~bzr-pqm/bzr/bzr.dev

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
# Copyright (C) 2006, 2007, 2009 Canonical Ltd
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

"""Blackbox tests for debugger breakin"""

try:
    import ctypes
    have_ctypes = True
except ImportError:
    have_ctypes = False
import errno
import os
import signal
import subprocess
import sys
import time

from bzrlib import (
    breakin,
    errors,
    tests,
    )


class TestBreakin(tests.TestCase):
    # FIXME: If something is broken, these tests may just hang indefinitely in
    # wait() waiting for the child to exit when it's not going to.

    def setUp(self):
        super(TestBreakin, self).setUp()
        if breakin.determine_signal() is None:
            raise tests.TestSkipped('this platform is missing SIGQUIT'
                                    ' or SIGBREAK')
        if sys.platform == 'win32':
            # Windows doesn't have os.kill, and we catch the SIGBREAK signal.
            # We trigger SIGBREAK via a Console api so we need ctypes to access
            # the function
            if not have_ctypes:
                raise tests.UnavailableFeature('ctypes')
            self._send_signal = self._send_signal_win32
        else:
            self._send_signal = self._send_signal_via_kill

    def _send_signal_via_kill(self, pid, sig_type):
        if sig_type == 'break':
            sig_num = signal.SIGQUIT
        elif sig_type == 'kill':
            sig_num = signal.SIGKILL
        else:
            raise ValueError("unknown signal type: %s" % (sig_type,))
        os.kill(pid, sig_num)

    def _send_signal_win32(self, pid, sig_type):
        """Send a 'signal' on Windows.

        Windows doesn't really have signals in the same way. All it really
        supports is:
            1) Sending SIGINT to the *current* process group (so self, and all
                children of self)
            2) Sending SIGBREAK to a process that shares the current console,
                which can be in its own process group.
        So we have start_bzr_subprocess to create a new process group for the
        spawned process, and then we map
            SIGQUIT to GenerateConsoleCtrlEvent(CTRL_BREAK_EVENT)
            SIGKILL to TerminateProcess
        """
        if sig_type == 'break':
            CTRL_BREAK_EVENT = 1
            # CTRL_C_EVENT = 0
            ret = ctypes.windll.kernel32.GenerateConsoleCtrlEvent(
                    CTRL_BREAK_EVENT, pid)
            if ret == 0: #error
                err = ctypes.FormatError()
                raise RuntimeError('failed to send CTRL_BREAK: %s'
                                   % (err,))
        elif sig_type == 'kill':
            # Does the exit code matter? For now we are just setting it to
            # something other than 0
            exit_code = breakin.determine_signal()
            ctypes.windll.kernel32.TerminateProcess(pid, exit_code)

    def _popen(self, *args, **kwargs):
        if sys.platform == 'win32':
            CREATE_NEW_PROCESS_GROUP = 512
            # This allows us to send a signal to the child, *without* also
            # sending it to ourselves
            kwargs['creationflags'] = CREATE_NEW_PROCESS_GROUP
        return super(TestBreakin, self)._popen(*args, **kwargs)

    def _dont_SIGQUIT_on_darwin(self):
        if sys.platform == 'darwin':
            # At least on Leopard and with python 2.6, this test will raise a
            # popup window asking if the python failure should be reported to
            # Apple... That's not the point of the test :) Marking the test as
            # not applicable Until we find a way to disable that intrusive
            # behavior... --vila20080611
            raise tests.TestNotApplicable(
                '%s raises a popup on OSX' % self.id())

    def _wait_for_process(self, pid, sig=None):
        # We don't know quite how long waiting for the process 'pid' will take,
        # but if it's more than 10s then it's probably not going to work.
        for i in range(100):
            time.sleep(0.1)
            if sig is not None:
                self._send_signal(pid, sig)
            # Use WNOHANG to ensure we don't get blocked, doing so, we may
            # leave the process continue after *we* die...
            # Win32 doesn't support WNOHANG, so we just pass 0
            opts = getattr(os, 'WNOHANG', 0)
            try:
                # note: waitpid is different on win32, but this test only runs
                # on unix
                # TODO: waitpid doesn't work well on windows, we might consider
                #       using WaitForSingleObject(proc._handle, TIMEOUT)
                #       instead. Most notably, the WNOHANG isn't allowed, so
                #       this can hang indefinitely.
                pid_killed, returncode = os.waitpid(pid, opts)
                if (pid_killed, returncode) != (0, 0):
                    if sig is not None:
                        # high bit in low byte says if core was dumped; we
                        # don't care
                        status, sig = (returncode >> 8, returncode & 0x7f)
                        return True, sig
            except OSError, e:
                if e.errno in (errno.ECHILD, errno.ESRCH):
                    # The process doesn't exist anymore
                    return True, None
                else:
                    raise

        return False, None

    # port 0 means to allocate any port
    _test_process_args = ['serve', '--port', 'localhost:0']

    def test_breakin(self):
        # Break in to a debugger while bzr is running
        # we need to test against a command that will wait for
        # a while -- bzr serve should do
        proc = self.start_bzr_subprocess(self._test_process_args,
                env_changes=dict(BZR_SIGQUIT_PDB=None))
        # wait for it to get started, and print the 'listening' line
        proc.stderr.readline()
        # first sigquit pops into debugger
        self._send_signal(proc.pid, 'break')
        # Wait for the debugger to acknowledge the signal reception
        # Note that it is possible for this to deadlock if the child doesn't
        # acknowlege the signal and write to stderr. Perhaps we should try
        # os.read(proc.stderr.fileno())?
        err = proc.stderr.readline()
        self.assertContainsRe(err, r'entering debugger')
        # Now that the debugger is entered, we can ask him to quit
        proc.stdin.write("q\n")
        # We wait a bit to let the child process handles our query and avoid
        # triggering deadlocks leading to hangs on multi-core hosts...
        dead, sig = self._wait_for_process(proc.pid)
        if not dead:
            # The process didn't finish, let's kill it before reporting failure
            dead, sig = self._wait_for_process(proc.pid, 'kill')
            if dead:
                raise tests.KnownFailure(
                    "subprocess wasn't terminated, it had to be killed")
            else:
                self.fail("subprocess %d wasn't terminated by repeated SIGKILL",
                          proc.pid)

    def test_breakin_harder(self):
        """SIGQUITting twice ends the process."""
        self._dont_SIGQUIT_on_darwin()
        proc = self.start_bzr_subprocess(self._test_process_args,
                env_changes=dict(BZR_SIGQUIT_PDB=None))
        # wait for it to get started, and print the 'listening' line
        proc.stderr.readline()
        # break into the debugger
        self._send_signal(proc.pid, 'break')
        # Wait for the debugger to acknowledge the signal reception (since we
        # want to send a second signal, we ensure it doesn't get lost by
        # validating the first get received and produce its effect).
        err = proc.stderr.readline()
        self.assertContainsRe(err, r'entering debugger')
        dead, sig = self._wait_for_process(proc.pid, 'break')
        self.assertTrue(dead)
        # Either the child was dead before we could read its status, or the
        # child was dead from the signal we sent it.
        self.assertTrue(sig in (None, breakin.determine_signal()))

    def test_breakin_disabled(self):
        self._dont_SIGQUIT_on_darwin()
        proc = self.start_bzr_subprocess(self._test_process_args,
                env_changes=dict(BZR_SIGQUIT_PDB='0'))
        # wait for it to get started, and print the 'listening' line
        proc.stderr.readline()
        # first hit should just kill it
        self._send_signal(proc.pid, 'break')
        proc.wait()