Some Jenkins plugins depend on other plugins, and their configuration section is a mix of both plugins. For Jenkins Job Builder, that means reusing one component directly from another one. Driving the generation of XML markup is the job of Base._dispatch. Unfortunately, components do not have access to their module object, and even if their could, _dispatch would still be a non-public method. Refactor Base._dispatch into ModuleRegistry.dispatch, which can be used from any place where the parser is available. Base and ModuleRegistry are extended so that the registry can discover which entry point must be used for each module, if appropriate. ModuleRegistry.dispatch signature can be simplified by dropping component_list_type parameter. Change-Id: Ie9d090817d0c2d464745b5634a22d3cea6a47ab1 Reviewed-on: https://review.openstack.org/26051 Reviewed-by: James E. Blair <corvus@inaugust.com> Reviewed-by: Jeremy Stanley <fungi@yuggoth.org> Approved: Clark Boylan <clark.boylan@gmail.com> Reviewed-by: Clark Boylan <clark.boylan@gmail.com> Tested-by: Jenkins
2.6 KiB
Extending
Jenkins Job Builder is quite modular. It is easy to add new attributes to existing components, a new module to support a Jenkins plugin, or include locally defined methods to deal with an idiosyncratic build system.
XML Processing
Most of the work of building XML from the YAML configuration file is
handled by individual functions that implement a single characteristic.
For example, see the jenkins_jobs/modules/builders.py
file
for the Python module that implements the standard Jenkins builders. The
shell
function at the top of the file implements the
standard Execute a shell build step. All
of the YAML to XML functions in Jenkins Job Builder have the same
signature:
- arg YAMLParser parser
the jenkins jobs YAML parser
- arg Element xml_parent
this attribute's parent XML element
- arg dict data
the YAML data structure for this attribute and below
The function is expected to examine the YAML data structure and create new XML nodes and attach them to the xml_parent element. This general pattern is applied throughout the included modules.
Modules
Nearly all of Jenkins Job Builder is implemented in modules. The main program has no concept of builders, publishers, properties, or any other aspects of job definition. Each of those building blocks is defined in a module, and due to the use of setuptools entry points, most modules are easily extensible with new components.
To add a new module, define a class that inherits from :pyjenkins_jobs.modules.base.Base
, and add it to the
jenkins_jobs.modules
entry point in your setup.py.
jenkins_jobs.modules.base.Base
Components
Most of the standard modules supply a number of components, and it's
easy to provide your own components for use by those modules. For
instance, the Builders module provides several builders, such as the
shell builder as well as the trigger_builds builder. If you wanted to add a
new builder, all you need to do is write a function that conforms to the
Component Interface <component_interface>
, and
then add that function to the appropriate entry point (via a setup.py
file).
Module Registry
All modules and their associated components are registered in the module registry. It can be accessed either from modules via the registry field, or via the parser parameter of components.
jenkins_jobs.builder.ModuleRegistry