291
261
<http://starship.python.net/crew/mwh/bzrlibapi/>.
293
263
See also the `Bazaar Architectural Overview
294
<http://doc.bazaar.canonical.com/developers/overview.html>`_.
264
<http://doc.bazaar-vcs.org/developers/overview.html>`_.
267
The Code Review Process
268
#######################
270
All code changes coming in to Bazaar are reviewed by someone else.
271
Normally changes by core contributors are reviewed by one other core
272
developer, and changes from other people are reviewed by two core
273
developers. Use intelligent discretion if the patch is trivial.
275
Good reviews do take time. They also regularly require a solid
276
understanding of the overall code base. In practice, this means a small
277
number of people often have a large review burden - with knowledge comes
278
responsibility. No one likes their merge requests sitting in a queue going
279
nowhere, so reviewing sooner rather than later is strongly encouraged.
288
Please put a "cover letter" on your merge request explaining:
290
* the reason **why** you're making this change
292
* **how** this change achieves this purpose
294
* anything else you may have fixed in passing
296
* anything significant that you thought of doing, such as a more
297
extensive fix or a different approach, but didn't or couldn't do now
299
A good cover letter makes reviewers' lives easier because they can decide
300
from the letter whether they agree with the purpose and approach, and then
301
assess whether the patch actually does what the cover letter says.
302
Explaining any "drive-by fixes" or roads not taken may also avoid queries
303
from the reviewer. All in all this should give faster and better reviews.
304
Sometimes writing the cover letter helps the submitter realize something
305
else they need to do. The size of the cover letter should be proportional
306
to the size and complexity of the patch.
309
Reviewing proposed changes
310
==========================
312
Anyone is welcome to review code, and reply to the thread with their
315
The simplest way to review a proposed change is to just read the patch on
316
the list or in Bundle Buggy. For more complex changes it may be useful
317
to make a new working tree or branch from trunk, and merge the proposed
318
change into it, so you can experiment with the code or look at a wider
321
There are three main requirements for code to get in:
323
* Doesn't reduce test coverage: if it adds new methods or commands,
324
there should be tests for them. There is a good test framework
325
and plenty of examples to crib from, but if you are having trouble
326
working out how to test something feel free to post a draft patch
329
* Doesn't reduce design clarity, such as by entangling objects
330
we're trying to separate. This is mostly something the more
331
experienced reviewers need to help check.
333
* Improves bugs, features, speed, or code simplicity.
335
Code that goes in should not degrade any of these aspects. Patches are
336
welcome that only cleanup the code without changing the external
337
behaviour. The core developers take care to keep the code quality high
338
and understandable while recognising that perfect is sometimes the enemy
341
It is easy for reviews to make people notice other things which should be
342
fixed but those things should not hold up the original fix being accepted.
343
New things can easily be recorded in the Bug Tracker instead.
345
It's normally much easier to review several smaller patches than one large
346
one. You might want to use ``bzr-loom`` to maintain threads of related
347
work, or submit a preparatory patch that will make your "real" change
351
Checklist for reviewers
352
=======================
354
* Do you understand what the code's doing and why?
356
* Will it perform reasonably for large inputs, both in memory size and
357
run time? Are there some scenarios where performance should be
360
* Is it tested, and are the tests at the right level? Are there both
361
blackbox (command-line level) and API-oriented tests?
363
* If this change will be visible to end users or API users, is it
364
appropriately documented in NEWS?
366
* Does it meet the coding standards below?
368
* If it changes the user-visible behaviour, does it update the help
369
strings and user documentation?
371
* If it adds a new major concept or standard practice, does it update the
372
developer documentation?
374
* (your ideas here...)
380
From May 2009 on, we prefer people to propose code reviews through
383
* <https://launchpad.net/+tour/code-review>
385
* <https://help.launchpad.net/Code/Review>
387
Anyone can propose or comment on a merge proposal just by creating a
390
There are two ways to create a new merge proposal: through the web
391
interface or by email.
394
Proposing a merge through the web
395
---------------------------------
397
To create the proposal through the web, first push your branch to Launchpad.
398
For example, a branch dealing with documentation belonging to the Launchpad
399
User mbp could be pushed as ::
401
bzr push lp:~mbp/bzr/doc
403
Then go to the branch's web page, which in this case would be
404
<https://code.launchpad.net/~mbp/bzr/doc>. You can simplify this step by just
409
You can then click "Propose for merging into another branch", and enter your
410
cover letter (see above) into the web form. Typically you'll want to merge
411
into ``~bzr/bzr/trunk`` which will be the default; you might also want to
412
nominate merging into a release branch for a bug fix. There is the option to
413
specify a specific reviewer or type of review, and you shouldn't normally
416
Submitting the form takes you to the new page about the merge proposal
417
containing the diff of the changes, comments by interested people, and
418
controls to comment or vote on the change.
420
Proposing a merge by mail
421
-------------------------
423
To propose a merge by mail, send a bundle to ``merge@code.launchpad.net``.
425
You can generate a merge request like this::
427
bzr send -o bug-1234.diff
429
``bzr send`` can also send mail directly if you prefer; see the help.
434
From <https://code.launchpad.net/bzr/+activereviews> you can see all
435
currently active reviews, and choose one to comment on. This page also
436
shows proposals that are now approved and should be merged by someone with
440
Reviews through Bundle Buggy
441
============================
443
The Bundle Buggy tool used up to May 2009 is still available as a review
446
Sending patches for review
447
--------------------------
449
If you'd like to propose a change, please post to the
450
bazaar@lists.canonical.com list with a bundle, patch, or link to a
451
branch. Put ``[PATCH]`` or ``[MERGE]`` in the subject so Bundle Buggy
452
can pick it out, and explain the change in the email message text.
453
Remember to update the NEWS file as part of your change if it makes any
454
changes visible to users or plugin developers. Please include a diff
455
against mainline if you're giving a link to a branch.
457
You can generate a merge request like this::
459
bzr send -o bug-1234.patch
461
A ``.patch`` extension is recommended instead of .bundle as many mail clients
462
will send the latter as a binary file.
464
``bzr send`` can also send mail directly if you prefer; see the help.
466
Please do **NOT** put [PATCH] or [MERGE] in the subject line if you don't
467
want it to be merged. If you want comments from developers rather than
468
to be merged, you can put ``[RFC]`` in the subject line.
470
If this change addresses a bug, please put the bug number in the subject
471
line too, in the form ``[#1]`` so that Bundle Buggy can recognize it.
473
If the change is intended for a particular release mark that in the
474
subject too, e.g. ``[1.6]``.
475
Anyone can "vote" on the mailing list by expressing an opinion. Core
476
developers can also vote using Bundle Buggy. Here are the voting codes and
479
:approve: Reviewer wants this submission merged.
480
:tweak: Reviewer wants this submission merged with small changes. (No
482
:abstain: Reviewer does not intend to vote on this patch.
483
:resubmit: Please make changes and resubmit for review.
484
:reject: Reviewer doesn't want this kind of change merged.
485
:comment: Not really a vote. Reviewer just wants to comment, for now.
487
If a change gets two approvals from core reviewers, and no rejections,
488
then it's OK to come in. Any of the core developers can bring it into the
489
bzr.dev trunk and backport it to maintenance branches if required. The
490
Release Manager will merge the change into the branch for a pending
491
release, if any. As a guideline, core developers usually merge their own
492
changes and volunteer to merge other contributions if they were the second
493
reviewer to agree to a change.
495
To track the progress of proposed changes, use Bundle Buggy. See
496
http://bundlebuggy.aaronbentley.com/help for a link to all the
497
outstanding merge requests together with an explanation of the columns.
498
Bundle Buggy will also mail you a link to track just your change.
500
Coding Style Guidelines
501
#######################
506
``hasattr`` should not be used because it swallows exceptions including
507
``KeyboardInterrupt``. Instead, say something like ::
509
if getattr(thing, 'name', None) is None
515
Please write PEP-8__ compliant code.
517
__ http://www.python.org/peps/pep-0008.html
519
One often-missed requirement is that the first line of docstrings
520
should be a self-contained one-sentence summary.
522
We use 4 space indents for blocks, and never use tab characters. (In vim,
525
Trailing white space should be avoided, but is allowed.
526
You should however not make lots of unrelated white space changes.
528
Unix style newlines (LF) are used.
530
Each file must have a newline at the end of it.
532
Lines should be no more than 79 characters if at all possible.
533
Lines that continue a long statement may be indented in either of
536
within the parenthesis or other character that opens the block, e.g.::
542
or indented by four spaces::
548
The first is considered clearer by some people; however it can be a bit
549
harder to maintain (e.g. when the method name changes), and it does not
550
work well if the relevant parenthesis is already far to the right. Avoid
553
self.legbone.kneebone.shinbone.toebone.shake_it(one,
559
self.legbone.kneebone.shinbone.toebone.shake_it(one,
565
self.legbone.kneebone.shinbone.toebone.shake_it(
568
For long lists, we like to add a trailing comma and put the closing
569
character on the following line. This makes it easier to add new items in
572
from bzrlib.goo import (
578
There should be spaces between function parameters, but not between the
579
keyword name and the value::
581
call(1, 3, cheese=quark)
585
;(defface my-invalid-face
586
; '((t (:background "Red" :underline t)))
587
; "Face used to highlight invalid constructs or other uglyties"
590
(defun my-python-mode-hook ()
591
;; setup preferred indentation style.
592
(setq fill-column 79)
593
(setq indent-tabs-mode nil) ; no tabs, never, I will not repeat
594
; (font-lock-add-keywords 'python-mode
595
; '(("^\\s *\t" . 'my-invalid-face) ; Leading tabs
596
; ("[ \t]+$" . 'my-invalid-face) ; Trailing spaces
597
; ("^[ \t]+$" . 'my-invalid-face)); Spaces only
601
(add-hook 'python-mode-hook 'my-python-mode-hook)
603
The lines beginning with ';' are comments. They can be activated
604
if one want to have a strong notice of some tab/space usage
611
* Imports should be done at the top-level of the file, unless there is
612
a strong reason to have them lazily loaded when a particular
613
function runs. Import statements have a cost, so try to make sure
614
they don't run inside hot functions.
616
* Module names should always be given fully-qualified,
617
i.e. ``bzrlib.hashcache`` not just ``hashcache``.
623
Functions, methods or members that are relatively private are given
624
a leading underscore prefix. Names without a leading underscore are
625
public not just across modules but to programmers using bzrlib as an
628
We prefer class names to be concatenated capital words (``TestCase``)
629
and variables, methods and functions to be lowercase words joined by
630
underscores (``revision_id``, ``get_revision``).
632
For the purposes of naming some names are treated as single compound
633
words: "filename", "revno".
635
Consider naming classes as nouns and functions/methods as verbs.
637
Try to avoid using abbreviations in names, because there can be
638
inconsistency if other people use the full name.
644
``revision_id`` not ``rev_id`` or ``revid``
646
Functions that transform one thing to another should be named ``x_to_y``
647
(not ``x2y`` as occurs in some old code.)
653
Python destructors (``__del__``) work differently to those of other
654
languages. In particular, bear in mind that destructors may be called
655
immediately when the object apparently becomes unreferenced, or at some
656
later time, or possibly never at all. Therefore we have restrictions on
657
what can be done inside them.
659
0. If you think you need to use a ``__del__`` method ask another
660
developer for alternatives. If you do need to use one, explain
663
1. Never rely on a ``__del__`` method running. If there is code that
664
must run, do it from a ``finally`` block instead.
666
2. Never ``import`` from inside a ``__del__`` method, or you may crash the
669
3. In some places we raise a warning from the destructor if the object
670
has not been cleaned up or closed. This is considered OK: the warning
671
may not catch every case but it's still useful sometimes.
677
Often when something has failed later code, including cleanups invoked
678
from ``finally`` blocks, will fail too. These secondary failures are
679
generally uninteresting compared to the original exception. So use the
680
``only_raises`` decorator (from ``bzrlib.decorators``) for methods that
681
are typically called in ``finally`` blocks, such as ``unlock`` methods.
682
For example, ``@only_raises(LockNotHeld, LockBroken)``. All errors that
683
are unlikely to be a knock-on failure from a previous failure should be
690
In some places we have variables which point to callables that construct
691
new instances. That is to say, they can be used a lot like class objects,
692
but they shouldn't be *named* like classes:
694
> I think that things named FooBar should create instances of FooBar when
695
> called. Its plain confusing for them to do otherwise. When we have
696
> something that is going to be used as a class - that is, checked for via
697
> isinstance or other such idioms, them I would call it foo_class, so that
698
> it is clear that a callable is not sufficient. If it is only used as a
699
> factory, then yes, foo_factory is what I would use.
705
Several places in Bazaar use (or will use) a registry, which is a
706
mapping from names to objects or classes. The registry allows for
707
loading in registered code only when it's needed, and keeping
708
associated information such as a help string or description.
711
InterObject and multiple dispatch
712
=================================
714
The ``InterObject`` provides for two-way `multiple dispatch`__: matching
715
up for example a source and destination repository to find the right way
716
to transfer data between them.
718
.. __: http://en.wikipedia.org/wiki/Multiple_dispatch
720
There is a subclass ``InterObject`` classes for each type of object that is
721
dispatched this way, e.g. ``InterRepository``. Calling ``.get()`` on this
722
class will return an ``InterObject`` instance providing the best match for
723
those parameters, and this instance then has methods for operations
726
inter = InterRepository.get(source_repo, target_repo)
727
inter.fetch(revision_id)
729
``InterRepository`` also acts as a registry-like object for its
730
subclasses, and they can be added through ``.register_optimizer``. The
731
right one to run is selected by asking each class, in reverse order of
732
registration, whether it ``.is_compatible`` with the relevant objects.
737
To make startup time faster, we use the ``bzrlib.lazy_import`` module to
738
delay importing modules until they are actually used. ``lazy_import`` uses
739
the same syntax as regular python imports. So to import a few modules in a
742
from bzrlib.lazy_import import lazy_import
743
lazy_import(globals(), """
752
revision as _mod_revision,
754
import bzrlib.transport
758
At this point, all of these exist as a ``ImportReplacer`` object, ready to
759
be imported once a member is accessed. Also, when importing a module into
760
the local namespace, which is likely to clash with variable names, it is
761
recommended to prefix it as ``_mod_<module>``. This makes it clearer that
762
the variable is a module, and these object should be hidden anyway, since
763
they shouldn't be imported into other namespaces.
765
While it is possible for ``lazy_import()`` to import members of a module
766
when using the ``from module import member`` syntax, it is recommended to
767
only use that syntax to load sub modules ``from module import submodule``.
768
This is because variables and classes can frequently be used without
769
needing a sub-member for example::
771
lazy_import(globals(), """
772
from module import MyClass
776
return isinstance(x, MyClass)
778
This will incorrectly fail, because ``MyClass`` is a ``ImportReplacer``
779
object, rather than the real class.
781
It also is incorrect to assign ``ImportReplacer`` objects to other variables.
782
Because the replacer only knows about the original name, it is unable to
783
replace other variables. The ``ImportReplacer`` class will raise an
784
``IllegalUseOfScopeReplacer`` exception if it can figure out that this
785
happened. But it requires accessing a member more than once from the new
786
variable, so some bugs are not detected right away.
792
The null revision is the ancestor of all revisions. Its revno is 0, its
793
revision-id is ``null:``, and its tree is the empty tree. When referring
794
to the null revision, please use ``bzrlib.revision.NULL_REVISION``. Old
795
code sometimes uses ``None`` for the null revision, but this practice is
799
Object string representations
800
=============================
802
Python prints objects using their ``__repr__`` method when they are
803
written to logs, exception tracebacks, or the debugger. We want
804
objects to have useful representations to help in determining what went
807
If you add a new class you should generally add a ``__repr__`` method
808
unless there is an adequate method in a parent class. There should be a
811
Representations should typically look like Python constructor syntax, but
812
they don't need to include every value in the object and they don't need
813
to be able to actually execute. They're to be read by humans, not
814
machines. Don't hardcode the classname in the format, so that we get the
815
correct value if the method is inherited by a subclass. If you're
816
printing attributes of the object, including strings, you should normally
817
use ``%r`` syntax (to call their repr in turn).
819
Try to avoid the representation becoming more than one or two lines long.
820
(But balance this against including useful information, and simplicity of
823
Because repr methods are often called when something has already gone
824
wrong, they should be written somewhat more defensively than most code.
825
The object may be half-initialized or in some other way in an illegal
826
state. The repr method shouldn't raise an exception, or it may hide the
827
(probably more useful) underlying exception.
832
return '%s(%r)' % (self.__class__.__name__,
839
A bare ``except`` statement will catch all exceptions, including ones that
840
really should terminate the program such as ``MemoryError`` and
841
``KeyboardInterrupt``. They should rarely be used unless the exception is
842
later re-raised. Even then, think about whether catching just
843
``Exception`` (which excludes system errors in Python2.5 and later) would
850
All code should be exercised by the test suite. See the `Bazaar Testing
851
Guide <http://doc.bazaar-vcs.org/developers/testing.html>`_ for detailed
852
information about writing tests.
885
1456
how to set it up and configure it.
1465
Of the many workflows supported by Bazaar, the one adopted for Bazaar
1466
development itself is known as "Decentralized with automatic gatekeeper".
1467
To repeat the explanation of this given on
1468
http://bazaar-vcs.org/Workflows:
1471
In this workflow, each developer has their own branch or
1472
branches, plus read-only access to the mainline. A software gatekeeper
1473
(e.g. PQM) has commit rights to the main branch. When a developer wants
1474
their work merged, they request the gatekeeper to merge it. The gatekeeper
1475
does a merge, a compile, and runs the test suite. If the code passes, it
1476
is merged into the mainline.
1478
In a nutshell, here's the overall submission process:
1480
#. get your work ready (including review except for trivial changes)
1481
#. push to a public location
1482
#. ask PQM to merge from that location
1485
At present, PQM always takes the changes to merge from a branch
1486
at a URL that can be read by it. For Bazaar, that means a public,
1487
typically http, URL.
1489
As a result, the following things are needed to use PQM for submissions:
1491
#. A publicly available web server
1492
#. Your OpenPGP key registered with PQM (contact RobertCollins for this)
1493
#. The PQM plugin installed and configured (not strictly required but
1494
highly recommended).
1497
Selecting a Public Branch Location
1498
----------------------------------
1500
If you don't have your own web server running, branches can always be
1501
pushed to Launchpad. Here's the process for doing that:
1503
Depending on your location throughout the world and the size of your
1504
repository though, it is often quicker to use an alternative public
1505
location to Launchpad, particularly if you can set up your own repo and
1506
push into that. By using an existing repo, push only needs to send the
1507
changes, instead of the complete repository every time. Note that it is
1508
easy to register branches in other locations with Launchpad so no benefits
1509
are lost by going this way.
1512
For Canonical staff, http://people.ubuntu.com/~<user>/ is one
1513
suggestion for public http branches. Contact your manager for information
1514
on accessing this system if required.
1516
It should also be noted that best practice in this area is subject to
1517
change as things evolve. For example, once the Bazaar smart server on
1518
Launchpad supports server-side branching, the performance situation will
1519
be very different to what it is now (Jun 2007).
1522
Configuring the PQM Plug-In
1523
---------------------------
1525
While not strictly required, the PQM plugin automates a few things and
1526
reduces the chance of error. Before looking at the plugin, it helps to
1527
understand a little more how PQM operates. Basically, PQM requires an
1528
email indicating what you want it to do. The email typically looks like
1531
star-merge source-branch target-branch
1535
star-merge http://bzr.arbash-meinel.com/branches/bzr/jam-integration http://bazaar-vcs.org/bzr/bzr.dev
1537
Note that the command needs to be on one line. The subject of the email
1538
will be used for the commit message. The email also needs to be ``gpg``
1539
signed with a key that PQM accepts.
1541
The advantages of using the PQM plugin are:
1543
#. You can use the config policies to make it easy to set up public
1544
branches, so you don't have to ever type the full paths you want to merge
1547
#. It checks to make sure the public branch last revision matches the
1548
local last revision so you are submitting what you think you are.
1550
#. It uses the same public_branch and smtp sending settings as bzr-email,
1551
so if you have one set up, you have the other mostly set up.
1553
#. Thunderbird refuses to not wrap lines, and request lines are usually
1554
pretty long (you have 2 long URLs in there).
1556
Here are sample configuration settings for the PQM plugin. Here are the
1557
lines in bazaar.conf::
1560
email = Joe Smith <joe.smith@internode.on.net>
1561
smtp_server=mail.internode.on.net:25
1563
And here are the lines in ``locations.conf`` (or ``branch.conf`` for
1564
dirstate-tags branches)::
1566
[/home/joe/bzr/my-integration]
1567
push_location = sftp://joe-smith@bazaar.launchpad.net/%7Ejoe-smith/bzr/my-integration/
1568
push_location:policy = norecurse
1569
public_branch = http://bazaar.launchpad.net/~joe-smith/bzr/my-integration/
1570
public_branch:policy = appendpath
1571
pqm_email = Bazaar PQM <pqm@bazaar-vcs.org>
1572
pqm_branch = http://bazaar-vcs.org/bzr/bzr.dev
1574
Note that the push settings will be added by the first ``push`` on
1575
a branch. Indeed the preferred way to generate the lines above is to use
1576
``push`` with an argument, then copy-and-paste the other lines into
1583
Here is one possible recipe once the above environment is set up:
1585
#. pull bzr.dev => my-integration
1586
#. merge patch => my-integration
1587
#. fix up any final merge conflicts (NEWS being the big killer here).
1593
The ``push`` step is not required if ``my-integration`` is a checkout of
1596
Because of defaults, you can type a single message into commit and
1597
pqm-commit will reuse that.
1600
Tracking Change Acceptance
1601
--------------------------
1603
The web interface to PQM is https://pqm.bazaar-vcs.org/. After submitting
1604
a change, you can visit this URL to confirm it was received and placed in
1607
When PQM completes processing a change, an email is sent to you with the
1611
Reviewing Blueprints
1612
====================
1614
Blueprint Tracking Using Launchpad
1615
----------------------------------
1617
New features typically require a fair amount of discussion, design and
1618
debate. For Bazaar, that information is often captured in a so-called
1619
"blueprint" on our Wiki. Overall tracking of blueprints and their status
1620
is done using Launchpad's relevant tracker,
1621
https://blueprints.launchpad.net/bzr/. Once a blueprint for ready for
1622
review, please announce it on the mailing list.
1624
Alternatively, send an email beginning with [RFC] with the proposal to the
1625
list. In some cases, you may wish to attach proposed code or a proposed
1626
developer document if that best communicates the idea. Debate can then
1627
proceed using the normal merge review processes.
1630
Recording Blueprint Review Feedback
1631
-----------------------------------
1633
Unlike its Bug Tracker, Launchpad's Blueprint Tracker doesn't currently
1634
(Jun 2007) support a chronological list of comment responses. Review
1635
feedback can either be recorded on the Wiki hosting the blueprints or by
1636
using Launchpad's whiteboard feature.
889
1639
Planning Releases
890
1640
=================
1643
Using Releases and Milestones in Launchpad
1644
------------------------------------------
1646
TODO ... (Exact policies still under discussion)
added
removed
doc/developers/HACKING.txt
