bzr selftest

Running particular tests

Currently, bzr selftest is used to invoke tests. You can provide a pattern argument to run a subset. For example, to run just the blackbox tests, run:

./bzr selftest -v blackbox

To skip a particular test (or set of tests), use the –exclude option (shorthand -x) like so:

./bzr selftest -v -x blackbox

To ensure that all tests are being run and succeeding, you can use the –strict option which will fail if there are any missing features or known failures, like so:

./bzr selftest --strict

To list tests without running them, use the –list-only option like so:

./bzr selftest --list-only

This option can be combined with other selftest options (like -x) and filter patterns to understand their effect.

Once you understand how to create a list of tests, you can use the –load-list option to run only a restricted set of tests that you kept in a file, one test id by line. Keep in mind that this will never be sufficient to validate your modifications, you still need to run the full test suite for that, but using it can help in some cases (like running only the failed tests for some time):

./bzr selftest -- load-list my_failing_tests

This option can also be combined with other selftest options, including patterns. It has some drawbacks though, the list can become out of date pretty quick when doing Test Driven Development.

To address this concern, there is another way to run a restricted set of tests: the –starting-with option will run only the tests whose name starts with the specified string. It will also avoid loading the other tests and as a consequence starts running your tests quicker:

./bzr selftest --starting-with bzrlib.blackbox

This option can be combined with all the other selftest options including –load-list. The later is rarely used but allows to run a subset of a list of failing tests for example.

Disabling plugins

To test only the bzr core, ignoring any plugins you may have installed, use:

./bzr --no-plugins selftest

Disabling crash reporting

By default Bazaar uses apport to report program crashes. In developing Bazaar it’s normal and expected to have it crash from time to time, at least because a test failed if for no other reason.

Therefore you should probably add debug_flags = no_apport to your bazaar.conf file (in ~/.bazaar/ on Unix), so that failures just print a traceback rather than writing a crash file.

Test suite debug flags

Similar to the global -Dfoo debug options, bzr selftest accepts -E=foo debug flags. These flags are:

Using subunit

Bazaar can optionally produce output in the machine-readable subunit format, so that test output can be post-processed by various tools. To generate a subunit test stream:

$ ./bzr selftest --subunit

Processing such a stream can be done using a variety of tools including:

• The builtin subunit2pyunit, subunit-filter, subunit-ls, subunit2junitxml from the subunit project.
• tribunal, a GUI for showing test results.
• testrepository, a tool for gathering and managing test runs.

Using testrepository

Bazaar ships with a config file for testrepository. This can be very useful for keeping track of failing tests and doing general workflow support. To run tests using testrepository:

$ testr run

To run only failing tests:

$ testr run --failing

To run only some tests, without plugins:

$ test run test_selftest -- --no-plugins

See the testrepository documentation for more details.

Where should I put a new test?

Bzrlib’s tests are organised by the type of test. Most of the tests in bzr’s test suite belong to one of these categories:

• Unit tests
• Blackbox (UI) tests
• Per-implementation tests
• Doctests

A quick description of these test types and where they belong in bzrlib’s source follows. Not all tests fall neatly into one of these categories; in those cases use your judgement.

Shell-like tests

bzrlib/tests/script.py allows users to write tests in a syntax very close to a shell session, using a restricted and limited set of commands that should be enough to mimic most of the behaviours.

A script is a set of commands, each command is composed of:

• one mandatory command line,
• one optional set of input lines to feed the command,
• one optional set of output expected lines,
• one optional set of error expected lines.

Input, output and error lines can be specified in any order.

Except for the expected output, all lines start with a special string (based on their origin when used under a Unix shell):

• ‘$ ‘ for the command,
• ‘<’ for input,
• nothing for output,
• ‘2>’ for errors,

Comments can be added anywhere, they start with ‘#’ and end with the line.

The execution stops as soon as an expected output or an expected error is not matched.

When no output is specified, any ouput from the command is accepted and execution continue.

If an error occurs and no expected error is specified, the execution stops.

An error is defined by a returned status different from zero, not by the presence of text on the error stream.

The matching is done on a full string comparison basis unless ‘...’ is used, in which case expected output/errors can be less precise.

Examples:

The following will succeeds only if ‘bzr add’ outputs ‘adding file’:

$ bzr add file
>adding file

If you want the command to succeed for any output, just use:

$ bzr add file

The following will stop with an error:

$ bzr not-a-command

If you want it to succeed, use:

$ bzr not-a-command
2> bzr: ERROR: unknown command "not-a-command"

You can use ellipsis (...) to replace any piece of text you don’t want to be matched exactly:

$ bzr branch not-a-branch
2>bzr: ERROR: Not a branch...not-a-branch/".

This can be used to ignore entire lines too:

$ cat
<first line
<second line
<third line
# And here we explain that surprising fourth line
<fourth line
<last line
>first line
>...
>last line

You can check the content of a file with cat:

$ cat <file
>expected content

You can also check the existence of a file with cat, the following will fail if the file doesn’t exist:

$ cat file

The actual use of ScriptRunner within a TestCase looks something like this:

def test_unshelve_keep(self):
        # some setup here
        sr = ScriptRunner()
        sr.run_script(self, '''
$ bzr add file
$ bzr shelve --all -m Foo
$ bzr shelve --list
1: Foo
$ bzr unshelve --keep
$ bzr shelve --list
1: Foo
$ cat file
contents of file
''')

Import tariff tests

bzrlib.tests.test_import_tariff has some tests that measure how many Python modules are loaded to run some representative commands.

We want to avoid loading code unnecessarily, for reasons including:

• Python modules are interpreted when they’re loaded, either to define classes or modules or perhaps to initialize some structures.
• With a cold cache we may incur blocking real disk IO for each module.
• Some modules depend on many others.
• Some optional modules such as testtools are meant to be soft dependencies and only needed for particular cases. If they’re loaded in other cases then bzr may break for people who don’t have those modules.

test_import_tarrif allows us to check that removal of imports doesn’t regress.

This is done by running the command in a subprocess with --profile-imports. Starting a whole Python interpreter is pretty slow, so we don’t want exhaustive testing here, but just enough to guard against distinct fixed problems.

Assertions about precisely what is loaded tend to be brittle so we instead make assertions that particular things aren’t loaded.

Unless selftest is run with --no-plugins, modules will be loaded in the usual way and checks made on what they cause to be loaded. This is probably worth checking into, because many bzr users have at least some plugins installed (and they’re included in binary installers).

In theory, plugins might have a good reason to load almost anything: someone might write a plugin that opens a network connection or pops up a gui window every time you run ‘bzr status’. However, it’s more likely that the code to do these things is just being loaded accidentally. We might eventually need to have a way to make exceptions for particular plugins.

Some things to check:

• non-GUI commands shouldn’t load GUI libraries
• operations on bzr native formats sholudn’t load foreign branch libraries
• network code shouldn’t be loaded for purely local operations
• particularly expensive Python built-in modules shouldn’t be loaded unless there is a good reason

Testing locking behaviour

You may want to write tests that particular objects are or aren’t locked during particular operations: see for example bug 498409.

The TestCase base class registers hooks that record lock actions into ._lock_actions in this format:

[
  ('acquired', LockResult(file:///tmp/testbzr-J2pcy2.tmp/.bzr/branch-lockc4au55ppz8wdym11z1aq)),
  ('released', LockResult(file:///tmp/testbzr-J2pcy2.tmp/.bzr/branch-lockc4au55ppz8wdym11z1aq)),
  ('acquired', LockResult(file:///tmp/testbzr-J2pcy2.tmp/.bzr/repository/lockyxb3rn4sw1oyx1jzkt45)),
  ('released', LockResult(file:///tmp/testbzr-J2pcy2.tmp/.bzr/repository/lockyxb3rn4sw1oyx1jzkt45)),
  ('acquired', LockResult(file:///tmp/testbzr-J2pcy2.tmp/.bzr/branch/lockh8c6t28rcjdkgxtndbje)),
  ('released', LockResult(file:///tmp/testbzr-J2pcy2.tmp/.bzr/branch/lockh8c6t28rcjdkgxtndbje)),
  ...

Alternatively you can register your own hooks to make custom assertions: see TestCase._check_locks for an example.

Skipping tests

In our enhancements to unittest we allow for some addition results beyond just success or failure.

If a test can’t be run, it can say that it’s skipped by raising a special exception. This is typically used in parameterized tests — for example if a transport doesn’t support setting permissions, we’ll skip the tests that relating to that.

try:
    return self.branch_format.initialize(repo.bzrdir)
except errors.UninitializableFormat:
    raise tests.TestSkipped('Uninitializable branch format')

Raising TestSkipped is a good idea when you want to make it clear that the test was not run, rather than just returning which makes it look as if it was run and passed.

Several different cases are distinguished:

TestSkipped

Generic skip; the only type that was present up to bzr 0.18.

TestNotApplicable

The test doesn’t apply to the parameters with which it was run. This is typically used when the test is being applied to all implementations of an interface, but some aspects of the interface are optional and not present in particular concrete implementations. (Some tests that should raise this currently either silently return or raise TestSkipped.) Another option is to use more precise parameterization to avoid generating the test at all.

UnavailableFeature

The test can’t be run because a dependency (typically a Python library) is not available in the test environment. These are in general things that the person running the test could fix by installing the library. It’s OK if some of these occur when an end user runs the tests or if we’re specifically testing in a limited environment, but a full test should never see them.

See Test feature dependencies below.

KnownFailure

The test exists but is known to fail, for example this might be appropriate to raise if you’ve committed a test for a bug but not the fix for it, or if something works on Unix but not on Windows.

Raising this allows you to distinguish these failures from the ones that are not expected to fail. If the test would fail because of something we don’t expect or intend to fix, KnownFailure is not appropriate, and TestNotApplicable might be better.

KnownFailure should be used with care as we don’t want a proliferation of quietly broken tests.

We plan to support three modes for running the test suite to control the interpretation of these results. Strict mode is for use in situations like merges to the mainline and releases where we want to make sure that everything that can be tested has been tested. Lax mode is for use by developers who want to temporarily tolerate some known failures. The default behaviour is obtained by bzr selftest with no options, and also (if possible) by running under another unittest harness.

Test feature dependencies

class TestStrace(TestCaseWithTransport):

    _test_needs_features = [StraceFeature]

self.requireFeature(StraceFeature)

class _SymlinkFeature(Feature):

    def _probe(self):
        return osutils.has_symlinks()

    def feature_name(self):
        return 'symlinks'

SymlinkFeature = _SymlinkFeature()

# in bzrlib/tests/features.py
apport = tests.ModuleAvailableFeature('apport')


# then in bzrlib/tests/test_apport.py
class TestApportReporting(TestCaseInTempDir):

    _test_needs_features = [features.apport]

Testing exceptions and errors

It’s important to test handling of errors and exceptions. Because this code is often not hit in ad-hoc testing it can often have hidden bugs – it’s particularly common to get NameError because the exception code references a variable that has since been renamed.

In general we want to test errors at two levels:

1. A test in test_errors.py checking that when the exception object is constructed with known parameters it produces an expected string form. This guards against mistakes in writing the format string, or in the str representations of its parameters. There should be one for each exception class.
2. Tests that when an api is called in a particular situation, it raises an error of the expected class. You should typically use assertRaises, which in the Bazaar test suite returns the exception object to allow you to examine its parameters.

In some cases blackbox tests will also want to check error reporting. But it can be difficult to provoke every error through the commandline interface, so those tests are only done as needed — eg in response to a particular bug or if the error is reported in an unusual way(?) Blackbox tests should mostly be testing how the command-line interface works, so should only test errors if there is something particular to the cli in how they’re displayed or handled.

Testing warnings

The Python warnings module is used to indicate a non-fatal code problem. Code that’s expected to raise a warning can be tested through callCatchWarnings.

The test suite can be run with -Werror to check no unexpected errors occur.

However, warnings should be used with discretion. It’s not an appropriate way to give messages to the user, because the warning is normally shown only once per source line that causes the problem. You should also think about whether the warning is serious enought that it should be visible to users who may not be able to fix it.

Interface implementation testing and test scenarios

There are several cases in Bazaar of multiple implementations of a common conceptual interface. (“Conceptual” because it’s not necessary for all the implementations to share a base class, though they often do.) Examples include transports and the working tree, branch and repository classes.

In these cases we want to make sure that every implementation correctly fulfils the interface requirements. For example, every Transport should support the has() and get() and clone() methods. We have a sub-suite of tests in test_transport_implementations. (Most per-implementation tests are in submodules of bzrlib.tests, but not the transport tests at the moment.)

These tests are repeated for each registered Transport, by generating a new TestCase instance for the cross product of test methods and transport implementations. As each test runs, it has transport_class and transport_server set to the class it should test. Most tests don’t access these directly, but rather use self.get_transport which returns a transport of the appropriate type.

The goal is to run per-implementation only the tests that relate to that particular interface. Sometimes we discover a bug elsewhere that happens with only one particular transport. Once it’s isolated, we can consider whether a test should be added for that particular implementation, or for all implementations of the interface.

The multiplication of tests for different implementations is normally accomplished by overriding the load_tests function used to load tests from a module. This function typically loads all the tests, then applies a TestProviderAdapter to them, which generates a longer suite containing all the test variations.

See also Per-implementation tests (above).

Test scenarios

Some utilities are provided for generating variations of tests. This can be used for per-implementation tests, or other cases where the same test code needs to run several times on different scenarios.

The general approach is to define a class that provides test methods, which depend on attributes of the test object being pre-set with the values to which the test should be applied. The test suite should then also provide a list of scenarios in which to run the tests.

Typically multiply_tests_from_modules should be called from the test module’s load_tests function.

Test support

We have a rich collection of tools to support writing tests. Please use them in preference to ad-hoc solutions as they provide portability and performance benefits.

builder = self.make_branch_builder('relpath')
builder.build_commit()
builder.build_commit()
builder.build_commit()
branch = builder.get_branch()

tree = self.make_branch_and_tree('path')
builder = TreeBuilder()
builder.start_tree(tree)
builder.build(['foo', "bar/", "bar/file"])
tree.commit('commit the tree')
builder.finish_tree()