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/en/developer-guide/HACKING.txt`` in the source tree, or at
16
http://doc.bazaar-vcs.org/bzr.dev/en/developer-guide/HACKING.html
19
`Bazaar Developer Documentation Catalog <../../developers/index.html>`_.
27
Exploring the Bazaar Platform
28
=============================
30
Before making changes, it's a good idea to explore the work already
31
done by others. Perhaps the new feature or improvement you're looking
32
for is available in another plug-in already? If you find a bug,
33
perhaps someone else has already fixed it?
35
To answer these questions and more, take a moment to explore the
36
overall Bazaar Platform. Here are some links to browse:
38
* The Plugins page on the Wiki - http://bazaar-vcs.org/BzrPlugins
40
* The Bazaar product family on Launchpad - https://launchpad.net/bazaar
42
* Bug Tracker for the core product - https://bugs.launchpad.net/bzr/
44
* Blueprint Tracker for the core product - https://blueprints.launchpad.net/bzr/
46
If nothing else, perhaps you'll find inspiration in how other developers
47
have solved their challenges.
49
Finding Something To Do
50
=======================
52
Ad-hoc performance work can also be done. One useful tool is the 'evil' debug
53
flag. For instance running ``bzr -Devil commit -m "test"`` will log a backtrace
54
to the bzr log file for every method call which triggers a slow or non-scalable
55
part of the bzr library. So checking that a given command with ``-Devil`` has
56
no backtraces logged to the log file is a good way to find problem function
57
calls that might be nested deep in the code base.
59
Planning and Discussing Changes
60
===============================
62
There is a very active community around Bazaar. Mostly we meet on IRC
63
(#bzr on irc.freenode.net) and on the mailing list. To join the Bazaar
64
community, see http://bazaar-vcs.org/BzrSupport.
66
If you are planning to make a change, it's a very good idea to mention it
67
on the IRC channel and/or on the mailing list. There are many advantages
68
to involving the community before you spend much time on a change.
71
* you get to build on the wisdom on others, saving time
73
* if others can direct you to similar code, it minimises the work to be done
75
* it assists everyone in coordinating direction, priorities and effort.
77
In summary, maximising the input from others typically minimises the
78
total effort required to get your changes merged. The community is
79
friendly, helpful and always keen to welcome newcomers.
82
Bazaar Development in a Nutshell
83
================================
85
Looking for a 10 minute introduction to submitting a change?
86
See http://bazaar-vcs.org/BzrGivingBack.
88
TODO: Merge that Wiki page into this document.
91
Understanding the Development Process
92
=====================================
94
The development team follows many practices including:
96
* a public roadmap and planning process in which anyone can participate
98
* time based milestones everyone can work towards and plan around
100
* extensive code review and feedback to contributors
102
* complete and rigorous test coverage on any code contributed
104
* automated validation that all tests still pass before code is merged
105
into the main code branch.
107
The key tools we use to enable these practices are:
109
* Launchpad - https://launchpad.net/
111
* Bazaar - http://bazaar-vcs.org/
113
* Bundle Buggy - http://bundlebuggy.aaronbentley.com/
115
* Patch Queue Manager - https://launchpad.net/pqm/
117
For further information, see http://bazaar-vcs.org/BzrDevelopment.
122
Preparing a Sandbox for Making Changes to Bazaar
123
================================================
125
Bazaar supports many ways of organising your work. See
126
http://bazaar-vcs.org/SharedRepositoryLayouts for a summary of the
127
popular alternatives.
129
Of course, the best choice for you will depend on numerous factors:
130
the number of changes you may be making, the complexity of the changes, etc.
131
As a starting suggestion though:
133
* create a local copy of the main development branch (bzr.dev) by using
136
bzr branch http://bazaar-vcs.org/bzr/bzr.dev/ bzr.dev
138
* keep your copy of bzr.dev prestine (by not developing in it) and keep
139
it up to date (by using bzr pull)
141
* create a new branch off your local bzr.dev copy for each issue
142
(bug or feature) you are working on.
144
This approach makes it easy to go back and make any required changes
145
after a code review. Resubmitting the change is then simple with no
146
risk of accidentially including edits related to other issues you may
147
be working on. After the changes for an issue are accepted and merged,
148
the associated branch can be deleted or archived as you wish.
151
Navigating the Code Base
152
========================
154
.. Was at <http://bazaar-vcs.org/NewDeveloperIntroduction>
156
Some of the key files in this directory are:
159
The command you run to start Bazaar itself. This script is pretty
160
short and just does some checks then jumps into bzrlib.
163
This file covers a brief introduction to Bazaar and lists some of its
167
Summary of changes in each Bazaar release that can affect users or
171
Installs Bazaar system-wide or to your home directory. To perform
172
development work on Bazaar it is not required to run this file - you
173
can simply run the bzr command from the top level directory of your
174
development copy. Note: That if you run setup.py this will create a
175
'build' directory in your development branch. There's nothing wrong
176
with this but don't be confused by it. The build process puts a copy
177
of the main code base into this build directory, along with some other
178
files. You don't need to go in here for anything discussed in this
182
Possibly the most exciting folder of all, bzrlib holds the main code
183
base. This is where you will go to edit python files and contribute to
187
Holds documentation on a whole range of things on Bazaar from the
188
origination of ideas within the project to information on Bazaar
189
features and use cases. Within this directory there is a subdirectory
190
for each translation into a human language. All the documentation
191
is in the ReStructuredText markup language.
194
Documentation specifically targetted at Bazaar and plugin developers.
195
(Including this document.)
199
Automatically-generated API reference information is available at
200
<http://starship.python.net/crew/mwh/bzrlibapi/>.
202
See also the `Bazaar Architectural Overview <../../developers/overview.html>`_.
205
The Code Review Process
206
#######################
208
All code changes coming in to Bazaar are reviewed by someone else.
209
Normally changes by core contributors are reviewed by one other core
210
developer, and changes from other people are reviewed by two core
211
developers. Use intelligent discretion if the patch is trivial.
213
Good reviews do take time. They also regularly require a solid
214
understanding of the overall code base. In practice, this means a small
215
number of people often have a large review burden - with knowledge comes
216
responsibility. No one like their merge requests sitting in a queue going
217
nowhere, so reviewing sooner rather than later is strongly encouraged.
226
Please put a "cover letter" on your merge request explaining:
228
* the reason **why** you're making this change
230
* **how** this change achieves this purpose
232
* anything else you may have fixed in passing
234
* anything significant that you thought of doing, such as a more
235
extensive fix or a different approach, but didn't or couldn't do now
237
A good cover letter makes reviewers' lives easier because they can decide
238
from the letter whether they agree with the purpose and approach, and then
239
assess whether the patch actually does what the cover letter says.
240
Explaining any "drive-by fixes" or roads not taken may also avoid queries
241
from the reviewer. All in all this should give faster and better reviews.
242
Sometimes writing the cover letter helps the submitter realize something
243
else they need to do. The size of the cover letter should be proportional
244
to the size and complexity of the patch.
247
Reviewing proposed changes
248
==========================
250
Anyone is welcome to review code, and reply to the thread with their
253
The simplest way to review a proposed change is to just read the patch on
254
the list or in Bundle Buggy. For more complex changes it may be useful
255
to make a new working tree or branch from trunk, and merge the proposed
256
change into it, so you can experiment with the code or look at a wider
259
There are three main requirements for code to get in:
261
* Doesn't reduce test coverage: if it adds new methods or commands,
262
there should be tests for them. There is a good test framework
263
and plenty of examples to crib from, but if you are having trouble
264
working out how to test something feel free to post a draft patch
267
* Doesn't reduce design clarity, such as by entangling objects
268
we're trying to separate. This is mostly something the more
269
experienced reviewers need to help check.
271
* Improves bugs, features, speed, or code simplicity.
273
Code that goes in should not degrade any of these aspects. Patches are
274
welcome that only cleanup the code without changing the external
275
behaviour. The core developers take care to keep the code quality high
276
and understandable while recognising that perfect is sometimes the enemy
279
It is easy for reviews to make people notice other things which should be
280
fixed but those things should not hold up the original fix being accepted.
281
New things can easily be recorded in the Bug Tracker instead.
283
It's normally much easier to review several smaller patches than one large
284
one. You might want to use ``bzr-loom`` to maintain threads of related
285
work, or submit a preparatory patch that will make your "real" change
289
Checklist for reviewers
290
=======================
292
* Do you understand what the code's doing and why?
294
* Will it perform reasonably for large inputs, both in memory size and
295
run time? Are there some scenarios where performance should be
298
* Is it tested, and are the tests at the right level? Are there both
299
blackbox (command-line level) and API-oriented tests?
301
* If this change will be visible to end users or API users, is it
302
appropriately documented in NEWS?
304
* Does it meet the coding standards below?
306
* If it changes the user-visible behaviour, does it update the help
307
strings and user documentation?
309
* If it adds a new major concept or standard practice, does it update the
310
developer documentation?
312
* (your ideas here...)
318
From May 2009 on, we prefer people to propose code reviews through
321
* <https://launchpad.net/+tour/code-review>
323
* <https://help.launchpad.net/Code/Review>
325
Anyone can propose or comment on a merge propsal just by creating a
328
There are two ways to create a new merge proposal: through the web
329
interface or by email.
332
Proposing a merge through the web
333
---------------------------------
335
To create the propsal through the web: push your branch to Launchpad, eg::
337
bzr push lp:~mbp/bzr/doc
339
then go to the branch's web page, which in this case would be
340
<https://code.launchpad.net/~mbp/bzr/doc>. You can automate that by just
345
You can then click "Propose for merging into another branch", and enter a
346
cover letter into the web form. Typically you'll want to merge into
347
``~bzr/bzr/trunk`` which will be the default; you might also want to
348
nominate merging into a release branch for a bug fix. There is the option
349
to specify a specific reviewer or type of review, and you shouldn't
350
normally change those.
352
Submitting the form takes you to the new page about the merge proposal
353
containing the diff of the changes, comments by interested people, and
354
controls to comment or vote on the change.
356
Proposing a merge by mail
357
-------------------------
359
To propose a merge by mail, send a bundle to ``merge@code.launchpad.net``.
361
You can generate a merge request like this::
363
bzr send -o bug-1234.diff
365
``bzr send`` can also send mail directly if you prefer; see the help.
370
From <https://code.launchpad.net/bzr/+activereviews> you can see all
371
currently active reviews, and choose one to comment on. This page also
372
shows proposals that are now approved and should be merged by someone with
376
Reviews through Bundle Buggy
377
============================
379
The Bundle Buggy tool used up to May 2009 is still available as a review
382
Sending patches for review
383
--------------------------
385
If you'd like to propose a change, please post to the
386
bazaar@lists.canonical.com list with a bundle, patch, or link to a
387
branch. Put ``[PATCH]`` or ``[MERGE]`` in the subject so Bundle Buggy
388
can pick it out, and explain the change in the email message text.
389
Remember to update the NEWS file as part of your change if it makes any
390
changes visible to users or plugin developers. Please include a diff
391
against mainline if you're giving a link to a branch.
393
You can generate a merge request like this::
395
bzr send -o bug-1234.patch
397
A ``.patch`` extension is recommended instead of .bundle as many mail clients
398
will send the latter as a binary file.
400
``bzr send`` can also send mail directly if you prefer; see the help.
402
Please do **NOT** put [PATCH] or [MERGE] in the subject line if you don't
403
want it to be merged. If you want comments from developers rather than
404
to be merged, you can put ``[RFC]`` in the subject line.
406
If this change addresses a bug, please put the bug number in the subject
407
line too, in the form ``[#1]`` so that Bundle Buggy can recognize it.
409
If the change is intended for a particular release mark that in the
410
subject too, e.g. ``[1.6]``.
411
Anyone can "vote" on the mailing list by expressing an opinion. Core
412
developers can also vote using Bundle Buggy. Here are the voting codes and
415
:approve: Reviewer wants this submission merged.
416
:tweak: Reviewer wants this submission merged with small changes. (No
418
:abstain: Reviewer does not intend to vote on this patch.
419
:resubmit: Please make changes and resubmit for review.
420
:reject: Reviewer doesn't want this kind of change merged.
421
:comment: Not really a vote. Reviewer just wants to comment, for now.
423
If a change gets two approvals from core reviewers, and no rejections,
424
then it's OK to come in. Any of the core developers can bring it into the
425
bzr.dev trunk and backport it to maintenance branches if required. The
426
Release Manager will merge the change into the branch for a pending
427
release, if any. As a guideline, core developers usually merge their own
428
changes and volunteer to merge other contributions if they were the second
429
reviewer to agree to a change.
431
To track the progress of proposed changes, use Bundle Buggy. See
432
http://bundlebuggy.aaronbentley.com/help for a link to all the
433
outstanding merge requests together with an explanation of the columns.
434
Bundle Buggy will also mail you a link to track just your change.
436
Coding Style Guidelines
437
#######################
442
``hasattr`` should not be used because it swallows exceptions including
443
``KeyboardInterrupt``. Instead, say something like ::
445
if getattr(thing, 'name', None) is None
451
Please write PEP-8__ compliant code.
453
__ http://www.python.org/peps/pep-0008.html
455
One often-missed requirement is that the first line of docstrings
456
should be a self-contained one-sentence summary.
458
We use 4 space indents for blocks, and never use tab characters. (In vim,
461
Trailing white space should be avoided, but is allowed.
462
You should however not make lots of unrelated white space changes.
464
Unix style newlines (LF) are used.
466
Each file must have a newline at the end of it.
468
Lines should be no more than 79 characters if at all possible.
469
Lines that continue a long statement may be indented in either of
472
within the parenthesis or other character that opens the block, e.g.::
478
or indented by four spaces::
484
The first is considered clearer by some people; however it can be a bit
485
harder to maintain (e.g. when the method name changes), and it does not
486
work well if the relevant parenthesis is already far to the right. Avoid
489
self.legbone.kneebone.shinbone.toebone.shake_it(one,
495
self.legbone.kneebone.shinbone.toebone.shake_it(one,
501
self.legbone.kneebone.shinbone.toebone.shake_it(
504
For long lists, we like to add a trailing comma and put the closing
505
character on the following line. This makes it easier to add new items in
508
from bzrlib.goo import (
514
There should be spaces between function paramaters, but not between the
515
keyword name and the value::
517
call(1, 3, cheese=quark)
521
;(defface my-invalid-face
522
; '((t (:background "Red" :underline t)))
523
; "Face used to highlight invalid constructs or other uglyties"
526
(defun my-python-mode-hook ()
527
;; setup preferred indentation style.
528
(setq fill-column 79)
529
(setq indent-tabs-mode nil) ; no tabs, never, I will not repeat
530
; (font-lock-add-keywords 'python-mode
531
; '(("^\\s *\t" . 'my-invalid-face) ; Leading tabs
532
; ("[ \t]+$" . 'my-invalid-face) ; Trailing spaces
533
; ("^[ \t]+$" . 'my-invalid-face)); Spaces only
537
(add-hook 'python-mode-hook 'my-python-mode-hook)
539
The lines beginning with ';' are comments. They can be activated
540
if one want to have a strong notice of some tab/space usage
547
* Imports should be done at the top-level of the file, unless there is
548
a strong reason to have them lazily loaded when a particular
549
function runs. Import statements have a cost, so try to make sure
550
they don't run inside hot functions.
552
* Module names should always be given fully-qualified,
553
i.e. ``bzrlib.hashcache`` not just ``hashcache``.
559
Functions, methods or members that are "private" to bzrlib are given
560
a leading underscore prefix. Names without a leading underscore are
561
public not just across modules but to programmers using bzrlib as an
562
API. As a consequence, a leading underscore is appropriate for names
563
exposed across modules but that are not to be exposed to bzrlib API
566
We prefer class names to be concatenated capital words (``TestCase``)
567
and variables, methods and functions to be lowercase words joined by
568
underscores (``revision_id``, ``get_revision``).
570
For the purposes of naming some names are treated as single compound
571
words: "filename", "revno".
573
Consider naming classes as nouns and functions/methods as verbs.
575
Try to avoid using abbreviations in names, because there can be
576
inconsistency if other people use the full name.
582
``revision_id`` not ``rev_id`` or ``revid``
584
Functions that transform one thing to another should be named ``x_to_y``
585
(not ``x2y`` as occurs in some old code.)
591
Python destructors (``__del__``) work differently to those of other
592
languages. In particular, bear in mind that destructors may be called
593
immediately when the object apparently becomes unreferenced, or at some
594
later time, or possibly never at all. Therefore we have restrictions on
595
what can be done inside them.
597
0. If you think you need to use a ``__del__`` method ask another
598
developer for alternatives. If you do need to use one, explain
601
1. Never rely on a ``__del__`` method running. If there is code that
602
must run, do it from a ``finally`` block instead.
604
2. Never ``import`` from inside a ``__del__`` method, or you may crash the
607
3. In some places we raise a warning from the destructor if the object
608
has not been cleaned up or closed. This is considered OK: the warning
609
may not catch every case but it's still useful sometimes.
615
In some places we have variables which point to callables that construct
616
new instances. That is to say, they can be used a lot like class objects,
617
but they shouldn't be *named* like classes:
619
> I think that things named FooBar should create instances of FooBar when
620
> called. Its plain confusing for them to do otherwise. When we have
621
> something that is going to be used as a class - that is, checked for via
622
> isinstance or other such idioms, them I would call it foo_class, so that
623
> it is clear that a callable is not sufficient. If it is only used as a
624
> factory, then yes, foo_factory is what I would use.
630
Several places in Bazaar use (or will use) a registry, which is a
631
mapping from names to objects or classes. The registry allows for
632
loading in registered code only when it's needed, and keeping
633
associated information such as a help string or description.
636
InterObject and multiple dispatch
637
=================================
639
The ``InterObject`` provides for two-way `multiple dispatch`__: matching
640
up for example a source and destination repository to find the right way
641
to transfer data between them.
643
.. __: http://en.wikipedia.org/wiki/Multiple_dispatch
645
There is a subclass ``InterObject`` classes for each type of object that is
646
dispatched this way, e.g. ``InterRepository``. Calling ``.get()`` on this
647
class will return an ``InterObject`` instance providing the best match for
648
those parameters, and this instance then has methods for operations
651
inter = InterRepository.get(source_repo, target_repo)
652
inter.fetch(revision_id)
654
``InterRepository`` also acts as a registry-like object for its
655
subclasses, and they can be added through ``.register_optimizer``. The
656
right one to run is selected by asking each class, in reverse order of
657
registration, whether it ``.is_compatible`` with the relevant objects.
662
To make startup time faster, we use the ``bzrlib.lazy_import`` module to
663
delay importing modules until they are actually used. ``lazy_import`` uses
664
the same syntax as regular python imports. So to import a few modules in a
667
from bzrlib.lazy_import import lazy_import
668
lazy_import(globals(), """
677
revision as _mod_revision,
679
import bzrlib.transport
683
At this point, all of these exist as a ``ImportReplacer`` object, ready to
684
be imported once a member is accessed. Also, when importing a module into
685
the local namespace, which is likely to clash with variable names, it is
686
recommended to prefix it as ``_mod_<module>``. This makes it clearer that
687
the variable is a module, and these object should be hidden anyway, since
688
they shouldn't be imported into other namespaces.
690
While it is possible for ``lazy_import()`` to import members of a module
691
when using the ``from module import member`` syntax, it is recommended to
692
only use that syntax to load sub modules ``from module import submodule``.
693
This is because variables and classes can frequently be used without
694
needing a sub-member for example::
696
lazy_import(globals(), """
697
from module import MyClass
701
return isinstance(x, MyClass)
703
This will incorrectly fail, because ``MyClass`` is a ``ImportReplacer``
704
object, rather than the real class.
706
It also is incorrect to assign ``ImportReplacer`` objects to other variables.
707
Because the replacer only knows about the original name, it is unable to
708
replace other variables. The ``ImportReplacer`` class will raise an
709
``IllegalUseOfScopeReplacer`` exception if it can figure out that this
710
happened. But it requires accessing a member more than once from the new
711
variable, so some bugs are not detected right away.
717
The null revision is the ancestor of all revisions. Its revno is 0, its
718
revision-id is ``null:``, and its tree is the empty tree. When referring
719
to the null revision, please use ``bzrlib.revision.NULL_REVISION``. Old
720
code sometimes uses ``None`` for the null revision, but this practice is
724
Object string representations
725
=============================
727
Python prints objects using their ``__repr__`` method when they are
728
written to logs, exception tracebacks, or the debugger. We want
729
objects to have useful representations to help in determining what went
732
If you add a new class you should generally add a ``__repr__`` method
733
unless there is an adequate method in a parent class. There should be a
736
Representations should typically look like Python constructor syntax, but
737
they don't need to include every value in the object and they don't need
738
to be able to actually execute. They're to be read by humans, not
739
machines. Don't hardcode the classname in the format, so that we get the
740
correct value if the method is inherited by a subclass. If you're
741
printing attributes of the object, including strings, you should normally
742
use ``%r`` syntax (to call their repr in turn).
744
Try to avoid the representation becoming more than one or two lines long.
745
(But balance this against including useful information, and simplicity of
748
Because repr methods are often called when something has already gone
749
wrong, they should be written somewhat more defensively than most code.
750
The object may be half-initialized or in some other way in an illegal
751
state. The repr method shouldn't raise an exception, or it may hide the
752
(probably more useful) underlying exception.
757
return '%s(%r)' % (self.__class__.__name__,
764
A bare ``except`` statement will catch all exceptions, including ones that
765
really should terminate the program such as ``MemoryError`` and
766
``KeyboardInterrupt``. They should rarely be used unless the exception is
767
later re-raised. Even then, think about whether catching just
768
``Exception`` (which excludes system errors in Python2.5 and later) would
775
All code should be exercised by the test suite. See `Guide to Testing
776
Bazaar <../../developers/testing.html>`_ for detailed information about writing tests.
785
We have a commitment to 6 months API stability - any supported symbol in a
786
release of bzr MUST NOT be altered in any way that would result in
787
breaking existing code that uses it. That means that method names,
788
parameter ordering, parameter names, variable and attribute names etc must
789
not be changed without leaving a 'deprecated forwarder' behind. This even
790
applies to modules and classes.
792
If you wish to change the behaviour of a supported API in an incompatible
793
way, you need to change its name as well. For instance, if I add an optional keyword
794
parameter to branch.commit - that's fine. On the other hand, if I add a
795
keyword parameter to branch.commit which is a *required* transaction
796
object, I should rename the API - i.e. to 'branch.commit_transaction'.
798
When renaming such supported API's, be sure to leave a deprecated_method (or
799
_function or ...) behind which forwards to the new API. See the
800
bzrlib.symbol_versioning module for decorators that take care of the
801
details for you - such as updating the docstring, and issuing a warning
802
when the old api is used.
804
For unsupported API's, it does not hurt to follow this discipline, but it's
805
not required. Minimally though, please try to rename things so that
806
callers will at least get an AttributeError rather than weird results.
809
Deprecation decorators
810
----------------------
812
``bzrlib.symbol_versioning`` provides decorators that can be attached to
813
methods, functions, and other interfaces to indicate that they should no
814
longer be used. For example::
816
@deprecated_method(deprecated_in((0, 1, 4)))
818
return self._new_foo()
820
To deprecate a static method you must call ``deprecated_function``
821
(**not** method), after the staticmethod call::
824
@deprecated_function(deprecated_in((0, 1, 4)))
825
def create_repository(base, shared=False, format=None):
827
When you deprecate an API, you should not just delete its tests, because
828
then we might introduce bugs in them. If the API is still present at all,
829
it should still work. The basic approach is to use
830
``TestCase.applyDeprecated`` which in one step checks that the API gives
831
the expected deprecation message, and also returns the real result from
832
the method, so that tests can keep running.
834
Deprecation warnings will be suppressed for final releases, but not for
835
development versions or release candidates, or when running ``bzr
836
selftest``. This gives developers information about whether their code is
837
using deprecated functions, but avoids confusing users about things they
844
Processing Command Lines
845
------------------------
847
bzrlib has a standard framework for parsing command lines and calling
848
processing routines associated with various commands. See builtins.py
849
for numerous examples.
852
Standard Parameter Types
853
------------------------
855
There are some common requirements in the library: some parameters need to be
856
unicode safe, some need byte strings, and so on. At the moment we have
857
only codified one specific pattern: Parameters that need to be unicode
858
should be checked via ``bzrlib.osutils.safe_unicode``. This will coerce the
859
input into unicode in a consistent fashion, allowing trivial strings to be
860
used for programmer convenience, but not performing unpredictably in the
861
presence of different locales.
867
(The strategy described here is what we want to get to, but it's not
868
consistently followed in the code at the moment.)
870
bzrlib is intended to be a generically reusable library. It shouldn't
871
write messages to stdout or stderr, because some programs that use it
872
might want to display that information through a GUI or some other
875
We can distinguish two types of output from the library:
877
1. Structured data representing the progress or result of an
878
operation. For example, for a commit command this will be a list
879
of the modified files and the finally committed revision number
882
These should be exposed either through the return code or by calls
883
to a callback parameter.
885
A special case of this is progress indicators for long-lived
886
operations, where the caller should pass a ProgressBar object.
888
2. Unstructured log/debug messages, mostly for the benefit of the
889
developers or users trying to debug problems. This should always
890
be sent through ``bzrlib.trace`` and Python ``logging``, so that
891
it can be redirected by the client.
893
The distinction between the two is a bit subjective, but in general if
894
there is any chance that a library would want to see something as
895
structured data, we should make it so.
897
The policy about how output is presented in the text-mode client
898
should be only in the command-line tool.
901
Progress and Activity Indications
902
---------------------------------
904
bzrlib has a way for code to display to the user that stuff is happening
905
during a long operation. There are two particular types: *activity* which
906
means that IO is happening on a Transport, and *progress* which means that
907
higher-level application work is occurring. Both are drawn together by
910
Transport objects are responsible for calling `report_transport_activity`
913
Progress uses a model/view pattern: application code acts on a
914
`ProgressTask` object, which notifies the UI when it needs to be
915
displayed. Progress tasks form a stack. To create a new progress task on
916
top of the stack, call `bzrlib.ui.ui_factory.nested_progress_bar()`, then
917
call `update()` on the returned ProgressTask. It can be updated with just
918
a text description, with a numeric count, or with a numeric count and
919
expected total count. If an expected total count is provided the view
920
can show the progress moving along towards the expected total.
922
The user should call `finish` on the `ProgressTask` when the logical
923
operation has finished, so it can be removed from the stack.
925
Progress tasks have a complex relatioship with generators: it's a very
926
good place to use them, but because python2.4 does not allow ``finally``
927
blocks in generators it's hard to clean them up properly. In this case
928
it's probably better to have the code calling the generator allocate a
929
progress task for its use and then call `finalize` when it's done, which
930
will close it if it was not already closed. The generator should also
931
finish the progress task when it exits, because it may otherwise be a long
932
time until the finally block runs.
938
Bazaar has online help for various topics through ``bzr help COMMAND`` or
939
equivalently ``bzr command -h``. We also have help on command options,
940
and on other help topics. (See ``help_topics.py``.)
942
As for python docstrings, the first paragraph should be a single-sentence
943
synopsis of the command.
945
The help for options should be one or more proper sentences, starting with
946
a capital letter and finishing with a full stop (period).
948
All help messages and documentation should have two spaces between
952
Handling Errors and Exceptions
953
==============================
955
Commands should return non-zero when they encounter circumstances that
956
the user should really pay attention to - which includes trivial shell
959
Recommended values are:
962
1. Conflicts in merge-like operations, or changes are present in
963
diff-like operations.
964
2. Unrepresentable diff changes (i.e. binary files that we cannot show
966
3. An error or exception has occurred.
967
4. An internal error occurred (one that shows a traceback.)
969
Errors are handled through Python exceptions. Exceptions should be defined
970
inside bzrlib.errors, so that we can see the whole tree at a glance.
972
We broadly classify errors as either being either internal or not,
973
depending on whether ``internal_error`` is set or not. If we think it's our
974
fault, we show a backtrace, an invitation to report the bug, and possibly
975
other details. This is the default for errors that aren't specifically
976
recognized as being caused by a user error. Otherwise we show a briefer
977
message, unless -Derror was given.
979
Many errors originate as "environmental errors" which are raised by Python
980
or builtin libraries -- for example IOError. These are treated as being
981
our fault, unless they're caught in a particular tight scope where we know
982
that they indicate a user errors. For example if the repository format
983
is not found, the user probably gave the wrong path or URL. But if one of
984
the files inside the repository is not found, then it's our fault --
985
either there's a bug in bzr, or something complicated has gone wrong in
986
the environment that means one internal file was deleted.
988
Many errors are defined in ``bzrlib/errors.py`` but it's OK for new errors
989
to be added near the place where they are used.
991
Exceptions are formatted for the user by conversion to a string
992
(eventually calling their ``__str__`` method.) As a convenience the
993
``._fmt`` member can be used as a template which will be mapped to the
994
error's instance dict.
996
New exception classes should be defined when callers might want to catch
997
that exception specifically, or when it needs a substantially different
1000
#. If it is something that a caller can recover from, a custom exception
1003
#. If it is a data consistency issue, using a builtin like
1004
``ValueError``/``TypeError`` is reasonable.
1006
#. If it is a programmer error (using an api incorrectly)
1007
``AssertionError`` is reasonable.
1009
#. Otherwise, use ``BzrError`` or ``InternalBzrError``.
1011
Exception strings should start with a capital letter and should not have a
1012
final fullstop. If long, they may contain newlines to break the text.
1018
Do not use the Python ``assert`` statement, either in tests or elsewhere.
1019
A source test checks that it is not used. It is ok to explicitly raise
1024
* It makes the behaviour vary depending on whether bzr is run with -O
1025
or not, therefore giving a chance for bugs that occur in one case or
1026
the other, several of which have already occurred: assertions with
1027
side effects, code which can't continue unless the assertion passes,
1028
cases where we should give the user a proper message rather than an
1030
* It's not that much shorter than an explicit if/raise.
1031
* It tends to lead to fuzzy thinking about whether the check is
1032
actually needed or not, and whether it's an internal error or not
1033
* It tends to cause look-before-you-leap patterns.
1034
* It's unsafe if the check is needed to protect the integrity of the
1036
* It tends to give poor messages since the developer can get by with
1037
no explanatory text at all.
1038
* We can't rely on people always running with -O in normal use, so we
1039
can't use it for tests that are actually expensive.
1040
* Expensive checks that help developers are better turned on from the
1041
test suite or a -D flag.
1042
* If used instead of ``self.assert*()`` in tests it makes them falsely pass with -O.
1048
When you change bzrlib, please update the relevant documentation for the
1049
change you made: Changes to commands should update their help, and
1050
possibly end user tutorials; changes to the core library should be
1051
reflected in API documentation.
1056
If you make a user-visible change, please add a note to the NEWS file.
1057
The description should be written to make sense to someone who's just
1058
a user of bzr, not a developer: new functions or classes shouldn't be
1059
mentioned, but new commands, changes in behaviour or fixed nontrivial
1060
bugs should be listed. See the existing entries for an idea of what
1063
Within each release, entries in the news file should have the most
1064
user-visible changes first. So the order should be approximately:
1066
* changes to existing behaviour - the highest priority because the
1067
user's existing knowledge is incorrect
1068
* new features - should be brought to their attention
1069
* bug fixes - may be of interest if the bug was affecting them, and
1070
should include the bug number if any
1071
* major documentation changes
1072
* changes to internal interfaces
1074
People who made significant contributions to each change are listed in
1075
parenthesis. This can include reporting bugs (particularly with good
1076
details or reproduction recipes), submitting patches, etc.
1081
The docstring of a command is used by ``bzr help`` to generate help output
1082
for the command. The list 'takes_options' attribute on a command is used by
1083
``bzr help`` to document the options for the command - the command
1084
docstring does not need to document them. Finally, the '_see_also'
1085
attribute on a command can be used to reference other related help topics.
1090
Functions, methods, classes and modules should have docstrings
1091
describing how they are used.
1093
The first line of the docstring should be a self-contained sentence.
1095
For the special case of Command classes, this acts as the user-visible
1096
documentation shown by the help command.
1098
The docstrings should be formatted as reStructuredText_ (like this
1099
document), suitable for processing using the epydoc_ tool into HTML
1102
.. _reStructuredText: http://docutils.sourceforge.net/rst.html
1103
.. _epydoc: http://epydoc.sourceforge.net/
1112
The copyright policy for bzr was recently made clear in this email (edited
1113
for grammatical correctness)::
1115
The attached patch cleans up the copyright and license statements in
1116
the bzr source. It also adds tests to help us remember to add them
1117
with the correct text.
1119
We had the problem that lots of our files were "Copyright Canonical
1120
Development Ltd" which is not a real company, and some other variations
1121
on this theme. Also, some files were missing the GPL statements.
1123
I want to be clear about the intent of this patch, since copyright can
1124
be a little controversial.
1126
1) The big motivation for this is not to shut out the community, but
1127
just to clean up all of the invalid copyright statements.
1129
2) It has been the general policy for bzr that we want a single
1130
copyright holder for all of the core code. This is following the model
1131
set by the FSF, which makes it easier to update the code to a new
1132
license in case problems are encountered. (For example, if we want to
1133
upgrade the project universally to GPL v3 it is much simpler if there is
1134
a single copyright holder). It also makes it clearer if copyright is
1135
ever debated, there is a single holder, which makes it easier to defend
1136
in court, etc. (I think the FSF position is that if you assign them
1137
copyright, they can defend it in court rather than you needing to, and
1138
I'm sure Canonical would do the same).
1139
As such, Canonical has requested copyright assignments from all of the
1142
3) If someone wants to add code and not attribute it to Canonical, there
1143
is a specific list of files that are excluded from this check. And the
1144
test failure indicates where that is, and how to update it.
1146
4) If anyone feels that I changed a copyright statement incorrectly, just
1147
let me know, and I'll be happy to correct it. Whenever you have large
1148
mechanical changes like this, it is possible to make some mistakes.
1150
Just to reiterate, this is a community project, and it is meant to stay
1151
that way. Core bzr code is copyright Canonical for legal reasons, and
1152
the tests are just there to help us maintain that.
1155
Miscellaneous Topics
1156
####################
1161
Bazaar has a few facilities to help debug problems by going into pdb_, the
1164
.. _pdb: http://docs.python.org/lib/debugger-commands.html
1166
If the ``BZR_PDB`` environment variable is set
1167
then bzr will go into pdb post-mortem mode when an unhandled exception
1170
If you send a SIGQUIT signal to bzr, which can be done by pressing
1171
Ctrl-\\ on Unix, bzr will go into the debugger immediately. You can
1172
continue execution by typing ``c``. This can be disabled if necessary
1173
by setting the environment variable ``BZR_SIGQUIT_PDB=0``.
1179
Bazaar accepts some global options starting with ``-D`` such as
1180
``-Dhpss``. These set a value in `bzrlib.debug.debug_flags`, and
1181
typically cause more information to be written to the trace file. Most
1182
`mutter` calls should be guarded by a check of those flags so that we
1183
don't write out too much information if it's not needed.
1185
Debug flags may have effects other than just emitting trace messages.
1187
Run ``bzr help global-options`` to see them all.
1189
These flags may also be set as a comma-separated list in the
1190
``debug_flags`` option in e.g. ``~/.bazaar/bazaar.conf``. (Note that it
1191
must be in this global file, not in the branch or location configuration,
1192
because it's currently only loaded at startup time.) For instance you may
1193
want to always record hpss traces and to see full error tracebacks::
1195
debug_flags = hpss, error
1202
Integer identifier for a revision on the main line of a branch.
1203
Revision 0 is always the null revision; others are 1-based
1204
indexes into the branch's revision history.
1207
Unicode and Encoding Support
1208
============================
1210
This section discusses various techniques that Bazaar uses to handle
1211
characters that are outside the ASCII set.
1216
When a ``Command`` object is created, it is given a member variable
1217
accessible by ``self.outf``. This is a file-like object, which is bound to
1218
``sys.stdout``, and should be used to write information to the screen,
1219
rather than directly writing to ``sys.stdout`` or calling ``print``.
1220
This file has the ability to translate Unicode objects into the correct
1221
representation, based on the console encoding. Also, the class attribute
1222
``encoding_type`` will effect how unprintable characters will be
1223
handled. This parameter can take one of 3 values:
1226
Unprintable characters will be represented with a suitable replacement
1227
marker (typically '?'), and no exception will be raised. This is for
1228
any command which generates text for the user to review, rather than
1229
for automated processing.
1230
For example: ``bzr log`` should not fail if one of the entries has text
1231
that cannot be displayed.
1234
Attempting to print an unprintable character will cause a UnicodeError.
1235
This is for commands that are intended more as scripting support, rather
1236
than plain user review.
1237
For exampl: ``bzr ls`` is designed to be used with shell scripting. One
1238
use would be ``bzr ls --null --unknows | xargs -0 rm``. If ``bzr``
1239
printed a filename with a '?', the wrong file could be deleted. (At the
1240
very least, the correct file would not be deleted). An error is used to
1241
indicate that the requested action could not be performed.
1244
Do not attempt to automatically convert Unicode strings. This is used
1245
for commands that must handle conversion themselves.
1246
For example: ``bzr diff`` needs to translate Unicode paths, but should
1247
not change the exact text of the contents of the files.
1250
``bzrlib.urlutils.unescape_for_display``
1251
----------------------------------------
1253
Because Transports work in URLs (as defined earlier), printing the raw URL
1254
to the user is usually less than optimal. Characters outside the standard
1255
set are printed as escapes, rather than the real character, and local
1256
paths would be printed as ``file://`` urls. The function
1257
``unescape_for_display`` attempts to unescape a URL, such that anything
1258
that cannot be printed in the current encoding stays an escaped URL, but
1259
valid characters are generated where possible.
1265
The ``bzrlib.osutils`` module has many useful helper functions, including
1266
some more portable variants of functions in the standard library.
1268
In particular, don't use ``shutil.rmtree`` unless it's acceptable for it
1269
to fail on Windows if some files are readonly or still open elsewhere.
1270
Use ``bzrlib.osutils.rmtree`` instead.
1276
We write some extensions in C using pyrex. We design these to work in
1279
* User with no C compiler
1280
* User with C compiler
1283
The recommended way to install bzr is to have a C compiler so that the
1284
extensions can be built, but if no C compiler is present, the pure python
1285
versions we supply will work, though more slowly.
1287
For developers we recommend that pyrex be installed, so that the C
1288
extensions can be changed if needed.
1290
For the C extensions, the extension module should always match the
1291
original python one in all respects (modulo speed). This should be
1292
maintained over time.
1294
To create an extension, add rules to setup.py for building it with pyrex,
1295
and with distutils. Now start with an empty .pyx file. At the top add
1296
"include 'yourmodule.py'". This will import the contents of foo.py into this
1297
file at build time - remember that only one module will be loaded at
1298
runtime. Now you can subclass classes, or replace functions, and only your
1299
changes need to be present in the .pyx file.
1301
Note that pyrex does not support all 2.4 programming idioms, so some
1302
syntax changes may be required. I.e.
1304
- 'from foo import (bar, gam)' needs to change to not use the brackets.
1305
- 'import foo.bar as bar' needs to be 'import foo.bar; bar = foo.bar'
1307
If the changes are too dramatic, consider
1308
maintaining the python code twice - once in the .pyx, and once in the .py,
1309
and no longer including the .py file.
1312
Making Installers for OS Windows
1313
================================
1314
To build a win32 installer, see the instructions on the wiki page:
1315
http://bazaar-vcs.org/BzrWin32Installer
1318
Core Developer Tasks
1319
####################
1324
What is a Core Developer?
1325
-------------------------
1327
While everyone in the Bazaar community is welcome and encouraged to
1328
propose and submit changes, a smaller team is reponsible for pulling those
1329
changes together into a cohesive whole. In addition to the general developer
1330
stuff covered above, "core" developers have responsibility for:
1333
* reviewing blueprints
1335
* managing releases (see the `Releasing Bazaar <../../developers/releasing.html>`_)
1338
Removing barriers to community participation is a key reason for adopting
1339
distributed VCS technology. While DVCS removes many technical barriers,
1340
a small number of social barriers are often necessary instead.
1341
By documenting how the above things are done, we hope to
1342
encourage more people to participate in these activities, keeping the
1343
differences between core and non-core contributors to a minimum.
1346
Communicating and Coordinating
1347
------------------------------
1349
While it has many advantages, one of the challenges of distributed
1350
development is keeping everyone else aware of what you're working on.
1351
There are numerous ways to do this:
1353
#. Assign bugs to yourself in Launchpad
1354
#. Mention it on the mailing list
1355
#. Mention it on IRC
1357
As well as the email notifcations that occur when merge requests are sent
1358
and reviewed, you can keep others informed of where you're spending your
1359
energy by emailing the **bazaar-commits** list implicitly. To do this,
1360
install and configure the Email plugin. One way to do this is add these
1361
configuration settings to your central configuration file (e.g.
1362
``~/.bazaar/bazaar.conf`` on Linux)::
1365
email = Joe Smith <joe.smith@internode.on.net>
1366
smtp_server = mail.internode.on.net:25
1368
Then add these lines for the relevant branches in ``locations.conf``::
1370
post_commit_to = bazaar-commits@lists.canonical.com
1371
post_commit_mailer = smtplib
1373
While attending a sprint, RobertCollins' Dbus plugin is useful for the
1374
same reason. See the documentation within the plugin for information on
1375
how to set it up and configure it.
1384
Of the many workflows supported by Bazaar, the one adopted for Bazaar
1385
development itself is known as "Decentralized with automatic gatekeeper".
1386
To repeat the explanation of this given on
1387
http://bazaar-vcs.org/Workflows:
1390
In this workflow, each developer has their own branch or
1391
branches, plus read-only access to the mainline. A software gatekeeper
1392
(e.g. PQM) has commit rights to the main branch. When a developer wants
1393
their work merged, they request the gatekeeper to merge it. The gatekeeper
1394
does a merge, a compile, and runs the test suite. If the code passes, it
1395
is merged into the mainline.
1397
In a nutshell, here's the overall submission process:
1399
#. get your work ready (including review except for trivial changes)
1400
#. push to a public location
1401
#. ask PQM to merge from that location
1404
At present, PQM always takes the changes to merge from a branch
1405
at a URL that can be read by it. For Bazaar, that means a public,
1406
typically http, URL.
1408
As a result, the following things are needed to use PQM for submissions:
1410
#. A publicly available web server
1411
#. Your OpenPGP key registered with PQM (contact RobertCollins for this)
1412
#. The PQM plugin installed and configured (not strictly required but
1413
highly recommended).
1416
Selecting a Public Branch Location
1417
----------------------------------
1419
If you don't have your own web server running, branches can always be
1420
pushed to Launchpad. Here's the process for doing that:
1422
Depending on your location throughout the world and the size of your
1423
repository though, it is often quicker to use an alternative public
1424
location to Launchpad, particularly if you can set up your own repo and
1425
push into that. By using an existing repo, push only needs to send the
1426
changes, instead of the complete repository every time. Note that it is
1427
easy to register branches in other locations with Launchpad so no benefits
1428
are lost by going this way.
1431
For Canonical staff, http://people.ubuntu.com/~<user>/ is one
1432
suggestion for public http branches. Contact your manager for information
1433
on accessing this system if required.
1435
It should also be noted that best practice in this area is subject to
1436
change as things evolve. For example, once the Bazaar smart server on
1437
Launchpad supports server-side branching, the performance situation will
1438
be very different to what it is now (Jun 2007).
1441
Configuring the PQM Plug-In
1442
---------------------------
1444
While not strictly required, the PQM plugin automates a few things and
1445
reduces the chance of error. Before looking at the plugin, it helps to
1446
understand a little more how PQM operates. Basically, PQM requires an
1447
email indicating what you want it to do. The email typically looks like
1450
star-merge source-branch target-branch
1454
star-merge http://bzr.arbash-meinel.com/branches/bzr/jam-integration http://bazaar-vcs.org/bzr/bzr.dev
1456
Note that the command needs to be on one line. The subject of the email
1457
will be used for the commit message. The email also needs to be ``gpg``
1458
signed with a key that PQM accepts.
1460
The advantages of using the PQM plugin are:
1462
#. You can use the config policies to make it easy to set up public
1463
branches, so you don't have to ever type the full paths you want to merge
1466
#. It checks to make sure the public branch last revision matches the
1467
local last revision so you are submitting what you think you are.
1469
#. It uses the same public_branch and smtp sending settings as bzr-email,
1470
so if you have one set up, you have the other mostly set up.
1472
#. Thunderbird refuses to not wrap lines, and request lines are usually
1473
pretty long (you have 2 long URLs in there).
1475
Here are sample configuration settings for the PQM plugin. Here are the
1476
lines in bazaar.conf::
1479
email = Joe Smith <joe.smith@internode.on.net>
1480
smtp_server=mail.internode.on.net:25
1482
And here are the lines in ``locations.conf`` (or ``branch.conf`` for
1483
dirstate-tags branches)::
1485
[/home/joe/bzr/my-integration]
1486
push_location = sftp://joe-smith@bazaar.launchpad.net/%7Ejoe-smith/bzr/my-integration/
1487
push_location:policy = norecurse
1488
public_branch = http://bazaar.launchpad.net/~joe-smith/bzr/my-integration/
1489
public_branch:policy = appendpath
1490
pqm_email = Bazaar PQM <pqm@bazaar-vcs.org>
1491
pqm_branch = http://bazaar-vcs.org/bzr/bzr.dev
1493
Note that the push settings will be added by the first ``push`` on
1494
a branch. Indeed the preferred way to generate the lines above is to use
1495
``push`` with an argument, then copy-and-paste the other lines into
1502
Here is one possible recipe once the above environment is set up:
1504
#. pull bzr.dev => my-integration
1505
#. merge patch => my-integration
1506
#. fix up any final merge conflicts (NEWS being the big killer here).
1512
The ``push`` step is not required if ``my-integration`` is a checkout of
1515
Because of defaults, you can type a single message into commit and
1516
pqm-commit will reuse that.
1519
Tracking Change Acceptance
1520
--------------------------
1522
The web interface to PQM is https://pqm.bazaar-vcs.org/. After submitting
1523
a change, you can visit this URL to confirm it was received and placed in
1526
When PQM completes processing a change, an email is sent to you with the
1530
Reviewing Blueprints
1531
====================
1533
Blueprint Tracking Using Launchpad
1534
----------------------------------
1536
New features typically require a fair amount of discussion, design and
1537
debate. For Bazaar, that information is often captured in a so-called
1538
"blueprint" on our Wiki. Overall tracking of blueprints and their status
1539
is done using Launchpad's relevant tracker,
1540
https://blueprints.launchpad.net/bzr/. Once a blueprint for ready for
1541
review, please announce it on the mailing list.
1543
Alternatively, send an email begining with [RFC] with the proposal to the
1544
list. In some cases, you may wish to attach proposed code or a proposed
1545
developer document if that best communicates the idea. Debate can then
1546
proceed using the normal merge review processes.
1549
Recording Blueprint Review Feedback
1550
-----------------------------------
1552
Unlike its Bug Tracker, Launchpad's Blueprint Tracker doesn't currently
1553
(Jun 2007) support a chronological list of comment responses. Review
1554
feedback can either be recorded on the Wiki hosting the blueprints or by
1555
using Launchpad's whiteboard feature.
1562
Using Releases and Milestones in Launchpad
1563
------------------------------------------
1565
TODO ... (Exact policies still under discussion)
1571
Keeping on top of bugs reported is an important part of ongoing release
1572
planning. Everyone in the community is welcome and encouraged to raise
1573
bugs, confirm bugs raised by others, and nominate a priority. Practically
1574
though, a good percentage of bug triage is often done by the core
1575
developers, partially because of their depth of product knowledge.
1577
With respect to bug triage, core developers are encouraged to play an
1578
active role with particular attention to the following tasks:
1580
* keeping the number of unconfirmed bugs low
1581
* ensuring the priorities are generally right (everything as critical - or
1582
medium - is meaningless)
1583
* looking out for regressions and turning those around sooner rather than later.
1586
As well as prioritizing bugs and nominating them against a
1587
target milestone, Launchpad lets core developers offer to mentor others in
1592
vim: ft=rst tw=74 ai
added
removed
doc/en/developer-guide/HACKING.txt
