~bzr-pqm/bzr/bzr.dev

======================

Bazaar Developer Guide

======================

.. contents::

(The current version of this document is available in the file 

``doc/developers/HACKING.txt`` in the source tree, or at

http://doc.bazaar-vcs.org/bzr.dev/en/developer-guide/HACKING.html)

Getting Started

###############

Exploring the Bazaar Platform

=============================

Before making changes, it's a good idea to explore the work already

done by others. Perhaps the new feature or improvement you're looking

for is available in another plug-in already? If you find a bug,

perhaps someone else has already fixed it?

To answer these questions and more, take a moment to explore the

overall Bazaar Platform. Here are some links to browse:

* The Plugins page on the Wiki - http://bazaar-vcs.org/BzrPlugins

* The Bazaar product family on Launchpad - https://launchpad.net/bazaar

* Bug Tracker for the core product - https://bugs.launchpad.net/bzr/

* Blueprint Tracker for the core product - https://blueprints.launchpad.net/bzr/

If nothing else, perhaps you'll find inspiration in how other developers

have solved their challenges.

Planning and Discussing Changes

===============================

There is a very active community around Bazaar. Mostly we meet on IRC

(#bzr on irc.freenode.net) and on the mailing list. To join the Bazaar

community, see http://bazaar-vcs.org/BzrSupport.

If you are planning to make a change, it's a very good idea to mention it

on the IRC channel and/or on the mailing list. There are many advantages

to involving the community before you spend much time on a change.

These include:

* you get to build on the wisdom on others, saving time

* if others can direct you to similar code, it minimises the work to be done 

* it assists everyone in coordinating direction, priorities and effort.

In summary, maximising the input from others typically minimises the

total effort required to get your changes merged. The community is

friendly, helpful and always keen to welcome newcomers.

Bazaar Development in a Nutshell

================================

Looking for a 10 minute introduction to submitting a change?

See http://bazaar-vcs.org/BzrGivingBack.

TODO: Merge that Wiki page into this document.

Understanding the Development Process

=====================================

The development team follows many best-practices including:

* a public roadmap and planning process in which anyone can participate

* time based milestones everyone can work towards and plan around

* extensive code review and feedback to contributors

* complete and rigorous test coverage on any code contributed

* automated validation that all tests still pass before code is merged

  into the main code branch.

The key tools we use to enable these practices are:

* Launchpad - https://launchpad.net/

* Bazaar - http://bazaar-vcs.org/

* Bundle Buggy - http://bundlebuggy.aaronbentley.com/

* Patch Queue Manager - https://launchpad.net/pqm/

For further information, see http://bazaar-vcs.org/BzrDevelopment.

A Closer Look at the Merge & Review Process

===========================================

If you'd like to propose a change, please post to the

bazaar@lists.canonical.com list with a bundle, patch, or link to a

branch. Put '[PATCH]' or '[MERGE]' in the subject so Bundle Buggy

can pick it out, and explain the change in the email message text.

Remember to update the NEWS file as part of your change if it makes any

changes visible to users or plugin developers. Please include a diff

against mainline if you're giving a link to a branch.

You can generate a bundle like this::

  bzr bundle > mybundle.patch

A .patch extension is recommended instead of .bundle as many mail clients

will send the latter as a binary file. If a bundle would be too long or your

mailer mangles whitespace (e.g. implicitly converts Unix newlines to DOS

newlines), use the merge-directive command instead like this::

  bzr merge-directive http://bazaar-vcs.org http://example.org/my_branch > my_directive.patch

See the help for details on the arguments to merge-directive.

Please do **NOT** put [PATCH] or [MERGE] in the subject line if you don't

want it to be merged. If you want comments from developers rather than

to be merged, you can put '[RFC]' in the subject line.

Anyone is welcome to review code.  There are broadly three gates for

code to get in:

 * Doesn't reduce test coverage: if it adds new methods or commands,

   there should be tests for them.  There is a good test framework

   and plenty of examples to crib from, but if you are having trouble

   working out how to test something feel free to post a draft patch

   and ask for help.

 * Doesn't reduce design clarity, such as by entangling objects

   we're trying to separate.  This is mostly something the more

   experienced reviewers need to help check.

 * Improves bugs, features, speed, or code simplicity.

Code that goes in should pass all three. The core developers take care

to keep the code quality high and understandable while recognising that

perfect is sometimes the enemy of good. (It is easy for reviews to make

people notice other things which should be fixed but those things should

not hold up the original fix being accepted. New things can easily be

recorded in the Bug Tracker instead.)

Anyone can "vote" on the mailing list. Core developers can also vote using

Bundle Buggy. Here are the voting codes and their explanations.

:approve:  Reviewer wants this submission merged.

:tweak:    Reviewer wants this submission merged with small changes. (No

  re-review required.)

:abstain:  Reviewer does not intend to vote on this patch.

:resubmit: Please make changes and resubmit for review.

:reject:   Reviewer doesn't want this kind of change merged.

:comment:  Not really a vote. Reviewer just wants to comment, for now.

If a change gets two approvals from core reviewers, and no rejections,

then it's OK to come in.  Any of the core developers can bring it into the

bzr.dev trunk and backport it to maintenance branches if required.  The

Release Manager will merge the change into the branch for a pending

release, if any. As a guideline, core developers usually merge their own

changes and volunteer to merge other contributions if they were the second

reviewer to agree to a change.

To track the progress of proposed changes, use Bundle Buggy. See

http://bundlebuggy.aaronbentley.com/help for a link to all the

outstanding merge requests together with an explanation of the columns.

Bundle Buggy will also mail you a link to track just your change.

Preparing a Sandbox for Making Changes to Bazaar

================================================

Bazaar supports many ways of organising your work. See

http://bazaar-vcs.org/SharedRepositoryLayouts for a summary of the

popular alternatives.

Of course, the best choice for you will depend on numerous factors:

the number of changes you may be making, the complexity of the changes, etc.

As a starting suggestion though:

* create a local copy of the main development branch (bzr.dev) by using

  this command::

    bzr branch http://bazaar-vcs.org/bzr/bzr.dev/ bzr.dev

* keep your copy of bzr.dev prestine (by not developing in it) and keep

  it up to date (by using bzr pull)

* create a new branch off your local bzr.dev copy for each issue

  (bug or feature) you are working on.

This approach makes it easy to go back and make any required changes

after a code review. Resubmitting the change is then simple with no

risk of accidentially including edits related to other issues you may

be working on. After the changes for an issue are accepted and merged,

the associated branch can be deleted or archived as you wish.

Navigating the Code Base

========================

TODO: List and describe in one line the purpose of each directory

inside an installation of bzr.

TODO: Refer to a central location holding an up to date copy of the API

documentation generated by epydoc, e.g. something like

http://starship.python.net/crew/mwh/bzrlibapi/bzrlib.html.

Testing Bazaar

##############

The Importance of Testing

=========================

Reliability is a critical success factor for any Version Control System.

We want Bazaar to be highly reliable across multiple platforms while

evolving over time to meet the needs of its community. 

In a nutshell, this is want we expect and encourage:

* New functionality should have test cases.  Preferably write the

  test before writing the code.

  In general, you can test at either the command-line level or the

  internal API level.  See Writing tests below for more detail.

* Try to practice Test-Driven Development: before fixing a bug, write a

  test case so that it does not regress.  Similarly for adding a new

  feature: write a test case for a small version of the new feature before

  starting on the code itself.  Check the test fails on the old code, then

  add the feature or fix and check it passes.

By doing these things, the Bazaar team gets increased confidence that

changes do what they claim to do, whether provided by the core team or

by community members. Equally importantly, we can be surer that changes

down the track do not break new features or bug fixes that you are

contributing today.

As of May 2007, Bazaar ships with a test suite containing over 6000 tests

and growing. We are proud of it and want to remain so. As community

members, we all benefit from it. Would you trust version control on

your project to a product *without* a test suite like Bazaar has?

Running the Test Suite

======================

Currently, bzr selftest is used to invoke tests.

You can provide a pattern argument to run a subset. For example, 

to run just the blackbox tests, run::

  ./bzr selftest -v blackbox

To skip a particular test (or set of tests), use the --exclude option

(shorthand -x) like so::

  ./bzr selftest -v -x blackbox  

To ensure that all tests are being run and succeeding, you can use the

--strict option which will fail if there are any missing features or known

failures, like so::

  ./bzr selftest --strict

To list tests without running them, use the --list-only option like so::

  ./bzr selftest --list-only

This option can be combined with other selftest options (like -x) and

filter patterns to understand their effect.

Writing Tests

=============

In general tests should be placed in a file named test_FOO.py where 

FOO is the logical thing under test. That file should be placed in the

tests subdirectory under the package being tested.

For example, tests for merge3 in bzrlib belong in bzrlib/tests/test_merge3.py.

See bzrlib/tests/test_sampler.py for a template test script.

Tests can be written for the UI or for individual areas of the library.

Choose whichever is appropriate: if adding a new command, or a new command 

option, then you should be writing a UI test.  If you are both adding UI

functionality and library functionality, you will want to write tests for 

both the UI and the core behaviours.  We call UI tests 'blackbox' tests

and they are found in ``bzrlib/tests/blackbox/*.py``. 

When writing blackbox tests please honour the following conventions:

 1. Place the tests for the command 'name' in

    bzrlib/tests/blackbox/test_name.py. This makes it easy for developers

    to locate the test script for a faulty command.

 2. Use the 'self.run_bzr("name")' utility function to invoke the command

    rather than running bzr in a subprocess or invoking the

    cmd_object.run() method directly. This is a lot faster than

    subprocesses and generates the same logging output as running it in a

    subprocess (which invoking the method directly does not).

 3. Only test the one command in a single test script. Use the bzrlib 

    library when setting up tests and when evaluating the side-effects of

    the command. We do this so that the library api has continual pressure

    on it to be as functional as the command line in a simple manner, and

    to isolate knock-on effects throughout the blackbox test suite when a

    command changes its name or signature. Ideally only the tests for a

    given command are affected when a given command is changed.

 4. If you have a test which does actually require running bzr in a

    subprocess you can use ``run_bzr_subprocess``. By default the spawned

    process will not load plugins unless ``--allow-plugins`` is supplied.

Doctests

--------

We make selective use of doctests__.  In general they should provide 

*examples* within the API documentation which can incidentally be tested.  We 

don't try to test every important case using doctests -- regular Python

tests are generally a better solution.

Most of these are in ``bzrlib/doc/api``.  More additions are welcome.

  __ http://docs.python.org/lib/module-doctest.html

Skipping tests and test requirements

------------------------------------

In our enhancements to unittest we allow for some addition results beyond

just success or failure.

If a test can't be run, it can say that it's skipped.  This is typically

used in parameterized tests - for example if a transport doesn't support

setting permissions, we'll skip the tests that relating to that.  ::

    try:

        return self.branch_format.initialize(repo.bzrdir)

    except errors.UninitializableFormat:

        raise tests.TestSkipped('Uninitializable branch format')

Raising TestSkipped is a good idea when you want to make it clear that the

test was not run, rather than just returning which makes it look as if it

was run and passed.

Several different cases are distinguished:

TestSkipped

        Generic skip; the only type that was present up to bzr 0.18.

TestNotApplicable

        The test doesn't apply to the parameters with which it was run.

        This is typically used when the test is being applied to all

        implementations of an interface, but some aspects of the interface

        are optional and not present in particular concrete

        implementations.  (Some tests that should raise this currently

        either silently return or raise TestSkipped.)  Another option is

        to use more precise parameterization to avoid generating the test

        at all.

TestPlatformLimit

        **(Not implemented yet)**

        The test can't be run because of an inherent limitation of the

        environment, such as not having symlinks or not supporting

        unicode.

UnavailableFeature

        The test can't be run because a dependency (typically a Python

        library) is not available in the test environment.  These

        are in general things that the person running the test could fix 

        by installing the library.  It's OK if some of these occur when 

        an end user runs the tests or if we're specifically testing in a

        limited environment, but a full test should never see them.

KnownFailure

        The test exists but is known to fail, for example because the 

        code to fix it hasn't been run yet.  Raising this allows 

        you to distinguish these failures from the ones that are not 

        expected to fail.  This could be conditionally raised if something

        is broken on some platforms but not on others.

We plan to support three modes for running the test suite to control the

interpretation of these results.  Strict mode is for use in situations

like merges to the mainline and releases where we want to make sure that

everything that can be tested has been tested.  Lax mode is for use by

developers who want to temporarily tolerate some known failures.  The

default behaviour is obtained by ``bzr selftest`` with no options, and

also (if possible) by running under another unittest harness.

======================= ======= ======= ========

result                  strict  default lax

======================= ======= ======= ========

TestSkipped             pass    pass    pass

TestNotApplicable       pass    pass    pass

TestPlatformLimit       pass    pass    pass

TestDependencyMissing   fail    pass    pass

KnownFailure            fail    pass    pass

======================= ======= ======= ========

Test feature dependencies

-------------------------

Rather than manually checking the environment in each test, a test class

can declare its dependence on some test features.  The feature objects are

checked only once for each run of the whole test suite.

For historical reasons, as of May 2007 many cases that should depend on

features currently raise TestSkipped.)

::

    class TestStrace(TestCaseWithTransport):

        _test_needs_features = [StraceFeature]

This means all tests in this class need the feature.  The feature itself

should provide a ``_probe`` method which is called once to determine if

it's available.

These should generally be equivalent to either TestDependencyMissing or

sometimes TestPlatformLimit.

Known failures

--------------

Known failures are when a test exists but we know it currently doesn't

work, allowing the test suite to still pass.  These should be used with

care, we don't want a proliferation of quietly broken tests.  It might be

appropriate to use them if you've committed a test for a bug but not the

fix for it, or if something works on Unix but not on Windows.

Testing exceptions and errors

-----------------------------

It's important to test handling of errors and exceptions.  Because this

code is often not hit in ad-hoc testing it can often have hidden bugs --

it's particularly common to get NameError because the exception code

references a variable that has since been renamed.

.. TODO: Something about how to provoke errors in the right way?

In general we want to test errors at two levels:

1. A test in ``test_errors.py`` checking that when the exception object is

   constructed with known parameters it produces an expected string form.

   This guards against mistakes in writing the format string, or in the

   ``str`` representations of its parameters.  There should be one for

   each exception class.

2. Tests that when an api is called in a particular situation, it raises

   an error of the expected class.  You should typically use

   ``assertRaises``, which in the Bazaar test suite returns the exception

   object to allow you to examine its parameters.  

In some cases blackbox tests will also want to check error reporting.  But

it can be difficult to provoke every error through the commandline

interface, so those tests are only done as needed -- eg in response to a

particular bug or if the error is reported in an unusual way(?)  Blackbox

tests should mostly be testing how the command-line interface works, so

should only test errors if there is something particular to the cli in how

they're displayed or handled.

Testing warnings

----------------

The Python ``warnings`` module is used to indicate a non-fatal code

problem.  Code that's expected to raise a warning can be tested through

callCatchWarnings.

The test suite can be run with ``-Werror`` to check no unexpected errors

occur.

However, warnings should be used with discretion.  It's not an appropriate

way to give messages to the user, because the warning is normally shown

only once per source line that causes the problem.  You should also think

about whether the warning is serious enought that it should be visible to

users who may not be able to fix it.

Interface implementation testing and test scenarios

---------------------------------------------------

There are several cases in Bazaar of multiple implementations of a common 

conceptual interface.  ("Conceptual" because 

it's not necessary for all the implementations to share a base class,

though they often do.)  Examples include transports and the working tree,

branch and repository classes. 

In these cases we want to make sure that every implementation correctly

fulfils the interface requirements.  For example, every Transport should

support the ``has()`` and ``get()`` and ``clone()`` methods.  We have a

sub-suite of tests in ``test_transport_implementations``.  (Most

per-implementation tests are in submodules of ``bzrlib.tests``, but not

the transport tests at the moment.)  

These tests are repeated for each registered Transport, by generating a

new TestCase instance for the cross product of test methods and transport

implementations.  As each test runs, it has ``transport_class`` and

``transport_server`` set to the class it should test.  Most tests don't

access these directly, but rather use ``self.get_transport`` which returns

a transport of the appropriate type.

The goal is to run per-implementation only tests that relate to that

particular interface.  Sometimes we discover a bug elsewhere that happens

with only one particular transport.  Once it's isolated, we can consider 

whether a test should be added for that particular implementation,

or for all implementations of the interface.

The multiplication of tests for different implementations is normally 

accomplished by overriding the ``test_suite`` function used to load 

tests from a module.  This function typically loads all the tests,

then applies a TestProviderAdapter to them, which generates a longer 

suite containing all the test variations.

Test scenarios

--------------

Some utilities are provided for generating variations of tests.  This can

be used for per-implementation tests, or other cases where the same test

code needs to run several times on different scenarios.

The general approach is to define a class that provides test methods,

which depend on attributes of the test object being pre-set with the

values to which the test should be applied.  The test suite should then

also provide a list of scenarios in which to run the tests.

Typically ``multiply_tests_from_modules`` should be called from the test

module's ``test_suite`` function.

Essential Domain Classes

########################

Introducing the Object Model

============================

The core domain objects within the bazaar model are:

* Transport

* Branch

* Repository

* WorkingTree

Transports are explained below. See http://bazaar-vcs.org/Classes/

for an introduction to the other key classes.

Using Transports

================

The ``Transport`` layer handles access to local or remote directories.

Each Transport object acts like a logical connection to a particular

directory, and it allows various operations on files within it.  You can

*clone* a transport to get a new Transport connected to a subdirectory or

parent directory.

Transports are not used for access to the working tree.  At present

working trees are always local and they are accessed through the regular

Python file io mechanisms.

Filenames vs URLs

-----------------

Transports work in URLs.  Take note that URLs are by definition only

ASCII - the decision of how to encode a Unicode string into a URL must be

taken at a higher level, typically in the Store.  (Note that Stores also

escape filenames which cannot be safely stored on all filesystems, but

this is a different level.)

The main reason for this is that it's not possible to safely roundtrip a

URL into Unicode and then back into the same URL.  The URL standard

gives a way to represent non-ASCII bytes in ASCII (as %-escapes), but

doesn't say how those bytes represent non-ASCII characters.  (They're not

guaranteed to be UTF-8 -- that is common but doesn't happen everywhere.)

For example if the user enters the url ``http://example/%e0`` there's no

way to tell whether that character represents "latin small letter a with

grave" in iso-8859-1, or "latin small letter r with acute" in iso-8859-2

or malformed UTF-8.  So we can't convert their URL to Unicode reliably.

Equally problematic if we're given a url-like string containing non-ascii

characters (such as the accented a) we can't be sure how to convert that

to the correct URL, because we don't know what encoding the server expects

for those characters.  (Although this is not totally reliable we might still

accept these and assume they should be put into UTF-8.)

A similar edge case is that the url ``http://foo/sweet%2Fsour`` contains

one directory component whose name is "sweet/sour".  The escaped slash is

not a directory separator.  If we try to convert URLs to regular Unicode

paths this information will be lost.

This implies that Transports must natively deal with URLs; for simplicity

they *only* deal with URLs and conversion of other strings to URLs is done

elsewhere.  Information they return, such as from ``list_dir``, is also in

the form of URL components.

Core Topics

###########

Evolving Interfaces

===================

We have a commitment to 6 months API stability - any supported symbol in a

release of bzr MUST NOT be altered in any way that would result in

breaking existing code that uses it. That means that method names,

parameter ordering, parameter names, variable and attribute names etc must

not be changed without leaving a 'deprecated forwarder' behind. This even

applies to modules and classes.

If you wish to change the behaviour of a supported API in an incompatible

way, you need to change its name as well. For instance, if I add an optional keyword

parameter to branch.commit - that's fine. On the other hand, if I add a

keyword parameter to branch.commit which is a *required* transaction

object, I should rename the API - i.e. to 'branch.commit_transaction'. 

When renaming such supported API's, be sure to leave a deprecated_method (or

_function or ...) behind which forwards to the new API. See the

bzrlib.symbol_versioning module for decorators that take care of the

details for you - such as updating the docstring, and issuing a warning

when the old api is used.

For unsupported API's, it does not hurt to follow this discipline, but it's

not required. Minimally though, please try to rename things so that

callers will at least get an AttributeError rather than weird results.

Deprecation decorators

----------------------

``bzrlib.symbol_versioning`` provides decorators that can be attached to

methods, functions, and other interfaces to indicate that they should no

longer be used.

To deprecate a static method you must call ``deprecated_function``

(**not** method), after the staticmethod call::

    @staticmethod

    @deprecated_function(zero_ninetyone)

    def create_repository(base, shared=False, format=None):

When you deprecate an API, you should not just delete its tests, because

then we might introduce bugs in them.  If the API is still present at all,

it should still work.  The basic approach is to use

``TestCase.applyDeprecated`` which in one step checks that the API gives

the expected deprecation message, and also returns the real result from

the method, so that tests can keep running.

Coding Style Guidelines

=======================

General Python rules

--------------------

``hasattr`` should not be used because it swallows exceptions including

``KeyboardInterrupt``.  Instead, say something like ::

  if getattr(thing, 'name', None) is None

Code layout

-----------

Please write PEP-8__ compliant code.  

__ http://www.python.org/peps/pep-0008.html

One often-missed requirement is that the first line of docstrings

should be a self-contained one-sentence summary.

We use 4 space indents for blocks, and never use tab characters.  (In vim,

``set expandtab``.)

Lines should be no more than 79 characters if at all possible.

Lines that continue a long statement may be indented in either of 

two ways:

within the parenthesis or other character that opens the block, e.g.::

    my_long_method(arg1,

                   arg2,

                   arg3)

or indented by four spaces::

    my_long_method(arg1,

        arg2,

        arg3)

The first is considered clearer by some people; however it can be a bit

harder to maintain (e.g. when the method name changes), and it does not

work well if the relevant parenthesis is already far to the right.  Avoid

this::

     self.legbone.kneebone.shinbone.toebone.shake_it(one,

                                                     two,

                                                     three)

but rather ::

     self.legbone.kneebone.shinbone.toebone.shake_it(one,

         two,

         three)

or ::

     self.legbone.kneebone.shinbone.toebone.shake_it(

         one, two, three)

For long lists, we like to add a trailing comma and put the closing

character on the following line.  This makes it easier to add new items in

future::

    from bzrlib.goo import (

        jam,

        jelly,

        marmalade,

        )

There should be spaces between function paramaters, but not between the

keyword name and the value::

    call(1, 3, cheese=quark)

In emacs::

    ;(defface my-invalid-face

    ;  '((t (:background "Red" :underline t)))

    ;  "Face used to highlight invalid constructs or other uglyties"

    ;  )

    (defun my-python-mode-hook ()

     ;; setup preferred indentation style.

     (setq fill-column 79)

     (setq indent-tabs-mode nil) ; no tabs, never, I will not repeat

    ;  (font-lock-add-keywords 'python-mode

    ;                         '(("^\\s *\t" . 'my-invalid-face) ; Leading tabs

    ;                            ("[ \t]+$" . 'my-invalid-face)  ; Trailing spaces

    ;                            ("^[ \t]+$" . 'my-invalid-face)); Spaces only

    ;                          )

     )

    (add-hook 'python-mode-hook 'my-python-mode-hook)

The lines beginning with ';' are comments. They can be activated

if one want to have a strong notice of some tab/space usage

violations.

Module Imports

--------------

* Imports should be done at the top-level of the file, unless there is

  a strong reason to have them lazily loaded when a particular

  function runs.  Import statements have a cost, so try to make sure

  they don't run inside hot functions.

* Module names should always be given fully-qualified,

  i.e. ``bzrlib.hashcache`` not just ``hashcache``.

Naming

------

Functions, methods or members that are "private" to bzrlib are given

a leading underscore prefix.  Names without a leading underscore are

public not just across modules but to programmers using bzrlib as an

API. As a consequence, a leading underscore is appropriate for names

exposed across modules but that are not to be exposed to bzrlib API

programmers.

We prefer class names to be concatenated capital words (``TestCase``)

and variables, methods and functions to be lowercase words joined by

underscores (``revision_id``, ``get_revision``).

For the purposes of naming some names are treated as single compound

words: "filename", "revno".

Consider naming classes as nouns and functions/methods as verbs.

Try to avoid using abbreviations in names, because there can be

inconsistency if other people use the full name.

Standard Names

--------------

``revision_id`` not ``rev_id`` or ``revid``

Functions that transform one thing to another should be named ``x_to_y``

(not ``x2y`` as occurs in some old code.)

Destructors

-----------

Python destructors (``__del__``) work differently to those of other

languages.  In particular, bear in mind that destructors may be called

immediately when the object apparently becomes unreferenced, or at some

later time, or possibly never at all.  Therefore we have restrictions on

what can be done inside them.

 0. Never use a __del__ method without asking Martin/Robert first.

 1. Never rely on a ``__del__`` method running.  If there is code that

    must run, do it from a ``finally`` block instead.

 2. Never ``import`` from inside a ``__del__`` method, or you may crash the

    interpreter!!

 3. In some places we raise a warning from the destructor if the object

    has not been cleaned up or closed.  This is considered OK: the warning

    may not catch every case but it's still useful sometimes.

Factories

---------

In some places we have variables which point to callables that construct

new instances.  That is to say, they can be used a lot like class objects,

but they shouldn't be *named* like classes:

> I think that things named FooBar should create instances of FooBar when

> called. Its plain confusing for them to do otherwise. When we have

> something that is going to be used as a class - that is, checked for via

> isinstance or other such idioms, them I would call it foo_class, so that

> it is clear that a callable is not sufficient. If it is only used as a

> factory, then yes, foo_factory is what I would use.

Registries

----------

Several places in Bazaar use (or will use) a registry, which is a 

mapping from names to objects or classes.  The registry allows for 

loading in registered code only when it's needed, and keeping

associated information such as a help string or description.

Lazy Imports

------------

To make startup time faster, we use the ``bzrlib.lazy_import`` module to

delay importing modules until they are actually used. ``lazy_import`` uses

the same syntax as regular python imports. So to import a few modules in a

lazy fashion do::

  from bzrlib.lazy_import import lazy_import

  lazy_import(globals(), """

  import os

  import subprocess

  import sys

  import time

  from bzrlib import (

     errors,

     transport,

     revision as _mod_revision,

     )

  import bzrlib.transport

  import bzrlib.xml5

  """)

At this point, all of these exist as a ``ImportReplacer`` object, ready to

be imported once a member is accessed. Also, when importing a module into

the local namespace, which is likely to clash with variable names, it is

recommended to prefix it as ``_mod_<module>``. This makes it clearer that

the variable is a module, and these object should be hidden anyway, since

they shouldn't be imported into other namespaces.

Modules versus Members

~~~~~~~~~~~~~~~~~~~~~~

While it is possible for ``lazy_import()`` to import members of a module

when using the ``from module import member`` syntax, it is recommended to

only use that syntax to load sub modules ``from module import submodule``.

This is because variables and classes can frequently be used without

needing a sub-member for example::

  lazy_import(globals(), """

  from module import MyClass

  """)

  def test(x):

      return isinstance(x, MyClass)

This will incorrectly fail, because ``MyClass`` is a ``ImportReplacer``

object, rather than the real class.

Passing to Other Variables

~~~~~~~~~~~~~~~~~~~~~~~~~~

It also is incorrect to assign ``ImportReplacer`` objects to other variables.

Because the replacer only knows about the original name, it is unable to

replace other variables. The ``ImportReplacer`` class will raise an

``IllegalUseOfScopeReplacer`` exception if it can figure out that this

happened. But it requires accessing a member more than once from the new

variable, so some bugs are not detected right away.

The Null revision

-----------------

The null revision is the ancestor of all revisions.  Its revno is 0, its

revision-id is ``null:``, and its tree is the empty tree.  When referring

to the null revision, please use ``bzrlib.revision.NULL_REVISION``.  Old

code sometimes uses ``None`` for the null revision, but this practice is

being phased out.

Getting Input

=============

Processing Command Lines

------------------------

bzrlib has a standard framework for parsing command lines and calling

processing routines associated with various commands. See builtins.py

for numerous examples.

Standard Parameter Types

------------------------

There are some common requirements in the library: some parameters need to be

unicode safe, some need byte strings, and so on. At the moment we have

only codified one specific pattern: Parameters that need to be unicode

should be checked via ``bzrlib.osutils.safe_unicode``. This will coerce the

input into unicode in a consistent fashion, allowing trivial strings to be

used for programmer convenience, but not performing unpredictably in the

presence of different locales.

Writing Output

==============

(The strategy described here is what we want to get to, but it's not

consistently followed in the code at the moment.)

bzrlib is intended to be a generically reusable library.  It shouldn't

write messages to stdout or stderr, because some programs that use it

might want to display that information through a GUI or some other

mechanism.

We can distinguish two types of output from the library:

 1. Structured data representing the progress or result of an

    operation.  For example, for a commit command this will be a list

    of the modified files and the finally committed revision number

    and id.

    These should be exposed either through the return code or by calls

    to a callback parameter.

    A special case of this is progress indicators for long-lived

    operations, where the caller should pass a ProgressBar object.

 2. Unstructured log/debug messages, mostly for the benefit of the

    developers or users trying to debug problems.  This should always

    be sent through ``bzrlib.trace`` and Python ``logging``, so that

    it can be redirected by the client.

The distinction between the two is a bit subjective, but in general if

there is any chance that a library would want to see something as

structured data, we should make it so.

The policy about how output is presented in the text-mode client

should be only in the command-line tool.

Displaying help

===============

Bazaar has online help for various topics through ``bzr help COMMAND`` or

equivalently ``bzr command -h``.  We also have help on command options,

and on other help topics.  (See ``help_topics.py``.)

As for python docstrings, the first paragraph should be a single-sentence

synopsis of the command.

The help for options should be one or more proper sentences, starting with

a capital letter and finishing with a full stop (period).

All help messages and documentation should have two spaces between

sentences.

Writing tests

=============

In general tests should be placed in a file named test_FOO.py where 

FOO is the logical thing under test. That file should be placed in the

tests subdirectory under the package being tested.

For example, tests for merge3 in bzrlib belong in bzrlib/tests/test_merge3.py.

See bzrlib/tests/test_sampler.py for a template test script.

Tests can be written for the UI or for individual areas of the library.

Choose whichever is appropriate: if adding a new command, or a new command 

option, then you should be writing a UI test.  If you are both adding UI

functionality and library functionality, you will want to write tests for 

both the UI and the core behaviours.  We call UI tests 'blackbox' tests

and they are found in ``bzrlib/tests/blackbox/*.py``. 

When writing blackbox tests please honour the following conventions:

 1. Place the tests for the command 'name' in

    bzrlib/tests/blackbox/test_name.py. This makes it easy for developers

    to locate the test script for a faulty command.

 2. Use the 'self.run_bzr("name")' utility function to invoke the command

    rather than running bzr in a subprocess or invoking the

    cmd_object.run() method directly. This is a lot faster than

    subprocesses and generates the same logging output as running it in a

    subprocess (which invoking the method directly does not).

 3. Only test the one command in a single test script. Use the bzrlib 

    library when setting up tests and when evaluating the side-effects of

    the command. We do this so that the library api has continual pressure

    on it to be as functional as the command line in a simple manner, and

    to isolate knock-on effects throughout the blackbox test suite when a

    command changes its name or signature. Ideally only the tests for a

    given command are affected when a given command is changed.

 4. If you have a test which does actually require running bzr in a

    subprocess you can use ``run_bzr_subprocess``. By default the spawned

    process will not load plugins unless ``--allow-plugins`` is supplied.

Test support

------------

We have a rich collection of tools to support writing tests. Please use

them in preference to ad-hoc solutions as they provide portability and

performance benefits.

TreeBuilder

~~~~~~~~~~~

The ``TreeBuilder`` interface allows the construction of arbitrary trees

with a declarative interface. A sample session might look like::

  tree = self.make_branch_and_tree('path')

  builder = TreeBuilder()

  builder.start_tree(tree)

  builder.build(['foo', "bar/", "bar/file"])

  tree.commit('commit the tree')

  builder.finish_tree()

Please see bzrlib.treebuilder for more details.

BranchBuilder

~~~~~~~~~~~~~

The ``BranchBuilder`` interface allows the creation of test branches in a

quick and easy manner. A sample session::

  builder = BranchBuilder(self.get_transport().clone('relpath'))

  builder.build_commit()

  builder.build_commit()

  builder.build_commit()

  branch = builder.get_branch()

Please see bzrlib.branchbuilder for more details.

Doctests

--------

We make selective use of doctests__.  In general they should provide 

*examples* within the API documentation which can incidentally be tested.  We 

don't try to test every important case using doctests -- regular Python

tests are generally a better solution.

Most of these are in ``bzrlib/doc/api``.  More additions are welcome.

  __ http://docs.python.org/lib/module-doctest.html

Running tests

=============

Currently, bzr selftest is used to invoke tests.

You can provide a pattern argument to run a subset. For example, 

to run just the blackbox tests, run::

  ./bzr selftest -v blackbox

To skip a particular test (or set of tests), use the --exclude option

(shorthand -x) like so::

  ./bzr selftest -v -x blackbox  

To list tests without running them, use the --list-only option like so::

  ./bzr selftest --list-only

This option can be combined with other selftest options (like -x) and

filter patterns to understand their effect.

Handling Errors and Exceptions

==============================

Commands should return non-zero when they encounter circumstances that

the user should really pay attention to - which includes trivial shell

pipelines.

Recommended values are:

    0. OK.

    1. Conflicts in merge-like operations, or changes are present in

       diff-like operations. 

    2. Unrepresentable diff changes (i.e. binary files that we cannot show 

       a diff of).

    3. An error or exception has occurred.

    4. An internal error occurred (one that shows a traceback.)

Errors are handled through Python exceptions. Exceptions should be defined

inside bzrlib.errors, so that we can see the whole tree at a glance.

We broadly classify errors as either being either internal or not,

depending on whether ``internal_error`` is set or not.  If we think it's our

fault, we show a backtrace, an invitation to report the bug, and possibly

other details.  This is the default for errors that aren't specifically

recognized as being caused by a user error.  Otherwise we show a briefer

message, unless -Derror was given.

Many errors originate as "environmental errors" which are raised by Python

or builtin libraries -- for example IOError.  These are treated as being

our fault, unless they're caught in a particular tight scope where we know

that they indicate a user errors.  For example if the repository format

is not found, the user probably gave the wrong path or URL.  But if one of

the files inside the repository is not found, then it's our fault --

either there's a bug in bzr, or something complicated has gone wrong in

the environment that means one internal file was deleted.

Many errors are defined in ``bzrlib/errors.py`` but it's OK for new errors

to be added near the place where they are used.

Exceptions are formatted for the user by conversion to a string

(eventually calling their ``__str__`` method.)  As a convenience the

``._fmt`` member can be used as a template which will be mapped to the

error's instance dict.

New exception classes should be defined when callers might want to catch

that exception specifically, or when it needs a substantially different

format string.

Exception strings should start with a capital letter and should not have a

final fullstop.  If long, they may contain newlines to break the text.

Documenting Changes

===================

When you change bzrlib, please update the relevant documentation for the

change you made: Changes to commands should update their help, and

possibly end user tutorials; changes to the core library should be

reflected in API documentation.

NEWS File

---------

If you make a user-visible change, please add a note to the NEWS file.

The description should be written to make sense to someone who's just

a user of bzr, not a developer: new functions or classes shouldn't be

mentioned, but new commands, changes in behaviour or fixed nontrivial

bugs should be listed.  See the existing entries for an idea of what

should be done.

Within each release, entries in the news file should have the most

user-visible changes first.  So the order should be approximately:

 * changes to existing behaviour - the highest priority because the 

   user's existing knowledge is incorrect

 * new features - should be brought to their attention

 * bug fixes - may be of interest if the bug was affecting them, and

   should include the bug number if any

 * major documentation changes

 * changes to internal interfaces

People who made significant contributions to each change are listed in

parenthesis.  This can include reporting bugs (particularly with good

details or reproduction recipes), submitting patches, etc.

Commands

--------

The docstring of a command is used by ``bzr help`` to generate help output

for the command. The list 'takes_options' attribute on a command is used by

``bzr help`` to document the options for the command - the command

docstring does not need to document them. Finally, the '_see_also'