1
============================
2
Guidelines for modifying bzr
3
============================
5
This document describes the Bazaar internals and the development process.
6
It's meant for people interested in developing Bazaar, and some parts will
7
also be useful to people developing Bazaar plugins.
9
If you have any questions or something seems to be incorrect, unclear or
10
missing, please talk to us in ``irc://irc.freenode.net/#bzr``, or write to
11
the Bazaar mailing list. To propose a correction or addition to this
12
document, send a merge request or new text to the mailing list.
14
The current version of this document is available in the file
15
``doc/developers/HACKING.txt`` in the source tree, or at
16
http://doc.bazaar-vcs.org/bzr.dev/en/developer-guide/HACKING.html
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)
24
Exploring the Bazaar Platform
25
=============================
27
Before making changes, it's a good idea to explore the work already
28
done by others. Perhaps the new feature or improvement you're looking
29
for is available in another plug-in already? If you find a bug,
30
perhaps someone else has already fixed it?
32
To answer these questions and more, take a moment to explore the
33
overall Bazaar Platform. Here are some links to browse:
35
* The Plugins page on the Wiki - http://bazaar-vcs.org/BzrPlugins
37
* The Bazaar product family on Launchpad - https://launchpad.net/bazaar
39
* Bug Tracker for the core product - https://bugs.launchpad.net/bzr/
41
* Blueprint Tracker for the core product - https://blueprints.launchpad.net/bzr/
43
If nothing else, perhaps you'll find inspiration in how other developers
44
have solved their challenges.
47
Planning and Discussing Changes
48
===============================
50
There is a very active community around Bazaar. Mostly we meet on IRC
51
(#bzr on irc.freenode.net) and on the mailing list. To join the Bazaar
52
community, see http://bazaar-vcs.org/BzrSupport.
54
If you are planning to make a change, it's a very good idea to mention it
55
on the IRC channel and/or on the mailing list. There are many advantages
56
to involving the community before you spend much time on a change.
59
* you get to build on the wisdom on others, saving time
61
* if others can direct you to similar code, it minimises the work to be done
63
* it assists everyone in coordinating direction, priorities and effort.
65
In summary, maximising the input from others typically minimises the
66
total effort required to get your changes merged. The community is
67
friendly, helpful and always keen to welcome newcomers.
70
Bazaar Development in a Nutshell
71
================================
73
Looking for a 10 minute introduction to submitting a change?
74
See http://bazaar-vcs.org/BzrGivingBack.
76
TODO: Merge that Wiki page into this document.
79
Understanding the Development Process
80
=====================================
82
The development team follows many best-practices including:
84
* a public roadmap and planning process in which anyone can participate
86
* time based milestones everyone can work towards and plan around
88
* extensive code review and feedback to contributors
90
* complete and rigorous test coverage on any code contributed
92
* automated validation that all tests still pass before code is merged
93
into the main code branch.
95
The key tools we use to enable these practices are:
97
* Launchpad - https://launchpad.net/
99
* Bazaar - http://bazaar-vcs.org/
101
* Bundle Buggy - http://bundlebuggy.aaronbentley.com/
103
* Patch Queue Manager - https://launchpad.net/pqm/
105
For further information, see http://bazaar-vcs.org/BzrDevelopment.
108
A Closer Look at the Merge & Review Process
109
===========================================
111
If you'd like to propose a change, please post to the
112
bazaar@lists.canonical.com list with a bundle, patch, or link to a
113
branch. Put '[PATCH]' or '[MERGE]' in the subject so Bundle Buggy
114
can pick it out, and explain the change in the email message text.
115
Remember to update the NEWS file as part of your change if it makes any
116
changes visible to users or plugin developers. Please include a diff
117
against mainline if you're giving a link to a branch.
119
You can generate a bundle like this::
121
bzr bundle > mybundle.patch
123
A .patch extension is recommended instead of .bundle as many mail clients
124
will send the latter as a binary file. If a bundle would be too long or your
125
mailer mangles whitespace (e.g. implicitly converts Unix newlines to DOS
126
newlines), use the merge-directive command instead like this::
128
bzr merge-directive http://bazaar-vcs.org http://example.org/my_branch > my_directive.patch
130
See the help for details on the arguments to merge-directive.
132
Please do **NOT** put [PATCH] or [MERGE] in the subject line if you don't
133
want it to be merged. If you want comments from developers rather than
134
to be merged, you can put '[RFC]' in the subject line.
136
Anyone is welcome to review code. There are broadly three gates for
139
* Doesn't reduce test coverage: if it adds new methods or commands,
140
there should be tests for them. There is a good test framework
141
and plenty of examples to crib from, but if you are having trouble
142
working out how to test something feel free to post a draft patch
145
* Doesn't reduce design clarity, such as by entangling objects
146
we're trying to separate. This is mostly something the more
147
experienced reviewers need to help check.
149
* Improves bugs, features, speed, or code simplicity.
151
Code that goes in should pass all three. The core developers take care
152
to keep the code quality high and understandable while recognising that
153
perfect is sometimes the enemy of good. (It is easy for reviews to make
154
people notice other things which should be fixed but those things should
155
not hold up the original fix being accepted. New things can easily be
156
recorded in the Bug Tracker instead.)
158
Anyone can "vote" on the mailing list. Core developers can also vote using
159
Bundle Buggy. Here are the voting codes and their explanations.
161
:approve: Reviewer wants this submission merged.
162
:tweak: Reviewer wants this submission merged with small changes. (No
164
:abstain: Reviewer does not intend to vote on this patch.
165
:resubmit: Please make changes and resubmit for review.
166
:reject: Reviewer doesn't want this kind of change merged.
167
:comment: Not really a vote. Reviewer just wants to comment, for now.
169
If a change gets two approvals from core reviewers, and no rejections,
170
then it's OK to come in. Any of the core developers can bring it into the
171
bzr.dev trunk and backport it to maintenance branches if required. The
172
Release Manager will merge the change into the branch for a pending
173
release, if any. As a guideline, core developers usually merge their own
174
changes and volunteer to merge other contributions if they were the second
175
reviewer to agree to a change.
177
To track the progress of proposed changes, use Bundle Buggy. See
178
http://bundlebuggy.aaronbentley.com/help for a link to all the
179
outstanding merge requests together with an explanation of the columns.
180
Bundle Buggy will also mail you a link to track just your change.
183
Preparing a Sandbox for Making Changes to Bazaar
184
================================================
186
Bazaar supports many ways of organising your work. See
187
http://bazaar-vcs.org/SharedRepositoryLayouts for a summary of the
188
popular alternatives.
190
Of course, the best choice for you will depend on numerous factors:
191
the number of changes you may be making, the complexity of the changes, etc.
192
As a starting suggestion though:
194
* create a local copy of the main development branch (bzr.dev) by using
197
bzr branch http://bazaar-vcs.org/bzr/bzr.dev/ bzr.dev
199
* keep your copy of bzr.dev prestine (by not developing in it) and keep
200
it up to date (by using bzr pull)
202
* create a new branch off your local bzr.dev copy for each issue
203
(bug or feature) you are working on.
205
This approach makes it easy to go back and make any required changes
206
after a code review. Resubmitting the change is then simple with no
207
risk of accidentially including edits related to other issues you may
208
be working on. After the changes for an issue are accepted and merged,
209
the associated branch can be deleted or archived as you wish.
212
Navigating the Code Base
213
========================
215
TODO: List and describe in one line the purpose of each directory
216
inside an installation of bzr.
218
TODO: Refer to a central location holding an up to date copy of the API
219
documentation generated by epydoc, e.g. something like
220
http://starship.python.net/crew/mwh/bzrlibapi/bzrlib.html.
226
The Importance of Testing
227
=========================
229
Reliability is a critical success factor for any Version Control System.
230
We want Bazaar to be highly reliable across multiple platforms while
231
evolving over time to meet the needs of its community.
233
In a nutshell, this is want we expect and encourage:
13
235
* New functionality should have test cases. Preferably write the
14
236
test before writing the code.
16
238
In general, you can test at either the command-line level or the
17
internal API level. See Writing Tests below for more detail.
239
internal API level. See Writing tests below for more detail.
19
* Try to practice Test-Driven Development. before fixing a bug, write a
241
* Try to practice Test-Driven Development: before fixing a bug, write a
20
242
test case so that it does not regress. Similarly for adding a new
21
243
feature: write a test case for a small version of the new feature before
22
244
starting on the code itself. Check the test fails on the old code, then
23
245
add the feature or fix and check it passes.
25
* Exceptions should be defined inside bzrlib.errors, so that we can
26
see the whole tree at a glance.
247
By doing these things, the Bazaar team gets increased confidence that
248
changes do what they claim to do, whether provided by the core team or
249
by community members. Equally importantly, we can be surer that changes
250
down the track do not break new features or bug fixes that you are
253
As of May 2008, Bazaar ships with a test suite containing over 12000 tests
254
and growing. We are proud of it and want to remain so. As community
255
members, we all benefit from it. Would you trust version control on
256
your project to a product *without* a test suite like Bazaar has?
259
Running the Test Suite
260
======================
262
Currently, bzr selftest is used to invoke tests.
263
You can provide a pattern argument to run a subset. For example,
264
to run just the blackbox tests, run::
266
./bzr selftest -v blackbox
268
To skip a particular test (or set of tests), use the --exclude option
269
(shorthand -x) like so::
271
./bzr selftest -v -x blackbox
273
To ensure that all tests are being run and succeeding, you can use the
274
--strict option which will fail if there are any missing features or known
277
./bzr selftest --strict
279
To list tests without running them, use the --list-only option like so::
281
./bzr selftest --list-only
283
This option can be combined with other selftest options (like -x) and
284
filter patterns to understand their effect.
286
Once you understand how to create a list of tests, you can use the --load-list
287
option to run only a restricted set of tests that you kept in a file, one test
288
id by line. Keep in mind that this will never be sufficient to validate your
289
modifications, you still need to run the full test suite for that, but using it
290
can help in some cases (like running only the failed tests for some time)::
292
./bzr selftest -- load-list my_failing_tests
294
This option can also be combined with other selftest options, including
295
patterns. It has some drawbacks though, the list can become out of date pretty
296
quick when doing Test Driven Development.
298
To address this concern, there is another way to run a restricted set of tests:
299
the --starting-with option will run only the tests whose name starts with the
300
specified string. It will also avoid loading the other tests and as a
301
consequence starts running your tests quicker::
303
./bzr selftest --starting-with bzrlib.blackbox
305
This option can be combined with all the other selftest options including
306
--load-list. The later is rarely used but allows to run a subset of a list of
307
failing tests for example.
309
Test suite debug flags
310
----------------------
312
Similar to the global ``-Dfoo`` debug options, bzr selftest accepts
313
``-E=foo`` debug flags. These flags are:
315
:allow_debug: do *not* clear the global debug flags when running a test.
316
This can provide useful logging to help debug test failures when used
317
with e.g. ``bzr -Dhpss selftest -E=allow_debug``
323
In general tests should be placed in a file named test_FOO.py where
324
FOO is the logical thing under test. That file should be placed in the
325
tests subdirectory under the package being tested.
327
For example, tests for merge3 in bzrlib belong in bzrlib/tests/test_merge3.py.
328
See bzrlib/tests/test_sampler.py for a template test script.
330
Tests can be written for the UI or for individual areas of the library.
331
Choose whichever is appropriate: if adding a new command, or a new command
332
option, then you should be writing a UI test. If you are both adding UI
333
functionality and library functionality, you will want to write tests for
334
both the UI and the core behaviours. We call UI tests 'blackbox' tests
335
and they are found in ``bzrlib/tests/blackbox/*.py``.
337
When writing blackbox tests please honour the following conventions:
339
1. Place the tests for the command 'name' in
340
bzrlib/tests/blackbox/test_name.py. This makes it easy for developers
341
to locate the test script for a faulty command.
343
2. Use the 'self.run_bzr("name")' utility function to invoke the command
344
rather than running bzr in a subprocess or invoking the
345
cmd_object.run() method directly. This is a lot faster than
346
subprocesses and generates the same logging output as running it in a
347
subprocess (which invoking the method directly does not).
349
3. Only test the one command in a single test script. Use the bzrlib
350
library when setting up tests and when evaluating the side-effects of
351
the command. We do this so that the library api has continual pressure
352
on it to be as functional as the command line in a simple manner, and
353
to isolate knock-on effects throughout the blackbox test suite when a
354
command changes its name or signature. Ideally only the tests for a
355
given command are affected when a given command is changed.
357
4. If you have a test which does actually require running bzr in a
358
subprocess you can use ``run_bzr_subprocess``. By default the spawned
359
process will not load plugins unless ``--allow-plugins`` is supplied.
365
We make selective use of doctests__. In general they should provide
366
*examples* within the API documentation which can incidentally be tested. We
367
don't try to test every important case using doctests -- regular Python
368
tests are generally a better solution.
370
Most of these are in ``bzrlib/doc/api``. More additions are welcome.
372
__ http://docs.python.org/lib/module-doctest.html
375
Skipping tests and test requirements
376
------------------------------------
378
In our enhancements to unittest we allow for some addition results beyond
379
just success or failure.
381
If a test can't be run, it can say that it's skipped. This is typically
382
used in parameterized tests - for example if a transport doesn't support
383
setting permissions, we'll skip the tests that relating to that. ::
386
return self.branch_format.initialize(repo.bzrdir)
387
except errors.UninitializableFormat:
388
raise tests.TestSkipped('Uninitializable branch format')
390
Raising TestSkipped is a good idea when you want to make it clear that the
391
test was not run, rather than just returning which makes it look as if it
394
Several different cases are distinguished:
397
Generic skip; the only type that was present up to bzr 0.18.
400
The test doesn't apply to the parameters with which it was run.
401
This is typically used when the test is being applied to all
402
implementations of an interface, but some aspects of the interface
403
are optional and not present in particular concrete
404
implementations. (Some tests that should raise this currently
405
either silently return or raise TestSkipped.) Another option is
406
to use more precise parameterization to avoid generating the test
410
**(Not implemented yet)**
411
The test can't be run because of an inherent limitation of the
412
environment, such as not having symlinks or not supporting
416
The test can't be run because a dependency (typically a Python
417
library) is not available in the test environment. These
418
are in general things that the person running the test could fix
419
by installing the library. It's OK if some of these occur when
420
an end user runs the tests or if we're specifically testing in a
421
limited environment, but a full test should never see them.
424
The test exists but is known to fail, for example because the
425
code to fix it hasn't been run yet. Raising this allows
426
you to distinguish these failures from the ones that are not
427
expected to fail. This could be conditionally raised if something
428
is broken on some platforms but not on others.
430
We plan to support three modes for running the test suite to control the
431
interpretation of these results. Strict mode is for use in situations
432
like merges to the mainline and releases where we want to make sure that
433
everything that can be tested has been tested. Lax mode is for use by
434
developers who want to temporarily tolerate some known failures. The
435
default behaviour is obtained by ``bzr selftest`` with no options, and
436
also (if possible) by running under another unittest harness.
438
======================= ======= ======= ========
439
result strict default lax
440
======================= ======= ======= ========
441
TestSkipped pass pass pass
442
TestNotApplicable pass pass pass
443
TestPlatformLimit pass pass pass
444
TestDependencyMissing fail pass pass
445
KnownFailure fail pass pass
446
======================= ======= ======= ========
449
Test feature dependencies
450
-------------------------
452
Rather than manually checking the environment in each test, a test class
453
can declare its dependence on some test features. The feature objects are
454
checked only once for each run of the whole test suite.
456
For historical reasons, as of May 2007 many cases that should depend on
457
features currently raise TestSkipped.)
461
class TestStrace(TestCaseWithTransport):
463
_test_needs_features = [StraceFeature]
465
This means all tests in this class need the feature. The feature itself
466
should provide a ``_probe`` method which is called once to determine if
469
These should generally be equivalent to either TestDependencyMissing or
470
sometimes TestPlatformLimit.
476
Known failures are when a test exists but we know it currently doesn't
477
work, allowing the test suite to still pass. These should be used with
478
care, we don't want a proliferation of quietly broken tests. It might be
479
appropriate to use them if you've committed a test for a bug but not the
480
fix for it, or if something works on Unix but not on Windows.
483
Testing exceptions and errors
484
-----------------------------
486
It's important to test handling of errors and exceptions. Because this
487
code is often not hit in ad-hoc testing it can often have hidden bugs --
488
it's particularly common to get NameError because the exception code
489
references a variable that has since been renamed.
491
.. TODO: Something about how to provoke errors in the right way?
493
In general we want to test errors at two levels:
495
1. A test in ``test_errors.py`` checking that when the exception object is
496
constructed with known parameters it produces an expected string form.
497
This guards against mistakes in writing the format string, or in the
498
``str`` representations of its parameters. There should be one for
499
each exception class.
501
2. Tests that when an api is called in a particular situation, it raises
502
an error of the expected class. You should typically use
503
``assertRaises``, which in the Bazaar test suite returns the exception
504
object to allow you to examine its parameters.
506
In some cases blackbox tests will also want to check error reporting. But
507
it can be difficult to provoke every error through the commandline
508
interface, so those tests are only done as needed -- eg in response to a
509
particular bug or if the error is reported in an unusual way(?) Blackbox
510
tests should mostly be testing how the command-line interface works, so
511
should only test errors if there is something particular to the cli in how
512
they're displayed or handled.
518
The Python ``warnings`` module is used to indicate a non-fatal code
519
problem. Code that's expected to raise a warning can be tested through
522
The test suite can be run with ``-Werror`` to check no unexpected errors
525
However, warnings should be used with discretion. It's not an appropriate
526
way to give messages to the user, because the warning is normally shown
527
only once per source line that causes the problem. You should also think
528
about whether the warning is serious enought that it should be visible to
529
users who may not be able to fix it.
532
Interface implementation testing and test scenarios
533
---------------------------------------------------
535
There are several cases in Bazaar of multiple implementations of a common
536
conceptual interface. ("Conceptual" because
537
it's not necessary for all the implementations to share a base class,
538
though they often do.) Examples include transports and the working tree,
539
branch and repository classes.
541
In these cases we want to make sure that every implementation correctly
542
fulfils the interface requirements. For example, every Transport should
543
support the ``has()`` and ``get()`` and ``clone()`` methods. We have a
544
sub-suite of tests in ``test_transport_implementations``. (Most
545
per-implementation tests are in submodules of ``bzrlib.tests``, but not
546
the transport tests at the moment.)
548
These tests are repeated for each registered Transport, by generating a
549
new TestCase instance for the cross product of test methods and transport
550
implementations. As each test runs, it has ``transport_class`` and
551
``transport_server`` set to the class it should test. Most tests don't
552
access these directly, but rather use ``self.get_transport`` which returns
553
a transport of the appropriate type.
555
The goal is to run per-implementation only tests that relate to that
556
particular interface. Sometimes we discover a bug elsewhere that happens
557
with only one particular transport. Once it's isolated, we can consider
558
whether a test should be added for that particular implementation,
559
or for all implementations of the interface.
561
The multiplication of tests for different implementations is normally
562
accomplished by overriding the ``test_suite`` function used to load
563
tests from a module. This function typically loads all the tests,
564
then applies a TestProviderAdapter to them, which generates a longer
565
suite containing all the test variations.
571
Some utilities are provided for generating variations of tests. This can
572
be used for per-implementation tests, or other cases where the same test
573
code needs to run several times on different scenarios.
575
The general approach is to define a class that provides test methods,
576
which depend on attributes of the test object being pre-set with the
577
values to which the test should be applied. The test suite should then
578
also provide a list of scenarios in which to run the tests.
580
Typically ``multiply_tests_from_modules`` should be called from the test
581
module's ``test_suite`` function.
584
Essential Domain Classes
585
########################
587
Introducing the Object Model
588
============================
590
The core domain objects within the bazaar model are:
600
Transports are explained below. See http://bazaar-vcs.org/Classes/
601
for an introduction to the other key classes.
606
The ``Transport`` layer handles access to local or remote directories.
607
Each Transport object acts like a logical connection to a particular
608
directory, and it allows various operations on files within it. You can
609
*clone* a transport to get a new Transport connected to a subdirectory or
612
Transports are not used for access to the working tree. At present
613
working trees are always local and they are accessed through the regular
614
Python file io mechanisms.
619
Transports work in URLs. Take note that URLs are by definition only
620
ASCII - the decision of how to encode a Unicode string into a URL must be
621
taken at a higher level, typically in the Store. (Note that Stores also
622
escape filenames which cannot be safely stored on all filesystems, but
623
this is a different level.)
625
The main reason for this is that it's not possible to safely roundtrip a
626
URL into Unicode and then back into the same URL. The URL standard
627
gives a way to represent non-ASCII bytes in ASCII (as %-escapes), but
628
doesn't say how those bytes represent non-ASCII characters. (They're not
629
guaranteed to be UTF-8 -- that is common but doesn't happen everywhere.)
631
For example if the user enters the url ``http://example/%e0`` there's no
632
way to tell whether that character represents "latin small letter a with
633
grave" in iso-8859-1, or "latin small letter r with acute" in iso-8859-2
634
or malformed UTF-8. So we can't convert their URL to Unicode reliably.
636
Equally problematic if we're given a url-like string containing non-ascii
637
characters (such as the accented a) we can't be sure how to convert that
638
to the correct URL, because we don't know what encoding the server expects
639
for those characters. (Although this is not totally reliable we might still
640
accept these and assume they should be put into UTF-8.)
642
A similar edge case is that the url ``http://foo/sweet%2Fsour`` contains
643
one directory component whose name is "sweet/sour". The escaped slash is
644
not a directory separator. If we try to convert URLs to regular Unicode
645
paths this information will be lost.
647
This implies that Transports must natively deal with URLs; for simplicity
648
they *only* deal with URLs and conversion of other strings to URLs is done
649
elsewhere. Information they return, such as from ``list_dir``, is also in
650
the form of URL components.
653
Coding Style Guidelines
654
#######################
659
``hasattr`` should not be used because it swallows exceptions including
660
``KeyboardInterrupt``. Instead, say something like ::
662
if getattr(thing, 'name', None) is None
668
Please write PEP-8__ compliant code.
670
__ http://www.python.org/peps/pep-0008.html
672
One often-missed requirement is that the first line of docstrings
673
should be a self-contained one-sentence summary.
675
We use 4 space indents for blocks, and never use tab characters. (In vim,
678
Lines should be no more than 79 characters if at all possible.
679
Lines that continue a long statement may be indented in either of
682
within the parenthesis or other character that opens the block, e.g.::
688
or indented by four spaces::
694
The first is considered clearer by some people; however it can be a bit
695
harder to maintain (e.g. when the method name changes), and it does not
696
work well if the relevant parenthesis is already far to the right. Avoid
699
self.legbone.kneebone.shinbone.toebone.shake_it(one,
705
self.legbone.kneebone.shinbone.toebone.shake_it(one,
711
self.legbone.kneebone.shinbone.toebone.shake_it(
714
For long lists, we like to add a trailing comma and put the closing
715
character on the following line. This makes it easier to add new items in
718
from bzrlib.goo import (
724
There should be spaces between function paramaters, but not between the
725
keyword name and the value::
727
call(1, 3, cheese=quark)
731
;(defface my-invalid-face
732
; '((t (:background "Red" :underline t)))
733
; "Face used to highlight invalid constructs or other uglyties"
736
(defun my-python-mode-hook ()
737
;; setup preferred indentation style.
738
(setq fill-column 79)
739
(setq indent-tabs-mode nil) ; no tabs, never, I will not repeat
740
; (font-lock-add-keywords 'python-mode
741
; '(("^\\s *\t" . 'my-invalid-face) ; Leading tabs
742
; ("[ \t]+$" . 'my-invalid-face) ; Trailing spaces
743
; ("^[ \t]+$" . 'my-invalid-face)); Spaces only
747
(add-hook 'python-mode-hook 'my-python-mode-hook)
749
The lines beginning with ';' are comments. They can be activated
750
if one want to have a strong notice of some tab/space usage
28
757
* Imports should be done at the top-level of the file, unless there is
29
758
a strong reason to have them lazily loaded when a particular
33
762
* Module names should always be given fully-qualified,
34
763
i.e. ``bzrlib.hashcache`` not just ``hashcache``.
36
* Commands should return non-zero when they encounter circumstances that
37
the user should really pay attention to - which includes trivial shell
40
Recommended values are
42
1- Conflicts in merge-like operations, or changes are present in
44
2- Unrepresentable diff changes (i.e. binary files that we cannot show
46
3- An error or exception has occurred.
769
Functions, methods or members that are "private" to bzrlib are given
770
a leading underscore prefix. Names without a leading underscore are
771
public not just across modules but to programmers using bzrlib as an
772
API. As a consequence, a leading underscore is appropriate for names
773
exposed across modules but that are not to be exposed to bzrlib API
776
We prefer class names to be concatenated capital words (``TestCase``)
777
and variables, methods and functions to be lowercase words joined by
778
underscores (``revision_id``, ``get_revision``).
780
For the purposes of naming some names are treated as single compound
781
words: "filename", "revno".
783
Consider naming classes as nouns and functions/methods as verbs.
785
Try to avoid using abbreviations in names, because there can be
786
inconsistency if other people use the full name.
792
``revision_id`` not ``rev_id`` or ``revid``
794
Functions that transform one thing to another should be named ``x_to_y``
795
(not ``x2y`` as occurs in some old code.)
801
Python destructors (``__del__``) work differently to those of other
802
languages. In particular, bear in mind that destructors may be called
803
immediately when the object apparently becomes unreferenced, or at some
804
later time, or possibly never at all. Therefore we have restrictions on
805
what can be done inside them.
807
0. If you think you need to use a ``__del__`` method ask another
808
developer for alternatives. If you do need to use one, explain
811
1. Never rely on a ``__del__`` method running. If there is code that
812
must run, do it from a ``finally`` block instead.
814
2. Never ``import`` from inside a ``__del__`` method, or you may crash the
817
3. In some places we raise a warning from the destructor if the object
818
has not been cleaned up or closed. This is considered OK: the warning
819
may not catch every case but it's still useful sometimes.
825
In some places we have variables which point to callables that construct
826
new instances. That is to say, they can be used a lot like class objects,
827
but they shouldn't be *named* like classes:
829
> I think that things named FooBar should create instances of FooBar when
830
> called. Its plain confusing for them to do otherwise. When we have
831
> something that is going to be used as a class - that is, checked for via
832
> isinstance or other such idioms, them I would call it foo_class, so that
833
> it is clear that a callable is not sufficient. If it is only used as a
834
> factory, then yes, foo_factory is what I would use.
840
Several places in Bazaar use (or will use) a registry, which is a
841
mapping from names to objects or classes. The registry allows for
842
loading in registered code only when it's needed, and keeping
843
associated information such as a help string or description.
849
To make startup time faster, we use the ``bzrlib.lazy_import`` module to
850
delay importing modules until they are actually used. ``lazy_import`` uses
851
the same syntax as regular python imports. So to import a few modules in a
854
from bzrlib.lazy_import import lazy_import
855
lazy_import(globals(), """
864
revision as _mod_revision,
866
import bzrlib.transport
870
At this point, all of these exist as a ``ImportReplacer`` object, ready to
871
be imported once a member is accessed. Also, when importing a module into
872
the local namespace, which is likely to clash with variable names, it is
873
recommended to prefix it as ``_mod_<module>``. This makes it clearer that
874
the variable is a module, and these object should be hidden anyway, since
875
they shouldn't be imported into other namespaces.
877
While it is possible for ``lazy_import()`` to import members of a module
878
when using the ``from module import member`` syntax, it is recommended to
879
only use that syntax to load sub modules ``from module import submodule``.
880
This is because variables and classes can frequently be used without
881
needing a sub-member for example::
883
lazy_import(globals(), """
884
from module import MyClass
888
return isinstance(x, MyClass)
890
This will incorrectly fail, because ``MyClass`` is a ``ImportReplacer``
891
object, rather than the real class.
893
It also is incorrect to assign ``ImportReplacer`` objects to other variables.
894
Because the replacer only knows about the original name, it is unable to
895
replace other variables. The ``ImportReplacer`` class will raise an
896
``IllegalUseOfScopeReplacer`` exception if it can figure out that this
897
happened. But it requires accessing a member more than once from the new
898
variable, so some bugs are not detected right away.
904
The null revision is the ancestor of all revisions. Its revno is 0, its
905
revision-id is ``null:``, and its tree is the empty tree. When referring
906
to the null revision, please use ``bzrlib.revision.NULL_REVISION``. Old
907
code sometimes uses ``None`` for the null revision, but this practice is
911
Object string representations
912
=============================
914
Python prints objects using their ``__repr__`` method when they are
915
written to logs, exception tracebacks, or the debugger. We want
916
objects to have useful representations to help in determining what went
919
If you add a new class you should generally add a ``__repr__`` method
920
unless there is an adequate method in a parent class. There should be a
923
Representations should typically look like Python constructor syntax, but
924
they don't need to include every value in the object and they don't need
925
to be able to actually execute. They're to be read by humans, not
926
machines. Don't hardcode the classname in the format, so that we get the
927
correct value if the method is inherited by a subclass. If you're
928
printing attributes of the object, including strings, you should normally
929
use ``%r`` syntax (to call their repr in turn).
931
Try to avoid the representation becoming more than one or two lines long.
932
(But balance this against including useful information, and simplicity of
935
Because repr methods are often called when something has already gone
936
wrong, they should be written more defensively than most code. The object
937
may be half-initialized or in some other way in an illegal state. The
938
repr method shouldn't raise an exception, or it may hide the (probably
939
more useful) underlying exception.
945
return '%s(%r)' % (self.__class__.__name__,
948
return 'FooObject(**unprintable**)'
51
957
We have a commitment to 6 months API stability - any supported symbol in a
52
958
release of bzr MUST NOT be altered in any way that would result in
67
973
details for you - such as updating the docstring, and issuing a warning
68
974
when the old api is used.
70
For unsupported API's, it does not hurt to follow this discipline, but its
976
For unsupported API's, it does not hurt to follow this discipline, but it's
71
977
not required. Minimally though, please try to rename things so that
72
978
callers will at least get an AttributeError rather than weird results.
75
Standard parameter types
981
Deprecation decorators
982
----------------------
984
``bzrlib.symbol_versioning`` provides decorators that can be attached to
985
methods, functions, and other interfaces to indicate that they should no
986
longer be used. For example::
988
@deprecated_method(deprecated_in((0, 1, 4)))
990
return self._new_foo()
992
To deprecate a static method you must call ``deprecated_function``
993
(**not** method), after the staticmethod call::
996
@deprecated_function(deprecated_in((0, 1, 4)))
997
def create_repository(base, shared=False, format=None):
999
When you deprecate an API, you should not just delete its tests, because
1000
then we might introduce bugs in them. If the API is still present at all,
1001
it should still work. The basic approach is to use
1002
``TestCase.applyDeprecated`` which in one step checks that the API gives
1003
the expected deprecation message, and also returns the real result from
1004
the method, so that tests can keep running.
1006
Deprecation warnings will be suppressed for final releases, but not for
1007
development versions or release candidates, or when running ``bzr
1008
selftest``. This gives developers information about whether their code is
1009
using deprecated functions, but avoids confusing users about things they
1016
Processing Command Lines
1017
------------------------
1019
bzrlib has a standard framework for parsing command lines and calling
1020
processing routines associated with various commands. See builtins.py
1021
for numerous examples.
1024
Standard Parameter Types
76
1025
------------------------
78
1027
There are some common requirements in the library: some parameters need to be
79
1028
unicode safe, some need byte strings, and so on. At the moment we have
80
1029
only codified one specific pattern: Parameters that need to be unicode
81
should be check via 'bzrlib.osutils.safe_unicode'. This will coerce the
1030
should be checked via ``bzrlib.osutils.safe_unicode``. This will coerce the
82
1031
input into unicode in a consistent fashion, allowing trivial strings to be
83
1032
used for programmer convenience, but not performing unpredictably in the
84
1033
presence of different locales.
89
If you change the behaviour of a command, please update its docstring
90
in bzrlib/commands.py. This is displayed by the 'bzr help' command.
1039
(The strategy described here is what we want to get to, but it's not
1040
consistently followed in the code at the moment.)
1042
bzrlib is intended to be a generically reusable library. It shouldn't
1043
write messages to stdout or stderr, because some programs that use it
1044
might want to display that information through a GUI or some other
1047
We can distinguish two types of output from the library:
1049
1. Structured data representing the progress or result of an
1050
operation. For example, for a commit command this will be a list
1051
of the modified files and the finally committed revision number
1054
These should be exposed either through the return code or by calls
1055
to a callback parameter.
1057
A special case of this is progress indicators for long-lived
1058
operations, where the caller should pass a ProgressBar object.
1060
2. Unstructured log/debug messages, mostly for the benefit of the
1061
developers or users trying to debug problems. This should always
1062
be sent through ``bzrlib.trace`` and Python ``logging``, so that
1063
it can be redirected by the client.
1065
The distinction between the two is a bit subjective, but in general if
1066
there is any chance that a library would want to see something as
1067
structured data, we should make it so.
1069
The policy about how output is presented in the text-mode client
1070
should be only in the command-line tool.
1077
Bazaar has online help for various topics through ``bzr help COMMAND`` or
1078
equivalently ``bzr command -h``. We also have help on command options,
1079
and on other help topics. (See ``help_topics.py``.)
1081
As for python docstrings, the first paragraph should be a single-sentence
1082
synopsis of the command.
1084
The help for options should be one or more proper sentences, starting with
1085
a capital letter and finishing with a full stop (period).
1087
All help messages and documentation should have two spaces between
1094
In general tests should be placed in a file named test_FOO.py where
1095
FOO is the logical thing under test. That file should be placed in the
1096
tests subdirectory under the package being tested.
1098
For example, tests for merge3 in bzrlib belong in bzrlib/tests/test_merge3.py.
1099
See bzrlib/tests/test_sampler.py for a template test script.
1101
Tests can be written for the UI or for individual areas of the library.
1102
Choose whichever is appropriate: if adding a new command, or a new command
1103
option, then you should be writing a UI test. If you are both adding UI
1104
functionality and library functionality, you will want to write tests for
1105
both the UI and the core behaviours. We call UI tests 'blackbox' tests
1106
and they are found in ``bzrlib/tests/blackbox/*.py``.
1108
When writing blackbox tests please honour the following conventions:
1110
1. Place the tests for the command 'name' in
1111
bzrlib/tests/blackbox/test_name.py. This makes it easy for developers
1112
to locate the test script for a faulty command.
1114
2. Use the 'self.run_bzr("name")' utility function to invoke the command
1115
rather than running bzr in a subprocess or invoking the
1116
cmd_object.run() method directly. This is a lot faster than
1117
subprocesses and generates the same logging output as running it in a
1118
subprocess (which invoking the method directly does not).
1120
3. Only test the one command in a single test script. Use the bzrlib
1121
library when setting up tests and when evaluating the side-effects of
1122
the command. We do this so that the library api has continual pressure
1123
on it to be as functional as the command line in a simple manner, and
1124
to isolate knock-on effects throughout the blackbox test suite when a
1125
command changes its name or signature. Ideally only the tests for a
1126
given command are affected when a given command is changed.
1128
4. If you have a test which does actually require running bzr in a
1129
subprocess you can use ``run_bzr_subprocess``. By default the spawned
1130
process will not load plugins unless ``--allow-plugins`` is supplied.
1136
We have a rich collection of tools to support writing tests. Please use
1137
them in preference to ad-hoc solutions as they provide portability and
1138
performance benefits.
1143
The ``TreeBuilder`` interface allows the construction of arbitrary trees
1144
with a declarative interface. A sample session might look like::
1146
tree = self.make_branch_and_tree('path')
1147
builder = TreeBuilder()
1148
builder.start_tree(tree)
1149
builder.build(['foo', "bar/", "bar/file"])
1150
tree.commit('commit the tree')
1151
builder.finish_tree()
1153
Please see bzrlib.treebuilder for more details.
1158
The ``BranchBuilder`` interface allows the creation of test branches in a
1159
quick and easy manner. A sample session::
1161
builder = BranchBuilder(self.get_transport().clone('relpath'))
1162
builder.build_commit()
1163
builder.build_commit()
1164
builder.build_commit()
1165
branch = builder.get_branch()
1167
Please see bzrlib.branchbuilder for more details.
1172
We make selective use of doctests__. In general they should provide
1173
*examples* within the API documentation which can incidentally be tested. We
1174
don't try to test every important case using doctests -- regular Python
1175
tests are generally a better solution.
1177
Most of these are in ``bzrlib/doc/api``. More additions are welcome.
1179
__ http://docs.python.org/lib/module-doctest.html
1184
Currently, bzr selftest is used to invoke tests.
1185
You can provide a pattern argument to run a subset. For example,
1186
to run just the blackbox tests, run::
1188
./bzr selftest -v blackbox
1190
To skip a particular test (or set of tests), use the --exclude option
1191
(shorthand -x) like so::
1193
./bzr selftest -v -x blackbox
1195
To list tests without running them, use the --list-only option like so::
1197
./bzr selftest --list-only
1199
This option can be combined with other selftest options (like -x) and
1200
filter patterns to understand their effect.
1203
Handling Errors and Exceptions
1204
==============================
1206
Commands should return non-zero when they encounter circumstances that
1207
the user should really pay attention to - which includes trivial shell
1210
Recommended values are:
1213
1. Conflicts in merge-like operations, or changes are present in
1214
diff-like operations.
1215
2. Unrepresentable diff changes (i.e. binary files that we cannot show
1217
3. An error or exception has occurred.
1218
4. An internal error occurred (one that shows a traceback.)
1220
Errors are handled through Python exceptions. Exceptions should be defined
1221
inside bzrlib.errors, so that we can see the whole tree at a glance.
1223
We broadly classify errors as either being either internal or not,
1224
depending on whether ``internal_error`` is set or not. If we think it's our
1225
fault, we show a backtrace, an invitation to report the bug, and possibly
1226
other details. This is the default for errors that aren't specifically
1227
recognized as being caused by a user error. Otherwise we show a briefer
1228
message, unless -Derror was given.
1230
Many errors originate as "environmental errors" which are raised by Python
1231
or builtin libraries -- for example IOError. These are treated as being
1232
our fault, unless they're caught in a particular tight scope where we know
1233
that they indicate a user errors. For example if the repository format
1234
is not found, the user probably gave the wrong path or URL. But if one of
1235
the files inside the repository is not found, then it's our fault --
1236
either there's a bug in bzr, or something complicated has gone wrong in
1237
the environment that means one internal file was deleted.
1239
Many errors are defined in ``bzrlib/errors.py`` but it's OK for new errors
1240
to be added near the place where they are used.
1242
Exceptions are formatted for the user by conversion to a string
1243
(eventually calling their ``__str__`` method.) As a convenience the
1244
``._fmt`` member can be used as a template which will be mapped to the
1245
error's instance dict.
1247
New exception classes should be defined when callers might want to catch
1248
that exception specifically, or when it needs a substantially different
1251
Exception strings should start with a capital letter and should not have a
1252
final fullstop. If long, they may contain newlines to break the text.
1258
Do not use the Python ``assert`` statement, either in tests or elsewhere.
1259
A source test checks that it is not used. It is ok to explicitly raise
1264
* It makes the behaviour vary depending on whether bzr is run with -O
1265
or not, therefore giving a chance for bugs that occur in one case or
1266
the other, several of which have already occurred: assertions with
1267
side effects, code which can't continue unless the assertion passes,
1268
cases where we should give the user a proper message rather than an
1270
* It's not that much shorter than an explicit if/raise.
1271
* It tends to lead to fuzzy thinking about whether the check is
1272
actually needed or not, and whether it's an internal error or not
1273
* It tends to cause look-before-you-leap patterns.
1274
* It's unsafe if the check is needed to protect the integrity of the
1276
* It tends to give poor messages since the developer can get by with
1277
no explanatory text at all.
1278
* We can't rely on people always running with -O in normal use, so we
1279
can't use it for tests that are actually expensive.
1280
* Expensive checks that help developers are better turned on from the
1281
test suite or a -D flag.
1282
* If used instead of ``self.assert*()`` in tests it makes them falsely pass with -O.
1288
When you change bzrlib, please update the relevant documentation for the
1289
change you made: Changes to commands should update their help, and
1290
possibly end user tutorials; changes to the core library should be
1291
reflected in API documentation.
95
1296
If you make a user-visible change, please add a note to the NEWS file.
133
1343
.. _epydoc: http://epydoc.sourceforge.net/
140
Please write PEP-8__ compliant code.
142
One often-missed requirement is that the first line of docstrings
143
should be a self-contained one-sentence summary.
145
__ http://www.python.org/peps/pep-0008.html
152
Functions, methods or members that are in some sense "private" are given
153
a leading underscore prefix. This is just a hint that code outside the
154
implementation should probably not use that interface.
156
We prefer class names to be concatenated capital words (``TestCase``)
157
and variables, methods and functions to be lowercase words joined by
158
underscores (``revision_id``, ``get_revision``).
160
For the purposes of naming some names are treated as single compound
161
words: "filename", "revno".
163
Consider naming classes as nouns and functions/methods as verbs.
169
``revision_id`` not ``rev_id`` or ``revid``
171
Functions that transform one thing to another should be named ``x_to_y``
172
(not ``x2y`` as occurs in some old code.)
178
Python destructors (``__del__``) work differently to those of other
179
languages. In particular, bear in mind that destructors may be called
180
immediately when the object apparently becomes unreferenced, or at some
181
later time, or possibly never at all. Therefore we have restrictions on
182
what can be done inside them.
184
0. Never use a __del__ method without asking Martin/Robert first.
186
1. Never rely on a ``__del__`` method running. If there is code that
187
must run, do it from a ``finally`` block instead.
189
2. Never ``import`` from inside a ``__del__`` method, or you may crash the
192
3. In some places we raise a warning from the destructor if the object
193
has not been cleaned up or closed. This is considered OK: the warning
194
may not catch every case but it's still useful sometimes.
200
(The strategy described here is what we want to get to, but it's not
201
consistently followed in the code at the moment.)
203
bzrlib is intended to be a generically reusable library. It shouldn't
204
write messages to stdout or stderr, because some programs that use it
205
might want to display that information through a GUI or some other
208
We can distinguish two types of output from the library:
210
1. Structured data representing the progress or result of an
211
operation. For example, for a commit command this will be a list
212
of the modified files and the finally committed revision number
215
These should be exposed either through the return code or by calls
216
to a callback parameter.
218
A special case of this is progress indicators for long-lived
219
operations, where the caller should pass a ProgressBar object.
221
2. Unstructured log/debug messages, mostly for the benefit of the
222
developers or users trying to debug problems. This should always
223
be sent through ``bzrlib.trace`` and Python ``logging``, so that
224
it can be redirected by the client.
226
The distinction between the two is a bit subjective, but in general if
227
there is any chance that a library would want to see something as
228
structured data, we should make it so.
230
The policy about how output is presented in the text-mode client
231
should be only in the command-line tool.
236
In general tests should be placed in a file named test_FOO.py where
237
FOO is the logical thing under test. That file should be placed in the
238
tests subdirectory under the package being tested.
240
For example, tests for merge3 in bzrlib belong in bzrlib/tests/test_merge3.py.
241
See bzrlib/selftest/test_sampler.py for a template test script.
243
Tests can be written for the UI or for individual areas of the library.
244
Choose whichever is appropriate: if adding a new command, or a new command
245
option, then you should be writing a UI test. If you are both adding UI
246
functionality and library functionality, you will want to write tests for
247
both the UI and the core behaviours. We call UI tests 'blackbox' tests
248
and they are found in bzrlib/tests/blackbox/*.py.
250
When writing blackbox tests please honour the following conventions:
252
1. Place the tests for the command 'name' in
253
bzrlib/tests/blackbox/test_name.py. This makes it easy for developers
254
to locate the test script for a faulty command.
256
2. Use the 'self.run_bzr("name")' utility function to invoke the command
257
rather than running bzr in a subprocess or invoking the
258
cmd_object.run() method directly. This is a lot faster than
259
subprocesses and generates the same logging output as running it in a
260
subprocess (which invoking the method directly does not).
262
3. Only test the one command in a single test script. Use the bzrlib
263
library when setting up tests and when evaluating the side-effects of
264
the command. We do this so that the library api has continual pressure
265
on it to be as functional as the command line in a simple manner, and
266
to isolate knock-on effects throughout the blackbox test suite when a
267
command changes it name or signature. Ideally only the tests for a
268
given command are affected when a given command is changed.
272
Currently, bzr selftest is used to invoke tests.
273
You can provide a pattern argument to run a subset. For example,
274
to run just the blackbox tests, run::
276
./bzr selftest -v blackbox
279
Errors and exceptions
280
=====================
282
Errors are handled through Python exceptions. They can represent user
283
errors, environmental errors or program bugs. Sometimes we can't be sure
284
at the time it's raised which case applies. See bzrlib/errors.py for
285
details on the error-handling practices.
1352
The copyright policy for bzr was recently made clear in this email (edited
1353
for grammatical correctness)::
1355
The attached patch cleans up the copyright and license statements in
1356
the bzr source. It also adds tests to help us remember to add them
1357
with the correct text.
1359
We had the problem that lots of our files were "Copyright Canonical
1360
Development Ltd" which is not a real company, and some other variations
1361
on this theme. Also, some files were missing the GPL statements.
1363
I want to be clear about the intent of this patch, since copyright can
1364
be a little controversial.
1366
1) The big motivation for this is not to shut out the community, but
1367
just to clean up all of the invalid copyright statements.
1369
2) It has been the general policy for bzr that we want a single
1370
copyright holder for all of the core code. This is following the model
1371
set by the FSF, which makes it easier to update the code to a new
1372
license in case problems are encountered. (For example, if we want to
1373
upgrade the project universally to GPL v3 it is much simpler if there is
1374
a single copyright holder). It also makes it clearer if copyright is
1375
ever debated, there is a single holder, which makes it easier to defend
1376
in court, etc. (I think the FSF position is that if you assign them
1377
copyright, they can defend it in court rather than you needing to, and
1378
I'm sure Canonical would do the same).
1379
As such, Canonical has requested copyright assignments from all of the
1382
3) If someone wants to add code and not attribute it to Canonical, there
1383
is a specific list of files that are excluded from this check. And the
1384
test failure indicates where that is, and how to update it.
1386
4) If anyone feels that I changed a copyright statement incorrectly, just
1387
let me know, and I'll be happy to correct it. Whenever you have large
1388
mechanical changes like this, it is possible to make some mistakes.
1390
Just to reiterate, this is a community project, and it is meant to stay
1391
that way. Core bzr code is copyright Canonical for legal reasons, and
1392
the tests are just there to help us maintain that.
1395
Miscellaneous Topics
1396
####################
1401
Bazaar has a few facilities to help debug problems by going into pdb_, the
1404
.. _pdb: http://docs.python.org/lib/debugger-commands.html
1406
If the ``BZR_PDB`` environment variable is set
1407
then bzr will go into pdb post-mortem mode when an unhandled exception
1410
If you send a SIGQUIT signal to bzr, which can be done by pressing
1411
Ctrl-\\ on Unix, bzr will go into the debugger immediately. You can
1412
continue execution by typing ``c``. This can be disabled if necessary
1413
by setting the environment variable ``BZR_SIGQUIT_PDB=0``.
294
1422
indexes into the branch's revision history.
1425
Unicode and Encoding Support
1426
============================
1428
This section discusses various techniques that Bazaar uses to handle
1429
characters that are outside the ASCII set.
1434
When a ``Command`` object is created, it is given a member variable
1435
accessible by ``self.outf``. This is a file-like object, which is bound to
1436
``sys.stdout``, and should be used to write information to the screen,
1437
rather than directly writing to ``sys.stdout`` or calling ``print``.
1438
This file has the ability to translate Unicode objects into the correct
1439
representation, based on the console encoding. Also, the class attribute
1440
``encoding_type`` will effect how unprintable characters will be
1441
handled. This parameter can take one of 3 values:
1444
Unprintable characters will be represented with a suitable replacement
1445
marker (typically '?'), and no exception will be raised. This is for
1446
any command which generates text for the user to review, rather than
1447
for automated processing.
1448
For example: ``bzr log`` should not fail if one of the entries has text
1449
that cannot be displayed.
1452
Attempting to print an unprintable character will cause a UnicodeError.
1453
This is for commands that are intended more as scripting support, rather
1454
than plain user review.
1455
For exampl: ``bzr ls`` is designed to be used with shell scripting. One
1456
use would be ``bzr ls --null --unknows | xargs -0 rm``. If ``bzr``
1457
printed a filename with a '?', the wrong file could be deleted. (At the
1458
very least, the correct file would not be deleted). An error is used to
1459
indicate that the requested action could not be performed.
1462
Do not attempt to automatically convert Unicode strings. This is used
1463
for commands that must handle conversion themselves.
1464
For example: ``bzr diff`` needs to translate Unicode paths, but should
1465
not change the exact text of the contents of the files.
1468
``bzrlib.urlutils.unescape_for_display``
1469
----------------------------------------
1471
Because Transports work in URLs (as defined earlier), printing the raw URL
1472
to the user is usually less than optimal. Characters outside the standard
1473
set are printed as escapes, rather than the real character, and local
1474
paths would be printed as ``file://`` urls. The function
1475
``unescape_for_display`` attempts to unescape a URL, such that anything
1476
that cannot be printed in the current encoding stays an escaped URL, but
1477
valid characters are generated where possible.
1483
The ``bzrlib.osutils`` module has many useful helper functions, including
1484
some more portable variants of functions in the standard library.
1486
In particular, don't use ``shutil.rmtree`` unless it's acceptable for it
1487
to fail on Windows if some files are readonly or still open elsewhere.
1488
Use ``bzrlib.osutils.rmtree`` instead.
1494
We write some extensions in C using pyrex. We design these to work in
1497
* User with no C compiler
1498
* User with C compiler
1501
The recommended way to install bzr is to have a C compiler so that the
1502
extensions can be built, but if no C compiler is present, the pure python
1503
versions we supply will work, though more slowly.
1505
For developers we recommend that pyrex be installed, so that the C
1506
extensions can be changed if needed.
1508
For the C extensions, the extension module should always match the
1509
original python one in all respects (modulo speed). This should be
1510
maintained over time.
1512
To create an extension, add rules to setup.py for building it with pyrex,
1513
and with distutils. Now start with an empty .pyx file. At the top add
1514
"include 'yourmodule.py'". This will import the contents of foo.py into this
1515
file at build time - remember that only one module will be loaded at
1516
runtime. Now you can subclass classes, or replace functions, and only your
1517
changes need to be present in the .pyx file.
1519
Note that pyrex does not support all 2.4 programming idioms, so some
1520
syntax changes may be required. I.e.
1522
- 'from foo import (bar, gam)' needs to change to not use the brackets.
1523
- 'import foo.bar as bar' needs to be 'import foo.bar; bar = foo.bar'
1525
If the changes are too dramatic, consider
1526
maintaining the python code twice - once in the .pyx, and once in the .py,
1527
and no longer including the .py file.
1530
Making Installers for OS Windows
1531
================================
1532
To build a win32 installer, see the instructions on the wiki page:
1533
http://bazaar-vcs.org/BzrWin32Installer
1536
Core Developer Tasks
1537
####################
1542
What is a Core Developer?
1543
-------------------------
1545
While everyone in the Bazaar community is welcome and encouraged to
1546
propose and submit changes, a smaller team is reponsible for pulling those
1547
changes together into a cohesive whole. In addition to the general developer
1548
stuff covered above, "core" developers have responsibility for:
1551
* reviewing blueprints
1553
* managing releases.
1556
Removing barriers to community participation is a key reason for adopting
1557
distributed VCS technology. While DVCS removes many technical barriers,
1558
a small number of social barriers are often necessary instead.
1559
By documenting how the above things are done, we hope to
1560
encourage more people to participate in these activities, keeping the
1561
differences between core and non-core contributors to a minimum.
1564
The Development Lifecycle
1565
-------------------------
1567
As a rule, Bazaar development follows a 4 week cycle:
1569
* 2 weeks - general changes
1570
* 1 week - feature freeze
1571
* 1 week+ - Release Candidate stabilization
1573
During the FeatureFreeze week, the trunk (bzr.dev) is open in a limited
1574
way: only low risk changes, critical and high priority fixes are accepted
1575
during this time. At the end of FeatureFreeze, a branch is created for the
1576
first Release Candidate and the trunk is reopened for general development
1577
on the *next* release. A week or so later, the final release is packaged
1578
assuming no serious problems were encountered with the one or more Release
1582
There is a one week overlap between the start of one release and
1583
the end of the previous one.
1586
Communicating and Coordinating
1587
------------------------------
1589
While it has many advantages, one of the challenges of distributed
1590
development is keeping everyone else aware of what you're working on.
1591
There are numerous ways to do this:
1593
#. Assign bugs to yourself in Launchpad
1594
#. Mention it on the mailing list
1595
#. Mention it on IRC
1597
As well as the email notifcations that occur when merge requests are sent
1598
and reviewed, you can keep others informed of where you're spending your
1599
energy by emailing the **bazaar-commits** list implicitly. To do this,
1600
install and configure the Email plugin. One way to do this is add these
1601
configuration settings to your central configuration file (e.g.
1602
``~/.bazaar/bazaar.conf`` on Linux)::
1605
email = Joe Smith <joe.smith@internode.on.net>
1606
smtp_server = mail.internode.on.net:25
1608
Then add these lines for the relevant branches in ``locations.conf``::
1610
post_commit_to = bazaar-commits@lists.canonical.com
1611
post_commit_mailer = smtplib
1613
While attending a sprint, RobertCollins' Dbus plugin is useful for the
1614
same reason. See the documentation within the plugin for information on
1615
how to set it up and configure it.
1621
Setting Up Your Workspace for Reviews
1622
-------------------------------------
1624
TODO: Incorporate John Arbash Meinel's detailed email to Ian C on the
1625
numerous ways of setting up integration branches.
1628
The Review Checklist
1629
--------------------
1631
See `A Closer Look at the Merge & Review Process`_
1632
for information on the gates used to decide whether code can be merged
1633
or not and details on how review results are recorded and communicated.
1636
The Importance of Timely Reviews
1637
--------------------------------
1639
Good reviews do take time. They also regularly require a solid
1640
understanding of the overall code base. In practice, this means a small
1641
number of people often have a large review burden - with knowledge comes
1642
responsibility. No one like their merge requests sitting in a queue going
1643
nowhere, so reviewing sooner rather than later is strongly encouraged.
1652
Of the many workflows supported by Bazaar, the one adopted for Bazaar
1653
development itself is known as "Decentralized with automatic gatekeeper".
1654
To repeat the explanation of this given on
1655
http://bazaar-vcs.org/Workflows:
1658
In this workflow, each developer has their own branch or
1659
branches, plus read-only access to the mainline. A software gatekeeper
1660
(e.g. PQM) has commit rights to the main branch. When a developer wants
1661
their work merged, they request the gatekeeper to merge it. The gatekeeper
1662
does a merge, a compile, and runs the test suite. If the code passes, it
1663
is merged into the mainline.
1665
In a nutshell, here's the overall submission process:
1667
#. get your work ready (including review except for trivial changes)
1668
#. push to a public location
1669
#. ask PQM to merge from that location
1672
At present, PQM always takes the changes to merge from a branch
1673
at a URL that can be read by it. For Bazaar, that means a public,
1674
typically http, URL.
1676
As a result, the following things are needed to use PQM for submissions:
1678
#. A publicly available web server
1679
#. Your OpenPGP key registered with PQM (contact RobertCollins for this)
1680
#. The PQM plugin installed and configured (not strictly required but
1681
highly recommended).
1684
Selecting a Public Branch Location
1685
----------------------------------
1687
If you don't have your own web server running, branches can always be
1688
pushed to Launchpad. Here's the process for doing that:
1690
Depending on your location throughout the world and the size of your
1691
repository though, it is often quicker to use an alternative public
1692
location to Launchpad, particularly if you can set up your own repo and
1693
push into that. By using an existing repo, push only needs to send the
1694
changes, instead of the complete repository every time. Note that it is
1695
easy to register branches in other locations with Launchpad so no benefits
1696
are lost by going this way.
1699
For Canonical staff, http://people.ubuntu.com/~<user>/ is one
1700
suggestion for public http branches. Contact your manager for information
1701
on accessing this system if required.
1703
It should also be noted that best practice in this area is subject to
1704
change as things evolve. For example, once the Bazaar smart server on
1705
Launchpad supports server-side branching, the performance situation will
1706
be very different to what it is now (Jun 2007).
1709
Configuring the PQM Plug-In
1710
---------------------------
1712
While not strictly required, the PQM plugin automates a few things and
1713
reduces the chance of error. Before looking at the plugin, it helps to
1714
understand a little more how PQM operates. Basically, PQM requires an
1715
email indicating what you want it to do. The email typically looks like
1718
star-merge source-branch target-branch
1722
star-merge http://bzr.arbash-meinel.com/branches/bzr/jam-integration http://bazaar-vcs.org/bzr/bzr.dev
1724
Note that the command needs to be on one line. The subject of the email
1725
will be used for the commit message. The email also needs to be ``gpg``
1726
signed with a key that PQM accepts.
1728
The advantages of using the PQM plugin are:
1730
#. You can use the config policies to make it easy to set up public
1731
branches, so you don't have to ever type the full paths you want to merge
1734
#. It checks to make sure the public branch last revision matches the
1735
local last revision so you are submitting what you think you are.
1737
#. It uses the same public_branch and smtp sending settings as bzr-email,
1738
so if you have one set up, you have the other mostly set up.
1740
#. Thunderbird refuses to not wrap lines, and request lines are usually
1741
pretty long (you have 2 long URLs in there).
1743
Here are sample configuration settings for the PQM plugin. Here are the
1744
lines in bazaar.conf::
1747
email = Joe Smith <joe.smith@internode.on.net>
1748
smtp_server=mail.internode.on.net:25
1750
And here are the lines in ``locations.conf`` (or ``branch.conf`` for
1751
dirstate-tags branches)::
1753
[/home/joe/bzr/my-integration]
1754
push_location = sftp://joe-smith@bazaar.launchpad.net/%7Ejoe-smith/bzr/my-integration/
1755
push_location:policy = norecurse
1756
public_branch = http://bazaar.launchpad.net/~joe-smith/bzr/my-integration/
1757
public_branch:policy = appendpath
1758
pqm_email = Bazaar PQM <pqm@bazaar-vcs.org>
1759
pqm_branch = http://bazaar-vcs.org/bzr/bzr.dev
1761
Note that the push settings will be added by the first ``push`` on
1762
a branch. Indeed the preferred way to generate the lines above is to use
1763
``push`` with an argument, then copy-and-paste the other lines into
1770
Here is one possible recipe once the above environment is set up:
1772
#. pull bzr.dev => my-integration
1773
#. merge patch => my-integration
1774
#. fix up any final merge conflicts (NEWS being the big killer here).
1780
The ``push`` step is not required if ``my-integration`` is a checkout of
1783
Because of defaults, you can type a single message into commit and
1784
pqm-commit will reuse that.
1787
Tracking Change Acceptance
1788
--------------------------
1790
The web interface to PQM is https://pqm.bazaar-vcs.org/. After submitting
1791
a change, you can visit this URL to confirm it was received and placed in
1794
When PQM completes processing a change, an email is sent to you with the
1798
Reviewing Blueprints
298
1799
====================
300
If you'd like to propose a change, please post to the
301
bazaar-ng@lists.canonical.com list with a patch, bzr changeset, or link to a
302
branch. Please put '[patch]' in the subject so we can pick them out, and
303
include some text explaining the change. Remember to put an update to the NEWS
304
file in your diff, if it makes any changes visible to users or plugin
305
developers. Please include a diff against mainline if you're giving a link to
308
Please indicate if you think the code is ready to merge, or if it's just a
309
draft or for discussion. If you want comments from many developers rather than
310
to be merged, you can put '[rfc]' in the subject lines.
312
Anyone is welcome to review code. There are broadly three gates for
315
* Doesn't reduce test coverage: if it adds new methods or commands,
316
there should be tests for them. There is a good test framework
317
and plenty of examples to crib from, but if you are having trouble
318
working out how to test something feel free to post a draft patch
321
* Doesn't reduce design clarity, such as by entangling objects
322
we're trying to separate. This is mostly something the more
323
experienced reviewers need to help check.
325
* Improves bugs, features, speed, or code simplicity.
327
Code that goes in should pass all three.
329
If you read a patch please reply and say so. We can use a numeric scale
330
of -1, -0, +0, +1, meaning respectively "really don't want it in current
331
form", "somewhat uncomfortable", "ok with me", and "please put it in".
332
Anyone can "vote". (It's not really voting, just a terse expression.)
334
If something gets say two +1 votes from core reviewers, and no
335
vetos, then it's OK to come in. Any of the core developers can bring it
336
into their integration branch, which I'll merge regularly. (If you do
337
so, please reply and say so.)
1801
Blueprint Tracking Using Launchpad
1802
----------------------------------
1804
New features typically require a fair amount of discussion, design and
1805
debate. For Bazaar, that information is often captured in a so-called
1806
"blueprint" on our Wiki. Overall tracking of blueprints and their status
1807
is done using Launchpad's relevant tracker,
1808
https://blueprints.launchpad.net/bzr/. Once a blueprint for ready for
1809
review, please announce it on the mailing list.
1811
Alternatively, send an email begining with [RFC] with the proposal to the
1812
list. In some cases, you may wish to attach proposed code or a proposed
1813
developer document if that best communicates the idea. Debate can then
1814
proceed using the normal merge review processes.
1817
Recording Blueprint Review Feedback
1818
-----------------------------------
1820
Unlike its Bug Tracker, Launchpad's Blueprint Tracker doesn't currently
1821
(Jun 2007) support a chronological list of comment responses. Review
1822
feedback can either be recorded on the Wiki hosting the blueprints or by
1823
using Launchpad's whiteboard feature.
1832
As the two senior developers, Martin Pool and Robert Collins coordinate
1833
the overall Bazaar product development roadmap. Core developers provide
1834
input and review into this, particularly during sprints. It's totally
1835
expected that community members ought to be working on things that
1836
interest them the most. The roadmap is valuable though because it provides
1837
context for understanding where the product is going as a whole and why.
1840
Using Releases and Milestones in Launchpad
1841
------------------------------------------
1843
TODO ... (Exact policies still under discussion)
1849
Keeping on top of bugs reported is an important part of ongoing release
1850
planning. Everyone in the community is welcome and encouraged to raise
1851
bugs, confirm bugs raised by others, and nominate a priority. Practically
1852
though, a good percentage of bug triage is often done by the core
1853
developers, partially because of their depth of product knowledge.
1855
With respect to bug triage, core developers are encouraged to play an
1856
active role with particular attention to the following tasks:
1858
* keeping the number of unconfirmed bugs low
1859
* ensuring the priorities are generally right (everything as critical - or
1860
medium - is meaningless)
1861
* looking out for regressions and turning those around sooner rather than later.
1864
As well as prioritizing bugs and nominating them against a
1865
target milestone, Launchpad lets core developers offer to mentor others in
1870
vim: ft=rst tw=74 ai
added
removed
doc/developers/HACKING.txt
