Combine no longer appends by default

Combine used to always load an existing .coverage file. This lead to confusing
results and extra tox-clean steps.  Now the default is to not load the existing
file, though a new --append switch on coverage combine gets you that behavior
if you need it.

This also pointed up an issue with concurrency=multiprocessing, which is that
the child processes automatically used parallel=True, but the parent process
did not.  Now concurrency=multiprocessing implies parallel=True.

6 files changed, 28 insertions, 6 deletions

diff --git a/CHANGES.rst b/CHANGES.rst
index a231a69..fcb61a9 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -8,6 +8,14 @@ Change history for Coverage.py
 Unreleased
 ----------
 
+- BACKWARD INCOMPATIBILITY: the ``coverage combine`` command now ignores an
+  existing ``.coverage`` data file.  It used to include that file in its
+  combining.  This caused confusing results, and extra tox "clean" steps.  If
+  you want the old behavior, use the new ``coverage combine --append`` option.
+
+- Using ``--concurrency=multiprocessing`` now implies ``--parallel`` so that
+  the main program is measured similarly to the sub-processes.
+
 - When using `automatic subprocess measurement`_, running coverage commands
   would create spurious data files.  This is now fixed, thanks to diagnosis and
   testing by Dan Riti.  Closes `issue 492`_.
diff --git a/coverage/cmdline.py b/coverage/cmdline.py
index e2b79fe..9ff39f4 100644
--- a/coverage/cmdline.py
+++ b/coverage/cmdline.py
@@ -274,7 +274,9 @@ CMDS = {
 
     'combine': CmdOptionParser(
         "combine",
-        GLOBAL_ARGS,
+        [
+            Opts.append,
+            ] + GLOBAL_ARGS,
         usage="<path1> <path2> ... <pathN>",
         description=(
             "Combine data from multiple coverage files collected "
@@ -484,7 +486,8 @@ class CoverageScript(object):
             return self.do_run(options, args)
 
         elif options.action == "combine":
-            self.coverage.load()
+            if options.append:
+                self.coverage.load()
             data_dirs = args or None
             self.coverage.combine(data_dirs)
             self.coverage.save()
diff --git a/coverage/control.py b/coverage/control.py
index 2f2989c..f350b3d 100644
--- a/coverage/control.py
+++ b/coverage/control.py
@@ -247,6 +247,9 @@ class Coverage(object):
         if concurrency == "multiprocessing":
             patch_multiprocessing()
             concurrency = None
+            # Multi-processing uses parallel for the subprocesses, so also use
+            # it for the main process.
+            self.config.parallel = True
 
         self.collector = Collector(
             should_trace=self._should_trace,
diff --git a/doc/cmd.rst b/doc/cmd.rst
index f77738b..90e2105 100644
--- a/doc/cmd.rst
+++ b/doc/cmd.rst
@@ -212,6 +212,11 @@ suffix.  Here are some examples of data files that can be combined::
     .coverage.20120807T212300
     .coverage.last_good_run.ok
 
+An existing combined data file is ignored and re-written. If you want to use
+**combine** to accumulate results into the .coverage data file over a number of
+runs, use the ``--append`` switch on the **combine** command.  This behavior
+was the default before version 4.2.
+
 The ``run --parallel-mode`` switch automatically creates separate data files
 for each run which can be combined later.  The file names include the machine
 name, the process id, and a random number::
diff --git a/tests/test_cmdline.py b/tests/test_cmdline.py
index 8e2840b..795a01f 100644
--- a/tests/test_cmdline.py
+++ b/tests/test_cmdline.py
@@ -190,6 +190,12 @@ class CmdLineTest(BaseCmdLineTest):
         # coverage combine with args
         self.cmd_executes("combine datadir1", """\
             .coverage()
+            .combine(["datadir1"])
+            .save()
+            """)
+        # coverage combine, appending
+        self.cmd_executes("combine --append datadir1", """\
+            .coverage()
             .load()
             .combine(["datadir1"])
             .save()
@@ -197,7 +203,6 @@ class CmdLineTest(BaseCmdLineTest):
         # coverage combine without args
         self.cmd_executes("combine", """\
             .coverage()
-            .load()
             .combine(None)
             .save()
             """)
@@ -206,13 +211,11 @@ class CmdLineTest(BaseCmdLineTest):
         # https://bitbucket.org/ned/coveragepy/issues/385/coverage-combine-doesnt-work-with-rcfile
         self.cmd_executes("combine --rcfile cov.ini", """\
             .coverage(config_file='cov.ini')
-            .load()
             .combine(None)
             .save()
             """)
         self.cmd_executes("combine --rcfile cov.ini data1 data2/more", """\
             .coverage(config_file='cov.ini')
-            .load()
             .combine(["data1", "data2/more"])
             .save()
             """)
diff --git a/tests/test_process.py b/tests/test_process.py
index 5e0014e..328bd9e 100644
--- a/tests/test_process.py
+++ b/tests/test_process.py
@@ -150,7 +150,7 @@ class ProcessTest(CoverageTest):
         self.assertEqual(self.number_of_data_files(), 2)
 
         # Combine the parallel coverage data files into .coverage .
-        self.run_command("coverage combine")
+        self.run_command("coverage combine --append")
         self.assert_exists(".coverage")
 
         # After combining, there should be only the .coverage file.