Contributing
Bug reports and code and documentation patches are greatly appreciated. You can also help by using the development version of todoman and reporting any bugs you might encounter here.
All participants must follow the pimutils Code of Conduct.
Before working on a new feature or a bug, please browse existing issues to see whether it has been previously discussed. If the change in question is a bigger one, it’s always good to open a new issue to discuss it before your starting working on it.
Hacking
Runtime dependencies are listed in setup.py
. We recommend that you use
virtualenv to make sure that no additional dependencies are required without
them being properly documented.
Run pip install -e .
to install todoman and its dependencies into a
virtualenv.
We use pre-commit
to run style and convention checks. Run pre-commit
install`
to install our git-hooks. These will check code style and inform you
of any issues when attempting to commit. This will also run black
to
reformat code that may have any inconsistencies.
Commits should follow Git Commit Guidelines whenever possible, including rewriting branch histories to remove any noise, and using a imperative present tense for commit summary messages (50 characters maximum).
All commits should pass all tests to facilitate bisecting in future.
An overview of the Todo lifecycle
This is a brief overview of the life cycles of todos (from the apps point of view) as they are read from disk, displayed, and or saved again.
When the app starts, it will read all todos from disk, and initialize from the
cache any further display (either list
, show
, edit
, etc) is then
done reading from the cache, which only contains the fields with which we
operate. This stage also assigns the id numbers to each todo.
When a Todo is edited, the entire cycle is:
File is read from disk and cached (if not already cached).
A Todo object is created by reading the cache.
If edition is interactive, show the UI now.
No matter how the edition occurs, apply changes to the Todo object.
- Start saving process:
Read file from disk (as a VTodo object).
Apply changes from fields to the VTodo object.
Write to disk.
The main goal of this is to simplify how many conversions we have. If we read from disk to the editor, we’d need an extra VTodo->Todo conversion code that skips the cache.
Running and testing locally
The easiest way to run tests, it to install tox
, and then simply run
tox
. By default, several python versions and environments are tested. If
you want to run a specific one use tox -e ENV
, where ENV
should be one
of the environments listed by tox -l
.
See the tox documentation for further details.
To run your modified copy of todoman
without installing it, it’s
recommended you set up a virtualenv, and run pip install -e .
to install
your checked-out copy into it (this’ll make todo
run your local copy while
the virtualenv is active).
Patch review checklist
Please follow this checklist when submitting new PRs (or reviewing PRs by others):
CI will automatically check these for us:
Do all tests pass?
Does the documentation build?
Do all linting and style checks pass?
Please keep an eye open for these other items:
Packaging
We appreciate distributions packaging todoman. Packaging should be relatively straightforward following usual Python package guidelines. We recommend that you git-clone tags, and build from source, since these tags are GPG signed.
Dependencies are listed in setup.py
. Please also try to include the
extras dependencies as optional dependencies (or what maps best for your
distribution).
You’ll need to run python setup.py build
to generate the
todoman/version.py
file which is necessary at runtime.
We recommend that you include the Man page in distribution packages. You can build this by running:
sphinx-build -b man docs/source docs/build/man
The man page will be saved as docs/build/man/todo.1.
Generating the man pages requires that todoman and its doc dependencies (see
requirements-docs.txt
) are either installed, or in the current
PYTHONPATH
.