Compare commits

...

2 Commits

Author SHA1 Message Date
Sebastian Crane 0a939da401 Document the release process in CONTRIBUTING.md
This commit adds documentation in the CONTRIBUTING.md file about how
releases of matchbot are made and published on Maven Central, and how
the version number changes between releases are determined.

Signed-off-by: Sebastian Crane <seabass-labrax@gmx.com>
2022-04-24 18:52:39 +01:00
Sebastian Crane 93306d6d34 Document the build system in CONTRIBUTING.md
This commit adds documentation in the CONTRIBUTING.md file about how to
generate POM, JAR and uberjar files using matchbot's build system.

Signed-off-by: Sebastian Crane <seabass-labrax@gmx.com>
2022-04-24 18:48:30 +01:00
1 changed files with 34 additions and 0 deletions

View File

@ -40,3 +40,37 @@ For example:
> `;; SPDX-FileCopyrightText: 2022 Joe Bloggs <joe@example.com>`
If the copyright to your contributions is held by your employer, put your employer's name in brackets after your own name.
## Build system
### Building a POM file
The POM file lists the dependencies needed to run `matchbot` as well as some additional information that can help people learn more about `matchbot`.
To generate a POM file, run `clojure -T:build pom`; you should find the generated `POM.xml` file in the `target/` directory.
### Building a JAR file
A JAR file contains all the source code of `matchbot` in a form that the JVM can load and pass to the Clojure compiler to run.
To generate a JAR file, run `clojure -T:build jar`; again, you should find the JAR file called something like `matchbot-x.x.x.jar` in the `target/` directory.
### Building an uberjar
An uberjar is much like a normal JAR file, but comes with all the dependencies of `matchbot` bundled in it.
This means that it can run directly on the JVM without Clojure being installed (it contains a copy of the Clojure compiler itself).
To generate an uberjar, run `clojure -T:build uber`; you should find the uberjar called something like `matchbot-x.x.x-standalone.jar` in the `target/` directory.
Please note that if you distribute an uberjar, you must not only comply with the licence of `matchbot`, but also the licences of all `matchbot`'s dependencies, both transitive and intransitive.
## Versioning and release process
### Semantic Versioning
`matchbot` uses [version 2.0.0 of the Semantic Versioning](https://semver.org/spec/v2.0.0.html) scheme, but there is still ambiguity in what exactly comprises the 'Public API' (used for determining the right part of the version to increment) for something like `matchbot`.
This 'Public API' is defined for `matchbot` as everything that is accessible by the end user or administrator of a `matchbot` instance.
For example, a change that requires the configuration file to be updated warrants a major version increment because it effects the administrator; however, a change to the structure of the internal namespaces would only require a patch level version increment because it doesn't affect either the administrator or the end user.
### Release process
At some point after a new feature has been added to `matchbot` or a bug has been fixed, a release will be made.
Once a suitable version increment for the type of changes has been determined, the `CHANGELOG.md` file at the root of the repository will be updated with release notes documenting the changes made in that version.
Then, a JAR file and a POM file will be produced using the build system (see above for more information), signed using GPG and finally uploaded to [Maven Central](https://central.sonatype.org/).
Currently this process is done by [Sebastian Crane](https://git.libregaming.org/seabass); if you would like to help with making releases, please familiarise yourself with the process (you can try everything locally except upload to Maven Central) and get in contact! 😀