/brz/remove-bazaar

To get this branch, use:
bzr branch http://gegoxaren.bato24.eu/bzr/brz/remove-bazaar
4070.8.3 by Martin Pool
Finish debug_flags support and add a test case
1
# Copyright (C) 2005, 2006, 2007, 2008, 2009 Canonical Ltd
1887.1.1 by Adeodato Simó
Do not separate paragraphs in the copyright statement with blank lines,
2
#
608 by Martin Pool
- Split selftests out into a new module and start changing them
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
1887.1.1 by Adeodato Simó
Do not separate paragraphs in the copyright statement with blank lines,
7
#
608 by Martin Pool
- Split selftests out into a new module and start changing them
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
11
# GNU General Public License for more details.
1887.1.1 by Adeodato Simó
Do not separate paragraphs in the copyright statement with blank lines,
12
#
608 by Martin Pool
- Split selftests out into a new module and start changing them
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
4183.7.1 by Sabin Iacob
update FSF mailing address
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
608 by Martin Pool
- Split selftests out into a new module and start changing them
16
609 by Martin Pool
- cleanup test code
17
1535 by Martin Pool
todo idea from John
18
# TODO: Perhaps there should be an API to find out if bzr running under the
19
# test suite -- some plugins might want to avoid making intrusive changes if
20
# this is the case.  However, we want behaviour under to test to diverge as
21
# little as possible, so this should be used rarely if it's added at all.
22
# (Suggestion from j-a-meinel, 2005-11-24)
23
1553.5.68 by Martin Pool
Add new TestCaseWithTransport.assertIsDirectory() and tests
24
# NOTE: Some classes in here use camelCaseNaming() rather than
25
# underscore_naming().  That's for consistency with unittest; it's not the
26
# general style of bzrlib.  Please continue that consistency when adding e.g.
27
# new assertFoo() methods.
28
2485.6.6 by Martin Pool
Put test root directory (containing per-test directories) in TMPDIR
29
import atexit
1534.4.41 by Robert Collins
Branch now uses BzrDir reasonably sanely.
30
import codecs
4650.1.5 by Robert Collins
Use copy, not deepcopy, to clone tests in test parameterisation.
31
from copy import copy
1185.1.29 by Robert Collins
merge merge tweaks from aaron, which includes latest .dev
32
from cStringIO import StringIO
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
33
import difflib
1740.6.1 by Martin Pool
Remove Scratch objects used by doctests
34
import doctest
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
35
import errno
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
36
import logging
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
37
import math
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
38
import os
2255.2.185 by Martin Pool
assertEqual uses pformat to show results
39
from pprint import pformat
2394.2.2 by Ian Clatworthy
Add --randomize and update help
40
import random
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
41
import re
2530.3.1 by Martin Pool
Cleanup old variations on run_bzr in the test suite
42
import shlex
1530.1.17 by Robert Collins
Move check_mode to TestCase.assertMode to make it generally accessible.
43
import stat
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
44
from subprocess import Popen, PIPE, STDOUT
1185.1.58 by Robert Collins
make selftest -v show the elapsed time for each test run.
45
import sys
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
46
import tempfile
3406.1.2 by Vincent Ladeuil
Fix as per Robert's review.
47
import threading
3084.1.1 by Andrew Bennetts
Add a --coverage option to selftest.
48
import time
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
49
import unittest
2485.6.5 by Martin Pool
Remove keep_output option
50
import warnings
1534.4.41 by Robert Collins
Branch now uses BzrDir reasonably sanely.
51
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
52
2095.5.3 by Martin Pool
Disable all debug_flags when running blackbox tests
53
from bzrlib import (
3567.4.12 by John Arbash Meinel
Expose the branch building framework to the test suite.
54
    branchbuilder,
2095.5.3 by Martin Pool
Disable all debug_flags when running blackbox tests
55
    bzrdir,
4634.90.2 by Andrew Bennetts
Clear chk_map page cache in TestCase._run_bzr_core, causes blackbox.test_log to fail without fix in previous revision.
56
    chk_map,
4695.3.2 by Vincent Ladeuil
Simplified and claried as per Robert's review.
57
    config,
2095.5.3 by Martin Pool
Disable all debug_flags when running blackbox tests
58
    debug,
59
    errors,
4119.3.1 by Robert Collins
Create a single registry of all Hooks classes, removing the test suite knowledge of such hooks and allowing plugins to sensibly and safely define new hooks.
60
    hooks,
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
61
    lock as _mod_lock,
2095.5.3 by Martin Pool
Disable all debug_flags when running blackbox tests
62
    memorytree,
63
    osutils,
64
    progress,
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
65
    ui,
2095.5.3 by Martin Pool
Disable all debug_flags when running blackbox tests
66
    urlutils,
3649.6.2 by Vincent Ladeuil
Replace aliases in selftest --starting-with option.
67
    registry,
2292.1.1 by Marius Kruger
"bzr remove" and "bzr rm" will now remove the working file.
68
    workingtree,
2095.5.3 by Martin Pool
Disable all debug_flags when running blackbox tests
69
    )
1513 by Robert Collins
Blackbox tests are maintained within the bzrlib.tests.blackbox directory.
70
import bzrlib.branch
974.1.26 by aaron.bentley at utoronto
merged mbp@sourcefrog.net-20050817233101-0939da1cf91f2472
71
import bzrlib.commands
1551.12.28 by Aaron Bentley
Move bundle timestamp code to timestamp
72
import bzrlib.timestamp
2024.2.3 by John Arbash Meinel
Move out export tests from test_too_much, refactor
73
import bzrlib.export
1513 by Robert Collins
Blackbox tests are maintained within the bzrlib.tests.blackbox directory.
74
import bzrlib.inventory
1185.67.6 by Aaron Bentley
Added tests and fixes for LockableFiles.put_utf8(); imported IterableFile
75
import bzrlib.iterablefile
1553.5.19 by Martin Pool
Run lockdir doctests
76
import bzrlib.lockdir
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
77
try:
78
    import bzrlib.lsprof
79
except ImportError:
80
    # lsprof not available
81
    pass
1649.1.1 by Robert Collins
* 'pull' and 'push' now normalise the revision history, so that any two
82
from bzrlib.merge import merge_inner
1513 by Robert Collins
Blackbox tests are maintained within the bzrlib.tests.blackbox directory.
83
import bzrlib.merge3
84
import bzrlib.plugin
4160.2.4 by Andrew Bennetts
Use BzrDir pre_open hook to jail request code from accessing transports other than the backing transport.
85
from bzrlib.smart import client, request, server
1513 by Robert Collins
Blackbox tests are maintained within the bzrlib.tests.blackbox directory.
86
import bzrlib.store
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
87
from bzrlib import symbol_versioning
2530.3.1 by Martin Pool
Cleanup old variations on run_bzr in the test suite
88
from bzrlib.symbol_versioning import (
2921.6.5 by Robert Collins
* The ``exclude_pattern`` parameter to the ``bzrlib.tests.`` functions
89
    DEPRECATED_PARAMETER,
2921.6.6 by Robert Collins
* The ``exclude_pattern`` and ``random_order`` parameters to the function
90
    deprecated_function,
91
    deprecated_method,
2921.6.5 by Robert Collins
* The ``exclude_pattern`` parameter to the ``bzrlib.tests.`` functions
92
    deprecated_passed,
2530.3.1 by Martin Pool
Cleanup old variations on run_bzr in the test suite
93
    )
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
94
import bzrlib.trace
4634.43.20 by Andrew Bennetts
Merge from bzr.dev, resolving conflicts.
95
from bzrlib.transport import get_transport, pathfilter
1534.4.3 by Robert Collins
Implement BranchTestProviderAdapter, so tests now run across all branch formats.
96
import bzrlib.transport
1951.2.1 by Martin Pool
Change to using LocalURLServer for testing.
97
from bzrlib.transport.local import LocalURLServer
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
98
from bzrlib.transport.memory import MemoryServer
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
99
from bzrlib.transport.readonly import ReadonlyServer
2095.4.1 by Martin Pool
Better progress bars during tests
100
from bzrlib.trace import mutter, note
1773.4.1 by Martin Pool
Add pyflakes makefile target; fix many warnings
101
from bzrlib.tests import TestUtil
2929.3.7 by Vincent Ladeuil
Rename bzrlib/test/HttpServer.py to bzrlib/tests/http_server.py and fix uses.
102
from bzrlib.tests.http_server import HttpServer
1773.4.1 by Martin Pool
Add pyflakes makefile target; fix many warnings
103
from bzrlib.tests.TestUtil import (
104
                          TestSuite,
105
                          TestLoader,
106
                          )
1514 by Robert Collins
Unbreak self.build_tree_shape in tests.
107
from bzrlib.tests.treeshape import build_tree_contents
4580.2.1 by Martin Pool
TestUIFactory no longer needs to pretend to be its own ProgressView
108
from bzrlib.ui import NullProgressView
4449.3.18 by Martin Pool
Fuse CLIUIFactory and TextUIFactory and deprecate the old name
109
from bzrlib.ui.text import TextUIFactory
2948.4.1 by Lukáš Lalinský
Custom template-based version info formatter.
110
import bzrlib.version_info_formats.format_custom
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
111
from bzrlib.workingtree import WorkingTree, WorkingTreeFormat2
1147 by Martin Pool
- split builtin commands into separate module bzrlib.builtins;
112
2387.2.1 by Robert Collins
Mark bzrlib.tests as providing assertFOO helper functions by adding a __unittest global attribute. (Robert Collins, Andrew Bennetts, Martin Pool, Jonathan Lange)
113
# Mark this python module as being part of the implementation
114
# of unittest: this gives us better tracebacks where the last
115
# shown frame is the test code, not our assertXYZ.
2598.5.7 by Aaron Bentley
Updates from review
116
__unittest = 1
2387.2.1 by Robert Collins
Mark bzrlib.tests as providing assertFOO helper functions by adding a __unittest global attribute. (Robert Collins, Andrew Bennetts, Martin Pool, Jonathan Lange)
117
1951.2.1 by Martin Pool
Change to using LocalURLServer for testing.
118
default_transport = LocalURLServer
1534.4.25 by Robert Collins
Add a --transport parameter to the test suite to set the default transport to be used in the test suite.
119
4573.2.3 by Robert Collins
Support python 2.4.
120
# Subunit result codes, defined here to prevent a hard dependency on subunit.
121
SUBUNIT_SEEK_SET = 0
122
SUBUNIT_SEEK_CUR = 1
123
1185.82.7 by John Arbash Meinel
Adding patches.py into bzrlib, including the tests into the test suite.
124
2095.4.1 by Martin Pool
Better progress bars during tests
125
class ExtendedTestResult(unittest._TextTestResult):
126
    """Accepts, reports and accumulates the results of running tests.
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
127
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
128
    Compared to the unittest version this class adds support for
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
129
    profiling, benchmarking, stopping as soon as a test fails,  and
130
    skipping tests.  There are further-specialized subclasses for
131
    different types of display.
132
133
    When a test finishes, in whatever way, it calls one of the addSuccess,
134
    addFailure or addError classes.  These in turn may redirect to a more
135
    specific case for the special test results supported by our extended
136
    tests.
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
137
138
    Note that just one of these objects is fed the results from many tests.
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
139
    """
2095.4.1 by Martin Pool
Better progress bars during tests
140
1185.62.21 by John Arbash Meinel
Allow bzr selftest --one to continue, even if we have a Skipped test.
141
    stop_early = False
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
142
2095.4.1 by Martin Pool
Better progress bars during tests
143
    def __init__(self, stream, descriptions, verbosity,
144
                 bench_history=None,
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
145
                 strict=False,
2095.4.1 by Martin Pool
Better progress bars during tests
146
                 ):
1819.1.8 by Martin Pool
Improved reporting of bzrlib revision_id
147
        """Construct new TestResult.
148
149
        :param bench_history: Optionally, a writable file object to accumulate
150
            benchmark results.
151
        """
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
152
        unittest._TextTestResult.__init__(self, stream, descriptions, verbosity)
1819.1.1 by Carl Friedrich Bolz
(lifeless, cfbolz, hpk): Give the test result object an optional benchmark
153
        if bench_history is not None:
1819.1.8 by Martin Pool
Improved reporting of bzrlib revision_id
154
            from bzrlib.version import _get_bzr_source_tree
155
            src_tree = _get_bzr_source_tree()
156
            if src_tree:
1908.7.6 by Robert Collins
Deprecate WorkingTree.last_revision.
157
                try:
158
                    revision_id = src_tree.get_parent_ids()[0]
159
                except IndexError:
160
                    # XXX: if this is a brand new tree, do the same as if there
161
                    # is no branch.
162
                    revision_id = ''
1819.1.8 by Martin Pool
Improved reporting of bzrlib revision_id
163
            else:
164
                # XXX: If there's no branch, what should we do?
165
                revision_id = ''
1819.1.4 by Jan Balster
save the revison id for every benchmark run in .perf-history
166
            bench_history.write("--date %s %s\n" % (time.time(), revision_id))
1819.1.3 by Carl Friedrich Bolz
(lifeless, cfbolz): Add recording of benchmark results to the benchmark history
167
        self._bench_history = bench_history
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
168
        self.ui = ui.ui_factory
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
169
        self.num_tests = 0
2095.4.1 by Martin Pool
Better progress bars during tests
170
        self.error_count = 0
171
        self.failure_count = 0
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
172
        self.known_failure_count = 0
2095.4.1 by Martin Pool
Better progress bars during tests
173
        self.skip_count = 0
2729.1.1 by Martin Pool
Add TestNotApplicable exception and handling of it; document test parameterization
174
        self.not_applicable_count = 0
2367.1.5 by Robert Collins
Implement reporting of Unsupported tests in the bzr test result and runner
175
        self.unsupported = {}
2095.4.1 by Martin Pool
Better progress bars during tests
176
        self.count = 0
177
        self._overall_start_time = time.time()
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
178
        self._strict = strict
179
4650.1.6 by Robert Collins
Fix interface skew between bzr selftest and python unittest - use stopTestRun not done to end test runs.
180
    def stopTestRun(self):
4650.1.7 by Robert Collins
Push result reporting thoroughly into TestResult.
181
        run = self.testsRun
182
        actionTaken = "Ran"
183
        stopTime = time.time()
184
        timeTaken = stopTime - self.startTime
185
        self.printErrors()
186
        self.stream.writeln(self.separator2)
187
        self.stream.writeln("%s %d test%s in %.3fs" % (actionTaken,
188
                            run, run != 1 and "s" or "", timeTaken))
189
        self.stream.writeln()
190
        if not self.wasSuccessful():
191
            self.stream.write("FAILED (")
192
            failed, errored = map(len, (self.failures, self.errors))
193
            if failed:
194
                self.stream.write("failures=%d" % failed)
195
            if errored:
196
                if failed: self.stream.write(", ")
197
                self.stream.write("errors=%d" % errored)
198
            if self.known_failure_count:
199
                if failed or errored: self.stream.write(", ")
200
                self.stream.write("known_failure_count=%d" %
201
                    self.known_failure_count)
202
            self.stream.writeln(")")
203
        else:
204
            if self.known_failure_count:
205
                self.stream.writeln("OK (known_failures=%d)" %
206
                    self.known_failure_count)
207
            else:
208
                self.stream.writeln("OK")
209
        if self.skip_count > 0:
210
            skipped = self.skip_count
211
            self.stream.writeln('%d test%s skipped' %
212
                                (skipped, skipped != 1 and "s" or ""))
213
        if self.unsupported:
214
            for feature, count in sorted(self.unsupported.items()):
215
                self.stream.writeln("Missing feature '%s' skipped %d tests." %
216
                    (feature, count))
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
217
        if self._strict:
218
            ok = self.wasStrictlySuccessful()
219
        else:
220
            ok = self.wasSuccessful()
4271.2.2 by Robert Collins
Move thread leak reporting to ExtendedTestResult.
221
        if TestCase._first_thread_leaker_id:
222
            self.stream.write(
223
                '%s is leaking threads among %d leaking tests.\n' % (
224
                TestCase._first_thread_leaker_id,
225
                TestCase._leaking_threads_tests))
4732.2.1 by Vincent Ladeuil
Clearer thread leaks reports.
226
            # We don't report the main thread as an active one.
227
            self.stream.write(
228
                '%d non-main threads were left active in the end.\n'
229
                % (TestCase._active_threads - 1))
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
230
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
231
    def _extractBenchmarkTime(self, testCase):
1707.2.3 by Robert Collins
Add a setBenchmarkTime method to the bzrlib test result allowing introduction of granular benchmarking. (Robert Collins, Martin Pool).
232
        """Add a benchmark time for the current test case."""
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
233
        return getattr(testCase, "_benchtime", None)
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
234
1707.2.3 by Robert Collins
Add a setBenchmarkTime method to the bzrlib test result allowing introduction of granular benchmarking. (Robert Collins, Martin Pool).
235
    def _elapsedTestTimeString(self):
236
        """Return a time string for the overall time the current test has taken."""
237
        return self._formatTime(time.time() - self._start_time)
238
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
239
    def _testTimeString(self, testCase):
240
        benchmark_time = self._extractBenchmarkTime(testCase)
241
        if benchmark_time is not None:
4536.5.2 by Martin Pool
Reserve less space for test elapsed time; more space for test name
242
            return self._formatTime(benchmark_time) + "*"
1707.2.3 by Robert Collins
Add a setBenchmarkTime method to the bzrlib test result allowing introduction of granular benchmarking. (Robert Collins, Martin Pool).
243
        else:
4536.5.5 by Martin Pool
More selftest display test tweaks
244
            return self._elapsedTestTimeString()
1707.2.3 by Robert Collins
Add a setBenchmarkTime method to the bzrlib test result allowing introduction of granular benchmarking. (Robert Collins, Martin Pool).
245
246
    def _formatTime(self, seconds):
247
        """Format seconds as milliseconds with leading spaces."""
2196.1.1 by Martin Pool
better formatting of benchmark output so it doesn't wrap
248
        # some benchmarks can take thousands of seconds to run, so we need 8
249
        # places
250
        return "%8dms" % (1000 * seconds)
1185.1.58 by Robert Collins
make selftest -v show the elapsed time for each test run.
251
2095.4.1 by Martin Pool
Better progress bars during tests
252
    def _shortened_test_description(self, test):
253
        what = test.id()
2196.1.1 by Martin Pool
better formatting of benchmark output so it doesn't wrap
254
        what = re.sub(r'^bzrlib\.(tests|benchmarks)\.', '', what)
2095.4.1 by Martin Pool
Better progress bars during tests
255
        return what
1534.11.3 by Robert Collins
Show test names and status in the progress bar.
256
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
257
    def startTest(self, test):
258
        unittest.TestResult.startTest(self, test)
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
259
        if self.count == 0:
260
            self.startTests()
2095.4.1 by Martin Pool
Better progress bars during tests
261
        self.report_test_start(test)
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
262
        test.number = self.count
1707.2.3 by Robert Collins
Add a setBenchmarkTime method to the bzrlib test result allowing introduction of granular benchmarking. (Robert Collins, Martin Pool).
263
        self._recordTestStartTime()
264
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
265
    def startTests(self):
4595.7.2 by Martin Pool
Also put platform identifier in --version and selftest output
266
        import platform
4574.2.1 by Alexander Belchenko
Show actual path to bzr.exe in tests (workaround for invoking bzr selftest from Makefile).
267
        if getattr(sys, 'frozen', None) is None:
268
            bzr_path = osutils.realpath(sys.argv[0])
269
        else:
270
            bzr_path = sys.executable
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
271
        self.stream.write(
4574.2.1 by Alexander Belchenko
Show actual path to bzr.exe in tests (workaround for invoking bzr selftest from Makefile).
272
            'testing: %s\n' % (bzr_path,))
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
273
        self.stream.write(
4595.7.2 by Martin Pool
Also put platform identifier in --version and selftest output
274
            '   %s\n' % (
275
                    bzrlib.__path__[0],))
276
        self.stream.write(
277
            '   bzr-%s python-%s %s\n' % (
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
278
                    bzrlib.version_string,
279
                    bzrlib._format_version_tuple(sys.version_info),
4595.7.3 by Martin Pool
Use platform(aliased=1)
280
                    platform.platform(aliased=1),
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
281
                    ))
282
        self.stream.write('\n')
283
1707.2.3 by Robert Collins
Add a setBenchmarkTime method to the bzrlib test result allowing introduction of granular benchmarking. (Robert Collins, Martin Pool).
284
    def _recordTestStartTime(self):
285
        """Record that a test has started."""
1185.1.58 by Robert Collins
make selftest -v show the elapsed time for each test run.
286
        self._start_time = time.time()
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
287
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
288
    def _cleanupLogFile(self, test):
289
        # We can only do this if we have one of our TestCases, not if
290
        # we have a doctest.
291
        setKeepLogfile = getattr(test, 'setKeepLogfile', None)
292
        if setKeepLogfile is not None:
293
            setKeepLogfile()
294
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
295
    def addError(self, test, err):
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
296
        """Tell result that test finished with an error.
297
298
        Called from the TestCase run() method when the test
299
        fails with an unexpected error.
300
        """
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
301
        self._post_mortem()
302
        unittest.TestResult.addError(self, test, err)
303
        self.error_count += 1
304
        self.report_error(test, err)
305
        if self.stop_early:
306
            self.stop()
307
        self._cleanupLogFile(test)
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
308
309
    def addFailure(self, test, err):
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
310
        """Tell result that test failed.
311
312
        Called from the TestCase run() method when the test
313
        fails because e.g. an assert() method failed.
314
        """
4780.1.4 by Robert Collins
Switch reporting of KnownFailure to be Python2.7 compatible.
315
        self._post_mortem()
316
        unittest.TestResult.addFailure(self, test, err)
317
        self.failure_count += 1
318
        self.report_failure(test, err)
319
        if self.stop_early:
320
            self.stop()
321
        self._cleanupLogFile(test)
2367.1.5 by Robert Collins
Implement reporting of Unsupported tests in the bzr test result and runner
322
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
323
    def addSuccess(self, test):
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
324
        """Tell result that test completed successfully.
325
326
        Called from the TestCase run()
327
        """
1819.1.3 by Carl Friedrich Bolz
(lifeless, cfbolz): Add recording of benchmark results to the benchmark history
328
        if self._bench_history is not None:
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
329
            benchmark_time = self._extractBenchmarkTime(test)
330
            if benchmark_time is not None:
1819.1.3 by Carl Friedrich Bolz
(lifeless, cfbolz): Add recording of benchmark results to the benchmark history
331
                self._bench_history.write("%s %s\n" % (
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
332
                    self._formatTime(benchmark_time),
1819.1.3 by Carl Friedrich Bolz
(lifeless, cfbolz): Add recording of benchmark results to the benchmark history
333
                    test.id()))
2095.4.1 by Martin Pool
Better progress bars during tests
334
        self.report_success(test)
3199.1.1 by Vincent Ladeuil
Get rid of ~1000 useless log files out of 10.000 tests in /tmp.
335
        self._cleanupLogFile(test)
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
336
        unittest.TestResult.addSuccess(self, test)
3224.4.4 by Andrew Bennetts
Tweak clearing of _log_contents (idea from John).
337
        test._log_contents = ''
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
338
4780.1.4 by Robert Collins
Switch reporting of KnownFailure to be Python2.7 compatible.
339
    def addExpectedFailure(self, test, err):
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
340
        self.known_failure_count += 1
341
        self.report_known_failure(test, err)
342
343
    def addNotSupported(self, test, feature):
344
        """The test will not be run because of a missing feature.
345
        """
346
        # this can be called in two different ways: it may be that the
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
347
        # test started running, and then raised (through requireFeature)
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
348
        # UnavailableFeature.  Alternatively this method can be called
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
349
        # while probing for features before running the test code proper; in
350
        # that case we will see startTest and stopTest, but the test will
351
        # never actually run.
2695.1.1 by Martin Pool
Fix problem if the first test is missing a dependency
352
        self.unsupported.setdefault(str(feature), 0)
353
        self.unsupported[str(feature)] += 1
354
        self.report_unsupported(test, feature)
355
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
356
    def addSkip(self, test, reason):
357
        """A test has not run for 'reason'."""
358
        self.skip_count += 1
359
        self.report_skip(test, reason)
360
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
361
    def addNotApplicable(self, test, reason):
362
        self.not_applicable_count += 1
363
        self.report_not_applicable(test, reason)
1185.33.95 by Martin Pool
New TestSkipped facility, and tests for it.
364
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
365
    def printErrorList(self, flavour, errors):
366
        for test, err in errors:
367
            self.stream.writeln(self.separator1)
2321.3.2 by Alexander Belchenko
numbered dirs: printErrorList show test number for NUMBERED_DIRS
368
            self.stream.write("%s: " % flavour)
369
            self.stream.writeln(self.getDescription(test))
1534.4.3 by Robert Collins
Implement BranchTestProviderAdapter, so tests now run across all branch formats.
370
            if getattr(test, '_get_log', None) is not None:
4536.5.1 by Martin Pool
Only print selftest error log if it's non-empty
371
                log_contents = test._get_log()
372
                if log_contents:
373
                    self.stream.write('\n')
374
                    self.stream.write(
375
                            ('vvvv[log from %s]' % test.id()).ljust(78,'-'))
376
                    self.stream.write('\n')
377
                    self.stream.write(log_contents)
378
                    self.stream.write('\n')
379
                    self.stream.write(
380
                            ('^^^^[log from %s]' % test.id()).ljust(78,'-'))
381
                    self.stream.write('\n')
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
382
            self.stream.writeln(self.separator2)
383
            self.stream.writeln("%s" % err)
384
4685.1.1 by Robert Collins
Use BZR_TEST_PDB=1 to trigger post_mortems in test failures.
385
    def _post_mortem(self):
386
        """Start a PDB post mortem session."""
387
        if os.environ.get('BZR_TEST_PDB', None):
388
            import pdb;pdb.post_mortem()
389
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
390
    def progress(self, offset, whence):
391
        """The test is adjusting the count of tests to run."""
4573.2.3 by Robert Collins
Support python 2.4.
392
        if whence == SUBUNIT_SEEK_SET:
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
393
            self.num_tests = offset
4573.2.3 by Robert Collins
Support python 2.4.
394
        elif whence == SUBUNIT_SEEK_CUR:
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
395
            self.num_tests += offset
396
        else:
397
            raise errors.BzrError("Unknown whence %r" % whence)
398
2095.4.1 by Martin Pool
Better progress bars during tests
399
    def report_cleaning_up(self):
400
        pass
401
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
402
    def startTestRun(self):
4650.1.7 by Robert Collins
Push result reporting thoroughly into TestResult.
403
        self.startTime = time.time()
404
2095.4.1 by Martin Pool
Better progress bars during tests
405
    def report_success(self, test):
406
        pass
407
2658.3.1 by Daniel Watkins
Added ExtendedTestResult.wasStrictlySuccessful.
408
    def wasStrictlySuccessful(self):
409
        if self.unsupported or self.known_failure_count:
410
            return False
411
        return self.wasSuccessful()
412
413
2095.4.1 by Martin Pool
Better progress bars during tests
414
class TextTestResult(ExtendedTestResult):
415
    """Displays progress and results of tests in text form"""
416
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
417
    def __init__(self, stream, descriptions, verbosity,
418
                 bench_history=None,
419
                 pb=None,
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
420
                 strict=None,
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
421
                 ):
422
        ExtendedTestResult.__init__(self, stream, descriptions, verbosity,
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
423
            bench_history, strict)
4580.3.2 by Martin Pool
TextTestResult now clears off the pb when the tests are done
424
        # We no longer pass them around, but just rely on the UIFactory stack
425
        # for state
426
        if pb is not None:
427
            warnings.warn("Passing pb to TextTestResult is deprecated")
428
        self.pb = self.ui.nested_progress_bar()
2095.4.5 by mbp at sourcefrog
Use regular progress-bar classes, not a special mechanism
429
        self.pb.show_pct = False
430
        self.pb.show_spinner = False
2367.1.6 by Robert Collins
Allow per-test-fixture feature requirements via 'requireFeature'.(Robert Collins)
431
        self.pb.show_eta = False,
2095.4.5 by mbp at sourcefrog
Use regular progress-bar classes, not a special mechanism
432
        self.pb.show_count = False
433
        self.pb.show_bar = False
4580.3.3 by Martin Pool
Test progress bar has zero latency so it's more accurate
434
        self.pb.update_latency = 0
4580.3.5 by Martin Pool
selftest sets ProgressTask.show_transport_activity off
435
        self.pb.show_transport_activity = False
2095.4.5 by mbp at sourcefrog
Use regular progress-bar classes, not a special mechanism
436
4650.1.6 by Robert Collins
Fix interface skew between bzr selftest and python unittest - use stopTestRun not done to end test runs.
437
    def stopTestRun(self):
4580.3.6 by Martin Pool
TextTestResult should also clear pb before done()
438
        # called when the tests that are going to run have run
439
        self.pb.clear()
4650.1.7 by Robert Collins
Push result reporting thoroughly into TestResult.
440
        self.pb.finished()
4650.1.6 by Robert Collins
Fix interface skew between bzr selftest and python unittest - use stopTestRun not done to end test runs.
441
        super(TextTestResult, self).stopTestRun()
4580.3.6 by Martin Pool
TextTestResult should also clear pb before done()
442
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
443
    def startTestRun(self):
444
        super(TextTestResult, self).startTestRun()
4103.3.2 by Martin Pool
Remove trailing punctuation from progress messages
445
        self.pb.update('[test 0/%d] Starting' % (self.num_tests))
2095.4.1 by Martin Pool
Better progress bars during tests
446
4580.3.6 by Martin Pool
TextTestResult should also clear pb before done()
447
    def printErrors(self):
448
        # clear the pb to make room for the error listing
449
        self.pb.clear()
450
        super(TextTestResult, self).printErrors()
451
2095.4.1 by Martin Pool
Better progress bars during tests
452
    def _progress_prefix_text(self):
3297.1.1 by Martin Pool
More concise display of test progress bar
453
        # the longer this text, the less space we have to show the test
454
        # name...
455
        a = '[%d' % self.count              # total that have been run
456
        # tests skipped as known not to be relevant are not important enough
457
        # to show here
458
        ## if self.skip_count:
459
        ##     a += ', %d skip' % self.skip_count
460
        ## if self.known_failure_count:
461
        ##     a += '+%dX' % self.known_failure_count
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
462
        if self.num_tests:
2095.4.1 by Martin Pool
Better progress bars during tests
463
            a +='/%d' % self.num_tests
3297.1.1 by Martin Pool
More concise display of test progress bar
464
        a += ' in '
465
        runtime = time.time() - self._overall_start_time
466
        if runtime >= 60:
467
            a += '%dm%ds' % (runtime / 60, runtime % 60)
468
        else:
469
            a += '%ds' % runtime
3297.1.3 by Martin Pool
Fix up selftest progress tests
470
        if self.error_count:
471
            a += ', %d err' % self.error_count
472
        if self.failure_count:
473
            a += ', %d fail' % self.failure_count
4721.1.1 by Martin Pool
Stop showing the number of tests due to missing features in the test progress bar.
474
        # if self.unsupported:
475
        #     a += ', %d missing' % len(self.unsupported)
2095.4.3 by Martin Pool
Tweak test display a bit more
476
        a += ']'
2095.4.1 by Martin Pool
Better progress bars during tests
477
        return a
478
479
    def report_test_start(self, test):
480
        self.count += 1
2095.4.5 by mbp at sourcefrog
Use regular progress-bar classes, not a special mechanism
481
        self.pb.update(
2095.4.1 by Martin Pool
Better progress bars during tests
482
                self._progress_prefix_text()
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
483
                + ' '
2095.4.1 by Martin Pool
Better progress bars during tests
484
                + self._shortened_test_description(test))
485
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
486
    def _test_description(self, test):
2598.4.1 by Martin Pool
Remove obsolete --clean-output, --keep-output, --numbered-dirs selftest options (thanks Alexander)
487
        return self._shortened_test_description(test)
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
488
2095.4.3 by Martin Pool
Tweak test display a bit more
489
    def report_error(self, test, err):
4471.2.2 by Martin Pool
Deprecate ProgressTask.note
490
        ui.ui_factory.note('ERROR: %s\n    %s\n' % (
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
491
            self._test_description(test),
2095.4.3 by Martin Pool
Tweak test display a bit more
492
            err[1],
4471.2.2 by Martin Pool
Deprecate ProgressTask.note
493
            ))
2095.4.1 by Martin Pool
Better progress bars during tests
494
2095.4.3 by Martin Pool
Tweak test display a bit more
495
    def report_failure(self, test, err):
4471.2.2 by Martin Pool
Deprecate ProgressTask.note
496
        ui.ui_factory.note('FAIL: %s\n    %s\n' % (
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
497
            self._test_description(test),
2095.4.3 by Martin Pool
Tweak test display a bit more
498
            err[1],
4471.2.2 by Martin Pool
Deprecate ProgressTask.note
499
            ))
2095.4.1 by Martin Pool
Better progress bars during tests
500
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
501
    def report_known_failure(self, test, err):
4471.2.2 by Martin Pool
Deprecate ProgressTask.note
502
        ui.ui_factory.note('XFAIL: %s\n%s\n' % (
503
            self._test_description(test), err[1]))
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
504
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
505
    def report_skip(self, test, reason):
2729.1.1 by Martin Pool
Add TestNotApplicable exception and handling of it; document test parameterization
506
        pass
507
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
508
    def report_not_applicable(self, test, reason):
2729.1.1 by Martin Pool
Add TestNotApplicable exception and handling of it; document test parameterization
509
        pass
2095.4.1 by Martin Pool
Better progress bars during tests
510
2367.1.5 by Robert Collins
Implement reporting of Unsupported tests in the bzr test result and runner
511
    def report_unsupported(self, test, feature):
512
        """test cannot be run because feature is missing."""
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
513
2095.4.1 by Martin Pool
Better progress bars during tests
514
    def report_cleaning_up(self):
4103.3.2 by Martin Pool
Remove trailing punctuation from progress messages
515
        self.pb.update('Cleaning up')
2095.4.1 by Martin Pool
Better progress bars during tests
516
517
518
class VerboseTestResult(ExtendedTestResult):
519
    """Produce long output, with one line per test run plus times"""
520
521
    def _ellipsize_to_right(self, a_string, final_width):
522
        """Truncate and pad a string, keeping the right hand side"""
523
        if len(a_string) > final_width:
524
            result = '...' + a_string[3-final_width:]
525
        else:
526
            result = a_string
527
        return result.ljust(final_width)
528
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
529
    def startTestRun(self):
530
        super(VerboseTestResult, self).startTestRun()
2095.4.1 by Martin Pool
Better progress bars during tests
531
        self.stream.write('running %d tests...\n' % self.num_tests)
532
533
    def report_test_start(self, test):
534
        self.count += 1
535
        name = self._shortened_test_description(test)
4747.3.6 by Vincent Ladeuil
terminal_width can now returns None.
536
        width = osutils.terminal_width()
537
        if width is not None:
538
            # width needs space for 6 char status, plus 1 for slash, plus an
539
            # 11-char time string, plus a trailing blank
540
            # when NUMBERED_DIRS: plus 5 chars on test number, plus 1 char on
541
            # space
542
            self.stream.write(self._ellipsize_to_right(name, width-18))
543
        else:
544
            self.stream.write(name)
2095.4.1 by Martin Pool
Better progress bars during tests
545
        self.stream.flush()
546
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
547
    def _error_summary(self, err):
548
        indent = ' ' * 4
549
        return '%s%s' % (indent, err[1])
550
2095.4.3 by Martin Pool
Tweak test display a bit more
551
    def report_error(self, test, err):
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
552
        self.stream.writeln('ERROR %s\n%s'
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
553
                % (self._testTimeString(test),
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
554
                   self._error_summary(err)))
2095.4.1 by Martin Pool
Better progress bars during tests
555
2095.4.3 by Martin Pool
Tweak test display a bit more
556
    def report_failure(self, test, err):
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
557
        self.stream.writeln(' FAIL %s\n%s'
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
558
                % (self._testTimeString(test),
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
559
                   self._error_summary(err)))
2095.4.1 by Martin Pool
Better progress bars during tests
560
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
561
    def report_known_failure(self, test, err):
562
        self.stream.writeln('XFAIL %s\n%s'
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
563
                % (self._testTimeString(test),
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
564
                   self._error_summary(err)))
565
2095.4.1 by Martin Pool
Better progress bars during tests
566
    def report_success(self, test):
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
567
        self.stream.writeln('   OK %s' % self._testTimeString(test))
2095.4.1 by Martin Pool
Better progress bars during tests
568
        for bench_called, stats in getattr(test, '_benchcalls', []):
569
            self.stream.writeln('LSProf output for %s(%s, %s)' % bench_called)
570
            stats.pprint(file=self.stream)
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
571
        # flush the stream so that we get smooth output. This verbose mode is
572
        # used to show the output in PQM.
2095.4.1 by Martin Pool
Better progress bars during tests
573
        self.stream.flush()
574
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
575
    def report_skip(self, test, reason):
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
576
        self.stream.writeln(' SKIP %s\n%s'
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
577
                % (self._testTimeString(test), reason))
2095.4.1 by Martin Pool
Better progress bars during tests
578
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
579
    def report_not_applicable(self, test, reason):
580
        self.stream.writeln('  N/A %s\n    %s'
581
                % (self._testTimeString(test), reason))
2729.1.1 by Martin Pool
Add TestNotApplicable exception and handling of it; document test parameterization
582
2367.1.5 by Robert Collins
Implement reporting of Unsupported tests in the bzr test result and runner
583
    def report_unsupported(self, test, feature):
584
        """test cannot be run because feature is missing."""
585
        self.stream.writeln("NODEP %s\n    The feature '%s' is not available."
2695.1.2 by Martin Pool
_benchmarkTime should not be an attribute of ExtendedTestResult, because it only applies to the most recent test reported
586
                %(self._testTimeString(test), feature))
2367.1.5 by Robert Collins
Implement reporting of Unsupported tests in the bzr test result and runner
587
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
588
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
589
class TextTestRunner(object):
1185.16.58 by mbp at sourcefrog
- run all selftests by default
590
    stop_on_failure = False
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
591
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
592
    def __init__(self,
593
                 stream=sys.stderr,
594
                 descriptions=0,
595
                 verbosity=1,
2379.6.3 by Alexander Belchenko
Rework NUMBERED_DIRS usage to keep test_selftest.py passing the tests on win32
596
                 bench_history=None,
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
597
                 strict=False,
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
598
                 result_decorators=None,
2379.6.3 by Alexander Belchenko
Rework NUMBERED_DIRS usage to keep test_selftest.py passing the tests on win32
599
                 ):
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
600
        """Create a TextTestRunner.
601
602
        :param result_decorators: An optional list of decorators to apply
603
            to the result object being used by the runner. Decorators are
604
            applied left to right - the first element in the list is the 
605
            innermost decorator.
606
        """
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
607
        self.stream = unittest._WritelnDecorator(stream)
608
        self.descriptions = descriptions
609
        self.verbosity = verbosity
1819.1.2 by Carl Friedrich Bolz
(lifeless, cfbolz, hpk): Add a benchmark output parameter to TextTestRunner.
610
        self._bench_history = bench_history
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
611
        self._strict = strict
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
612
        self._result_decorators = result_decorators or []
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
613
614
    def run(self, test):
615
        "Run the given test case or test suite."
2095.4.1 by Martin Pool
Better progress bars during tests
616
        if self.verbosity == 1:
617
            result_class = TextTestResult
618
        elif self.verbosity >= 2:
619
            result_class = VerboseTestResult
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
620
        original_result = result_class(self.stream,
2095.4.1 by Martin Pool
Better progress bars during tests
621
                              self.descriptions,
622
                              self.verbosity,
623
                              bench_history=self._bench_history,
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
624
                              strict=self._strict,
2095.4.1 by Martin Pool
Better progress bars during tests
625
                              )
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
626
        # Signal to result objects that look at stop early policy to stop,
627
        original_result.stop_early = self.stop_on_failure
628
        result = original_result
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
629
        for decorator in self._result_decorators:
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
630
            result = decorator(result)
631
            result.stop_early = self.stop_on_failure
4650.1.7 by Robert Collins
Push result reporting thoroughly into TestResult.
632
        try:
633
            import testtools
634
        except ImportError:
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
635
            pass
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
636
        else:
4650.1.7 by Robert Collins
Push result reporting thoroughly into TestResult.
637
            if isinstance(test, testtools.ConcurrentTestSuite):
638
                # We need to catch bzr specific behaviors
4650.1.8 by Robert Collins
Push all starting up reporting down into startTestRun.
639
                result = BZRTransformingResult(result)
640
        result.startTestRun()
641
        try:
642
            test.run(result)
643
        finally:
644
            result.stopTestRun()
645
        # higher level code uses our extended protocol to determine
646
        # what exit code to give.
647
        return original_result
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
648
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
649
1393.1.46 by Martin Pool
- bzr selftest arguments can be partial ids of tests to run
650
def iter_suite_tests(suite):
651
    """Return all tests in a suite, recursing through nested suites"""
4084.5.1 by Robert Collins
Bulk update all test adaptation into a single approach, using multiply_tests rather than test adapters.
652
    if isinstance(suite, unittest.TestCase):
653
        yield suite
654
    elif isinstance(suite, unittest.TestSuite):
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
655
        for item in suite:
1393.1.46 by Martin Pool
- bzr selftest arguments can be partial ids of tests to run
656
            for r in iter_suite_tests(item):
657
                yield r
4084.5.1 by Robert Collins
Bulk update all test adaptation into a single approach, using multiply_tests rather than test adapters.
658
    else:
659
        raise Exception('unknown type %r for object %r'
660
                        % (type(suite), suite))
1393.1.46 by Martin Pool
- bzr selftest arguments can be partial ids of tests to run
661
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
662
663
class TestSkipped(Exception):
664
    """Indicates that a test was intentionally skipped, rather than failing."""
665
666
2729.1.1 by Martin Pool
Add TestNotApplicable exception and handling of it; document test parameterization
667
class TestNotApplicable(TestSkipped):
668
    """A test is not applicable to the situation where it was run.
669
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
670
    This is only normally raised by parameterized tests, if they find that
671
    the instance they're constructed upon does not support one aspect
2729.1.1 by Martin Pool
Add TestNotApplicable exception and handling of it; document test parameterization
672
    of its interface.
673
    """
674
675
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
676
class KnownFailure(AssertionError):
677
    """Indicates that a test failed in a precisely expected manner.
678
679
    Such failures dont block the whole test suite from passing because they are
680
    indicators of partially completed code or of future work. We have an
681
    explicit error for them so that we can ensure that they are always visible:
682
    KnownFailures are always shown in the output of bzr selftest.
683
    """
684
685
2367.1.6 by Robert Collins
Allow per-test-fixture feature requirements via 'requireFeature'.(Robert Collins)
686
class UnavailableFeature(Exception):
687
    """A feature required for this test was not available.
688
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
689
    This can be considered a specialised form of SkippedTest.
690
2367.1.6 by Robert Collins
Allow per-test-fixture feature requirements via 'requireFeature'.(Robert Collins)
691
    The feature should be used to construct the exception.
692
    """
693
694
1147 by Martin Pool
- split builtin commands into separate module bzrlib.builtins;
695
class CommandFailed(Exception):
696
    pass
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
697
1185.85.8 by John Arbash Meinel
Adding wrapper for sys.stdout so we can set the output encoding. Adding tests that 'bzr log' handles multiple encodings properly
698
699
class StringIOWrapper(object):
700
    """A wrapper around cStringIO which just adds an encoding attribute.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
701
1185.85.8 by John Arbash Meinel
Adding wrapper for sys.stdout so we can set the output encoding. Adding tests that 'bzr log' handles multiple encodings properly
702
    Internally we can check sys.stdout to see what the output encoding
703
    should be. However, cStringIO has no encoding attribute that we can
704
    set. So we wrap it instead.
705
    """
706
    encoding='ascii'
707
    _cstring = None
708
709
    def __init__(self, s=None):
710
        if s is not None:
711
            self.__dict__['_cstring'] = StringIO(s)
712
        else:
713
            self.__dict__['_cstring'] = StringIO()
714
715
    def __getattr__(self, name, getattr=getattr):
716
        return getattr(self.__dict__['_cstring'], name)
717
718
    def __setattr__(self, name, val):
719
        if name == 'encoding':
720
            self.__dict__['encoding'] = val
721
        else:
722
            return setattr(self._cstring, name, val)
723
724
4449.3.18 by Martin Pool
Fuse CLIUIFactory and TextUIFactory and deprecate the old name
725
class TestUIFactory(TextUIFactory):
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
726
    """A UI Factory for testing.
727
728
    Hide the progress bar but emit note()s.
729
    Redirect stdin.
730
    Allows get_password to be tested without real tty attached.
4580.2.1 by Martin Pool
TestUIFactory no longer needs to pretend to be its own ProgressView
731
732
    See also CannedInputUIFactory which lets you provide programmatic input in
733
    a structured way.
2294.4.1 by Vincent Ladeuil
Add a UIFactory.get_login method, fix tests.
734
    """
4580.3.9 by Martin Pool
doc
735
    # TODO: Capture progress events at the model level and allow them to be
736
    # observed by tests that care.
4580.3.11 by Martin Pool
merge trunk
737
    #
4580.2.1 by Martin Pool
TestUIFactory no longer needs to pretend to be its own ProgressView
738
    # XXX: Should probably unify more with CannedInputUIFactory or a
739
    # particular configuration of TextUIFactory, or otherwise have a clearer
740
    # idea of how they're supposed to be different.
4580.2.3 by Martin Pool
doc
741
    # See https://bugs.edge.launchpad.net/bzr/+bug/408213
2294.4.1 by Vincent Ladeuil
Add a UIFactory.get_login method, fix tests.
742
4237.2.1 by Vincent Ladeuil
Stop requiring a tty for CLIUIFactory and derivatives.
743
    def __init__(self, stdout=None, stderr=None, stdin=None):
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
744
        if stdin is not None:
745
            # We use a StringIOWrapper to be able to test various
746
            # encodings, but the user is still responsible to
747
            # encode the string and to set the encoding attribute
748
            # of StringIOWrapper.
4237.2.1 by Vincent Ladeuil
Stop requiring a tty for CLIUIFactory and derivatives.
749
            stdin = StringIOWrapper(stdin)
750
        super(TestUIFactory, self).__init__(stdin, stdout, stderr)
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
751
4237.2.1 by Vincent Ladeuil
Stop requiring a tty for CLIUIFactory and derivatives.
752
    def get_non_echoed_password(self):
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
753
        """Get password from stdin without trying to handle the echo mode"""
754
        password = self.stdin.readline()
755
        if not password:
756
            raise EOFError
757
        if password[-1] == '\n':
758
            password = password[:-1]
759
        return password
2294.4.1 by Vincent Ladeuil
Add a UIFactory.get_login method, fix tests.
760
4463.1.2 by Martin Pool
merge trunk
761
    def make_progress_view(self):
762
        return NullProgressView()
763
2294.4.1 by Vincent Ladeuil
Add a UIFactory.get_login method, fix tests.
764
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
765
class TestCase(unittest.TestCase):
766
    """Base class for bzr unit tests.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
767
768
    Tests that need access to disk resources should subclass
1141 by Martin Pool
- rename FunctionalTest to TestCaseInTempDir
769
    TestCaseInTempDir not TestCase.
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
770
771
    Error and debug log messages are redirected from their usual
772
    location into a temporary file, the contents of which can be
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
773
    retrieved by _get_log().  We use a real OS file, not an in-memory object,
774
    so that it can also capture file IO.  When the test completes this file
775
    is read into memory and removed from disk.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
776
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
777
    There are also convenience functions to invoke bzr's command-line
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
778
    routine, and to build and check bzr trees.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
779
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
780
    In addition to the usual method of overriding tearDown(), this class also
781
    allows subclasses to register functions into the _cleanups list, which is
782
    run in order as the object is torn down.  It's less likely this will be
783
    accidentally overlooked.
784
    """
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
785
3406.1.2 by Vincent Ladeuil
Fix as per Robert's review.
786
    _active_threads = None
787
    _leaking_threads_tests = 0
788
    _first_thread_leaker_id = None
1185.16.14 by Martin Pool
- make TestCase._get_log work even if setup was aborted
789
    _log_file_name = None
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
790
    _log_contents = ''
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
791
    _keep_log_file = False
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
792
    # record lsprof data when performing benchmark calls.
793
    _gather_lsprof_in_benchmarks = False
3405.1.1 by Robert Collins
(robertc) Preserve test ids correctly to aid debugging. (Robert Collins, Andrew Bennetts)
794
    attrs_to_keep = ('id', '_testMethodName', '_testMethodDoc',
3224.4.9 by Andrew Bennetts
Add _TestCase__testMethodName to attrs_to_keep, for compatibility with Python 2.4.
795
                     '_log_contents', '_log_file_name', '_benchtime',
4313.1.1 by Michael Hudson
add _TestCase__testMethodDoc to attrs_to_keep, do attr saving more sanelyb
796
                     '_TestCase__testMethodName', '_TestCase__testMethodDoc',)
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
797
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
798
    def __init__(self, methodName='testMethod'):
799
        super(TestCase, self).__init__(methodName)
800
        self._cleanups = []
4153.1.1 by Andrew Bennetts
Check that TestCase.setUp was called in TestCase.run. If not, fail the test.
801
        self._bzr_test_setUp_run = False
4153.1.3 by Andrew Bennetts
Check that bzrlib.tests.TestCase.tearDown is called too.
802
        self._bzr_test_tearDown_run = False
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
803
        self._directory_isolation = True
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
804
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
805
    def setUp(self):
806
        unittest.TestCase.setUp(self)
4153.1.1 by Andrew Bennetts
Check that TestCase.setUp was called in TestCase.run. If not, fail the test.
807
        self._bzr_test_setUp_run = True
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
808
        self._cleanEnvironment()
2095.4.1 by Martin Pool
Better progress bars during tests
809
        self._silenceUI()
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
810
        self._startLogFile()
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
811
        self._benchcalls = []
1707.2.4 by Robert Collins
Teach the bzrlib TestCase to report the time take by calls to self.time as benchmark time, allowing granular reporting of time during benchmarks. See bzrlib.benchmarks.bench_add. (Robert Collins, Martin Pool)
812
        self._benchtime = None
2423.1.1 by Martin Pool
fix import order dependency that broke benchmarks
813
        self._clear_hooks()
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
814
        self._track_transports()
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
815
        self._track_locks()
2560.1.1 by Robert Collins
Make debug.debug_flags be isolated for all tests.
816
        self._clear_debug_flags()
3406.1.2 by Vincent Ladeuil
Fix as per Robert's review.
817
        TestCase._active_threads = threading.activeCount()
818
        self.addCleanup(self._check_leaked_threads)
819
4084.2.1 by Robert Collins
Make accessing a branch.tags.get_tag_dict use a smart[er] method rather than VFS calls and real objects.
820
    def debug(self):
821
        # debug a frame up.
822
        import pdb
823
        pdb.Pdb().set_trace(sys._getframe().f_back)
824
3406.1.2 by Vincent Ladeuil
Fix as per Robert's review.
825
    def _check_leaked_threads(self):
826
        active = threading.activeCount()
827
        leaked_threads = active - TestCase._active_threads
828
        TestCase._active_threads = active
4732.2.1 by Vincent Ladeuil
Clearer thread leaks reports.
829
        # If some tests make the number of threads *decrease*, we'll consider
830
        # that they are just observing old threads dieing, not agressively kill
831
        # random threads. So we don't report these tests as leaking. The risk
832
        # is that we have false positives that way (the test see 2 threads
833
        # going away but leak one) but it seems less likely than the actual
834
        # false positives (the test see threads going away and does not leak).
835
        if leaked_threads > 0:
3406.1.2 by Vincent Ladeuil
Fix as per Robert's review.
836
            TestCase._leaking_threads_tests += 1
837
            if TestCase._first_thread_leaker_id is None:
838
                TestCase._first_thread_leaker_id = self.id()
2560.1.1 by Robert Collins
Make debug.debug_flags be isolated for all tests.
839
840
    def _clear_debug_flags(self):
841
        """Prevent externally set debug flags affecting tests.
3882.6.20 by John Arbash Meinel
Clear out the InventoryEntry caches as part of the test suite.
842
2560.1.1 by Robert Collins
Make debug.debug_flags be isolated for all tests.
843
        Tests that want to use debug flags can just set them in the
844
        debug_flags set during setup/teardown.
845
        """
3731.3.1 by Andrew Bennetts
Make the test suite pass when -Eallow_debug is used.
846
        self._preserved_debug_flags = set(debug.debug_flags)
3390.1.1 by Andrew Bennetts
Add --debugflags/-E option to selftest.
847
        if 'allow_debug' not in selftest_debug_flags:
3302.2.1 by Andrew Bennetts
Add -Dselftest_debug debug flag.
848
            debug.debug_flags.clear()
4523.4.12 by John Arbash Meinel
Update the test_selftest tests so that they pass again.
849
        if 'disable_lock_checks' not in selftest_debug_flags:
850
            debug.debug_flags.add('strict_locks')
3731.3.1 by Andrew Bennetts
Make the test suite pass when -Eallow_debug is used.
851
        self.addCleanup(self._restore_debug_flags)
2423.1.1 by Martin Pool
fix import order dependency that broke benchmarks
852
853
    def _clear_hooks(self):
2245.1.1 by Robert Collins
New Branch hooks facility, with one initial hook 'set_rh' which triggers
854
        # prevent hooks affecting tests
4119.3.1 by Robert Collins
Create a single registry of all Hooks classes, removing the test suite knowledge of such hooks and allowing plugins to sensibly and safely define new hooks.
855
        self._preserved_hooks = {}
856
        for key, factory in hooks.known_hooks.items():
857
            parent, name = hooks.known_hooks_key_to_parent_and_attribute(key)
858
            current_hooks = hooks.known_hooks_key_to_object(key)
859
            self._preserved_hooks[parent] = (name, current_hooks)
2245.1.1 by Robert Collins
New Branch hooks facility, with one initial hook 'set_rh' which triggers
860
        self.addCleanup(self._restoreHooks)
4119.3.1 by Robert Collins
Create a single registry of all Hooks classes, removing the test suite knowledge of such hooks and allowing plugins to sensibly and safely define new hooks.
861
        for key, factory in hooks.known_hooks.items():
862
            parent, name = hooks.known_hooks_key_to_parent_and_attribute(key)
863
            setattr(parent, name, factory())
4160.2.4 by Andrew Bennetts
Use BzrDir pre_open hook to jail request code from accessing transports other than the backing transport.
864
        # this hook should always be installed
865
        request._install_hook()
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
866
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
867
    def disable_directory_isolation(self):
868
        """Turn off directory isolation checks."""
869
        self._directory_isolation = False
870
871
    def enable_directory_isolation(self):
872
        """Enable directory isolation checks."""
873
        self._directory_isolation = True
874
2095.4.1 by Martin Pool
Better progress bars during tests
875
    def _silenceUI(self):
876
        """Turn off UI for duration of test"""
877
        # by default the UI is off; tests can turn it on if they want it.
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
878
        saved = ui.ui_factory
2095.4.1 by Martin Pool
Better progress bars during tests
879
        def _restore():
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
880
            ui.ui_factory = saved
881
        ui.ui_factory = ui.SilentUIFactory()
2095.4.1 by Martin Pool
Better progress bars during tests
882
        self.addCleanup(_restore)
883
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
884
    def _check_locks(self):
885
        """Check that all lock take/release actions have been paired."""
4523.4.9 by John Arbash Meinel
Change the flags around a bit.
886
        # We always check for mismatched locks. If a mismatch is found, we
887
        # fail unless -Edisable_lock_checks is supplied to selftest, in which
888
        # case we just print a warning.
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
889
        # unhook:
890
        acquired_locks = [lock for action, lock in self._lock_actions
4327.1.4 by Vincent Ladeuil
Fix lock test failures by taking lock breaking into account.
891
                          if action == 'acquired']
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
892
        released_locks = [lock for action, lock in self._lock_actions
4327.1.4 by Vincent Ladeuil
Fix lock test failures by taking lock breaking into account.
893
                          if action == 'released']
894
        broken_locks = [lock for action, lock in self._lock_actions
895
                        if action == 'broken']
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
896
        # trivially, given the tests for lock acquistion and release, if we
4327.1.4 by Vincent Ladeuil
Fix lock test failures by taking lock breaking into account.
897
        # have as many in each list, it should be ok. Some lock tests also
898
        # break some locks on purpose and should be taken into account by
899
        # considering that breaking a lock is just a dirty way of releasing it.
900
        if len(acquired_locks) != (len(released_locks) + len(broken_locks)):
901
            message = ('Different number of acquired and '
902
                       'released or broken locks. (%s, %s + %s)' %
903
                       (acquired_locks, released_locks, broken_locks))
4523.4.9 by John Arbash Meinel
Change the flags around a bit.
904
            if not self._lock_check_thorough:
905
                # Rather than fail, just warn
906
                print "Broken test %s: %s" % (self, message)
907
                return
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
908
            self.fail(message)
909
910
    def _track_locks(self):
911
        """Track lock activity during tests."""
912
        self._lock_actions = []
4523.4.9 by John Arbash Meinel
Change the flags around a bit.
913
        if 'disable_lock_checks' in selftest_debug_flags:
914
            self._lock_check_thorough = False
915
        else:
916
            self._lock_check_thorough = True
917
            
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
918
        self.addCleanup(self._check_locks)
4327.1.4 by Vincent Ladeuil
Fix lock test failures by taking lock breaking into account.
919
        _mod_lock.Lock.hooks.install_named_hook('lock_acquired',
920
                                                self._lock_acquired, None)
921
        _mod_lock.Lock.hooks.install_named_hook('lock_released',
922
                                                self._lock_released, None)
923
        _mod_lock.Lock.hooks.install_named_hook('lock_broken',
924
                                                self._lock_broken, None)
3331.4.1 by Robert Collins
* -Dlock when passed to the selftest (e.g. ``bzr -Dlock selftest``) will
925
926
    def _lock_acquired(self, result):
927
        self._lock_actions.append(('acquired', result))
928
929
    def _lock_released(self, result):
930
        self._lock_actions.append(('released', result))
931
4327.1.4 by Vincent Ladeuil
Fix lock test failures by taking lock breaking into account.
932
    def _lock_broken(self, result):
933
        self._lock_actions.append(('broken', result))
934
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
935
    def permit_dir(self, name):
936
        """Permit a directory to be used by this test. See permit_url."""
937
        name_transport = get_transport(name)
938
        self.permit_url(name)
939
        self.permit_url(name_transport.base)
940
941
    def permit_url(self, url):
942
        """Declare that url is an ok url to use in this test.
943
        
944
        Do this for memory transports, temporary test directory etc.
945
        
946
        Do not do this for the current working directory, /tmp, or any other
947
        preexisting non isolated url.
948
        """
949
        if not url.endswith('/'):
950
            url += '/'
951
        self._bzr_selftest_roots.append(url)
952
953
    def permit_source_tree_branch_repo(self):
954
        """Permit the source tree bzr is running from to be opened.
955
956
        Some code such as bzrlib.version attempts to read from the bzr branch
957
        that bzr is executing from (if any). This method permits that directory
958
        to be used in the test suite.
959
        """
960
        path = self.get_source_path()
4691.2.5 by Robert Collins
Handle attempted directory access to the source tree by tracing rather than post-success inspection.
961
        self.record_directory_isolation()
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
962
        try:
963
            try:
4691.2.5 by Robert Collins
Handle attempted directory access to the source tree by tracing rather than post-success inspection.
964
                workingtree.WorkingTree.open(path)
4691.2.4 by Robert Collins
Handle NotBranchError when looking up source tree.
965
            except (errors.NotBranchError, errors.NoWorkingTree):
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
966
                return
967
        finally:
968
            self.enable_directory_isolation()
969
970
    def _preopen_isolate_transport(self, transport):
971
        """Check that all transport openings are done in the test work area."""
4634.43.20 by Andrew Bennetts
Merge from bzr.dev, resolving conflicts.
972
        while isinstance(transport, pathfilter.PathFilteringTransport):
973
            # Unwrap pathfiltered transports
974
            transport = transport.server.backing_transport.clone(
975
                transport._filter('.'))
976
        url = transport.base
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
977
        # ReadonlySmartTCPServer_for_testing decorates the backing transport
978
        # urls it is given by prepending readonly+. This is appropriate as the
979
        # client shouldn't know that the server is readonly (or not readonly).
980
        # We could register all servers twice, with readonly+ prepending, but
981
        # that makes for a long list; this is about the same but easier to
982
        # read.
983
        if url.startswith('readonly+'):
984
            url = url[len('readonly+'):]
985
        self._preopen_isolate_url(url)
986
987
    def _preopen_isolate_url(self, url):
988
        if not self._directory_isolation:
989
            return
4691.2.5 by Robert Collins
Handle attempted directory access to the source tree by tracing rather than post-success inspection.
990
        if self._directory_isolation == 'record':
991
            self._bzr_selftest_roots.append(url)
992
            return
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
993
        # This prevents all transports, including e.g. sftp ones backed on disk
994
        # from working unless they are explicitly granted permission. We then
995
        # depend on the code that sets up test transports to check that they are
996
        # appropriately isolated and enable their use by calling
997
        # self.permit_transport()
998
        if not osutils.is_inside_any(self._bzr_selftest_roots, url):
999
            raise errors.BzrError("Attempt to escape test isolation: %r %r"
1000
                % (url, self._bzr_selftest_roots))
1001
4691.2.5 by Robert Collins
Handle attempted directory access to the source tree by tracing rather than post-success inspection.
1002
    def record_directory_isolation(self):
1003
        """Gather accessed directories to permit later access.
1004
        
1005
        This is used for tests that access the branch bzr is running from.
1006
        """
1007
        self._directory_isolation = "record"
1008
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
1009
    def start_server(self, transport_server, backing_server=None):
1010
        """Start transport_server for this test.
1011
1012
        This starts the server, registers a cleanup for it and permits the
1013
        server's urls to be used.
1014
        """
1015
        if backing_server is None:
1016
            transport_server.setUp()
1017
        else:
4659.1.3 by Robert Collins
Review feedback.
1018
            transport_server.setUp(backing_server)
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
1019
        self.addCleanup(transport_server.tearDown)
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
1020
        # Obtain a real transport because if the server supplies a password, it
1021
        # will be hidden from the base on the client side.
1022
        t = get_transport(transport_server.get_url())
1023
        # Some transport servers effectively chroot the backing transport;
1024
        # others like SFTPServer don't - users of the transport can walk up the
1025
        # transport to read the entire backing transport. This wouldn't matter
1026
        # except that the workdir tests are given - and that they expect the
1027
        # server's url to point at - is one directory under the safety net. So
1028
        # Branch operations into the transport will attempt to walk up one
1029
        # directory. Chrooting all servers would avoid this but also mean that
1030
        # we wouldn't be testing directly against non-root urls. Alternatively
1031
        # getting the test framework to start the server with a backing server
1032
        # at the actual safety net directory would work too, but this then
1033
        # means that the self.get_url/self.get_transport methods would need
1034
        # to transform all their results. On balance its cleaner to handle it
1035
        # here, and permit a higher url when we have one of these transports.
1036
        if t.base.endswith('/work/'):
1037
            # we have safety net/test root/work
1038
            t = t.clone('../..')
1039
        elif isinstance(transport_server, server.SmartTCPServer_for_testing):
1040
            # The smart server adds a path similar to work, which is traversed
1041
            # up from by the client. But the server is chrooted - the actual
1042
            # backing transport is not escaped from, and VFS requests to the
1043
            # root will error (because they try to escape the chroot).
1044
            t2 = t.clone('..')
1045
            while t2.base != t.base:
1046
                t = t2
1047
                t2 = t.clone('..')
1048
        self.permit_url(t.base)
1049
1050
    def _track_transports(self):
1051
        """Install checks for transport usage."""
1052
        # TestCase has no safe place it can write to.
1053
        self._bzr_selftest_roots = []
1054
        # Currently the easiest way to be sure that nothing is going on is to
1055
        # hook into bzr dir opening. This leaves a small window of error for
1056
        # transport tests, but they are well known, and we can improve on this
1057
        # step.
1058
        bzrdir.BzrDir.hooks.install_named_hook("pre_open",
1059
            self._preopen_isolate_transport, "Check bzr directories are safe.")
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
1060
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
1061
    def _ndiff_strings(self, a, b):
1185.16.67 by Martin Pool
- assertEqualDiff handles strings without trailing newline
1062
        """Return ndiff between two strings containing lines.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1063
1185.16.67 by Martin Pool
- assertEqualDiff handles strings without trailing newline
1064
        A trailing newline is added if missing to make the strings
1065
        print properly."""
1066
        if b and b[-1] != '\n':
1067
            b += '\n'
1068
        if a and a[-1] != '\n':
1069
            a += '\n'
1185.16.21 by Martin Pool
- tweak diff shown by assertEqualDiff
1070
        difflines = difflib.ndiff(a.splitlines(True),
1071
                                  b.splitlines(True),
1072
                                  linejunk=lambda x: False,
1073
                                  charjunk=lambda x: False)
1074
        return ''.join(difflines)
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1075
2255.2.190 by Martin Pool
assertEqual can take an option message
1076
    def assertEqual(self, a, b, message=''):
2360.1.2 by John Arbash Meinel
Add an overzealous test, for Unicode support of _iter_changes.
1077
        try:
1078
            if a == b:
1079
                return
1080
        except UnicodeError, e:
1081
            # If we can't compare without getting a UnicodeError, then
1082
            # obviously they are different
1083
            mutter('UnicodeError: %s', e)
2255.2.190 by Martin Pool
assertEqual can take an option message
1084
        if message:
1085
            message += '\n'
1086
        raise AssertionError("%snot equal:\na = %s\nb = %s\n"
1087
            % (message,
2477.1.4 by Martin Pool
fix up indenting in pformat of inequalities displayed by test suite
1088
               pformat(a), pformat(b)))
2255.2.185 by Martin Pool
assertEqual uses pformat to show results
1089
1090
    assertEquals = assertEqual
1091
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
1092
    def assertEqualDiff(self, a, b, message=None):
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
1093
        """Assert two texts are equal, if not raise an exception.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1094
1095
        This is intended for use with multi-line strings where it can
1185.16.16 by Martin Pool
- add TestCase.assertEqualDiffs helper
1096
        be hard to find the differences by eye.
1097
        """
1098
        # TODO: perhaps override assertEquals to call this for strings?
1099
        if a == b:
1100
            return
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
1101
        if message is None:
1102
            message = "texts not equal:\n"
4680.1.1 by Vincent Ladeuil
Surprisingly, assertEqualDiff was wrong.
1103
        if a + '\n' == b:
1104
            message = 'first string is missing a final newline.\n'
3468.2.2 by Martin Pool
Better message re final newlines from assertEqualDiff
1105
        if a == b + '\n':
4680.1.2 by Vincent Ladeuil
Blessed be the tests that protect the imprudent :)
1106
            message = 'second string is missing a final newline.\n'
2555.3.3 by Martin Pool
Simple lock tracing in LockDir
1107
        raise AssertionError(message +
1108
                             self._ndiff_strings(a, b))
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1109
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
1110
    def assertEqualMode(self, mode, mode_test):
1111
        self.assertEqual(mode, mode_test,
1112
                         'mode mismatch %o != %o' % (mode, mode_test))
1113
4807.2.2 by John Arbash Meinel
Move all the stat comparison and platform checkning code to assertEqualStat.
1114
    def assertEqualStat(self, expected, actual):
3709.3.2 by Robert Collins
Race-free stat-fingerprint updating during commit via a new method get_file_with_stat.
1115
        """assert that expected and actual are the same stat result.
1116
1117
        :param expected: A stat result.
1118
        :param actual: A stat result.
1119
        :raises AssertionError: If the expected and actual stat values differ
1120
            other than by atime.
1121
        """
4789.24.1 by John Arbash Meinel
os.lstat() != os.fstat() for the st_ino field on windows
1122
        self.assertEqual(expected.st_size, actual.st_size,
1123
                         'st_size did not match')
1124
        self.assertEqual(expected.st_mtime, actual.st_mtime,
1125
                         'st_mtime did not match')
1126
        self.assertEqual(expected.st_ctime, actual.st_ctime,
1127
                         'st_ctime did not match')
4807.2.3 by John Arbash Meinel
It helps if you get the platform check right.
1128
        if sys.platform != 'win32':
4807.2.2 by John Arbash Meinel
Move all the stat comparison and platform checkning code to assertEqualStat.
1129
            # On Win32 both 'dev' and 'ino' cannot be trusted. In python2.4 it
1130
            # is 'dev' that varies, in python 2.5 (6?) it is st_ino that is
1131
            # odd. Regardless we shouldn't actually try to assert anything
1132
            # about their values
1133
            self.assertEqual(expected.st_dev, actual.st_dev,
1134
                             'st_dev did not match')
4789.24.1 by John Arbash Meinel
os.lstat() != os.fstat() for the st_ino field on windows
1135
            self.assertEqual(expected.st_ino, actual.st_ino,
1136
                             'st_ino did not match')
1137
        self.assertEqual(expected.st_mode, actual.st_mode,
1138
                         'st_mode did not match')
3709.3.2 by Robert Collins
Race-free stat-fingerprint updating during commit via a new method get_file_with_stat.
1139
4144.1.1 by Robert Collins
New assertLength method based on one Martin has squirreled away somewhere.
1140
    def assertLength(self, length, obj_with_len):
1141
        """Assert that obj_with_len is of length length."""
1142
        if len(obj_with_len) != length:
1143
            self.fail("Incorrect length: wanted %d, got %d for %r" % (
1144
                length, len(obj_with_len), obj_with_len))
1145
4634.85.11 by Andrew Bennetts
Suppress most errors from Branch.unlock too.
1146
    def assertLogsError(self, exception_class, func, *args, **kwargs):
1147
        """Assert that func(*args, **kwargs) quietly logs a specific exception.
1148
        """
1149
        from bzrlib import trace
1150
        captured = []
1151
        orig_log_exception_quietly = trace.log_exception_quietly
1152
        try:
1153
            def capture():
1154
                orig_log_exception_quietly()
1155
                captured.append(sys.exc_info())
1156
            trace.log_exception_quietly = capture
1157
            func(*args, **kwargs)
1158
        finally:
1159
            trace.log_exception_quietly = orig_log_exception_quietly
1160
        self.assertLength(1, captured)
1161
        err = captured[0][1]
1162
        self.assertIsInstance(err, exception_class)
1163
        return err
1164
2474.1.68 by John Arbash Meinel
Review feedback from Martin, mostly documentation updates.
1165
    def assertPositive(self, val):
1166
        """Assert that val is greater than 0."""
1167
        self.assertTrue(val > 0, 'expected a positive value, but got %s' % val)
1168
1169
    def assertNegative(self, val):
1170
        """Assert that val is less than 0."""
1171
        self.assertTrue(val < 0, 'expected a negative value, but got %s' % val)
1172
1185.31.40 by John Arbash Meinel
Added osutils.mkdtemp()
1173
    def assertStartsWith(self, s, prefix):
1174
        if not s.startswith(prefix):
1175
            raise AssertionError('string %r does not start with %r' % (s, prefix))
1176
1177
    def assertEndsWith(self, s, suffix):
1692.3.1 by Robert Collins
Fix push to work with just a branch, no need for a working tree.
1178
        """Asserts that s ends with suffix."""
1179
        if not s.endswith(suffix):
1185.31.40 by John Arbash Meinel
Added osutils.mkdtemp()
1180
            raise AssertionError('string %r does not end with %r' % (s, suffix))
1185.16.42 by Martin Pool
- Add assertContainsRe
1181
3940.1.1 by Ian Clatworthy
support flags when asserting re's in tests
1182
    def assertContainsRe(self, haystack, needle_re, flags=0):
1185.16.42 by Martin Pool
- Add assertContainsRe
1183
        """Assert that a contains something matching a regular expression."""
3940.1.1 by Ian Clatworthy
support flags when asserting re's in tests
1184
        if not re.search(needle_re, haystack, flags):
2555.3.1 by Martin Pool
Better messages from assertContainsRe
1185
            if '\n' in haystack or len(haystack) > 60:
1186
                # a long string, format it in a more readable way
1187
                raise AssertionError(
1188
                        'pattern "%s" not found in\n"""\\\n%s"""\n'
1189
                        % (needle_re, haystack))
1190
            else:
1191
                raise AssertionError('pattern "%s" not found in "%s"'
1192
                        % (needle_re, haystack))
1442.1.70 by Robert Collins
Add assertFileEqual to TestCaseInTempDir.
1193
3940.1.1 by Ian Clatworthy
support flags when asserting re's in tests
1194
    def assertNotContainsRe(self, haystack, needle_re, flags=0):
1185.84.3 by Aaron Bentley
Hide diffs for old revisions in bundles
1195
        """Assert that a does not match a regular expression"""
3940.1.1 by Ian Clatworthy
support flags when asserting re's in tests
1196
        if re.search(needle_re, haystack, flags):
1185.84.3 by Aaron Bentley
Hide diffs for old revisions in bundles
1197
            raise AssertionError('pattern "%s" found in "%s"'
1198
                    % (needle_re, haystack))
1199
1553.5.3 by Martin Pool
[patch] Rename TestCase.AssertSubset to assertSubset for consistency (Jan Hudec)
1200
    def assertSubset(self, sublist, superlist):
1185.46.8 by Aaron Bentley
bzr add reports ignored patterns.
1201
        """Assert that every entry in sublist is present in superlist."""
2695.1.4 by Martin Pool
Much faster assertSubset using sets, not O(n**2)
1202
        missing = set(sublist) - set(superlist)
1185.46.8 by Aaron Bentley
bzr add reports ignored patterns.
1203
        if len(missing) > 0:
2695.1.4 by Martin Pool
Much faster assertSubset using sets, not O(n**2)
1204
            raise AssertionError("value(s) %r not present in container %r" %
1185.46.8 by Aaron Bentley
bzr add reports ignored patterns.
1205
                                 (missing, superlist))
1206
2004.1.29 by v.ladeuil+lp at free
New tests for http range requests handling.
1207
    def assertListRaises(self, excClass, func, *args, **kwargs):
1208
        """Fail unless excClass is raised when the iterator from func is used.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1209
2004.1.29 by v.ladeuil+lp at free
New tests for http range requests handling.
1210
        Many functions can return generators this makes sure
1211
        to wrap them in a list() call to make sure the whole generator
1212
        is run, and that the proper exception is raised.
1213
        """
1214
        try:
1215
            list(func(*args, **kwargs))
3287.20.1 by John Arbash Meinel
Update assertListRaises so that it returns the exception.
1216
        except excClass, e:
1217
            return e
2004.1.29 by v.ladeuil+lp at free
New tests for http range requests handling.
1218
        else:
1219
            if getattr(excClass,'__name__', None) is not None:
1220
                excName = excClass.__name__
1221
            else:
1222
                excName = str(excClass)
1223
            raise self.failureException, "%s not raised" % excName
1224
2399.1.7 by John Arbash Meinel
Cleanup bzrlib/benchmarks/* so that everything at least has a valid doc string.
1225
    def assertRaises(self, excClass, callableObj, *args, **kwargs):
2323.5.5 by Martin Pool
override TestCase.assertRaises to return the exception
1226
        """Assert that a callable raises a particular exception.
1227
2323.5.9 by Martin Pool
Clear up assertRaises (r=robert)
1228
        :param excClass: As for the except statement, this may be either an
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1229
            exception class, or a tuple of classes.
2399.1.7 by John Arbash Meinel
Cleanup bzrlib/benchmarks/* so that everything at least has a valid doc string.
1230
        :param callableObj: A callable, will be passed ``*args`` and
1231
            ``**kwargs``.
2323.5.9 by Martin Pool
Clear up assertRaises (r=robert)
1232
2323.5.5 by Martin Pool
override TestCase.assertRaises to return the exception
1233
        Returns the exception so that you can examine it.
1234
        """
1235
        try:
2399.1.10 by John Arbash Meinel
fix assertRaises to use the right parameter...
1236
            callableObj(*args, **kwargs)
2323.5.5 by Martin Pool
override TestCase.assertRaises to return the exception
1237
        except excClass, e:
1238
            return e
1239
        else:
1240
            if getattr(excClass,'__name__', None) is not None:
1241
                excName = excClass.__name__
1242
            else:
2323.5.9 by Martin Pool
Clear up assertRaises (r=robert)
1243
                # probably a tuple
2323.5.5 by Martin Pool
override TestCase.assertRaises to return the exception
1244
                excName = str(excClass)
1245
            raise self.failureException, "%s not raised" % excName
1246
2220.1.13 by Marius Kruger
Remove assertNone
1247
    def assertIs(self, left, right, message=None):
1185.68.1 by Aaron Bentley
test transactions
1248
        if not (left is right):
2220.1.13 by Marius Kruger
Remove assertNone
1249
            if message is not None:
1250
                raise AssertionError(message)
1251
            else:
1252
                raise AssertionError("%r is not %r." % (left, right))
1253
1254
    def assertIsNot(self, left, right, message=None):
1255
        if (left is right):
1256
            if message is not None:
1257
                raise AssertionError(message)
1258
            else:
1259
                raise AssertionError("%r is %r." % (left, right))
2220.1.4 by Marius Kruger
* bzrlib/tests/__init__
1260
1530.1.21 by Robert Collins
Review feedback fixes.
1261
    def assertTransportMode(self, transport, path, mode):
4031.3.1 by Frank Aspell
Fixing various typos
1262
        """Fail if a path does not have mode "mode".
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1263
1651.1.3 by Martin Pool
Use transport._can_roundtrip_unix_modebits to decide whether to check transport results
1264
        If modes are not supported on this transport, the assertion is ignored.
1530.1.17 by Robert Collins
Move check_mode to TestCase.assertMode to make it generally accessible.
1265
        """
1651.1.3 by Martin Pool
Use transport._can_roundtrip_unix_modebits to decide whether to check transport results
1266
        if not transport._can_roundtrip_unix_modebits():
1530.1.17 by Robert Collins
Move check_mode to TestCase.assertMode to make it generally accessible.
1267
            return
1268
        path_stat = transport.stat(path)
1269
        actual_mode = stat.S_IMODE(path_stat.st_mode)
3508.1.22 by Vincent Ladeuil
Fix python2.4 failures.
1270
        self.assertEqual(mode, actual_mode,
3508.1.15 by Vincent Ladeuil
Tweak chmod bits output for easier debug.
1271
                         'mode of %r incorrect (%s != %s)'
1272
                         % (path, oct(mode), oct(actual_mode)))
1530.1.17 by Robert Collins
Move check_mode to TestCase.assertMode to make it generally accessible.
1273
2823.1.4 by Vincent Ladeuil
Use assertIsSameRealPath to avoid OSX aliasing (specifically /tmp
1274
    def assertIsSameRealPath(self, path1, path2):
1275
        """Fail if path1 and path2 points to different files"""
2823.1.11 by Vincent Ladeuil
Review feedback.
1276
        self.assertEqual(osutils.realpath(path1),
1277
                         osutils.realpath(path2),
1278
                         "apparent paths:\na = %s\nb = %s\n," % (path1, path2))
2823.1.4 by Vincent Ladeuil
Use assertIsSameRealPath to avoid OSX aliasing (specifically /tmp
1279
4449.3.12 by Martin Pool
assertIsInstance can take an extra message
1280
    def assertIsInstance(self, obj, kls, msg=None):
1281
        """Fail if obj is not an instance of kls
1282
        
1283
        :param msg: Supplementary message to show if the assertion fails.
1284
        """
1540.3.22 by Martin Pool
[patch] Add TestCase.assertIsInstance
1285
        if not isinstance(obj, kls):
4449.3.12 by Martin Pool
assertIsInstance can take an extra message
1286
            m = "%r is an instance of %s rather than %s" % (
1287
                obj, obj.__class__, kls)
1288
            if msg:
1289
                m += ": " + msg
1290
            self.fail(m)
1540.3.22 by Martin Pool
[patch] Add TestCase.assertIsInstance
1291
1551.13.10 by Aaron Bentley
Changes from review (poolie)
1292
    def expectFailure(self, reason, assertion, *args, **kwargs):
1551.13.9 by Aaron Bentley
Implement TestCase.expectFailure
1293
        """Invoke a test, expecting it to fail for the given reason.
1294
1551.13.10 by Aaron Bentley
Changes from review (poolie)
1295
        This is for assertions that ought to succeed, but currently fail.
1551.13.11 by Aaron Bentley
Update docs
1296
        (The failure is *expected* but not *wanted*.)  Please be very precise
1297
        about the failure you're expecting.  If a new bug is introduced,
1298
        AssertionError should be raised, not KnownFailure.
1299
1300
        Frequently, expectFailure should be followed by an opposite assertion.
1301
        See example below.
1551.13.10 by Aaron Bentley
Changes from review (poolie)
1302
1551.13.9 by Aaron Bentley
Implement TestCase.expectFailure
1303
        Intended to be used with a callable that raises AssertionError as the
1551.13.10 by Aaron Bentley
Changes from review (poolie)
1304
        'assertion' parameter.  args and kwargs are passed to the 'assertion'.
1551.13.9 by Aaron Bentley
Implement TestCase.expectFailure
1305
1306
        Raises KnownFailure if the test fails.  Raises AssertionError if the
1307
        test succeeds.
1308
1309
        example usage::
1551.13.11 by Aaron Bentley
Update docs
1310
1311
          self.expectFailure('Math is broken', self.assertNotEqual, 54,
1312
                             dynamic_val)
1313
          self.assertEqual(42, dynamic_val)
1314
1315
          This means that a dynamic_val of 54 will cause the test to raise
1316
          a KnownFailure.  Once math is fixed and the expectFailure is removed,
1317
          only a dynamic_val of 42 will allow the test to pass.  Anything other
1318
          than 54 or 42 will cause an AssertionError.
1551.13.9 by Aaron Bentley
Implement TestCase.expectFailure
1319
        """
1320
        try:
1551.13.10 by Aaron Bentley
Changes from review (poolie)
1321
            assertion(*args, **kwargs)
1551.13.9 by Aaron Bentley
Implement TestCase.expectFailure
1322
        except AssertionError:
1323
            raise KnownFailure(reason)
1324
        else:
1325
            self.fail('Unexpected success.  Should have failed: %s' % reason)
1326
3173.1.10 by Martin Pool
Move assertFileEqual to TestCase base class as it's generally usable
1327
    def assertFileEqual(self, content, path):
1328
        """Fail if path does not contain 'content'."""
1329
        self.failUnlessExists(path)
1330
        f = file(path, 'rb')
1331
        try:
1332
            s = f.read()
1333
        finally:
1334
            f.close()
1335
        self.assertEqualDiff(content, s)
1336
3173.1.12 by Martin Pool
Add test_push_log_file
1337
    def failUnlessExists(self, path):
1338
        """Fail unless path or paths, which may be abs or relative, exist."""
1339
        if not isinstance(path, basestring):
1340
            for p in path:
1341
                self.failUnlessExists(p)
1342
        else:
1343
            self.failUnless(osutils.lexists(path),path+" does not exist")
1344
1345
    def failIfExists(self, path):
1346
        """Fail if path or paths, which may be abs or relative, exist."""
1347
        if not isinstance(path, basestring):
1348
            for p in path:
1349
                self.failIfExists(p)
1350
        else:
1351
            self.failIf(osutils.lexists(path),path+" exists")
1352
2592.3.243 by Martin Pool
Rename TestCase._capture_warnings
1353
    def _capture_deprecation_warnings(self, a_callable, *args, **kwargs):
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1354
        """A helper for callDeprecated and applyDeprecated.
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1355
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1356
        :param a_callable: A callable to call.
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1357
        :param args: The positional arguments for the callable
1358
        :param kwargs: The keyword arguments for the callable
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1359
        :return: A tuple (warnings, result). result is the result of calling
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1360
            a_callable(``*args``, ``**kwargs``).
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1361
        """
1362
        local_warnings = []
2067.3.3 by Martin Pool
merge bzr.dev and reconcile several changes, also some test fixes
1363
        def capture_warnings(msg, cls=None, stacklevel=None):
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1364
            # we've hooked into a deprecation specific callpath,
1365
            # only deprecations should getting sent via it.
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1366
            self.assertEqual(cls, DeprecationWarning)
1367
            local_warnings.append(msg)
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1368
        original_warning_method = symbol_versioning.warn
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1369
        symbol_versioning.set_warning_method(capture_warnings)
1370
        try:
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1371
            result = a_callable(*args, **kwargs)
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1372
        finally:
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1373
            symbol_versioning.set_warning_method(original_warning_method)
1374
        return (local_warnings, result)
1375
1376
    def applyDeprecated(self, deprecation_format, a_callable, *args, **kwargs):
1377
        """Call a deprecated callable without warning the user.
1378
2530.3.3 by Martin Pool
Clean up some callers that use varargs syntax for run_bzr, but don't
1379
        Note that this only captures warnings raised by symbol_versioning.warn,
1380
        not other callers that go direct to the warning module.
1381
2697.2.2 by Martin Pool
deprecate Branch.append_revision
1382
        To test that a deprecated method raises an error, do something like
1383
        this::
1384
3400.3.6 by Martin Pool
Remove code deprecated prior to 1.1 and its tests
1385
            self.assertRaises(errors.ReservedId,
1386
                self.applyDeprecated,
1387
                deprecated_in((1, 5, 0)),
1388
                br.append_revision,
1389
                'current:')
2697.2.2 by Martin Pool
deprecate Branch.append_revision
1390
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1391
        :param deprecation_format: The deprecation format that the callable
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1392
            should have been deprecated with. This is the same type as the
1393
            parameter to deprecated_method/deprecated_function. If the
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1394
            callable is not deprecated with this format, an assertion error
1395
            will be raised.
1396
        :param a_callable: A callable to call. This may be a bound method or
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1397
            a regular function. It will be called with ``*args`` and
1398
            ``**kwargs``.
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1399
        :param args: The positional arguments for the callable
1400
        :param kwargs: The keyword arguments for the callable
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1401
        :return: The result of a_callable(``*args``, ``**kwargs``)
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1402
        """
2592.3.243 by Martin Pool
Rename TestCase._capture_warnings
1403
        call_warnings, result = self._capture_deprecation_warnings(a_callable,
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1404
            *args, **kwargs)
1405
        expected_first_warning = symbol_versioning.deprecation_string(
1406
            a_callable, deprecation_format)
1407
        if len(call_warnings) == 0:
2255.7.47 by Robert Collins
Improve applyDeprecated warning message.
1408
            self.fail("No deprecation warning generated by call to %s" %
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1409
                a_callable)
1410
        self.assertEqual(expected_first_warning, call_warnings[0])
1411
        return result
1412
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1413
    def callCatchWarnings(self, fn, *args, **kw):
1414
        """Call a callable that raises python warnings.
1415
1416
        The caller's responsible for examining the returned warnings.
1417
1418
        If the callable raises an exception, the exception is not
1419
        caught and propagates up to the caller.  In that case, the list
1420
        of warnings is not available.
1421
1422
        :returns: ([warning_object, ...], fn_result)
1423
        """
2592.3.246 by Andrew Bennetts
Override warnings.filters in callCatchWarnings, to insulate it from -Werror.
1424
        # XXX: This is not perfect, because it completely overrides the
1425
        # warnings filters, and some code may depend on suppressing particular
1426
        # warnings.  It's the easiest way to insulate ourselves from -Werror,
1427
        # though.  -- Andrew, 20071062
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1428
        wlist = []
3734.5.3 by Vincent Ladeuil
Martin's review feedback.
1429
        def _catcher(message, category, filename, lineno, file=None, line=None):
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1430
            # despite the name, 'message' is normally(?) a Warning subclass
1431
            # instance
1432
            wlist.append(message)
1433
        saved_showwarning = warnings.showwarning
2592.3.246 by Andrew Bennetts
Override warnings.filters in callCatchWarnings, to insulate it from -Werror.
1434
        saved_filters = warnings.filters
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1435
        try:
1436
            warnings.showwarning = _catcher
2592.3.246 by Andrew Bennetts
Override warnings.filters in callCatchWarnings, to insulate it from -Werror.
1437
            warnings.filters = []
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1438
            result = fn(*args, **kw)
1439
        finally:
1440
            warnings.showwarning = saved_showwarning
2592.3.246 by Andrew Bennetts
Override warnings.filters in callCatchWarnings, to insulate it from -Werror.
1441
            warnings.filters = saved_filters
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1442
        return wlist, result
1443
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1444
    def callDeprecated(self, expected, callable, *args, **kwargs):
1445
        """Assert that a callable is deprecated in a particular way.
1446
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1447
        This is a very precise test for unusual requirements. The
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1448
        applyDeprecated helper function is probably more suited for most tests
1449
        as it allows you to simply specify the deprecation format being used
1450
        and will ensure that that is issued for the function being called.
1451
2530.3.3 by Martin Pool
Clean up some callers that use varargs syntax for run_bzr, but don't
1452
        Note that this only captures warnings raised by symbol_versioning.warn,
2592.3.242 by Martin Pool
New method TestCase.call_catch_warnings
1453
        not other callers that go direct to the warning module.  To catch
1454
        general warnings, use callCatchWarnings.
2530.3.3 by Martin Pool
Clean up some callers that use varargs syntax for run_bzr, but don't
1455
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1456
        :param expected: a list of the deprecation warnings expected, in order
1457
        :param callable: The callable to call
1458
        :param args: The positional arguments for the callable
1459
        :param kwargs: The keyword arguments for the callable
1460
        """
2592.3.243 by Martin Pool
Rename TestCase._capture_warnings
1461
        call_warnings, result = self._capture_deprecation_warnings(callable,
1982.3.2 by Robert Collins
New TestCase helper applyDeprecated. This allows you to call a callable
1462
            *args, **kwargs)
1463
        self.assertEqual(expected, call_warnings)
1910.2.9 by Aaron Bentley
Inroduce assertDeprecated, and use it to test old commitbuilder API
1464
        return result
1465
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
1466
    def _startLogFile(self):
1467
        """Send bzr and test log messages to a temporary file.
1468
1469
        The file is removed as the test is torn down.
1470
        """
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1471
        fileno, name = tempfile.mkstemp(suffix='.log', prefix='testbzr')
1948.1.9 by John Arbash Meinel
Change mutter() so that it doesn't try so hard to write out perfect utf8, instead, rather than using a utf8 file, it changes unicode to utf8 manually
1472
        self._log_file = os.fdopen(fileno, 'w+')
3173.1.9 by Martin Pool
tests should now call push/pop_test_log
1473
        self._log_memento = bzrlib.trace.push_log_file(self._log_file)
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1474
        self._log_file_name = name
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
1475
        self.addCleanup(self._finishLogFile)
1476
1477
    def _finishLogFile(self):
1478
        """Finished with the log file.
1479
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1480
        Close the file and delete it, unless setKeepLogfile was called.
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
1481
        """
1711.7.10 by John Arbash Meinel
Nothing to clean up if self._log_file is None
1482
        if self._log_file is None:
1483
            return
3173.1.9 by Martin Pool
tests should now call push/pop_test_log
1484
        bzrlib.trace.pop_log_file(self._log_memento)
1185.16.122 by Martin Pool
[patch] Close test log file before deleting, needed on Windows
1485
        self._log_file.close()
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1486
        self._log_file = None
2024.2.9 by John Arbash Meinel
Bring back log deletions
1487
        if not self._keep_log_file:
1488
            os.remove(self._log_file_name)
1489
            self._log_file_name = None
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1490
1491
    def setKeepLogfile(self):
1492
        """Make the logfile not be deleted when _finishLogFile is called."""
1493
        self._keep_log_file = True
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1494
4523.4.12 by John Arbash Meinel
Update the test_selftest tests so that they pass again.
1495
    def thisFailsStrictLockCheck(self):
4523.4.11 by John Arbash Meinel
Update the tests, adding a test for -Edisable_lock_checks.
1496
        """It is known that this test would fail with -Dstrict_locks.
1497
1498
        By default, all tests are run with strict lock checking unless
1499
        -Edisable_lock_checks is supplied. However there are some tests which
1500
        we know fail strict locks at this point that have not been fixed.
1501
        They should call this function to disable the strict checking.
4523.4.12 by John Arbash Meinel
Update the test_selftest tests so that they pass again.
1502
1503
        This should be used sparingly, it is much better to fix the locking
1504
        issues rather than papering over the problem by calling this function.
4523.4.11 by John Arbash Meinel
Update the tests, adding a test for -Edisable_lock_checks.
1505
        """
1506
        debug.debug_flags.discard('strict_locks')
1507
3620.1.1 by Jonathan Lange
Let addCleanup take arguments that are then passed to the cleanup.
1508
    def addCleanup(self, callable, *args, **kwargs):
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1509
        """Arrange to run a callable when this case is torn down.
1510
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1511
        Callables are run in the reverse of the order they are registered,
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1512
        ie last-in first-out.
1513
        """
3620.1.1 by Jonathan Lange
Let addCleanup take arguments that are then passed to the cleanup.
1514
        self._cleanups.append((callable, args, kwargs))
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1515
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
1516
    def _cleanEnvironment(self):
1185.38.3 by John Arbash Meinel
Refactored environment cleaning code
1517
        new_env = {
2055.1.2 by John Arbash Meinel
Clean up BZR_HOME in ENV for the test suite
1518
            'BZR_HOME': None, # Don't inherit BZR_HOME to all the tests.
1185.38.3 by John Arbash Meinel
Refactored environment cleaning code
1519
            'HOME': os.getcwd(),
3638.4.4 by Mark Hammond
fix typo in comments
1520
            # bzr now uses the Win32 API and doesn't rely on APPDATA, but the
3638.4.1 by Mark Hammond
Add win32utils.get_local_appdata_location() so bzr and plugins can
1521
            # tests do check our impls match APPDATA
2839.6.2 by Alexander Belchenko
changes after Martin's review
1522
            'BZR_EDITOR': None, # test_msgeditor manipulates this variable
4183.4.1 by Vincent Ladeuil
Fix bug #347130 by properly isolating test_msgeditor.MsgEditorTest tests.
1523
            'VISUAL': None,
1524
            'EDITOR': None,
1861.4.1 by Matthieu Moy
BZREMAIL renamed to BZR_EMAIL.
1525
            'BZR_EMAIL': None,
1912.2.1 by Adeodato Simó
Clear $BZREMAIL in tests, not only the newer $BZR_EMAIL, since the
1526
            'BZREMAIL': None, # may still be present in the environment
1185.38.3 by John Arbash Meinel
Refactored environment cleaning code
1527
            'EMAIL': None,
1963.1.6 by John Arbash Meinel
Use the new helper function in a few places
1528
            'BZR_PROGRESS_BAR': None,
3193.6.1 by Alexander Belchenko
``BZR_LOG=disable`` suppresses writing messages to .bzr.log.
1529
            'BZR_LOG': None,
3835.2.7 by Aaron Bentley
Add tests for plugins
1530
            'BZR_PLUGIN_PATH': None,
4766.3.7 by Vincent Ladeuil
Mix BZR_CONCURRENCY and --concurrency so both are available.
1531
            'BZR_CONCURRENCY': None,
4449.3.2 by Martin Pool
Also set TERM, COLUMNS, LINES for test cases
1532
            # Make sure that any text ui tests are consistent regardless of
1533
            # the environment the test case is run in; you may want tests that
1534
            # test other combinations.  'dumb' is a reasonable guess for tests
1535
            # going to a pipe or a StringIO.
1536
            'TERM': 'dumb',
1537
            'LINES': '25',
1538
            'COLUMNS': '80',
4747.3.7 by Vincent Ladeuil
Introduce BZR_COLUMNS since COLUMNS behaviour is too obscure.
1539
            'BZR_COLUMNS': '80',
2617.2.1 by Jelmer Vernooij
Sanitize SSH_AUTH_SOCK environment variable (#125955).
1540
            # SSH Agent
1541
            'SSH_AUTH_SOCK': None,
2167.3.6 by v.ladeuil+lp at free
Take John's comments into account and add more tests.
1542
            # Proxies
1543
            'http_proxy': None,
1544
            'HTTP_PROXY': None,
1545
            'https_proxy': None,
1546
            'HTTPS_PROXY': None,
1547
            'no_proxy': None,
1548
            'NO_PROXY': None,
1549
            'all_proxy': None,
1550
            'ALL_PROXY': None,
4229.2.1 by Vincent Ladeuil
Desambiguate comment before it became totally obscure.
1551
            # Nobody cares about ftp_proxy, FTP_PROXY AFAIK. So far at
2167.3.6 by v.ladeuil+lp at free
Take John's comments into account and add more tests.
1552
            # least. If you do (care), please update this comment
4229.2.1 by Vincent Ladeuil
Desambiguate comment before it became totally obscure.
1553
            # -- vila 20080401
2167.3.6 by v.ladeuil+lp at free
Take John's comments into account and add more tests.
1554
            'ftp_proxy': None,
1555
            'FTP_PROXY': None,
2512.3.1 by Daniel Watkins
Added BZR_REMOTE_PATH to the list of environment variables cleaned before running tests, fixing bug #111958.
1556
            'BZR_REMOTE_PATH': None,
1185.38.3 by John Arbash Meinel
Refactored environment cleaning code
1557
        }
1185.38.4 by John Arbash Meinel
Making old_env a private member
1558
        self.__old_env = {}
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
1559
        self.addCleanup(self._restoreEnvironment)
1185.38.3 by John Arbash Meinel
Refactored environment cleaning code
1560
        for name, value in new_env.iteritems():
1561
            self._captureVar(name, value)
1562
1563
    def _captureVar(self, name, newvalue):
1963.1.6 by John Arbash Meinel
Use the new helper function in a few places
1564
        """Set an environment variable, and reset it when finished."""
1565
        self.__old_env[name] = osutils.set_or_unset_env(name, newvalue)
1185.38.2 by John Arbash Meinel
[patch] Aaron Bentley's HOME fix.
1566
2560.1.1 by Robert Collins
Make debug.debug_flags be isolated for all tests.
1567
    def _restore_debug_flags(self):
2560.1.3 by Robert Collins
Allow 'from debug import debug_flags'
1568
        debug.debug_flags.clear()
1569
        debug.debug_flags.update(self._preserved_debug_flags)
2560.1.1 by Robert Collins
Make debug.debug_flags be isolated for all tests.
1570
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
1571
    def _restoreEnvironment(self):
1185.38.4 by John Arbash Meinel
Making old_env a private member
1572
        for name, value in self.__old_env.iteritems():
1963.1.6 by John Arbash Meinel
Use the new helper function in a few places
1573
            osutils.set_or_unset_env(name, value)
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
1574
2245.1.1 by Robert Collins
New Branch hooks facility, with one initial hook 'set_rh' which triggers
1575
    def _restoreHooks(self):
4119.3.1 by Robert Collins
Create a single registry of all Hooks classes, removing the test suite knowledge of such hooks and allowing plugins to sensibly and safely define new hooks.
1576
        for klass, (name, hooks) in self._preserved_hooks.items():
1577
            setattr(klass, name, hooks)
2245.1.1 by Robert Collins
New Branch hooks facility, with one initial hook 'set_rh' which triggers
1578
2367.1.2 by Robert Collins
Some minor cleanups of test code, and implement KnownFailure support as
1579
    def knownFailure(self, reason):
1580
        """This test has failed for some known reason."""
1581
        raise KnownFailure(reason)
1582
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1583
    def _do_skip(self, result, reason):
1584
        addSkip = getattr(result, 'addSkip', None)
1585
        if not callable(addSkip):
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1586
            result.addSuccess(result)
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1587
        else:
1588
            addSkip(self, reason)
1589
4780.1.4 by Robert Collins
Switch reporting of KnownFailure to be Python2.7 compatible.
1590
    def _do_known_failure(self, result):
1591
        err = sys.exc_info()
1592
        addExpectedFailure = getattr(result, 'addExpectedFailure', None)
1593
        if addExpectedFailure is not None:
1594
            addExpectedFailure(self, err)
1595
        else:
1596
            result.addSuccess(self)
1597
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
1598
    def _do_not_applicable(self, result, e):
4780.1.5 by Robert Collins
Fix fallback of NotApplicable to Skip.
1599
        if not e.args:
1600
            reason = 'No reason given'
1601
        else:
1602
            reason = e.args[0]
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
1603
        addNotApplicable = getattr(result, 'addNotApplicable', None)
1604
        if addNotApplicable is not None:
1605
            result.addNotApplicable(self, reason)
1606
        else:
1607
            self._do_skip(result, reason)
1608
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1609
    def _do_unsupported_or_skip(self, result, reason):
1610
        addNotSupported = getattr(result, 'addNotSupported', None)
1611
        if addNotSupported is not None:
1612
            result.addNotSupported(self, reason)
1613
        else:
1614
            self._do_skip(result, reason)
1615
2367.1.3 by Robert Collins
Add support for calling addNotSupported on TestResults to bzr TestCase's
1616
    def run(self, result=None):
1617
        if result is None: result = self.defaultTestResult()
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1618
        result.startTest(self)
1619
        try:
1620
            self._run(result)
1621
            return result
1622
        finally:
1623
            result.stopTest(self)
1624
1625
    def _run(self, result):
2367.1.3 by Robert Collins
Add support for calling addNotSupported on TestResults to bzr TestCase's
1626
        for feature in getattr(self, '_test_needs_features', []):
1627
            if not feature.available():
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1628
                return self._do_unsupported_or_skip(result, feature)
3224.4.1 by Andrew Bennetts
Prune __dict__ of TestCases after they have run to save memory.
1629
        try:
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1630
            absent_attr = object()
1631
            # Python 2.5
1632
            method_name = getattr(self, '_testMethodName', absent_attr)
1633
            if method_name is absent_attr:
1634
                # Python 2.4
1635
                method_name = getattr(self, '_TestCase__testMethodName')
1636
            testMethod = getattr(self, method_name)
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1637
            try:
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1638
                try:
1639
                    self.setUp()
1640
                    if not self._bzr_test_setUp_run:
1641
                        self.fail(
1642
                            "test setUp did not invoke "
1643
                            "bzrlib.tests.TestCase's setUp")
1644
                except KeyboardInterrupt:
1645
                    self._runCleanups()
1646
                    raise
4780.1.4 by Robert Collins
Switch reporting of KnownFailure to be Python2.7 compatible.
1647
                except KnownFailure:
1648
                    self._do_known_failure(result)
1649
                    self.tearDown()
1650
                    return
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
1651
                except TestNotApplicable, e:
1652
                    self._do_not_applicable(result, e)
1653
                    self.tearDown()
1654
                    return
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1655
                except TestSkipped, e:
1656
                    self._do_skip(result, e.args[0])
1657
                    self.tearDown()
1658
                    return result
1659
                except UnavailableFeature, e:
1660
                    self._do_unsupported_or_skip(result, e.args[0])
1661
                    self.tearDown()
1662
                    return
1663
                except:
1664
                    result.addError(self, sys.exc_info())
1665
                    self._runCleanups()
1666
                    return result
1667
1668
                ok = False
1669
                try:
1670
                    testMethod()
1671
                    ok = True
4780.1.4 by Robert Collins
Switch reporting of KnownFailure to be Python2.7 compatible.
1672
                except KnownFailure:
1673
                    self._do_known_failure(result)
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1674
                except self.failureException:
1675
                    result.addFailure(self, sys.exc_info())
4780.1.3 by Robert Collins
TestNotApplicable handling improved for compatibility with stdlib TestResult objects.
1676
                except TestNotApplicable, e:
1677
                    self._do_not_applicable(result, e)
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1678
                except TestSkipped, e:
1679
                    if not e.args:
1680
                        reason = "No reason given."
1681
                    else:
1682
                        reason = e.args[0]
1683
                    self._do_skip(result, reason)
1684
                except UnavailableFeature, e:
1685
                    self._do_unsupported_or_skip(result, e.args[0])
1686
                except KeyboardInterrupt:
1687
                    self._runCleanups()
1688
                    raise
1689
                except:
1690
                    result.addError(self, sys.exc_info())
1691
1692
                try:
1693
                    self.tearDown()
1694
                    if not self._bzr_test_tearDown_run:
1695
                        self.fail(
1696
                            "test tearDown did not invoke "
1697
                            "bzrlib.tests.TestCase's tearDown")
1698
                except KeyboardInterrupt:
1699
                    self._runCleanups()
1700
                    raise
1701
                except:
1702
                    result.addError(self, sys.exc_info())
1703
                    self._runCleanups()
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1704
                    ok = False
4780.1.1 by Robert Collins
Make addUnsupported more compatible with other TestResults.
1705
                if ok: result.addSuccess(self)
4314.2.1 by Robert Collins
Update lock debugging support patch.
1706
                return result
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1707
            except KeyboardInterrupt:
4314.2.3 by Robert Collins
Scatter a few _runCleanups in TestCase.run
1708
                self._runCleanups()
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1709
                raise
3224.4.1 by Andrew Bennetts
Prune __dict__ of TestCases after they have run to save memory.
1710
        finally:
1711
            saved_attrs = {}
1712
            for attr_name in self.attrs_to_keep:
4313.1.1 by Michael Hudson
add _TestCase__testMethodDoc to attrs_to_keep, do attr saving more sanelyb
1713
                if attr_name in self.__dict__:
1714
                    saved_attrs[attr_name] = self.__dict__[attr_name]
3224.4.1 by Andrew Bennetts
Prune __dict__ of TestCases after they have run to save memory.
1715
            self.__dict__ = saved_attrs
2367.1.3 by Robert Collins
Add support for calling addNotSupported on TestResults to bzr TestCase's
1716
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
1717
    def tearDown(self):
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1718
        self._runCleanups()
4063.1.1 by Robert Collins
Move skipped test detection to TestCase, and make reporting use an addSkip method as per testtools.
1719
        self._log_contents = ''
4314.2.3 by Robert Collins
Scatter a few _runCleanups in TestCase.run
1720
        self._bzr_test_tearDown_run = True
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1721
        unittest.TestCase.tearDown(self)
1722
1707.2.4 by Robert Collins
Teach the bzrlib TestCase to report the time take by calls to self.time as benchmark time, allowing granular reporting of time during benchmarks. See bzrlib.benchmarks.bench_add. (Robert Collins, Martin Pool)
1723
    def time(self, callable, *args, **kwargs):
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
1724
        """Run callable and accrue the time it takes to the benchmark time.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1725
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
1726
        If lsprofiling is enabled (i.e. by --lsprof-time to bzr selftest) then
1727
        this will cause lsprofile statistics to be gathered and stored in
1728
        self._benchcalls.
1729
        """
1707.2.4 by Robert Collins
Teach the bzrlib TestCase to report the time take by calls to self.time as benchmark time, allowing granular reporting of time during benchmarks. See bzrlib.benchmarks.bench_add. (Robert Collins, Martin Pool)
1730
        if self._benchtime is None:
1731
            self._benchtime = 0
1732
        start = time.time()
1733
        try:
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
1734
            if not self._gather_lsprof_in_benchmarks:
1735
                return callable(*args, **kwargs)
1736
            else:
1737
                # record this benchmark
1738
                ret, stats = bzrlib.lsprof.profile(callable, *args, **kwargs)
1739
                stats.sort()
1740
                self._benchcalls.append(((callable, args, kwargs), stats))
1741
                return ret
1707.2.4 by Robert Collins
Teach the bzrlib TestCase to report the time take by calls to self.time as benchmark time, allowing granular reporting of time during benchmarks. See bzrlib.benchmarks.bench_add. (Robert Collins, Martin Pool)
1742
        finally:
1743
            self._benchtime += time.time() - start
1744
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1745
    def _runCleanups(self):
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1746
        """Run registered cleanup functions.
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1747
1748
        This should only be called from TestCase.tearDown.
1749
        """
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1750
        # TODO: Perhaps this should keep running cleanups even if
1541 by Martin Pool
doc
1751
        # one of them fails?
2338.4.10 by Marien Zwart
Make a test skipped from setUp run tearDown again. Make calling _runCleanups twice safe. Clean up tests.
1752
1753
        # Actually pop the cleanups from the list so tearDown running
1754
        # twice is safe (this happens for skipped tests).
1755
        while self._cleanups:
3620.1.1 by Jonathan Lange
Let addCleanup take arguments that are then passed to the cleanup.
1756
            cleanup, args, kwargs = self._cleanups.pop()
1757
            cleanup(*args, **kwargs)
1185.16.108 by mbp at sourcefrog
Add TestCase.addCleanup method.
1758
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1759
    def log(self, *args):
1185.43.1 by Martin Pool
Remove direct logging calls from selftest
1760
        mutter(*args)
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1761
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1762
    def _get_log(self, keep_log_file=False):
2725.1.1 by Robert Collins
Add -Devil flag to highlight the use of problematic API calls.
1763
        """Get the log from bzrlib.trace calls from this test.
1764
1765
        :param keep_log_file: When True, if the log is still a file on disk
1766
            leave it as a file on disk. When False, if the log is still a file
1767
            on disk, the log file is deleted and the log preserved as
1768
            self._log_contents.
1769
        :return: A string containing the log.
1770
        """
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1771
        # flush the log file, to get all content
1772
        import bzrlib.trace
3778.4.1 by Martin Pool
test _get_log method should cope if .bzr.log is not open
1773
        if bzrlib.trace._trace_file:
1774
            bzrlib.trace._trace_file.flush()
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1775
        if self._log_contents:
3199.1.1 by Vincent Ladeuil
Get rid of ~1000 useless log files out of 10.000 tests in /tmp.
1776
            # XXX: this can hardly contain the content flushed above --vila
1777
            # 20080128
1185.16.109 by mbp at sourcefrog
Clean up test log files when tests complete.
1778
            return self._log_contents
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1779
        if self._log_file_name is not None:
1780
            logfile = open(self._log_file_name)
1781
            try:
1782
                log_contents = logfile.read()
1783
            finally:
1784
                logfile.close()
1785
            if not keep_log_file:
1786
                self._log_contents = log_contents
2309.2.7 by Alexander Belchenko
Skip permission denied (on win32) during selftest cleanup
1787
                try:
1788
                    os.remove(self._log_file_name)
1789
                except OSError, e:
1790
                    if sys.platform == 'win32' and e.errno == errno.EACCES:
2911.6.1 by Blake Winton
Change 'print >> f,'s to 'f.write('s.
1791
                        sys.stderr.write(('Unable to delete log file '
1792
                                             ' %r\n' % self._log_file_name))
2309.2.7 by Alexander Belchenko
Skip permission denied (on win32) during selftest cleanup
1793
                    else:
1794
                        raise
1927.3.1 by Carl Friedrich Bolz
Throw away on-disk logfile when possible.
1795
            return log_contents
1796
        else:
1797
            return "DELETED log file to reduce memory footprint"
1185.3.18 by Martin Pool
- add new helper TestBase.run_bzr_captured
1798
2367.1.6 by Robert Collins
Allow per-test-fixture feature requirements via 'requireFeature'.(Robert Collins)
1799
    def requireFeature(self, feature):
1800
        """This test requires a specific feature is available.
1801
1802
        :raises UnavailableFeature: When feature is not available.
1803
        """
1804
        if not feature.available():
1805
            raise UnavailableFeature(feature)
1806
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1807
    def _run_bzr_autosplit(self, args, retcode, encoding, stdin,
1808
            working_dir):
2530.3.3 by Martin Pool
Clean up some callers that use varargs syntax for run_bzr, but don't
1809
        """Run bazaar command line, splitting up a string command line."""
1810
        if isinstance(args, basestring):
2804.4.1 by Alexander Belchenko
some win32-specific fixes for selftest
1811
            # shlex don't understand unicode strings,
1812
            # so args should be plain string (bialix 20070906)
1813
            args = list(shlex.split(str(args)))
2530.3.3 by Martin Pool
Clean up some callers that use varargs syntax for run_bzr, but don't
1814
        return self._run_bzr_core(args, retcode=retcode,
1815
                encoding=encoding, stdin=stdin, working_dir=working_dir,
1816
                )
1817
1818
    def _run_bzr_core(self, args, retcode, encoding, stdin,
1819
            working_dir):
4634.90.2 by Andrew Bennetts
Clear chk_map page cache in TestCase._run_bzr_core, causes blackbox.test_log to fail without fix in previous revision.
1820
        # Clear chk_map page cache, because the contents are likely to mask
1821
        # locking errors.
1822
        chk_map.clear_cache()
1185.85.8 by John Arbash Meinel
Adding wrapper for sys.stdout so we can set the output encoding. Adding tests that 'bzr log' handles multiple encodings properly
1823
        if encoding is None:
3224.5.4 by Andrew Bennetts
Fix test suite, mainly weeding out uses of bzrlib.user_encoding.
1824
            encoding = osutils.get_user_encoding()
1185.85.8 by John Arbash Meinel
Adding wrapper for sys.stdout so we can set the output encoding. Adding tests that 'bzr log' handles multiple encodings properly
1825
        stdout = StringIOWrapper()
1826
        stderr = StringIOWrapper()
1827
        stdout.encoding = encoding
1828
        stderr.encoding = encoding
1829
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1830
        self.log('run bzr: %r', args)
1185.43.5 by Martin Pool
Update log message quoting
1831
        # FIXME: don't call into logging here
1185.3.20 by Martin Pool
- run_bzr_captured also includes logged errors in
1832
        handler = logging.StreamHandler(stderr)
1833
        handler.setLevel(logging.INFO)
1834
        logger = logging.getLogger('')
1835
        logger.addHandler(handler)
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
1836
        old_ui_factory = ui.ui_factory
1837
        ui.ui_factory = TestUIFactory(stdin=stdin, stdout=stdout, stderr=stderr)
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
1838
1839
        cwd = None
1840
        if working_dir is not None:
1841
            cwd = osutils.getcwd()
1842
            os.chdir(working_dir)
1843
1185.3.20 by Martin Pool
- run_bzr_captured also includes logged errors in
1844
        try:
4815.3.5 by Gordon Tyler
Fixed try/except/finally to try/try/except/finally for Python 2.4 compatibility.
1845
            try:
1846
                result = self.apply_redirected(ui.ui_factory.stdin,
1847
                    stdout, stderr,
1848
                    bzrlib.commands.run_bzr_catch_user_errors,
1849
                    args)
1850
            except KeyboardInterrupt:
4815.3.8 by Vincent Ladeuil
Fix too long lines.
1851
                # Reraise KeyboardInterrupt with contents of redirected stdout
1852
                # and stderr as arguments, for tests which are interested in
1853
                # stdout and stderr and are expecting the exception.
4815.3.5 by Gordon Tyler
Fixed try/except/finally to try/try/except/finally for Python 2.4 compatibility.
1854
                out = stdout.getvalue()
1855
                err = stderr.getvalue()
1856
                if out:
1857
                    self.log('output:\n%r', out)
1858
                if err:
1859
                    self.log('errors:\n%r', err)
1860
                raise KeyboardInterrupt(out, err)
1185.3.20 by Martin Pool
- run_bzr_captured also includes logged errors in
1861
        finally:
1862
            logger.removeHandler(handler)
2294.4.4 by Vincent Ladeuil
Provide a better implementation for testing passwords.
1863
            ui.ui_factory = old_ui_factory
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
1864
            if cwd is not None:
1865
                os.chdir(cwd)
1685.1.69 by Wouter van Heyst
merge bzr.dev 1740
1866
1185.3.18 by Martin Pool
- add new helper TestBase.run_bzr_captured
1867
        out = stdout.getvalue()
1868
        err = stderr.getvalue()
1869
        if out:
1185.85.72 by John Arbash Meinel
Fix some of the tests.
1870
            self.log('output:\n%r', out)
1185.3.18 by Martin Pool
- add new helper TestBase.run_bzr_captured
1871
        if err:
1185.85.72 by John Arbash Meinel
Fix some of the tests.
1872
            self.log('errors:\n%r', err)
1185.3.18 by Martin Pool
- add new helper TestBase.run_bzr_captured
1873
        if retcode is not None:
2292.1.32 by Marius Kruger
* tests/__init__.run_bzr
1874
            self.assertEquals(retcode, result,
1875
                              message='Unexpected return code')
4665.5.15 by Vincent Ladeuil
Catch the retcode for all commands.
1876
        return result, out, err
1185.3.18 by Martin Pool
- add new helper TestBase.run_bzr_captured
1877
2830.2.5 by Martin Pool
Deprecated ``run_bzr_decode``; use the new ``output_encoding`` parameter to
1878
    def run_bzr(self, args, retcode=0, encoding=None, stdin=None,
1879
                working_dir=None, error_regexes=[], output_encoding=None):
1119 by Martin Pool
doc
1880
        """Invoke bzr, as if it were run from the command line.
1881
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1882
        The argument list should not include the bzr program name - the
1883
        first argument is normally the bzr command.  Arguments may be
1884
        passed in three ways:
1885
1886
        1- A list of strings, eg ["commit", "a"].  This is recommended
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1887
        when the command contains whitespace or metacharacters, or
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1888
        is built up at run time.
1889
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1890
        2- A single string, eg "add a".  This is the most convenient
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1891
        for hardcoded commands.
1892
2530.3.4 by Martin Pool
Deprecate run_bzr_captured in favour of just run_bzr
1893
        This runs bzr through the interface that catches and reports
1894
        errors, and with logging set to something approximating the
1895
        default, so that error reporting can be checked.
1896
1119 by Martin Pool
doc
1897
        This should be the main method for tests that want to exercise the
1898
        overall behavior of the bzr application (rather than a unit test
1899
        or a functional test of the library.)
1900
1185.3.18 by Martin Pool
- add new helper TestBase.run_bzr_captured
1901
        This sends the stdout/stderr results into the test's log,
1902
        where it may be useful for debugging.  See also run_captured.
1687.1.2 by Robert Collins
Add stdin parameter to run_bzr and run_bzr_captured.
1903
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1904
        :keyword stdin: A string to be used as stdin for the command.
2399.1.17 by John Arbash Meinel
[merge] bzr.dev 2562
1905
        :keyword retcode: The status code the command should return;
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1906
            default 0.
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1907
        :keyword working_dir: The directory to run the command in
2399.1.17 by John Arbash Meinel
[merge] bzr.dev 2562
1908
        :keyword error_regexes: A list of expected error messages.  If
1909
            specified they must be seen in the error output of the command.
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
1910
        """
4665.5.15 by Vincent Ladeuil
Catch the retcode for all commands.
1911
        retcode, out, err = self._run_bzr_autosplit(
2830.2.5 by Martin Pool
Deprecated ``run_bzr_decode``; use the new ``output_encoding`` parameter to
1912
            args=args,
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1913
            retcode=retcode,
2830.2.5 by Martin Pool
Deprecated ``run_bzr_decode``; use the new ``output_encoding`` parameter to
1914
            encoding=encoding,
1915
            stdin=stdin,
1916
            working_dir=working_dir,
2530.3.2 by Martin Pool
Refactoring run_bzr code into more of a common base.
1917
            )
4325.4.6 by Vincent Ladeuil
Fixed as per John's and Markus reviews.
1918
        self.assertIsInstance(error_regexes, (list, tuple))
2292.1.27 by Marius Kruger
* tests/__init__.TestCase.run_bzr_captured
1919
        for regex in error_regexes:
1920
            self.assertContainsRe(err, regex)
1921
        return out, err
1922
1711.2.70 by John Arbash Meinel
Add run_bzr_errors alongside run_bzr, to make it easy to check the right error is occurring.
1923
    def run_bzr_error(self, error_regexes, *args, **kwargs):
1711.2.71 by John Arbash Meinel
Default to retcode=3, and add a test for run_bzr_error
1924
        """Run bzr, and check that stderr contains the supplied regexes
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1925
1926
        :param error_regexes: Sequence of regular expressions which
1711.7.11 by John Arbash Meinel
Clean up the documentation for run_bzr_error on Martin's suggestion.
1927
            must each be found in the error output. The relative ordering
1928
            is not enforced.
1929
        :param args: command-line arguments for bzr
1930
        :param kwargs: Keyword arguments which are interpreted by run_bzr
1931
            This function changes the default value of retcode to be 3,
1932
            since in most cases this is run when you expect bzr to fail.
2581.1.1 by Martin Pool
Merge more runbzr cleanups
1933
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1934
        :return: (out, err) The actual output of running the command (in case
1935
            you want to do more inspection)
1936
1937
        Examples of use::
1938
1711.7.11 by John Arbash Meinel
Clean up the documentation for run_bzr_error on Martin's suggestion.
1939
            # Make sure that commit is failing because there is nothing to do
1940
            self.run_bzr_error(['no changes to commit'],
2665.1.1 by Michael Hudson
make run_bzr stricter about the keyword arguments it takes.
1941
                               ['commit', '-m', 'my commit comment'])
1711.7.11 by John Arbash Meinel
Clean up the documentation for run_bzr_error on Martin's suggestion.
1942
            # Make sure --strict is handling an unknown file, rather than
1943
            # giving us the 'nothing to do' error
1944
            self.build_tree(['unknown'])
1945
            self.run_bzr_error(['Commit refused because there are unknown files'],
2665.1.1 by Michael Hudson
make run_bzr stricter about the keyword arguments it takes.
1946
                               ['commit', --strict', '-m', 'my commit comment'])
1711.2.71 by John Arbash Meinel
Default to retcode=3, and add a test for run_bzr_error
1947
        """
1948
        kwargs.setdefault('retcode', 3)
2552.2.3 by Vincent Ladeuil
Deprecate the varargs syntax and fix the tests.
1949
        kwargs['error_regexes'] = error_regexes
1950
        out, err = self.run_bzr(*args, **kwargs)
1711.2.71 by John Arbash Meinel
Default to retcode=3, and add a test for run_bzr_error
1951
        return out, err
1711.2.70 by John Arbash Meinel
Add run_bzr_errors alongside run_bzr, to make it easy to check the right error is occurring.
1952
1752.1.6 by Aaron Bentley
Rename run_bzr_external -> run_bzr_subprocess, add docstring
1953
    def run_bzr_subprocess(self, *args, **kwargs):
1954
        """Run bzr in a subprocess for testing.
1955
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1956
        This starts a new Python interpreter and runs bzr in there.
1752.1.6 by Aaron Bentley
Rename run_bzr_external -> run_bzr_subprocess, add docstring
1957
        This should only be used for tests that have a justifiable need for
1958
        this isolation: e.g. they are testing startup time, or signal
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
1959
        handling, or early startup code, etc.  Subprocess code can't be
1752.1.6 by Aaron Bentley
Rename run_bzr_external -> run_bzr_subprocess, add docstring
1960
        profiled or debugged so easily.
1752.1.7 by Aaron Bentley
Stop using shlex in run_bzr_subprocess
1961
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1962
        :keyword retcode: The status code that is expected.  Defaults to 0.  If
1963.1.1 by John Arbash Meinel
run_bzr_subprocess() can take an env_changes parameter
1963
            None is supplied, the status code is not checked.
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1964
        :keyword env_changes: A dictionary which lists changes to environment
1963.1.1 by John Arbash Meinel
run_bzr_subprocess() can take an env_changes parameter
1965
            variables. A value of None will unset the env variable.
1966
            The values must be strings. The change will only occur in the
1967
            child, so you don't need to fix the environment after running.
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
1968
        :keyword universal_newlines: Convert CRLF => LF
1969
        :keyword allow_plugins: By default the subprocess is run with
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
1970
            --no-plugins to ensure test reproducibility. Also, it is possible
2067.2.2 by John Arbash Meinel
Review comments from Robert
1971
            for system-wide plugins to create unexpected output on stderr,
1972
            which can cause unnecessary test failures.
1752.1.6 by Aaron Bentley
Rename run_bzr_external -> run_bzr_subprocess, add docstring
1973
        """
1963.1.2 by John Arbash Meinel
Cleanups suggested by Martin, add test that env_changes can remove an env variable
1974
        env_changes = kwargs.get('env_changes', {})
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
1975
        working_dir = kwargs.get('working_dir', None)
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
1976
        allow_plugins = kwargs.get('allow_plugins', False)
2665.4.1 by Aaron Bentley
teach run_bzr_subprocess to accept either a list of strings or a string
1977
        if len(args) == 1:
1978
            if isinstance(args[0], list):
1979
                args = args[0]
1980
            elif isinstance(args[0], basestring):
1981
                args = list(shlex.split(args[0]))
1982
        else:
3400.3.6 by Martin Pool
Remove code deprecated prior to 1.1 and its tests
1983
            raise ValueError("passing varargs to run_bzr_subprocess")
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
1984
        process = self.start_bzr_subprocess(args, env_changes=env_changes,
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
1985
                                            working_dir=working_dir,
1986
                                            allow_plugins=allow_plugins)
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
1987
        # We distinguish between retcode=None and retcode not passed.
1988
        supplied_retcode = kwargs.get('retcode', 0)
1989
        return self.finish_bzr_subprocess(process, retcode=supplied_retcode,
1990
            universal_newlines=kwargs.get('universal_newlines', False),
1991
            process_args=args)
1992
1910.17.9 by Andrew Bennetts
Add skip_if_plan_to_signal flag to start_bzr_subprocess.
1993
    def start_bzr_subprocess(self, process_args, env_changes=None,
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
1994
                             skip_if_plan_to_signal=False,
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
1995
                             working_dir=None,
1996
                             allow_plugins=False):
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
1997
        """Start bzr in a subprocess for testing.
1998
1999
        This starts a new Python interpreter and runs bzr in there.
2000
        This should only be used for tests that have a justifiable need for
2001
        this isolation: e.g. they are testing startup time, or signal
2002
        handling, or early startup code, etc.  Subprocess code can't be
2003
        profiled or debugged so easily.
2004
2005
        :param process_args: a list of arguments to pass to the bzr executable,
2399.1.7 by John Arbash Meinel
Cleanup bzrlib/benchmarks/* so that everything at least has a valid doc string.
2006
            for example ``['--version']``.
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2007
        :param env_changes: A dictionary which lists changes to environment
2008
            variables. A value of None will unset the env variable.
2009
            The values must be strings. The change will only occur in the
2010
            child, so you don't need to fix the environment after running.
1910.17.9 by Andrew Bennetts
Add skip_if_plan_to_signal flag to start_bzr_subprocess.
2011
        :param skip_if_plan_to_signal: raise TestSkipped when true and os.kill
2012
            is not available.
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
2013
        :param allow_plugins: If False (default) pass --no-plugins to bzr.
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2014
2015
        :returns: Popen object for the started process.
2016
        """
1910.17.9 by Andrew Bennetts
Add skip_if_plan_to_signal flag to start_bzr_subprocess.
2017
        if skip_if_plan_to_signal:
2018
            if not getattr(os, 'kill', None):
2019
                raise TestSkipped("os.kill not available.")
2020
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2021
        if env_changes is None:
2022
            env_changes = {}
1963.1.7 by John Arbash Meinel
Switch to directly setting the env, and cleaning it up. So that it works on all platforms
2023
        old_env = {}
2024
1963.1.2 by John Arbash Meinel
Cleanups suggested by Martin, add test that env_changes can remove an env variable
2025
        def cleanup_environment():
2026
            for env_var, value in env_changes.iteritems():
1963.1.7 by John Arbash Meinel
Switch to directly setting the env, and cleaning it up. So that it works on all platforms
2027
                old_env[env_var] = osutils.set_or_unset_env(env_var, value)
2028
2029
        def restore_environment():
2030
            for env_var, value in old_env.iteritems():
2031
                osutils.set_or_unset_env(env_var, value)
1963.1.1 by John Arbash Meinel
run_bzr_subprocess() can take an env_changes parameter
2032
2018.1.1 by Andrew Bennetts
Make bzr+ssh:// actually work (at least with absolute paths).
2033
        bzr_path = self.get_bzr_path()
1963.1.7 by John Arbash Meinel
Switch to directly setting the env, and cleaning it up. So that it works on all platforms
2034
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
2035
        cwd = None
2036
        if working_dir is not None:
2037
            cwd = osutils.getcwd()
2038
            os.chdir(working_dir)
2039
1963.1.7 by John Arbash Meinel
Switch to directly setting the env, and cleaning it up. So that it works on all platforms
2040
        try:
2041
            # win32 subprocess doesn't support preexec_fn
2042
            # so we will avoid using it on all platforms, just to
2043
            # make sure the code path is used, and we don't break on win32
2044
            cleanup_environment()
3616.2.1 by Mark Hammond
Fix how blackbox tests start bzr from frozen executables.
2045
            command = [sys.executable]
2046
            # frozen executables don't need the path to bzr
3616.2.7 by Mark Hammond
prefer getattr() over hasattr()
2047
            if getattr(sys, "frozen", None) is None:
3616.2.1 by Mark Hammond
Fix how blackbox tests start bzr from frozen executables.
2048
                command.append(bzr_path)
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
2049
            if not allow_plugins:
2050
                command.append('--no-plugins')
2051
            command.extend(process_args)
2052
            process = self._popen(command, stdin=PIPE, stdout=PIPE, stderr=PIPE)
1963.1.7 by John Arbash Meinel
Switch to directly setting the env, and cleaning it up. So that it works on all platforms
2053
        finally:
2054
            restore_environment()
2027.5.1 by John Arbash Meinel
Add working_dir=XX to run_bzr_* functions, and clean up tests
2055
            if cwd is not None:
2056
                os.chdir(cwd)
2057
1910.17.2 by Andrew Bennetts
Add start_bzr_subprocess and stop_bzr_subprocess to allow test code to continue
2058
        return process
2059
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
2060
    def _popen(self, *args, **kwargs):
2061
        """Place a call to Popen.
2067.2.2 by John Arbash Meinel
Review comments from Robert
2062
2063
        Allows tests to override this method to intercept the calls made to
2064
        Popen for introspection.
2067.2.1 by John Arbash Meinel
Change run_bzr_subprocess to default to supplying --no-plugins.
2065
        """
2066
        return Popen(*args, **kwargs)
2067
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2068
    def get_source_path(self):
2069
        """Return the path of the directory containing bzrlib."""
2070
        return os.path.dirname(os.path.dirname(bzrlib.__file__))
2071
2018.1.1 by Andrew Bennetts
Make bzr+ssh:// actually work (at least with absolute paths).
2072
    def get_bzr_path(self):
2018.1.9 by Andrew Bennetts
Implement ParamikoVendor.connect_ssh
2073
        """Return the path of the 'bzr' executable for this test suite."""
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2074
        bzr_path = self.get_source_path()+'/bzr'
2018.1.1 by Andrew Bennetts
Make bzr+ssh:// actually work (at least with absolute paths).
2075
        if not os.path.isfile(bzr_path):
2076
            # We are probably installed. Assume sys.argv is the right file
2077
            bzr_path = sys.argv[0]
2078
        return bzr_path
2079
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2080
    def finish_bzr_subprocess(self, process, retcode=0, send_signal=None,
2081
                              universal_newlines=False, process_args=None):
1910.17.2 by Andrew Bennetts
Add start_bzr_subprocess and stop_bzr_subprocess to allow test code to continue
2082
        """Finish the execution of process.
2083
2084
        :param process: the Popen object returned from start_bzr_subprocess.
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2085
        :param retcode: The status code that is expected.  Defaults to 0.  If
2086
            None is supplied, the status code is not checked.
1910.17.2 by Andrew Bennetts
Add start_bzr_subprocess and stop_bzr_subprocess to allow test code to continue
2087
        :param send_signal: an optional signal to send to the process.
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2088
        :param universal_newlines: Convert CRLF => LF
1910.17.2 by Andrew Bennetts
Add start_bzr_subprocess and stop_bzr_subprocess to allow test code to continue
2089
        :returns: (stdout, stderr)
2090
        """
2091
        if send_signal is not None:
2092
            os.kill(process.pid, send_signal)
1910.17.8 by Andrew Bennetts
Refactor run_bzr_subprocess to use start_bzr_subprocess and finish_bzr_subprocess.
2093
        out, err = process.communicate()
2094
2095
        if universal_newlines:
2096
            out = out.replace('\r\n', '\n')
2097
            err = err.replace('\r\n', '\n')
2098
2099
        if retcode is not None and retcode != process.returncode:
2100
            if process_args is None:
2101
                process_args = "(unknown args)"
2102
            mutter('Output of bzr %s:\n%s', process_args, out)
2103
            mutter('Error for bzr %s:\n%s', process_args, err)
2104
            self.fail('Command bzr %s failed with retcode %s != %s'
2105
                      % (process_args, retcode, process.returncode))
2106
        return [out, err]
1910.17.2 by Andrew Bennetts
Add start_bzr_subprocess and stop_bzr_subprocess to allow test code to continue
2107
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2108
    def check_inventory_shape(self, inv, shape):
1291 by Martin Pool
- add test for moving files between directories
2109
        """Compare an inventory to a list of expected names.
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2110
2111
        Fail if they are not precisely equal.
2112
        """
2113
        extras = []
2114
        shape = list(shape)             # copy
2115
        for path, ie in inv.entries():
2116
            name = path.replace('\\', '/')
2545.3.1 by James Westby
Fix detection of directory entries in the inventory.
2117
            if ie.kind == 'directory':
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2118
                name = name + '/'
2119
            if name in shape:
2120
                shape.remove(name)
2121
            else:
2122
                extras.append(name)
2123
        if shape:
2124
            self.fail("expected paths not found in inventory: %r" % shape)
2125
        if extras:
2126
            self.fail("unexpected paths found in inventory: %r" % extras)
2127
1141 by Martin Pool
- rename FunctionalTest to TestCaseInTempDir
2128
    def apply_redirected(self, stdin=None, stdout=None, stderr=None,
2129
                         a_callable=None, *args, **kwargs):
2130
        """Call callable with redirected std io pipes.
2131
2132
        Returns the return code."""
2133
        if not callable(a_callable):
2134
            raise ValueError("a_callable must be callable.")
2135
        if stdin is None:
2136
            stdin = StringIO("")
2137
        if stdout is None:
1534.4.3 by Robert Collins
Implement BranchTestProviderAdapter, so tests now run across all branch formats.
2138
            if getattr(self, "_log_file", None) is not None:
974.1.70 by Aaron Bentley
Fixed selftest spewage (Brian M. Carlson)
2139
                stdout = self._log_file
2140
            else:
2141
                stdout = StringIO()
1141 by Martin Pool
- rename FunctionalTest to TestCaseInTempDir
2142
        if stderr is None:
1534.4.3 by Robert Collins
Implement BranchTestProviderAdapter, so tests now run across all branch formats.
2143
            if getattr(self, "_log_file", None is not None):
974.1.70 by Aaron Bentley
Fixed selftest spewage (Brian M. Carlson)
2144
                stderr = self._log_file
2145
            else:
2146
                stderr = StringIO()
1141 by Martin Pool
- rename FunctionalTest to TestCaseInTempDir
2147
        real_stdin = sys.stdin
2148
        real_stdout = sys.stdout
2149
        real_stderr = sys.stderr
2150
        try:
2151
            sys.stdout = stdout
2152
            sys.stderr = stderr
2153
            sys.stdin = stdin
1160 by Martin Pool
- tiny refactoring
2154
            return a_callable(*args, **kwargs)
1141 by Martin Pool
- rename FunctionalTest to TestCaseInTempDir
2155
        finally:
2156
            sys.stdout = real_stdout
2157
            sys.stderr = real_stderr
2158
            sys.stdin = real_stdin
2159
2018.5.78 by Andrew Bennetts
Implement RemoteRepository.lock_write/unlock to expect and send tokens over the
2160
    def reduceLockdirTimeout(self):
2161
        """Reduce the default lock timeout for the duration of the test, so that
2162
        if LockContention occurs during a test, it does so quickly.
2163
2164
        Tests that expect to provoke LockContention errors should call this.
2165
        """
2166
        orig_timeout = bzrlib.lockdir._DEFAULT_TIMEOUT_SECONDS
2167
        def resetTimeout():
2168
            bzrlib.lockdir._DEFAULT_TIMEOUT_SECONDS = orig_timeout
2169
        self.addCleanup(resetTimeout)
2170
        bzrlib.lockdir._DEFAULT_TIMEOUT_SECONDS = 0
1141 by Martin Pool
- rename FunctionalTest to TestCaseInTempDir
2171
2717.1.1 by Lukáš Lalinsky
Use UTF-8 encoded StringIO for log tests to avoid failures on non-ASCII committer names.
2172
    def make_utf8_encoded_stringio(self, encoding_type=None):
2173
        """Return a StringIOWrapper instance, that will encode Unicode
2174
        input to UTF-8.
2175
        """
2176
        if encoding_type is None:
2177
            encoding_type = 'strict'
2178
        sio = StringIO()
2179
        output_encoding = 'utf-8'
2180
        sio = codecs.getwriter(output_encoding)(sio, errors=encoding_type)
2181
        sio.encoding = output_encoding
2182
        return sio
2183
4476.3.32 by Andrew Bennetts
Move disable_verb into base TestCase to remove duplication, fix trivial test failures due to new insert_stream verb in test_remote (and also add some trivial tests for the new verb).
2184
    def disable_verb(self, verb):
2185
        """Disable a smart server verb for one test."""
2186
        from bzrlib.smart import request
2187
        request_handlers = request.request_handlers
2188
        orig_method = request_handlers.get(verb)
2189
        request_handlers.remove(verb)
2190
        def restoreVerb():
2191
            request_handlers.register(verb, orig_method)
2192
        self.addCleanup(restoreVerb)
2193
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2194
4070.3.1 by Robert Collins
Alter branch sprouting with an alternate fix for stacked branches that does not require multiple copy_content_into and set_parent calls, reducing IO and round trips.
2195
class CapturedCall(object):
2196
    """A helper for capturing smart server calls for easy debug analysis."""
2197
2198
    def __init__(self, params, prefix_length):
2199
        """Capture the call with params and skip prefix_length stack frames."""
2200
        self.call = params
2201
        import traceback
2202
        # The last 5 frames are the __init__, the hook frame, and 3 smart
2203
        # client frames. Beyond this we could get more clever, but this is good
2204
        # enough for now.
2205
        stack = traceback.extract_stack()[prefix_length:-5]
2206
        self.stack = ''.join(traceback.format_list(stack))
2207
2208
    def __str__(self):
2209
        return self.call.method
2210
2211
    def __repr__(self):
2212
        return self.call.method
2213
2214
    def stack(self):
2215
        return self.stack
2216
2217
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2218
class TestCaseWithMemoryTransport(TestCase):
2219
    """Common test class for tests that do not need disk resources.
2220
2221
    Tests that need disk resources should derive from TestCaseWithTransport.
2222
2223
    TestCaseWithMemoryTransport sets the TEST_ROOT variable for all bzr tests.
2224
2225
    For TestCaseWithMemoryTransport the test_home_dir is set to the name of
2226
    a directory which does not exist. This serves to help ensure test isolation
2227
    is preserved. test_dir is set to the TEST_ROOT, as is cwd, because they
2228
    must exist. However, TestCaseWithMemoryTransport does not offer local
2229
    file defaults for the transport in tests, nor does it obey the command line
2230
    override, so tests that accidentally write to the common directory should
2231
    be rare.
2485.6.6 by Martin Pool
Put test root directory (containing per-test directories) in TMPDIR
2232
2233
    :cvar TEST_ROOT: Directory containing all temporary directories, plus
2234
    a .bzr directory that stops us ascending higher into the filesystem.
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2235
    """
2236
2237
    TEST_ROOT = None
2238
    _TEST_NAME = 'test'
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2239
1986.2.5 by Robert Collins
Unbreak transport tests.
2240
    def __init__(self, methodName='runTest'):
3128.1.3 by Vincent Ladeuil
Since we are there s/parameteris.*/parameteriz&/.
2241
        # allow test parameterization after test construction and before test
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2242
        # execution. Variables that the parameterizer sets need to be
1986.2.5 by Robert Collins
Unbreak transport tests.
2243
        # ones that are not set by setUp, or setUp will trash them.
2244
        super(TestCaseWithMemoryTransport, self).__init__(methodName)
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2245
        self.vfs_transport_factory = default_transport
2246
        self.transport_server = None
1986.2.5 by Robert Collins
Unbreak transport tests.
2247
        self.transport_readonly_server = None
2018.5.44 by Andrew Bennetts
Small changes to help a couple more tests pass.
2248
        self.__vfs_server = None
1986.2.5 by Robert Collins
Unbreak transport tests.
2249
2520.3.1 by Vincent Ladeuil
Fix 110448 by adding a relpath parameter to get_transport.
2250
    def get_transport(self, relpath=None):
2251
        """Return a writeable transport.
2252
2253
        This transport is for the test scratch space relative to
2592.2.5 by Jonathan Lange
Make UnicodeFilename feature less insane. Add a simple test for it too.
2254
        "self._test_root"
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2255
2520.3.1 by Vincent Ladeuil
Fix 110448 by adding a relpath parameter to get_transport.
2256
        :param relpath: a path relative to the base url.
2257
        """
2258
        t = get_transport(self.get_url(relpath))
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2259
        self.assertFalse(t.is_readonly())
2260
        return t
2261
2520.3.1 by Vincent Ladeuil
Fix 110448 by adding a relpath parameter to get_transport.
2262
    def get_readonly_transport(self, relpath=None):
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2263
        """Return a readonly transport for the test scratch space
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2264
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2265
        This can be used to test that operations which should only need
2266
        readonly access in fact do not try to write.
2520.3.1 by Vincent Ladeuil
Fix 110448 by adding a relpath parameter to get_transport.
2267
2268
        :param relpath: a path relative to the base url.
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2269
        """
2520.3.1 by Vincent Ladeuil
Fix 110448 by adding a relpath parameter to get_transport.
2270
        t = get_transport(self.get_readonly_url(relpath))
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2271
        self.assertTrue(t.is_readonly())
2272
        return t
2273
2004.1.19 by v.ladeuil+lp at free
Test protocol version in http responses.
2274
    def create_transport_readonly_server(self):
2275
        """Create a transport server from class defined at init.
2276
2145.1.1 by mbp at sourcefrog
merge urllib keepalive etc
2277
        This is mostly a hook for daughter classes.
2004.1.19 by v.ladeuil+lp at free
Test protocol version in http responses.
2278
        """
2279
        return self.transport_readonly_server()
2280
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2281
    def get_readonly_server(self):
2282
        """Get the server instance for the readonly transport
2283
2284
        This is useful for some tests with specific servers to do diagnostics.
2285
        """
2286
        if self.__readonly_server is None:
2287
            if self.transport_readonly_server is None:
2288
                # readonly decorator requested
2289
                self.__readonly_server = ReadonlyServer()
2290
            else:
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2291
                # explicit readonly transport.
2004.1.19 by v.ladeuil+lp at free
Test protocol version in http responses.
2292
                self.__readonly_server = self.create_transport_readonly_server()
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2293
            self.start_server(self.__readonly_server,
2294
                self.get_vfs_only_server())
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2295
        return self.__readonly_server
2296
2297
    def get_readonly_url(self, relpath=None):
2298
        """Get a URL for the readonly transport.
2299
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2300
        This will either be backed by '.' or a decorator to the transport
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2301
        used by self.get_url()
2302
        relpath provides for clients to get a path relative to the base url.
2303
        These should only be downwards relative, not upwards.
2304
        """
2305
        base = self.get_readonly_server().get_url()
2520.3.1 by Vincent Ladeuil
Fix 110448 by adding a relpath parameter to get_transport.
2306
        return self._adjust_url(base, relpath)
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2307
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2308
    def get_vfs_only_server(self):
2018.5.44 by Andrew Bennetts
Small changes to help a couple more tests pass.
2309
        """Get the vfs only read/write server instance.
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2310
2311
        This is useful for some tests with specific servers that need
2312
        diagnostics.
2313
2314
        For TestCaseWithMemoryTransport this is always a MemoryServer, and there
2315
        is no means to override it.
2316
        """
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2317
        if self.__vfs_server is None:
2318
            self.__vfs_server = MemoryServer()
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2319
            self.start_server(self.__vfs_server)
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2320
        return self.__vfs_server
2321
2322
    def get_server(self):
2323
        """Get the read/write server instance.
2324
2325
        This is useful for some tests with specific servers that need
2326
        diagnostics.
2327
2328
        This is built from the self.transport_server factory. If that is None,
2329
        then the self.get_vfs_server is returned.
2330
        """
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2331
        if self.__server is None:
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2332
            if (self.transport_server is None or self.transport_server is
2333
                self.vfs_transport_factory):
2334
                self.__server = self.get_vfs_only_server()
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2335
            else:
2336
                # bring up a decorated means of access to the vfs only server.
2337
                self.__server = self.transport_server()
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2338
                self.start_server(self.__server, self.get_vfs_only_server())
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2339
        return self.__server
2340
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2341
    def _adjust_url(self, base, relpath):
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2342
        """Get a URL (or maybe a path) for the readwrite transport.
2343
2344
        This will either be backed by '.' or to an equivalent non-file based
2345
        facility.
2346
        relpath provides for clients to get a path relative to the base url.
2347
        These should only be downwards relative, not upwards.
2348
        """
2349
        if relpath is not None and relpath != '.':
2350
            if not base.endswith('/'):
2351
                base = base + '/'
2352
            # XXX: Really base should be a url; we did after all call
2353
            # get_url()!  But sometimes it's just a path (from
2354
            # LocalAbspathServer), and it'd be wrong to append urlescaped data
2355
            # to a non-escaped local path.
2356
            if base.startswith('./') or base.startswith('/'):
2357
                base += relpath
2358
            else:
2359
                base += urlutils.escape(relpath)
2360
        return base
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2361
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2362
    def get_url(self, relpath=None):
2363
        """Get a URL (or maybe a path) for the readwrite transport.
2364
2365
        This will either be backed by '.' or to an equivalent non-file based
2366
        facility.
2367
        relpath provides for clients to get a path relative to the base url.
2368
        These should only be downwards relative, not upwards.
2369
        """
2370
        base = self.get_server().get_url()
2371
        return self._adjust_url(base, relpath)
2372
2373
    def get_vfs_only_url(self, relpath=None):
2374
        """Get a URL (or maybe a path for the plain old vfs transport.
2375
2018.18.9 by Martin Pool
remote Repository.tarball builds a temporary directory and tars that
2376
        This will never be a smart protocol.  It always has all the
2377
        capabilities of the local filesystem, but it might actually be a
2378
        MemoryTransport or some other similar virtual filesystem.
2379
2399.1.16 by John Arbash Meinel
[merge] bzr.dev 2466
2380
        This is the backing transport (if any) of the server returned by
2018.18.9 by Martin Pool
remote Repository.tarball builds a temporary directory and tars that
2381
        get_url and get_readonly_url.
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
2382
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2383
        :param relpath: provides for clients to get a path relative to the base
2384
            url.  These should only be downwards relative, not upwards.
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
2385
        :return: A URL
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2386
        """
2387
        base = self.get_vfs_only_server().get_url()
2388
        return self._adjust_url(base, relpath)
2389
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2390
    def _create_safety_net(self):
2391
        """Make a fake bzr directory.
2392
2875.1.1 by Vincent Ladeuil
Fix #147986 by monitoring a safety .bzr directory.
2393
        This prevents any tests propagating up onto the TEST_ROOT directory's
2394
        real branch.
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2395
        """
2396
        root = TestCaseWithMemoryTransport.TEST_ROOT
2397
        bzrdir.BzrDir.create_standalone_workingtree(root)
2398
2399
    def _check_safety_net(self):
2400
        """Check that the safety .bzr directory have not been touched.
2401
2402
        _make_test_root have created a .bzr directory to prevent tests from
2403
        propagating. This method ensures than a test did not leaked.
2404
        """
2405
        root = TestCaseWithMemoryTransport.TEST_ROOT
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2406
        self.permit_url(get_transport(root).base)
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2407
        wt = workingtree.WorkingTree.open(root)
2408
        last_rev = wt.last_revision()
2409
        if last_rev != 'null:':
2410
            # The current test have modified the /bzr directory, we need to
2411
            # recreate a new one or all the followng tests will fail.
2412
            # If you need to inspect its content uncomment the following line
2413
            # import pdb; pdb.set_trace()
4807.3.3 by John Arbash Meinel
Report the test-id when we fail to delete a testing dir.
2414
            _rmtree_temp_dir(root + '/.bzr', test_id=self.id())
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2415
            self._create_safety_net()
2416
            raise AssertionError('%s/.bzr should not be modified' % root)
2417
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2418
    def _make_test_root(self):
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2419
        if TestCaseWithMemoryTransport.TEST_ROOT is None:
4707.1.1 by Vincent Ladeuil
Fix OSX and FreeBSD failures.
2420
            # Watch out for tricky test dir (on OSX /tmp -> /private/tmp)
2421
            root = osutils.realpath(osutils.mkdtemp(prefix='testbzr-',
2422
                                                    suffix='.tmp'))
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2423
            TestCaseWithMemoryTransport.TEST_ROOT = root
2424
2425
            self._create_safety_net()
2426
2427
            # The same directory is used by all tests, and we're not
2428
            # specifically told when all tests are finished.  This will do.
2429
            atexit.register(_rmtree_temp_dir, root)
2430
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2431
        self.permit_dir(TestCaseWithMemoryTransport.TEST_ROOT)
2817.5.1 by Vincent Ladeuil
Catch leaking tests.
2432
        self.addCleanup(self._check_safety_net)
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2433
2434
    def makeAndChdirToTestDir(self):
2435
        """Create a temporary directories for this one test.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2436
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2437
        This must set self.test_home_dir and self.test_dir and chdir to
2438
        self.test_dir.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2439
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2440
        For TestCaseWithMemoryTransport we chdir to the TEST_ROOT for this test.
2441
        """
2442
        os.chdir(TestCaseWithMemoryTransport.TEST_ROOT)
2443
        self.test_dir = TestCaseWithMemoryTransport.TEST_ROOT
2444
        self.test_home_dir = self.test_dir + "/MemoryTransportMissingHomeDir"
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2445
        self.permit_dir(self.test_dir)
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2446
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2447
    def make_branch(self, relpath, format=None):
2448
        """Create a branch on the transport at relpath."""
2449
        repo = self.make_repository(relpath, format=format)
2450
        return repo.bzrdir.create_branch()
2451
2452
    def make_bzrdir(self, relpath, format=None):
2453
        try:
2454
            # might be a relative or absolute path
2455
            maybe_a_url = self.get_url(relpath)
2456
            segments = maybe_a_url.rsplit('/', 1)
2457
            t = get_transport(maybe_a_url)
2458
            if len(segments) > 1 and segments[-1] not in ('', '.'):
2475.3.3 by John Arbash Meinel
Change calls to try/mkdir('.')/except FileExists to ensure_base()
2459
                t.ensure_base()
2230.3.22 by Aaron Bentley
Make test suite use format registry default, not BzrDir default
2460
            if format is None:
2461
                format = 'default'
2204.4.11 by Aaron Bentley
deprecate Repository.set_default_format, update upgrade tests
2462
            if isinstance(format, basestring):
2463
                format = bzrdir.format_registry.make_bzrdir(format)
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2464
            return format.initialize_on_transport(t)
2465
        except errors.UninitializableFormat:
2466
            raise TestSkipped("Format %s is not initializable." % format)
2467
2468
    def make_repository(self, relpath, shared=False, format=None):
2018.18.9 by Martin Pool
remote Repository.tarball builds a temporary directory and tars that
2469
        """Create a repository on our default transport at relpath.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2470
2018.18.9 by Martin Pool
remote Repository.tarball builds a temporary directory and tars that
2471
        Note that relpath must be a relative path, not a full url.
2472
        """
2473
        # FIXME: If you create a remoterepository this returns the underlying
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2474
        # real format, which is incorrect.  Actually we should make sure that
2018.18.9 by Martin Pool
remote Repository.tarball builds a temporary directory and tars that
2475
        # RemoteBzrDir returns a RemoteRepository.
2476
        # maybe  mbp 20070410
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2477
        made_control = self.make_bzrdir(relpath, format=format)
2478
        return made_control.create_repository(shared=shared)
2479
4145.1.2 by Robert Collins
Add a refresh_data method on Repository allowing cleaner handling of insertions into RemoteRepository objects with _real_repository instances.
2480
    def make_smart_server(self, path):
2481
        smart_server = server.SmartTCPServer_for_testing()
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2482
        self.start_server(smart_server, self.get_server())
4145.1.2 by Robert Collins
Add a refresh_data method on Repository allowing cleaner handling of insertions into RemoteRepository objects with _real_repository instances.
2483
        remote_transport = get_transport(smart_server.get_url()).clone(path)
2484
        return remote_transport
2485
1986.4.9 by Robert Collins
``TestCase.make_branch_and_memory_tree`` now takes a format
2486
    def make_branch_and_memory_tree(self, relpath, format=None):
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2487
        """Create a branch on the default transport and a MemoryTree for it."""
1986.4.9 by Robert Collins
``TestCase.make_branch_and_memory_tree`` now takes a format
2488
        b = self.make_branch(relpath, format=format)
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2489
        return memorytree.MemoryTree.create_on_branch(b)
2490
4476.3.59 by Andrew Bennetts
Undo changes that aren't needed anymore.
2491
    def make_branch_builder(self, relpath, format=None):
4257.3.8 by Andrew Bennetts
Fix TestCase.make_branch_builder to make a branch in the specified format. Also add an interrepo test scenario for KnitPack1 -> KnitPack6RichRoot, which fails.
2492
        branch = self.make_branch(relpath, format=format)
4476.3.59 by Andrew Bennetts
Undo changes that aren't needed anymore.
2493
        return branchbuilder.BranchBuilder(branch=branch)
3567.4.12 by John Arbash Meinel
Expose the branch building framework to the test suite.
2494
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2495
    def overrideEnvironmentForTesting(self):
4815.2.2 by Michael Hudson
another approach
2496
        test_home_dir = self.test_home_dir
2497
        if isinstance(test_home_dir, unicode):
2498
            test_home_dir = test_home_dir.encode(sys.getfilesystemencoding())
2499
        os.environ['HOME'] = test_home_dir
2500
        os.environ['BZR_HOME'] = test_home_dir
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2501
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2502
    def setUp(self):
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2503
        super(TestCaseWithMemoryTransport, self).setUp()
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2504
        self._make_test_root()
1185.16.110 by mbp at sourcefrog
Refactor test setup/teardown into cleanup callbacks
2505
        _currentdir = os.getcwdu()
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2506
        def _leaveDirectory():
2507
            os.chdir(_currentdir)
2508
        self.addCleanup(_leaveDirectory)
2509
        self.makeAndChdirToTestDir()
2510
        self.overrideEnvironmentForTesting()
2511
        self.__readonly_server = None
2512
        self.__server = None
2381.1.3 by Robert Collins
Review feedback.
2513
        self.reduceLockdirTimeout()
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2514
4017.2.1 by Robert Collins
Add BzrDirFormatMeta1 test for the amount of rpc calls made initializing over the network.
2515
    def setup_smart_server_with_call_log(self):
2516
        """Sets up a smart server as the transport server with a call log."""
2517
        self.transport_server = server.SmartTCPServer_for_testing
2518
        self.hpss_calls = []
4070.3.1 by Robert Collins
Alter branch sprouting with an alternate fix for stacked branches that does not require multiple copy_content_into and set_parent calls, reducing IO and round trips.
2519
        import traceback
2520
        # Skip the current stack down to the caller of
2521
        # setup_smart_server_with_call_log
2522
        prefix_length = len(traceback.extract_stack()) - 2
4017.2.1 by Robert Collins
Add BzrDirFormatMeta1 test for the amount of rpc calls made initializing over the network.
2523
        def capture_hpss_call(params):
4070.3.1 by Robert Collins
Alter branch sprouting with an alternate fix for stacked branches that does not require multiple copy_content_into and set_parent calls, reducing IO and round trips.
2524
            self.hpss_calls.append(
2525
                CapturedCall(params, prefix_length))
4017.2.1 by Robert Collins
Add BzrDirFormatMeta1 test for the amount of rpc calls made initializing over the network.
2526
        client._SmartClient.hooks.install_named_hook(
2527
            'call', capture_hpss_call, None)
2528
2529
    def reset_smart_call_log(self):
2530
        self.hpss_calls = []
2531
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2532
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2533
class TestCaseInTempDir(TestCaseWithMemoryTransport):
2534
    """Derived class that runs a test within a temporary directory.
2535
2536
    This is useful for tests that need to create a branch, etc.
2537
2538
    The directory is created in a slightly complex way: for each
2539
    Python invocation, a new temporary top-level directory is created.
2540
    All test cases create their own directory within that.  If the
2541
    tests complete successfully, the directory is removed.
2542
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2543
    :ivar test_base_dir: The path of the top-level directory for this
2485.6.3 by Martin Pool
TestCaseInTempDir takes responsibility for cleaning up its own test dir
2544
    test, which contains a home directory and a work directory.
2545
2546
    :ivar test_home_dir: An initially empty directory under test_base_dir
2547
    which is used as $HOME for this test.
2548
2549
    :ivar test_dir: A directory under test_base_dir used as the current
2550
    directory when the test proper is run.
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2551
    """
2552
2553
    OVERRIDE_PYTHON = 'python'
2554
2555
    def check_file_contents(self, filename, expect):
2556
        self.log("check contents of file %s" % filename)
2557
        contents = file(filename, 'r').read()
2558
        if contents != expect:
2559
            self.log("expected: %r" % expect)
2560
            self.log("actually: %r" % contents)
2561
            self.fail("contents of %s not as expected" % filename)
2562
3549.2.4 by Martin Pool
Rename _getTestDirPrefix not to look like a test
2563
    def _getTestDirPrefix(self):
3549.2.1 by Martin Pool
Use test names in the temporary directory name
2564
        # create a directory within the top level test directory
4615.3.1 by Martin
Extend work around for path length limitations in selftest to cygwin
2565
        if sys.platform in ('win32', 'cygwin'):
3549.2.1 by Martin Pool
Use test names in the temporary directory name
2566
            name_prefix = re.sub('[<>*=+",:;_/\\-]', '_', self.id())
2567
            # windows is likely to have path-length limits so use a short name
2568
            name_prefix = name_prefix[-30:]
2569
        else:
2570
            name_prefix = re.sub('[/]', '_', self.id())
2571
        return name_prefix
2572
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2573
    def makeAndChdirToTestDir(self):
2574
        """See TestCaseWithMemoryTransport.makeAndChdirToTestDir().
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2575
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2576
        For TestCaseInTempDir we create a temporary directory based on the test
2577
        name and then create two subdirs - test and home under it.
2578
        """
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
2579
        name_prefix = osutils.pathjoin(TestCaseWithMemoryTransport.TEST_ROOT,
2580
            self._getTestDirPrefix())
3549.2.1 by Martin Pool
Use test names in the temporary directory name
2581
        name = name_prefix
2582
        for i in range(100):
2583
            if os.path.exists(name):
2584
                name = name_prefix + '_' + str(i)
2585
            else:
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2586
                # now create test and home directories within this dir
2587
                self.test_base_dir = name
2588
                self.addCleanup(self.deleteTestDir)
2589
                os.mkdir(self.test_base_dir)
3549.2.1 by Martin Pool
Use test names in the temporary directory name
2590
                break
4691.2.1 by Robert Collins
Add stronger test isolation by interception BzrDir.open and checking the thing being opened is known to the test suite.
2591
        self.permit_dir(self.test_base_dir)
2592
        # 'sprouting' and 'init' of a branch both walk up the tree to find
2593
        # stacking policy to honour; create a bzr dir with an unshared
2594
        # repository (but not a branch - our code would be trying to escape
2595
        # then!) to stop them, and permit it to be read.
2596
        # control = bzrdir.BzrDir.create(self.test_base_dir)
2597
        # control.create_repository()
2485.6.3 by Martin Pool
TestCaseInTempDir takes responsibility for cleaning up its own test dir
2598
        self.test_home_dir = self.test_base_dir + '/home'
2485.6.1 by Martin Pool
Remove duplication in TestCaseInTempDir.makeAndChdirToTestDir
2599
        os.mkdir(self.test_home_dir)
2485.6.3 by Martin Pool
TestCaseInTempDir takes responsibility for cleaning up its own test dir
2600
        self.test_dir = self.test_base_dir + '/work'
2485.6.1 by Martin Pool
Remove duplication in TestCaseInTempDir.makeAndChdirToTestDir
2601
        os.mkdir(self.test_dir)
2602
        os.chdir(self.test_dir)
2603
        # put name of test inside
2485.6.3 by Martin Pool
TestCaseInTempDir takes responsibility for cleaning up its own test dir
2604
        f = file(self.test_base_dir + '/name', 'w')
2485.6.1 by Martin Pool
Remove duplication in TestCaseInTempDir.makeAndChdirToTestDir
2605
        try:
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
2606
            f.write(self.id())
2485.6.1 by Martin Pool
Remove duplication in TestCaseInTempDir.makeAndChdirToTestDir
2607
        finally:
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
2608
            f.close()
2485.6.3 by Martin Pool
TestCaseInTempDir takes responsibility for cleaning up its own test dir
2609
2610
    def deleteTestDir(self):
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
2611
        os.chdir(TestCaseWithMemoryTransport.TEST_ROOT)
4807.3.3 by John Arbash Meinel
Report the test-id when we fail to delete a testing dir.
2612
        _rmtree_temp_dir(self.test_base_dir, test_id=self.id())
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2613
2193.2.1 by Alexander Belchenko
selftest: build tree for test with binary line-endings by default
2614
    def build_tree(self, shape, line_endings='binary', transport=None):
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2615
        """Build a test tree according to a pattern.
2616
2617
        shape is a sequence of file specifications.  If the final
2618
        character is '/', a directory is created.
2619
1755.3.1 by Robert Collins
Tune the time to build our kernel_like tree : make LocalTransport.put faster, AtomicFile faster, LocalTransport.append faster.
2620
        This assumes that all the elements in the tree being built are new.
2621
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2622
        This doesn't add anything to a branch.
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
2623
3034.4.8 by Alexander Belchenko
TestCaseInTempDir.build_tree now checks type of shape argument.
2624
        :type shape:    list or tuple.
1185.38.7 by John Arbash Meinel
Updated build_tree to use fixed line-endings for tests which read the file contents and compare
2625
        :param line_endings: Either 'binary' or 'native'
2399.1.6 by John Arbash Meinel
Cleanup bzrlib/tests/__init__.py so that epydoc doesn't complain.
2626
            in binary mode, exact contents are written in native mode, the
2627
            line endings match the default platform endings.
2628
        :param transport: A transport to write to, for building trees on VFS's.
2629
            If the transport is readonly or None, "." is opened automatically.
2630
        :return: None
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2631
        """
3044.1.1 by Martin Pool
Fix up calls to TestCase.build_tree passing a string rather than a list
2632
        if type(shape) not in (list, tuple):
2633
            raise AssertionError("Parameter 'shape' should be "
2634
                "a list or a tuple. Got %r instead" % (shape,))
1755.3.1 by Robert Collins
Tune the time to build our kernel_like tree : make LocalTransport.put faster, AtomicFile faster, LocalTransport.append faster.
2635
        # It's OK to just create them using forward slashes on windows.
1530.1.3 by Robert Collins
transport implementations now tested consistently.
2636
        if transport is None or transport.is_readonly():
1553.5.9 by Martin Pool
Add TestCaseWithTransport.get_transport and get_readonly_transport
2637
            transport = get_transport(".")
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2638
        for name in shape:
3757.3.2 by Vincent Ladeuil
Add a credential store for '.netrc'.
2639
            self.assertIsInstance(name, basestring)
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2640
            if name[-1] == '/':
1685.1.45 by John Arbash Meinel
Moved url functions into bzrlib.urlutils
2641
                transport.mkdir(urlutils.escape(name[:-1]))
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2642
            else:
1185.38.7 by John Arbash Meinel
Updated build_tree to use fixed line-endings for tests which read the file contents and compare
2643
                if line_endings == 'binary':
1530.1.3 by Robert Collins
transport implementations now tested consistently.
2644
                    end = '\n'
1185.38.7 by John Arbash Meinel
Updated build_tree to use fixed line-endings for tests which read the file contents and compare
2645
                elif line_endings == 'native':
1530.1.3 by Robert Collins
transport implementations now tested consistently.
2646
                    end = os.linesep
1185.38.7 by John Arbash Meinel
Updated build_tree to use fixed line-endings for tests which read the file contents and compare
2647
                else:
2227.2.2 by v.ladeuil+lp at free
Cleanup.
2648
                    raise errors.BzrError(
2649
                        'Invalid line ending request %r' % line_endings)
1185.85.76 by John Arbash Meinel
Adding an InvalidURL so transports can report they expect utf-8 quoted paths. Updated tests
2650
                content = "contents of %s%s" % (name.encode('utf-8'), end)
1955.3.27 by John Arbash Meinel
rename non_atomic_put_* to put_*non_atomic, and re-order the functions
2651
                transport.put_bytes_non_atomic(urlutils.escape(name), content)
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2652
1185.16.53 by Martin Pool
- annotate improvements from Goffreddo, with extra bug fixes and tests
2653
    def build_tree_contents(self, shape):
1514 by Robert Collins
Unbreak self.build_tree_shape in tests.
2654
        build_tree_contents(shape)
1185.16.53 by Martin Pool
- annotate improvements from Goffreddo, with extra bug fixes and tests
2655
2655.2.5 by Marius Kruger
* Improve BzrRemoveChangedFilesError message.
2656
    def assertInWorkingTree(self, path, root_path='.', tree=None):
2292.1.1 by Marius Kruger
"bzr remove" and "bzr rm" will now remove the working file.
2657
        """Assert whether path or paths are in the WorkingTree"""
2658
        if tree is None:
2659
            tree = workingtree.WorkingTree.open(root_path)
2660
        if not isinstance(path, basestring):
2661
            for p in path:
3585.2.1 by Robert Collins
Create acceptance test for bug 150438.
2662
                self.assertInWorkingTree(p, tree=tree)
2292.1.1 by Marius Kruger
"bzr remove" and "bzr rm" will now remove the working file.
2663
        else:
2664
            self.assertIsNot(tree.path2id(path), None,
2665
                path+' not in working tree.')
2666
2655.2.5 by Marius Kruger
* Improve BzrRemoveChangedFilesError message.
2667
    def assertNotInWorkingTree(self, path, root_path='.', tree=None):
2292.1.1 by Marius Kruger
"bzr remove" and "bzr rm" will now remove the working file.
2668
        """Assert whether path or paths are not in the WorkingTree"""
2669
        if tree is None:
2670
            tree = workingtree.WorkingTree.open(root_path)
2671
        if not isinstance(path, basestring):
2672
            for p in path:
2673
                self.assertNotInWorkingTree(p,tree=tree)
2674
        else:
2675
            self.assertIs(tree.path2id(path), None, path+' in working tree.')
2227.2.1 by v.ladeuil+lp at free
Small fixes to test suite in the hope that it will facilitate the
2676
1123 by Martin Pool
* move bzr-specific code from testsweet into bzrlib.selftest
2677
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
2678
class TestCaseWithTransport(TestCaseInTempDir):
2679
    """A test case that provides get_url and get_readonly_url facilities.
2680
2681
    These back onto two transport servers, one for readonly access and one for
2682
    read write access.
2683
2684
    If no explicit class is provided for readonly access, a
2685
    ReadonlyTransportDecorator is used instead which allows the use of non disk
2686
    based read write transports.
2687
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2688
    If an explicit class is provided for readonly access, that server and the
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
2689
    readwrite one must both define get_url() as resolving to os.getcwd().
2690
    """
2691
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2692
    def get_vfs_only_server(self):
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2693
        """See TestCaseWithMemoryTransport.
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
2694
2695
        This is useful for some tests with specific servers that need
2696
        diagnostics.
2697
        """
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2698
        if self.__vfs_server is None:
2699
            self.__vfs_server = self.vfs_transport_factory()
4659.1.2 by Robert Collins
Refactor creation and shutdown of test servers to use a common helper,
2700
            self.start_server(self.__vfs_server)
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2701
        return self.__vfs_server
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
2702
1666.1.4 by Robert Collins
* 'Metadir' is now the default disk format. This improves behaviour in
2703
    def make_branch_and_tree(self, relpath, format=None):
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
2704
        """Create a branch on the transport and a tree locally.
2705
1910.7.17 by Andrew Bennetts
Various cosmetic changes.
2706
        If the transport is not a LocalTransport, the Tree can't be created on
2381.1.2 by Robert Collins
Fixup the test changes made for hpss to be clean and self contained.
2707
        the transport.  In that case if the vfs_transport_factory is
2708
        LocalURLServer the working tree is created in the local
2018.5.88 by Andrew Bennetts
Clarify make_branch_and_tree docstring a little.
2709
        directory backing the transport, and the returned tree's branch and
2381.1.2 by Robert Collins
Fixup the test changes made for hpss to be clean and self contained.
2710
        repository will also be accessed locally. Otherwise a lightweight
2711
        checkout is created and returned.
1910.7.17 by Andrew Bennetts
Various cosmetic changes.
2712
4650.1.2 by Robert Collins
Remove unnecessary use of an SFTP server connection to test the behaviour of TestCase.make_branch_and_tree.
2713
        We do this because we can't physically create a tree in the local
2714
        path, with a branch reference to the transport_factory url, and
2715
        a branch + repository in the vfs_transport, unless the vfs_transport
2716
        namespace is distinct from the local disk - the two branch objects
2717
        would collide. While we could construct a tree with its branch object
2718
        pointing at the transport_factory transport in memory, reopening it
2719
        would behaving unexpectedly, and has in the past caused testing bugs
2720
        when we tried to do it that way.
2721
1910.7.17 by Andrew Bennetts
Various cosmetic changes.
2722
        :param format: The BzrDirFormat.
2723
        :returns: the WorkingTree.
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
2724
        """
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
2725
        # TODO: always use the local disk path for the working tree,
2726
        # this obviously requires a format that supports branch references
2727
        # so check for that by checking bzrdir.BzrDirFormat.get_default_format()
2728
        # RBC 20060208
1666.1.4 by Robert Collins
* 'Metadir' is now the default disk format. This improves behaviour in
2729
        b = self.make_branch(relpath, format=format)
1534.4.50 by Robert Collins
Got the bzrdir api straightened out, plenty of refactoring to use it pending, but the api is up and running.
2730
        try:
2731
            return b.bzrdir.create_workingtree()
2732
        except errors.NotLocalUrl:
1910.14.1 by Andrew Bennetts
Fix to make_branch_and_tree's behavior when used with an sftp transport.
2733
            # We can only make working trees locally at the moment.  If the
2018.5.87 by Andrew Bennetts
Make make_branch_and_tree fall back to creating a local checkout if the transport doesn't support working trees, allowing several more Remote tests to pass.
2734
            # transport can't support them, then we keep the non-disk-backed
2735
            # branch and create a local checkout.
2381.1.1 by Robert Collins
Split out hpss test fixes which dont depend on new or altered API's.
2736
            if self.vfs_transport_factory is LocalURLServer:
2737
                # the branch is colocated on disk, we cannot create a checkout.
2738
                # hopefully callers will expect this.
2739
                local_controldir= bzrdir.BzrDir.open(self.get_vfs_only_url(relpath))
3489.2.1 by Andrew Bennetts
Fix make_branch_and_tree to return a tree whose .branch is always the right format.
2740
                wt = local_controldir.create_workingtree()
3489.2.5 by Andrew Bennetts
Tweak suggested by John's review.
2741
                if wt.branch._format != b._format:
3489.2.1 by Andrew Bennetts
Fix make_branch_and_tree to return a tree whose .branch is always the right format.
2742
                    wt._branch = b
2743
                    # Make sure that assigning to wt._branch fixes wt.branch,
2744
                    # in case the implementation details of workingtree objects
2745
                    # change.
2746
                    self.assertIs(b, wt.branch)
2747
                return wt
2381.1.1 by Robert Collins
Split out hpss test fixes which dont depend on new or altered API's.
2748
            else:
2749
                return b.create_checkout(relpath, lightweight=True)
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
2750
1553.5.68 by Martin Pool
Add new TestCaseWithTransport.assertIsDirectory() and tests
2751
    def assertIsDirectory(self, relpath, transport):
2752
        """Assert that relpath within transport is a directory.
2753
2754
        This may not be possible on all transports; in that case it propagates
2755
        a TransportNotPossible.
2756
        """
2757
        try:
2758
            mode = transport.stat(relpath).st_mode
2759
        except errors.NoSuchFile:
2760
            self.fail("path %s is not a directory; no such file"
2761
                      % (relpath))
2762
        if not stat.S_ISDIR(mode):
2763
            self.fail("path %s is not a directory; has mode %#o"
2764
                      % (relpath, mode))
2765
1852.13.18 by Robert Collins
Write top level acceptance test for dirstate.
2766
    def assertTreesEqual(self, left, right):
2767
        """Check that left and right have the same content and properties."""
2768
        # we use a tree delta to check for equality of the content, and we
2769
        # manually check for equality of other things such as the parents list.
2770
        self.assertEqual(left.get_parent_ids(), right.get_parent_ids())
2771
        differences = left.changes_from(right)
2255.2.3 by Robert Collins
Split out working tree format 4 to its own file, create stub dirstate revision object, start working on dirstate.set_parent_trees - a key failure point.
2772
        self.assertFalse(differences.has_changed(),
2773
            "Trees %r and %r are different: %r" % (left, right, differences))
1852.13.18 by Robert Collins
Write top level acceptance test for dirstate.
2774
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2775
    def setUp(self):
2776
        super(TestCaseWithTransport, self).setUp()
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2777
        self.__vfs_server = None
1986.2.3 by Robert Collins
New test base class TestCaseWithMemoryTransport offers memory-only
2778
4695.3.2 by Vincent Ladeuil
Simplified and claried as per Robert's review.
2779
    def disable_missing_extensions_warning(self):
2780
        """Some tests expect a precise stderr content.
2781
2782
        There is no point in forcing them to duplicate the extension related
2783
        warning.
2784
        """
2785
        config.GlobalConfig().set_user_option('ignore_missing_extensions', True)
2786
1534.4.10 by Robert Collins
Add TestCaseWithTransport class that provides tests with read and write transport pairs.
2787
1534.4.31 by Robert Collins
cleanedup test_outside_wt
2788
class ChrootedTestCase(TestCaseWithTransport):
2789
    """A support class that provides readonly urls outside the local namespace.
2790
2791
    This is done by checking if self.transport_server is a MemoryServer. if it
2792
    is then we are chrooted already, if it is not then an HttpServer is used
2793
    for readonly urls.
2794
2795
    TODO RBC 20060127: make this an option to TestCaseWithTransport so it can
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2796
                       be used without needed to redo it when a different
1534.4.31 by Robert Collins
cleanedup test_outside_wt
2797
                       subclass is in use ?
2798
    """
2799
2800
    def setUp(self):
2801
        super(ChrootedTestCase, self).setUp()
2018.5.42 by Robert Collins
Various hopefully improvements, but wsgi is broken, handing over to spiv :).
2802
        if not self.vfs_transport_factory == MemoryServer:
2004.1.25 by v.ladeuil+lp at free
Shuffle http related test code. Hopefully it ends up at the right place :)
2803
            self.transport_readonly_server = HttpServer
1534.4.31 by Robert Collins
cleanedup test_outside_wt
2804
2805
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2806
def condition_id_re(pattern):
2807
    """Create a condition filter which performs a re check on a test's id.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2808
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2809
    :param pattern: A regular expression string.
2810
    :return: A callable that returns True if the re matches.
2811
    """
4183.6.5 by Martin Pool
selftest should use re_compile_checked too
2812
    filter_re = osutils.re_compile_checked(pattern, 0,
2813
        'test filter')
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2814
    def condition(test):
2815
        test_id = test.id()
2816
        return filter_re.search(test_id)
2817
    return condition
2818
2819
2921.6.8 by Robert Collins
* New helper function ``bzrlib.tests.condition_isinstance`` which helps
2820
def condition_isinstance(klass_or_klass_list):
2821
    """Create a condition filter which returns isinstance(param, klass).
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2822
2921.6.8 by Robert Collins
* New helper function ``bzrlib.tests.condition_isinstance`` which helps
2823
    :return: A callable which when called with one parameter obj return the
2824
        result of isinstance(obj, klass_or_klass_list).
2825
    """
2826
    def condition(obj):
2827
        return isinstance(obj, klass_or_klass_list)
2828
    return condition
2829
2830
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
2831
def condition_id_in_list(id_list):
3193.1.2 by Vincent Ladeuil
Add condition_id_in_list and filter_suite_by_id_list capabilities.
2832
    """Create a condition filter which verify that test's id in a list.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2833
3302.8.4 by Vincent Ladeuil
Cosmetic changes.
2834
    :param id_list: A TestIdList object.
3193.1.2 by Vincent Ladeuil
Add condition_id_in_list and filter_suite_by_id_list capabilities.
2835
    :return: A callable that returns True if the test's id appears in the list.
2836
    """
2837
    def condition(test):
3302.8.3 by Vincent Ladeuil
Use better names for TestIdList methods.
2838
        return id_list.includes(test.id())
3193.1.2 by Vincent Ladeuil
Add condition_id_in_list and filter_suite_by_id_list capabilities.
2839
    return condition
2840
2841
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2842
def condition_id_startswith(starts):
3302.11.1 by Vincent Ladeuil
Create a new selftest filter allowing loading only one module/class/test.
2843
    """Create a condition filter verifying that test's id starts with a string.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2844
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2845
    :param starts: A list of string.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2846
    :return: A callable that returns True if the test's id starts with one of
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2847
        the given strings.
3302.11.1 by Vincent Ladeuil
Create a new selftest filter allowing loading only one module/class/test.
2848
    """
2849
    def condition(test):
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2850
        for start in starts:
2851
            if test.id().startswith(start):
2852
                return True
2853
        return False
3302.11.1 by Vincent Ladeuil
Create a new selftest filter allowing loading only one module/class/test.
2854
    return condition
2855
2856
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2857
def exclude_tests_by_condition(suite, condition):
2858
    """Create a test suite which excludes some tests from suite.
2859
2860
    :param suite: The suite to get tests from.
2861
    :param condition: A callable whose result evaluates True when called with a
2862
        test case which should be excluded from the result.
2863
    :return: A suite which contains the tests found in suite that fail
2864
        condition.
2865
    """
2866
    result = []
2867
    for test in iter_suite_tests(suite):
2868
        if not condition(test):
2869
            result.append(test)
2870
    return TestUtil.TestSuite(result)
2871
2872
2921.6.7 by Robert Collins
* New helper function ``bzrlib.tests.filter_suite_by_condition`` which
2873
def filter_suite_by_condition(suite, condition):
2874
    """Create a test suite by filtering another one.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2875
2921.6.7 by Robert Collins
* New helper function ``bzrlib.tests.filter_suite_by_condition`` which
2876
    :param suite: The source suite.
2877
    :param condition: A callable whose result evaluates True when called with a
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2878
        test case which should be included in the result.
2921.6.7 by Robert Collins
* New helper function ``bzrlib.tests.filter_suite_by_condition`` which
2879
    :return: A suite which contains the tests found in suite that pass
2880
        condition.
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2881
    """
2921.6.7 by Robert Collins
* New helper function ``bzrlib.tests.filter_suite_by_condition`` which
2882
    result = []
2883
    for test in iter_suite_tests(suite):
2884
        if condition(test):
2885
            result.append(test)
2886
    return TestUtil.TestSuite(result)
2887
2888
3400.3.6 by Martin Pool
Remove code deprecated prior to 1.1 and its tests
2889
def filter_suite_by_re(suite, pattern):
2394.2.8 by Ian Clatworthy
incorporate feedback from jam
2890
    """Create a test suite by filtering another one.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2891
2394.2.8 by Ian Clatworthy
incorporate feedback from jam
2892
    :param suite:           the source suite
2893
    :param pattern:         pattern that names must match
2894
    :returns: the newly created suite
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2895
    """
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2896
    condition = condition_id_re(pattern)
2921.6.7 by Robert Collins
* New helper function ``bzrlib.tests.filter_suite_by_condition`` which
2897
    result_suite = filter_suite_by_condition(suite, condition)
2921.6.4 by Robert Collins
Move the filter implementation of sort_tests_by_re back to filter_tests_by_re.
2898
    return result_suite
2394.2.8 by Ian Clatworthy
incorporate feedback from jam
2899
2900
3193.1.2 by Vincent Ladeuil
Add condition_id_in_list and filter_suite_by_id_list capabilities.
2901
def filter_suite_by_id_list(suite, test_id_list):
2902
    """Create a test suite by filtering another one.
2903
2904
    :param suite: The source suite.
2905
    :param test_id_list: A list of the test ids to keep as strings.
2906
    :returns: the newly created suite
2907
    """
2908
    condition = condition_id_in_list(test_id_list)
2909
    result_suite = filter_suite_by_condition(suite, condition)
2910
    return result_suite
2911
2912
3302.11.1 by Vincent Ladeuil
Create a new selftest filter allowing loading only one module/class/test.
2913
def filter_suite_by_id_startswith(suite, start):
2914
    """Create a test suite by filtering another one.
2915
2916
    :param suite: The source suite.
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2917
    :param start: A list of string the test id must start with one of.
3302.11.1 by Vincent Ladeuil
Create a new selftest filter allowing loading only one module/class/test.
2918
    :returns: the newly created suite
2919
    """
2920
    condition = condition_id_startswith(start)
2921
    result_suite = filter_suite_by_condition(suite, condition)
2922
    return result_suite
2923
2924
2921.6.2 by Robert Collins
* New helper method ``bzrlib.tests.exclude_tests_by_re`` which gives a new
2925
def exclude_tests_by_re(suite, pattern):
2926
    """Create a test suite which excludes some tests from suite.
2927
2928
    :param suite: The suite to get tests from.
2929
    :param pattern: A regular expression string. Test ids that match this
2930
        pattern will be excluded from the result.
2931
    :return: A TestSuite that contains all the tests from suite without the
2932
        tests that matched pattern. The order of tests is the same as it was in
2933
        suite.
2934
    """
2921.6.9 by Robert Collins
* New helper function ``bzrlib.tests.condition_id_re`` which helps
2935
    return exclude_tests_by_condition(suite, condition_id_re(pattern))
2921.6.2 by Robert Collins
* New helper method ``bzrlib.tests.exclude_tests_by_re`` which gives a new
2936
2937
2921.6.6 by Robert Collins
* The ``exclude_pattern`` and ``random_order`` parameters to the function
2938
def preserve_input(something):
2939
    """A helper for performing test suite transformation chains.
2940
2941
    :param something: Anything you want to preserve.
2942
    :return: Something.
2943
    """
2944
    return something
2945
2946
3128.1.2 by Vincent Ladeuil
Tweak as per review feedback: s/randomise.*/randomize&/, 0.92 -> 1.0.
2947
def randomize_suite(suite):
2921.6.6 by Robert Collins
* The ``exclude_pattern`` and ``random_order`` parameters to the function
2948
    """Return a new TestSuite with suite's tests in random order.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2949
2921.6.6 by Robert Collins
* The ``exclude_pattern`` and ``random_order`` parameters to the function
2950
    The tests in the input suite are flattened into a single suite in order to
2951
    accomplish this. Any nested TestSuites are removed to provide global
2952
    randomness.
2953
    """
2921.6.3 by Robert Collins
* New helper method ``bzrlib.tests.randomise_suite`` which returns a
2954
    tests = list(iter_suite_tests(suite))
2955
    random.shuffle(tests)
2956
    return TestUtil.TestSuite(tests)
2957
2958
3350.5.1 by Robert Collins
* New helper function for splitting test suites ``split_suite_by_condition``.
2959
def split_suite_by_condition(suite, condition):
2960
    """Split a test suite into two by a condition.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2961
3350.5.1 by Robert Collins
* New helper function for splitting test suites ``split_suite_by_condition``.
2962
    :param suite: The suite to split.
2963
    :param condition: The condition to match on. Tests that match this
2964
        condition are returned in the first test suite, ones that do not match
2965
        are in the second suite.
2966
    :return: A tuple of two test suites, where the first contains tests from
2967
        suite matching the condition, and the second contains the remainder
2968
        from suite. The order within each output suite is the same as it was in
2969
        suite.
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2970
    """
3350.5.1 by Robert Collins
* New helper function for splitting test suites ``split_suite_by_condition``.
2971
    matched = []
2972
    did_not_match = []
2973
    for test in iter_suite_tests(suite):
2974
        if condition(test):
2975
            matched.append(test)
2976
        else:
2977
            did_not_match.append(test)
2978
    return TestUtil.TestSuite(matched), TestUtil.TestSuite(did_not_match)
2979
2980
2921.6.1 by Robert Collins
* New helper method ``bzrlib.tests.split_suite_by_re`` which splits a test
2981
def split_suite_by_re(suite, pattern):
2982
    """Split a test suite into two by a regular expression.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
2983
2921.6.1 by Robert Collins
* New helper method ``bzrlib.tests.split_suite_by_re`` which splits a test
2984
    :param suite: The suite to split.
2985
    :param pattern: A regular expression string. Test ids that match this
2986
        pattern will be in the first test suite returned, and the others in the
2987
        second test suite returned.
2988
    :return: A tuple of two test suites, where the first contains tests from
2989
        suite matching pattern, and the second contains the remainder from
2990
        suite. The order within each output suite is the same as it was in
2991
        suite.
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
2992
    """
3350.5.1 by Robert Collins
* New helper function for splitting test suites ``split_suite_by_condition``.
2993
    return split_suite_by_condition(suite, condition_id_re(pattern))
2213.2.1 by Martin Pool
Add selftest --first flag
2994
2995
1185.16.58 by mbp at sourcefrog
- run all selftests by default
2996
def run_suite(suite, name='test', verbose=False, pattern=".*",
2485.6.6 by Martin Pool
Put test root directory (containing per-test directories) in TMPDIR
2997
              stop_on_failure=False,
2213.2.1 by Martin Pool
Add selftest --first flag
2998
              transport=None, lsprof_timed=None, bench_history=None,
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
2999
              matching_tests_first=None,
2394.2.1 by Ian Clatworthy
--list and --exclude first cut
3000
              list_only=False,
2394.2.2 by Ian Clatworthy
Add --randomize and update help
3001
              random_seed=None,
2418.4.1 by John Arbash Meinel
(Ian Clatworthy) Bugs #102679, #102686. Add --exclude and --randomize to 'bzr selftest'
3002
              exclude_pattern=None,
4000.2.1 by Robert Collins
Add library level support for different test runners to bzrlib.
3003
              strict=False,
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3004
              runner_class=None,
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
3005
              suite_decorators=None,
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3006
              stream=None,
3007
              result_decorators=None,
3008
              ):
4000.2.1 by Robert Collins
Add library level support for different test runners to bzrlib.
3009
    """Run a test suite for bzr selftest.
3010
3011
    :param runner_class: The class of runner to use. Must support the
3012
        constructor arguments passed by run_suite which are more than standard
3013
        python uses.
3014
    :return: A boolean indicating success.
3015
    """
2379.6.3 by Alexander Belchenko
Rework NUMBERED_DIRS usage to keep test_selftest.py passing the tests on win32
3016
    TestCase._gather_lsprof_in_benchmarks = lsprof_timed
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
3017
    if verbose:
3018
        verbosity = 2
3019
    else:
3020
        verbosity = 1
4000.2.1 by Robert Collins
Add library level support for different test runners to bzrlib.
3021
    if runner_class is None:
3022
        runner_class = TextTestRunner
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
3023
    if stream is None:
3024
        stream = sys.stdout
3025
    runner = runner_class(stream=stream,
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
3026
                            descriptions=0,
1534.11.1 by Robert Collins
Teach bzr selftest to use a progress bar in non verbose mode.
3027
                            verbosity=verbosity,
2379.6.3 by Alexander Belchenko
Rework NUMBERED_DIRS usage to keep test_selftest.py passing the tests on win32
3028
                            bench_history=bench_history,
4271.2.1 by Robert Collins
Move test prelude and suffix output to ExtendedTestResult
3029
                            strict=strict,
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3030
                            result_decorators=result_decorators,
2379.6.3 by Alexander Belchenko
Rework NUMBERED_DIRS usage to keep test_selftest.py passing the tests on win32
3031
                            )
1185.16.58 by mbp at sourcefrog
- run all selftests by default
3032
    runner.stop_on_failure=stop_on_failure
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3033
    # built in decorator factories:
3034
    decorators = [
3035
        random_order(random_seed, runner),
3036
        exclude_tests(exclude_pattern),
3037
        ]
3038
    if matching_tests_first:
3039
        decorators.append(tests_first(pattern))
3040
    else:
3041
        decorators.append(filter_tests(pattern))
3042
    if suite_decorators:
3043
        decorators.extend(suite_decorators)
4618.1.1 by Vincent Ladeuil
Make --parallel=fork work again.
3044
    # tell the result object how many tests will be running: (except if
4618.1.2 by Vincent Ladeuil
Fixed as per John's review.
3045
    # --parallel=fork is being used. Robert said he will provide a better
4618.1.1 by Vincent Ladeuil
Make --parallel=fork work again.
3046
    # progress design later -- vila 20090817)
4618.1.2 by Vincent Ladeuil
Fixed as per John's review.
3047
    if fork_decorator not in decorators:
4618.1.1 by Vincent Ladeuil
Make --parallel=fork work again.
3048
        decorators.append(CountingDecorator)
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3049
    for decorator in decorators:
3050
        suite = decorator(suite)
4266.2.1 by Robert Collins
Remove noise from bzr selftest --list-only so that it is easier to use in scripts.
3051
    if list_only:
4650.1.9 by Robert Collins
Detangle test listing: its more part of the ui layer not the execute-this-test-layer.
3052
        # Done after test suite decoration to allow randomisation etc
3053
        # to take effect, though that is of marginal benefit.
3054
        if verbosity >= 2:
3055
            stream.write("Listing tests only ...\n")
3056
        for t in iter_suite_tests(suite):
3057
            stream.write("%s\n" % (t.id()))
4266.2.1 by Robert Collins
Remove noise from bzr selftest --list-only so that it is easier to use in scripts.
3058
        return True
4650.1.9 by Robert Collins
Detangle test listing: its more part of the ui layer not the execute-this-test-layer.
3059
    result = runner.run(suite)
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3060
    if strict:
3061
        return result.wasStrictlySuccessful()
3062
    else:
3063
        return result.wasSuccessful()
3064
3065
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3066
# A registry where get() returns a suite decorator.
3067
parallel_registry = registry.Registry()
4229.3.1 by Vincent Ladeuil
Fix selftest --parallel for ConcurrentTestSuite uses.
3068
3069
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3070
def fork_decorator(suite):
4398.4.3 by Vincent Ladeuil
Detect # cores on win32 and Solaris too.
3071
    concurrency = osutils.local_concurrency()
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3072
    if concurrency == 1:
3073
        return suite
3074
    from testtools import ConcurrentTestSuite
3075
    return ConcurrentTestSuite(suite, fork_for_tests)
3076
parallel_registry.register('fork', fork_decorator)
4229.3.1 by Vincent Ladeuil
Fix selftest --parallel for ConcurrentTestSuite uses.
3077
3078
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3079
def subprocess_decorator(suite):
4398.4.3 by Vincent Ladeuil
Detect # cores on win32 and Solaris too.
3080
    concurrency = osutils.local_concurrency()
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3081
    if concurrency == 1:
3082
        return suite
3083
    from testtools import ConcurrentTestSuite
3084
    return ConcurrentTestSuite(suite, reinvoke_for_tests)
3085
parallel_registry.register('subprocess', subprocess_decorator)
3086
3087
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3088
def exclude_tests(exclude_pattern):
3089
    """Return a test suite decorator that excludes tests."""
3090
    if exclude_pattern is None:
3091
        return identity_decorator
3092
    def decorator(suite):
3093
        return ExcludeDecorator(suite, exclude_pattern)
3094
    return decorator
3095
3096
3097
def filter_tests(pattern):
3098
    if pattern == '.*':
3099
        return identity_decorator
3100
    def decorator(suite):
3101
        return FilterTestsDecorator(suite, pattern)
3102
    return decorator
3103
3104
3105
def random_order(random_seed, runner):
3106
    """Return a test suite decorator factory for randomising tests order.
3107
    
3108
    :param random_seed: now, a string which casts to a long, or a long.
3109
    :param runner: A test runner with a stream attribute to report on.
3110
    """
3111
    if random_seed is None:
3112
        return identity_decorator
3113
    def decorator(suite):
3114
        return RandomDecorator(suite, random_seed, runner.stream)
3115
    return decorator
3116
3117
3118
def tests_first(pattern):
3119
    if pattern == '.*':
3120
        return identity_decorator
3121
    def decorator(suite):
3122
        return TestFirstDecorator(suite, pattern)
3123
    return decorator
3124
3125
3126
def identity_decorator(suite):
3127
    """Return suite."""
3128
    return suite
3129
3130
3131
class TestDecorator(TestSuite):
3132
    """A decorator for TestCase/TestSuite objects.
3133
    
3134
    Usually, subclasses should override __iter__(used when flattening test
3135
    suites), which we do to filter, reorder, parallelise and so on, run() and
3136
    debug().
3137
    """
3138
3139
    def __init__(self, suite):
3140
        TestSuite.__init__(self)
3141
        self.addTest(suite)
3142
3143
    def countTestCases(self):
3144
        cases = 0
3145
        for test in self:
3146
            cases += test.countTestCases()
3147
        return cases
3148
3149
    def debug(self):
3150
        for test in self:
3151
            test.debug()
3152
3153
    def run(self, result):
3154
        # Use iteration on self, not self._tests, to allow subclasses to hook
3155
        # into __iter__.
3156
        for test in self:
3157
            if result.shouldStop:
3158
                break
3159
            test.run(result)
3160
        return result
3161
3162
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
3163
class CountingDecorator(TestDecorator):
3164
    """A decorator which calls result.progress(self.countTestCases)."""
3165
3166
    def run(self, result):
3167
        progress_method = getattr(result, 'progress', None)
3168
        if callable(progress_method):
4573.2.3 by Robert Collins
Support python 2.4.
3169
            progress_method(self.countTestCases(), SUBUNIT_SEEK_SET)
4573.2.1 by Robert Collins
Don't call countTestCases from TextTestRunner.run, rather let tests decide if they want to be counted.
3170
        return super(CountingDecorator, self).run(result)
3171
3172
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3173
class ExcludeDecorator(TestDecorator):
3174
    """A decorator which excludes test matching an exclude pattern."""
3175
3176
    def __init__(self, suite, exclude_pattern):
3177
        TestDecorator.__init__(self, suite)
3178
        self.exclude_pattern = exclude_pattern
3179
        self.excluded = False
3180
3181
    def __iter__(self):
3182
        if self.excluded:
3183
            return iter(self._tests)
3184
        self.excluded = True
3185
        suite = exclude_tests_by_re(self, self.exclude_pattern)
3186
        del self._tests[:]
3187
        self.addTests(suite)
3188
        return iter(self._tests)
3189
3190
3191
class FilterTestsDecorator(TestDecorator):
3192
    """A decorator which filters tests to those matching a pattern."""
3193
3194
    def __init__(self, suite, pattern):
3195
        TestDecorator.__init__(self, suite)
3196
        self.pattern = pattern
3197
        self.filtered = False
3198
3199
    def __iter__(self):
3200
        if self.filtered:
3201
            return iter(self._tests)
3202
        self.filtered = True
3203
        suite = filter_suite_by_re(self, self.pattern)
3204
        del self._tests[:]
3205
        self.addTests(suite)
3206
        return iter(self._tests)
3207
3208
3209
class RandomDecorator(TestDecorator):
3210
    """A decorator which randomises the order of its tests."""
3211
3212
    def __init__(self, suite, random_seed, stream):
3213
        TestDecorator.__init__(self, suite)
3214
        self.random_seed = random_seed
3215
        self.randomised = False
3216
        self.stream = stream
3217
3218
    def __iter__(self):
3219
        if self.randomised:
3220
            return iter(self._tests)
3221
        self.randomised = True
3222
        self.stream.writeln("Randomizing test order using seed %s\n" %
3223
            (self.actual_seed()))
3224
        # Initialise the random number generator.
3225
        random.seed(self.actual_seed())
3226
        suite = randomize_suite(self)
3227
        del self._tests[:]
3228
        self.addTests(suite)
3229
        return iter(self._tests)
3230
3231
    def actual_seed(self):
3232
        if self.random_seed == "now":
3233
            # We convert the seed to a long to make it reuseable across
3234
            # invocations (because the user can reenter it).
3235
            self.random_seed = long(time.time())
2394.2.2 by Ian Clatworthy
Add --randomize and update help
3236
        else:
3237
            # Convert the seed to a long if we can
3238
            try:
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3239
                self.random_seed = long(self.random_seed)
2394.2.2 by Ian Clatworthy
Add --randomize and update help
3240
            except:
3241
                pass
4205.3.1 by Robert Collins
Refactor tests.run_suite to be more data driven, making it shorter and able to be extended more easily.
3242
        return self.random_seed
3243
3244
3245
class TestFirstDecorator(TestDecorator):
3246
    """A decorator which moves named tests to the front."""
3247
3248
    def __init__(self, suite, pattern):
3249
        TestDecorator.__init__(self, suite)
3250
        self.pattern = pattern
3251
        self.filtered = False
3252
3253
    def __iter__(self):
3254
        if self.filtered:
3255
            return iter(self._tests)
3256
        self.filtered = True
3257
        suites = split_suite_by_re(self, self.pattern)
3258
        del self._tests[:]
3259
        self.addTests(suites)
3260
        return iter(self._tests)
1393.1.6 by Martin Pool
- fold testsweet into bzrlib.selftest
3261
3262
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3263
def partition_tests(suite, count):
3264
    """Partition suite into count lists of tests."""
3265
    result = []
3266
    tests = list(iter_suite_tests(suite))
3267
    tests_per_process = int(math.ceil(float(len(tests)) / count))
3268
    for block in range(count):
3269
        low_test = block * tests_per_process
3270
        high_test = low_test + tests_per_process
3271
        process_tests = tests[low_test:high_test]
3272
        result.append(process_tests)
3273
    return result
3274
3275
3276
def fork_for_tests(suite):
3277
    """Take suite and start up one runner per CPU by forking()
3278
3279
    :return: An iterable of TestCase-like objects which can each have
4229.3.1 by Vincent Ladeuil
Fix selftest --parallel for ConcurrentTestSuite uses.
3280
        run(result) called on them to feed tests to result.
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3281
    """
4398.4.3 by Vincent Ladeuil
Detect # cores on win32 and Solaris too.
3282
    concurrency = osutils.local_concurrency()
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3283
    result = []
3284
    from subunit import TestProtocolClient, ProtocolTestCase
3285
    class TestInOtherProcess(ProtocolTestCase):
3286
        # Should be in subunit, I think. RBC.
3287
        def __init__(self, stream, pid):
3288
            ProtocolTestCase.__init__(self, stream)
3289
            self.pid = pid
3290
3291
        def run(self, result):
3292
            try:
3293
                ProtocolTestCase.run(self, result)
3294
            finally:
3295
                os.waitpid(self.pid, os.WNOHANG)
3296
3297
    test_blocks = partition_tests(suite, concurrency)
3298
    for process_tests in test_blocks:
3299
        process_suite = TestSuite()
3300
        process_suite.addTests(process_tests)
3301
        c2pread, c2pwrite = os.pipe()
3302
        pid = os.fork()
3303
        if pid == 0:
3304
            try:
3305
                os.close(c2pread)
3306
                # Leave stderr and stdout open so we can see test noise
3307
                # Close stdin so that the child goes away if it decides to
3308
                # read from stdin (otherwise its a roulette to see what
4229.3.1 by Vincent Ladeuil
Fix selftest --parallel for ConcurrentTestSuite uses.
3309
                # child actually gets keystrokes for pdb etc).
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3310
                sys.stdin.close()
3311
                sys.stdin = None
4229.3.1 by Vincent Ladeuil
Fix selftest --parallel for ConcurrentTestSuite uses.
3312
                stream = os.fdopen(c2pwrite, 'wb', 1)
4780.1.6 by Vincent Ladeuil
Make --parallel=fork compatible with --subunit.
3313
                subunit_result = BzrAutoTimingTestResultDecorator(
4557.2.1 by Robert Collins
Support timestamping subunit streams.
3314
                    TestProtocolClient(stream))
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3315
                process_suite.run(subunit_result)
3316
            finally:
3317
                os._exit(0)
3318
        else:
3319
            os.close(c2pwrite)
3320
            stream = os.fdopen(c2pread, 'rb', 1)
3321
            test = TestInOtherProcess(stream, pid)
3322
            result.append(test)
3323
    return result
3324
3325
3326
def reinvoke_for_tests(suite):
3327
    """Take suite and start up one runner per CPU using subprocess().
3328
3329
    :return: An iterable of TestCase-like objects which can each have
4229.3.1 by Vincent Ladeuil
Fix selftest --parallel for ConcurrentTestSuite uses.
3330
        run(result) called on them to feed tests to result.
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3331
    """
4398.4.3 by Vincent Ladeuil
Detect # cores on win32 and Solaris too.
3332
    concurrency = osutils.local_concurrency()
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3333
    result = []
4496.3.6 by Andrew Bennetts
Tidy some more imports.
3334
    from subunit import ProtocolTestCase
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3335
    class TestInSubprocess(ProtocolTestCase):
3336
        def __init__(self, process, name):
3337
            ProtocolTestCase.__init__(self, process.stdout)
3338
            self.process = process
3339
            self.process.stdin.close()
3340
            self.name = name
3341
3342
        def run(self, result):
3343
            try:
3344
                ProtocolTestCase.run(self, result)
3345
            finally:
3346
                self.process.wait()
3347
                os.unlink(self.name)
3348
            # print "pid %d finished" % finished_process
3349
    test_blocks = partition_tests(suite, concurrency)
3350
    for process_tests in test_blocks:
3351
        # ugly; currently reimplement rather than reuses TestCase methods.
3352
        bzr_path = os.path.dirname(os.path.dirname(bzrlib.__file__))+'/bzr'
3353
        if not os.path.isfile(bzr_path):
3354
            # We are probably installed. Assume sys.argv is the right file
3355
            bzr_path = sys.argv[0]
4805.2.1 by Gordon Tyler
Fixed reinvoke_for_tests to work on win32.
3356
        bzr_path = [bzr_path]
3357
        if sys.platform == "win32":
3358
            # if we're on windows, we can't execute the bzr script directly
3359
            bzr_path = [sys.executable] + bzr_path
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3360
        fd, test_list_file_name = tempfile.mkstemp()
3361
        test_list_file = os.fdopen(fd, 'wb', 1)
3362
        for test in process_tests:
3363
            test_list_file.write(test.id() + '\n')
3364
        test_list_file.close()
3365
        try:
4805.2.1 by Gordon Tyler
Fixed reinvoke_for_tests to work on win32.
3366
            argv = bzr_path + ['selftest', '--load-list', test_list_file_name,
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3367
                '--subunit']
3368
            if '--no-plugins' in sys.argv:
3369
                argv.append('--no-plugins')
3370
            # stderr=STDOUT would be ideal, but until we prevent noise on
3371
            # stderr it can interrupt the subunit protocol.
3372
            process = Popen(argv, stdin=PIPE, stdout=PIPE, stderr=PIPE,
3373
                bufsize=1)
3374
            test = TestInSubprocess(process, test_list_file_name)
3375
            result.append(test)
3376
        except:
3377
            os.unlink(test_list_file_name)
3378
            raise
3379
    return result
3380
3381
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3382
class ForwardingResult(unittest.TestResult):
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3383
3384
    def __init__(self, target):
3385
        unittest.TestResult.__init__(self)
3386
        self.result = target
3387
3388
    def startTest(self, test):
3389
        self.result.startTest(test)
3390
3391
    def stopTest(self, test):
3392
        self.result.stopTest(test)
3393
4650.1.6 by Robert Collins
Fix interface skew between bzr selftest and python unittest - use stopTestRun not done to end test runs.
3394
    def startTestRun(self):
3395
        self.result.startTestRun()
3396
3397
    def stopTestRun(self):
4650.1.7 by Robert Collins
Push result reporting thoroughly into TestResult.
3398
        self.result.stopTestRun()
4650.1.6 by Robert Collins
Fix interface skew between bzr selftest and python unittest - use stopTestRun not done to end test runs.
3399
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3400
    def addSkip(self, test, reason):
3401
        self.result.addSkip(test, reason)
3402
3403
    def addSuccess(self, test):
3404
        self.result.addSuccess(test)
3405
3406
    def addError(self, test, err):
3407
        self.result.addError(test, err)
3408
3409
    def addFailure(self, test, err):
3410
        self.result.addFailure(test, err)
3411
3412
3413
class BZRTransformingResult(ForwardingResult):
3414
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3415
    def addError(self, test, err):
3416
        feature = self._error_looks_like('UnavailableFeature: ', err)
3417
        if feature is not None:
3418
            self.result.addNotSupported(test, feature)
3419
        else:
3420
            self.result.addError(test, err)
3421
3422
    def addFailure(self, test, err):
3423
        known = self._error_looks_like('KnownFailure: ', err)
3424
        if known is not None:
4780.1.4 by Robert Collins
Switch reporting of KnownFailure to be Python2.7 compatible.
3425
            self.result.addExpectedFailure(test,
3426
                [KnownFailure, KnownFailure(known), None])
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3427
        else:
3428
            self.result.addFailure(test, err)
3429
3430
    def _error_looks_like(self, prefix, err):
3431
        """Deserialize exception and returns the stringify value."""
3432
        import subunit
3433
        value = None
3434
        typ, exc, _ = err
3435
        if isinstance(exc, subunit.RemoteException):
3436
            # stringify the exception gives access to the remote traceback
3437
            # We search the last line for 'prefix'
3438
            lines = str(exc).split('\n')
4266.1.1 by Robert Collins
Sniff KnownFailure etc harder when reading subunit results.
3439
            while lines and not lines[-1]:
3440
                lines.pop(-1)
3441
            if lines:
3442
                if lines[-1].startswith(prefix):
3443
                    value = lines[-1][len(prefix):]
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3444
        return value
3445
3446
4780.1.6 by Vincent Ladeuil
Make --parallel=fork compatible with --subunit.
3447
try:
3448
    from subunit.test_results import AutoTimingTestResultDecorator
3449
    # Expected failure should be seen as a success not a failure Once subunit
3450
    # provide native support for that, BZRTransformingResult and this class
3451
    # will become useless.
3452
    class BzrAutoTimingTestResultDecorator(AutoTimingTestResultDecorator):
3453
3454
        def addExpectedFailure(self, test, err):
3455
            self._before_event()
3456
            return self._call_maybe("addExpectedFailure", self._degrade_skip,
3457
                                    test, err)
3458
except ImportError:
3459
    # Let's just define a no-op decorator
3460
    BzrAutoTimingTestResultDecorator = lambda x:x
3461
3462
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3463
class ProfileResult(ForwardingResult):
3464
    """Generate profiling data for all activity between start and success.
3465
    
3466
    The profile data is appended to the test's _benchcalls attribute and can
3467
    be accessed by the forwarded-to TestResult.
3468
3469
    While it might be cleaner do accumulate this in stopTest, addSuccess is
3470
    where our existing output support for lsprof is, and this class aims to
3471
    fit in with that: while it could be moved it's not necessary to accomplish
4641.3.4 by Robert Collins
Fix typo.
3472
    test profiling, nor would it be dramatically cleaner.
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3473
    """
3474
3475
    def startTest(self, test):
3476
        self.profiler = bzrlib.lsprof.BzrProfiler()
3477
        self.profiler.start()
3478
        ForwardingResult.startTest(self, test)
3479
3480
    def addSuccess(self, test):
3481
        stats = self.profiler.stop()
3482
        try:
3483
            calls = test._benchcalls
3484
        except AttributeError:
3485
            test._benchcalls = []
3486
            calls = test._benchcalls
3487
        calls.append(((test.id(), "", ""), stats))
3488
        ForwardingResult.addSuccess(self, test)
3489
3490
    def stopTest(self, test):
3491
        ForwardingResult.stopTest(self, test)
3492
        self.profiler = None
3493
3494
3390.1.1 by Andrew Bennetts
Add --debugflags/-E option to selftest.
3495
# Controlled by "bzr selftest -E=..." option
4523.4.9 by John Arbash Meinel
Change the flags around a bit.
3496
# Currently supported:
3497
#   -Eallow_debug           Will no longer clear debug.debug_flags() so it
3498
#                           preserves any flags supplied at the command line.
3499
#   -Edisable_lock_checks   Turns errors in mismatched locks into simple prints
3500
#                           rather than failing tests. And no longer raise
3501
#                           LockContention when fctnl locks are not being used
3502
#                           with proper exclusion rules.
3390.1.1 by Andrew Bennetts
Add --debugflags/-E option to selftest.
3503
selftest_debug_flags = set()
3504
3505
1185.35.20 by Aaron Bentley
Only keep test failure directories if --keep-output is specified
3506
def selftest(verbose=False, pattern=".*", stop_on_failure=True,
1707.2.1 by Robert Collins
'bzr selftest --benchmark' will run a new benchmarking selftest.
3507
             transport=None,
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
3508
             test_suite_factory=None,
1819.1.3 by Carl Friedrich Bolz
(lifeless, cfbolz): Add recording of benchmark results to the benchmark history
3509
             lsprof_timed=None,
2213.2.1 by Martin Pool
Add selftest --first flag
3510
             bench_history=None,
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
3511
             matching_tests_first=None,
2394.2.1 by Ian Clatworthy
--list and --exclude first cut
3512
             list_only=False,
2394.2.2 by Ian Clatworthy
Add --randomize and update help
3513
             random_seed=None,
2658.3.3 by Daniel Watkins
Added UI support for bzr selftest --strict.
3514
             exclude_pattern=None,
3515
             strict=False,
3193.1.8 by Vincent Ladeuil
Add '--load-list' option to selftest.
3516
             load_list=None,
3390.1.1 by Andrew Bennetts
Add --debugflags/-E option to selftest.
3517
             debug_flags=None,
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3518
             starting_with=None,
4000.2.3 by Robert Collins
Allow extra options to bzrlib.tests.selftest from plugins.
3519
             runner_class=None,
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3520
             suite_decorators=None,
4636.2.1 by Robert Collins
Test selftest --list-only and --randomize options using more precisely layers.
3521
             stream=None,
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3522
             lsprof_tests=False,
2658.3.3 by Daniel Watkins
Added UI support for bzr selftest --strict.
3523
             ):
1204 by Martin Pool
doc
3524
    """Run the whole test suite under the enhanced runner"""
1904.2.5 by Martin Pool
Fix format warning inside test suite and add test
3525
    # XXX: Very ugly way to do this...
3526
    # Disable warning about old formats because we don't want it to disturb
3527
    # any blackbox tests.
3528
    from bzrlib import repository
3529
    repository._deprecation_warning_done = True
3530
1534.4.25 by Robert Collins
Add a --transport parameter to the test suite to set the default transport to be used in the test suite.
3531
    global default_transport
3532
    if transport is None:
3533
        transport = default_transport
3534
    old_transport = default_transport
3535
    default_transport = transport
3390.1.1 by Andrew Bennetts
Add --debugflags/-E option to selftest.
3536
    global selftest_debug_flags
3537
    old_debug_flags = selftest_debug_flags
3538
    if debug_flags is not None:
3539
        selftest_debug_flags = set(debug_flags)
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
3540
    try:
3193.1.8 by Vincent Ladeuil
Add '--load-list' option to selftest.
3541
        if load_list is None:
3542
            keep_only = None
3543
        else:
3544
            keep_only = load_test_id_list(load_list)
4641.2.1 by Robert Collins
Resolve test aliases at the outermost level that test skip filtering is done.
3545
        if starting_with:
3546
            starting_with = [test_prefix_alias_registry.resolve_alias(start)
3547
                             for start in starting_with]
1707.2.1 by Robert Collins
'bzr selftest --benchmark' will run a new benchmarking selftest.
3548
        if test_suite_factory is None:
4636.2.4 by Robert Collins
Move selftest internals tests out of blackbox test space - they are not testing our selftest command line.
3549
            # Reduce loading time by loading modules based on the starting_with
3550
            # patterns.
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3551
            suite = test_suite(keep_only, starting_with)
1707.2.1 by Robert Collins
'bzr selftest --benchmark' will run a new benchmarking selftest.
3552
        else:
3553
            suite = test_suite_factory()
4636.2.4 by Robert Collins
Move selftest internals tests out of blackbox test space - they are not testing our selftest command line.
3554
        if starting_with:
3555
            # But always filter as requested.
3556
            suite = filter_suite_by_id_startswith(suite, starting_with)
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3557
        result_decorators = []
3558
        if lsprof_tests:
3559
            result_decorators.append(ProfileResult)
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
3560
        return run_suite(suite, 'testbzr', verbose=verbose, pattern=pattern,
2485.6.6 by Martin Pool
Put test root directory (containing per-test directories) in TMPDIR
3561
                     stop_on_failure=stop_on_failure,
1725.1.1 by Robert Collins
'bzr selftest --benchmark --lsprof-timed' will use lsprofile to generate
3562
                     transport=transport,
1819.1.3 by Carl Friedrich Bolz
(lifeless, cfbolz): Add recording of benchmark results to the benchmark history
3563
                     lsprof_timed=lsprof_timed,
2213.2.1 by Martin Pool
Add selftest --first flag
3564
                     bench_history=bench_history,
2321.2.1 by Alexander Belchenko
`bzr selftest --numbered-dirs` use numbered dirs for TestCaseInTempDir
3565
                     matching_tests_first=matching_tests_first,
2394.2.1 by Ian Clatworthy
--list and --exclude first cut
3566
                     list_only=list_only,
2394.2.2 by Ian Clatworthy
Add --randomize and update help
3567
                     random_seed=random_seed,
2658.3.3 by Daniel Watkins
Added UI support for bzr selftest --strict.
3568
                     exclude_pattern=exclude_pattern,
4000.2.3 by Robert Collins
Allow extra options to bzrlib.tests.selftest from plugins.
3569
                     strict=strict,
3570
                     runner_class=runner_class,
4205.3.2 by Robert Collins
Create fork and reinvoke parallel testing support.
3571
                     suite_decorators=suite_decorators,
4636.2.1 by Robert Collins
Test selftest --list-only and --randomize options using more precisely layers.
3572
                     stream=stream,
4641.3.3 by Robert Collins
Enable --lsprof-tests on bzr selftest.
3573
                     result_decorators=result_decorators,
4000.2.3 by Robert Collins
Allow extra options to bzrlib.tests.selftest from plugins.
3574
                     )
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
3575
    finally:
3576
        default_transport = old_transport
3390.1.1 by Andrew Bennetts
Add --debugflags/-E option to selftest.
3577
        selftest_debug_flags = old_debug_flags
1534.4.26 by Robert Collins
Move working tree initialisation out from Branch.initialize, deprecated Branch.initialize to Branch.create.
3578
1092.1.17 by Robert Collins
remove TEST_CLASSES dead code and provide a bzrlib.test_suite() convenience method
3579
3193.1.7 by Vincent Ladeuil
Load test id list from a text file.
3580
def load_test_id_list(file_name):
3581
    """Load a test id list from a text file.
3582
3583
    The format is one test id by line.  No special care is taken to impose
3584
    strict rules, these test ids are used to filter the test suite so a test id
3585
    that do not match an existing test will do no harm. This allows user to add
3586
    comments, leave blank lines, etc.
3587
    """
3588
    test_list = []
3589
    try:
3590
        ftest = open(file_name, 'rt')
3591
    except IOError, e:
3592
        if e.errno != errno.ENOENT:
3593
            raise
3594
        else:
3595
            raise errors.NoSuchFile(file_name)
3596
3597
    for test_name in ftest.readlines():
3598
        test_list.append(test_name.strip())
3599
    ftest.close()
3600
    return test_list
3601
3302.3.3 by Vincent Ladeuil
Fix PEP8 catched by Aaron. Update NEWS.
3602
3302.3.1 by Vincent Ladeuil
Help identify duplicates IDs in test suite and missing tests in id
3603
def suite_matches_id_list(test_suite, id_list):
3604
    """Warns about tests not appearing or appearing more than once.
3605
3606
    :param test_suite: A TestSuite object.
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
3607
    :param test_id_list: The list of test ids that should be found in
3302.3.1 by Vincent Ladeuil
Help identify duplicates IDs in test suite and missing tests in id
3608
         test_suite.
3609
3610
    :return: (absents, duplicates) absents is a list containing the test found
3611
        in id_list but not in test_suite, duplicates is a list containing the
3612
        test found multiple times in test_suite.
3613
3614
    When using a prefined test id list, it may occurs that some tests do not
3615
    exist anymore or that some tests use the same id. This function warns the
3616
    tester about potential problems in his workflow (test lists are volatile)
3617
    or in the test suite itself (using the same id for several tests does not
3618
    help to localize defects).
3619
    """
3620
    # Build a dict counting id occurrences
3621
    tests = dict()
3622
    for test in iter_suite_tests(test_suite):
3623
        id = test.id()
3624
        tests[id] = tests.get(id, 0) + 1
3625
3626
    not_found = []
3627
    duplicates = []
3628
    for id in id_list:
3629
        occurs = tests.get(id, 0)
3630
        if not occurs:
3631
            not_found.append(id)
3632
        elif occurs > 1:
3633
            duplicates.append(id)
3634
3635
    return not_found, duplicates
3636
3193.1.7 by Vincent Ladeuil
Load test id list from a text file.
3637
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
3638
class TestIdList(object):
3639
    """Test id list to filter a test suite.
3193.1.3 by Vincent Ladeuil
Create a TestIdListFilter helper object to make testing easier.
3640
3641
    Relying on the assumption that test ids are built as:
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
3642
    <module>[.<class>.<method>][(<param>+)], <module> being in python dotted
3643
    notation, this class offers methods to :
3193.1.4 by Vincent Ladeuil
Make TestTestIdListFilter aware that a test exists for a module or one of
3644
    - avoid building a test suite for modules not refered to in the test list,
3193.1.3 by Vincent Ladeuil
Create a TestIdListFilter helper object to make testing easier.
3645
    - keep only the tests listed from the module test suite.
3646
    """
3647
3648
    def __init__(self, test_id_list):
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
3649
        # When a test suite needs to be filtered against us we compare test ids
3650
        # for equality, so a simple dict offers a quick and simple solution.
3651
        self.tests = dict().fromkeys(test_id_list, True)
3652
3653
        # While unittest.TestCase have ids like:
3654
        # <module>.<class>.<method>[(<param+)],
3655
        # doctest.DocTestCase can have ids like:
3656
        # <module>
3657
        # <module>.<class>
3658
        # <module>.<function>
3659
        # <module>.<class>.<method>
3660
3661
        # Since we can't predict a test class from its name only, we settle on
3662
        # a simple constraint: a test id always begins with its module name.
3663
3664
        modules = {}
3193.1.3 by Vincent Ladeuil
Create a TestIdListFilter helper object to make testing easier.
3665
        for test_id in test_id_list:
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
3666
            parts = test_id.split('.')
3667
            mod_name = parts.pop(0)
3668
            modules[mod_name] = True
3669
            for part in parts:
3670
                mod_name += '.' + part
3671
                modules[mod_name] = True
3672
        self.modules = modules
3193.1.4 by Vincent Ladeuil
Make TestTestIdListFilter aware that a test exists for a module or one of
3673
3302.8.3 by Vincent Ladeuil
Use better names for TestIdList methods.
3674
    def refers_to(self, module_name):
3193.1.4 by Vincent Ladeuil
Make TestTestIdListFilter aware that a test exists for a module or one of
3675
        """Is there tests for the module or one of its sub modules."""
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
3676
        return self.modules.has_key(module_name)
3677
3302.8.3 by Vincent Ladeuil
Use better names for TestIdList methods.
3678
    def includes(self, test_id):
3193.1.11 by Vincent Ladeuil
Relax constraint on test ids, simplify implementation and update tests.
3679
        return self.tests.has_key(test_id)
3193.1.5 by Vincent Ladeuil
Add helper method to get only listed tests from a module test suite.
3680
3193.1.3 by Vincent Ladeuil
Create a TestIdListFilter helper object to make testing easier.
3681
3649.6.2 by Vincent Ladeuil
Replace aliases in selftest --starting-with option.
3682
class TestPrefixAliasRegistry(registry.Registry):
3683
    """A registry for test prefix aliases.
3684
3685
    This helps implement shorcuts for the --starting-with selftest
3686
    option. Overriding existing prefixes is not allowed but not fatal (a
3687
    warning will be emitted).
3688
    """
3689
3690
    def register(self, key, obj, help=None, info=None,
3691
                 override_existing=False):
3692
        """See Registry.register.
3693
3694
        Trying to override an existing alias causes a warning to be emitted,
3695
        not a fatal execption.
3696
        """
3697
        try:
3698
            super(TestPrefixAliasRegistry, self).register(
3699
                key, obj, help=help, info=info, override_existing=False)
3700
        except KeyError:
3701
            actual = self.get(key)
3702
            note('Test prefix alias %s is already used for %s, ignoring %s'
3703
                 % (key, actual, obj))
3704
3705
    def resolve_alias(self, id_start):
3706
        """Replace the alias by the prefix in the given string.
3707
3649.6.3 by Vincent Ladeuil
Fixed typos as per John's review.
3708
        Using an unknown prefix is an error to help catching typos.
3649.6.2 by Vincent Ladeuil
Replace aliases in selftest --starting-with option.
3709
        """
3710
        parts = id_start.split('.')
3711
        try:
3712
            parts[0] = self.get(parts[0])
3713
        except KeyError:
3714
            raise errors.BzrCommandError(
3715
                '%s is not a known test prefix alias' % parts[0])
3716
        return '.'.join(parts)
3717
3718
3719
test_prefix_alias_registry = TestPrefixAliasRegistry()
3649.6.3 by Vincent Ladeuil
Fixed typos as per John's review.
3720
"""Registry of test prefix aliases."""
3649.6.2 by Vincent Ladeuil
Replace aliases in selftest --starting-with option.
3721
3722
3723
# This alias allows to detect typos ('bzrlin.') by making all valid test ids
3724
# appear prefixed ('bzrlib.' is "replaced" by 'bzrlib.').
3725
test_prefix_alias_registry.register('bzrlib', 'bzrlib')
3726
3727
# Obvious higest levels prefixes, feel free to add your own via a plugin
3728
test_prefix_alias_registry.register('bd', 'bzrlib.doc')
3729
test_prefix_alias_registry.register('bu', 'bzrlib.utils')
3730
test_prefix_alias_registry.register('bt', 'bzrlib.tests')
3731
test_prefix_alias_registry.register('bb', 'bzrlib.tests.blackbox')
3732
test_prefix_alias_registry.register('bp', 'bzrlib.plugins')
3733
3734
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3735
def _test_suite_testmod_names():
3736
    """Return the standard list of test module names to test."""
3737
    return [
3738
        'bzrlib.doc',
3739
        'bzrlib.tests.blackbox',
3740
        'bzrlib.tests.commands',
3741
        'bzrlib.tests.per_branch',
3742
        'bzrlib.tests.per_bzrdir',
4585.1.5 by Jelmer Vernooij
Some review comments from John.
3743
        'bzrlib.tests.per_foreign_vcs',
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3744
        'bzrlib.tests.per_interrepository',
3745
        'bzrlib.tests.per_intertree',
3746
        'bzrlib.tests.per_inventory',
3747
        'bzrlib.tests.per_interbranch',
3748
        'bzrlib.tests.per_lock',
3749
        'bzrlib.tests.per_transport',
3750
        'bzrlib.tests.per_tree',
3751
        'bzrlib.tests.per_pack_repository',
3752
        'bzrlib.tests.per_repository',
3753
        'bzrlib.tests.per_repository_chk',
3754
        'bzrlib.tests.per_repository_reference',
4711.1.2 by Martin Pool
Start adding tests.per_uifactory
3755
        'bzrlib.tests.per_uifactory',
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3756
        'bzrlib.tests.per_versionedfile',
3757
        'bzrlib.tests.per_workingtree',
3758
        'bzrlib.tests.test__annotator',
3759
        'bzrlib.tests.test__chk_map',
3760
        'bzrlib.tests.test__dirstate_helpers',
3761
        'bzrlib.tests.test__groupcompress',
3762
        'bzrlib.tests.test__known_graph',
3763
        'bzrlib.tests.test__rio',
4679.3.76 by John Arbash Meinel
Rename StaticTupleInterner => SimpleSet.
3764
        'bzrlib.tests.test__simple_set',
4679.5.2 by John Arbash Meinel
Merge 2.1-export-c-api branch, and bring back the static_tuple code.
3765
        'bzrlib.tests.test__static_tuple',
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3766
        'bzrlib.tests.test__walkdirs_win32',
3767
        'bzrlib.tests.test_ancestry',
3768
        'bzrlib.tests.test_annotate',
3769
        'bzrlib.tests.test_api',
3770
        'bzrlib.tests.test_atomicfile',
3771
        'bzrlib.tests.test_bad_files',
3772
        'bzrlib.tests.test_bencode',
3773
        'bzrlib.tests.test_bisect_multi',
3774
        'bzrlib.tests.test_branch',
3775
        'bzrlib.tests.test_branchbuilder',
3776
        'bzrlib.tests.test_btree_index',
3777
        'bzrlib.tests.test_bugtracker',
3778
        'bzrlib.tests.test_bundle',
3779
        'bzrlib.tests.test_bzrdir',
3780
        'bzrlib.tests.test__chunks_to_lines',
3781
        'bzrlib.tests.test_cache_utf8',
3782
        'bzrlib.tests.test_chk_map',
3783
        'bzrlib.tests.test_chk_serializer',
3784
        'bzrlib.tests.test_chunk_writer',
3785
        'bzrlib.tests.test_clean_tree',
4634.85.12 by Andrew Bennetts
Merge lp:bzr.
3786
        'bzrlib.tests.test_cleanup',
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3787
        'bzrlib.tests.test_commands',
3788
        'bzrlib.tests.test_commit',
3789
        'bzrlib.tests.test_commit_merge',
3790
        'bzrlib.tests.test_config',
3791
        'bzrlib.tests.test_conflicts',
3792
        'bzrlib.tests.test_counted_lock',
3793
        'bzrlib.tests.test_crash',
3794
        'bzrlib.tests.test_decorators',
3795
        'bzrlib.tests.test_delta',
3796
        'bzrlib.tests.test_debug',
3797
        'bzrlib.tests.test_deprecated_graph',
3798
        'bzrlib.tests.test_diff',
3799
        'bzrlib.tests.test_directory_service',
3800
        'bzrlib.tests.test_dirstate',
3801
        'bzrlib.tests.test_email_message',
3802
        'bzrlib.tests.test_eol_filters',
3803
        'bzrlib.tests.test_errors',
3804
        'bzrlib.tests.test_export',
3805
        'bzrlib.tests.test_extract',
3806
        'bzrlib.tests.test_fetch',
3807
        'bzrlib.tests.test_fifo_cache',
3808
        'bzrlib.tests.test_filters',
3809
        'bzrlib.tests.test_ftp_transport',
3810
        'bzrlib.tests.test_foreign',
3811
        'bzrlib.tests.test_generate_docs',
3812
        'bzrlib.tests.test_generate_ids',
3813
        'bzrlib.tests.test_globbing',
3814
        'bzrlib.tests.test_gpg',
3815
        'bzrlib.tests.test_graph',
3816
        'bzrlib.tests.test_groupcompress',
3817
        'bzrlib.tests.test_hashcache',
3818
        'bzrlib.tests.test_help',
3819
        'bzrlib.tests.test_hooks',
3820
        'bzrlib.tests.test_http',
3821
        'bzrlib.tests.test_http_response',
3822
        'bzrlib.tests.test_https_ca_bundle',
3823
        'bzrlib.tests.test_identitymap',
3824
        'bzrlib.tests.test_ignores',
3825
        'bzrlib.tests.test_index',
3826
        'bzrlib.tests.test_info',
3827
        'bzrlib.tests.test_inv',
3828
        'bzrlib.tests.test_inventory_delta',
3829
        'bzrlib.tests.test_knit',
3830
        'bzrlib.tests.test_lazy_import',
3831
        'bzrlib.tests.test_lazy_regex',
3832
        'bzrlib.tests.test_lock',
3833
        'bzrlib.tests.test_lockable_files',
3834
        'bzrlib.tests.test_lockdir',
3835
        'bzrlib.tests.test_log',
3836
        'bzrlib.tests.test_lru_cache',
3837
        'bzrlib.tests.test_lsprof',
3838
        'bzrlib.tests.test_mail_client',
3839
        'bzrlib.tests.test_memorytree',
3840
        'bzrlib.tests.test_merge',
3841
        'bzrlib.tests.test_merge3',
3842
        'bzrlib.tests.test_merge_core',
3843
        'bzrlib.tests.test_merge_directive',
3844
        'bzrlib.tests.test_missing',
3845
        'bzrlib.tests.test_msgeditor',
3846
        'bzrlib.tests.test_multiparent',
3847
        'bzrlib.tests.test_mutabletree',
3848
        'bzrlib.tests.test_nonascii',
3849
        'bzrlib.tests.test_options',
3850
        'bzrlib.tests.test_osutils',
3851
        'bzrlib.tests.test_osutils_encodings',
3852
        'bzrlib.tests.test_pack',
3853
        'bzrlib.tests.test_patch',
3854
        'bzrlib.tests.test_patches',
3855
        'bzrlib.tests.test_permissions',
3856
        'bzrlib.tests.test_plugins',
3857
        'bzrlib.tests.test_progress',
3858
        'bzrlib.tests.test_read_bundle',
3859
        'bzrlib.tests.test_reconcile',
3860
        'bzrlib.tests.test_reconfigure',
3861
        'bzrlib.tests.test_registry',
3862
        'bzrlib.tests.test_remote',
3863
        'bzrlib.tests.test_rename_map',
3864
        'bzrlib.tests.test_repository',
3865
        'bzrlib.tests.test_revert',
3866
        'bzrlib.tests.test_revision',
3867
        'bzrlib.tests.test_revisionspec',
3868
        'bzrlib.tests.test_revisiontree',
3869
        'bzrlib.tests.test_rio',
3870
        'bzrlib.tests.test_rules',
3871
        'bzrlib.tests.test_sampler',
4665.5.1 by Vincent Ladeuil
Start some shell-like capability to write tests.
3872
        'bzrlib.tests.test_script',
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3873
        'bzrlib.tests.test_selftest',
3874
        'bzrlib.tests.test_serializer',
3875
        'bzrlib.tests.test_setup',
3876
        'bzrlib.tests.test_sftp_transport',
3877
        'bzrlib.tests.test_shelf',
3878
        'bzrlib.tests.test_shelf_ui',
3879
        'bzrlib.tests.test_smart',
3880
        'bzrlib.tests.test_smart_add',
3881
        'bzrlib.tests.test_smart_request',
3882
        'bzrlib.tests.test_smart_transport',
3883
        'bzrlib.tests.test_smtp_connection',
3884
        'bzrlib.tests.test_source',
3885
        'bzrlib.tests.test_ssh_transport',
3886
        'bzrlib.tests.test_status',
3887
        'bzrlib.tests.test_store',
3888
        'bzrlib.tests.test_strace',
3889
        'bzrlib.tests.test_subsume',
3890
        'bzrlib.tests.test_switch',
3891
        'bzrlib.tests.test_symbol_versioning',
3892
        'bzrlib.tests.test_tag',
3893
        'bzrlib.tests.test_testament',
3894
        'bzrlib.tests.test_textfile',
3895
        'bzrlib.tests.test_textmerge',
3896
        'bzrlib.tests.test_timestamp',
3897
        'bzrlib.tests.test_trace',
3898
        'bzrlib.tests.test_transactions',
3899
        'bzrlib.tests.test_transform',
3900
        'bzrlib.tests.test_transport',
3901
        'bzrlib.tests.test_transport_log',
3902
        'bzrlib.tests.test_tree',
3903
        'bzrlib.tests.test_treebuilder',
3904
        'bzrlib.tests.test_tsort',
3905
        'bzrlib.tests.test_tuned_gzip',
3906
        'bzrlib.tests.test_ui',
3907
        'bzrlib.tests.test_uncommit',
3908
        'bzrlib.tests.test_upgrade',
3909
        'bzrlib.tests.test_upgrade_stacked',
3910
        'bzrlib.tests.test_urlutils',
3911
        'bzrlib.tests.test_version',
3912
        'bzrlib.tests.test_version_info',
3913
        'bzrlib.tests.test_weave',
3914
        'bzrlib.tests.test_whitebox',
3915
        'bzrlib.tests.test_win32utils',
3916
        'bzrlib.tests.test_workingtree',
3917
        'bzrlib.tests.test_workingtree_4',
3918
        'bzrlib.tests.test_wsgi',
3919
        'bzrlib.tests.test_xml',
3920
        ]
3921
3922
3923
def _test_suite_modules_to_doctest():
3924
    """Return the list of modules to doctest."""   
3925
    return [
3926
        'bzrlib',
3927
        'bzrlib.branchbuilder',
3928
        'bzrlib.export',
3929
        'bzrlib.inventory',
3930
        'bzrlib.iterablefile',
3931
        'bzrlib.lockdir',
3932
        'bzrlib.merge3',
3933
        'bzrlib.option',
3934
        'bzrlib.symbol_versioning',
3935
        'bzrlib.tests',
3936
        'bzrlib.timestamp',
3937
        'bzrlib.version_info_formats.format_custom',
3938
        ]
3939
3940
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3941
def test_suite(keep_only=None, starting_with=None):
1707.2.1 by Robert Collins
'bzr selftest --benchmark' will run a new benchmarking selftest.
3942
    """Build and return TestSuite for the whole of bzrlib.
3193.1.6 by Vincent Ladeuil
Filter the whole test suite.
3943
3944
    :param keep_only: A list of test ids limiting the suite returned.
3945
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3946
    :param starting_with: An id limiting the suite returned to the tests
3947
         starting with it.
3948
1707.2.1 by Robert Collins
'bzr selftest --benchmark' will run a new benchmarking selftest.
3949
    This function can be replaced if you need to change the default test
3950
    suite on a global basis, but it is not encouraged.
3951
    """
3302.9.23 by Vincent Ladeuil
Simplify test_suite().
3952
1707.2.2 by Robert Collins
Start on bench_add, an add benchtest.
3953
    loader = TestUtil.TestLoader()
3193.1.6 by Vincent Ladeuil
Filter the whole test suite.
3954
4498.1.2 by Vincent Ladeuil
Fix selftest -s xxx --load yyy usage.
3955
    if keep_only is not None:
3956
        id_filter = TestIdList(keep_only)
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
3957
    if starting_with:
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3958
        # We take precedence over keep_only because *at loading time* using
3959
        # both options means we will load less tests for the same final result.
3302.11.5 by Vincent Ladeuil
Fixed as per John's review. Also added a NEWS entry.
3960
        def interesting_module(name):
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
3961
            for start in starting_with:
3962
                if (
3963
                    # Either the module name starts with the specified string
3964
                    name.startswith(start)
3965
                    # or it may contain tests starting with the specified string
3966
                    or start.startswith(name)
3967
                    ):
3968
                    return True
3969
            return False
3302.11.6 by Vincent Ladeuil
Fixed as per Martin and John reviews. Also fix a bug.
3970
        loader = TestUtil.FilteredByModuleTestLoader(interesting_module)
3302.11.5 by Vincent Ladeuil
Fixed as per John's review. Also added a NEWS entry.
3971
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3972
    elif keep_only is not None:
3302.8.12 by Vincent Ladeuil
Simplify tests.test_suite.
3973
        loader = TestUtil.FilteredByModuleTestLoader(id_filter.refers_to)
3302.11.5 by Vincent Ladeuil
Fixed as per John's review. Also added a NEWS entry.
3974
        def interesting_module(name):
3975
            return id_filter.refers_to(name)
3976
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3977
    else:
3978
        loader = TestUtil.TestLoader()
3302.11.5 by Vincent Ladeuil
Fixed as per John's review. Also added a NEWS entry.
3979
        def interesting_module(name):
3980
            # No filtering, all modules are interesting
3981
            return True
3982
3302.8.12 by Vincent Ladeuil
Simplify tests.test_suite.
3983
    suite = loader.suiteClass()
3193.1.6 by Vincent Ladeuil
Filter the whole test suite.
3984
3985
    # modules building their suite with loadTestsFromModuleNames
4650.1.1 by Robert Collins
Refactor test_suite to make stubbing out the list of tests to load possible without sacrificing coverage.
3986
    suite.addTest(loader.loadTestsFromModuleNames(_test_suite_testmod_names()))
3987
3988
    for mod in _test_suite_modules_to_doctest():
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
3989
        if not interesting_module(mod):
3302.8.13 by Vincent Ladeuil
DocTests can now be filtered at module level too.
3990
            # No tests to keep here, move along
3991
            continue
1996.3.5 by John Arbash Meinel
Cleanup, deprecated, and get the tests passing again.
3992
        try:
3943.8.1 by Marius Kruger
remove all trailing whitespace from bzr source
3993
            # note that this really does mean "report only" -- doctest
3825.3.4 by Martin Pool
Doc
3994
            # still runs the rest of the examples
3825.3.3 by Martin Pool
Doctests now stop on the first failure
3995
            doc_suite = doctest.DocTestSuite(mod,
3825.3.5 by Martin Pool
Fix tab damage
3996
                optionflags=doctest.REPORT_ONLY_FIRST_FAILURE)
1996.3.5 by John Arbash Meinel
Cleanup, deprecated, and get the tests passing again.
3997
        except ValueError, e:
3193.1.6 by Vincent Ladeuil
Filter the whole test suite.
3998
            print '**failed to get doctest for: %s\n%s' % (mod, e)
1996.3.5 by John Arbash Meinel
Cleanup, deprecated, and get the tests passing again.
3999
            raise
3825.3.1 by Martin Pool
selftest errors if modules said to have doctests don't have them
4000
        if len(doc_suite._tests) == 0:
4001
            raise errors.BzrError("no doctests found in %s" % (mod,))
3193.1.6 by Vincent Ladeuil
Filter the whole test suite.
4002
        suite.addTest(doc_suite)
4003
2762.2.1 by Robert Collins
* ``bzr plugins`` now lists the version number for each plugin in square
4004
    default_encoding = sys.getdefaultencoding()
3221.4.1 by Martin Pool
Treat failure to load plugin test suites as a fatal error
4005
    for name, plugin in bzrlib.plugin.plugins().items():
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
4006
        if not interesting_module(plugin.module.__name__):
4007
            continue
3221.4.1 by Martin Pool
Treat failure to load plugin test suites as a fatal error
4008
        plugin_suite = plugin.test_suite()
4009
        # We used to catch ImportError here and turn it into just a warning,
4010
        # but really if you don't have --no-plugins this should be a failure.
4011
        # mbp 20080213 - see http://bugs.launchpad.net/bugs/189771
3302.8.12 by Vincent Ladeuil
Simplify tests.test_suite.
4012
        if plugin_suite is None:
3302.8.21 by Vincent Ladeuil
Fixed as per Robert's review.
4013
            plugin_suite = plugin.load_plugin_tests(loader)
3221.4.1 by Martin Pool
Treat failure to load plugin test suites as a fatal error
4014
        if plugin_suite is not None:
4015
            suite.addTest(plugin_suite)
2762.2.1 by Robert Collins
* ``bzr plugins`` now lists the version number for each plugin in square
4016
        if default_encoding != sys.getdefaultencoding():
4017
            bzrlib.trace.warning(
4018
                'Plugin "%s" tried to reset default encoding to: %s', name,
4019
                sys.getdefaultencoding())
4020
            reload(sys)
4021
            sys.setdefaultencoding(default_encoding)
3302.3.1 by Vincent Ladeuil
Help identify duplicates IDs in test suite and missing tests in id
4022
4023
    if keep_only is not None:
3302.8.12 by Vincent Ladeuil
Simplify tests.test_suite.
4024
        # Now that the referred modules have loaded their tests, keep only the
4025
        # requested ones.
4026
        suite = filter_suite_by_id_list(suite, id_filter)
3302.3.1 by Vincent Ladeuil
Help identify duplicates IDs in test suite and missing tests in id
4027
        # Do some sanity checks on the id_list filtering
4028
        not_found, duplicates = suite_matches_id_list(suite, keep_only)
3649.6.4 by Vincent Ladeuil
selftest --starting-with now accepts multiple values.
4029
        if starting_with:
3302.11.5 by Vincent Ladeuil
Fixed as per John's review. Also added a NEWS entry.
4030
            # The tester has used both keep_only and starting_with, so he is
4031
            # already aware that some tests are excluded from the list, there
4032
            # is no need to tell him which.
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
4033
            pass
4034
        else:
3302.11.5 by Vincent Ladeuil
Fixed as per John's review. Also added a NEWS entry.
4035
            # Some tests mentioned in the list are not in the test suite. The
4036
            # list may be out of date, report to the tester.
3302.11.2 by Vincent Ladeuil
selftest now accepts --starting-ith <id> to load and execute only a module|class|test* reduced suite.
4037
            for id in not_found:
4038
                bzrlib.trace.warning('"%s" not found in the test suite', id)
3302.3.1 by Vincent Ladeuil
Help identify duplicates IDs in test suite and missing tests in id
4039
        for id in duplicates:
4040
            bzrlib.trace.warning('"%s" is used as an id by several tests', id)
4041
1092.1.17 by Robert Collins
remove TEST_CLASSES dead code and provide a bzrlib.test_suite() convenience method
4042
    return suite
764 by Martin Pool
- log messages from a particular test are printed if that test fails
4043
1185.51.1 by Martin Pool
Better message when failing to import a test suite.
4044
2745.6.58 by Andrew Bennetts
Slightly neater test parameterisation in repository_implementations; extract a 'multiply_scenarios' function.
4045
def multiply_scenarios(scenarios_left, scenarios_right):
4046
    """Multiply two sets of scenarios.
4047
4048
    :returns: the cartesian product of the two sets of scenarios, that is
4049
        a scenario for every possible combination of a left scenario and a
4050
        right scenario.
4051
    """
4052
    return [
4053
        ('%s,%s' % (left_name, right_name),
4054
         dict(left_dict.items() + right_dict.items()))
4055
        for left_name, left_dict in scenarios_left
4056
        for right_name, right_dict in scenarios_right]
4057
4058
4084.5.1 by Robert Collins
Bulk update all test adaptation into a single approach, using multiply_tests rather than test adapters.
4059
def multiply_tests(tests, scenarios, result):
4060
    """Multiply tests_list by scenarios into result.
4061
4062
    This is the core workhorse for test parameterisation.
4063
4064
    Typically the load_tests() method for a per-implementation test suite will
4065
    call multiply_tests and return the result.
4066
4067
    :param tests: The tests to parameterise.
4068
    :param scenarios: The scenarios to apply: pairs of (scenario_name,
4069
        scenario_param_dict).
4070
    :param result: A TestSuite to add created tests to.
4071
4072
    This returns the passed in result TestSuite with the cross product of all
4073
    the tests repeated once for each scenario.  Each test is adapted by adding
4074
    the scenario name at the end of its id(), and updating the test object's
4075
    __dict__ with the scenario_param_dict.
4076
4205.4.1 by Robert Collins
Import needed module for a doctest.
4077
    >>> import bzrlib.tests.test_sampler
4084.5.1 by Robert Collins
Bulk update all test adaptation into a single approach, using multiply_tests rather than test adapters.
4078
    >>> r = multiply_tests(
4079
    ...     bzrlib.tests.test_sampler.DemoTest('test_nothing'),
4080
    ...     [('one', dict(param=1)),
4081
    ...      ('two', dict(param=2))],
4082
    ...     TestSuite())
4083
    >>> tests = list(iter_suite_tests(r))
4084
    >>> len(tests)
4085
    2
4086
    >>> tests[0].id()
4087
    'bzrlib.tests.test_sampler.DemoTest.test_nothing(one)'
4088
    >>> tests[0].param
4089
    1
4090
    >>> tests[1].param
4091
    2
4092
    """
4093
    for test in iter_suite_tests(tests):
4094
        apply_scenarios(test, scenarios, result)
4095
    return result
4096
4097
4098
def apply_scenarios(test, scenarios, result):
4099
    """Apply the scenarios in scenarios to test and add to result.
4100
4101
    :param test: The test to apply scenarios to.
4102
    :param scenarios: An iterable of scenarios to apply to test.
4103
    :return: result
4104
    :seealso: apply_scenario
4105
    """
4106
    for scenario in scenarios:
4107
        result.addTest(apply_scenario(test, scenario))
4108
    return result
4109
4110
4111
def apply_scenario(test, scenario):
4112
    """Copy test and apply scenario to it.
4113
4114
    :param test: A test to adapt.
4115
    :param scenario: A tuple describing the scenarion.
4116
        The first element of the tuple is the new test id.
4117
        The second element is a dict containing attributes to set on the
4118
        test.
4119
    :return: The adapted test.
4120
    """
4121
    new_id = "%s(%s)" % (test.id(), scenario[0])
4122
    new_test = clone_test(test, new_id)
4123
    for name, value in scenario[1].items():
4124
        setattr(new_test, name, value)
4125
    return new_test
4126
4127
4128
def clone_test(test, new_id):
4129
    """Clone a test giving it a new id.
4130
4131
    :param test: The test to clone.
4132
    :param new_id: The id to assign to it.
4133
    :return: The new test.
4134
    """
4650.1.5 by Robert Collins
Use copy, not deepcopy, to clone tests in test parameterisation.
4135
    new_test = copy(test)
4084.5.1 by Robert Collins
Bulk update all test adaptation into a single approach, using multiply_tests rather than test adapters.
4136
    new_test.id = lambda: new_id
4137
    return new_test
3004.1.5 by Daniel Watkins
Added adapt_tests which will adapt tests at a finer-than-module level.
4138
4139
4807.3.3 by John Arbash Meinel
Report the test-id when we fail to delete a testing dir.
4140
def _rmtree_temp_dir(dirname, test_id=None):
2485.6.4 by Martin Pool
Move unicode handling code into _rmtree_temp_dir
4141
    # If LANG=C we probably have created some bogus paths
4142
    # which rmtree(unicode) will fail to delete
4143
    # so make sure we are using rmtree(str) to delete everything
4144
    # except on win32, where rmtree(str) will fail
4145
    # since it doesn't have the property of byte-stream paths
4146
    # (they are either ascii or mbcs)
4147
    if sys.platform == 'win32':
4148
        # make sure we are using the unicode win32 api
4149
        dirname = unicode(dirname)
4150
    else:
4151
        dirname = dirname.encode(sys.getfilesystemencoding())
2379.6.4 by Alexander Belchenko
Teach `bzr selftest --clean-output` to remove read-only files (win32-specific)
4152
    try:
4153
        osutils.rmtree(dirname)
4154
    except OSError, e:
4615.3.2 by Vincent Ladeuil
Allows selftest to finish even without fully cleaning the test dir.
4155
        # We don't want to fail here because some useful display will be lost
4156
        # otherwise. Polluting the tmp dir is bad, but not giving all the
4157
        # possible info to the test runner is even worse.
4807.3.3 by John Arbash Meinel
Report the test-id when we fail to delete a testing dir.
4158
        if test_id != None:
4159
            ui.ui_factory.clear_term()
4857.1.1 by John Arbash Meinel
Add an extra newline. It seems that calling clear_term() just isn't enough.
4160
            sys.stderr.write('\nWhile running: %s\n' % (test_id,))
4615.3.2 by Vincent Ladeuil
Allows selftest to finish even without fully cleaning the test dir.
4161
        sys.stderr.write('Unable to remove testing dir %s\n%s'
4162
                         % (os.path.basename(dirname), e))
2379.6.4 by Alexander Belchenko
Teach `bzr selftest --clean-output` to remove read-only files (win32-specific)
4163
4164
2367.1.4 by Robert Collins
Add operating system Feature model to bzrlib.tests to allow writing tests
4165
class Feature(object):
4166
    """An operating system Feature."""
4167
4168
    def __init__(self):
4169
        self._available = None
4170
4171
    def available(self):
4172
        """Is the feature available?
4173
4174
        :return: True if the feature is available.
4175
        """
4176
        if self._available is None:
4177
            self._available = self._probe()
4178
        return self._available
4179
4180
    def _probe(self):
4181
        """Implement this method in concrete features.
4182
4183
        :return: True if the feature is available.
4184
        """
4185
        raise NotImplementedError
4186
4187
    def __str__(self):
4188
        if getattr(self, 'feature_name', None):
4189
            return self.feature_name()
4190
        return self.__class__.__name__
2553.2.3 by Robert Collins
Split out the common test scenario support from the repository implementation specific code.
4191
4192
2776.1.5 by Robert Collins
Add reasonably comprehensive tests for path last modified and per file graph behaviour.
4193
class _SymlinkFeature(Feature):
4194
4195
    def _probe(self):
4196
        return osutils.has_symlinks()
4197
4198
    def feature_name(self):
4199
        return 'symlinks'
4200
4201
SymlinkFeature = _SymlinkFeature()
4202
4203
3136.1.1 by Aaron Bentley
Add support for hardlinks to TreeTransform
4204
class _HardlinkFeature(Feature):
4205
4206
    def _probe(self):
4207
        return osutils.has_hardlinks()
4208
4209
    def feature_name(self):
4210
        return 'hardlinks'
4211
4212
HardlinkFeature = _HardlinkFeature()
4213
4214
2949.5.2 by Alexander Belchenko
John's review
4215
class _OsFifoFeature(Feature):
4216
4217
    def _probe(self):
4218
        return getattr(os, 'mkfifo', None)
4219
4220
    def feature_name(self):
4221
        return 'filesystem fifos'
4222
4223
OsFifoFeature = _OsFifoFeature()
4224
4225
3477.1.2 by John Arbash Meinel
Rename UnicodeFilename => UnicodeFilenameFeature
4226
class _UnicodeFilenameFeature(Feature):
3477.1.1 by John Arbash Meinel
Move UnicodeFeature into a core 'tests' feature, rather than living in test_diff.
4227
    """Does the filesystem support Unicode filenames?"""
4228
4229
    def _probe(self):
4230
        try:
3526.2.1 by Martin von Gagern
Improved UnicodeFilenameFeature.
4231
            # Check for character combinations unlikely to be covered by any
4232
            # single non-unicode encoding. We use the characters
4233
            # - greek small letter alpha (U+03B1) and
4234
            # - braille pattern dots-123456 (U+283F).
4235
            os.stat(u'\u03b1\u283f')
3477.1.1 by John Arbash Meinel
Move UnicodeFeature into a core 'tests' feature, rather than living in test_diff.
4236
        except UnicodeEncodeError:
4237
            return False
4238
        except (IOError, OSError):
4239
            # The filesystem allows the Unicode filename but the file doesn't
4240
            # exist.
4241
            return True
4242
        else:
4243
            # The filesystem allows the Unicode filename and the file exists,
4244
            # for some reason.
4245
            return True
4246
3477.1.2 by John Arbash Meinel
Rename UnicodeFilename => UnicodeFilenameFeature
4247
UnicodeFilenameFeature = _UnicodeFilenameFeature()
3477.1.1 by John Arbash Meinel
Move UnicodeFeature into a core 'tests' feature, rather than living in test_diff.
4248
4249
4873.2.3 by John Arbash Meinel
Change from _ModuleFeature to ModuleAvailableFeature, per vila's review.
4250
class ModuleAvailableFeature(Feature):
4873.2.1 by John Arbash Meinel
Add a helper _ModuleFeature.
4251
    """This is a feature than describes a module we want to be available.
4252
4253
    Declare the name of the module in __init__(), and then after probing, the
4254
    module will be available as 'self.module'.
4255
4256
    :ivar module: The module if it is available, else None.
4257
    """
4258
4259
    def __init__(self, module_name):
4873.2.3 by John Arbash Meinel
Change from _ModuleFeature to ModuleAvailableFeature, per vila's review.
4260
        super(ModuleAvailableFeature, self).__init__()
4873.2.1 by John Arbash Meinel
Add a helper _ModuleFeature.
4261
        self.module_name = module_name
4262
4263
    def _probe(self):
4264
        try:
4265
            self._module = __import__(self.module_name, {}, {}, [''])
4266
            return True
4267
        except ImportError:
4268
            return False
4269
4270
    @property
4271
    def module(self):
4272
        if self.available(): # Make sure the probe has been done
4273
            return self._module
4274
        return None
4275
    
4276
    def feature_name(self):
4277
        return self.module_name
4278
4279
4280
2785.1.5 by Alexander Belchenko
support for non-ascii BZR_HOME in show_version()
4281
def probe_unicode_in_user_encoding():
4282
    """Try to encode several unicode strings to use in unicode-aware tests.
4283
    Return first successfull match.
4284
4285
    :return:  (unicode value, encoded plain string value) or (None, None)
4286
    """
4287
    possible_vals = [u'm\xb5', u'\xe1', u'\u0410']
4288
    for uni_val in possible_vals:
4289
        try:
3224.5.4 by Andrew Bennetts
Fix test suite, mainly weeding out uses of bzrlib.user_encoding.
4290
            str_val = uni_val.encode(osutils.get_user_encoding())
2785.1.5 by Alexander Belchenko
support for non-ascii BZR_HOME in show_version()
4291
        except UnicodeEncodeError:
4292
            # Try a different character
4293
            pass
4294
        else:
4295
            return uni_val, str_val
4296
    return None, None
2804.4.1 by Alexander Belchenko
some win32-specific fixes for selftest
4297
4298
2839.6.2 by Alexander Belchenko
changes after Martin's review
4299
def probe_bad_non_ascii(encoding):
2804.4.1 by Alexander Belchenko
some win32-specific fixes for selftest
4300
    """Try to find [bad] character with code [128..255]
2839.6.2 by Alexander Belchenko
changes after Martin's review
4301
    that cannot be decoded to unicode in some encoding.
2804.4.1 by Alexander Belchenko
some win32-specific fixes for selftest
4302
    Return None if all non-ascii characters is valid
2839.6.2 by Alexander Belchenko
changes after Martin's review
4303
    for given encoding.
2804.4.1 by Alexander Belchenko
some win32-specific fixes for selftest
4304
    """
4305
    for i in xrange(128, 256):
4306
        char = chr(i)
4307
        try:
2839.6.2 by Alexander Belchenko
changes after Martin's review
4308
            char.decode(encoding)
2804.4.1 by Alexander Belchenko
some win32-specific fixes for selftest
4309
        except UnicodeDecodeError:
4310
            return char
4311
    return None
2917.3.1 by Vincent Ladeuil
Separate transport from test server.
4312
4313
2929.3.10 by Vincent Ladeuil
Add a fake https server and test facilities.
4314
class _HTTPSServerFeature(Feature):
4315
    """Some tests want an https Server, check if one is available.
4316
2929.3.12 by Vincent Ladeuil
Implement an https server passing the same tests than http. Except
4317
    Right now, the only way this is available is under python2.6 which provides
4318
    an ssl module.
2929.3.10 by Vincent Ladeuil
Add a fake https server and test facilities.
4319
    """
4320
4321
    def _probe(self):
2929.3.12 by Vincent Ladeuil
Implement an https server passing the same tests than http. Except
4322
        try:
4323
            import ssl
4324
            return True
4325
        except ImportError:
4326
            return False
2929.3.10 by Vincent Ladeuil
Add a fake https server and test facilities.
4327
4328
    def feature_name(self):
4329
        return 'HTTPSServer'
4330
4331
4332
HTTPSServerFeature = _HTTPSServerFeature()
2929.3.14 by Vincent Ladeuil
Merge bzr.dev
4333
4334
4700.1.1 by Robert Collins
Focus and move out of blackbox the acceptance test for bzr+ssh connection handshake/process spawning.
4335
class _ParamikoFeature(Feature):
4336
    """Is paramiko available?"""
4337
4338
    def _probe(self):
4339
        try:
4340
            from bzrlib.transport.sftp import SFTPAbsoluteServer
4341
            return True
4342
        except errors.ParamikoNotPresent:
4343
            return False
4344
4345
    def feature_name(self):
4346
        return "Paramiko"
4347
4348
4349
ParamikoFeature = _ParamikoFeature()
4350
4351
3287.20.2 by John Arbash Meinel
Raise a clear error about the offending filename when there is a filename with bad characters.
4352
class _UnicodeFilename(Feature):
4353
    """Does the filesystem support Unicode filenames?"""
4354
4355
    def _probe(self):
4356
        try:
4357
            os.stat(u'\u03b1')
4358
        except UnicodeEncodeError:
4359
            return False
4360
        except (IOError, OSError):
4361
            # The filesystem allows the Unicode filename but the file doesn't
4362
            # exist.
4363
            return True
4364
        else:
4365
            # The filesystem allows the Unicode filename and the file exists,
4366
            # for some reason.
4367
            return True
4368
4369
UnicodeFilename = _UnicodeFilename()
4370
4371
4372
class _UTF8Filesystem(Feature):
4373
    """Is the filesystem UTF-8?"""
4374
4375
    def _probe(self):
4376
        if osutils._fs_enc.upper() in ('UTF-8', 'UTF8'):
4377
            return True
4378
        return False
4379
4380
UTF8Filesystem = _UTF8Filesystem()
4381
4382
4715.2.1 by Robert Collins
Stop caring about a clean shutdown of the process in tests for the interactive debugger break-in facility.
4383
class _BreakinFeature(Feature):
4384
    """Does this platform support the breakin feature?"""
4385
4386
    def _probe(self):
4387
        from bzrlib import breakin
4388
        if breakin.determine_signal() is None:
4389
            return False
4390
        if sys.platform == 'win32':
4391
            # Windows doesn't have os.kill, and we catch the SIGBREAK signal.
4392
            # We trigger SIGBREAK via a Console api so we need ctypes to
4393
            # access the function
4789.5.1 by John Arbash Meinel
Simple test suite fix for a variable that didn't exist.
4394
            try:
4395
                import ctypes
4396
            except OSError:
4715.2.1 by Robert Collins
Stop caring about a clean shutdown of the process in tests for the interactive debugger break-in facility.
4397
                return False
4398
        return True
4399
4400
    def feature_name(self):
4401
        return "SIGQUIT or SIGBREAK w/ctypes on win32"
4402
4403
4404
BreakinFeature = _BreakinFeature()
4405
4406
3794.5.26 by Mark Hammond
Make the CaseInsCasePresFilenameFeature mutually exclusive from CaseInsCasePresFilenameFeature as the semantics differ in meaningful ways.
4407
class _CaseInsCasePresFilenameFeature(Feature):
4408
    """Is the file-system case insensitive, but case-preserving?"""
4409
4410
    def _probe(self):
4411
        fileno, name = tempfile.mkstemp(prefix='MixedCase')
4412
        try:
4413
            # first check truly case-preserving for created files, then check
4414
            # case insensitive when opening existing files.
4415
            name = osutils.normpath(name)
4416
            base, rel = osutils.split(name)
4417
            found_rel = osutils.canonical_relpath(base, name)
3932.3.1 by Martin Pool
merge cicp patch, correct rest syntax and news typo
4418
            return (found_rel == rel
3794.5.26 by Mark Hammond
Make the CaseInsCasePresFilenameFeature mutually exclusive from CaseInsCasePresFilenameFeature as the semantics differ in meaningful ways.
4419
                    and os.path.isfile(name.upper())
4420
                    and os.path.isfile(name.lower()))
4421
        finally:
4422
            os.close(fileno)
4423
            os.remove(name)
4424
4425
    def feature_name(self):
4426
        return "case-insensitive case-preserving filesystem"
4427
4428
CaseInsCasePresFilenameFeature = _CaseInsCasePresFilenameFeature()
4429
4430
3063.1.1 by Alexander Belchenko
Catch OSError 17 (file exists) in final phase of tree transform and show filename to user (#111758).
4431
class _CaseInsensitiveFilesystemFeature(Feature):
3794.5.26 by Mark Hammond
Make the CaseInsCasePresFilenameFeature mutually exclusive from CaseInsCasePresFilenameFeature as the semantics differ in meaningful ways.
4432
    """Check if underlying filesystem is case-insensitive but *not* case
4433
    preserving.
3063.1.1 by Alexander Belchenko
Catch OSError 17 (file exists) in final phase of tree transform and show filename to user (#111758).
4434
    """
3794.5.26 by Mark Hammond
Make the CaseInsCasePresFilenameFeature mutually exclusive from CaseInsCasePresFilenameFeature as the semantics differ in meaningful ways.
4435
    # Note that on Windows, Cygwin, MacOS etc, the file-systems are far
4436
    # more likely to be case preserving, so this case is rare.
3063.1.1 by Alexander Belchenko
Catch OSError 17 (file exists) in final phase of tree transform and show filename to user (#111758).
4437
4438
    def _probe(self):
3794.5.26 by Mark Hammond
Make the CaseInsCasePresFilenameFeature mutually exclusive from CaseInsCasePresFilenameFeature as the semantics differ in meaningful ways.
4439
        if CaseInsCasePresFilenameFeature.available():
4440
            return False
4441
3063.1.1 by Alexander Belchenko
Catch OSError 17 (file exists) in final phase of tree transform and show filename to user (#111758).
4442
        if TestCaseWithMemoryTransport.TEST_ROOT is None:
4443
            root = osutils.mkdtemp(prefix='testbzr-', suffix='.tmp')
4444
            TestCaseWithMemoryTransport.TEST_ROOT = root
4445
        else:
4446
            root = TestCaseWithMemoryTransport.TEST_ROOT
4447
        tdir = osutils.mkdtemp(prefix='case-sensitive-probe-', suffix='',
4448
            dir=root)
4449
        name_a = osutils.pathjoin(tdir, 'a')
4450
        name_A = osutils.pathjoin(tdir, 'A')
4451
        os.mkdir(name_a)
4452
        result = osutils.isdir(name_A)
4453
        _rmtree_temp_dir(tdir)
4454
        return result
4455
4456
    def feature_name(self):
4457
        return 'case-insensitive filesystem'
4458
4459
CaseInsensitiveFilesystemFeature = _CaseInsensitiveFilesystemFeature()
4165.1.1 by Robert Collins
Add builtin subunit support.
4460
4461
4462
class _SubUnitFeature(Feature):
4463
    """Check if subunit is available."""
4464
4465
    def _probe(self):
4466
        try:
4467
            import subunit
4468
            return True
4469
        except ImportError:
4470
            return False
4471
4472
    def feature_name(self):
4473
        return 'subunit'
4474
4475
SubUnitFeature = _SubUnitFeature()
4476
# Only define SubUnitBzrRunner if subunit is available.
4477
try:
4478
    from subunit import TestProtocolClient
4479
    class SubUnitBzrRunner(TextTestRunner):
4480
        def run(self, test):
4780.1.6 by Vincent Ladeuil
Make --parallel=fork compatible with --subunit.
4481
            result = BzrAutoTimingTestResultDecorator(
4557.2.1 by Robert Collins
Support timestamping subunit streams.
4482
                TestProtocolClient(self.stream))
4165.1.1 by Robert Collins
Add builtin subunit support.
4483
            test.run(result)
4484
            return result
4485
except ImportError:
4486
    pass