This page contains information on how we make releases of j5 and what the process for releasing is.
Every version that will be released, with the exception of hotfix releases, will be added as a milestone on GitHub, such that one can see at a glance what work needs to be done before those features are released. All issues with the exception of patches are likely to be added to a milestone so that we know when it will be released to end users.
We release j5 to two major locations:
The most important of these two locations is PyPI, as this will allow users to specify j5 as a dependency for their API and pip will be able to resolve and download our package, and our dependencies also.
The release should also be created as a ‘release’ on GitHub, with a git tag, version number and description of the changes that have been made since the previous release. Any binary files associated with the release, such as wheels, should also be uploaded to GitHub at this point. It should be ensured that these binaries match those that are uploaded to PyPI.
As a general rule, j5 will follow Semantic Versioning.
During early development of j5, we will be using version numbers of the format 0.y.z, where y increments when new features are added and z increments when a patch version is released. During this phase of development, the API is considered to be unstable and subject to change.
After early development is finished, j5 will use a combination of Semantic Versioning and ideas taken from Git Flow. In particular, the concept of release branches for major versions and having multiple major and minor versions in maintenance at any one time. Those who are running a competition should never ship an increment to the major version number during a competition cycle as this will break the code of their teams. All versions where the major version number is greater than 0 should be considered to be stable and will undergo additional testing before release.