Alfred API abstracts away past and future changes to the Alfresco, across major and minor versions, providing a stable interface to Alfresco on which client-side applications can be built.
It also provides functional grouping of related operations from the Alfresco Public API, and additional endpoints that are not supported by the Alfresco Public API.
Alfred API is a part of the Xenit Open Source Tooling around Alfresco. Xenit is company with a deep expertise and strong team centered around Alfresco. If you'd like to learn more about our tools, services and products, please visit our website.
Full documentation can be found at the project's documentation.
- Common sense trumps all rules.
- For every pull request please extend the CHANGELOG.md.
- Do not make breaking changes since this is an API used by customers. Breaking changes include
adding, changing or removing endpoints or JSON objects used in requests and responses.
- If you are forced to make a breaking change:
- Notify maintainers
- Add a note to the changelog with upgrade instructions
- Notify all customers at the next release
- If you are forced to make a breaking change:
- When working in REST code, please comply to REST HTTP result codes policy outlined in the user guide.
- Prefer unit tests over integration tests to keep builds fast
- Avoid
this.
prefix for consistency (unless the scope is ambiguous). - Follow our coding styleguide and other active procedures.
- alfred-api-interface builds the interface of Alfred API. This part is agnostic of the Alfresco version used.
- alfred-api-rest builds the REST API of Alfred API.
- alfred-api-impl builds the Java code for each version of Alfresco.
- alfresco builds the AMP for each Alfresco version that is the main deliverable for Alfred API. The AMP contains
the JARs of alfred-api-interface and alfred-api-rest.
- alfresco/xx contains the correct properties for each Alfresco version.
- alfred-api-integrationtests-client contains the integration tests for each Alfresco version.
- alfred-api-integrationtests-server contains the Remote-JUnit runner for remote class loading.
- uses java serialization and HTTP for communication.
- We startup a CodeRunnerStandaloneServer, which starts a nanohttpd server, listening on a specific port (4578 by default)
- Using a static appicationContext to reach all necessary beans.
- SRC: https://github.com/ruediste/remote-junit
- uses java serialization and HTTP for communication.
The following command starts up all docker containers required for an Alfresco running Alfred API.
./gradlew :alfred-api-docker:docker-${VERSION}:composeUp --info
Where VERSION
is e.g. 231
.
./gradlew :alfred-api-integrationtests-client:alfresco:${VERSION}:integrationTest
Again, where VERSION
is e.g. 231
.
However, this starts (and afterwards stops) docker containers. This includes starting an Alfresco container, adding a startup time of several minutes. To circumvent this you also run the test on already running containers with for example:
./gradlew -x composeUp -x composeDown :alfred-api-integrationtests-client:alfresco:231:integrationTest -Pprotocol=http -Phost=localhost -Pport=8074
If you only want to run specific tests, you can specify this on the Gradle invocation with a pattern. For example:
./gradlew :alfred-api-integrationtests-client:alfresco:231:integrationTest -x composeDown --tests ContentServiceTestJavaApi.TestContentUrlExists
- Debugging settings are already added by
alfred-api-docker/${VERSION}/debug-extension.docker-compose.yml
, including a portmapping8000:8000
. This file does not get loaded when running in CI. - Prepare your remote debugger in IntelliJ and set breakpoints where you want in your tests (or Alfred API code).
- Run the integration tests (see section above).
- Wait until the container is started and healthy, then attach the debugger.
Again, where VERSION
is e.g. 231
.