Skip to content

RELEASE.md

Latest commit

 

History

History
33 lines (24 loc) · 2.83 KB

RELEASE.md

File metadata and controls

33 lines (24 loc) · 2.83 KB
Raw

Releasing function-agent-java

Step-by-step instructions for publishing a release. HANDBOOK.md (Release process) explains how the tooling works.

Before you start

  • Everything changed since the previous release is reviewed: the code for correctness, and the documentation and changelog for accuracy and brevity.
  • bash (Git Bash on Windows), Maven, Java 21, and changefrog (npm install -g changefrog), which writes the version section of CHANGELOG.md.
  • Push access to origin (https://gitlab.ilabt.imec.be/KNoWS/fno/proc/function-component).
  • You are on development, up to date with origin/development, with a clean working tree.
  • The tests pass: mvn verify.
  • ## Unreleased in CHANGELOG.md lists everything since the last release.
  • Pick the version X.Y.Z with Semantic Versioning. pom.xml holds the next version as <version>-SNAPSHOT: release that version, or a higher minor or major version when the changelog has new features or breaking changes.

Release

  1. Run ./bump-version.sh <version> and answer y to both questions. The script
    • sets the version in pom.xml and the version in README.md;
    • turns ## Unreleased into the version section of CHANGELOG.md;
    • commits "Update version to ", pushes development, and creates and pushes the tag v<version>;
    • moves pom.xml to the next patch -SNAPSHOT, and commits and pushes "Prepare for next development cycle".
  2. Move main to the release: git push origin v<version>^{commit}:main. main always points at the latest release; the push succeeds only as a fast-forward.
  3. The tag pipeline (https://gitlab.ilabt.imec.be/KNoWS/fno/proc/function-component/-/pipelines) builds the release with the release Maven profile, signs it and deploys it to Maven Central. The new version then appears at https://repo1.maven.org/maven2/be/ugent/idlab/knows/function-agent-java/ (this can take up to an hour).

After the release

  • GitLab mirrors the branches and tags to GitHub (https://github.com/FnOio/function-agent-java); check that the tag is there.
  • Update function-agent.version in MappingWeaver-java's pom.xml and the function-agent-java version in rmlmapper-java's pom.xml.

When something goes wrong

  • The script stops at the first failing command. When it stops before pushing, fix the cause, discard its changes (git reset --hard origin/development) and run it again. When it stops after pushing development, finish by hand: create the tag v<version> on the version commit if it is missing, push it (git push origin v<version>), then set the next -SNAPSHOT (mvn versions:set -DnewVersion=<next>-SNAPSHOT -DgenerateBackupPoms=false), commit and push.
  • Once the tag is pushed, keep it: fix the cause and retry the failed pipeline job. When the released code itself is broken, release the next patch version.