make --durations also show the execution times of setup/teardown calls. This requires a slight incompatibility - pytest_runtest_logreport now sees setup/teardown reports even if the tests passed.

CHANGELOG

@@ -1,8 +1,8 @@
-Changes between 2.1.3 and [next version]
+Changes between 2.1.3 and XXX 2.2.0
 ----------------------------------------
 
 - new feature to help optimizing your tests: --durations=N option for 
-  displaying N slowest test calls
+  displaying N slowest test calls and setup/teardown methods.
 - fix and cleanup pytest's own test suite to not leak FDs 
 - fix issue83: link to generated funcarg list
 - fix issue74: pyarg module names are now checked against imp.find_module false positives

_pytest/__init__.py

@@ -1,2 +1,2 @@
 #
-__version__ = '2.2.0.dev1'
+__version__ = '2.2.0.dev2'

_pytest/hookspec.py

@@ -149,7 +149,8 @@ def pytest_runtest_makereport(item, call):
 pytest_runtest_makereport.firstresult = True
 
 def pytest_runtest_logreport(report):
-    """ process item test report. """
+    """ process a test setup/call/teardown report relating to
+    the respective phase of executing a test. """
 
 # special handling for final teardown - somewhat internal for now
 def pytest__teardown_final(session):

_pytest/junitxml.py

@@ -166,6 +166,7 @@ class LogXML(object):
 
     def pytest_runtest_logreport(self, report):
         if report.passed:
+            if report.when == "call": # ignore setup/teardown
                 self.append_pass(report)
         elif report.failed:
             if report.when != "call":

_pytest/pytester.py

@@ -334,8 +334,8 @@ class TmpTestdir:
         l = list(args) + [p]
         reprec = self.inline_run(*l)
         reports = reprec.getreports("pytest_runtest_logreport")
-        assert len(reports) == 1, reports
-        return reports[0]
+        assert len(reports) == 3, reports # setup/call/teardown
+        return reports[1]
 
     def inline_genitems(self, *args):
         return self.inprocess_run(list(args) + ['--collectonly'])

_pytest/resultlog.py

@@ -63,6 +63,8 @@ class ResultLog(object):
         self.write_log_entry(testpath, lettercode, longrepr)
 
     def pytest_runtest_logreport(self, report):
+        if report.when != "call" and report.passed:
+            return
         res = self.config.hook.pytest_report_teststatus(report=report)
         code = res[1]
         if code == 'x':

_pytest/runner.py

@@ -26,25 +26,36 @@ def pytest_terminal_summary(terminalreporter):
         return
     tr = terminalreporter
     duration2rep = {}
+    alldurations = 0.0
     for key, replist in tr.stats.items():
         if key == "deselected":
             continue
         for rep in replist:
             duration2rep[rep.duration] = rep
+            alldurations += rep.duration
     if not duration2rep:
         return
-    d2 = duration2rep.items()
+    d2 = remaining = duration2rep.items()
     d2.sort()
     d2.reverse()
     if not durations:
         tr.write_sep("=", "slowest test durations")
     else:
         tr.write_sep("=", "slowest %s test durations" % durations)
+        remaining = d2[durations:]
         d2 = d2[:durations]
 
+    assert (alldurations/100) > 0
     for duration, rep in d2:
         nodeid = rep.nodeid.replace("::()::", "::")
-        tr.write_line("%2.2f %s %s" % (duration, rep.when, nodeid))
+        percent = rep.duration / (alldurations  / 100)
+        tr.write_line("%02.2fs %-02.2f%% %s %s" %
+            (duration, percent, rep.when, nodeid))
+    if remaining:
+        remsum = sum(map(lambda x: x[0], remaining))
+        tr.write_line("%02.2fs %-02.2f%% remaining in %d test phases" %(
+            remsum, remsum / (alldurations  / 100), len(remaining)))
+
 
 def pytest_sessionstart(session):
     session._setupstate = SetupState()
@@ -110,7 +121,7 @@ def call_and_report(item, when, log=True):
     call = call_runtest_hook(item, when)
     hook = item.ihook
     report = hook.pytest_runtest_makereport(item=item, call=call)
-    if log and (when == "call" or not report.passed):
+    if log:
         hook.pytest_runtest_logreport(report=report)
     return report
 

_pytest/terminal.py

@@ -426,6 +426,7 @@ class TerminalReporter:
                 keys.append(key)
         parts = []
         for key in keys:
+            if key: # setup/teardown reports have an empty key, ignore them
                 val = self.stats.get(key, None)
                 if val:
                     parts.append("%d %s" %(len(val), key))

doc/announce/release-2.2.0.txt

@@ -1,7 +1,7 @@
 py.test 2.2.0: new test duration profiling and bug fixes
 ===========================================================================
 
-pytest-2.2.0 is a backward compatible release of the popular
+pytest-2.2.0 is a quite backward compatible release of the popular
 py.test testing tool.  It introduces the new "--duration=N" option
 showing the N slowest test execution or setup/teardown calls.   The
 release also contains a few fixes and some cleanup of pytest's own test
@@ -16,6 +16,19 @@ If you want to install or upgrade pytest you might just type::
     pip install -U pytest # or
     easy_install -U pytest
 
+incompatible change
+------------------------------------
+
+* You need a new version of the pytest-xdist plugin (1.7) for distributing 
+  test runs.  
+
+* Also other plugins might need an upgrade if they implement
+  the ``pytest_runtest_logreport`` hook which now is called unconditionally
+  for the setup/teardown fixture phases of a test. You can just choose to
+  ignore them by inserting "if rep.when != 'call': return". Note that
+  most code probably "just" works because the hook was already called
+  for failing setup/teardown phases of a test.
+
 best,
 holger krekel
 

setup.py

@@ -24,7 +24,7 @@ def main():
         name='pytest',
         description='py.test: simple powerful testing with Python',
         long_description = long_description,
-        version='2.2.0.dev1',
+        version='2.2.0.dev2',
         url='http://pytest.org',
         license='MIT license',
         platforms=['unix', 'linux', 'osx', 'cygwin', 'win32'],

testing/acceptance_test.py

@@ -484,12 +484,15 @@ def test_duration_test(testdir):
         "*call*test_2*",
         "*call*test_1*",
     ])
+    assert "remaining in"  not in result.stdout.str()
+
     result = testdir.runpytest("--durations=2")
     assert result.ret == 0
     result.stdout.fnmatch_lines([
         "*durations*",
         "*call*test_3*",
         "*call*test_2*",
+        "*s*%*remaining in 7 test phases",
     ])
     assert "test_1" not in result.stdout.str()
     result = testdir.runpytest("--durations=0")
@@ -503,6 +506,28 @@ def test_duration_test(testdir):
             else:
                 raise AssertionError("not found %s %s" % (x,y))
 
+def test_duration_test_with_fixture(testdir):
+    testdir.makepyfile("""
+        import time
+        frag = 0.01
+        def setup_function(func):
+            time.sleep(frag * 3)
+
+        def test_1():
+            time.sleep(frag*2)
+        def test_2():
+            time.sleep(frag)
+    """)
+    result = testdir.runpytest("--durations=10")
+    assert result.ret == 0
+    result.stdout.fnmatch_lines([
+        "*durations*",
+        "*setup*test_1*",
+        "*setup*test_2*",
+        "*call*test_1*",
+        "*call*test_2*",
+    ])
+
     result = testdir.runpytest("--durations=2", "-k test_1")
     assert result.ret == 0
     result.stdout.fnmatch_lines([

testing/test_mark.py

@@ -77,7 +77,7 @@ class TestFunctional:
                 assert hasattr(test_hello, 'hello')
         """)
         result = testdir.runpytest(p)
-        result.stdout.fnmatch_lines(["*passed*"])
+        result.stdout.fnmatch_lines(["*1 passed*"])
 
     def test_mark_per_module(self, testdir):
         item = testdir.getitem("""