1
============================
2
Guidelines for modifying bzr
3
============================
7
(The current version of this document is available in the file ``HACKING``
8
in the source tree, or at http://bazaar-ng.org/hacking.html)
13
* New functionality should have test cases. Preferably write the
14
test before writing the code.
16
In general, you can test at either the command-line level or the
17
internal API level. Choose whichever is appropriate: if adding a
18
new command, or a new command option, then call through run_bzr().
19
It is not necessary to do both. Tests that test the command line level
20
are appropriate for checking the UI behaves well - bug fixes and
21
core improvements should be tested closer to the code that is doing the
22
work. Command line level tests should be placed in 'blackbox.py'.
24
* Try to practice Test-Driven Development. before fixing a bug, write a
25
test case so that it does not regress. Similarly for adding a new
26
feature: write a test case for a small version of the new feature before
27
starting on the code itself. Check the test fails on the old code, then
28
add the feature or fix and check it passes.
30
* Exceptions should be defined inside bzrlib.errors, so that we can
31
see the whole tree at a glance.
33
* Imports should be done at the top-level of the file, unless there is
34
a strong reason to have them lazily loaded when a particular
35
function runs. Import statements have a cost, so try to make sure
36
they don't run inside hot functions.
38
* Module names should always be given fully-qualified,
39
i.e. ``bzrlib.hashcache`` not just ``hashcache``.
41
* Commands should return non-zero when they encounter circumstances that
42
the user should really pay attention to - which includes trivial shell
45
Recommended values are
47
1- Conflicts in merge-like operations, or changes are present in
49
2- Unrepresentable diff changes (i.e. binary files that we cannot show
51
3- An error or exception has occurred.
56
If you change the behaviour of an API in an incompatible way, please
57
be sure to change its name as well. For instance, if I add a keyword
58
parameter to branch.commit - that's fine. On the other hand, if I add
59
a keyword parameter to branch.commit which is a *required* transaction
60
object, I should rename the API - i.e. to 'branch.commit_transaction'.
62
This will prevent users of the old API getting surprising results.
63
Instead, they will get an Attribute error as the API is missing, and
64
will know to update their code. If in doubt, just ask on #bzr.
69
If you change the behaviour of a command, please update its docstring
70
in bzrlib/commands.py. This is displayed by the 'bzr help' command.
75
If you make a user-visible change, please add a note to the NEWS file.
76
The description should be written to make sense to someone who's just
77
a user of bzr, not a developer: new functions or classes shouldn't be
78
mentioned, but new commands, changes in behaviour or fixed nontrivial
79
bugs should be listed. See the existing entries for an idea of what
82
Within each release, entries in the news file should have the most
83
user-visible changes first. So the order should be approximately:
85
* changes to existing behaviour - the highest priority because the
86
user's existing knowledge is incorrect
87
* new features - should be brought to their attention
88
* bug fixes - may be of interest if the bug was affecting them, and
89
should include the bug number if any
90
* major documentation changes
91
* changes to internal interfaces
93
People who made significant contributions to each change are listed in
94
parenthesis. This can include reporting bugs (particularly with good
95
details or reproduction recipes), submitting patches, etc.
100
Functions, methods, classes and modules should have docstrings
101
describing how they are used.
103
The first line of the docstring should be a self-contained sentence.
105
For the special case of Command classes, this acts as the user-visible
106
documentation shown by the help command.
108
The docstrings should be formatted as reStructuredText_ (like this
109
document), suitable for processing using the epydoc_ tool into HTML
112
.. _reStructuredText: http://docutils.sourceforge.net/rst.html
113
.. _epydoc: http://epydoc.sourceforge.net/
120
Please write PEP-8__ compliant code.
122
One often-missed requirement is that the first line of docstrings
123
should be a self-contained one-sentence summary.
125
__ http://www.python.org/peps/pep-0008.html
132
Functions, methods or members that are in some sense "private" are given
133
a leading underscore prefix. This is just a hint that code outside the
134
implementation should probably not use that interface.
136
We prefer class names to be concatenated capital words (``TestCase``)
137
and variables, methods and functions to be lowercase words joined by
138
underscores (``revision_id``, ``get_revision``).
140
For the purposes of naming some names are treated as single compound
141
words: "filename", "revno".
143
Consider naming classes as nouns and functions/methods as verbs.
149
``revision_id`` not ``rev_id`` or ``revid``
151
Functions that transform one thing to another should be named ``x_to_y``
152
(not ``x2y`` as occurs in some old code.)
158
Python destructors (``__del__``) work differently to those of other
159
languages. In particular, bear in mind that destructors may be called
160
immediately when the object apparently becomes unreferenced, or at some
161
later time, or possibly never at all. Therefore we have restrictions on
162
what can be done inside them.
164
0. Never use a __del__ method without asking Martin/Robert first.
166
1. Never rely on a ``__del__`` method running. If there is code that
167
must run, do it from a ``finally`` block instead.
169
2. Never ``import`` from inside a ``__del__`` method, or you may crash the
172
3. In some places we raise a warning from the destructor if the object
173
has not been cleaned up or closed. This is considered OK: the warning
174
may not catch every case but it's still useful sometimes.
180
(The strategy described here is what we want to get to, but it's not
181
consistently followed in the code at the moment.)
183
bzrlib is intended to be a generically reusable library. It shouldn't
184
write messages to stdout or stderr, because some programs that use it
185
might want to display that information through a GUI or some other
188
We can distinguish two types of output from the library:
190
1. Structured data representing the progress or result of an
191
operation. For example, for a commit command this will be a list
192
of the modified files and the finally committed revision number
195
These should be exposed either through the return code or by calls
196
to a callback parameter.
198
A special case of this is progress indicators for long-lived
199
operations, where the caller should pass a ProgressBar object.
201
2. Unstructured log/debug messages, mostly for the benefit of the
202
developers or users trying to debug problems. This should always
203
be sent through ``bzrlib.trace`` and Python ``logging``, so that
204
it can be redirected by the client.
206
The distinction between the two is a bit subjective, but in general if
207
there is any chance that a library would want to see something as
208
structured data, we should make it so.
210
The policy about how output is presented in the text-mode client
211
should be only in the command-line tool.
216
In general tests should be placed in a file named testFOO.py where
217
FOO is the logical thing under test. That file should be placed in the
218
tests subdirectory under the package being tested.
220
For example, tests for merge3 in bzrlib belong in bzrlib/tests/testmerge3.py.
221
See bzrlib/selftest/testsampler.py for a template test script.
226
Currently, bzr selftest is used to invoke tests.
227
You can provide a pattern argument to run a subset. For example,
228
to run just the whitebox tests, run::
230
bzr selftest -v whitebox
233
Errors and exceptions
234
=====================
236
Errors are handled through Python exceptions. They can represent user
237
errors, environmental errors or program bugs. Sometimes we can't be sure
238
at the time it's raised which case applies. See bzrlib/errors.py for
239
details on the error-handling practices.
246
Integer identifier for a revision on the main line of a branch.
247
Revision 0 is always the null revision; others are 1-based
248
indexes into the branch's revision history.
added
removed
HACKING
