Fixed #18436 -- Updated contributing docs for git.

Most of the credit for this large patch goes to Anssi Kääriäinen.
Many thanks to all the people who contributed to the discussion.
This commit is contained in:
Aymeric Augustin 2012-06-07 18:48:29 +02:00
parent 1a412dda62
commit 90fb6a4648
12 changed files with 626 additions and 483 deletions

View file

@ -1,171 +0,0 @@
=============
Branch policy
=============
In general, the trunk must be kept stable. People should be able to run
production sites against the trunk at any time. Additionally, commits to trunk
ought to be as atomic as possible -- smaller changes are better. Thus, large
feature changes -- that is, changes too large to be encapsulated in a single
patch, or changes that need multiple eyes on them -- must happen on dedicated
branches.
This means that if you want to work on a large feature -- anything that would
take more than a single patch, or requires large-scale refactoring -- you need
to do it on a feature branch. Our development process recognizes two options
for feature branches:
1. Feature branches using a distributed revision control system like
Git_, Mercurial_, Bazaar_, etc.
If you're familiar with one of these tools, this is probably your best
option since it doesn't require any support or buy-in from the Django
core developers.
However, do keep in mind that Django will continue to use Subversion
for the foreseeable future, and this will naturally limit the
recognition of your branch. Further, if your branch becomes eligible
for merging to trunk you'll need to find a core developer familiar
with your DVCS of choice who'll actually perform the merge.
If you do decided to start a distributed branch of Django and choose to
make it public, please add the branch to the `Django branches`_ wiki
page.
2. Feature branches using SVN have a higher bar. If you want a branch
in SVN itself, you'll need a "mentor" among the :doc:`core committers
</internals/committers>`. This person is responsible for actually
creating the branch, monitoring your process (see below), and
ultimately merging the branch into trunk.
If you want a feature branch in SVN, you'll need to ask in
`django-developers`_ for a mentor.
.. _git: http://git-scm.com/
.. _mercurial: http://mercurial.selenic.com/
.. _bazaar: http://bazaar.canonical.com/
.. _django branches: https://code.djangoproject.com/wiki/DjangoBranches
Branch rules
------------
We've got a few rules for branches born out of experience with what makes a
successful Django branch.
DVCS branches are obviously not under central control, so we have no way of
enforcing these rules. However, if you're using a DVCS, following these rules
will give you the best chance of having a successful branch (read: merged back
to trunk).
Developers with branches in SVN, however, **must** follow these rules. The
branch mentor will keep on eye on the branch and **will delete it** if these
rules are broken.
* Only branch entire copies of the Django tree, even if work is only
happening on part of that tree. This makes it painless to switch to a
branch.
* Merge changes from trunk no less than once a week, and preferably every
couple-three days.
In our experience, doing regular trunk merges is often the difference
between a successful branch and one that fizzles and dies.
If you're working on an SVN branch, you should be using `svnmerge.py`_
to track merges from trunk.
* Keep tests passing and documentation up-to-date. As with patches,
we'll only merge a branch that comes with tests and documentation.
.. _svnmerge.py: http://www.orcaware.com/svn/wiki/Svnmerge.py
Once the branch is stable and ready to be merged into the trunk, alert
`django-developers`_.
After a branch has been merged, it should be considered "dead"; write access
to it will be disabled, and old branches will be periodically "trimmed."
To keep our SVN wrangling to a minimum, we won't be merging from a given
branch into the trunk more than once.
Using branches
--------------
To use a branch, you'll need to do two things:
* Get the branch's code through Subversion.
* Point your Python ``site-packages`` directory at the branch's version of
the ``django`` package rather than the version you already have
installed.
Getting the code from Subversion
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To get the latest version of a branch's code, check it out using Subversion:
.. code-block:: bash
svn co https://code.djangoproject.com/svn/django/branches/<branch>/
...where ``<branch>`` is the branch's name. See the `list of branch names`_.
Alternatively, you can automatically convert an existing directory of the
Django source code as long as you've checked it out via Subversion. To do the
conversion, execute this command from within your ``django`` directory:
.. code-block:: bash
svn switch https://code.djangoproject.com/svn/django/branches/<branch>/
The advantage of using ``svn switch`` instead of ``svn co`` is that the
``switch`` command retains any changes you might have made to your local copy
of the code. It attempts to merge those changes into the "switched" code. The
disadvantage is that it may cause conflicts with your local changes if the
"switched" code has altered the same lines of code.
(Note that if you use ``svn switch``, you don't need to point Python at the
new version, as explained in the next section.)
.. _list of branch names: https://code.djangoproject.com/browser/django/branches
.. _pointing-python-at-the-new-django-version:
Pointing Python at the new Django version
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Once you've retrieved the branch's code, you'll need to change your Python
``site-packages`` directory so that it points to the branch version of the
``django`` directory. (The ``site-packages`` directory is somewhere such as
``/usr/lib/python2.7/site-packages`` or
``/usr/local/lib/python2.7/site-packages`` or ``C:\Python\site-packages``.)
The simplest way to do this is by renaming the old ``django`` directory to
``django.OLD`` and moving the trunk version of the code into the directory
and calling it ``django``.
Alternatively, you can use a symlink called ``django`` that points to the
location of the branch's ``django`` package. If you want to switch back, just
change the symlink to point to the old code.
A third option is to use a path file (``<something>.pth``). This is a feature of
the :mod:`site` module. First, make sure there are no files, directories or
symlinks named ``django`` in your ``site-packages`` directory. Then create a
text file named ``django.pth`` and save it to your ``site-packages`` directory.
That file should contain a path to your copy of Django on a single line and
optional comments. Here is an example that points to multiple branches. Just
uncomment the line for the branch you want to use ('trunk' in this example) and
make sure all other lines are commented::
# Trunk is a svn checkout of:
# https://code.djangoproject.com/svn/django/trunk/
#
/path/to/trunk
# <branch> is a svn checkout of:
# https://code.djangoproject.com/svn/django/branches/<branch>/
#
#/path/to/<branch>
# On windows a path may look like this:
# C:/path/to/<branch>
.. _django-developers: http://groups.google.com/group/django-developers

View file

@ -12,4 +12,4 @@ chances to be included in Django core:
coding-style
unit-tests
submitting-patches
branch-policy
working-with-git

View file

@ -6,6 +6,15 @@ We're always grateful for patches to Django's code. Indeed, bug reports
with associated patches will get fixed *far* more quickly than those
without patches.
Typo fixes and trivial documentation changes
--------------------------------------------
If you are fixing a really trivial issue, for example changing a word in the
documentation, the preferred way to provide the patch is using GitHub pull
requests without a Trac ticket. Trac tickets are still acceptable.
See the :doc:`working-with-git` for more details on how to use pull requests.
"Claiming" tickets
------------------
@ -69,38 +78,16 @@ Of course, going through the steps of claiming tickets is overkill in some
cases. In the case of small changes, such as typos in the documentation or
small bugs that will only take a few minutes to fix, you don't need to jump
through the hoops of claiming tickets. Just submit your patch and be done with
it.
it. Of course, it is always acceptable, regardless of the ticket's ownership
status, to submit patches to a ticket if you happen to have a patch ready.
.. _patch-style:
Patch style
-----------
* Make sure your code matches our :doc:`coding-style`.
* Submit patches in the format returned by the ``svn diff`` command.
An exception is for code changes that are described more clearly in
plain English than in code. Indentation is the most common example; it's
hard to read patches when the only difference in code is that it's
indented.
Patches in ``git diff`` format are also acceptable.
* When creating patches, always run ``svn diff`` from the top-level
``trunk`` directory -- i.e. the one that contains ``django``, ``docs``,
``tests``, ``AUTHORS``, etc. This makes it easy for other people to
apply your patches.
* Attach patches to a ticket in the `ticket tracker`_, using the "attach
file" button. Please *don't* put the patch in the ticket description
or comment unless it's a single line patch.
* Name the patch file with a ``.diff`` extension; this will let the ticket
tracker apply correct syntax highlighting, which is quite helpful.
* Check the "Has patch" box on the ticket details. This will make it
obvious that the ticket includes a patch, and it will add the ticket to
the `list of tickets with patches`_.
Make sure that any contribution you do fulfills at least the following
requirements:
* The code required to fix a problem or add a feature is an essential part
of a patch, but it is not the only part. A good patch should also
@ -114,6 +101,37 @@ Patch style
behavior of an existing feature, the patch should also contain
documentation.
You can use either GitHub branches and pull requests or direct patches
to publish your work. If you use the Git workflow, then you should
announce your branch in the ticket by including a link to your branch.
When you think your work is ready to be merged in create a pull request.
See the :doc:`working-with-git` documentation for mode details.
You can also use patches in Trac. When using this style, follow these
guidelines.
* Submit patches in the format returned by the ``git diff`` command.
An exception is for code changes that are described more clearly in
plain English than in code. Indentation is the most common example; it's
hard to read patches when the only difference in code is that it's
indented.
* Attach patches to a ticket in the `ticket tracker`_, using the "attach
file" button. Please *don't* put the patch in the ticket description
or comment unless it's a single line patch.
* Name the patch file with a ``.diff`` extension; this will let the ticket
tracker apply correct syntax highlighting, which is quite helpful.
Regardless of the way you submit your work, follow these steps.
* Make sure your code matches our :doc:`coding-style`.
* Check the "Has patch" box on the ticket details. This will make it
obvious that the ticket includes a patch, and it will add the ticket to
the `list of tickets with patches`_.
Non-trivial patches
-------------------

View file

@ -36,9 +36,7 @@ with this sample ``settings`` module, ``cd`` into the Django
./runtests.py --settings=test_sqlite
If you get an ``ImportError: No module named django.contrib`` error,
you need to add your install of Django to your ``PYTHONPATH``. For
more details on how to do this, read
:ref:`pointing-python-at-the-new-django-version`.
you need to add your install of Django to your ``PYTHONPATH``.
Using another ``settings`` module
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

View file

@ -0,0 +1,222 @@
Working with Git and GitHub
===========================
Django uses `Git`_ for its source control. You can `download
<http://git-scm.com/download>`_ Git, but it's often easier to install with
your operating system's package manager.
Django's `Git repository`_ is hosted on `GitHub`_, and it is recommended
that you also work using GitHub.
After installing Git the first thing you should do is setup your name and
email::
$ git config --global user.name "Firstname Lastname"
$ git config --global user.email "your_email@youremail.com"
Note that ``user.name`` should be your real name, not your GitHub nick. GitHub
should know the email you use in the ``user.email`` field, as this will be
used to associate your commits with your GitHub account.
Now we are going to show how to create a GitHub pull request containing the
changes for Trac ticket #xxxxx. By creating a fully ready pull request you
will make the committers' job easier, and thus your work is more likely to be
merged into Django. You can also upload a traditional patch to Trac, but it's
less practical for reviews.
.. _Git: http://git-scm.com/
.. _GitHub: https://github.com/
.. _Git repository: https://github.com/django/django/
Setting up local repository
---------------------------
When you have created a GitHub account, with the nick "github_nick", and
forked Django's repository, you should create a local copy of your fork::
git clone git@github.com:github_nick/django.git
This will create a new directory "django" containing a clone of your GitHub
repository. Your GitHub repository will be called "origin" in Git. You should
also setup django/django as an "upstream" remote::
git remote add upstream git@github.com:django/django.git
git fetch upstream
You can add other remotes similarly, for example::
git remote add akaariai git@github.com:akaariai/django.git
Working on a ticket
-------------------
When working on a ticket you will almost always want to create a new branch
for the work, and base that work on upstream/master::
git checkout -b ticket_xxxxx upstream/master
If you are working for a fix on the 1.4 branch, you would instead do::
git checkout -b ticket_xxxxx_1_4 upstream/stable/1.4.x
Assume the work is carried on ticket_xxxxx branch. Make some changes and
commit them::
git commit
When writing the commit message, you should follow the :ref:`commit message
guidelines <committing-guidlines>` to ease the work of the committer. If
you're uncomfortable with English, try at least to describe precisely what the
commit does.
If you need to do additional work on your branch, commit as often as
necessary::
git commit -m 'Added two more tests for edge cases'
Publishing work
~~~~~~~~~~~~~~~
You can publish your work on GitHub by just using::
git push origin ticket_xxxxx
When you go to your GitHub page you will notice a new branch has been created.
If you are working on a Trac ticket, you should mention in the ticket that
your work is available from branch ticket_xxxxx of your github repo. Include a
link to your branch.
Note that the above branch is called a "topic branch" in Git parlance. This
means that other people should not base their work on your branch. In
particular this means you are free to rewrite the history of this branch (by
using ``git rebase`` for example). There are also "public branches". These are
branches other people are supposed to fork, and thus their history should
never change. Good examples of public branches are the ``master`` and
``stable/A.B.x`` branches in the django/django repository.
When you think your work is ready to be pulled into Django, you should create
a pull request at GitHub. A good pull request contains:
* Commits with one logical change in each, following the
:doc:`coding style <coding-style>`.
* Well formed messages for each commit: a summary line and then paragraphs
wrapped at 72 characters thereafter. See the :ref:`committing guidelines
<committing-guidlines>` for more details.
* Documentation and tests, if needed. Actually tests are always needed, except
for documentation changes.
* The test suite passes and the documentation builds without warnings.
Once you have created your pull request, you should add a comment in the
related Trac ticket explaining what you've done. In particular you should tell
in which environment you've run the tests, for instance: "all tests pass under
SQLite and MySQL".
Your pull request should be ready for merging into Django. Pull requests at
GitHub have only two states: open and closed. The committers who deals with
your pull request has only two options: merge it or close it. For this reason,
it isn't useful to make a pull request until the code is ready for merging --
or sufficiently close that a committer will finish it himself.
Rebasing branches
~~~~~~~~~~~~~~~~~
In the example above you created two commits, the "Fixed ticket_xxxxx" commit
and "Added two more tests" commit. We do not want to have the "Added two more
tests" commit in the Django's repository as it would just be useless noise.
Instead, we would like to only have one commit. To rework the history of your
branch you can squash the commits into one by using interactive rebase::
git rebase -i HEAD~2
The HEAD~2 above is shorthand for two latest commits. The above command
will open an editor showing the two commits, prefixed with the word "pick".
You should change the second line to "squash" instead. This will keep the
first commit, and squash the second commit to the first one. Save and quit
the editor. A second editor window should open. Here you can reword the
commit message for the commit.
You can also use the "edit" option in rebase. This way you can change a single
commit. For example::
git rebase -i HEAD~3
# Choose edit, pick, pick for the commits
# Now you are able to rework the commit (use git add normally to add changes)
# When finished, commit work with "--amend" and continue
git commit --amend
# reword the commit message if needed
git rebase --continue
# The second and third commit should be applied.
If you need to change an already published topic branch at GitHub, you will
need to force-push the changes::
git push -f origin ticket_xxxxx
Note that this will rewrite history of ticket_xxxxx - if you check the commit
hashes before and after the operation at GitHub you will notice that the
commit hashes do not match any more. This is acceptable, as the branch is topic
branch, and nobody should be basing their work on this branch.
After upstream has changed
~~~~~~~~~~~~~~~~~~~~~~~~~~
When upstream (django/django) has changed, you should rebase your work. To
do this, use::
git fetch upstream
git rebase
The work is automatically rebased using the branch you forked on, in the
example case using upstream/master.
The rebase command removes all your local commits temporarily, applies the
upstream commits, and then applies your local commits again on the work. If
there are merge conflicts you will need to resolve them and then use ``git
rebase --continue``. At any point you can use ``git rebase --abort`` to return
to the original state.
Note that you want to rebase on upstream, not merge the upstream. The reason
for this is that by rebasing, your commits will always be on top of the
upstream's work, not mixed with the changes in the upstream. This way your
branch only contains commits related to its topic, and this makes squashing
easier.
After review
------------
It is unusual to get any non-trivial amount of code into core without changes
requested by reviewers. In this case, it is often a good idea to add the
changes as one incremental commit to your work. This allows the reviewer to
easily check what changes you have done::
# Do changes required by the reviewer, commit often.
# Before publishing the changes, rebase your work. Assume you added two
# commits to the work.
git rebase -i HEAD~2
# squash the second commit into the first, write a commit message something
# like this:
Made changes asked in review by the_reviewer
- Fixed whitespace errors in foo/bar
- Reworded the doc string of the_method()
# Push your work back to your github repo, there should not be any need
# for force (-f) push, as you didn't touch the public commits in the rebase.
git push origin ticket_xxxxx
# Check your pull request, it should now contain the new commit, too.
The committer is likely to squash the review commit into the previous commit
when committing the code.
Summary
-------
* Work on GitHub if possible.
* Announce your work on the Trac ticket by linking to your GitHub branch.
* When you have something ready, make a pull request.
* Make your pull requests as good as you can.
* When doing fixes to your work, use ``git rebase -i`` to squash the commits.
* When upstream has changed, do ``git fetch upstream; git rebase``.