Warning

These instrctions come directly from CPython devguide and need to be evaluated for Jython specifics. While most of this applies, our actual practices may vary from these a bit.

16. Committing and Pushing Changes

16.1. Is the change ready for committing?

Before a change is committed, you must make sure it is ready to enter the public source tree. Draft commits are prohibited. Therefore, you must ensure your changes fulfill several mandatory criteria.

Note

If you want to share your work-in-progress code on a feature or bugfix, either publish patches or create a public fork of the repository.

16.1.1. Does the test suite still pass?

You must run the whole test suite to ensure that it passes before pushing any code changes.

Note

You really need to run the entire test suite. Running a single test is not enough as your changes may have unforeseen effects on other tests or library modules.

Running the entire test suite doesn’t guarantee that your changes will pass the continuous integration tests, as those will exercise more possibilities still (such as different platforms or build options). But it will at least catch non-build specific, non-platform specific errors, therefore minimizing the chance for breakage.

16.1.2. Patch checklist

Apart from running the tests, there’s a simple patch checklist that make patchcheck (or ./python.exe Tools/scripts/patchcheck.py on Windows) will run through:

  • Are there any whitespace problems in Python files? (using Tools/scripts/reindent.py)
  • Are there any whitespace problems in C files?
  • Are there any whitespace problems in the documentation? (using Tools/scripts/reindent-rst.py)
  • Has the documentation been updated?
  • Has the test suite been updated?
  • Has Misc/NEWS been updated?
  • Has Misc/ACKS been updated?
  • Has configure been regenerated, if necessary?
  • Has pyconfig.h.in been regenerated, if necessary?

Note that the automated patch check can’t actually answer all of these questions, and even if it could, it still wouldn’t know whether or not those answers were appropriate. Aside from the whitespace checks, it is just a memory aid to help with remembering the various elements that can go into making a complete patch.

16.2. Commit Style

Once a change patch is ready and tested, it can be committed to the repository. We usually prefer to put a whole feature or bugfix into a single commit, but no more. In particular:

  • Do not fix more than one issue in the same commit (except, of course, if one code change fixes all of them).
  • Do not do cosmetic changes to unrelated code in the same commit as some feature/bugfix.

It is of course okay to pile up several commits to one branch and merge them into another in one merge commit.

16.3. Handling Others’ Code

As a core developer you will occasionally want to commit a patch created by someone else. When doing so you will want to make sure of some things.

First, make sure the patch is in a good state. Both patch and Helping Triage Issues explain what is to be expected of a patch. Typically patches that get cleared by triagers are good to go except maybe lacking Misc/ACKS and Misc/NEWS entries.

Second, make sure the patch does not break backwards-compatibility without a good reason. This means running the entire test suite to make sure everything still passes. It also means that if semantics do change there must be a good reason for the breakage of code the change will cause (and it will break someone’s code). If you are unsure if the breakage is worth it, ask on python-dev.

Third, ensure the patch is attributed correctly by adding the contributor’s name to Misc/ACKS if they aren’t already there (and didn’t add themselves in their patch) and by mentioning “Patch by <x>” in the Misc/NEWS entry and the checkin message. If the patch has been heavily modified then “Initial patch by <x>” is an appropriate alternate wording.

If you omit correct attribution in the initial checkin, then update ACKS and NEWS in a subsequent checkin (don’t worry about trying to fix the original checkin message in that case).

Finally, especially for larger patches, check if the submitter of the patch has a CLA in place (indicated by an asterisk following their username in the issue tracker). If the asterisk is missing and the patch is non-trivial, direct them to the electronic Contributor Licensing Agreement to ensure the PSF has the appropriate authorizations in place to relicense and redistribute their code.

16.4. Contributor Licensing Agreements

It’s unlikely bug fixes will require a Contributor Licensing Agreement unless they touch a lot of code. For new features, it is preferable to ask that the contributor submit a signed CLA to the PSF as the associated comments, docstrings and documentation are far more likely to reach a copyrightable standard.

These days, the CLA can be signed electronically through the form linked above, and this process is strongly preferred to the old mechanism that involved sending a scanned copy of the signed paper form.

As discussed on the PSF Contribution page, it is the CLA itself that gives the PSF the necessary relicensing rights to redistribute contributions under the Python license stack. This is an additional permission granted above and beyond the normal permissions provided by the chosen open source license.

Some developers may object to the relicensing permissions granted to the PSF by the CLA. They’re entirely within their rights to refuse to sign the CLA on that basis, but that refusal does mean we can’t accept their patches for inclusion.

16.5. NEWS Entries

Almost all changes made to the code base deserve an entry in Misc/NEWS. If the change is particularly interesting for end users (e.g. new features, significant improvements, or backwards-incompatible changes), an entry in the What's New in Python document (in Doc/whatsnew/) should be added as well. There are two notable exceptions to this general principle, and they both relate to changes that already have a NEWS entry, and have not yet been included in any formal release (including alpha and beta releases). These exceptions are:

  • If a change is reverted prior to release, then the corresponding entry is simply removed. Otherwise, a new entry must be added noting that the change has been reverted (e.g. when a feature is released in an alpha and then cut prior to the first beta).
  • If a change is a fix (or other adjustment) to an earlier unreleased change and the original NEWS entry remains valid, then no additional entry is needed.

New NEWS entries are customarily added at or near the top of their respective sections, so that entries within a section appear in approximate order from newest to oldest. However, this is customary and not a requirement.

The NEWS file is now read by Sphinx to produce the “Changelog” page; accordingly it should be valid reStructuredText. The “default role” (single backticks) can be used to refer to objects in the documentation. Example NEWS entry:

- Issue #15304: Fix warning message when `os.chdir()` fails inside
  `test.support.temp_cwd()`.  Patch by Chris Jerdonek.

(In all other .rst files, the single backticks should not be used. They are allowed here because NEWS is meant to be as readable as possible unprocessed.)

A nice trick to make Mercurial’s automatic file merge work more smoothly is to put a new entry after the first or first two entries rather than at the very top. This way if you commit, pull new changesets and merge, the merge will succeed automatically.

16.6. Commit Messages

Every commit has a commit message to document why a change was made and to communicate that reason to other core developers. Python core developers have developed a standard way of formatting commit messages that everyone is expected to follow.

Our usual convention mimics that used in the Misc/NEWS file. Actually, it is common to simply paste the NEWS entry into the commit message. Here is an example:

Issue #42: the spam module is now more spammy.
The spam module sporadically came up short on spam. This change
raises the amount of spam in the module by making it more spammy.
Thanks to Monty Python for the patch.

The first line or sentence is meant to be a dense, to-the-point explanation of what the purpose of the commit is. If this is not enough detail for a commit, a new paragraph(s) can be added to explain in proper depth what has happened (detail should be good enough that a core developer reading the commit message understands the justification for the change). Also, if a non-core developer contributed to the resolution, it is good practice to credit them.

16.6.1. Mercurial hooks

Special hooks have been added to the Mercurial repository to enable notifying the issue tracker of a commit related to an issue.

A commit message can mention one or several issues in one of the following ways:

#12345
issue12345
issue 12345
bug12345
bug 12345

where 12345 is the number of the issue. The commit details (including its changeset, branch and commit message) will then be posted as a message to the issue’s page in the tracker, for each mentioned issue.

If “closes” (or “closed”, or “closing”) is prepended, the issue is automatically closed as “fixed”.

17. Working with Mercurial

As a core developer, the ability to push changes to the official Python repositories means you have to be more careful with your workflow:

  • You should not push new named branches to the main repository. You can still use them in clones that you use for development of patches; you can also push these branches to a separate public repository that will be dedicated to maintenance of the work before the work gets integrated in the main repository.
  • You should collapse changesets of a single feature or bugfix before pushing the result to the main repository. The reason is that we don’t want the history to be full of intermediate commits recording the private history of the person working on a patch. If you are using the rebase extension, consider adding the --collapse option to hg rebase. The collapse extension is another choice.

Because of these constraints, it can be practical to use other approaches such as mq (Mercurial Queues), in order to maintain patches in a single local repository and to push them seamlessly when they are ready.

It can also be useful to keep a pristine clone of the main repository around, as it allows simple reversion of all local changes (even “committed” ones) if your local clone gets into a state you aren’t happy with.

17.1. Minimal Configuration

To use Mercurial as a committer (both of your and others’ patches), you should set up some basic options in your configuration file. Under Windows, TortoiseHg has a graphical settings dialog for most options, meaning you don’t need to edit the file directly (it is still available in %USERPROFILE%\Mercurial.ini). Under other platforms, you must edit ~/.hgrc.

Here are the minimal options you need to activate:

  • your username: this setting defines the name that will be used when you commit changes. The usual convention is to also include an e-mail contact address in there:

    [ui]
    username = Your Name <[email protected]>
    
  • extended diffing: this setting enables an extended diff format which is more useful than the standard unified diff format as it includes metadata about file copies, permission bits, and is able to represent binary files:

    [diff]
    git = on
    

Under Windows, you should also enable the eol extension, which will fix any Windows-specific line endings your text editor might insert when you create or modify versioned files. The public repository has a hook which will reject all changesets having the wrong line endings, so enabling this extension on your local computer is in your best interest.

17.2. Clones Setup

There are several possible ways to set up your Mercurial clone(s). If you are a core developer, you often need to work on the different branches, so the best approach is to have a separate clone/directory for each active branch. If you are a contributor, having a single clone might be enough.

17.2.1. Single Clone Approach

This approach has the advantage of being simpler because it requires a single clone/directory, but, on the other hand, it requires you to recompile Python every time you need to switch branch. For this reason, this approach is not suggested to core developers, but it’s usually suitable for contributors.

See Getting the Source Code to find information about cloning and switching branches.

17.2.2. Multiple Clones Approach

This approach requires you to keep a separate clone/directory for each active branch, but, on the other hand, it doesn’t require you to switch branches and recompile Python, so it saves times while merging and testing a patch on the different branches. For this reason, this approach is suggested to core developers.

The easiest way to do this is by using the share extension, that can be enabled by adding the following lines to your ~/.hgrc:

[extensions]
share =

Once you have cloned the hg.python.org/jython repository you can create the other shared clones using:

$ hg share jython 2.7  # create a new shared clone
$ cd 2.7                # enter the directory
$ hg up 2.7             # switch to the 2.7 branch

You can then repeat the same operation for the other active branches. This will create different clones/directories that share the same history. This means that once you commit or pull new changesets in one of the clones, they will be immediately available in all the other clones (note however that while you only need to use hg pull once, you still need to use hg up in each clone to update its working copy).

If you don’t want to specify ssh://hg@hg.python.org/jython every time you pull or push, you should add to the .hg/hgrc files of the clones:

[paths]
default = ssh://[email protected]/jython

In order to apply a patch, commit, and merge it on all the branches, you can do as follow:

$ cd 2.7
$ hg pull
$ hg up
$ hg import --no-c http://bugs.python.org/url/to/the/patch.diff
$ # review, run tests, run `make patchcheck`
$ hg ci -m '#12345: fix some issue.'
$ # switch to 3.4 and port the changeset using `hg graft`
$ cd ../3.4
$ hg up
$ hg graft 2.7
$ # switch to 3.x, merge, commit, and push everything
$ cd ../3.x
$ hg up
$ hg merge 3.4
$ hg ci -m '#12345: merge with 3.4.'
$ hg push

Unless noted otherwise, the rest of the page will assume you are using the multiple clone approach, and explain in more detail these basic steps.

For more advanced explanations about null merges, heads merges, merge conflicts, etc., see the FAQs for core developers.

17.3. Active branches

If you do hg branches you will see a list of branches. default is the in-development branch, and is the only branch that receives new features. The other branches only receive bug fixes or security fixes. Depending on what you are committing (feature, bug fix, or security fix), you should commit to the oldest branch applicable, and then forward-port until the in-development branch.

17.4. Merging order

There are two separate lines of development: one for Python 2 (the 2.x branches) and one for Python 3 (the 3.x branches and default). You should never merge between the two major versions (2.x and 3.x) — only between minor versions (e.g. 3.x->3.y). The merge always happens from the oldest applicable branch to the newest branch within the same major Python version.

17.5. Merging between different branches (within the same major version)

Assume that Python 3.5 is the current in-development version of Python and that you have a patch that should also be applied to Python 3.4. To properly port the patch to both versions of Python, you should first apply the patch to Python 3.4:

cd 3.4
hg import --no-commit patch.diff
# Compile; run the test suite
hg ci -m '#12345: fix some issue.'

Then you can switch to the 3.5 clone, merge, run the tests and commit:

cd ../3.5
hg merge 3.4
# Fix any conflicts (e.g. ``hg revert -r default Misc/NEWS``); compile; run the test suite
hg ci -m '#12345: merge with 3.4.'

If you are not using the share extension, you will need to use hg pull ../3.4 before being able to merge.

Note

Even when porting an already committed patch, you should still check the test suite runs successfully before committing the patch to another branch. Subtle differences between two branches sometimes make a patch bogus if ported without any modifications.

17.6. Porting changesets between the two major Python versions (2.x and 3.x)

Assume you just committed something on 2.7, and want to port it to 3.4. You can use hg graft as follow:

cd ../3.4
hg graft 2.7

This will port the latest changeset committed in the 2.7 clone to the 3.4 clone. hg graft always commits automatically, except in case of conflicts, when you have to resolve them and run hg graft --continue afterwards. Instead of the branch name you can also specify a changeset id, and you can also graft changesets from 3.x to 2.7.

On older version of Mercurial where hg graft is not available, you can use:

cd ../3.4
hg export 2.7 | hg import -

The result will be the same, but in case of conflict this will create .rej files rather than using Mercurial merge capabilities.

A third option is to apply manually the patch on 3.4. This is convenient when there are too many differences with 2.7 or when there is already a specific patch for 3.4.

Warning

Never use hg merge to port changes between 2.x and 3.x (or vice versa).

17.7. Long-term development of features

If you want to work on a feature long-term (perhaps you’re implementing a PEP), you will probably want to publish your work in a dedicated repository. The following instructions will help you do so on hg.python.org‘s infrastructure without requiring a lot of upload bandwidth.

Go to the main repository’s Web page (http://hg.python.org/jython/); there you find a button labelled “server-side clone”, which you can click on to display a Web form. Enter the relative path of the repository you want to create on the server, for example features/mywork; and press the button. A new repository gets created on the server with all the changesets of the original repository (it will seem very fast; this is normal).

You can now do a local clone of this repository on your disk:

$ hg clone ssh://[email protected]/features/mywork
$ cd mywork

It is recommended that you create a new named branch for your work, so as to easily track changes. That named branch will exist in your feature repository, but not in the main repository:

$ hg branch mywork
$ hg commit -m "Creating branch mywork"
$ hg push --new-branch

You can now work on your feature, commit changes as you will, and push them when desired:

$ hg push

When you push them, they will land in the public repository at ssh://hg@hg.python.org/features/mywork (or http://hg.python.org/features/mywork for the read-only URL). Other people can clone the public repository and work on the code too.

When you want to synchronize with CPython’s upstream changes, you can pull from the main repository, either from its remote URL:

$ hg pull http://hg.python.org/jython

or from a local clone that you may have on your disk (which is of course faster):

$ hg pull ../jython

and merge all new changes from branch default to branch mywork:

$ hg branch
mywork
$ hg merge default

Rather than using a clone on python.org (which isn’t particularly useful for collaboration with folks that don’t already have CPython commit rights), Bitbucket also maintain an up to date clone of the main cpython repository that can be used as the basis for a new clone or patch queue.

17.7.1. Uploading a patch for review

In this scheme, your work will probably consist of many commits (some of them merges). If you want to upload a patch for review somewhere, you need a single aggregate patch. This is where having a dedicated named branch mywork gets handy.

First ensure that you have pulled and merged all changes from the main repository, as explained above. Then, assuming your currently checked out branch is still mywork, simply do:

$ hg diff -r default > mywork.patch

This will write to mywork.patch all the changes between default and mywork.