refactor cpu number autodetection to avoid the regression

Add --max-slave-restart option

.hgignore

@@ -22,7 +22,10 @@ dist/
 include/
 lib/
 bin/
+xdist/_version.py
 pytest_xdist.egg-info
 issue/
 3rdparty/
+pytestdebug.log
 .tox
+.cache

.hgtags

@@ -14,3 +14,7 @@ cd44a941c833c098e4899fe3d42a96703754d0d5 1.5
 0d1c00018008433956aa7d93007bab6ea7de96e4 1.8
 1d27987c267577899350a25ba5828d55d87083ad 1.8
 5c5cb6d59e12e566fbb0217aea718dc31578bee1 1.9
+4406fc2a6427fadc021ed7e43e7aa5032b1ea91f 1.10
+220f6e46eb71a6212ccbe6b67b9e6edcf8ee4fa5 1.11
+39ef85dbc893cc63dede11601208098a667b58e9 1.12
+4e25f4c568be2d7cb4d1739638a9e66bbf28f588 v1.13

CHANGELOG

@@ -1,3 +1,59 @@
+1.13.1
+-------
+
+- fix a regression -n 0 now disables xdist again
+
+
+1.13
+-------------------------
+
+- extended the tox matrix with the supported py.test versions
+
+- split up the plugin into 3 plugin's
+  to prepare the departure of boxed and looponfail.
+
+  looponfail will be a part of core
+  and forked boxed will be replaced
+  with a more reliable primitive based on xdist
+
+- conforming with new pytest-2.8 behavior of returning non-zero when all
+  tests were skipped or deselected.
+
+- new "--max-slave-restart" option that can be used to control maximum
+  number of times pytest-xdist can restart slaves due to crashes. Thanks to
+  Anatoly Bubenkov for the report and Bruno Oliveira for the PR.
+
+- release as wheel
+
+- "-n" option now can be set to "auto" for automatic detection of number
+  of cpus in the host system. Thanks Suloev Dmitry for the PR.
+
+1.12
+-------------------------
+
+- fix issue594: properly report errors when the test collection
+  is random.  Thanks Bruno Oliveira.
+
+- some internal test suite adaptation (to become forward
+  compatible with the upcoming pytest-2.8)
+
+
+1.11
+-------------------------
+
+- fix pytest/xdist issue485 (also depends on py-1.4.22):
+  attach stdout/stderr on --boxed processes that die.
+
+- fix pytest/xdist issue503: make sure that a node has usually
+  two items to execute to avoid scoped fixtures to be torn down
+  pre-maturely (fixture teardown/setup is "nextitem" sensitive).
+  Thanks to Andreas Pelme for bug analysis and failing test.
+
+- restart crashed nodes by internally refactoring setup handling
+  of nodes.  Also includes better code documentation.
+  Many thanks to Floris Bruynooghe for the complete PR.
+
+
 1.10
 -------------------------
 
@@ -7,7 +63,7 @@
 - fix pytest issue382 - produce "pytest_runtest_logstart" event again
   in master. Thanks Aron Curzon.
 
-- fix pytest issue419 by sending/receiving indices into the test 
+- fix pytest issue419 by sending/receiving indices into the test
   collection instead of node ids (which are not neccessarily unique
   for functions parametrized with duplicate values)
 
@@ -120,4 +176,3 @@
 - cleaned up termination handling
 - make -x cause hard killing of test nodes to decrease wait time
   until the traceback shows up on first failure
-

README.txt

@@ -1,212 +1,221 @@
-xdist: pytest distributed testing plugin
-===============================================================
-
-The `pytest-xdist`_ plugin extends py.test with some unique
-test execution modes:
-
-* test run parallelization_: if you have multiple CPUs or hosts you can use
-  those for a combined test run.  This allows to speed up
-  development or to use special resources of `remote machines`_.
-
-* ``--boxed``: (not available on Windows) run each test in a boxed_
-  subprocess to survive ``SEGFAULTS`` or otherwise dying processes
-
-* ``--looponfail``: run your tests repeatedly in a subprocess.  After each run
-  py.test waits until a file in your project changes and then re-runs
-  the previously failing tests.  This is repeated until all tests pass
-  after which again a full run is performed.
-
-* `Multi-Platform`_ coverage: you can specify different Python interpreters
-  or different platforms and run tests in parallel on all of them.
-
-Before running tests remotely, ``py.test`` efficiently "rsyncs" your
-program source code to the remote place.  All test results
-are reported back and displayed to your local terminal.
-You may specify different Python versions and interpreters.
-
-
-Installation
------------------------
-
-Install the plugin with::
-
-    easy_install pytest-xdist
-
-    # or
-
-    pip install pytest-xdist
-
-or use the package in develope/in-place mode with
-a checkout of the `pytest-xdist repository`_ ::
-
-    python setup.py develop
-
-Usage examples
----------------------
-
-.. _parallelization:
-
-Speed up test runs by sending tests to multiple CPUs
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-To send tests to multiple CPUs, type::
-
-    py.test -n NUM
-
-Especially for longer running tests or tests requiring
-a lot of IO this can lead to considerable speed ups.
-
-
-Running tests in a Python subprocess
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-To instantiate a python2.5 sub process and send tests to it, you may type::
-
-    py.test -d --tx popen//python=python2.5
-
-This will start a subprocess which is run with the "python2.5"
-Python interpreter, found in your system binary lookup path.
-
-If you prefix the --tx option value like this::
-
-    --tx 3*popen//python=python2.5
-
-then three subprocesses would be created and tests
-will be load-balanced across these three processes.
-
-.. _boxed:
-
-Running tests in a boxed subprocess
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-If you have tests involving C or C++ libraries you might have to deal
-with tests crashing the process.  For this case you may use the boxing
-options::
-
-    py.test --boxed
-
-which will run each test in a subprocess and will report if a test
-crashed the process.  You can also combine this option with
-running multiple processes to speed up the test run and use your CPU cores::
-
-    py.test -n3 --boxed
-
-this would run 3 testing subprocesses in parallel which each
-create new boxed subprocesses for each test.
-
-
-.. _`remote machines`:
-
-Sending tests to remote SSH accounts
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-Suppose you have a package ``mypkg`` which contains some
-tests that you can successfully run locally. And you
-have a ssh-reachable machine ``myhost``.  Then
-you can ad-hoc distribute your tests by typing::
-
-    py.test -d --tx ssh=myhostpopen --rsyncdir mypkg mypkg
-
-This will synchronize your ``mypkg`` package directory
-to an remote ssh account and then locally collect tests
-and send them to remote places for execution.
-
-You can specify multiple ``--rsyncdir`` directories
-to be sent to the remote side.
-
-**NOTE:** For py.test to collect and send tests correctly
-you not only need to make sure all code and tests
-directories are rsynced, but that any test (sub) directory
-also has an ``__init__.py`` file because internally
-py.test references tests as a fully qualified python
-module path.  **You will otherwise get strange errors**
-during setup of the remote side.
-
-You can specify multiple ``--rsyncignore`` glob-patterns
-to be ignored when file are sent to the remote side.
-There are also internal ignores: .*, *.pyc, *.pyo, *~
-Those you cannot override using rsyncignore command-line or
-ini-file option(s).
-
-
-Sending tests to remote Socket Servers
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-Download the single-module `socketserver.py`_ Python program
-and run it like this::
-
-    python socketserver.py
-
-It will tell you that it starts listening on the default
-port.  You can now on your home machine specify this
-new socket host with something like this::
-
-    py.test -d --tx socket=192.168.1.102:8888 --rsyncdir mypkg mypkg
-
-
-.. _`atonce`:
-.. _`Multi-Platform`:
-
-
-Running tests on many platforms at once
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-The basic command to run tests on multiple platforms is::
-
-    py.test --dist=each --tx=spec1 --tx=spec2
-
-If you specify a windows host, an OSX host and a Linux
-environment this command will send each tests to all
-platforms - and report back failures from all platforms
-at once.   The specifications strings use the `xspec syntax`_.
-
-.. _`xspec syntax`: http://codespeak.net/execnet/trunk/basics.html#xspec
-
-.. _`socketserver.py`: http://bitbucket.org/hpk42/execnet/raw/2af991418160/execnet/script/socketserver.py
-
-.. _`execnet`: http://codespeak.net/execnet
-
-Specifying test exec environments in an ini file
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-pytest (since version 2.0) supports ini-style cofiguration.
-You can for example make running with three subprocesses
-your default like this::
-
-    [pytest]
-    addopts = -n3
-
-You can also add default environments like this::
-
-    [pytest]
-    addopts = --tx ssh=myhost//python=python2.5 --tx ssh=myhost//python=python2.6
-
-and then just type::
-
-    py.test --dist=each
-
-to run tests in each of the environments.
-
-Specifying "rsync" dirs in an ini-file
-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-In a ``tox.ini`` or ``setup.cfg`` file in your root project directory
-you may specify directories to include or to exclude in synchronisation::
-
-    [pytest]
-    rsyncdirs = . mypkg helperpkg
-    rsyncignore = .hg
-
-These directory specifications are relative to the directory
-where the configuration file was found.
-
-.. _`pytest-xdist`: http://pypi.python.org/pypi/pytest-xdist
-.. _`pytest-xdist repository`: http://bitbucket.org/hpk42/pytest-xdist
-.. _`pytest`: http://pytest.org
-
-Issue and Bug Tracker
-------------------------
-
-Please use the pytest issue tracker for bugs in this plugin, see https://bitbucket.org/hpk42/pytest/issues .
-
-
+.. image:: https://drone.io/bitbucket.org/pytest-dev/pytest-xdist/status.png
+   :target: https://drone.io/bitbucket.org/pytest-dev/pytest-xdist/latest
+.. image:: https://pypip.in/v/pytest-xdist/badge.png
+   :target: https://pypi.python.org/pypi/pytest-xdist
+
+xdist: pytest distributed testing plugin
+============================
+
+The `pytest-xdist`_ plugin extends py.test with some unique
+test execution modes:
+
+* test run parallelization_: if you have multiple CPUs or hosts you can use
+  those for a combined test run.  This allows to speed up
+  development or to use special resources of `remote machines`_.
+
+* ``--boxed``: (not available on Windows) run each test in a boxed_
+  subprocess to survive ``SEGFAULTS`` or otherwise dying processes
+
+* ``--looponfail``: run your tests repeatedly in a subprocess.  After each run
+  py.test waits until a file in your project changes and then re-runs
+  the previously failing tests.  This is repeated until all tests pass
+  after which again a full run is performed.
+
+* `Multi-Platform`_ coverage: you can specify different Python interpreters
+  or different platforms and run tests in parallel on all of them.
+
+Before running tests remotely, ``py.test`` efficiently "rsyncs" your
+program source code to the remote place.  All test results
+are reported back and displayed to your local terminal.
+You may specify different Python versions and interpreters.
+
+
+Installation
+-----------------------
+
+Install the plugin with::
+
+    easy_install pytest-xdist
+
+    # or
+
+    pip install pytest-xdist
+
+or use the package in develope/in-place mode with
+a checkout of the `pytest-xdist repository`_ ::
+
+    python setup.py develop
+
+Usage examples
+---------------------
+
+.. _parallelization:
+
+Speed up test runs by sending tests to multiple CPUs
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+To send tests to multiple CPUs, type::
+
+    py.test -n NUM
+
+Especially for longer running tests or tests requiring
+a lot of IO this can lead to considerable speed ups. This option can
+also be set to ``auto`` for automatic detection of the number of CPUs.
+
+If a test crashes the interpreter, pytest-xdist will automatically restart
+that slave and report the failure as usual. You can use the
+``--max-slave-restart`` option to limit the number of slaves that can
+be restarted, or disable restarting altogether using ``--max-slave-restart=0``.
+
+
+Running tests in a Python subprocess
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+To instantiate a python2.5 sub process and send tests to it, you may type::
+
+    py.test -d --tx popen//python=python2.5
+
+This will start a subprocess which is run with the "python2.5"
+Python interpreter, found in your system binary lookup path.
+
+If you prefix the --tx option value like this::
+
+    --tx 3*popen//python=python2.5
+
+then three subprocesses would be created and tests
+will be load-balanced across these three processes.
+
+.. _boxed:
+
+Running tests in a boxed subprocess
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+If you have tests involving C or C++ libraries you might have to deal
+with tests crashing the process.  For this case you may use the boxing
+options::
+
+    py.test --boxed
+
+which will run each test in a subprocess and will report if a test
+crashed the process.  You can also combine this option with
+running multiple processes to speed up the test run and use your CPU cores::
+
+    py.test -n3 --boxed
+
+this would run 3 testing subprocesses in parallel which each
+create new boxed subprocesses for each test.
+
+
+.. _`remote machines`:
+
+Sending tests to remote SSH accounts
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+Suppose you have a package ``mypkg`` which contains some
+tests that you can successfully run locally. And you
+have a ssh-reachable machine ``myhost``.  Then
+you can ad-hoc distribute your tests by typing::
+
+    py.test -d --tx ssh=myhostpopen --rsyncdir mypkg mypkg
+
+This will synchronize your ``mypkg`` package directory
+to an remote ssh account and then locally collect tests
+and send them to remote places for execution.
+
+You can specify multiple ``--rsyncdir`` directories
+to be sent to the remote side.
+
+**NOTE:** For py.test to collect and send tests correctly
+you not only need to make sure all code and tests
+directories are rsynced, but that any test (sub) directory
+also has an ``__init__.py`` file because internally
+py.test references tests as a fully qualified python
+module path.  **You will otherwise get strange errors**
+during setup of the remote side.
+
+You can specify multiple ``--rsyncignore`` glob-patterns
+to be ignored when file are sent to the remote side.
+There are also internal ignores: .*, *.pyc, *.pyo, *~
+Those you cannot override using rsyncignore command-line or
+ini-file option(s).
+
+
+Sending tests to remote Socket Servers
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+Download the single-module `socketserver.py`_ Python program
+and run it like this::
+
+    python socketserver.py
+
+It will tell you that it starts listening on the default
+port.  You can now on your home machine specify this
+new socket host with something like this::
+
+    py.test -d --tx socket=192.168.1.102:8888 --rsyncdir mypkg mypkg
+
+
+.. _`atonce`:
+.. _`Multi-Platform`:
+
+
+Running tests on many platforms at once
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+The basic command to run tests on multiple platforms is::
+
+    py.test --dist=each --tx=spec1 --tx=spec2
+
+If you specify a windows host, an OSX host and a Linux
+environment this command will send each tests to all
+platforms - and report back failures from all platforms
+at once.   The specifications strings use the `xspec syntax`_.
+
+.. _`xspec syntax`: http://codespeak.net/execnet/trunk/basics.html#xspec
+
+.. _`socketserver.py`: http://bitbucket.org/hpk42/execnet/raw/2af991418160/execnet/script/socketserver.py
+
+.. _`execnet`: http://codespeak.net/execnet
+
+Specifying test exec environments in an ini file
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+pytest (since version 2.0) supports ini-style cofiguration.
+You can for example make running with three subprocesses
+your default like this::
+
+    [pytest]
+    addopts = -n3
+
+You can also add default environments like this::
+
+    [pytest]
+    addopts = --tx ssh=myhost//python=python2.5 --tx ssh=myhost//python=python2.6
+
+and then just type::
+
+    py.test --dist=each
+
+to run tests in each of the environments.
+
+Specifying "rsync" dirs in an ini-file
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+
+In a ``tox.ini`` or ``setup.cfg`` file in your root project directory
+you may specify directories to include or to exclude in synchronisation::
+
+    [pytest]
+    rsyncdirs = . mypkg helperpkg
+    rsyncignore = .hg
+
+These directory specifications are relative to the directory
+where the configuration file was found.
+
+.. _`pytest-xdist`: http://pypi.python.org/pypi/pytest-xdist
+.. _`pytest-xdist repository`: http://bitbucket.org/pytest-dev/pytest-xdist
+.. _`pytest`: http://pytest.org
+
+Issue and Bug Tracker
+------------------------
+
+Please use the pytest issue tracker for bugs in this plugin, see https://bitbucket.org/hpk42/pytest/issues .

setup.cfg

@@ -0,0 +1,2 @@
+[bdist_wheel]
+universal = 1

setup.py

@@ -2,7 +2,7 @@ from setuptools import setup
 
 setup(
     name="pytest-xdist",
-    version='1.10',
+    use_scm_version={'write_to': 'xdist/_version.py'},
     description='py.test xdist plugin for distributed testing and loop-on-failing modes',
     long_description=open('README.txt').read(),
     license='MIT',
@@ -11,20 +11,27 @@ setup(
     url='http://bitbucket.org/hpk42/pytest-xdist',
     platforms=['linux', 'osx', 'win32'],
     packages = ['xdist'],
-    entry_points = {'pytest11': ['xdist = xdist.plugin'],},
+    entry_points = {
+        'pytest11': [
+            'xdist = xdist.plugin',
+            'xdist.looponfail = xdist.looponfail',
+            'xdist.boxed = xdist.boxed',
+        ],
+    },
     zip_safe=False,
-    install_requires = ['execnet>=1.1', 'pytest>=2.4.2'],
+    install_requires=['execnet>=1.1', 'pytest>=2.4.2', 'py>=1.4.22'],
+    setup_requires=['setuptools_scm'],
     classifiers=[
-    'Development Status :: 5 - Production/Stable',
-    'Intended Audience :: Developers',
-    'License :: OSI Approved :: MIT License',
-    'Operating System :: POSIX',
-    'Operating System :: Microsoft :: Windows',
-    'Operating System :: MacOS :: MacOS X',
-    'Topic :: Software Development :: Testing',
-    'Topic :: Software Development :: Quality Assurance',
-    'Topic :: Utilities',
-    'Programming Language :: Python',
-    'Programming Language :: Python :: 3',
+        'Development Status :: 5 - Production/Stable',
+        'Intended Audience :: Developers',
+        'License :: OSI Approved :: MIT License',
+        'Operating System :: POSIX',
+        'Operating System :: Microsoft :: Windows',
+        'Operating System :: MacOS :: MacOS X',
+        'Topic :: Software Development :: Testing',
+        'Topic :: Software Development :: Quality Assurance',
+        'Topic :: Utilities',
+        'Programming Language :: Python',
+        'Programming Language :: Python :: 3',
     ],
 )

testing/acceptance_test.py

@@ -1,5 +1,6 @@
 import py
-import sys
+import pytest
+
 
 class TestDistribution:
     def test_n1_pass(self, testdir):
@@ -112,7 +113,7 @@ class TestDistribution:
                 import py
                 assert tmpdir.relto(py.path.local(%r)), tmpdir
         """ % str(testdir.tmpdir))
-        result = testdir.runpytest(p1, "-n1")
+        result = testdir.runpytest_subprocess(p1, "-n1")
         assert result.ret == 0
         result.stdout.fnmatch_lines([
             "*1 passed*",
@@ -193,7 +194,7 @@ class TestDistribution:
         assert dest.join(subdir.basename).check(dir=1)
 
     def test_data_exchange(self, testdir):
-        c1 = testdir.makeconftest("""
+        testdir.makeconftest("""
             # This hook only called on master.
             def pytest_configure_node(node):
                 node.slaveinput['a'] = 42
@@ -242,7 +243,7 @@ class TestDistribution:
                 print ("s2call-finished")
         """)
         args = ["-n1", "--debug"]
-        result = testdir.runpytest(*args)
+        result = testdir.runpytest_subprocess(*args)
         s = result.stdout.str()
         assert result.ret == 2
         assert 's2call' in s
@@ -250,14 +251,13 @@ class TestDistribution:
 
     def test_keyboard_interrupt_dist(self, testdir):
         # xxx could be refined to check for return code
-        p = testdir.makepyfile("""
+        testdir.makepyfile("""
             def test_sleep():
                 import time
                 time.sleep(10)
         """)
-        child = testdir.spawn_pytest("-n1")
-        py.std.time.sleep(0.1)
-        child.expect(".*test session starts.*")
+        child = testdir.spawn_pytest("-n1 -v")
+        child.expect(".*test_sleep.*")
         child.kill(2) # keyboard interrupt
         child.expect(".*KeyboardInterrupt.*")
         #child.expect(".*seconds.*")
@@ -270,7 +270,7 @@ class TestDistEach:
             def test_hello():
                 pass
         """)
-        result = testdir.runpytest("--debug", "--dist=each", "--tx=2*popen")
+        result = testdir.runpytest_subprocess("--debug", "--dist=each", "--tx=2*popen")
         assert not result.ret
         result.stdout.fnmatch_lines(["*2 pass*"])
 
@@ -300,7 +300,7 @@ class TestDistEach:
 
 class TestTerminalReporting:
     def test_pass_skip_fail(self, testdir):
-        p = testdir.makepyfile("""
+        testdir.makepyfile("""
             import py
             def test_ok():
                 pass
@@ -311,9 +311,9 @@ class TestTerminalReporting:
         """)
         result = testdir.runpytest("-n1", "-v")
         result.stdout.fnmatch_lines_random([
-            "*PASS*test_pass_skip_fail.py:2: *test_ok*",
-            "*SKIP*test_pass_skip_fail.py:4: *test_skip*",
-            "*FAIL*test_pass_skip_fail.py:6: *test_func*",
+            "*PASS*test_pass_skip_fail.py*test_ok*",
+            "*SKIP*test_pass_skip_fail.py*test_skip*",
+            "*FAIL*test_pass_skip_fail.py*test_func*",
         ])
         result.stdout.fnmatch_lines([
             "*def test_func():",
@@ -322,13 +322,13 @@ class TestTerminalReporting:
         ])
 
     def test_fail_platinfo(self, testdir):
-        p = testdir.makepyfile("""
+        testdir.makepyfile("""
             def test_func():
                 assert 0
         """)
         result = testdir.runpytest("-n1", "-v")
         result.stdout.fnmatch_lines([
-            "*FAIL*test_fail_platinfo.py:1: *test_func*",
+            "*FAIL*test_fail_platinfo.py*test_func*",
             "*0*Python*",
             "*def test_func():",
             ">       assert 0",
@@ -351,7 +351,7 @@ def test_teardownfails_one_function(testdir):
 @py.test.mark.xfail
 def test_terminate_on_hangingnode(testdir):
     p = testdir.makeconftest("""
-        def pytest_sessionfinishes(session):
+        def pytest_sessionfinish(session):
             if session.nodeid == "my": # running on slave
                 import time
                 time.sleep(3)
@@ -363,7 +363,20 @@ def test_terminate_on_hangingnode(testdir):
     ])
 
 
+def test_auto_detect_cpus(testdir, monkeypatch):
+    import multiprocessing
+    monkeypatch.setattr(multiprocessing, 'cpu_count', lambda: 3)
+    testdir.makeconftest("""
+        def pytest_unconfigure(config):
+            with open('cpus', 'w') as f:
+                f.write('cpus = %s' % config.option.numprocesses)
+    """)
+    testdir.inline_run('-n=auto')
+    cpus_file = testdir.tmpdir.join('cpus')
+    assert cpus_file.read() == 'cpus = 3'
 
+
+@pytest.mark.xfail(reason="works if run outside test suite", run=False)
 def test_session_hooks(testdir):
     testdir.makeconftest("""
         import sys
@@ -397,6 +410,31 @@ def test_session_hooks(testdir):
     assert testdir.tmpdir.join("slave").check()
     assert testdir.tmpdir.join("master").check()
 
+
+def test_session_testscollected(testdir):
+    """
+    Make sure master node is updating the session object with the number
+    of tests collected from the slaves.
+    """
+    testdir.makepyfile(test_foo="""
+        import pytest
+        @pytest.mark.parametrize('i', range(3))
+        def test_ok(i):
+            pass
+    """)
+    testdir.makeconftest("""
+        def pytest_sessionfinish(session):
+            collected = getattr(session, 'testscollected', None)
+            with open('testscollected', 'w') as f:
+                f.write('collected = %s' % collected)
+    """)
+    result = testdir.inline_run("-n1")
+    result.assertoutcome(passed=3)
+    collected_file = testdir.tmpdir.join('testscollected')
+    assert collected_file.isfile()
+    assert collected_file.read() == 'collected = 3'
+
+
 def test_funcarg_teardown_failure(testdir):
     p = testdir.makepyfile("""
         def pytest_funcarg__myarg(request):
@@ -407,7 +445,7 @@ def test_funcarg_teardown_failure(testdir):
         def test_hello(myarg):
             pass
     """)
-    result = testdir.runpytest("--debug", p) # , "-n1")
+    result = testdir.runpytest_subprocess("--debug", p) # , "-n1")
     result.stdout.fnmatch_lines([
         "*ValueError*42*",
         "*1 passed*1 error*",
@@ -454,8 +492,142 @@ def test_issue34_pluginloading_in_subprocess(testdir):
         def test_hello():
             assert pytest.sample_variable == "testing"
     """)
-    result = testdir.runpytest("-n1", "-p", "plugin123")
+    result = testdir.runpytest_subprocess("-n1", "-p", "plugin123")
     assert result.ret == 0
     result.stdout.fnmatch_lines([
         "*1 passed*",
     ])
+
+
+def test_fixture_scope_caching_issue503(testdir):
+    p1 = testdir.makepyfile("""
+            import pytest
+
+            @pytest.fixture(scope='session')
+            def fix():
+                assert fix.counter == 0, 'session fixture was invoked multiple times'
+                fix.counter += 1
+            fix.counter = 0
+
+            def test_a(fix):
+                pass
+
+            def test_b(fix):
+                pass
+    """)
+    result = testdir.runpytest(p1, '-v', '-n1')
+    assert result.ret == 0
+    result.stdout.fnmatch_lines([
+    "*2 passed*",
+    ])
+
+
+def test_issue_594_random_parametrize(testdir):
+    """
+    Make sure that tests that are randomly parametrized display an appropriate
+    error message, instead of silently skipping the entire test run.
+    """
+    p1 = testdir.makepyfile("""
+        import pytest
+        import random
+
+        xs = list(range(10))
+        random.shuffle(xs)
+        @pytest.mark.parametrize('x', xs)
+        def test_foo(x):
+            assert 1
+    """)
+    result = testdir.runpytest(p1, '-v', '-n4')
+    assert result.ret == 1
+    result.stdout.fnmatch_lines([
+        "Different tests were collected between gw* and gw*",
+    ])
+
+
+class TestNodeFailure:
+
+    def test_load_single(self, testdir):
+        f = testdir.makepyfile("""
+            import os
+            def test_a(): os._exit(1)
+            def test_b(): pass
+        """)
+        res = testdir.runpytest(f, '-n1')
+        res.stdout.fnmatch_lines([
+            "*Replacing crashed slave*",
+            "*Slave*crashed while running*",
+            "*1 failed*1 passed*",
+        ])
+
+    def test_load_multiple(self, testdir):
+        f = testdir.makepyfile("""
+            import os
+            def test_a(): pass
+            def test_b(): os._exit(1)
+            def test_c(): pass
+            def test_d(): pass
+        """)
+        res = testdir.runpytest(f, '-n2')
+        res.stdout.fnmatch_lines([
+            "*Replacing crashed slave*",
+            "*Slave*crashed while running*",
+            "*1 failed*3 passed*",
+        ])
+
+    def test_each_single(self, testdir):
+        f = testdir.makepyfile("""
+            import os
+            def test_a(): os._exit(1)
+            def test_b(): pass
+        """)
+        res = testdir.runpytest(f, '--dist=each', '--tx=popen')
+        res.stdout.fnmatch_lines([
+            "*Replacing crashed slave*",
+            "*Slave*crashed while running*",
+            "*1 failed*1 passed*",
+        ])
+
+    def test_each_multiple(self, testdir):
+        f = testdir.makepyfile("""
+            import os
+            def test_a(): os._exit(1)
+            def test_b(): pass
+        """)
+        res = testdir.runpytest(f, '--dist=each', '--tx=2*popen')
+        res.stdout.fnmatch_lines([
+            "*Replacing crashed slave*",
+            "*Slave*crashed while running*",
+            "*2 failed*2 passed*",
+        ])
+
+    def test_max_slave_restart(self, testdir):
+        f = testdir.makepyfile("""
+            import os
+            def test_a(): pass
+            def test_b(): os._exit(1)
+            def test_c(): os._exit(1)
+            def test_d(): pass
+        """)
+        res = testdir.runpytest(f, '-n4', '--max-slave-restart=1')
+        res.stdout.fnmatch_lines([
+            "*Replacing crashed slave*",
+            "*Maximum crashed slaves reached: 1*",
+            "*Slave*crashed while running*",
+            "*Slave*crashed while running*",
+            "*2 failed*2 passed*",
+        ])
+
+
+    def test_disable_restart(self, testdir):
+        f = testdir.makepyfile("""
+            import os
+            def test_a(): pass
+            def test_b(): os._exit(1)
+            def test_c(): pass
+        """)
+        res = testdir.runpytest(f, '-n4', '--max-slave-restart=0')
+        res.stdout.fnmatch_lines([
+            "*Slave restarting disabled*",
+            "*Slave*crashed while running*",
+            "*1 failed*2 passed*",
+        ])

testing/conftest.py

@@ -1,20 +1,43 @@
 import py
+import pytest
 import execnet
 
+@pytest.fixture(scope="session", autouse=True)
+def _ensure_imports():
+    # we import some modules because pytest-2.8's testdir fixture
+    # will unload all modules after each test and this cause
+    # (unknown) problems with execnet.Group()
+    execnet.Group
+    execnet.makegateway
+
 pytest_plugins = "pytester"
 
 #rsyncdirs = ['.', '../xdist', py.path.local(execnet.__file__).dirpath()]
 
+@pytest.fixture(autouse=True)
+def _divert_atexit(request, monkeypatch):
+    import atexit
+    l = []
+    def finish():
+        while l:
+            l.pop()()
+    monkeypatch.setattr(atexit, "register", l.append)
+    request.addfinalizer(finish)
+
 def pytest_addoption(parser):
     parser.addoption('--gx',
-       action="append", dest="gspecs", default=None,
+       action="append", dest="gspecs",
        help=("add a global test environment, XSpec-syntax. "))
 
 def pytest_funcarg__specssh(request):
     return getspecssh(request.config)
-def getgspecs(config):
-    return [execnet.XSpec(spec)
-                for spec in config.getvalueorskip("gspecs")]
+
+@pytest.fixture
+def testdir(testdir):
+    # pytest before 2.8 did not have a runpytest_subprocess
+    if not hasattr(testdir, "runpytest_subprocess"):
+        testdir.runpytest_subprocess = testdir.runpytest
+    return testdir
 
 # configuration information for tests
 def getgspecs(config):

testing/test_boxed.py

@@ -1,6 +1,10 @@
-import py
+import pytest
+import os
 
-@py.test.mark.skipif("not hasattr(os, 'fork')")
+needsfork = pytest.mark.skipif(not hasattr(os, "fork"),
+                               reason="os.fork required")
+
+@needsfork
 def test_functional_boxed(testdir):
     p1 = testdir.makepyfile("""
         import os
@@ -13,12 +17,36 @@ def test_functional_boxed(testdir):
         "*1 failed*"
     ])
 
+@needsfork
+@pytest.mark.parametrize("capmode", [
+    "no",
+    pytest.mark.xfail("sys", reason="capture cleanup needed"),
+    pytest.mark.xfail("fd", reason="capture cleanup needed")])
+def test_functional_boxed_capturing(testdir, capmode):
+    p1 = testdir.makepyfile("""
+        import os
+        import sys
+        def test_function():
+            sys.stdout.write("hello\\n")
+            sys.stderr.write("world\\n")
+            os.kill(os.getpid(), 15)
+    """)
+    result = testdir.runpytest(p1, "--boxed", "--capture=%s" % capmode)
+    result.stdout.fnmatch_lines("""
+        *CRASHED*
+        *stdout*
+        hello
+        *stderr*
+        world
+        *1 failed*
+    """)
+
 class TestOptionEffects:
     def test_boxed_option_default(self, testdir):
         tmpdir = testdir.tmpdir.ensure("subdir", dir=1)
         config = testdir.parseconfig()
         assert not config.option.boxed
-        py.test.importorskip("execnet")
+        pytest.importorskip("execnet")
         config = testdir.parseconfig('-d', tmpdir)
         assert not config.option.boxed
 

testing/test_dsession.py

@@ -4,7 +4,6 @@ from xdist.dsession import (
     EachScheduling,
     report_collection_diff,
 )
-from _pytest import main as outcome
 import py
 import pytest
 import execnet
@@ -84,11 +83,10 @@ class TestEachScheduling:
 
 class TestLoadScheduling:
     def test_schedule_load_simple(self):
-        node1 = MockNode()
-        node2 = MockNode()
         sched = LoadScheduling(2)
-        sched.addnode(node1)
-        sched.addnode(node2)
+        sched.addnode(MockNode())
+        sched.addnode(MockNode())
+        node1, node2 = sched.nodes
         collection = ["a.py::test_1", "a.py::test_2"]
         assert not sched.collection_is_completed
         sched.addnode_collection(node1, collection)
@@ -98,38 +96,40 @@ class TestLoadScheduling:
         assert sched.node2collection[node1] == collection
         assert sched.node2collection[node2] == collection
         sched.init_distribute()
-        assert sched.tests_finished()
-        assert len(node1.sent) == 1
-        assert len(node2.sent) == 1
-        x = sorted(node1.sent + node2.sent)
-        assert x == [0, 1]
-        sched.remove_item(node1, node1.sent[0])
-        sched.remove_item(node2, node2.sent[0])
-        assert sched.tests_finished()
         assert not sched.pending
+        assert not sched.tests_finished()
+        assert len(node1.sent) == 2
+        assert len(node2.sent) == 0
+        assert node1.sent == [0, 1]
+        sched.remove_item(node1, node1.sent[0])
+        assert sched.tests_finished()
+        sched.remove_item(node1, node1.sent[1])
+        assert sched.tests_finished()
 
     def test_init_distribute_chunksize(self):
         sched = LoadScheduling(2)
-        node1 = MockNode()
-        node2 = MockNode()
-        sched.addnode(node1)
-        sched.addnode(node2)
-        col = ["xyz"] * (3)
+        sched.addnode(MockNode())
+        sched.addnode(MockNode())
+        node1, node2 = sched.nodes
+        col = ["xyz"] * (6)
         sched.addnode_collection(node1, col)
         sched.addnode_collection(node2, col)
         sched.init_distribute()
         #assert not sched.tests_finished()
         sent1 = node1.sent
         sent2 = node2.sent
-        chunkitems = col[:1]
-        assert (sent1 == [0] and sent2 == [1]) or (
-                sent1 == [1] and sent2 == [0])
+        assert sent1 == [0, 1]
+        assert sent2 == [2, 3]
+        assert sched.pending == [4, 5]
         assert sched.node2pending[node1] == sent1
         assert sched.node2pending[node2] == sent2
-        assert len(sched.pending) == 1
-        for node in (node1, node2):
-            for i in sched.node2pending[node]:
-                sched.remove_item(node, i)
+        assert len(sched.pending) == 2
+        sched.remove_item(node1, 0)
+        assert node1.sent == [0, 1, 4]
+        assert sched.pending == [5]
+        assert node2.sent == [2, 3]
+        sched.remove_item(node1, 1)
+        assert node1.sent == [0, 1, 4, 5]
         assert not sched.pending
 
     def test_add_remove_node(self):
@@ -144,6 +144,37 @@ class TestLoadScheduling:
         crashitem = sched.remove_node(node)
         assert crashitem == collection[0]
 
+    def test_different_tests_collected(self, testdir):
+        """
+        Test that LoadScheduling is reporting collection errors when
+        different test ids are collected by slaves.
+        """
+        class CollectHook(object):
+            """
+            Dummy hook that stores collection reports.
+            """
+
+            def __init__(self):
+                self.reports = []
+
+            def pytest_collectreport(self, report):
+                self.reports.append(report)
+
+        collect_hook = CollectHook()
+        config = testdir.parseconfig()
+        config.pluginmanager.register(collect_hook, "collect_hook")
+        node1 = MockNode()
+        node2 = MockNode()
+        sched = LoadScheduling(2, config=config)
+        sched.addnode(node1)
+        sched.addnode(node2)
+        sched.addnode_collection(node1, ["a.py::test_1"])
+        sched.addnode_collection(node2, ["a.py::test_2"])
+        sched.init_distribute()
+        assert len(collect_hook.reports) == 1
+        rep = collect_hook.reports[0]
+        assert 'Different tests were collected between' in rep.longrepr
+
 
 class TestDistReporter:
 
@@ -179,7 +210,7 @@ class TestDistReporter:
 def test_report_collection_diff_equal():
     """Test reporting of equal collections."""
     from_collection = to_collection = ['aaa', 'bbb', 'ccc']
-    assert report_collection_diff(from_collection, to_collection, 1, 2)
+    assert report_collection_diff(from_collection, to_collection, 1, 2) is None
 
 
 def test_report_collection_diff_different():
@@ -202,10 +233,8 @@ def test_report_collection_diff_different():
         '-YYY'
     )
 
-    try:
-        report_collection_diff(from_collection, to_collection, 1, 2)
-    except AssertionError as e:
-        assert py.builtin._totext(e) == error_message
+    msg = report_collection_diff(from_collection, to_collection, 1, 2)
+    assert msg == error_message
 
 @pytest.mark.xfail(reason="duplicate test ids not supported yet")
 def test_pytest_issue419(testdir):

testing/test_looponfail.py

@@ -42,7 +42,7 @@ class TestStatRecorder:
 
     def test_dirchange(self, tmpdir):
         tmp = tmpdir
-        hello = tmp.ensure("dir", "hello.py")
+        tmp.ensure("dir", "hello.py")
         sd = StatRecorder([tmp])
         assert not sd.fil(tmp.join("dir"))
 

testing/test_plugin.py

@@ -13,7 +13,7 @@ def test_dist_incompatibility_messages(testdir):
     assert "incompatible" in result.stderr.str()
 
 def test_dist_options(testdir):
-    from xdist.plugin import check_options
+    from xdist.plugin import pytest_cmdline_main as check_options
     config = testdir.parseconfigure("-n 2")
     check_options(config)
     assert config.option.dist == "load"
@@ -67,4 +67,3 @@ class TestDistOptions:
         assert py.path.local('y') in roots
         assert py.path.local('z') in roots
         assert testdir.tmpdir.join('x') in roots
-

testing/test_remote.py

@@ -3,7 +3,6 @@ from xdist.slavemanage import SlaveController, unserialize_report
 from xdist.remote import serialize_report
 import execnet
 queue = py.builtin._tryimport("queue", "Queue")
-from py.builtin import print_
 import marshal
 
 WAIT_TIMEOUT = 10.0
@@ -26,7 +25,7 @@ class SlaveSetup:
     use_callback = False
 
     def __init__(self, request):
-        self.testdir = testdir = request.getfuncargvalue("testdir")
+        self.testdir = request.getfuncargvalue("testdir")
         self.request = request
         self.events = queue.Queue()
 
@@ -140,7 +139,7 @@ class TestReportSerialization:
 
 class TestSlaveInteractor:
     def test_basic_collect_and_runtests(self, slave):
-        p = slave.testdir.makepyfile("""
+        slave.testdir.makepyfile("""
             def test_func():
                 pass
         """)
@@ -170,7 +169,7 @@ class TestSlaveInteractor:
         assert 'slaveoutput' in ev.kwargs
 
     def test_remote_collect_skip(self, slave):
-        p = slave.testdir.makepyfile("""
+        slave.testdir.makepyfile("""
             import py
             py.test.skip("hello")
         """)
@@ -187,7 +186,7 @@ class TestSlaveInteractor:
         assert not ev.kwargs['ids']
 
     def test_remote_collect_fail(self, slave):
-        p = slave.testdir.makepyfile("""aasd qwe""")
+        slave.testdir.makepyfile("""aasd qwe""")
         slave.setup()
         ev = slave.popevent("collectionstart")
         assert not ev.kwargs
@@ -201,7 +200,7 @@ class TestSlaveInteractor:
         assert not ev.kwargs['ids']
 
     def test_runtests_all(self, slave):
-        p = slave.testdir.makepyfile("""
+        slave.testdir.makepyfile("""
             def test_func(): pass
             def test_func2(): pass
         """)
@@ -228,7 +227,7 @@ class TestSlaveInteractor:
     def test_happy_run_events_converted(self, testdir, slave):
         py.test.xfail("implement a simple test for event production")
         assert not slave.use_callback
-        p = slave.testdir.makepyfile("""
+        slave.testdir.makepyfile("""
             def test_func():
                 pass
         """)

testing/test_slavemanage.py

@@ -1,28 +1,35 @@
 import py
-import os
+import pytest
 import execnet
+from _pytest.pytester import HookRecorder
+from xdist import slavemanage, newhooks
 from xdist.slavemanage import HostRSync, NodeManager
 
-pytest_plugins = "pytester",
+pytest_plugins = "pytester"
 
-def pytest_funcarg__hookrecorder(request):
-    _pytest = request.getfuncargvalue('_pytest')
-    config = request.getfuncargvalue('config')
-    return _pytest.gethookrecorder(config.hook)
+def pytest_funcarg__hookrecorder(request, config):
+    hookrecorder = HookRecorder(config.pluginmanager)
+    if hasattr(hookrecorder, "start_recording"):
+        hookrecorder.start_recording(newhooks)
+    request.addfinalizer(hookrecorder.finish_recording)
+    return hookrecorder
 
-def pytest_funcarg__config(request):
-    testdir = request.getfuncargvalue("testdir")
-    config = testdir.parseconfig()
-    return config
+def pytest_funcarg__config(testdir):
+    return testdir.parseconfig()
 
-def pytest_funcarg__mysetup(request):
+def pytest_funcarg__mysetup(tmpdir):
     class mysetup:
-        def __init__(self, request):
-            temp = request.getfuncargvalue("tmpdir")
-            self.source = temp.mkdir("source")
-            self.dest = temp.mkdir("dest")
-            request.getfuncargvalue("_pytest")
-    return mysetup(request)
+        source = tmpdir.mkdir("source")
+        dest = tmpdir.mkdir("dest")
+    return mysetup()
+
+@pytest.fixture
+def slavecontroller(monkeypatch):
+    class MockController(object):
+        def __init__(self, *args): pass
+        def setup(self): pass
+    monkeypatch.setattr(slavemanage, 'SlaveController', MockController)
+    return MockController
 
 class TestNodeManagerPopen:
     def test_popen_no_default_chdir(self, config):
@@ -36,9 +43,9 @@ class TestNodeManagerPopen:
         for spec in NodeManager(config, l, defaultchdir="abc").specs:
             assert spec.chdir == "abc"
 
-    def test_popen_makegateway_events(self, config, hookrecorder, _pytest):
+    def test_popen_makegateway_events(self, config, hookrecorder, slavecontroller):
         hm = NodeManager(config, ["popen"] * 2)
-        hm.makegateways()
+        hm.setup_nodes(None)
         call = hookrecorder.popcall("pytest_xdist_setupnodes")
         assert len(call.specs) == 2
 
@@ -51,10 +58,10 @@ class TestNodeManagerPopen:
         hm.teardown_nodes()
         assert not len(hm.group)
 
-    def test_popens_rsync(self, config, mysetup):
+    def test_popens_rsync(self, config, mysetup, slavecontroller):
         source = mysetup.source
         hm = NodeManager(config, ["popen"] * 2)
-        hm.makegateways()
+        hm.setup_nodes(None)
         assert len(hm.group) == 2
         for gw in hm.group:
             class pseudoexec:
@@ -65,19 +72,21 @@ class TestNodeManagerPopen:
                     pass
             gw.remote_exec = pseudoexec
         l = []
-        hm.rsync(source, notify=lambda *args: l.append(args))
+        for gw in hm.group:
+            hm.rsync(gw, source, notify=lambda *args: l.append(args))
         assert not l
         hm.teardown_nodes()
         assert not len(hm.group)
         assert "sys.path.insert" in gw.remote_exec.args[0]
 
-    def test_rsync_popen_with_path(self, config, mysetup):
+    def test_rsync_popen_with_path(self, config, mysetup, slavecontroller):
         source, dest = mysetup.source, mysetup.dest
-        hm = NodeManager(config, ["popen//chdir=%s" %dest] * 1)
-        hm.makegateways()
+        hm = NodeManager(config, ["popen//chdir=%s" % dest] * 1)
+        hm.setup_nodes(None)
         source.ensure("dir1", "dir2", "hello")
         l = []
-        hm.rsync(source, notify=lambda *args: l.append(args))
+        for gw in hm.group:
+            hm.rsync(gw, source, notify=lambda *args: l.append(args))
         assert len(l) == 1
         assert l[0] == ("rsyncrootready", hm.group['gw0'].spec, source)
         hm.teardown_nodes()
@@ -86,12 +95,15 @@ class TestNodeManagerPopen:
         assert dest.join("dir1", "dir2").check()
         assert dest.join("dir1", "dir2", 'hello').check()
 
-    def test_rsync_same_popen_twice(self, config, mysetup, hookrecorder):
+    def test_rsync_same_popen_twice(self, config, mysetup,
+                                    hookrecorder, slavecontroller):
         source, dest = mysetup.source, mysetup.dest
-        hm = NodeManager(config, ["popen//chdir=%s" %dest] * 2)
-        hm.makegateways()
+        hm = NodeManager(config, ["popen//chdir=%s" % dest] * 2)
+        hm.roots = []
+        hm.setup_nodes(None)
         source.ensure("dir1", "dir2", "hello")
-        hm.rsync(source)
+        gw = hm.group[0]
+        hm.rsync(gw, source)
         call = hookrecorder.popcall("pytest_xdist_rsyncstart")
         assert call.source == source
         assert len(call.gateways) == 1
@@ -99,16 +111,8 @@ class TestNodeManagerPopen:
         call = hookrecorder.popcall("pytest_xdist_rsyncfinish")
 
 class TestHRSync:
-    def pytest_funcarg__mysetup(self, request):
-        class mysetup:
-            def __init__(self, request):
-                tmp = request.getfuncargvalue('tmpdir')
-                self.source = tmp.mkdir("source")
-                self.dest = tmp.mkdir("dest")
-        return mysetup(request)
-
     def test_hrsync_filter(self, mysetup):
-        source, dest = mysetup.source, mysetup.dest
+        source, _ = mysetup.source, mysetup.dest  # noqa
         source.ensure("dir", "file.txt")
         source.ensure(".svn", "entries")
         source.ensure(".somedotfile", "moreentries")
@@ -136,10 +140,10 @@ class TestHRSync:
 
 
 class TestNodeManager:
-    @py.test.mark.xfail
+    @py.test.mark.xfail(run=False)
     def test_rsync_roots_no_roots(self, testdir, mysetup):
         mysetup.source.ensure("dir1", "file1").write("hello")
-        config = testdir.parseconfig(source)
+        config = testdir.parseconfig(mysetup.source)
         nodemanager = NodeManager(config, ["popen//chdir=%s" % mysetup.dest])
         #assert nodemanager.config.topdir == source == config.topdir
         nodemanager.makegateways()
@@ -152,7 +156,7 @@ class TestNodeManager:
         assert p.join("dir1").check()
         assert p.join("dir1", "file1").check()
 
-    def test_popen_rsync_subdir(self, testdir, mysetup):
+    def test_popen_rsync_subdir(self, testdir, mysetup, slavecontroller):
         source, dest = mysetup.source, mysetup.dest
         dir1 = mysetup.source.mkdir("dir1")
         dir2 = dir1.mkdir("dir2")
@@ -164,8 +168,7 @@ class TestNodeManager:
                 "--rsyncdir", rsyncroot,
                 source,
             ))
-            nodemanager.makegateways()
-            nodemanager.rsync_roots()
+            nodemanager.setup_nodes(None)  # calls .rsync_roots()
             if rsyncroot == source:
                 dest = dest.join("source")
             assert dest.join("dir1").check()
@@ -173,7 +176,7 @@ class TestNodeManager:
             assert dest.join("dir1", "dir2", 'hello').check()
             nodemanager.teardown_nodes()
 
-    def test_init_rsync_roots(self, testdir, mysetup):
+    def test_init_rsync_roots(self, testdir, mysetup, slavecontroller):
         source, dest = mysetup.source, mysetup.dest
         dir2 = source.ensure("dir1", "dir2", dir=1)
         source.ensure("dir1", "somefile", dir=1)
@@ -185,20 +188,19 @@ class TestNodeManager:
         """))
         config = testdir.parseconfig(source)
         nodemanager = NodeManager(config, ["popen//chdir=%s" % dest])
-        nodemanager.makegateways()
-        nodemanager.rsync_roots()
+        nodemanager.setup_nodes(None)  # calls .rsync_roots()
         assert dest.join("dir2").check()
         assert not dest.join("dir1").check()
         assert not dest.join("bogus").check()
 
-    def test_rsyncignore(self, testdir, mysetup):
+    def test_rsyncignore(self, testdir, mysetup, slavecontroller):
         source, dest = mysetup.source, mysetup.dest
         dir2 = source.ensure("dir1", "dir2", dir=1)
-        dir5 = source.ensure("dir5", "dir6", "bogus")
-        dirf = source.ensure("dir5", "file")
+        source.ensure("dir5", "dir6", "bogus")
+        source.ensure("dir5", "file")
         dir2.ensure("hello")
-        dirfoo = source.ensure("foo", "bar")
-        dirbar = source.ensure("bar", "foo")
+        source.ensure("foo", "bar")
+        source.ensure("bar", "foo")
         source.join("tox.ini").write(py.std.textwrap.dedent("""
             [pytest]
             rsyncdirs = dir1 dir5
@@ -207,24 +209,22 @@ class TestNodeManager:
         config = testdir.parseconfig(source)
         config.option.rsyncignore = ['bar']
         nodemanager = NodeManager(config, ["popen//chdir=%s" % dest])
-        nodemanager.makegateways()
-        nodemanager.rsync_roots()
+        nodemanager.setup_nodes(None)  # calls .rsync_roots()
         assert dest.join("dir1").check()
         assert not dest.join("dir1", "dir2").check()
-        assert dest.join("dir5","file").check()
+        assert dest.join("dir5", "file").check()
         assert not dest.join("dir6").check()
         assert not dest.join('foo').check()
         assert not dest.join('bar').check()
 
-    def test_optimise_popen(self, testdir, mysetup):
-        source, dest = mysetup.source, mysetup.dest
+    def test_optimise_popen(self, testdir, mysetup, slavecontroller):
+        source = mysetup.source
         specs = ["popen"] * 3
         source.join("conftest.py").write("rsyncdirs = ['a']")
         source.ensure('a', dir=1)
         config = testdir.parseconfig(source)
         nodemanager = NodeManager(config, specs)
-        nodemanager.makegateways()
-        nodemanager.rsync_roots()
+        nodemanager.setup_nodes(None)  # calls .rysnc_roots()
         for gwspec in nodemanager.specs:
             assert gwspec._samefilesystem()
             assert not gwspec.chdir
@@ -235,8 +235,6 @@ class TestNodeManager:
                 pass
         """)
         reprec = testdir.inline_run("-d", "--rsyncdir=%s" % testdir.tmpdir,
-                "--tx", specssh, testdir.tmpdir)
+                                    "--tx", specssh, testdir.tmpdir)
         rep, = reprec.getreports("pytest_runtest_logreport")
         assert rep.passed
-
-

tox.ini

@@ -1,26 +1,27 @@
 [tox]
-envlist=py26,py32,py33,py27,py27-pexpect,py33-pexpect,py26,py26-old,py33-old
+envlist=
+  py{26,33,34,27}-pytest2{4,5,6,7},py{27,34}-pytest27-pexpect,flakes
+
 
 [testenv]
 changedir=testing
-deps=pytest>=2.5.1
-commands= py.test --junitxml={envlogdir}/junit-{envname}.xml []
+deps =
+    pycmd
+    pytest24: pytest~=2.4.0
+    pytest25: pytest~=2.5.0
 
-[testenv:py27-pexpect]
-deps={[testenv]deps}
-    pexpect
-[testenv:py33-pexpect]
-deps={[testenv]deps}
-    pexpect
+    pytest26: pytest~=2.6.1
+    pytest27: pytest~=2.7.2
+    pexpect: pexpect
+commands=
+  # always clean to avoid code unmarshal mismatch on old python/pytest
+  py.cleanup -aq
+  py.test {posargs}
 
-[testenv:py26-old]
-deps=
-    pytest==2.4.2
-
-[testenv:py33-old]
-basepython = python3.3
-deps=
-    pytest==2.4.2
+[testenv:flakes]
+changedir=
+deps = pytest-flakes>=0.2
+commands = py.test --flakes -m flakes testing xdist
 
 [pytest]
 addopts = -rsfxX

xdist/__init__.py

@@ -1,2 +1,2 @@
-#
-__version__ = '1.10'
+__all__ = ['__version__']
+from xdist._version import version as __version__

xdist/boxed.py

@@ -0,0 +1,56 @@
+
+import py
+
+
+def pytest_addoption(parser):
+    group = parser.getgroup("xdist", "distributed and subprocess testing")
+    group.addoption('--boxed',
+           action="store_true", dest="boxed", default=False,
+           help="box each test run in a separate process (unix)")
+
+
+def pytest_runtest_protocol(item):
+    if item.config.getvalue("boxed"):
+        reports = forked_run_report(item)
+        for rep in reports:
+            item.ihook.pytest_runtest_logreport(report=rep)
+        return True
+
+def forked_run_report(item):
+    # for now, we run setup/teardown in the subprocess
+    # XXX optionally allow sharing of setup/teardown
+    from _pytest.runner import runtestprotocol
+    EXITSTATUS_TESTEXIT = 4
+    import marshal
+    from xdist.remote import serialize_report
+    from xdist.slavemanage import unserialize_report
+    def runforked():
+        try:
+            reports = runtestprotocol(item, log=False)
+        except KeyboardInterrupt:
+            py.std.os._exit(EXITSTATUS_TESTEXIT)
+        return marshal.dumps([serialize_report(x) for x in reports])
+
+    ff = py.process.ForkedFunc(runforked)
+    result = ff.waitfinish()
+    if result.retval is not None:
+        report_dumps = marshal.loads(result.retval)
+        return [unserialize_report("testreport", x) for x in report_dumps]
+    else:
+        if result.exitstatus == EXITSTATUS_TESTEXIT:
+            py.test.exit("forked test item %s raised Exit" %(item,))
+        return [report_process_crash(item, result)]
+
+def report_process_crash(item, result):
+    path, lineno = item._getfslineno()
+    info = ("%s:%s: running the test CRASHED with signal %d" %
+            (path, lineno, result.signal))
+    from _pytest import runner
+    call = runner.CallInfo(lambda: 0/0, "???")
+    call.excinfo = info
+    rep = runner.pytest_runtest_makereport(item, call)
+    if result.out:
+        rep.sections.append(("captured stdout", result.out))
+    if result.err:
+        rep.sections.append(("captured stderr", result.err))
+    return rep

xdist/dsession.py

@@ -1,5 +1,5 @@
-import sys
 import difflib
+from _pytest.runner import CollectReport
 
 import pytest
 import py
@@ -10,35 +10,96 @@ queue = py.builtin._tryimport('queue', 'Queue')
 
 
 class EachScheduling:
+    """Implement scheduling of test items on all nodes
+
+    If a node gets added after the test run is started then it is
+    assumed to replace a node which got removed before it finished
+    it's collection.  In this case it will only be used if a a node
+    with the same spec got removed earlier.
+
+    Any nodes added after the run is started will only get items
+    assigned if a node with matching spec was removed before it
+    finished all it's pending items.  The new node will then be
+    assigned the remaining items from the removed node.
+    """
 
     def __init__(self, numnodes, log=None):
         self.numnodes = numnodes
         self.node2collection = {}
         self.node2pending = {}
+        self._started = []
+        self._removed2pending = {}
         if log is None:
             self.log = py.log.Producer("eachsched")
         else:
-            self.log = log.loadsched
+            self.log = log.eachsched
         self.collection_is_completed = False
 
+    @property
+    def nodes(self):
+        """A list of all nodes in the scheduler."""
+        return list(self.node2pending.keys())
+
     def hasnodes(self):
         return bool(self.node2pending)
 
+    def haspending(self):
+        """Return True if there are pending test items
+
+        This indicates that collection has finished and nodes are
+        still processing test items, so can be thought of as "the
+        scheduler is active".
+        """
+        for pending in self.node2pending.values():
+            if pending:
+                return True
+        return False
+
     def addnode(self, node):
-        self.node2collection[node] = None
+        assert node not in self.node2pending
+        self.node2pending[node] = []
 
     def tests_finished(self):
         if not self.collection_is_completed:
             return False
+        if self._removed2pending:
+            return False
+        for pending in self.node2pending.values():
+            if len(pending) >= 2:
+                return False
         return True
 
     def addnode_collection(self, node, collection):
-        assert not self.collection_is_completed
-        assert self.node2collection[node] is None
-        self.node2collection[node] = list(collection)
-        self.node2pending[node] = []
-        if len(self.node2pending) >= self.numnodes:
-            self.collection_is_completed = True
+        """Add the collected test items from a node
+
+        Collection is complete once all nodes have submitted their
+        collection.  In this case it's peding list is set to an empty
+        list.  When the collection is already completed this
+        submission is from a node which was restarted to replace a
+        dead node.  In this case we already assing the pending items
+        here.  In either case ``.init_distribute()`` will instruct the
+        node to start running the required tests.
+        """
+        assert node in self.node2pending
+        if not self.collection_is_completed:
+            self.node2collection[node] = list(collection)
+            self.node2pending[node] = []
+            if len(self.node2collection) >= self.numnodes:
+                self.collection_is_completed = True
+        elif self._removed2pending:
+            for deadnode in self._removed2pending:
+                if deadnode.gateway.spec == node.gateway.spec:
+                    dead_collection = self.node2collection[deadnode]
+                    if collection != dead_collection:
+                        msg = report_collection_diff(dead_collection,
+                                                     collection,
+                                                     deadnode.gateway.id,
+                                                     node.gateway.id)
+                        self.log(msg)
+                        return
+                    pending = self._removed2pending.pop(deadnode)
+                    self.node2pending[node] = pending
+                    break
 
     def remove_item(self, node, item_index, duration=0):
         self.node2pending[node].remove(item_index)
@@ -49,129 +110,309 @@ class EachScheduling:
         if not pending:
             return
         crashitem = self.node2collection[node][pending.pop(0)]
-        # XXX what about the rest of pending?
+        if pending:
+            self._removed2pending[node] = pending
         return crashitem
 
     def init_distribute(self):
+        """Schedule the test items on the nodes
+
+        If the node's pending list is empty it is a new node which
+        needs to run all the tests.  If the pending list is already
+        populated (by ``.addnode_collection()``) then it replaces a
+        died node and we only need to run those tests.
+        """
         assert self.collection_is_completed
         for node, pending in self.node2pending.items():
-            node.send_runtest_all()
-            pending[:] = range(len(self.node2collection[node]))
+            if node in self._started:
+                continue
+            if not pending:
+                pending[:] = range(len(self.node2collection[node]))
+                node.send_runtest_all()
+            else:
+                node.send_runtest_some(pending)
+            self._started.append(node)
+
 
 class LoadScheduling:
-    def __init__(self, numnodes, log=None):
+    """Implement load scheduling accross nodes.
+
+    This distributes the tests collected across all nodes so each test
+    is run just once.  All nodes collect and submit the test suit and
+    when all collections are received it is verified they are
+    identical collections.  Then the collection gets devided up in
+    chunks and chunks get submitted to nodes.  Whenver a node finishes
+    an item they call ``.remove_item()`` which will trigger the
+    scheduler to assign more tests if the number of pending tests for
+    the node falls below a low-watermark.
+
+    When created ``numnodes`` defines how many nodes are expected to
+    submit a collection, this is used to know when all nodes have
+    finished collection or how large the chunks need to be created.
+
+    Attributes:
+
+    :numnodes: The expected number of nodes taking part.  The actual
+       number of nodes will vary during the scheduler's lifetime as
+       nodes are added by the DSession as they are brought up and
+       removed either because of a died node or normal shutdown.  This
+       number is primarily used to know when the initial collection is
+       completed.
+
+    :node2collection: Map of nodes and their test collection.  All
+       collections should always be identical.
+
+    :node2pending: Map of nodes and the indices of their pending
+       tests.  The indices are an index into ``.pending`` (which is
+       identical to their own collection stored in
+       ``.node2collection``).
+
+    :collection: The one collection once it is validated to be
+       identical between all the nodes.  It is initialised to None
+       until ``.init_distribute()`` is called.
+
+    :pending: List of indices of globally pending tests.  These are
+       tests which have not yet been allocated to a chunk for a node
+       to process.
+
+    :log: A py.log.Producer instance.
+
+    :config: Config object, used for handling hooks.
+    """
+
+    def __init__(self, numnodes, log=None, config=None):
         self.numnodes = numnodes
-        self.node2pending = {}
         self.node2collection = {}
+        self.node2pending = {}
         self.pending = []
+        self.collection = None
         if log is None:
             self.log = py.log.Producer("loadsched")
         else:
             self.log = log.loadsched
-        self.collection_is_completed = False
+        self.config = config
+
+    @property
+    def nodes(self):
+        """A list of all nodes in the scheduler."""
+        return list(self.node2pending.keys())
+
+    @property
+    def collection_is_completed(self):
+        """Boolean indication initial test collection is complete.
+
+        This is a boolean indicating all initial participating nodes
+        have finished collection.  The required number of initial
+        nodes is defined by ``.numnodes``.
+        """
+        return len(self.node2collection) >= self.numnodes
+
+    def haspending(self):
+        """Return True if there are pending test items
+
+        This indicates that collection has finished and nodes are
+        still processing test items, so can be thought of as "the
+        scheduler is active".
+        """
+        if self.pending:
+            return True
+        for pending in self.node2pending.values():
+            if pending:
+                return True
+        return False
 
     def hasnodes(self):
+        """Return True if nodes exist in the scheduler."""
         return bool(self.node2pending)
 
     def addnode(self, node):
+        """Add a new node in the scheduler.
+
+        From now on the node will be allocated chunks of tests to
+        execute.
+
+        Called by the ``DSession.slave_slaveready`` hook when it
+        sucessfully bootstrapped a new node.
+        """
+        assert node not in self.node2pending
         self.node2pending[node] = []
 
     def tests_finished(self):
-        if not self.collection_is_completed or self.pending:
+        """Return True if all tests have been executed by the nodes."""
+        if not self.collection_is_completed:
             return False
-        #for items in self.node2pending.values():
-        #    if items:
-        #        return False
+        if self.pending:
+            return False
+        for pending in self.node2pending.values():
+            if len(pending) >= 2:
+                return False
         return True
 
     def addnode_collection(self, node, collection):
-        assert not self.collection_is_completed
+        """Add the collected test items from a node
+
+        The collection is stored in the ``.node2collection`` map.
+        Called by the ``DSession.slave_collectionfinish`` hook.
+        """
         assert node in self.node2pending
+        if self.collection_is_completed:
+            # A new node has been added later, perhaps an original one died.
+            assert self.collection  # .init_distribute() should have
+                                    # been called by now
+            if collection != self.collection:
+                other_node = next(iter(self.node2collection.keys()))
+                msg = report_collection_diff(self.collection,
+                                             collection,
+                                             other_node.gateway.id,
+                                             node.gateway.id)
+                self.log(msg)
+                return
         self.node2collection[node] = list(collection)
-        if len(self.node2collection) >= self.numnodes:
-            self.collection_is_completed = True
 
     def remove_item(self, node, item_index, duration=0):
-        node_pending = self.node2pending[node]
-        node_pending.remove(item_index)
-        # pre-load items-to-test if the node may become ready
+        """Mark test item as completed by node
 
+        The duration it took to execute the item is used as a hint to
+        the scheduler.
+
+        This is called by the ``DSession.slave_testreport`` hook.
+        """
+        self.node2pending[node].remove(item_index)
+        self.check_schedule(node, duration=duration)
+
+    def check_schedule(self, node, duration=0):
+        """Maybe schedule new items on the node
+
+        If there are any globally pending nodes left then this will
+        check if the given node should be given any more tests.  The
+        ``duration`` of the last test is optionally used as a
+        heuristic to influence how many tests the node is assigned.
+        """
         if self.pending:
-            if duration >= 0.1 and node_pending:
-                # seems the node is doing long-running tests
-                # so let's rather wait with sending new items
-                return
-            # how many nodes do we have remaining per node roughly?
+            # how many nodes do we have?
             num_nodes = len(self.node2pending)
             # if our node goes below a heuristic minimum, fill it out to
             # heuristic maximum
-            items_per_node_min = max(
-                    1, len(self.pending) // num_nodes // 4)
-            items_per_node_max = max(
-                    1, len(self.pending) // num_nodes // 2)
-            if len(node_pending) <= items_per_node_min:
-                num_send = items_per_node_max - len(node_pending) + 1
+            items_per_node_min = max(2, len(self.pending) // num_nodes // 4)
+            items_per_node_max = max(2, len(self.pending) // num_nodes // 2)
+            node_pending = self.node2pending[node]
+            if len(node_pending) < items_per_node_min:
+                if duration >= 0.1 and len(node_pending) >= 2:
+                    # seems the node is doing long-running tests
+                    # and has enough items to continue
+                    # so let's rather wait with sending new items
+                    return
+                num_send = items_per_node_max - len(node_pending)
                 self._send_tests(node, num_send)
-
         self.log("num items waiting for node:", len(self.pending))
-        #self.log("node2pending:", self.node2pending)
 
     def remove_node(self, node):
+        """Remove an node from the scheduler
+
+        This should be called either when the node crashed or at
+        shutdown time.  In the former case any pending items assigned
+        to the node will be re-scheduled.  Called by the
+        ``DSession.slave_slavefinished`` and
+        ``DSession.slave_errordown`` hooks.
+
+        Return the item which was being executing while the node
+        crashed or None if the node has no more pending items.
+
+        """
         pending = self.node2pending.pop(node)
         if not pending:
             return
-        # the node must have crashed on the item if there are pending ones
+
+        # The node crashed, reassing pending items
         crashitem = self.collection[pending.pop(0)]
         self.pending.extend(pending)
+        for node in self.node2pending:
+            self.check_schedule(node)
         return crashitem
 
     def init_distribute(self):
-        assert self.collection_is_completed
-        # XXX allow nodes to have different collections
-        node_collection_items = list(self.node2collection.items())
-        first_node, col = node_collection_items[0]
-        for node, collection in node_collection_items[1:]:
-            report_collection_diff(
-                col,
-                collection,
-                first_node.gateway.id,
-                node.gateway.id,
-            )
+        """Initiate distribution of the test collection
 
-        # all collections are the same, good.
-        # we now create an index
-        self.collection = col
-        self.pending[:] = range(len(col))
-        if not col:
+        Initiate scheduling of the items across the nodes.  If this
+        gets called again later it behaves the same as calling
+        ``.check_schedule()`` on all nodes so that newly added nodes
+        will start to be used.
+
+        This is called by the ``DSession.slave_collectionfinish`` hook
+        if ``.collection_is_completed`` is True.
+
+        XXX Perhaps this method should have been called ".schedule()".
+        """
+        assert self.collection_is_completed
+
+        # Initial distribution already happend, reschedule on all nodes
+        if self.collection is not None:
+            for node in self.nodes:
+                self.check_schedule(node)
             return
+
+        # XXX allow nodes to have different collections
+        if not self._check_nodes_have_same_collection():
+            self.log('**Different tests collected, aborting run**')
+            return
+
+        # Collections are identical, create the index of pending items.
+        self.collection = list(self.node2collection.values())[0]
+        self.pending[:] = range(len(self.collection))
+        if not self.collection:
+            return
+
         # how many items per node do we have about?
         items_per_node = len(self.collection) // len(self.node2pending)
         # take a fraction of tests for initial distribution
-        node_chunksize = max(items_per_node // 4, 1)
+        node_chunksize = max(items_per_node // 4, 2)
         # and initialize each node with a chunk of tests
-        for node in self.node2pending:
+        for node in self.nodes:
             self._send_tests(node, node_chunksize)
 
-    #f = open("/tmp/sent", "w")
     def _send_tests(self, node, num):
         tests_per_node = self.pending[:num]
-        #print >>self.f, "sent", node, tests_per_node
         if tests_per_node:
             del self.pending[:num]
             self.node2pending[node].extend(tests_per_node)
             node.send_runtest_some(tests_per_node)
 
+    def _check_nodes_have_same_collection(self):
+        """Return True if all nodes have collected the same items.
+
+        If collections differ, this method returns False while logging
+        the collection differences and posting collection errors to
+        pytest_collectreport hook.
+        """
+        node_collection_items = list(self.node2collection.items())
+        first_node, col = node_collection_items[0]
+        same_collection = True
+        for node, collection in node_collection_items[1:]:
+            msg = report_collection_diff(
+                col,
+                collection,
+                first_node.gateway.id,
+                node.gateway.id,
+            )
+            if msg:
+                same_collection = False
+                self.log(msg)
+                if self.config is not None:
+                    rep = CollectReport(node.gateway.id, 'failed', longrepr=msg,
+                                        result=[])
+                    self.config.hook.pytest_collectreport(report=rep)
+
+        return same_collection
+
+
 def report_collection_diff(from_collection, to_collection, from_id, to_id):
     """Report the collected test difference between two nodes.
 
-    :returns: True if collections are equal.
-
-    :raises: AssertionError with a detailed error message describing the
-             difference between the collections.
-
+    :returns: detailed message describing the difference between the given
+    collections, or None if they are equal.
     """
     if from_collection == to_collection:
-        return True
+        return None
 
     diff = difflib.unified_diff(
         from_collection,
@@ -185,13 +426,26 @@ def report_collection_diff(from_collection, to_collection, from_id, to_id):
         '{diff}'
     ).format(from_id=from_id, to_id=to_id, diff='\n'.join(diff))
     msg = "\n".join([x.rstrip() for x in error_message.split("\n")])
-    raise AssertionError(msg)
+    return msg
 
 
 class Interrupted(KeyboardInterrupt):
     """ signals an immediate interruption. """
 
+
 class DSession:
+    """A py.test plugin which runs a distributed test session
+
+    At the beginning of the test session this creates a NodeManager
+    instance which creates and starts all nodes.  Nodes then emit
+    events processed in the pytest_runtestloop hook using the slave_*
+    methods.
+
+    Once a node is started it will automatically start running the
+    py.test mainloop with some custom hooks.  This means a node
+    automatically starts collecting tests.  Once tests are collected
+    it will wait for instructions.
+    """
     def __init__(self, config):
         self.config = config
         self.log = py.log.Producer("dsession")
@@ -201,7 +455,13 @@ class DSession:
         self.countfailures = 0
         self.maxfail = config.getvalue("maxfail")
         self.queue = queue.Queue()
+        self._session = None
         self._failed_collection_errors = {}
+        self._active_nodes = set()
+        self._failed_nodes_count = 0
+        self._max_slave_restart = self.config.getoption('max_slave_restart')
+        if self._max_slave_restart is not None:
+            self._max_slave_restart = int(self._max_slave_restart)
         try:
             self.terminal = config.pluginmanager.getplugin("terminalreporter")
         except KeyError:
@@ -210,20 +470,37 @@ class DSession:
             self.trdist = TerminalDistReporter(config)
             config.pluginmanager.register(self.trdist, "terminaldistreporter")
 
+    @property
+    def session_finished(self):
+        """Return True if the distributed session has finished
+
+        This means all nodes have executed all test items.  This is
+        used to by pytest_runtestloop to break out of it's loop.
+        """
+        return bool(self.shuttingdown and not self._active_nodes)
+
     def report_line(self, line):
         if self.terminal and self.config.option.verbose >= 0:
             self.terminal.write_line(line)
 
     @pytest.mark.trylast
     def pytest_sessionstart(self, session):
+        """Creates and starts the nodes.
+
+        The nodes are setup to put their events onto self.queue.  As
+        soon as nodes start they will emit the slave_slaveready event.
+        """
         self.nodemanager = NodeManager(self.config)
-        self.nodemanager.setup_nodes(putevent=self.queue.put)
+        nodes = self.nodemanager.setup_nodes(putevent=self.queue.put)
+        self._active_nodes.update(nodes)
+        self._session = session
 
     def pytest_sessionfinish(self, session):
-        """ teardown any resources after a test run. """
-        nm = getattr(self, 'nodemanager', None) # if not fully initialized
+        """Shutdown all nodes."""
+        nm = getattr(self, 'nodemanager', None)  # if not fully initialized
         if nm is not None:
             nm.teardown_nodes()
+        self._session = None
 
     def pytest_collection(self):
         # prohibit collection of test items in master process
@@ -233,13 +510,13 @@ class DSession:
         numnodes = len(self.nodemanager.specs)
         dist = self.config.getvalue("dist")
         if dist == "load":
-            self.sched = LoadScheduling(numnodes, log=self.log)
+            self.sched = LoadScheduling(numnodes, log=self.log,
+                                        config=self.config)
         elif dist == "each":
             self.sched = EachScheduling(numnodes, log=self.log)
         else:
             assert 0, dist
         self.shouldstop = False
-        self.session_finished = False
         while not self.session_finished:
             self.loop_once()
             if self.shouldstop:
@@ -247,7 +524,7 @@ class DSession:
         return True
 
     def loop_once(self):
-        """ process one callback from one of the slaves. """
+        """Process one callback from one of the slaves."""
         while 1:
             try:
                 eventcall = self.queue.get(timeout=2.0)
@@ -268,26 +545,40 @@ class DSession:
     #
 
     def slave_slaveready(self, node, slaveinfo):
+        """Emitted when a node first starts up.
+
+        This adds the node to the scheduler, nodes continue with
+        collection without any further input.
+        """
         node.slaveinfo = slaveinfo
         node.slaveinfo['id'] = node.gateway.id
         node.slaveinfo['spec'] = node.gateway.spec
         self.config.hook.pytest_testnodeready(node=node)
-        self.sched.addnode(node)
         if self.shuttingdown:
             node.shutdown()
+        else:
+            self.sched.addnode(node)
 
     def slave_slavefinished(self, node):
+        """Emitted when node executes its pytest_sessionfinish hook.
+
+        Removes the node from the scheduler.
+
+        The node might not be the scheduler if it had not emitted
+        slaveready before shutdown was triggered.
+        """
         self.config.hook.pytest_testnodedown(node=node, error=None)
-        if node.slaveoutput['exitstatus'] == 2: # keyboard-interrupt
+        if node.slaveoutput['exitstatus'] == 2:  # keyboard-interrupt
             self.shouldstop = "%s received keyboard-interrupt" % (node,)
             self.slave_errordown(node, "keyboard-interrupt")
             return
-        crashitem = self.sched.remove_node(node)
-        #assert not crashitem, (crashitem, node)
-        if self.shuttingdown and not self.sched.hasnodes():
-            self.session_finished = True
+        if node in self.sched.nodes:
+            crashitem = self.sched.remove_node(node)
+            assert not crashitem, (crashitem, node)
+        self._active_nodes.remove(node)
 
     def slave_errordown(self, node, error):
+        """Emitted by the SlaveController when a node dies."""
         self.config.hook.pytest_testnodedown(node=node, error=error)
         try:
             crashitem = self.sched.remove_node(node)
@@ -296,41 +587,85 @@ class DSession:
         else:
             if crashitem:
                 self.handle_crashitem(crashitem, node)
-                #self.report_line("item crashed on node: %s" % crashitem)
-        if not self.sched.hasnodes():
-            self.session_finished = True
+
+        self._failed_nodes_count += 1
+        maximum_reached = (self._max_slave_restart is not None and
+                           self._failed_nodes_count > self._max_slave_restart)
+        if maximum_reached:
+            if self._max_slave_restart == 0:
+                msg = 'Slave restarting disabled'
+            else:
+                msg = "Maximum crashed slaves reached: %d" % \
+                      self._max_slave_restart
+            self.report_line(msg)
+        else:
+            self.report_line("Replacing crashed slave %s" % node.gateway.id)
+            self._clone_node(node)
+        self._active_nodes.remove(node)
 
     def slave_collectionfinish(self, node, ids):
+        """Slave has finished test collection.
+
+        This adds the collection for this node to the scheduler.  If
+        the scheduler indicates collection is finished (i.e. all
+        initial nodes have submitted their collection), then tells the
+        scheduler to schedule the collected items.  When initiating
+        scheduling the first time it logs which scheduler is in use.
+        """
+        if self.shuttingdown:
+            return
+        # tell session which items were effectively collected otherwise
+        # the master node will finish the session with EXIT_NOTESTSCOLLECTED
+        self._session.testscollected = len(ids)
         self.sched.addnode_collection(node, ids)
         if self.terminal:
-            self.trdist.setstatus(node.gateway.spec, "[%d]" %(len(ids)))
-
+            self.trdist.setstatus(node.gateway.spec, "[%d]" % (len(ids)))
         if self.sched.collection_is_completed:
-            if self.terminal:
+            if self.terminal and not self.sched.haspending():
                 self.trdist.ensure_show_status()
                 self.terminal.write_line("")
-                self.terminal.write_line("scheduling tests via %s" %(
+                self.terminal.write_line("scheduling tests via %s" % (
                     self.sched.__class__.__name__))
-
             self.sched.init_distribute()
 
     def slave_logstart(self, node, nodeid, location):
+        """Emitted when a node calls the pytest_runtest_logstart hook."""
         self.config.hook.pytest_runtest_logstart(
             nodeid=nodeid, location=location)
 
     def slave_testreport(self, node, rep):
-        if not (rep.passed and rep.when != "call"):
-            if rep.when in ("setup", "call"):
-                self.sched.remove_item(node, rep.item_index, rep.duration)
+        """Emitted when a node calls the pytest_runtest_logreport hook.
+
+        If the node indicates it is finished with a test item remove
+        the item from the pending list in the scheduler.
+        """
+        if rep.when == "call" or (rep.when == "setup" and not rep.passed):
+            self.sched.remove_item(node, rep.item_index, rep.duration)
         #self.report_line("testreport %s: %s" %(rep.id, rep.status))
         rep.node = node
         self.config.hook.pytest_runtest_logreport(report=rep)
         self._handlefailures(rep)
 
     def slave_collectreport(self, node, rep):
+        """Emitted when a node calls the pytest_collectreport hook."""
         if rep.failed:
             self._failed_slave_collectreport(node, rep)
 
+    def _clone_node(self, node):
+        """Return new node based on an existing one.
+
+        This is normally for when a node died, this will copy the spec
+        of the existing node and create a new one with a new id.  The
+        new node will have been setup so will start calling the
+        "slave_*" hooks and do work soon.
+        """
+        spec = node.gateway.spec
+        spec.id = None
+        self.nodemanager.group.allocate_id(spec)
+        node = self.nodemanager.setup_node(spec, self.queue.put)
+        self._active_nodes.add(node)
+        return node
+
     def _failed_slave_collectreport(self, node, rep):
         # Check we haven't already seen this report (from
         # another slave).
@@ -349,19 +684,21 @@ class DSession:
     def triggershutdown(self):
         self.log("triggering shutdown")
         self.shuttingdown = True
-        for node in self.sched.node2pending:
+        for node in self.sched.nodes:
             node.shutdown()
 
     def handle_crashitem(self, nodeid, slave):
         # XXX get more reporting info by recording pytest_runtest_logstart?
+        # XXX count no of failures and retry N times
         runner = self.config.pluginmanager.getplugin("runner")
         fspath = nodeid.split("::")[0]
-        msg = "Slave %r crashed while running %r" %(slave.gateway.id, nodeid)
-        rep = runner.TestReport(nodeid, (fspath, None, fspath), (),
-            "failed", msg, "???")
+        msg = "Slave %r crashed while running %r" % (slave.gateway.id, nodeid)
+        rep = runner.TestReport(nodeid, (fspath, None, fspath),
+                                (), "failed", msg, "???")
         rep.node = slave
         self.config.hook.pytest_runtest_logreport(report=rep)
 
+
 class TerminalDistReporter:
     def __init__(self, config):
         self.config = config

xdist/looponfail.py

@@ -11,6 +11,21 @@ import py, pytest
 import sys
 import execnet
 
+def pytest_addoption(parser):
+    group = parser.getgroup("xdist", "distributed and subprocess testing")
+    group._addoption('-f', '--looponfail',
+           action="store_true", dest="looponfail", default=False,
+           help="run tests in subprocess, wait for modified files "
+                "and re-run failing test set until all pass.")
+
+def pytest_cmdline_main(config):
+    
+    if config.getoption("looponfail"):
+        looponfail_main(config)
+        return 2 # looponfail only can get stop with ctrl-C anyway
+
+
+
 def looponfail_main(config):
     remotecontrol = RemoteControl(config)
     rootdirs = config.getini("looponfailroots")
@@ -87,7 +102,7 @@ class RemoteControl(object):
         result = self.runsession()
         failures, reports, collection_failed = result
         if collection_failed:
-            reports = ["Collection failed, keeping previous failure set"]
+            pass # "Collection failed, keeping previous failure set"
         else:
             uniq_failures = []
             for failure in failures:
@@ -109,7 +124,6 @@ def repr_pytest_looponfailinfo(failreports, rootdirs):
 
 def init_slave_session(channel, args, option_dict):
     import os, sys
-    import py
     outchannel = channel.gateway.newchannel()
     sys.stdout = sys.stderr = outchannel.makefile('w')
     channel.send(outchannel)
@@ -228,4 +242,3 @@ class StatRecorder:
             changed = True
         self.statcache = newstat
         return changed
-

xdist/plugin.py

@@ -1,18 +1,26 @@
 import py
 import pytest
 
+
+def parse_numprocesses(s):
+    if s == 'auto':
+        import multiprocessing
+        return multiprocessing.cpu_count()
+    else:
+        return int(s)
+
+
 def pytest_addoption(parser):
     group = parser.getgroup("xdist", "distributed and subprocess testing")
-    group._addoption('-f', '--looponfail',
-           action="store_true", dest="looponfail", default=False,
-           help="run tests in subprocess, wait for modified files "
-                "and re-run failing test set until all pass.")
     group._addoption('-n', dest="numprocesses", metavar="numprocesses",
-           action="store", type="int",
-           help="shortcut for '--dist=load --tx=NUM*popen'")
-    group.addoption('--boxed',
-           action="store_true", dest="boxed", default=False,
-           help="box each test run in a separate process (unix)")
+           action="store",
+           type=parse_numprocesses,
+           help="shortcut for '--dist=load --tx=NUM*popen', "
+                "you can use 'auto' here for auto detection CPUs number on "
+                "host system")
+    group._addoption('--max-slave-restart', action="store", default=None,
+                     help="maximum number of slaves that can be restarted "
+                          "when crashed (set to zero to disable this feature)")
     group._addoption('--dist', metavar="distmode",
            action="store", choices=['load', 'each', 'no'],
            type="choice", dest="dist", default="no",
@@ -45,32 +53,31 @@ def pytest_addoption(parser):
 # -------------------------------------------------------------------------
 def pytest_addhooks(pluginmanager):
     from xdist import newhooks
-    pluginmanager.addhooks(newhooks)
+    # avoid warnings with pytest-2.8
+    method = getattr(pluginmanager, "add_hookspecs", None)
+    if method is None:
+        method = pluginmanager.addhooks
+    method(newhooks)
 
 # -------------------------------------------------------------------------
 # distributed testing initialization
 # -------------------------------------------------------------------------
 
-def pytest_cmdline_main(config):
-    check_options(config)
-    if config.getvalue("looponfail"):
-        from xdist.looponfail import looponfail_main
-        looponfail_main(config)
-        return 2 # looponfail only can get stop with ctrl-C anyway
 
-def pytest_configure(config, __multicall__):
-    __multicall__.execute()
-    if config.getvalue("dist") != "no":
+@pytest.mark.trylast
+def pytest_configure(config):
+    if config.getoption("dist") != "no":
         from xdist.dsession import DSession
         session = DSession(config)
         config.pluginmanager.register(session, "dsession")
         tr = config.pluginmanager.getplugin("terminalreporter")
         tr.showfspath = False
 
-def check_options(config):
+@pytest.mark.tryfirst
+def pytest_cmdline_main(config):
     if config.option.numprocesses:
         config.option.dist = "load"
-        config.option.tx = ['popen'] * int(config.option.numprocesses)
+        config.option.tx = ['popen'] * config.option.numprocesses
     if config.option.distload:
         config.option.dist = "load"
     val = config.getvalue
@@ -82,46 +89,3 @@ def check_options(config):
         elif val("dist") != "no":
             if usepdb:
                 raise pytest.UsageError("--pdb incompatible with distributing tests.")
-
-
-def pytest_runtest_protocol(item):
-    if item.config.getvalue("boxed"):
-        reports = forked_run_report(item)
-        for rep in reports:
-            item.ihook.pytest_runtest_logreport(report=rep)
-        return True
-
-def forked_run_report(item):
-    # for now, we run setup/teardown in the subprocess
-    # XXX optionally allow sharing of setup/teardown
-    from _pytest.runner import runtestprotocol
-    EXITSTATUS_TESTEXIT = 4
-    import marshal
-    from xdist.remote import serialize_report
-    from xdist.slavemanage import unserialize_report
-    def runforked():
-        try:
-            reports = runtestprotocol(item, log=False)
-        except KeyboardInterrupt:
-            py.std.os._exit(EXITSTATUS_TESTEXIT)
-        return marshal.dumps([serialize_report(x) for x in reports])
-
-    ff = py.process.ForkedFunc(runforked)
-    result = ff.waitfinish()
-    if result.retval is not None:
-        report_dumps = marshal.loads(result.retval)
-        return [unserialize_report("testreport", x) for x in report_dumps]
-    else:
-        if result.exitstatus == EXITSTATUS_TESTEXIT:
-            py.test.exit("forked test item %s raised Exit" %(item,))
-        return [report_process_crash(item, result)]
-
-def report_process_crash(item, result):
-    path, lineno = item._getfslineno()
-    info = "%s:%s: running the test CRASHED with signal %d" %(
-            path, lineno, result.signal)
-    from _pytest import runner
-    call = runner.CallInfo(lambda: 0/0, "???")
-    call.excinfo = info
-    rep = runner.pytest_runtest_makereport(item, call)
-    return rep

xdist/remote.py

@@ -51,9 +51,12 @@ class SlaveInteractor:
             elif name == "runtests_all":
                 torun.extend(range(len(session.items)))
             self.log("items to run:", torun)
-            while torun:
+            # only run if we have an item and a next item
+            while len(torun) >= 2:
                 self.run_tests(torun)
             if name == "shutdown":
+                if torun:
+                    self.run_tests(torun)
                 break
         return True
 
@@ -125,6 +128,7 @@ def remote_initconfig(option_dict, args):
 
 
 if __name__ == '__channelexec__':
+    channel = channel  # noqa
     # python3.2 is not concurrent import safe, so let's play it safe
     # https://bitbucket.org/hpk42/pytest/issue/347/pytest-xdist-and-python-32
     if sys.version_info[:2] == (3,2):

xdist/slavemanage.py

@@ -28,34 +28,32 @@ class NodeManager(object):
             self.specs.append(spec)
         self.roots = self._getrsyncdirs()
         self.rsyncoptions = self._getrsyncoptions()
+        self._rsynced_specs = py.builtin.set()
 
-    def rsync_roots(self):
-        """ make sure that all remote gateways
-            have the same set of roots in their
-            current directory.
-        """
+    def rsync_roots(self, gateway):
+        """Rsync the set of roots to the node's gateway cwd."""
         if self.roots:
-            # send each rsync root
             for root in self.roots:
-                self.rsync(root, **self.rsyncoptions)
-
-    def makegateways(self):
-        assert not list(self.group)
-        self.config.hook.pytest_xdist_setupnodes(config=self.config,
-            specs=self.specs)
-        for spec in self.specs:
-            gw = self.group.makegateway(spec)
-            self.config.hook.pytest_xdist_newgateway(gateway=gw)
+                self.rsync(gateway, root, **self.rsyncoptions)
 
     def setup_nodes(self, putevent):
-        self.makegateways()
-        self.rsync_roots()
+        self.config.hook.pytest_xdist_setupnodes(config=self.config,
+                                                 specs=self.specs)
         self.trace("setting up nodes")
-        for gateway in self.group:
-            node = SlaveController(self, gateway, self.config, putevent)
-            gateway.node = node  # to keep node alive
-            node.setup()
-            self.trace("started node %r" % node)
+        nodes = []
+        for spec in self.specs:
+            nodes.append(self.setup_node(spec, putevent))
+        return nodes
+
+    def setup_node(self, spec, putevent):
+        gw = self.group.makegateway(spec)
+        self.config.hook.pytest_xdist_newgateway(gateway=gw)
+        self.rsync_roots(gw)
+        node = SlaveController(self, gw, self.config, putevent)
+        gw.node = node          # keep the node alive
+        node.setup()
+        self.trace("started node %r" % node)
+        return node
 
     def teardown_nodes(self):
         self.group.terminate(self.EXIT_TIMEOUT)
@@ -110,49 +108,43 @@ class NodeManager(object):
             'verbose': self.config.option.verbose,
         }
 
-
-    def rsync(self, source, notify=None, verbose=False, ignores=None):
-        """ perform rsync to all remote hosts.
-        """
+    def rsync(self, gateway, source, notify=None, verbose=False, ignores=None):
+        """Perform rsync to remote hosts for node."""
+        # XXX This changes the calling behaviour of
+        #     pytest_xdist_rsyncstart and pytest_xdist_rsyncfinish to
+        #     be called once per rsync target.
         rsync = HostRSync(source, verbose=verbose, ignores=ignores)
-        seen = py.builtin.set()
-        gateways = []
-        for gateway in self.group:
-            spec = gateway.spec
-            if spec.popen and not spec.chdir:
-                # XXX this assumes that sources are python-packages
-                # and that adding the basedir does not hurt
-                gateway.remote_exec("""
-                    import sys ; sys.path.insert(0, %r)
-                """ % os.path.dirname(str(source))).waitclose()
-                continue
-            if spec not in seen:
-                def finished():
-                    if notify:
-                        notify("rsyncrootready", spec, source)
-                rsync.add_target_host(gateway, finished=finished)
-                seen.add(spec)
-                gateways.append(gateway)
-        if seen:
-            self.config.hook.pytest_xdist_rsyncstart(
-                source=source,
-                gateways=gateways,
-            )
-            rsync.send()
-            self.config.hook.pytest_xdist_rsyncfinish(
-                source=source,
-                gateways=gateways,
-            )
+        spec = gateway.spec
+        if spec.popen and not spec.chdir:
+            # XXX This assumes that sources are python-packages
+            #     and that adding the basedir does not hurt.
+            gateway.remote_exec("""
+                import sys ; sys.path.insert(0, %r)
+            """ % os.path.dirname(str(source))).waitclose()
+            return
+        if (spec, source) in self._rsynced_specs:
+            return
+        def finished():
+            if notify:
+                notify("rsyncrootready", spec, source)
+        rsync.add_target_host(gateway, finished=finished)
+        self._rsynced_specs.add((spec, source))
+        self.config.hook.pytest_xdist_rsyncstart(
+            source=source,
+            gateways=[gateway],
+        )
+        rsync.send()
+        self.config.hook.pytest_xdist_rsyncfinish(
+            source=source,
+            gateways=[gateway],
+        )
 
 class HostRSync(execnet.RSync):
     """ RSyncer that filters out common files
     """
     def __init__(self, sourcedir, *args, **kwargs):
         self._synced = {}
-        ignores= None
-        if 'ignores' in kwargs:
-            ignores = kwargs.pop('ignores')
-        self._ignores = ignores or []
+        self._ignores = kwargs.pop('ignores', None) or []
         super(HostRSync, self).__init__(sourcedir=sourcedir, **kwargs)
 
     def filter(self, path):