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 latest developer documentation can be found online at
15
http://doc.bazaar-vcs.org/developers/.
21
Exploring the Bazaar Platform
22
=============================
24
Before making changes, it's a good idea to explore the work already
25
done by others. Perhaps the new feature or improvement you're looking
26
for is available in another plug-in already? If you find a bug,
27
perhaps someone else has already fixed it?
29
To answer these questions and more, take a moment to explore the
30
overall Bazaar Platform. Here are some links to browse:
32
* The Plugins page on the Wiki - http://bazaar-vcs.org/BzrPlugins
34
* The Bazaar product family on Launchpad - https://launchpad.net/bazaar
36
* Bug Tracker for the core product - https://bugs.launchpad.net/bzr/
38
* Blueprint Tracker for the core product - https://blueprints.launchpad.net/bzr/
40
If nothing else, perhaps you'll find inspiration in how other developers
41
have solved their challenges.
43
Finding Something To Do
44
=======================
46
Ad-hoc performance work can also be done. One useful tool is the 'evil' debug
47
flag. For instance running ``bzr -Devil commit -m "test"`` will log a backtrace
48
to the bzr log file for every method call which triggers a slow or non-scalable
49
part of the bzr library. So checking that a given command with ``-Devil`` has
50
no backtraces logged to the log file is a good way to find problem function
51
calls that might be nested deep in the code base.
53
Planning and Discussing Changes
54
===============================
56
There is a very active community around Bazaar. Mostly we meet on IRC
57
(#bzr on irc.freenode.net) and on the mailing list. To join the Bazaar
58
community, see http://bazaar-vcs.org/BzrSupport.
60
If you are planning to make a change, it's a very good idea to mention it
61
on the IRC channel and/or on the mailing list. There are many advantages
62
to involving the community before you spend much time on a change.
65
* you get to build on the wisdom of others, saving time
67
* if others can direct you to similar code, it minimises the work to be done
69
* it assists everyone in coordinating direction, priorities and effort.
71
In summary, maximising the input from others typically minimises the
72
total effort required to get your changes merged. The community is
73
friendly, helpful and always keen to welcome newcomers.
76
Bazaar Development in a Nutshell
77
================================
79
.. was from bazaar-vcs.org/BzrGivingBack
81
One of the fun things about working on a version control system like Bazaar is
82
that the users have a high level of proficiency in contributing back into
83
the tool. Consider the following very brief introduction to contributing back
84
to Bazaar. More detailed instructions are in the following sections.
89
First, get a local copy of the development mainline (See `Why make a local
95
$ bzr branch http://bazaar-vcs.org/bzr/bzr.dev/ bzr.dev
97
Now make your own branch::
99
$ bzr branch bzr.dev 123456-my-bugfix
101
This will give you a branch called "123456-my-bugfix" that you can work on
102
and commit in. Here, you can study the code, make a fix or a new feature.
103
Feel free to commit early and often (after all, it's your branch!).
105
Documentation improvements are an easy place to get started giving back to the
106
Bazaar project. The documentation is in the `doc/` subdirectory of the Bazaar
109
When you are done, make sure that you commit your last set of changes as well!
110
Once you are happy with your changes, ask for them to be merged, as described
113
Making a Merge Proposal
114
-----------------------
116
The Bazaar developers use Launchpad to further enable a truly distributed
117
style of development. Anyone can propose a branch for merging into the Bazaar
118
trunk. To start this process, you need to push your branch to Launchpad. To
119
do this, you will need a Launchpad account and user name, e.g.
120
`your_lp_username`. You can push your branch to Launchpad directly from
123
$ bzr push lp:~your_lp_username/bzr/meaningful_name_here
125
After you have pushed your branch, you will need to propose it for merging to
126
the Bazaar trunk. Go to
127
<https://launchpad.net/your_lp_username/bzr/meaningful_name_here> and choose
128
"Propose for merging into another branch". Select "~bzr/bzr/trunk" to hand
129
your changes off to the Bazaar developers for review and merging.
131
Using a meaningful name for your branch will help you and the reviewer(s)
132
better track the submission. Use a very succint description of your submission
133
and prefix it with bug number if needed (lp:~mbp/bzr/484558-merge-directory
134
for example). Alternatively, you can suffix with the bug number
135
(lp:~jameinel/bzr/export-file-511987).
138
Why make a local copy of bzr.dev?
139
---------------------------------
141
Making a local mirror of bzr.dev is not strictly necessary, but it means
143
- You can use that copy of bzr.dev as your main bzr executable, and keep it
144
up-to-date using ``bzr pull``.
145
- Certain operations are faster, and can be done when offline. For example:
148
- ``bzr diff -r ancestor:...``
151
- When it's time to create your next branch, it's more convenient. When you
152
have further contributions to make, you should do them in their own branch::
155
$ bzr branch bzr.dev additional_fixes
156
$ cd additional_fixes # hack, hack, hack
160
Understanding the Development Process
161
=====================================
163
The development team follows many practices including:
165
* a public roadmap and planning process in which anyone can participate
167
* time based milestones everyone can work towards and plan around
169
* extensive code review and feedback to contributors
171
* complete and rigorous test coverage on any code contributed
173
* automated validation that all tests still pass before code is merged
174
into the main code branch.
176
The key tools we use to enable these practices are:
178
* Launchpad - https://launchpad.net/
180
* Bazaar - http://bazaar-vcs.org/
182
* Patch Queue Manager - https://launchpad.net/pqm/
184
For further information, see <http://wiki.bazaar.canonical.com/BzrDevelopment>.
189
Preparing a Sandbox for Making Changes to Bazaar
190
================================================
192
Bazaar supports many ways of organising your work. See
193
http://bazaar-vcs.org/SharedRepositoryLayouts for a summary of the
194
popular alternatives.
196
Of course, the best choice for you will depend on numerous factors:
197
the number of changes you may be making, the complexity of the changes, etc.
198
As a starting suggestion though:
200
* create a local copy of the main development branch (bzr.dev) by using
203
bzr branch http://bazaar-vcs.org/bzr/bzr.dev/ bzr.dev
205
* keep your copy of bzr.dev pristine (by not developing in it) and keep
206
it up to date (by using bzr pull)
208
* create a new branch off your local bzr.dev copy for each issue
209
(bug or feature) you are working on.
211
This approach makes it easy to go back and make any required changes
212
after a code review. Resubmitting the change is then simple with no
213
risk of accidentally including edits related to other issues you may
214
be working on. After the changes for an issue are accepted and merged,
215
the associated branch can be deleted or archived as you wish.
218
Navigating the Code Base
219
========================
221
.. Was at <http://bazaar-vcs.org/NewDeveloperIntroduction>
223
Some of the key files in this directory are:
226
The command you run to start Bazaar itself. This script is pretty
227
short and just does some checks then jumps into bzrlib.
230
This file covers a brief introduction to Bazaar and lists some of its
234
Summary of changes in each Bazaar release that can affect users or
238
Installs Bazaar system-wide or to your home directory. To perform
239
development work on Bazaar it is not required to run this file - you
240
can simply run the bzr command from the top level directory of your
241
development copy. Note: That if you run setup.py this will create a
242
'build' directory in your development branch. There's nothing wrong
243
with this but don't be confused by it. The build process puts a copy
244
of the main code base into this build directory, along with some other
245
files. You don't need to go in here for anything discussed in this
249
Possibly the most exciting folder of all, bzrlib holds the main code
250
base. This is where you will go to edit python files and contribute to
254
Holds documentation on a whole range of things on Bazaar from the
255
origination of ideas within the project to information on Bazaar
256
features and use cases. Within this directory there is a subdirectory
257
for each translation into a human language. All the documentation
258
is in the ReStructuredText markup language.
261
Documentation specifically targeted at Bazaar and plugin developers.
262
(Including this document.)
266
Automatically-generated API reference information is available at
267
<http://starship.python.net/crew/mwh/bzrlibapi/>.
269
See also the `Bazaar Architectural Overview
270
<http://doc.bazaar-vcs.org/developers/overview.html>`_.
273
Coding Style Guidelines
274
#######################
279
``hasattr`` should not be used because it swallows exceptions including
280
``KeyboardInterrupt``. Instead, say something like ::
282
if getattr(thing, 'name', None) is None
288
Please write PEP-8__ compliant code.
290
__ http://www.python.org/peps/pep-0008.html
292
One often-missed requirement is that the first line of docstrings
293
should be a self-contained one-sentence summary.
295
We use 4 space indents for blocks, and never use tab characters. (In vim,
298
Trailing white space should be avoided, but is allowed.
299
You should however not make lots of unrelated white space changes.
301
Unix style newlines (LF) are used.
303
Each file must have a newline at the end of it.
305
Lines should be no more than 79 characters if at all possible.
306
Lines that continue a long statement may be indented in either of
309
within the parenthesis or other character that opens the block, e.g.::
315
or indented by four spaces::
321
The first is considered clearer by some people; however it can be a bit
322
harder to maintain (e.g. when the method name changes), and it does not
323
work well if the relevant parenthesis is already far to the right. Avoid
326
self.legbone.kneebone.shinbone.toebone.shake_it(one,
332
self.legbone.kneebone.shinbone.toebone.shake_it(one,
338
self.legbone.kneebone.shinbone.toebone.shake_it(
341
For long lists, we like to add a trailing comma and put the closing
342
character on the following line. This makes it easier to add new items in
345
from bzrlib.goo import (
351
There should be spaces between function parameters, but not between the
352
keyword name and the value::
354
call(1, 3, cheese=quark)
358
;(defface my-invalid-face
359
; '((t (:background "Red" :underline t)))
360
; "Face used to highlight invalid constructs or other uglyties"
363
(defun my-python-mode-hook ()
364
;; setup preferred indentation style.
365
(setq fill-column 79)
366
(setq indent-tabs-mode nil) ; no tabs, never, I will not repeat
367
; (font-lock-add-keywords 'python-mode
368
; '(("^\\s *\t" . 'my-invalid-face) ; Leading tabs
369
; ("[ \t]+$" . 'my-invalid-face) ; Trailing spaces
370
; ("^[ \t]+$" . 'my-invalid-face)); Spaces only
374
(add-hook 'python-mode-hook 'my-python-mode-hook)
376
The lines beginning with ';' are comments. They can be activated
377
if one want to have a strong notice of some tab/space usage
384
* Imports should be done at the top-level of the file, unless there is
385
a strong reason to have them lazily loaded when a particular
386
function runs. Import statements have a cost, so try to make sure
387
they don't run inside hot functions.
389
* Module names should always be given fully-qualified,
390
i.e. ``bzrlib.hashcache`` not just ``hashcache``.
396
Functions, methods or members that are relatively private are given
397
a leading underscore prefix. Names without a leading underscore are
398
public not just across modules but to programmers using bzrlib as an
401
We prefer class names to be concatenated capital words (``TestCase``)
402
and variables, methods and functions to be lowercase words joined by
403
underscores (``revision_id``, ``get_revision``).
405
For the purposes of naming some names are treated as single compound
406
words: "filename", "revno".
408
Consider naming classes as nouns and functions/methods as verbs.
410
Try to avoid using abbreviations in names, because there can be
411
inconsistency if other people use the full name.
417
``revision_id`` not ``rev_id`` or ``revid``
419
Functions that transform one thing to another should be named ``x_to_y``
420
(not ``x2y`` as occurs in some old code.)
426
Python destructors (``__del__``) work differently to those of other
427
languages. In particular, bear in mind that destructors may be called
428
immediately when the object apparently becomes unreferenced, or at some
429
later time, or possibly never at all. Therefore we have restrictions on
430
what can be done inside them.
432
0. If you think you need to use a ``__del__`` method ask another
433
developer for alternatives. If you do need to use one, explain
436
1. Never rely on a ``__del__`` method running. If there is code that
437
must run, do it from a ``finally`` block instead.
439
2. Never ``import`` from inside a ``__del__`` method, or you may crash the
442
3. In some places we raise a warning from the destructor if the object
443
has not been cleaned up or closed. This is considered OK: the warning
444
may not catch every case but it's still useful sometimes.
450
Often when something has failed later code, including cleanups invoked
451
from ``finally`` blocks, will fail too. These secondary failures are
452
generally uninteresting compared to the original exception. So use the
453
``only_raises`` decorator (from ``bzrlib.decorators``) for methods that
454
are typically called in ``finally`` blocks, such as ``unlock`` methods.
455
For example, ``@only_raises(LockNotHeld, LockBroken)``. All errors that
456
are unlikely to be a knock-on failure from a previous failure should be
463
In some places we have variables which point to callables that construct
464
new instances. That is to say, they can be used a lot like class objects,
465
but they shouldn't be *named* like classes::
467
> I think that things named FooBar should create instances of FooBar when
468
> called. Its plain confusing for them to do otherwise. When we have
469
> something that is going to be used as a class - that is, checked for via
470
> isinstance or other such idioms, them I would call it foo_class, so that
471
> it is clear that a callable is not sufficient. If it is only used as a
472
> factory, then yes, foo_factory is what I would use.
478
Several places in Bazaar use (or will use) a registry, which is a
479
mapping from names to objects or classes. The registry allows for
480
loading in registered code only when it's needed, and keeping
481
associated information such as a help string or description.
484
InterObject and multiple dispatch
485
=================================
487
The ``InterObject`` provides for two-way `multiple dispatch`__: matching
488
up for example a source and destination repository to find the right way
489
to transfer data between them.
491
.. __: http://en.wikipedia.org/wiki/Multiple_dispatch
493
There is a subclass ``InterObject`` classes for each type of object that is
494
dispatched this way, e.g. ``InterRepository``. Calling ``.get()`` on this
495
class will return an ``InterObject`` instance providing the best match for
496
those parameters, and this instance then has methods for operations
501
inter = InterRepository.get(source_repo, target_repo)
502
inter.fetch(revision_id)
504
``InterRepository`` also acts as a registry-like object for its
505
subclasses, and they can be added through ``.register_optimizer``. The
506
right one to run is selected by asking each class, in reverse order of
507
registration, whether it ``.is_compatible`` with the relevant objects.
512
To make startup time faster, we use the ``bzrlib.lazy_import`` module to
513
delay importing modules until they are actually used. ``lazy_import`` uses
514
the same syntax as regular python imports. So to import a few modules in a
517
from bzrlib.lazy_import import lazy_import
518
lazy_import(globals(), """
527
revision as _mod_revision,
529
import bzrlib.transport
533
At this point, all of these exist as a ``ImportReplacer`` object, ready to
534
be imported once a member is accessed. Also, when importing a module into
535
the local namespace, which is likely to clash with variable names, it is
536
recommended to prefix it as ``_mod_<module>``. This makes it clearer that
537
the variable is a module, and these object should be hidden anyway, since
538
they shouldn't be imported into other namespaces.
540
While it is possible for ``lazy_import()`` to import members of a module
541
when using the ``from module import member`` syntax, it is recommended to
542
only use that syntax to load sub modules ``from module import submodule``.
543
This is because variables and classes can frequently be used without
544
needing a sub-member for example::
546
lazy_import(globals(), """
547
from module import MyClass
551
return isinstance(x, MyClass)
553
This will incorrectly fail, because ``MyClass`` is a ``ImportReplacer``
554
object, rather than the real class.
556
It also is incorrect to assign ``ImportReplacer`` objects to other variables.
557
Because the replacer only knows about the original name, it is unable to
558
replace other variables. The ``ImportReplacer`` class will raise an
559
``IllegalUseOfScopeReplacer`` exception if it can figure out that this
560
happened. But it requires accessing a member more than once from the new
561
variable, so some bugs are not detected right away.
567
The null revision is the ancestor of all revisions. Its revno is 0, its
568
revision-id is ``null:``, and its tree is the empty tree. When referring
569
to the null revision, please use ``bzrlib.revision.NULL_REVISION``. Old
570
code sometimes uses ``None`` for the null revision, but this practice is
574
Object string representations
575
=============================
577
Python prints objects using their ``__repr__`` method when they are
578
written to logs, exception tracebacks, or the debugger. We want
579
objects to have useful representations to help in determining what went
582
If you add a new class you should generally add a ``__repr__`` method
583
unless there is an adequate method in a parent class. There should be a
586
Representations should typically look like Python constructor syntax, but
587
they don't need to include every value in the object and they don't need
588
to be able to actually execute. They're to be read by humans, not
589
machines. Don't hardcode the classname in the format, so that we get the
590
correct value if the method is inherited by a subclass. If you're
591
printing attributes of the object, including strings, you should normally
592
use ``%r`` syntax (to call their repr in turn).
594
Try to avoid the representation becoming more than one or two lines long.
595
(But balance this against including useful information, and simplicity of
598
Because repr methods are often called when something has already gone
599
wrong, they should be written somewhat more defensively than most code.
600
The object may be half-initialized or in some other way in an illegal
601
state. The repr method shouldn't raise an exception, or it may hide the
602
(probably more useful) underlying exception.
607
return '%s(%r)' % (self.__class__.__name__,
614
A bare ``except`` statement will catch all exceptions, including ones that
615
really should terminate the program such as ``MemoryError`` and
616
``KeyboardInterrupt``. They should rarely be used unless the exception is
617
later re-raised. Even then, think about whether catching just
618
``Exception`` (which excludes system errors in Python2.5 and later) would
625
All code should be exercised by the test suite. See the `Bazaar Testing
626
Guide <http://doc.bazaar-vcs.org/developers/testing.html>`_ for detailed
627
information about writing tests.
636
We don't change APIs in stable branches: any supported symbol in a stable
637
release of bzr must not be altered in any way that would result in
638
breaking existing code that uses it. That means that method names,
639
parameter ordering, parameter names, variable and attribute names etc must
640
not be changed without leaving a 'deprecated forwarder' behind. This even
641
applies to modules and classes.
643
If you wish to change the behaviour of a supported API in an incompatible
644
way, you need to change its name as well. For instance, if I add an optional keyword
645
parameter to branch.commit - that's fine. On the other hand, if I add a
646
keyword parameter to branch.commit which is a *required* transaction
647
object, I should rename the API - i.e. to 'branch.commit_transaction'.
649
(Actually, that may break code that provides a new implementation of
650
``commit`` and doesn't expect to receive the parameter.)
652
When renaming such supported API's, be sure to leave a deprecated_method (or
653
_function or ...) behind which forwards to the new API. See the
654
bzrlib.symbol_versioning module for decorators that take care of the
655
details for you - such as updating the docstring, and issuing a warning
656
when the old API is used.
658
For unsupported API's, it does not hurt to follow this discipline, but it's
659
not required. Minimally though, please try to rename things so that
660
callers will at least get an AttributeError rather than weird results.
663
Deprecation decorators
664
----------------------
666
``bzrlib.symbol_versioning`` provides decorators that can be attached to
667
methods, functions, and other interfaces to indicate that they should no
668
longer be used. For example::
670
@deprecated_method(deprecated_in((0, 1, 4)))
672
return self._new_foo()
674
To deprecate a static method you must call ``deprecated_function``
675
(**not** method), after the staticmethod call::
678
@deprecated_function(deprecated_in((0, 1, 4)))
679
def create_repository(base, shared=False, format=None):
681
When you deprecate an API, you should not just delete its tests, because
682
then we might introduce bugs in them. If the API is still present at all,
683
it should still work. The basic approach is to use
684
``TestCase.applyDeprecated`` which in one step checks that the API gives
685
the expected deprecation message, and also returns the real result from
686
the method, so that tests can keep running.
688
Deprecation warnings will be suppressed for final releases, but not for
689
development versions or release candidates, or when running ``bzr
690
selftest``. This gives developers information about whether their code is
691
using deprecated functions, but avoids confusing users about things they
698
Processing Command Lines
699
------------------------
701
bzrlib has a standard framework for parsing command lines and calling
702
processing routines associated with various commands. See builtins.py
703
for numerous examples.
706
Standard Parameter Types
707
------------------------
709
There are some common requirements in the library: some parameters need to be
710
unicode safe, some need byte strings, and so on. At the moment we have
711
only codified one specific pattern: Parameters that need to be unicode
712
should be checked via ``bzrlib.osutils.safe_unicode``. This will coerce the
713
input into unicode in a consistent fashion, allowing trivial strings to be
714
used for programmer convenience, but not performing unpredictably in the
715
presence of different locales.
721
(The strategy described here is what we want to get to, but it's not
722
consistently followed in the code at the moment.)
724
bzrlib is intended to be a generically reusable library. It shouldn't
725
write messages to stdout or stderr, because some programs that use it
726
might want to display that information through a GUI or some other
729
We can distinguish two types of output from the library:
731
1. Structured data representing the progress or result of an
732
operation. For example, for a commit command this will be a list
733
of the modified files and the finally committed revision number
736
These should be exposed either through the return code or by calls
737
to a callback parameter.
739
A special case of this is progress indicators for long-lived
740
operations, where the caller should pass a ProgressBar object.
742
2. Unstructured log/debug messages, mostly for the benefit of the
743
developers or users trying to debug problems. This should always
744
be sent through ``bzrlib.trace`` and Python ``logging``, so that
745
it can be redirected by the client.
747
The distinction between the two is a bit subjective, but in general if
748
there is any chance that a library would want to see something as
749
structured data, we should make it so.
751
The policy about how output is presented in the text-mode client
752
should be only in the command-line tool.
755
Progress and Activity Indications
756
---------------------------------
758
bzrlib has a way for code to display to the user that stuff is happening
759
during a long operation. There are two particular types: *activity* which
760
means that IO is happening on a Transport, and *progress* which means that
761
higher-level application work is occurring. Both are drawn together by
764
Transport objects are responsible for calling `report_transport_activity`
767
Progress uses a model/view pattern: application code acts on a
768
`ProgressTask` object, which notifies the UI when it needs to be
769
displayed. Progress tasks form a stack. To create a new progress task on
770
top of the stack, call `bzrlib.ui.ui_factory.nested_progress_bar()`, then
771
call `update()` on the returned ProgressTask. It can be updated with just
772
a text description, with a numeric count, or with a numeric count and
773
expected total count. If an expected total count is provided the view
774
can show the progress moving along towards the expected total.
776
The user should call `finish` on the `ProgressTask` when the logical
777
operation has finished, so it can be removed from the stack.
779
Progress tasks have a complex relationship with generators: it's a very
780
good place to use them, but because python2.4 does not allow ``finally``
781
blocks in generators it's hard to clean them up properly. In this case
782
it's probably better to have the code calling the generator allocate a
783
progress task for its use and then call `finalize` when it's done, which
784
will close it if it was not already closed. The generator should also
785
finish the progress task when it exits, because it may otherwise be a long
786
time until the finally block runs.
792
When filenames or similar variables are presented inline within a message,
793
they should be enclosed in double quotes (ascii 0x22, not chiral unicode
796
bzr: ERROR: No such file "asdf"
798
When we print just a list of filenames there should not be any quoting:
801
.. _bug 544297: https://bugs.launchpad.net/bugs/544297
803
https://wiki.ubuntu.com/UnitsPolicy provides a good explanation about
804
which unit should be used when. Roughly speaking, IEC standard applies
805
for base-2 units and SI standard applies for base-10 units:
807
* for network bandwidth and disk sizes, use base-10 (Mbits/s, kB/s, GB)
809
* for RAM sizes, use base-2 (GiB, TiB)
816
Bazaar has online help for various topics through ``bzr help COMMAND`` or
817
equivalently ``bzr command -h``. We also have help on command options,
818
and on other help topics. (See ``help_topics.py``.)
820
As for python docstrings, the first paragraph should be a single-sentence
821
synopsis of the command. These are user-visible and should be prefixed with
822
``__doc__ =`` so help works under ``python -OO`` with docstrings stripped.
824
The help for options should be one or more proper sentences, starting with
825
a capital letter and finishing with a full stop (period).
827
All help messages and documentation should have two spaces between
831
Handling Errors and Exceptions
832
==============================
834
Commands should return non-zero when they encounter circumstances that
835
the user should really pay attention to - which includes trivial shell
838
Recommended values are:
841
1. Conflicts in merge-like operations, or changes are present in
842
diff-like operations.
843
2. Unrepresentable diff changes (i.e. binary files that we cannot show
845
3. An error or exception has occurred.
846
4. An internal error occurred (one that shows a traceback.)
848
Errors are handled through Python exceptions. Exceptions should be defined
849
inside bzrlib.errors, so that we can see the whole tree at a glance.
851
We broadly classify errors as either being either internal or not,
852
depending on whether ``internal_error`` is set or not. If we think it's our
853
fault, we show a backtrace, an invitation to report the bug, and possibly
854
other details. This is the default for errors that aren't specifically
855
recognized as being caused by a user error. Otherwise we show a briefer
856
message, unless -Derror was given.
858
Many errors originate as "environmental errors" which are raised by Python
859
or builtin libraries -- for example IOError. These are treated as being
860
our fault, unless they're caught in a particular tight scope where we know
861
that they indicate a user errors. For example if the repository format
862
is not found, the user probably gave the wrong path or URL. But if one of
863
the files inside the repository is not found, then it's our fault --
864
either there's a bug in bzr, or something complicated has gone wrong in
865
the environment that means one internal file was deleted.
867
Many errors are defined in ``bzrlib/errors.py`` but it's OK for new errors
868
to be added near the place where they are used.
870
Exceptions are formatted for the user by conversion to a string
871
(eventually calling their ``__str__`` method.) As a convenience the
872
``._fmt`` member can be used as a template which will be mapped to the
873
error's instance dict.
875
New exception classes should be defined when callers might want to catch
876
that exception specifically, or when it needs a substantially different
879
#. If it is something that a caller can recover from, a custom exception
882
#. If it is a data consistency issue, using a builtin like
883
``ValueError``/``TypeError`` is reasonable.
885
#. If it is a programmer error (using an api incorrectly)
886
``AssertionError`` is reasonable.
888
#. Otherwise, use ``BzrError`` or ``InternalBzrError``.
890
Exception strings should start with a capital letter and should not have a
891
final fullstop. If long, they may contain newlines to break the text.
897
Do not use the Python ``assert`` statement, either in tests or elsewhere.
898
A source test checks that it is not used. It is ok to explicitly raise
903
* It makes the behaviour vary depending on whether bzr is run with -O
904
or not, therefore giving a chance for bugs that occur in one case or
905
the other, several of which have already occurred: assertions with
906
side effects, code which can't continue unless the assertion passes,
907
cases where we should give the user a proper message rather than an
909
* It's not that much shorter than an explicit if/raise.
910
* It tends to lead to fuzzy thinking about whether the check is
911
actually needed or not, and whether it's an internal error or not
912
* It tends to cause look-before-you-leap patterns.
913
* It's unsafe if the check is needed to protect the integrity of the
915
* It tends to give poor messages since the developer can get by with
916
no explanatory text at all.
917
* We can't rely on people always running with -O in normal use, so we
918
can't use it for tests that are actually expensive.
919
* Expensive checks that help developers are better turned on from the
920
test suite or a -D flag.
921
* If used instead of ``self.assert*()`` in tests it makes them falsely pass with -O.
927
When you change bzrlib, please update the relevant documentation for the
928
change you made: Changes to commands should update their help, and
929
possibly end user tutorials; changes to the core library should be
930
reflected in API documentation.
935
If you make a user-visible change, please add a note to the NEWS file.
936
The description should be written to make sense to someone who's just
937
a user of bzr, not a developer: new functions or classes shouldn't be
938
mentioned, but new commands, changes in behaviour or fixed nontrivial
939
bugs should be listed. See the existing entries for an idea of what
942
Within each release, entries in the news file should have the most
943
user-visible changes first. So the order should be approximately:
945
* changes to existing behaviour - the highest priority because the
946
user's existing knowledge is incorrect
947
* new features - should be brought to their attention
948
* bug fixes - may be of interest if the bug was affecting them, and
949
should include the bug number if any
950
* major documentation changes, including fixed documentation bugs
951
* changes to internal interfaces
953
People who made significant contributions to each change are listed in
954
parenthesis. This can include reporting bugs (particularly with good
955
details or reproduction recipes), submitting patches, etc.
957
To help with merging, NEWS entries should be sorted lexicographically
963
The docstring of a command is used by ``bzr help`` to generate help output
964
for the command. The list 'takes_options' attribute on a command is used by
965
``bzr help`` to document the options for the command - the command
966
docstring does not need to document them. Finally, the '_see_also'
967
attribute on a command can be used to reference other related help topics.
972
Functions, methods, classes and modules should have docstrings
973
describing how they are used.
975
The first line of the docstring should be a self-contained sentence.
977
For the special case of Command classes, this acts as the user-visible
978
documentation shown by the help command.
980
The docstrings should be formatted as reStructuredText_ (like this
981
document), suitable for processing using the epydoc_ tool into HTML
984
.. _reStructuredText: http://docutils.sourceforge.net/rst.html
985
.. _epydoc: http://epydoc.sourceforge.net/
994
The copyright policy for bzr was recently made clear in this email (edited
995
for grammatical correctness)::
997
The attached patch cleans up the copyright and license statements in
998
the bzr source. It also adds tests to help us remember to add them
999
with the correct text.
1001
We had the problem that lots of our files were "Copyright Canonical
1002
Development Ltd" which is not a real company, and some other variations
1003
on this theme. Also, some files were missing the GPL statements.
1005
I want to be clear about the intent of this patch, since copyright can
1006
be a little controversial.
1008
1) The big motivation for this is not to shut out the community, but
1009
just to clean up all of the invalid copyright statements.
1011
2) It has been the general policy for bzr that we want a single
1012
copyright holder for all of the core code. This is following the model
1013
set by the FSF, which makes it easier to update the code to a new
1014
license in case problems are encountered. (For example, if we want to
1015
upgrade the project universally to GPL v3 it is much simpler if there is
1016
a single copyright holder). It also makes it clearer if copyright is
1017
ever debated, there is a single holder, which makes it easier to defend
1018
in court, etc. (I think the FSF position is that if you assign them
1019
copyright, they can defend it in court rather than you needing to, and
1020
I'm sure Canonical would do the same).
1021
As such, Canonical has requested copyright assignments from all of the
1024
3) If someone wants to add code and not attribute it to Canonical, there
1025
is a specific list of files that are excluded from this check. And the
1026
test failure indicates where that is, and how to update it.
1028
4) If anyone feels that I changed a copyright statement incorrectly, just
1029
let me know, and I'll be happy to correct it. Whenever you have large
1030
mechanical changes like this, it is possible to make some mistakes.
1032
Just to reiterate, this is a community project, and it is meant to stay
1033
that way. Core bzr code is copyright Canonical for legal reasons, and
1034
the tests are just there to help us maintain that.
1037
Miscellaneous Topics
1038
####################
1043
Bazaar has a few facilities to help debug problems by going into pdb_, the
1046
.. _pdb: http://docs.python.org/lib/debugger-commands.html
1048
If the ``BZR_PDB`` environment variable is set
1049
then bzr will go into pdb post-mortem mode when an unhandled exception
1052
If you send a SIGQUIT or SIGBREAK signal to bzr then it will drop into the
1053
debugger immediately. SIGQUIT can be generated by pressing Ctrl-\\ on
1054
Unix. SIGBREAK is generated with Ctrl-Pause on Windows (some laptops have
1055
this as Fn-Pause). You can continue execution by typing ``c``. This can
1056
be disabled if necessary by setting the environment variable
1057
``BZR_SIGQUIT_PDB=0``.
1063
Bazaar accepts some global options starting with ``-D`` such as
1064
``-Dhpss``. These set a value in `bzrlib.debug.debug_flags`, and
1065
typically cause more information to be written to the trace file. Most
1066
`mutter` calls should be guarded by a check of those flags so that we
1067
don't write out too much information if it's not needed.
1069
Debug flags may have effects other than just emitting trace messages.
1071
Run ``bzr help global-options`` to see them all.
1073
These flags may also be set as a comma-separated list in the
1074
``debug_flags`` option in e.g. ``~/.bazaar/bazaar.conf``. (Note that it
1075
must be in this global file, not in the branch or location configuration,
1076
because it's currently only loaded at startup time.) For instance you may
1077
want to always record hpss traces and to see full error tracebacks::
1079
debug_flags = hpss, error
1086
Integer identifier for a revision on the main line of a branch.
1087
Revision 0 is always the null revision; others are 1-based
1088
indexes into the branch's revision history.
1091
Unicode and Encoding Support
1092
============================
1094
This section discusses various techniques that Bazaar uses to handle
1095
characters that are outside the ASCII set.
1100
When a ``Command`` object is created, it is given a member variable
1101
accessible by ``self.outf``. This is a file-like object, which is bound to
1102
``sys.stdout``, and should be used to write information to the screen,
1103
rather than directly writing to ``sys.stdout`` or calling ``print``.
1104
This file has the ability to translate Unicode objects into the correct
1105
representation, based on the console encoding. Also, the class attribute
1106
``encoding_type`` will effect how unprintable characters will be
1107
handled. This parameter can take one of 3 values:
1110
Unprintable characters will be represented with a suitable replacement
1111
marker (typically '?'), and no exception will be raised. This is for
1112
any command which generates text for the user to review, rather than
1113
for automated processing.
1114
For example: ``bzr log`` should not fail if one of the entries has text
1115
that cannot be displayed.
1118
Attempting to print an unprintable character will cause a UnicodeError.
1119
This is for commands that are intended more as scripting support, rather
1120
than plain user review.
1121
For example: ``bzr ls`` is designed to be used with shell scripting. One
1122
use would be ``bzr ls --null --unknowns | xargs -0 rm``. If ``bzr``
1123
printed a filename with a '?', the wrong file could be deleted. (At the
1124
very least, the correct file would not be deleted). An error is used to
1125
indicate that the requested action could not be performed.
1128
Do not attempt to automatically convert Unicode strings. This is used
1129
for commands that must handle conversion themselves.
1130
For example: ``bzr diff`` needs to translate Unicode paths, but should
1131
not change the exact text of the contents of the files.
1134
``bzrlib.urlutils.unescape_for_display``
1135
----------------------------------------
1137
Because Transports work in URLs (as defined earlier), printing the raw URL
1138
to the user is usually less than optimal. Characters outside the standard
1139
set are printed as escapes, rather than the real character, and local
1140
paths would be printed as ``file://`` urls. The function
1141
``unescape_for_display`` attempts to unescape a URL, such that anything
1142
that cannot be printed in the current encoding stays an escaped URL, but
1143
valid characters are generated where possible.
1149
The ``bzrlib.osutils`` module has many useful helper functions, including
1150
some more portable variants of functions in the standard library.
1152
In particular, don't use ``shutil.rmtree`` unless it's acceptable for it
1153
to fail on Windows if some files are readonly or still open elsewhere.
1154
Use ``bzrlib.osutils.rmtree`` instead.
1160
We write some extensions in C using pyrex. We design these to work in
1163
* User with no C compiler
1164
* User with C compiler
1167
The recommended way to install bzr is to have a C compiler so that the
1168
extensions can be built, but if no C compiler is present, the pure python
1169
versions we supply will work, though more slowly.
1171
For developers we recommend that pyrex be installed, so that the C
1172
extensions can be changed if needed.
1174
For the C extensions, the extension module should always match the
1175
original python one in all respects (modulo speed). This should be
1176
maintained over time.
1178
To create an extension, add rules to setup.py for building it with pyrex,
1179
and with distutils. Now start with an empty .pyx file. At the top add
1180
"include 'yourmodule.py'". This will import the contents of foo.py into this
1181
file at build time - remember that only one module will be loaded at
1182
runtime. Now you can subclass classes, or replace functions, and only your
1183
changes need to be present in the .pyx file.
1185
Note that pyrex does not support all 2.4 programming idioms, so some
1186
syntax changes may be required. I.e.
1188
- 'from foo import (bar, gam)' needs to change to not use the brackets.
1189
- 'import foo.bar as bar' needs to be 'import foo.bar; bar = foo.bar'
1191
If the changes are too dramatic, consider
1192
maintaining the python code twice - once in the .pyx, and once in the .py,
1193
and no longer including the .py file.
1196
Making Installers for OS Windows
1197
================================
1198
To build a win32 installer, see the instructions on the wiki page:
1199
http://bazaar-vcs.org/BzrWin32Installer
1202
Core Developer Tasks
1203
####################
1208
What is a Core Developer?
1209
-------------------------
1211
While everyone in the Bazaar community is welcome and encouraged to
1212
propose and submit changes, a smaller team is reponsible for pulling those
1213
changes together into a cohesive whole. In addition to the general developer
1214
stuff covered above, "core" developers have responsibility for:
1217
* reviewing blueprints
1219
* managing releases (see `Releasing Bazaar <http://doc.bazaar-vcs.org/developers/releasing.html>`_)
1222
Removing barriers to community participation is a key reason for adopting
1223
distributed VCS technology. While DVCS removes many technical barriers,
1224
a small number of social barriers are often necessary instead.
1225
By documenting how the above things are done, we hope to
1226
encourage more people to participate in these activities, keeping the
1227
differences between core and non-core contributors to a minimum.
1230
Communicating and Coordinating
1231
------------------------------
1233
While it has many advantages, one of the challenges of distributed
1234
development is keeping everyone else aware of what you're working on.
1235
There are numerous ways to do this:
1237
#. Assign bugs to yourself in Launchpad
1238
#. Mention it on the mailing list
1239
#. Mention it on IRC
1241
As well as the email notifcations that occur when merge requests are sent
1242
and reviewed, you can keep others informed of where you're spending your
1243
energy by emailing the **bazaar-commits** list implicitly. To do this,
1244
install and configure the Email plugin. One way to do this is add these
1245
configuration settings to your central configuration file (e.g.
1246
``~/.bazaar/bazaar.conf`` on Linux)::
1249
email = Joe Smith <joe.smith@internode.on.net>
1250
smtp_server = mail.internode.on.net:25
1252
Then add these lines for the relevant branches in ``locations.conf``::
1254
post_commit_to = bazaar-commits@lists.canonical.com
1255
post_commit_mailer = smtplib
1257
While attending a sprint, RobertCollins' Dbus plugin is useful for the
1258
same reason. See the documentation within the plugin for information on
1259
how to set it up and configure it.
1268
Of the many workflows supported by Bazaar, the one adopted for Bazaar
1269
development itself is known as "Decentralized with automatic gatekeeper".
1270
To repeat the explanation of this given on
1271
http://bazaar-vcs.org/Workflows:
1274
In this workflow, each developer has their own branch or
1275
branches, plus read-only access to the mainline. A software gatekeeper
1276
(e.g. PQM) has commit rights to the main branch. When a developer wants
1277
their work merged, they request the gatekeeper to merge it. The gatekeeper
1278
does a merge, a compile, and runs the test suite. If the code passes, it
1279
is merged into the mainline.
1281
In a nutshell, here's the overall submission process:
1283
#. get your work ready (including review except for trivial changes)
1284
#. push to a public location
1285
#. ask PQM to merge from that location
1288
At present, PQM always takes the changes to merge from a branch
1289
at a URL that can be read by it. For Bazaar, that means a public,
1290
typically http, URL.
1292
As a result, the following things are needed to use PQM for submissions:
1294
#. A publicly available web server
1295
#. Your OpenPGP key registered with PQM (contact RobertCollins for this)
1296
#. The PQM plugin installed and configured (not strictly required but
1297
highly recommended).
1300
Selecting a Public Branch Location
1301
----------------------------------
1303
If you don't have your own web server running, branches can always be
1304
pushed to Launchpad. Here's the process for doing that:
1306
Depending on your location throughout the world and the size of your
1307
repository though, it is often quicker to use an alternative public
1308
location to Launchpad, particularly if you can set up your own repo and
1309
push into that. By using an existing repo, push only needs to send the
1310
changes, instead of the complete repository every time. Note that it is
1311
easy to register branches in other locations with Launchpad so no benefits
1312
are lost by going this way.
1315
For Canonical staff, http://people.ubuntu.com/~<user>/ is one
1316
suggestion for public http branches. Contact your manager for information
1317
on accessing this system if required.
1319
It should also be noted that best practice in this area is subject to
1320
change as things evolve. For example, once the Bazaar smart server on
1321
Launchpad supports server-side branching, the performance situation will
1322
be very different to what it is now (Jun 2007).
1325
Configuring the PQM Plug-In
1326
---------------------------
1328
While not strictly required, the PQM plugin automates a few things and
1329
reduces the chance of error. Before looking at the plugin, it helps to
1330
understand a little more how PQM operates. Basically, PQM requires an
1331
email indicating what you want it to do. The email typically looks like
1334
star-merge source-branch target-branch
1338
star-merge http://bzr.arbash-meinel.com/branches/bzr/jam-integration http://bazaar-vcs.org/bzr/bzr.dev
1340
Note that the command needs to be on one line. The subject of the email
1341
will be used for the commit message. The email also needs to be ``gpg``
1342
signed with a key that PQM accepts.
1344
The advantages of using the PQM plugin are:
1346
#. You can use the config policies to make it easy to set up public
1347
branches, so you don't have to ever type the full paths you want to merge
1350
#. It checks to make sure the public branch last revision matches the
1351
local last revision so you are submitting what you think you are.
1353
#. It uses the same public_branch and smtp sending settings as bzr-email,
1354
so if you have one set up, you have the other mostly set up.
1356
#. Thunderbird refuses to not wrap lines, and request lines are usually
1357
pretty long (you have 2 long URLs in there).
1359
Here are sample configuration settings for the PQM plugin. Here are the
1360
lines in bazaar.conf::
1363
email = Joe Smith <joe.smith@internode.on.net>
1364
smtp_server=mail.internode.on.net:25
1366
And here are the lines in ``locations.conf`` (or ``branch.conf`` for
1367
dirstate-tags branches)::
1369
[/home/joe/bzr/my-integration]
1370
push_location = sftp://joe-smith@bazaar.launchpad.net/%7Ejoe-smith/bzr/my-integration/
1371
push_location:policy = norecurse
1372
public_branch = http://bazaar.launchpad.net/~joe-smith/bzr/my-integration/
1373
public_branch:policy = appendpath
1374
pqm_email = Bazaar PQM <pqm@bazaar-vcs.org>
1375
pqm_branch = http://bazaar-vcs.org/bzr/bzr.dev
1377
Note that the push settings will be added by the first ``push`` on
1378
a branch. Indeed the preferred way to generate the lines above is to use
1379
``push`` with an argument, then copy-and-paste the other lines into
1386
Here is one possible recipe once the above environment is set up:
1388
#. pull bzr.dev => my-integration
1389
#. merge patch => my-integration
1390
#. fix up any final merge conflicts (NEWS being the big killer here).
1396
The ``push`` step is not required if ``my-integration`` is a checkout of
1399
Because of defaults, you can type a single message into commit and
1400
pqm-commit will reuse that.
1403
Tracking Change Acceptance
1404
--------------------------
1406
The web interface to PQM is https://pqm.bazaar-vcs.org/. After submitting
1407
a change, you can visit this URL to confirm it was received and placed in
1410
When PQM completes processing a change, an email is sent to you with the
1414
Reviewing Blueprints
1415
====================
1417
Blueprint Tracking Using Launchpad
1418
----------------------------------
1420
New features typically require a fair amount of discussion, design and
1421
debate. For Bazaar, that information is often captured in a so-called
1422
"blueprint" on our Wiki. Overall tracking of blueprints and their status
1423
is done using Launchpad's relevant tracker,
1424
https://blueprints.launchpad.net/bzr/. Once a blueprint for ready for
1425
review, please announce it on the mailing list.
1427
Alternatively, send an email beginning with [RFC] with the proposal to the
1428
list. In some cases, you may wish to attach proposed code or a proposed
1429
developer document if that best communicates the idea. Debate can then
1430
proceed using the normal merge review processes.
1433
Recording Blueprint Review Feedback
1434
-----------------------------------
1436
Unlike its Bug Tracker, Launchpad's Blueprint Tracker doesn't currently
1437
(Jun 2007) support a chronological list of comment responses. Review
1438
feedback can either be recorded on the Wiki hosting the blueprints or by
1439
using Launchpad's whiteboard feature.
1446
Using Releases and Milestones in Launchpad
1447
------------------------------------------
1449
TODO ... (Exact policies still under discussion)
1455
Keeping on top of bugs reported is an important part of ongoing release
1456
planning. Everyone in the community is welcome and encouraged to raise
1457
bugs, confirm bugs raised by others, and nominate a priority. Practically
1458
though, a good percentage of bug triage is often done by the core
1459
developers, partially because of their depth of product knowledge.
1461
With respect to bug triage, core developers are encouraged to play an
1462
active role with particular attention to the following tasks:
1464
* keeping the number of unconfirmed bugs low
1465
* ensuring the priorities are generally right (everything as critical - or
1466
medium - is meaningless)
1467
* looking out for regressions and turning those around sooner rather than later.
1470
As well as prioritizing bugs and nominating them against a
1471
target milestone, Launchpad lets core developers offer to mentor others in
1476
vim: ft=rst tw=74 ai
added
removed
doc/developers/HACKING.txt
