Branch

Once the release is ready, the release can be made using the following Git commands:

```
git checkout master
git merge --no-ff release-1.2
git tag -a 1.2.0
git push --follow-tags
```

### Releasing a PATCH Revision

Patch releases for a given MAJOR.MINOR version will be marked by annotated tags. This allows the same release branch to be used for multiple PATCH releases.

Once a patch release is ready, the release can be made using the following Git commands:

```
git checkout master
git merge --no-ff release-1.2
git tag -a 1.2.1
git push --follow-tags
```

## Feature Branches

A *feature branch* provides means to integrate a set of features that are a work in progress and the release target of a given set of features is unknown or unpredictable. Work on such a set of features can proceed independent of work targeted at the next release in `develop`.

Feature branches represent major development topics.
- Branched off of `develop`.
- Merged back into `develop` when the feature work is completed.
- Feature branches will be named `feature-*`, where the `*` is a brief dash-separated label describing the feature.

If multiple committers are working on a feature, then each committer must work in a personal branch and submit a PR to the feature branch when their work is complete.

### Creating a Feature Branch

A feature branch can be created by issuing the following Git command:

```
git checkout -b feature-NAME develop
git push --set-upstream upstream feature-NAME
```

where *feature-NAME* will follow the pattern `feature-*`.

### Syncing a Feature Branch with `develop`

It may be necessary to periodically sync a feature branch with the latest in `develop`. You can do this using the following Git commands:

```
# switch to the feature branch
git checkout feature-NAME
# get the latest from upstream
git pull --ff-only upstream feature-NAME
# get the latest from develop
git pull -r upstream develop
git push --force-with-lease upstream feature-NAME
```

### Merging a Feature Branch

The following Git commands will be used to merge a feature branch into `develop`:

```
# switch to the develop branch
git checkout develop
# merge the feature branch
git merge --no-ff feature-myfeature
# delete the feature branch once it is merged
git branch -d feature-myfeature
# push the branch to the upstream repository
git push upstream develop
```

## Personal Working Branches

All individual work will be done in branches in a personal fork of the `upstream` repository.

Personal branches should be named using either:

- `<issue #>-brief-dashed-name` (preferred)
- `wip-brief-dashed-name`

Once work is complete on a personal branch, the branch should be interactively rebased to tidy any commits. Then a PR should be opened against the target `feature-*` branch or the `develop` branch if the changes are to be included in the next release.