Skip to content
This repository was archived by the owner on May 24, 2022. It is now read-only.

developing.md

Latest commit

 

History

History
60 lines (38 loc) · 2.29 KB

developing.md

File metadata and controls

60 lines (38 loc) · 2.29 KB
Raw

Developer documentation

Project structure

ORM uses orm/__main__.py as entry point which does the following:

  • Lists all rule files (according to fileglob argument --orm-rules-path).
  • Validates rule files.
  • Parses rule files into an internal data structure.
  • Renders configuration.

Rule parsing

orm/parser.py is:

  • Listing rule files
  • Parsing rule files into internal data structures.

The internal data structures are fed into the configuration renderers.

Rule validation

Rules are validated in orm/validator.py by doing:

  • ORM schema validation
  • Additional ORM constraints checks

ORM schema validation is performed according to versioned schemas in orm/schemas/. The jsonschema python library is used as validator, with JSON Schema draft 4.

Additional ORM constraints, not covererable by prior validations, are specified in orm/validator.py:validate_rule_constraints.

Configuration renderers

Renderers extends the class RenderOutput in orm/render.py.

Currently there are configuration renderers for HAProxy (orm/renderhaproxy.py) and Varnish (orm/rendervarnish.py).

Tests

Unit tests are located in test/.

Deployment tests are performed by using the rules and tests in orm-rules-tests against the ORM deployment image. HTTP echo servers running in the ORM deployment container are used as backends when needed to verify rule functionality. See the Makefile target deployment-test for details.

Updating ORM dependencies

Run make list-outdated-deps to check for candidates.

Edit install_requires in setup.py.in and run make update-deps to generate requirements.txt.

Updating ORM deployment image

The folder lxd contains tools for building an ORM deployment, packaged into an LXD image. For more information, see lxd/README.md.

Contributing (developer workflow)

  • Fork and create a feature branch.
  • Test and implement the new feature.
  • Rebase to master.
  • Pass tests. make release-test
  • Document changes in README.md, docs/ and CHANGELOG.md
  • Create a merge request.
  • Let maintainers review changes. +1 required for merge to master.
  • Maintainers are responsible for merging.