python-efl/CODING

TODO
====
* Automatically build and publish binary wheels on pypi
  - for py: 36,37,38,39,310
  - for linux x86 and rpi
  - see https://packaging.python.org/guides/publishing-package-distribution-releases-using-github-actions-ci-cd-workflows/
  - see https://github.com/pypa/manylinux
  - also include efl itsef in the wheels? PySide is doing this

* Drop python 2 compatibility
  - Switch to cython3
  - remove all the workarounds for py2 (see strings and enums)

* Improved editors support, at least on VSCode and PyCharm

* Try to implement better the elementary module, removing
  the __init__ module hack (that create problems in the docs)

* Make the examples a first citizen!
  - install them with python efl
  - make them easily runnable
  - show them in the docs
  - show the code inside the examples themself?
  - make the code editable on the fly?


Style
=====

* For indentation, use *four space characters* per level of indentation. Keep
  lines under the 80 chars limit (only exception are the functions definition)

* When comparing C pointers with NULL, use == and != instead of the python
  operator "is". This makes a visual distinction between python and C code.

* For long lines that do not fit in the 80 cols please use only the first
  raccomandation from PEP-8 (Aligned with opening delimiter). Example:
  Yes:
    foo = long_function_name(var_one, var_two,
                             var_three, var_four)
  No:
    foo = long_function_name(
        var_one, var_two,
        var_three, var_four)

  This to keep new code consistent with the rest of the bindings and to
  try to match the style of the C efl code style as much as possible.
  ...also because I found it more readable and I like it more :P -davemds-


Documentation cheatsheet
========================

* Links:

  :class:`List`        (for classes)
  :func:`elm_list_go`  (for functions)
  :attr:`homogeneous`  (for properties)
  :ref:`Elm_List_Mode` (for enums)
  :data:`ELM_LIST_LIMIT` (for enum values)

  :func:`efl.evas.Object.delete`  (for items not in current scope)
  :func:`~efl.evas.Object.delete` (will show it short, just "delete")

* Formatting:

  ``ELM_LIST_SCROLL``  (for enum values, bools and None)


* Versions:

.. versionadded:: 1.15

.. versionchanged:: 1.15
    Description of the change.

.. deprecated:: 1.15
    Use the blah, blah way instead.

* Notes:

    .. seealso:: :py:attr:`mode`

    .. note:: Some text to be noted, blah, blah, blah,
        some more information for this note, etc.

    .. warning:: Same as note, but the box will be red
        some more information for this warning, etc.


Design patterns
===============

 * From "The Zen of Python":

    Beautiful is better than ugly.
    Explicit is better than implicit.
    Simple is better than complex.
    Complex is better than complicated.
    Flat is better than nested.
    Sparse is better than dense.
    Readability counts.
    Special cases aren't special enough to break the rules.
    Although practicality beats purity.
    Errors should never pass silently.
    Unless explicitly silenced.
    In the face of ambiguity, refuse the temptation to guess.
    There should be one-- and preferably only one --obvious way to do it.
    Although that way may not be obvious at first unless you're Dutch.
    Now is better than never.
    Although never is often better than *right* now.
    If the implementation is hard to explain, it's a bad idea.
    If the implementation is easy to explain, it may be a good idea.
    Namespaces are one honking great idea -- let's do more of those!


Tips
====

* cython does automatic dict <-> struct conversion with basic struct members


Release process instructions
============================

* use "api_coverage.py --python elementary" to see the missing bindings

* Version:
  - change versions in efl/__init__.py (ex: 1.9.0)
  - update EFL_MIN_VER in setup.py (if needed)

* Update the ChangeLog file:
  - setup.py build_doc -b changes ...and manually merge from the html file

* Generate the source (tar) and binary (wheel) distributions:
  - make maintainer-clean  (just to be sure nothing is cached)
  - make dist

* Test the generated tarball and wheel:
  - the targz must work by extracting and running: python setup.py install [--user]
  - the targz must be installable using: pip install python-efl-1.9.0.tar.gz
  - the wheel must be installable using: pip install python_efl-1.9.0-xxxx.whl

* Publish the 2 tar archive on e.org:
  - scp tarballs & md5sums to:
    download.enlightenment.org:/srv/web/download.enlightenment.org/rel/bindings/python/
  - update download link on the wiki (www.enlightenment.org/download)

* Publish the .tar.gz archive on pypi:
  - twine upload python-efl-1.9.0.tar.gz [--repository testpypy]
  - TODO: also upload binary wheels for linux py36,37,38,39,310 (see manylinux on pypa)

* Documentation:
  - make doc  (check that inheritance graphs are there)
  - scp the generated html documentation to:
    download.enlightenment.org:/srv/web/docs.enlightenment.org/python-efl/1.9.0/
  - update the 'current' link on the server (ssh)

* Create and push the tag for the release
  - git tag -a v1.9.0 && git push origin v1.9.0

* Create and push the branch for stable backporting
    git branch python-efl-1.9 && git push origin python-efl-1.9

* Announce the release to release@lists.enlightenment.org and
  enlightenment-release@lists.sourceforge.net

* Change versions again in efl/__init__.py (ex: 1.9.99)

for more info:
--------------
 * packaging.python.org
 * phab.enlightenment.org/w/release_procedure/
 * phab.enlightenment.org/w/hosting/ssh/