104
86
--load-list. The later is rarely used but allows to run a subset of a list of
105
87
failing tests for example.
110
To test only the bzr core, ignoring any plugins you may have installed,
113
./bzr --no-plugins selftest
115
Disabling crash reporting
116
-------------------------
118
By default Bazaar uses apport_ to report program crashes. In developing
119
Bazaar it's normal and expected to have it crash from time to time, at
120
least because a test failed if for no other reason.
122
Therefore you should probably add ``debug_flags = no_apport`` to your
123
``bazaar.conf`` file (in ``~/.bazaar/`` on Unix), so that failures just
124
print a traceback rather than writing a crash file.
126
.. _apport: https://launchpad.net/apport/
129
90
Test suite debug flags
130
91
----------------------
136
97
This can provide useful logging to help debug test failures when used
137
98
with e.g. ``bzr -Dhpss selftest -E=allow_debug``
139
Note that this will probably cause some tests to fail, because they
140
don't expect to run with any debug flags on.
146
Bazaar can optionally produce output in the machine-readable subunit_
147
format, so that test output can be post-processed by various tools. To
148
generate a subunit test stream::
150
$ ./bzr selftest --subunit
152
Processing such a stream can be done using a variety of tools including:
154
* The builtin ``subunit2pyunit``, ``subunit-filter``, ``subunit-ls``,
155
``subunit2junitxml`` from the subunit project.
157
* tribunal_, a GUI for showing test results.
159
* testrepository_, a tool for gathering and managing test runs.
161
.. _subunit: https://launchpad.net/subunit/
162
.. _tribunal: https://launchpad.net/tribunal/
168
Bazaar ships with a config file for testrepository_. This can be very
169
useful for keeping track of failing tests and doing general workflow
170
support. To run tests using testrepository::
174
To run only failing tests::
176
$ testr run --failing
178
To run only some tests, without plugins::
180
$ test run test_selftest -- --no-plugins
182
See the testrepository documentation for more details.
184
.. _testrepository: https://launchpad.net/testrepository
187
Babune continuous integration
188
-----------------------------
190
We have a Hudson continuous-integration system that automatically runs
191
tests across various platforms. In the future we plan to add more
192
combinations including testing plugins. See
193
<http://babune.ladeuil.net:24842/>. (Babune = Bazaar Buildbot Network.)
196
Running tests in parallel
197
-------------------------
199
Bazaar can use subunit to spawn multiple test processes. There is
200
slightly more chance you will hit ordering or timing-dependent bugs but
203
$ ./bzr selftest --parallel=fork
205
Note that you will need the Subunit library
206
<https://launchpad.net/subunit/> to use this, which is in
207
``python-subunit`` on Ubuntu.
210
Running tests from a ramdisk
211
----------------------------
213
The tests create and delete a lot of temporary files. In some cases you
214
can make the test suite run much faster by running it on a ramdisk. For
218
$ sudo mount -t tmpfs none /ram
219
$ TMPDIR=/ram ./bzr selftest ...
221
You could also change ``/tmp`` in ``/etc/fstab`` to have type ``tmpfs``,
222
if you don't mind possibly losing other files in there when the machine
223
restarts. Add this line (if there is none for ``/tmp`` already)::
225
none /tmp tmpfs defaults 0 0
227
With a 6-core machine and ``--parallel=fork`` using a tmpfs doubles the
228
test execution speed.
234
Normally you should add or update a test for all bug fixes or new features
238
104
Where should I put a new test?
239
105
------------------------------
308
174
Per-implementation tests are tests that are defined once and then run
309
175
against multiple implementations of an interface. For example,
310
``per_transport.py`` defines tests that all Transport implementations
311
(local filesystem, HTTP, and so on) must pass. They are found in
312
``bzrlib/tests/per_*/*.py``, and ``bzrlib/tests/per_*.py``.
176
``test_transport_implementations.py`` defines tests that all Transport
177
implementations (local filesystem, HTTP, and so on) must pass.
179
They are found in ``bzrlib/tests/*_implementations/test_*.py``,
180
``bzrlib/tests/per_*/*.py``, and
181
``bzrlib/tests/test_*_implementations.py``.
314
183
These are really a sub-category of unit tests, but an important one.
316
Along the same lines are tests for extension modules. We generally have
317
both a pure-python and a compiled implementation for each module. As such,
318
we want to run the same tests against both implementations. These can
319
generally be found in ``bzrlib/tests/*__*.py`` since extension modules are
320
usually prefixed with an underscore. Since there are only two
321
implementations, we have a helper function
322
``bzrlib.tests.permute_for_extension``, which can simplify the
323
``load_tests`` implementation.
329
189
We make selective use of doctests__. In general they should provide
330
190
*examples* within the API documentation which can incidentally be tested. We
331
191
don't try to test every important case using doctests |--| regular Python
332
tests are generally a better solution. That is, we just use doctests to make
333
our documentation testable, rather than as a way to make tests. Be aware that
334
doctests are not as well isolated as the unit tests, if you need more
335
isolation, you're likely want to write unit tests anyway if only to get a
336
better control of the test environment.
192
tests are generally a better solution. That is, we just use doctests to
193
make our documentation testable, rather than as a way to make tests.
338
195
Most of these are in ``bzrlib/doc/api``. More additions are welcome.
340
197
__ http://docs.python.org/lib/module-doctest.html
342
There is an `assertDoctestExampleMatches` method in
343
`bzrlib.tests.TestCase` that allows you to match against doctest-style
344
string templates (including ``...`` to skip sections) from regular Python
351
``bzrlib/tests/script.py`` allows users to write tests in a syntax very
352
close to a shell session, using a restricted and limited set of commands
353
that should be enough to mimic most of the behaviours.
203
``bzrlib/tests/script.py`` allows users to write tests in a syntax very close to a shell session,
204
using a restricted and limited set of commands that should be enough to mimic
205
most of the behaviours.
355
207
A script is a set of commands, each command is composed of:
445
You can run files containing shell-like scripts with::
447
$ bzr test-script <script>
449
where ``<script>`` is the path to the file containing the shell-like script.
451
The actual use of ScriptRunner within a TestCase looks something like
454
from bzrlib.tests import script
456
def test_unshelve_keep(self):
458
script.run_script(self, '''
460
$ bzr shelve -q --all -m Foo
463
$ bzr unshelve -q --keep
470
You can also test commands that read user interaction::
472
def test_confirm_action(self):
473
"""You can write tests that demonstrate user confirmation"""
474
commands.builtin_command_registry.register(cmd_test_confirm)
475
self.addCleanup(commands.builtin_command_registry.remove, 'test-confirm')
478
2>Really do it? [y/n]:
483
To avoid having to specify "-q" for all commands whose output is
484
irrelevant, the run_script() method may be passed the keyword argument
485
``null_output_matches_anything=True``. For example::
487
def test_ignoring_null_output(self):
490
$ bzr ci -m 'first revision' --unchanged
493
""", null_output_matches_anything=True)
499
`bzrlib.tests.test_import_tariff` has some tests that measure how many
500
Python modules are loaded to run some representative commands.
502
We want to avoid loading code unnecessarily, for reasons including:
504
* Python modules are interpreted when they're loaded, either to define
505
classes or modules or perhaps to initialize some structures.
507
* With a cold cache we may incur blocking real disk IO for each module.
509
* Some modules depend on many others.
511
* Some optional modules such as `testtools` are meant to be soft
512
dependencies and only needed for particular cases. If they're loaded in
513
other cases then bzr may break for people who don't have those modules.
515
`test_import_tariff` allows us to check that removal of imports doesn't
518
This is done by running the command in a subprocess with
519
``PYTHON_VERBOSE=1``. Starting a whole Python interpreter is pretty slow,
520
so we don't want exhaustive testing here, but just enough to guard against
521
distinct fixed problems.
523
Assertions about precisely what is loaded tend to be brittle so we instead
524
make assertions that particular things aren't loaded.
526
Unless selftest is run with ``--no-plugins``, modules will be loaded in
527
the usual way and checks made on what they cause to be loaded. This is
528
probably worth checking into, because many bzr users have at least some
529
plugins installed (and they're included in binary installers).
531
In theory, plugins might have a good reason to load almost anything:
532
someone might write a plugin that opens a network connection or pops up a
533
gui window every time you run 'bzr status'. However, it's more likely
534
that the code to do these things is just being loaded accidentally. We
535
might eventually need to have a way to make exceptions for particular
538
Some things to check:
540
* non-GUI commands shouldn't load GUI libraries
542
* operations on bzr native formats sholudn't load foreign branch libraries
544
* network code shouldn't be loaded for purely local operations
546
* particularly expensive Python built-in modules shouldn't be loaded
547
unless there is a good reason
550
Testing locking behaviour
551
-------------------------
553
In order to test the locking behaviour of commands, it is possible to install
554
a hook that is called when a write lock is: acquired, released or broken.
555
(Read locks also exist, they cannot be discovered in this way.)
557
A hook can be installed by calling bzrlib.lock.Lock.hooks.install_named_hook.
558
The three valid hooks are: `lock_acquired`, `lock_released` and `lock_broken`.
565
lock.Lock.hooks.install_named_hook('lock_acquired',
566
locks_acquired.append, None)
567
lock.Lock.hooks.install_named_hook('lock_released',
568
locks_released.append, None)
570
`locks_acquired` will now receive a LockResult instance for all locks acquired
571
since the time the hook is installed.
573
The last part of the `lock_url` allows you to identify the type of object that is locked.
575
- BzrDir: `/branch-lock`
576
- Working tree: `/checkout/lock`
577
- Branch: `/branch/lock`
578
- Repository: `/repository/lock`
580
To test if a lock is a write lock on a working tree, one can do the following::
582
self.assertEndsWith(locks_acquired[0].lock_url, "/checkout/lock")
584
See bzrlib/tests/commands/test_revert.py for an example of how to use this for
720
416
``_probe`` and ``feature_name`` methods. For example::
722
418
class _SymlinkFeature(Feature):
724
420
def _probe(self):
725
421
return osutils.has_symlinks()
727
423
def feature_name(self):
728
424
return 'symlinks'
730
426
SymlinkFeature = _SymlinkFeature()
732
A helper for handling running tests based on whether a python
733
module is available. This can handle 3rd-party dependencies (is
734
``paramiko`` available?) as well as stdlib (``termios``) or
735
extension modules (``bzrlib._groupcompress_pyx``). You create a
736
new feature instance with::
738
# in bzrlib/tests/features.py
739
apport = tests.ModuleAvailableFeature('apport')
742
# then in bzrlib/tests/test_apport.py
743
class TestApportReporting(TestCaseInTempDir):
745
_test_needs_features = [features.apport]
748
Testing deprecated code
749
-----------------------
751
When code is deprecated, it is still supported for some length of time,
752
usually until the next major version. The ``applyDeprecated`` helper
753
wraps calls to deprecated code to verify that it is correctly issuing the
754
deprecation warning, and also prevents the warnings from being printed
757
Typically patches that apply the ``@deprecated_function`` decorator should
758
update the accompanying tests to use the ``applyDeprecated`` wrapper.
760
``applyDeprecated`` is defined in ``bzrlib.tests.TestCase``. See the API
761
docs for more details.
764
429
Testing exceptions and errors
765
430
-----------------------------
839
504
whether a test should be added for that particular implementation,
840
505
or for all implementations of the interface.
507
The multiplication of tests for different implementations is normally
508
accomplished by overriding the ``load_tests`` function used to load tests
509
from a module. This function typically loads all the tests, then applies
510
a TestProviderAdapter to them, which generates a longer suite containing
511
all the test variations.
842
513
See also `Per-implementation tests`_ (above).
845
Test scenarios and variations
846
-----------------------------
848
519
Some utilities are provided for generating variations of tests. This can
849
520
be used for per-implementation tests, or other cases where the same test
854
525
values to which the test should be applied. The test suite should then
855
526
also provide a list of scenarios in which to run the tests.
857
A single *scenario* is defined by a `(name, parameter_dict)` tuple. The
858
short string name is combined with the name of the test method to form the
859
test instance name. The parameter dict is merged into the instance's
864
load_tests = load_tests_apply_scenarios
866
class TestCheckout(TestCase):
868
scenarios = multiply_scenarios(
869
VaryByRepositoryFormat(),
873
The `load_tests` declaration or definition should be near the top of the
874
file so its effect can be seen.
528
Typically ``multiply_tests_from_modules`` should be called from the test
529
module's ``load_tests`` function.
892
547
A base TestCase that extends the Python standard library's
893
TestCase in several ways. TestCase is build on
894
``testtools.TestCase``, which gives it support for more assertion
895
methods (e.g. ``assertContainsRe``), ``addCleanup``, and other
896
features (see its API docs for details). It also has a ``setUp`` that
897
makes sure that global state like registered hooks and loggers won't
898
interfere with your test. All tests should use this base class
899
(whether directly or via a subclass). Note that we are trying not to
900
add more assertions at this point, and instead to build up a library
901
of ``bzrlib.tests.matchers``.
548
TestCase in several ways. It adds more assertion methods (e.g.
549
``assertContainsRe``), ``addCleanup``, and other features (see its API
550
docs for details). It also has a ``setUp`` that makes sure that
551
global state like registered hooks and loggers won't interfere with
552
your test. All tests should use this base class (whether directly or
903
555
TestCaseWithMemoryTransport
904
556
Extends TestCase and adds methods like ``get_transport``,
977
629
Please see bzrlib.treebuilder for more details.
982
PreviewTrees are based on TreeTransforms. This means they can represent
983
virtually any state that a WorkingTree can have, including unversioned files.
984
They can be used to test the output of anything that produces TreeTransforms,
985
such as merge algorithms and revert. They can also be used to test anything
986
that takes arbitrary Trees as its input.
990
# Get an empty tree to base the transform on.
991
b = self.make_branch('.')
992
empty_tree = b.repository.revision_tree(_mod_revision.NULL_REVISION)
993
tt = TransformPreview(empty_tree)
994
self.addCleanup(tt.finalize)
995
# Empty trees don't have a root, so add it first.
996
root = tt.new_directory('', ROOT_PARENT, 'tree-root')
997
# Set the contents of a file.
998
tt.new_file('new-file', root, 'contents', 'file-id')
999
preview = tt.get_preview_tree()
1000
# Test the contents.
1001
self.assertEqual('contents', preview.get_file_text('file-id'))
1003
PreviewTrees can stack, with each tree falling back to the previous::
1005
tt2 = TransformPreview(preview)
1006
self.addCleanup(tt2.finalize)
1007
tt2.new_file('new-file2', tt2.root, 'contents2', 'file-id2')
1008
preview2 = tt2.get_preview_tree()
1009
self.assertEqual('contents', preview2.get_file_text('file-id'))
1010
self.assertEqual('contents2', preview2.get_file_text('file-id2'))
1013
Temporarily changing state
1014
~~~~~~~~~~~~~~~~~~~~~~~~~~
1016
If your test needs to temporarily mutate some global state, and you need
1017
it restored at the end, you can say for example::
1019
self.overrideAttr(osutils, '_cached_user_encoding', 'latin-1')
1021
Temporarily changing environment variables
1022
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1024
If yout test needs to temporarily change some environment variable value
1025
(which generally means you want it restored at the end), you can use::
1027
self.overrideEnv('BZR_ENV_VAR', 'new_value')
1029
If you want to remove a variable from the environment, you should use the
1030
special ``None`` value::
1032
self.overrideEnv('PATH', None)
1034
If you add a new feature which depends on a new environment variable, make
1035
sure it behaves properly when this variable is not defined (if applicable) and
1036
if you need to enforce a specific default value, check the
1037
``TestCase._cleanEnvironment`` in ``bzrlib.tests.__init__.py`` which defines a
1038
proper set of values for all tests.
1043
Our base ``TestCase`` class provides an ``addCleanup`` method, which
1044
should be used instead of ``tearDown``. All the cleanups are run when the
1045
test finishes, regardless of whether it passes or fails. If one cleanup
1046
fails, later cleanups are still run.
1048
(The same facility is available outside of tests through
1049
``bzrlib.cleanup``.)
1055
Generally we prefer automated testing but sometimes a manual test is the
1056
right thing, especially for performance tests that want to measure elapsed
1057
time rather than effort.
1059
Simulating slow networks
1060
------------------------
1062
To get realistically slow network performance for manually measuring
1063
performance, we can simulate 500ms latency (thus 1000ms round trips)::
1065
$ sudo tc qdisc add dev lo root netem delay 500ms
1067
Normal system behaviour is restored with ::
1069
$ sudo tc qdisc del dev lo root
1071
A more precise version that only filters traffic to port 4155 is::
1073
tc qdisc add dev lo root handle 1: prio
1074
tc qdisc add dev lo parent 1:3 handle 30: netem delay 500ms
1075
tc qdisc add dev lo parent 30:1 handle 40: prio
1076
tc filter add dev lo protocol ip parent 1:0 prio 3 u32 match ip dport 4155 0xffff flowid 1:3 handle 800::800
1077
tc filter add dev lo protocol ip parent 1:0 prio 3 u32 match ip sport 4155 0xffff flowid 1:3 handle 800::801
1079
and to remove this::
1081
tc filter del dev lo protocol ip parent 1: pref 3 u32
1082
tc qdisc del dev lo root handle 1:
1084
You can use similar code to add additional delay to a real network
1085
interface, perhaps only when talking to a particular server or pointing at
1086
a VM. For more information see <http://lartc.org/>.
1089
632
.. |--| unicode:: U+2014
1092
vim: ft=rst tw=74 ai et sw=4
added
removed
doc/developers/testing.txt
