Made run_cmd return out and err

7 files changed, 401 insertions, 459 deletions

diff --git a/tests/conftest.py b/tests/conftest.py
index f41afb64..56538718 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -5,6 +5,7 @@ Cmd2 unit/functional testing
 Copyright 2016 Federico Ceratto <federico.ceratto@gmail.com>
 Released under MIT license, see LICENSE file
 """
+import sys
 from typing import Optional
 from unittest import mock
 
@@ -13,6 +14,13 @@ from pytest import fixture
 import cmd2
 from cmd2.utils import StdSim
 
+# Python 3.4 require contextlib2 for temporarily redirecting stderr and stdout
+if sys.version_info < (3, 5):
+    # noinspection PyUnresolvedReferences
+    from contextlib2 import redirect_stdout, redirect_stderr
+else:
+    from contextlib import redirect_stdout, redirect_stderr
+
 # Prefer statically linked gnureadline if available (for macOS compatibility due to issues with libedit)
 try:
     import gnureadline as readline
@@ -126,19 +134,33 @@ def normalize(block):
 
 
 def run_cmd(app, cmd):
-    """ Clear StdSim buffer, run the command, extract the buffer contents, """
-    app.stdout.clear()
-    app.onecmd_plus_hooks(cmd)
-    out = app.stdout.getvalue()
-    app.stdout.clear()
-    return normalize(out)
+    """ Clear out and err StdSim buffers, run the command, and return out and err """
+    saved_sysout = sys.stdout
+    sys.stdout = app.stdout
+
+    # This will be used to capture app.stdout and sys.stdout
+    copy_cmd_stdout = StdSim(app.stdout)
+
+    # This will be used to capture sys.stderr
+    copy_stderr = StdSim(sys.stderr)
+
+    try:
+        app.stdout = copy_cmd_stdout
+        with redirect_stdout(copy_cmd_stdout):
+            with redirect_stderr(copy_stderr):
+                app.onecmd_plus_hooks(cmd)
+    finally:
+        app.stdout = copy_cmd_stdout.inner_stream
+        sys.stdout = saved_sysout
+
+    out = copy_cmd_stdout.getvalue()
+    err = copy_stderr.getvalue()
+    return normalize(out), normalize(err)
 
 
 @fixture
 def base_app():
-    c = cmd2.Cmd()
-    c.stdout = StdSim(c.stdout)
-    return c
+    return cmd2.Cmd()
 
 
 def complete_tester(text: str, line: str, begidx: int, endidx: int, app) -> Optional[str]:
diff --git a/tests/test_argparse.py b/tests/test_argparse.py
index 74a03e28..d716c68d 100644
--- a/tests/test_argparse.py
+++ b/tests/test_argparse.py
@@ -112,62 +112,60 @@ class ArgparseApp(cmd2.Cmd):
 @pytest.fixture
 def argparse_app():
     app = ArgparseApp()
-    app.stdout = StdSim(app.stdout)
     return app
 
 
-def test_invalid_syntax(argparse_app, capsys):
-    run_cmd(argparse_app, 'speak "')
-    out, err = capsys.readouterr()
-    assert err == "Invalid syntax: No closing quotation\n"
+def test_invalid_syntax(argparse_app):
+    out, err = run_cmd(argparse_app, 'speak "')
+    assert err[0] == "Invalid syntax: No closing quotation"
 
 def test_argparse_basic_command(argparse_app):
-    out = run_cmd(argparse_app, 'say hello')
+    out, err = run_cmd(argparse_app, 'say hello')
     assert out == ['hello']
 
 def test_argparse_quoted_arguments(argparse_app):
-    out = run_cmd(argparse_app, 'say "hello there"')
+    out, err = run_cmd(argparse_app, 'say "hello there"')
     assert out == ['hello there']
 
 def test_argparse_with_list(argparse_app):
-    out = run_cmd(argparse_app, 'speak -s hello world!')
+    out, err = run_cmd(argparse_app, 'speak -s hello world!')
     assert out == ['HELLO WORLD!']
 
 def test_argparse_with_list_and_empty_doc(argparse_app):
-    out = run_cmd(argparse_app, 'speak -s hello world!')
+    out, err = run_cmd(argparse_app, 'speak -s hello world!')
     assert out == ['HELLO WORLD!']
 
 def test_argparser_correct_args_with_quotes_and_midline_options(argparse_app):
-    out = run_cmd(argparse_app, "speak 'This  is a' -s test of the emergency broadcast system!")
+    out, err = run_cmd(argparse_app, "speak 'This  is a' -s test of the emergency broadcast system!")
     assert out == ['THIS  IS A TEST OF THE EMERGENCY BROADCAST SYSTEM!']
 
 def test_argparse_quoted_arguments_multiple(argparse_app):
-    out = run_cmd(argparse_app, 'say "hello  there" "rick & morty"')
+    out, err = run_cmd(argparse_app, 'say "hello  there" "rick & morty"')
     assert out == ['hello  there rick & morty']
 
 def test_argparse_help_docstring(argparse_app):
-    out = run_cmd(argparse_app, 'help say')
+    out, err = run_cmd(argparse_app, 'help say')
     assert out[0].startswith('usage: say')
     assert out[1] == ''
     assert out[2] == 'Repeat what you tell me to.'
 
 def test_argparse_help_description(argparse_app):
-    out = run_cmd(argparse_app, 'help tag')
+    out, err = run_cmd(argparse_app, 'help tag')
     assert out[0].startswith('usage: tag')
     assert out[1] == ''
     assert out[2] == 'create a html tag'
 
 def test_argparse_prog(argparse_app):
-    out = run_cmd(argparse_app, 'help tag')
+    out, err = run_cmd(argparse_app, 'help tag')
     progname = out[0].split(' ')[1]
     assert progname == 'tag'
 
 def test_arglist(argparse_app):
-    out = run_cmd(argparse_app, 'arglist "we  should" get these')
+    out, err = run_cmd(argparse_app, 'arglist "we  should" get these')
     assert out[0] == 'True'
 
 def test_preservelist(argparse_app):
-    out = run_cmd(argparse_app, 'preservelist foo "bar baz"')
+    out, err = run_cmd(argparse_app, 'preservelist foo "bar baz"')
     assert out[0] == "['foo', '\"bar baz\"']"
 
 
@@ -215,39 +213,36 @@ class SubcommandApp(cmd2.Cmd):
 @pytest.fixture
 def subcommand_app():
     app = SubcommandApp()
-    app.stdout = StdSim(app.stdout)
     return app
 
 
 def test_subcommand_foo(subcommand_app):
-    out = run_cmd(subcommand_app, 'base foo -x2 5.0')
+    out, err = run_cmd(subcommand_app, 'base foo -x2 5.0')
     assert out == ['10.0']
 
 
 def test_subcommand_bar(subcommand_app):
-    out = run_cmd(subcommand_app, 'base bar baz')
+    out, err = run_cmd(subcommand_app, 'base bar baz')
     assert out == ['((baz))']
 
-def test_subcommand_invalid(subcommand_app, capsys):
-    run_cmd(subcommand_app, 'base baz')
-    out, err = capsys.readouterr()
-    err = err.splitlines()
+def test_subcommand_invalid(subcommand_app):
+    out, err = run_cmd(subcommand_app, 'base baz')
     assert err[0].startswith('usage: base')
     assert err[1].startswith("base: error: invalid choice: 'baz'")
 
 def test_subcommand_base_help(subcommand_app):
-    out = run_cmd(subcommand_app, 'help base')
+    out, err = run_cmd(subcommand_app, 'help base')
     assert out[0].startswith('usage: base')
     assert out[1] == ''
     assert out[2] == 'Base command help'
 
 def test_subcommand_help(subcommand_app):
-    out = run_cmd(subcommand_app, 'help base foo')
+    out, err = run_cmd(subcommand_app, 'help base foo')
     assert out[0].startswith('usage: base foo')
     assert out[1] == ''
     assert out[2] == 'positional arguments:'
 
 
 def test_subcommand_invalid_help(subcommand_app):
-    out = run_cmd(subcommand_app, 'help base baz')
+    out, err = run_cmd(subcommand_app, 'help base baz')
     assert out[0].startswith('usage: base')
diff --git a/tests/test_autocompletion.py b/tests/test_autocompletion.py
index 2b9e0458..8aeac6b1 100644
--- a/tests/test_autocompletion.py
+++ b/tests/test_autocompletion.py
@@ -52,12 +52,9 @@ required arguments:
 optional arguments:
   -h, --help            show this help message and exit'''
 
-def test_help_required_group(cmd2_app, capsys):
-    run_cmd(cmd2_app, 'suggest -h')
-    out, err = capsys.readouterr()
-    out1 = normalize(str(out))
-
-    out2 = run_cmd(cmd2_app, 'help suggest')
+def test_help_required_group(cmd2_app):
+    out1, err1 = run_cmd(cmd2_app, 'suggest -h')
+    out2, err2 = run_cmd(cmd2_app, 'help suggest')
 
     assert out1 == out2
     assert out1[0].startswith('Usage: suggest')
@@ -66,12 +63,9 @@ def test_help_required_group(cmd2_app, capsys):
     assert out1 == normalize(SUGGEST_HELP)
 
 
-def test_help_required_group_long(cmd2_app, capsys):
-    run_cmd(cmd2_app, 'media movies add -h')
-    out, err = capsys.readouterr()
-    out1 = normalize(str(out))
-
-    out2 = run_cmd(cmd2_app, 'help media movies add')
+def test_help_required_group_long(cmd2_app):
+    out1, err1 = run_cmd(cmd2_app, 'media movies add -h')
+    out2, err2 = run_cmd(cmd2_app, 'help media movies add')
 
     assert out1 == out2
     assert out1[0].startswith('Usage: media movies add')
@@ -156,11 +150,9 @@ def test_autocomp_flags_narg_max(cmd2_app):
     assert first_match is None
 
 
-def test_autcomp_narg_beyond_max(cmd2_app, capsys):
-    run_cmd(cmd2_app, 'suggest -t movie -d 3 4 5')
-    out, err = capsys.readouterr()
-
-    assert 'Error: unrecognized arguments: 5' in err
+def test_autcomp_narg_beyond_max(cmd2_app):
+    out, err = run_cmd(cmd2_app, 'suggest -t movie -d 3 4 5')
+    assert 'Error: unrecognized arguments: 5' in err[0]
 
 
 def test_autocomp_subcmd_nested(cmd2_app):
diff --git a/tests/test_cmd2.py b/tests/test_cmd2.py
index 6d2daa89..acb9d0f7 100644
--- a/tests/test_cmd2.py
+++ b/tests/test_cmd2.py
@@ -29,21 +29,27 @@ from .conftest import run_cmd, normalize, BASE_HELP, BASE_HELP_VERBOSE, \
     HELP_HISTORY, SHORTCUTS_TXT, SHOW_TXT, SHOW_LONG
 
 
+@pytest.fixture
+def outsim_app():
+    c = cmd2.Cmd()
+    c.stdout = utils.StdSim(c.stdout)
+    return c
+
 def test_version(base_app):
     assert cmd2.__version__
 
 def test_empty_statement(base_app):
-    out = run_cmd(base_app, '')
+    out, err = run_cmd(base_app, '')
     expected = normalize('')
     assert out == expected
 
 def test_base_help(base_app):
-    out = run_cmd(base_app, 'help')
+    out, err = run_cmd(base_app, 'help')
     expected = normalize(BASE_HELP)
     assert out == expected
 
 def test_base_help_verbose(base_app):
-    out = run_cmd(base_app, 'help -v')
+    out, err = run_cmd(base_app, 'help -v')
     expected = normalize(BASE_HELP_VERBOSE)
     assert out == expected
 
@@ -52,32 +58,26 @@ def test_base_help_verbose(base_app):
     help_doc += "\n:param fake param"
     base_app.do_help.__func__.__doc__ = help_doc
 
-    out = run_cmd(base_app, 'help --verbose')
+    out, err = run_cmd(base_app, 'help --verbose')
     assert out == expected
 
-def test_base_argparse_help(base_app, capsys):
+def test_base_argparse_help(base_app):
     # Verify that "set -h" gives the same output as "help set" and that it starts in a way that makes sense
-    run_cmd(base_app, 'set -h')
-    out, err = capsys.readouterr()
-    out1 = normalize(str(out))
-
-    out2 = run_cmd(base_app, 'help set')
+    out1, err1 = run_cmd(base_app, 'set -h')
+    out2, err2 = run_cmd(base_app, 'help set')
 
     assert out1 == out2
     assert out1[0].startswith('Usage: set')
     assert out1[1] == ''
     assert out1[2].startswith('Set a settable parameter')
 
-def test_base_invalid_option(base_app, capsys):
-    run_cmd(base_app, 'set -z')
-    out, err = capsys.readouterr()
-    out = normalize(out)
-    err = normalize(err)
+def test_base_invalid_option(base_app):
+    out, err = run_cmd(base_app, 'set -z')
     assert 'Error: unrecognized arguments: -z' in err[0]
     assert out[0] == 'Usage: set [-h] [-a] [-l] [param] [value]'
 
 def test_base_shortcuts(base_app):
-    out = run_cmd(base_app, 'shortcuts')
+    out, err = run_cmd(base_app, 'shortcuts')
     expected = normalize(SHORTCUTS_TXT)
     assert out == expected
 
@@ -85,7 +85,7 @@ def test_base_shortcuts(base_app):
 def test_base_show(base_app):
     # force editor to be 'vim' so test is repeatable across platforms
     base_app.editor = 'vim'
-    out = run_cmd(base_app, 'set')
+    out, err = run_cmd(base_app, 'set')
     expected = normalize(SHOW_TXT)
     assert out == expected
 
@@ -93,14 +93,14 @@ def test_base_show(base_app):
 def test_base_show_long(base_app):
     # force editor to be 'vim' so test is repeatable across platforms
     base_app.editor = 'vim'
-    out = run_cmd(base_app, 'set -l')
+    out, err = run_cmd(base_app, 'set -l')
     expected = normalize(SHOW_LONG)
     assert out == expected
 
 
 def test_base_show_readonly(base_app):
     base_app.editor = 'vim'
-    out = run_cmd(base_app, 'set -a')
+    out, err = run_cmd(base_app, 'set -a')
     expected = normalize(SHOW_TXT + '\nRead only settings:' + """
         Commands may be terminated with: {}
         Arguments at invocation allowed: {}
@@ -149,33 +149,32 @@ def test_cast_problems(capsys):
 
 
 def test_base_set(base_app):
-    out = run_cmd(base_app, 'set quiet True')
+    out, err = run_cmd(base_app, 'set quiet True')
     expected = normalize("""
 quiet - was: False
 now: True
 """)
     assert out == expected
 
-    out = run_cmd(base_app, 'set quiet')
+    out, err = run_cmd(base_app, 'set quiet')
     assert out == ['quiet: True']
 
-def test_set_not_supported(base_app, capsys):
-    run_cmd(base_app, 'set qqq True')
-    out, err = capsys.readouterr()
+def test_set_not_supported(base_app):
+    out, err = run_cmd(base_app, 'set qqq True')
     expected = normalize("""
 Parameter 'qqq' not supported (type 'set' for list of parameters).
 """)
-    assert normalize(str(err)) == expected
+    assert err == expected
 
 def test_set_quiet(base_app):
-    out = run_cmd(base_app, 'set quie True')
+    out, err = run_cmd(base_app, 'set quie True')
     expected = normalize("""
 quiet - was: False
 now: True
 """)
     assert out == expected
 
-    out = run_cmd(base_app, 'set quiet')
+    out, err = run_cmd(base_app, 'set quiet')
     assert out == ['quiet: True']
 
 
@@ -190,11 +189,10 @@ class OnChangeHookApp(cmd2.Cmd):
 @pytest.fixture
 def onchange_app():
     app = OnChangeHookApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_set_onchange_hook(onchange_app):
-    out = run_cmd(onchange_app, 'set quiet True')
+    out, err = run_cmd(onchange_app, 'set quiet True')
     expected = normalize("""
 quiet - was: False
 now: True
@@ -206,61 +204,55 @@ You changed quiet
 def test_base_shell(base_app, monkeypatch):
     m = mock.Mock()
     monkeypatch.setattr("{}.Popen".format('subprocess'), m)
-    out = run_cmd(base_app, 'shell echo a')
+    out, err = run_cmd(base_app, 'shell echo a')
     assert out == []
     assert m.called
 
-def test_base_py(base_app, capsys):
+def test_base_py(base_app):
     # Create a variable and make sure we can see it
-    run_cmd(base_app, 'py qqq=3')
-    out, err = capsys.readouterr()
-    assert out == ''
-    run_cmd(base_app, 'py print(qqq)')
-    out, err = capsys.readouterr()
-    assert out.rstrip() == '3'
+    out, err = run_cmd(base_app, 'py qqq=3')
+    assert not out
+
+    out, err = run_cmd(base_app, 'py print(qqq)')
+    assert out[0].rstrip() == '3'
 
     # Add a more complex statement
-    run_cmd(base_app, 'py print("spaces" + " in this " + "command")')
-    out, err = capsys.readouterr()
-    assert out.rstrip() == 'spaces in this command'
+    out, err = run_cmd(base_app, 'py print("spaces" + " in this " + "command")')
+    assert out[0].rstrip() == 'spaces in this command'
 
     # Set locals_in_py to True and make sure we see self
-    out = run_cmd(base_app, 'set locals_in_py True')
+    out, err = run_cmd(base_app, 'set locals_in_py True')
     assert 'now: True' in out
 
-    run_cmd(base_app, 'py print(self)')
-    out, err = capsys.readouterr()
-    assert 'cmd2.cmd2.Cmd object' in out
+    out, err = run_cmd(base_app, 'py print(self)')
+    assert 'cmd2.cmd2.Cmd object' in out[0]
 
     # Set locals_in_py to False and make sure we can't see self
-    out = run_cmd(base_app, 'set locals_in_py False')
+    out, err = run_cmd(base_app, 'set locals_in_py False')
     assert 'now: False' in out
 
-    run_cmd(base_app, 'py print(self)')
-    out, err = capsys.readouterr()
-    assert "name 'self' is not defined" in err
+    out, err = run_cmd(base_app, 'py print(self)')
+    assert "NameError: name 'self' is not defined" in err
 
 
 @pytest.mark.skipif(sys.platform == 'win32',
                     reason="Unit test doesn't work on win32, but feature does")
-def test_base_run_python_script(base_app, capsys, request):
+def test_base_run_python_script(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     python_script = os.path.join(test_dir, 'script.py')
-    expected = 'This is a python script running ...\n'
+    expected = 'This is a python script running ...'
 
-    run_cmd(base_app, "py run('{}')".format(python_script))
-    out, err = capsys.readouterr()
-    assert out == expected
+    out, err = run_cmd(base_app, "py run('{}')".format(python_script))
+    assert expected in out
 
 
 def test_base_run_pyscript(base_app, capsys, request):
     test_dir = os.path.dirname(request.module.__file__)
     python_script = os.path.join(test_dir, 'script.py')
-    expected = 'This is a python script running ...\n'
+    expected = 'This is a python script running ...'
 
-    run_cmd(base_app, "pyscript {}".format(python_script))
-    out, err = capsys.readouterr()
-    assert out == expected
+    out, err = run_cmd(base_app, "pyscript {}".format(python_script))
+    assert expected in out
 
 def test_recursive_pyscript_not_allowed(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
@@ -271,30 +263,26 @@ def test_recursive_pyscript_not_allowed(base_app, request):
     err = base_app._last_result.stderr
     assert err == expected
 
-def test_pyscript_with_nonexist_file(base_app, capsys):
+def test_pyscript_with_nonexist_file(base_app):
     python_script = 'does_not_exist.py'
-    run_cmd(base_app, "pyscript {}".format(python_script))
-    out, err = capsys.readouterr()
-    assert "Error opening script file" in err
+    out, err = run_cmd(base_app, "pyscript {}".format(python_script))
+    assert "Error opening script file" in err[0]
 
-def test_pyscript_with_exception(base_app, capsys, request):
+def test_pyscript_with_exception(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     python_script = os.path.join(test_dir, 'scripts', 'raises_exception.py')
-    run_cmd(base_app, "pyscript {}".format(python_script))
-    out, err = capsys.readouterr()
-    assert err.startswith('Traceback')
-    assert err.endswith("TypeError: unsupported operand type(s) for +: 'int' and 'str'\n")
+    out, err = run_cmd(base_app, "pyscript {}".format(python_script))
+    assert err[0].startswith('Traceback')
+    assert "TypeError: unsupported operand type(s) for +: 'int' and 'str'" in err[-1]
 
-def test_pyscript_requires_an_argument(base_app, capsys):
-    run_cmd(base_app, "pyscript")
-    out, err = capsys.readouterr()
-    assert "the following arguments are required: script_path" in err
+def test_pyscript_requires_an_argument(base_app):
+    out, err = run_cmd(base_app, "pyscript")
+    assert "the following arguments are required: script_path" in err[0]
 
 
-def test_base_error(base_app, capsys):
-    run_cmd(base_app, 'meow')
-    out, err = capsys.readouterr()
-    assert "is not a recognized command" in err
+def test_base_error(base_app):
+    out, err = run_cmd(base_app, 'meow')
+    assert "is not a recognized command" in err[0]
 
 
 def test_base_load(base_app, request):
@@ -313,62 +301,57 @@ def test_base_load(base_app, request):
     assert base_app._script_dir == [sdir]
     assert base_app._current_script_dir == sdir
 
-def test_load_with_empty_args(base_app, capsys):
+def test_load_with_empty_args(base_app):
     # The way the load command works, we can't directly capture its stdout or stderr
-    run_cmd(base_app, 'load')
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'load')
 
     # The load command requires a file path argument, so we should get an error message
-    assert "the following arguments are required" in str(err)
+    assert "the following arguments are required" in err[0]
     assert base_app.cmdqueue == []
 
 
 def test_load_with_nonexistent_file(base_app, capsys):
     # The way the load command works, we can't directly capture its stdout or stderr
-    run_cmd(base_app, 'load does_not_exist.txt')
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'load does_not_exist.txt')
 
     # The load command requires a path to an existing file
-    assert "does not exist" in str(err)
+    assert "does not exist" in err[0]
     assert base_app.cmdqueue == []
 
-def test_load_with_directory(base_app, capsys, request):
+def test_load_with_directory(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
 
     # The way the load command works, we can't directly capture its stdout or stderr
-    run_cmd(base_app, 'load {}'.format(test_dir))
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'load {}'.format(test_dir))
 
-    assert "is not a file" in str(err)
+    assert "is not a file" in err[0]
     assert base_app.cmdqueue == []
 
-def test_load_with_empty_file(base_app, capsys, request):
+def test_load_with_empty_file(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     filename = os.path.join(test_dir, 'scripts', 'empty.txt')
 
     # The way the load command works, we can't directly capture its stdout or stderr
-    run_cmd(base_app, 'load {}'.format(filename))
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'load {}'.format(filename))
 
     # The load command requires non-empty script files
-    assert "is empty" in str(err)
+    assert "is empty" in err[0]
     assert base_app.cmdqueue == []
 
 
-def test_load_with_binary_file(base_app, capsys, request):
+def test_load_with_binary_file(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     filename = os.path.join(test_dir, 'scripts', 'binary.bin')
 
     # The way the load command works, we can't directly capture its stdout or stderr
-    run_cmd(base_app, 'load {}'.format(filename))
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'load {}'.format(filename))
 
     # The load command requires non-empty scripts files
-    assert "is not an ASCII or UTF-8 encoded text file" in str(err)
+    assert "is not an ASCII or UTF-8 encoded text file" in err[0]
     assert base_app.cmdqueue == []
 
 
-def test_load_with_utf8_file(base_app, capsys, request):
+def test_load_with_utf8_file(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     filename = os.path.join(test_dir, 'scripts', 'utf8.txt')
 
@@ -411,7 +394,8 @@ help
 shortcuts
 _relative_load postcmds.txt
 set colors Never""" % initial_load
-    assert run_cmd(base_app, 'history -s') == normalize(expected)
+    out, err = run_cmd(base_app, 'history -s')
+    assert out == normalize(expected)
 
 
 def test_base_runcmds_plus_hooks(base_app, request):
@@ -433,7 +417,9 @@ help
 shortcuts
 load %s
 set colors Never""" % (prefilepath, postfilepath)
-    assert run_cmd(base_app, 'history -s') == normalize(expected)
+
+    out, err = run_cmd(base_app, 'history -s')
+    assert out == normalize(expected)
 
 
 def test_base_relative_load(base_app, request):
@@ -452,10 +438,9 @@ def test_base_relative_load(base_app, request):
     assert base_app._script_dir == [sdir]
     assert base_app._current_script_dir == sdir
 
-def test_relative_load_requires_an_argument(base_app, capsys):
-    run_cmd(base_app, '_relative_load')
-    out, err = capsys.readouterr()
-    assert 'Error: the following arguments' in err
+def test_relative_load_requires_an_argument(base_app):
+    out, err = run_cmd(base_app, '_relative_load')
+    assert 'Error: the following arguments' in err[0]
     assert base_app.cmdqueue == []
 
 
@@ -538,19 +523,19 @@ def test_feedback_to_output_true(base_app):
         os.remove(filename)
 
 
-def test_feedback_to_output_false(base_app, capsys):
+def test_feedback_to_output_false(base_app):
     base_app.feedback_to_output = False
     base_app.timing = True
     f, filename = tempfile.mkstemp(prefix='feedback_to_output', suffix='.txt')
     os.close(f)
 
     try:
-        run_cmd(base_app, 'help > {}'.format(filename))
-        out, err = capsys.readouterr()
+        out, err = run_cmd(base_app, 'help > {}'.format(filename))
+
         with open(filename) as f:
             content = f.readlines()
         assert not content[-1].startswith('Elapsed: ')
-        assert err.startswith('Elapsed')
+        assert err[0].startswith('Elapsed')
     except:
         raise
     finally:
@@ -564,7 +549,7 @@ def test_disallow_redirection(base_app):
     filename = 'test_allow_redirect.txt'
 
     # Verify output wasn't redirected
-    out = run_cmd(base_app, 'help > {}'.format(filename))
+    out, err = run_cmd(base_app, 'help > {}'.format(filename))
     expected = normalize(BASE_HELP)
     assert out == expected
 
@@ -580,16 +565,14 @@ def test_pipe_to_shell(base_app):
         # Get help on help and pipe it's output to the input of the word count shell command
         command = 'help help | wc'
 
-    sys.stderr = utils.StdSim(sys.stderr)
-    out = run_cmd(base_app, command)
-    assert out and not sys.stderr.getvalue()
+    out, err = run_cmd(base_app, command)
+    assert out and not err
 
 def test_pipe_to_shell_error(base_app):
     # Try to pipe command output to a shell command that doesn't exist in order to produce an error
-    sys.stderr = utils.StdSim(sys.stderr)
-    out = run_cmd(base_app, 'help | foobarbaz.this_does_not_exist')
+    out, err = run_cmd(base_app, 'help | foobarbaz.this_does_not_exist')
     assert not out
-    assert "No such file or directory" in sys.stderr.getvalue()
+    assert "No such file or directory" in err[0]
 
 @pytest.mark.skipif(not clipboard.can_clip,
                     reason="Pyperclip could not find a copy/paste mechanism for your system")
@@ -605,18 +588,18 @@ def test_send_to_paste_buffer(base_app):
     assert normalize(cmd2.cmd2.get_paste_buffer()) == expected
 
 
-def test_base_timing(base_app, capsys):
+def test_base_timing(base_app):
     base_app.feedback_to_output = False
-    out = run_cmd(base_app, 'set timing True')
+    out, err = run_cmd(base_app, 'set timing True')
     expected = normalize("""timing - was: False
 now: True
 """)
     assert out == expected
-    out, err = capsys.readouterr()
+
     if sys.platform == 'win32':
-        assert err.startswith('Elapsed: 0:00:00')
+        assert err[0].startswith('Elapsed: 0:00:00')
     else:
-        assert err.startswith('Elapsed: 0:00:00.0')
+        assert err[0].startswith('Elapsed: 0:00:00.0')
 
 
 def _expected_no_editor_error():
@@ -632,19 +615,18 @@ To enable full traceback, run the following command:  'set debug true'
 
     return expected_text
 
-def test_base_debug(base_app, capsys):
+def test_base_debug(base_app):
     # Purposely set the editor to None
     base_app.editor = None
 
     # Make sure we get an exception, but cmd2 handles it
-    run_cmd(base_app, 'edit')
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'edit')
 
     expected = _expected_no_editor_error()
-    assert normalize(str(err)) == expected
+    assert err == expected
 
     # Set debug true
-    out = run_cmd(base_app, 'set debug True')
+    out, err = run_cmd(base_app, 'set debug True')
     expected = normalize("""
 debug - was: False
 now: True
@@ -652,9 +634,8 @@ now: True
     assert out == expected
 
     # Verify that we now see the exception traceback
-    run_cmd(base_app, 'edit')
-    out, err = capsys.readouterr()
-    assert str(err).startswith('Traceback (most recent call last):')
+    out, err = run_cmd(base_app, 'edit')
+    assert err[0].startswith('Traceback (most recent call last):')
 
 def test_edit_file(base_app, request, monkeypatch):
     # Set a fake editor just to make sure we have one.  We aren't really going to call it due to the mock
@@ -713,30 +694,26 @@ def test_base_py_interactive(base_app):
     m.assert_called_once()
 
 
-def test_base_cmdloop_with_queue():
+def test_base_cmdloop_with_queue(outsim_app):
     # Create a cmd2.Cmd() instance and make sure basic settings are like we want for test
-    app = cmd2.Cmd()
-    app.use_rawinput = True
+    outsim_app.use_rawinput = True
     intro = 'Hello World, this is an intro ...'
-    app.cmdqueue.append('quit\n')
-    app.stdout = utils.StdSim(app.stdout)
+    outsim_app.cmdqueue.append('quit\n')
 
     # Need to patch sys.argv so cmd2 doesn't think it was called with arguments equal to the py.test args
     testargs = ["prog"]
     expected = intro + '\n'
     with mock.patch.object(sys, 'argv', testargs):
         # Run the command loop with custom intro
-        app.cmdloop(intro=intro)
-    out = app.stdout.getvalue()
+        outsim_app.cmdloop(intro=intro)
+    out = outsim_app.stdout.getvalue()
     assert out == expected
 
 
-def test_base_cmdloop_without_queue():
+def test_base_cmdloop_without_queue(outsim_app):
     # Create a cmd2.Cmd() instance and make sure basic settings are like we want for test
-    app = cmd2.Cmd()
-    app.use_rawinput = True
-    app.intro = 'Hello World, this is an intro ...'
-    app.stdout = utils.StdSim(app.stdout)
+    outsim_app.use_rawinput = True
+    outsim_app.intro = 'Hello World, this is an intro ...'
 
     # Mock out the input call so we don't actually wait for a user's response on stdin
     m = mock.MagicMock(name='input', return_value='quit')
@@ -744,21 +721,19 @@ def test_base_cmdloop_without_queue():
 
     # Need to patch sys.argv so cmd2 doesn't think it was called with arguments equal to the py.test args
     testargs = ["prog"]
-    expected = app.intro + '\n'
+    expected = outsim_app.intro + '\n'
     with mock.patch.object(sys, 'argv', testargs):
         # Run the command loop
-        app.cmdloop()
-    out = app.stdout.getvalue()
+        outsim_app.cmdloop()
+    out = outsim_app.stdout.getvalue()
     assert out == expected
 
 
-def test_cmdloop_without_rawinput():
+def test_cmdloop_without_rawinput(outsim_app):
     # Create a cmd2.Cmd() instance and make sure basic settings are like we want for test
-    app = cmd2.Cmd()
-    app.use_rawinput = False
-    app.echo = False
-    app.intro = 'Hello World, this is an intro ...'
-    app.stdout = utils.StdSim(app.stdout)
+    outsim_app.use_rawinput = False
+    outsim_app.echo = False
+    outsim_app.intro = 'Hello World, this is an intro ...'
 
     # Mock out the input call so we don't actually wait for a user's response on stdin
     m = mock.MagicMock(name='input', return_value='quit')
@@ -766,11 +741,11 @@ def test_cmdloop_without_rawinput():
 
     # Need to patch sys.argv so cmd2 doesn't think it was called with arguments equal to the py.test args
     testargs = ["prog"]
-    expected = app.intro + '\n'
+    expected = outsim_app.intro + '\n'
     with mock.patch.object(sys, 'argv', testargs):
         # Run the command loop
-        app.cmdloop()
-    out = app.stdout.getvalue()
+        outsim_app.cmdloop()
+    out = outsim_app.stdout.getvalue()
     assert out == expected
 
 
@@ -788,7 +763,6 @@ class HookFailureApp(cmd2.Cmd):
 @pytest.fixture
 def hook_failure():
     app = HookFailureApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_precmd_hook_success(base_app):
@@ -858,7 +832,7 @@ def test_default_to_shell(base_app, monkeypatch):
     base_app.default_to_shell = True
     m = mock.Mock()
     monkeypatch.setattr("{}.Popen".format('subprocess'), m)
-    out = run_cmd(base_app, line)
+    out, err = run_cmd(base_app, line)
     assert out == []
     assert m.called
 
@@ -912,16 +886,15 @@ class HelpApp(cmd2.Cmd):
 @pytest.fixture
 def help_app():
     app = HelpApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_custom_command_help(help_app):
-    out = run_cmd(help_app, 'help squat')
+    out, err = run_cmd(help_app, 'help squat')
     expected = normalize('This command does diddly squat...')
     assert out == expected
 
 def test_custom_help_menu(help_app):
-    out = run_cmd(help_app, 'help')
+    out, err = run_cmd(help_app, 'help')
     expected = normalize("""
 Documented commands (type help <topic>):
 ========================================
@@ -934,13 +907,12 @@ undoc
 """)
     assert out == expected
 
-def test_help_undocumented(help_app, capsys):
-    run_cmd(help_app, 'help undoc')
-    out, err = capsys.readouterr()
-    assert err.startswith("No help on undoc")
+def test_help_undocumented(help_app):
+    out, err = run_cmd(help_app, 'help undoc')
+    assert err[0].startswith("No help on undoc")
 
 def test_help_overridden_method(help_app):
-    out = run_cmd(help_app, 'help edit')
+    out, err = run_cmd(help_app, 'help edit')
     expected = normalize('This overrides the edit command and does nothing.')
     assert out == expected
 
@@ -980,11 +952,10 @@ class HelpCategoriesApp(cmd2.Cmd):
 @pytest.fixture
 def helpcat_app():
     app = HelpCategoriesApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_help_cat_base(helpcat_app):
-    out = run_cmd(helpcat_app, 'help')
+    out, err = run_cmd(helpcat_app, 'help')
     expected = normalize("""Documented commands (type help <topic>):
 
 Custom Category
@@ -1006,7 +977,7 @@ undoc
     assert out == expected
 
 def test_help_cat_verbose(helpcat_app):
-    out = run_cmd(helpcat_app, 'help --verbose')
+    out, err = run_cmd(helpcat_app, 'help --verbose')
     expected = normalize("""Documented commands (type help <topic>):
 
 Custom Category
@@ -1074,7 +1045,6 @@ class SelectApp(cmd2.Cmd):
 @pytest.fixture
 def select_app():
     app = SelectApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_select_options(select_app):
@@ -1083,7 +1053,7 @@ def test_select_options(select_app):
     builtins.input = m
 
     food = 'bacon'
-    out = run_cmd(select_app, "eat {}".format(food))
+    out, err = run_cmd(select_app, "eat {}".format(food))
     expected = normalize("""
    1. sweet
    2. salty
@@ -1104,7 +1074,7 @@ def test_select_invalid_option_too_big(select_app):
     builtins.input = m
 
     food = 'fish'
-    out = run_cmd(select_app, "eat {}".format(food))
+    out, err = run_cmd(select_app, "eat {}".format(food))
     expected = normalize("""
    1. sweet
    2. salty
@@ -1128,7 +1098,7 @@ def test_select_invalid_option_too_small(select_app):
     builtins.input = m
 
     food = 'fish'
-    out = run_cmd(select_app, "eat {}".format(food))
+    out, err = run_cmd(select_app, "eat {}".format(food))
     expected = normalize("""
    1. sweet
    2. salty
@@ -1149,7 +1119,7 @@ def test_select_list_of_strings(select_app):
     m = mock.MagicMock(name='input', return_value='2')
     builtins.input = m
 
-    out = run_cmd(select_app, "study")
+    out, err = run_cmd(select_app, "study")
     expected = normalize("""
    1. math
    2. science
@@ -1167,7 +1137,7 @@ def test_select_list_of_tuples(select_app):
     m = mock.MagicMock(name='input', return_value='2')
     builtins.input = m
 
-    out = run_cmd(select_app, "procrastinate")
+    out, err = run_cmd(select_app, "procrastinate")
     expected = normalize("""
    1. Netflix
    2. WebSurfing
@@ -1186,7 +1156,7 @@ def test_select_uneven_list_of_tuples(select_app):
     m = mock.MagicMock(name='input', return_value='2')
     builtins.input = m
 
-    out = run_cmd(select_app, "play")
+    out, err = run_cmd(select_app, "play")
     expected = normalize("""
    1. Electric Guitar
    2. Drums
@@ -1262,7 +1232,6 @@ class MultilineApp(cmd2.Cmd):
 @pytest.fixture
 def multiline_app():
     app = MultilineApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_multiline_complete_empty_statement_raises_exception(multiline_app):
@@ -1329,7 +1298,6 @@ class CommandResultApp(cmd2.Cmd):
 @pytest.fixture
 def commandresult_app():
     app = CommandResultApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_commandresult_truthy(commandresult_app):
@@ -1489,7 +1457,7 @@ def test_pseudo_raw_input_piped_rawinput_false_echo_false(capsys):
     assert firstline.startswith('colors:')
     assert not '{}{}'.format(app.prompt, command) in out
 
-#
+
 # other input tests
 def test_raw_input(base_app):
     base_app.use_raw_input = True
@@ -1526,50 +1494,49 @@ def test_empty_stdin_input():
     line = app.pseudo_raw_input('(cmd2)')
     assert line == 'eof'
 
-
-def test_poutput_string(base_app):
+def test_poutput_string(outsim_app):
     msg = 'This is a test'
-    base_app.poutput(msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.poutput(msg)
+    out = outsim_app.stdout.getvalue()
     expected = msg + '\n'
     assert out == expected
 
-def test_poutput_zero(base_app):
+def test_poutput_zero(outsim_app):
     msg = 0
-    base_app.poutput(msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.poutput(msg)
+    out = outsim_app.stdout.getvalue()
     expected = str(msg) + '\n'
     assert out == expected
 
-def test_poutput_empty_string(base_app):
+def test_poutput_empty_string(outsim_app):
     msg = ''
-    base_app.poutput(msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.poutput(msg)
+    out = outsim_app.stdout.getvalue()
     expected = msg
     assert out == expected
 
-def test_poutput_none(base_app):
+def test_poutput_none(outsim_app):
     msg = None
-    base_app.poutput(msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.poutput(msg)
+    out = outsim_app.stdout.getvalue()
     expected = ''
     assert out == expected
 
-def test_poutput_color_always(base_app):
+def test_poutput_color_always(outsim_app):
     msg = 'Hello World'
     color = Fore.CYAN
-    base_app.colors = 'Always'
-    base_app.poutput(msg, color=color)
-    out = base_app.stdout.getvalue()
+    outsim_app.colors = 'Always'
+    outsim_app.poutput(msg, color=color)
+    out = outsim_app.stdout.getvalue()
     expected = color + msg + '\n' + Fore.RESET
     assert out == expected
 
-def test_poutput_color_never(base_app):
+def test_poutput_color_never(outsim_app):
     msg = 'Hello World'
     color = Fore.CYAN
-    base_app.colors = 'Never'
-    base_app.poutput(msg, color=color)
-    out = base_app.stdout.getvalue()
+    outsim_app.colors = 'Never'
+    outsim_app.poutput(msg, color=color)
+    out = outsim_app.stdout.getvalue()
     expected = msg + '\n'
     assert out == expected
 
@@ -1604,88 +1571,82 @@ def test_get_macro_names(base_app):
 def test_get_settable_names(base_app):
     assert sorted(base_app.get_settable_names()) == sorted(base_app.settable.keys())
 
-def test_alias_no_subcommand(base_app, capsys):
-    out = run_cmd(base_app, 'alias')
+def test_alias_no_subcommand(base_app):
+    out, err = run_cmd(base_app, 'alias')
     assert "Usage: alias [-h]" in out[0]
 
-def test_alias_create(base_app, capsys):
+def test_alias_create(base_app):
     # Create the alias
-    out = run_cmd(base_app, 'alias create fake pyscript')
+    out, err = run_cmd(base_app, 'alias create fake pyscript')
     assert out == normalize("Alias 'fake' created")
 
     # Use the alias
-    run_cmd(base_app, 'fake')
-    out, err = capsys.readouterr()
-    assert "the following arguments are required: script_path" in err
+    out, err = run_cmd(base_app, 'fake')
+    assert "the following arguments are required: script_path" in err[0]
 
     # See a list of aliases
-    out = run_cmd(base_app, 'alias list')
+    out, err = run_cmd(base_app, 'alias list')
     assert out == normalize('alias create fake pyscript')
 
     # Look up the new alias
-    out = run_cmd(base_app, 'alias list fake')
+    out, err = run_cmd(base_app, 'alias list fake')
     assert out == normalize('alias create fake pyscript')
 
-def test_alias_create_with_quoted_value(base_app, capsys):
+def test_alias_create_with_quoted_value(base_app):
     """Demonstrate that quotes in alias value will be preserved (except for redirectors)"""
 
     # Create the alias
-    out = run_cmd(base_app, 'alias create fake help ">" "out file.txt"')
+    out, err = run_cmd(base_app, 'alias create fake help ">" "out file.txt"')
     assert out == normalize("Alias 'fake' created")
 
     # Look up the new alias (Only the redirector should be unquoted)
-    out = run_cmd(base_app, 'alias list fake')
+    out, err = run_cmd(base_app, 'alias list fake')
     assert out == normalize('alias create fake help > "out file.txt"')
 
 @pytest.mark.parametrize('alias_name', invalid_command_name)
 def test_alias_create_invalid_name(base_app, alias_name, capsys):
-    run_cmd(base_app, 'alias create {} help'.format(alias_name))
-    out, err = capsys.readouterr()
-    assert "Invalid alias name" in err
+    out, err = run_cmd(base_app, 'alias create {} help'.format(alias_name))
+    assert "Invalid alias name" in err[0]
 
-def test_alias_create_with_macro_name(base_app, capsys):
+def test_alias_create_with_macro_name(base_app):
     macro = "my_macro"
     run_cmd(base_app, 'macro create {} help'.format(macro))
-    run_cmd(base_app, 'alias create {} help'.format(macro))
-    out, err = capsys.readouterr()
-    assert "Alias cannot have the same name as a macro" in err
+    out, err = run_cmd(base_app, 'alias create {} help'.format(macro))
+    assert "Alias cannot have the same name as a macro" in err[0]
 
-def test_alias_that_resolves_into_comment(base_app, capsys):
+def test_alias_that_resolves_into_comment(base_app):
     # Create the alias
-    out = run_cmd(base_app, 'alias create fake ' + constants.COMMENT_CHAR + ' blah blah')
+    out, err = run_cmd(base_app, 'alias create fake ' + constants.COMMENT_CHAR + ' blah blah')
     assert out == normalize("Alias 'fake' created")
 
     # Use the alias
-    run_cmd(base_app, 'fake')
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'fake')
     assert not out
     assert not err
 
-def test_alias_list_invalid_alias(base_app, capsys):
+def test_alias_list_invalid_alias(base_app):
     # Look up invalid alias
-    out = run_cmd(base_app, 'alias list invalid')
-    out, err = capsys.readouterr()
-    assert "Alias 'invalid' not found" in err
+    out, err = run_cmd(base_app, 'alias list invalid')
+    assert "Alias 'invalid' not found" in err[0]
 
 def test_alias_delete(base_app):
     # Create an alias
     run_cmd(base_app, 'alias create fake pyscript')
 
     # Delete the alias
-    out = run_cmd(base_app, 'alias delete fake')
+    out, err = run_cmd(base_app, 'alias delete fake')
     assert out == normalize("Alias 'fake' deleted")
 
 def test_alias_delete_all(base_app):
-    out = run_cmd(base_app, 'alias delete --all')
+    out, err = run_cmd(base_app, 'alias delete --all')
     assert out == normalize("All aliases deleted")
 
-def test_alias_delete_non_existing(base_app, capsys):
-    run_cmd(base_app, 'alias delete fake')
-    out, err = capsys.readouterr()
-    assert "Alias 'fake' does not exist" in err
+def test_alias_delete_non_existing(base_app):
+    out, err = run_cmd(base_app, 'alias delete fake')
+    assert "Alias 'fake' does not exist" in err[0]
 
-def test_alias_delete_no_name(base_app, capsys):
-    out = run_cmd(base_app, 'alias delete')
+def test_alias_delete_no_name(base_app):
+    out, err = run_cmd(base_app, 'alias delete')
     assert "Usage: alias delete" in out[0]
 
 def test_multiple_aliases(base_app):
@@ -1693,167 +1654,154 @@ def test_multiple_aliases(base_app):
     alias2 = 'h2'
     run_cmd(base_app, 'alias create {} help'.format(alias1))
     run_cmd(base_app, 'alias create {} help -v'.format(alias2))
-    out = run_cmd(base_app, alias1)
+    out, err = run_cmd(base_app, alias1)
     expected = normalize(BASE_HELP)
     assert out == expected
 
-    out = run_cmd(base_app, alias2)
+    out, err = run_cmd(base_app, alias2)
     expected = normalize(BASE_HELP_VERBOSE)
     assert out == expected
 
-def test_macro_no_subcommand(base_app, capsys):
-    out = run_cmd(base_app, 'macro')
+def test_macro_no_subcommand(base_app):
+    out, err = run_cmd(base_app, 'macro')
     assert "Usage: macro [-h]" in out[0]
 
-def test_macro_create(base_app, capsys):
+def test_macro_create(base_app):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake pyscript')
+    out, err = run_cmd(base_app, 'macro create fake pyscript')
     assert out == normalize("Macro 'fake' created")
 
     # Use the macro
-    run_cmd(base_app, 'fake')
-    out, err = capsys.readouterr()
-    assert "the following arguments are required: script_path" in err
+    out, err = run_cmd(base_app, 'fake')
+    assert "the following arguments are required: script_path" in err[0]
 
     # See a list of macros
-    out = run_cmd(base_app, 'macro list')
+    out, err = run_cmd(base_app, 'macro list')
     assert out == normalize('macro create fake pyscript')
 
     # Look up the new macro
-    out = run_cmd(base_app, 'macro list fake')
+    out, err = run_cmd(base_app, 'macro list fake')
     assert out == normalize('macro create fake pyscript')
 
-def test_macro_create_with_quoted_value(base_app, capsys):
+def test_macro_create_with_quoted_value(base_app):
     """Demonstrate that quotes in macro value will be preserved (except for redirectors)"""
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake help ">" "out file.txt"')
+    out, err = run_cmd(base_app, 'macro create fake help ">" "out file.txt"')
     assert out == normalize("Macro 'fake' created")
 
     # Look up the new macro (Only the redirector should be unquoted)
-    out = run_cmd(base_app, 'macro list fake')
+    out, err = run_cmd(base_app, 'macro list fake')
     assert out == normalize('macro create fake help > "out file.txt"')
 
 @pytest.mark.parametrize('macro_name', invalid_command_name)
-def test_macro_create_invalid_name(base_app, macro_name, capsys):
-    run_cmd(base_app, 'macro create {} help'.format(macro_name))
-    out, err = capsys.readouterr()
-    assert "Invalid macro name" in err
+def test_macro_create_invalid_name(base_app, macro_name):
+    out, err = run_cmd(base_app, 'macro create {} help'.format(macro_name))
+    assert "Invalid macro name" in err[0]
 
-def test_macro_create_with_alias_name(base_app, capsys):
+def test_macro_create_with_alias_name(base_app):
     macro = "my_macro"
     run_cmd(base_app, 'alias create {} help'.format(macro))
-    run_cmd(base_app, 'macro create {} help'.format(macro))
-    out, err = capsys.readouterr()
-    assert "Macro cannot have the same name as an alias" in err
+    out, err = run_cmd(base_app, 'macro create {} help'.format(macro))
+    assert "Macro cannot have the same name as an alias" in err[0]
 
-def test_macro_create_with_command_name(base_app, capsys):
+def test_macro_create_with_command_name(base_app):
     macro = "my_macro"
-    run_cmd(base_app, 'macro create help stuff')
-    out, err = capsys.readouterr()
-    assert "Macro cannot have the same name as a command" in err
+    out, err = run_cmd(base_app, 'macro create help stuff')
+    assert "Macro cannot have the same name as a command" in err[0]
 
-def test_macro_create_with_args(base_app, capsys):
+def test_macro_create_with_args(base_app):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake {1} {2}')
+    out, err = run_cmd(base_app, 'macro create fake {1} {2}')
     assert out == normalize("Macro 'fake' created")
 
     # Run the macro
-    out = run_cmd(base_app, 'fake help -v')
+    out, err = run_cmd(base_app, 'fake help -v')
     expected = normalize(BASE_HELP_VERBOSE)
     assert out == expected
 
-def test_macro_create_with_escaped_args(base_app, capsys):
+def test_macro_create_with_escaped_args(base_app):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake help {{1}}')
+    out, err = run_cmd(base_app, 'macro create fake help {{1}}')
     assert out == normalize("Macro 'fake' created")
 
     # Run the macro
-    run_cmd(base_app, 'fake')
-    out, err = capsys.readouterr()
-    assert err.startswith('No help on {1}')
+    out, err = run_cmd(base_app, 'fake')
+    assert err[0].startswith('No help on {1}')
 
-def test_macro_usage_with_missing_args(base_app, capsys):
+def test_macro_usage_with_missing_args(base_app):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake help {1} {2}')
+    out, err = run_cmd(base_app, 'macro create fake help {1} {2}')
     assert out == normalize("Macro 'fake' created")
 
     # Run the macro
-    run_cmd(base_app, 'fake arg1')
-    out, err = capsys.readouterr()
-    assert "expects at least 2 argument(s)" in err
+    out, err = run_cmd(base_app, 'fake arg1')
+    assert "expects at least 2 argument(s)" in err[0]
 
-def test_macro_usage_with_exta_args(base_app, capsys):
+def test_macro_usage_with_exta_args(base_app):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake help {1}')
+    out, err = run_cmd(base_app, 'macro create fake help {1}')
     assert out == normalize("Macro 'fake' created")
 
     # Run the macro
-    out = run_cmd(base_app, 'fake alias create')
+    out, err = run_cmd(base_app, 'fake alias create')
     assert "Usage: alias create" in out[0]
 
-def test_macro_create_with_missing_arg_nums(base_app, capsys):
+def test_macro_create_with_missing_arg_nums(base_app):
     # Create the macro
-    run_cmd(base_app, 'macro create fake help {1} {3}')
-    out, err = capsys.readouterr()
-    assert "Not all numbers between 1 and 3" in err
+    out, err = run_cmd(base_app, 'macro create fake help {1} {3}')
+    assert "Not all numbers between 1 and 3" in err[0]
 
-def test_macro_create_with_invalid_arg_num(base_app, capsys):
+def test_macro_create_with_invalid_arg_num(base_app):
     # Create the macro
-    run_cmd(base_app, 'macro create fake help {1} {-1} {0}')
-    out, err = capsys.readouterr()
-    assert "Argument numbers must be greater than 0" in err
+    out, err = run_cmd(base_app, 'macro create fake help {1} {-1} {0}')
+    assert "Argument numbers must be greater than 0" in err[0]
 
-def test_macro_create_with_unicode_numbered_arg(base_app, capsys):
+def test_macro_create_with_unicode_numbered_arg(base_app):
     # Create the macro expecting 1 argument
-    out = run_cmd(base_app, 'macro create fake help {\N{ARABIC-INDIC DIGIT ONE}}')
+    out, err = run_cmd(base_app, 'macro create fake help {\N{ARABIC-INDIC DIGIT ONE}}')
     assert out == normalize("Macro 'fake' created")
 
     # Run the macro
-    out = run_cmd(base_app, 'fake')
-    out, err = capsys.readouterr()
-    assert "expects at least 1 argument(s)" in err
+    out, err = run_cmd(base_app, 'fake')
+    assert "expects at least 1 argument(s)" in err[0]
 
-def test_macro_create_with_missing_unicode_arg_nums(base_app, capsys):
-    run_cmd(base_app, 'macro create fake help {1} {\N{ARABIC-INDIC DIGIT THREE}}')
-    out, err = capsys.readouterr()
-    assert "Not all numbers between 1 and 3" in err
+def test_macro_create_with_missing_unicode_arg_nums(base_app):
+    out, err = run_cmd(base_app, 'macro create fake help {1} {\N{ARABIC-INDIC DIGIT THREE}}')
+    assert "Not all numbers between 1 and 3" in err[0]
 
-def test_macro_that_resolves_into_comment(base_app, capsys):
+def test_macro_that_resolves_into_comment(base_app):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake {1} blah blah')
+    out, err = run_cmd(base_app, 'macro create fake {1} blah blah')
     assert out == normalize("Macro 'fake' created")
 
     # Use the macro
-    run_cmd(base_app, 'fake ' + constants.COMMENT_CHAR)
-    out, err = capsys.readouterr()
+    out, err = run_cmd(base_app, 'fake ' + constants.COMMENT_CHAR)
     assert not out
     assert not err
 
-def test_macro_list_invalid_macro(base_app, capsys):
+def test_macro_list_invalid_macro(base_app):
     # Look up invalid macro
-    run_cmd(base_app, 'macro list invalid')
-    out, err = capsys.readouterr()
-    assert "Macro 'invalid' not found" in err
+    out, err = run_cmd(base_app, 'macro list invalid')
+    assert "Macro 'invalid' not found" in err[0]
 
 def test_macro_delete(base_app):
     # Create an macro
     run_cmd(base_app, 'macro create fake pyscript')
 
     # Delete the macro
-    out = run_cmd(base_app, 'macro delete fake')
+    out, err = run_cmd(base_app, 'macro delete fake')
     assert out == normalize("Macro 'fake' deleted")
 
 def test_macro_delete_all(base_app):
-    out = run_cmd(base_app, 'macro delete --all')
+    out, err = run_cmd(base_app, 'macro delete --all')
     assert out == normalize("All macros deleted")
 
-def test_macro_delete_non_existing(base_app, capsys):
-    run_cmd(base_app, 'macro delete fake')
-    out, err = capsys.readouterr()
-    assert "Macro 'fake' does not exist" in err
+def test_macro_delete_non_existing(base_app):
+    out, err = run_cmd(base_app, 'macro delete fake')
+    assert "Macro 'fake' does not exist" in err[0]
 
-def test_macro_delete_no_name(base_app, capsys):
-    out = run_cmd(base_app, 'macro delete')
+def test_macro_delete_no_name(base_app):
+    out, err = run_cmd(base_app, 'macro delete')
     assert "Usage: macro delete" in out[0]
 
 def test_multiple_macros(base_app):
@@ -1861,15 +1809,15 @@ def test_multiple_macros(base_app):
     macro2 = 'h2'
     run_cmd(base_app, 'macro create {} help'.format(macro1))
     run_cmd(base_app, 'macro create {} help -v'.format(macro2))
-    out = run_cmd(base_app, macro1)
+    out, err = run_cmd(base_app, macro1)
     expected = normalize(BASE_HELP)
     assert out == expected
 
-    out = run_cmd(base_app, macro2)
+    out, err = run_cmd(base_app, macro2)
     expected = normalize(BASE_HELP_VERBOSE)
     assert out == expected
 
-def test_nonexistent_macro(base_app, capsys):
+def test_nonexistent_macro(base_app):
     from cmd2.parsing import StatementParser
     exception = None
 
@@ -1881,29 +1829,29 @@ def test_nonexistent_macro(base_app, capsys):
     assert exception is not None
 
 
-def test_ppaged(base_app):
+def test_ppaged(outsim_app):
     msg = 'testing...'
     end = '\n'
-    base_app.ppaged(msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.ppaged(msg)
+    out = outsim_app.stdout.getvalue()
     assert out == msg + end
 
-def test_ppaged_strips_color_when_redirecting(base_app):
+def test_ppaged_strips_color_when_redirecting(outsim_app):
     msg = 'testing...'
     end = '\n'
-    base_app.colors = cmd2.constants.COLORS_TERMINAL
-    base_app.redirecting = True
-    base_app.ppaged(Fore.RED + msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.colors = cmd2.constants.COLORS_TERMINAL
+    outsim_app.redirecting = True
+    outsim_app.ppaged(Fore.RED + msg)
+    out = outsim_app.stdout.getvalue()
     assert out == msg + end
 
-def test_ppaged_strips_color_when_redirecting_if_always(base_app):
+def test_ppaged_strips_color_when_redirecting_if_always(outsim_app):
     msg = 'testing...'
     end = '\n'
-    base_app.colors = cmd2.constants.COLORS_ALWAYS
-    base_app.redirecting = True
-    base_app.ppaged(Fore.RED + msg)
-    out = base_app.stdout.getvalue()
+    outsim_app.colors = cmd2.constants.COLORS_ALWAYS
+    outsim_app.redirecting = True
+    outsim_app.ppaged(Fore.RED + msg)
+    out = outsim_app.stdout.getvalue()
     assert out == Fore.RED + msg + end
 
 # we override cmd.parseline() so we always get consistent
@@ -1925,17 +1873,17 @@ def test_parseline(base_app):
     assert line == statement.strip()
 
 
-def test_onecmd_raw_str_continue(base_app):
+def test_onecmd_raw_str_continue(outsim_app):
     line = "help"
-    stop = base_app.onecmd(line)
-    out = base_app.stdout.getvalue()
+    stop = outsim_app.onecmd(line)
+    out = outsim_app.stdout.getvalue()
     assert not stop
     assert normalize(out) == normalize(BASE_HELP)
 
-def test_onecmd_raw_str_quit(base_app):
+def test_onecmd_raw_str_quit(outsim_app):
     line = "quit"
-    stop = base_app.onecmd(line)
-    out = base_app.stdout.getvalue()
+    stop = outsim_app.onecmd(line)
+    out = outsim_app.stdout.getvalue()
     assert stop
     assert out == ''
 
@@ -1986,13 +1934,13 @@ Usage:  exit [exit_code]
 @pytest.fixture
 def exit_code_repl():
     app = ReplWithExitCode()
+    app.stdout = utils.StdSim(app.stdout)
     return app
 
 def test_exit_code_default(exit_code_repl):
     # Create a cmd2.Cmd() instance and make sure basic settings are like we want for test
     app = exit_code_repl
     app.use_rawinput = True
-    app.stdout = utils.StdSim(app.stdout)
 
     # Mock out the input call so we don't actually wait for a user's response on stdin
     m = mock.MagicMock(name='input', return_value='exit')
@@ -2011,7 +1959,6 @@ def test_exit_code_nonzero(exit_code_repl):
     # Create a cmd2.Cmd() instance and make sure basic settings are like we want for test
     app = exit_code_repl
     app.use_rawinput = True
-    app.stdout = utils.StdSim(app.stdout)
 
     # Mock out the input call so we don't actually wait for a user's response on stdin
     m = mock.MagicMock(name='input', return_value='exit 23')
@@ -2178,31 +2125,26 @@ class DisableCommandsApp(cmd2.Cmd):
 @pytest.fixture
 def disable_commands_app():
     app = DisableCommandsApp()
-    app.stdout = utils.StdSim(app.stdout)
     return app
 
 
-def test_disable_and_enable_category(disable_commands_app, capsys):
+def test_disable_and_enable_category(disable_commands_app):
     # Disable the category
     message_to_print = 'These commands are currently disabled'
     disable_commands_app.disable_category(disable_commands_app.category_name, message_to_print)
 
     # Make sure all the commands and help on those commands displays the message
-    run_cmd(disable_commands_app, 'has_help_func')
-    out, err = capsys.readouterr()
-    assert err.startswith(message_to_print)
+    out, err = run_cmd(disable_commands_app, 'has_help_func')
+    assert err[0].startswith(message_to_print)
 
-    run_cmd(disable_commands_app, 'help has_help_func')
-    out, err = capsys.readouterr()
-    assert err.startswith(message_to_print)
+    out, err = run_cmd(disable_commands_app, 'help has_help_func')
+    assert err[0].startswith(message_to_print)
 
-    run_cmd(disable_commands_app, 'has_no_help_func')
-    out, err = capsys.readouterr()
-    assert err.startswith(message_to_print)
+    out, err = run_cmd(disable_commands_app, 'has_no_help_func')
+    assert err[0].startswith(message_to_print)
 
-    run_cmd(disable_commands_app, 'help has_no_help_func')
-    out, err = capsys.readouterr()
-    assert err.startswith(message_to_print)
+    out, err = run_cmd(disable_commands_app, 'help has_no_help_func')
+    assert err[0].startswith(message_to_print)
 
     visible_commands = disable_commands_app.get_visible_commands()
     assert 'has_help_func' not in visible_commands
@@ -2212,17 +2154,17 @@ def test_disable_and_enable_category(disable_commands_app, capsys):
     disable_commands_app.enable_category(disable_commands_app.category_name)
 
     # Make sure all the commands and help on those commands are restored
-    out = run_cmd(disable_commands_app, 'has_help_func')
-    assert out == ["The real has_help_func"]
+    out, err = run_cmd(disable_commands_app, 'has_help_func')
+    assert out[0] == "The real has_help_func"
 
-    out = run_cmd(disable_commands_app, 'help has_help_func')
-    assert out == ["Help for has_help_func"]
+    out, err = run_cmd(disable_commands_app, 'help has_help_func')
+    assert out[0] == "Help for has_help_func"
 
-    out = run_cmd(disable_commands_app, 'has_no_help_func')
-    assert out == ["The real has_no_help_func"]
+    out, err = run_cmd(disable_commands_app, 'has_no_help_func')
+    assert out[0] == "The real has_no_help_func"
 
-    out = run_cmd(disable_commands_app, 'help has_no_help_func')
-    assert out == ["Help for has_no_help_func"]
+    out, err = run_cmd(disable_commands_app, 'help has_no_help_func')
+    assert out[0] == "Help for has_no_help_func"
 
     visible_commands = disable_commands_app.get_visible_commands()
     assert 'has_help_func' in visible_commands
@@ -2263,10 +2205,9 @@ def test_disabled_command_not_in_history(disable_commands_app):
     run_cmd(disable_commands_app, 'has_help_func')
     assert saved_len == len(disable_commands_app.history)
 
-def test_disabled_message_command_name(disable_commands_app, capsys):
+def test_disabled_message_command_name(disable_commands_app):
     message_to_print = '{} is currently disabled'.format(cmd2.cmd2.COMMAND_NAME)
     disable_commands_app.disable_command('has_help_func', message_to_print)
 
-    run_cmd(disable_commands_app, 'has_help_func')
-    out, err = capsys.readouterr()
-    assert err.startswith('has_help_func is currently disabled')
+    out, err = run_cmd(disable_commands_app, 'has_help_func')
+    assert err[0].startswith('has_help_func is currently disabled')
diff --git a/tests/test_completion.py b/tests/test_completion.py
index 47a7a9d6..0916e500 100644
--- a/tests/test_completion.py
+++ b/tests/test_completion.py
@@ -122,7 +122,7 @@ def test_complete_bogus_command(cmd2_app):
 
 def test_complete_macro(base_app, request):
     # Create the macro
-    out = run_cmd(base_app, 'macro create fake pyscript {1}')
+    out, err = run_cmd(base_app, 'macro create fake pyscript {1}')
     assert out == normalize("Macro 'fake' created")
 
     # Macros do path completion
diff --git a/tests/test_history.py b/tests/test_history.py
index 9046e877..058386a3 100644
--- a/tests/test_history.py
+++ b/tests/test_history.py
@@ -16,13 +16,11 @@ except ImportError:
     from unittest import mock
 
 import cmd2
-from cmd2 import clipboard
-from cmd2 import utils
 from .conftest import run_cmd, normalize, HELP_HISTORY
 
 
 def test_base_help_history(base_app):
-    out = run_cmd(base_app, 'help history')
+    out, err = run_cmd(base_app, 'help history')
     assert out == normalize(HELP_HISTORY)
 
 def test_exclude_from_history(base_app, monkeypatch):
@@ -40,14 +38,14 @@ def test_exclude_from_history(base_app, monkeypatch):
     run_cmd(base_app, 'history')
 
     # Verify that the history is empty
-    out = run_cmd(base_app, 'history')
+    out, err = run_cmd(base_app, 'history')
     assert out == []
 
     # Now run a command which isn't excluded from the history
     run_cmd(base_app, 'help')
 
     # And verify we have a history now ...
-    out = run_cmd(base_app, 'history')
+    out, err = run_cmd(base_app, 'history')
     expected = normalize("""    1  help""")
     assert out == expected
 
@@ -126,20 +124,20 @@ def test_history_regex_search(hist):
 def test_base_history(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
-    out = run_cmd(base_app, 'history')
+    out, err = run_cmd(base_app, 'history')
     expected = normalize("""
     1  help
     2  shortcuts
 """)
     assert out == expected
 
-    out = run_cmd(base_app, 'history he')
+    out, err = run_cmd(base_app, 'history he')
     expected = normalize("""
     1  help
 """)
     assert out == expected
 
-    out = run_cmd(base_app, 'history sh')
+    out, err = run_cmd(base_app, 'history sh')
     expected = normalize("""
     2  shortcuts
 """)
@@ -148,7 +146,7 @@ def test_base_history(base_app):
 def test_history_script_format(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
-    out = run_cmd(base_app, 'history -s')
+    out, err = run_cmd(base_app, 'history -s')
     expected = normalize("""
 help
 shortcuts
@@ -159,7 +157,7 @@ def test_history_with_string_argument(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
     run_cmd(base_app, 'help history')
-    out = run_cmd(base_app, 'history help')
+    out, err = run_cmd(base_app, 'history help')
     expected = normalize("""
     1  help
     3  help history
@@ -171,7 +169,7 @@ def test_history_expanded_with_string_argument(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'help history')
     run_cmd(base_app, 'sc')
-    out = run_cmd(base_app, 'history -v shortcuts')
+    out, err = run_cmd(base_app, 'history -v shortcuts')
     expected = normalize("""
     1  alias create sc shortcuts
     4  sc
@@ -184,7 +182,7 @@ def test_history_expanded_with_regex_argument(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'help history')
     run_cmd(base_app, 'sc')
-    out = run_cmd(base_app, 'history -v /sh.*cuts/')
+    out, err = run_cmd(base_app, 'history -v /sh.*cuts/')
     expected = normalize("""
     1  alias create sc shortcuts
     4  sc
@@ -195,7 +193,7 @@ def test_history_expanded_with_regex_argument(base_app):
 def test_history_with_integer_argument(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
-    out = run_cmd(base_app, 'history 1')
+    out, err = run_cmd(base_app, 'history 1')
     expected = normalize("""
     1  help
 """)
@@ -206,7 +204,7 @@ def test_history_with_integer_span(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
     run_cmd(base_app, 'help history')
-    out = run_cmd(base_app, 'history 1..2')
+    out, err = run_cmd(base_app, 'history 1..2')
     expected = normalize("""
     1  help
     2  shortcuts
@@ -217,7 +215,7 @@ def test_history_with_span_start(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
     run_cmd(base_app, 'help history')
-    out = run_cmd(base_app, 'history 2:')
+    out, err = run_cmd(base_app, 'history 2:')
     expected = normalize("""
     2  shortcuts
     3  help history
@@ -228,7 +226,7 @@ def test_history_with_span_end(base_app):
     run_cmd(base_app, 'help')
     run_cmd(base_app, 'shortcuts')
     run_cmd(base_app, 'help history')
-    out = run_cmd(base_app, 'history :2')
+    out, err = run_cmd(base_app, 'history :2')
     expected = normalize("""
     1  help
     2  shortcuts
@@ -240,8 +238,8 @@ def test_history_with_span_index_error(base_app):
     run_cmd(base_app, 'help history')
     run_cmd(base_app, '!ls -hal :')
     with pytest.raises(ValueError):
-        _, err = run_cmd(base_app, 'history "hal :"')
-        assert "ValueError" in err
+        out, err = run_cmd(base_app, 'history "hal :"')
+        assert "ValueError" in err[0]
 
 def test_history_output_file(base_app):
     run_cmd(base_app, 'help')
@@ -275,7 +273,7 @@ def test_history_edit(base_app, monkeypatch):
 def test_history_run_all_commands(base_app):
     # make sure we refuse to run all commands as a default
     run_cmd(base_app, 'shortcuts')
-    out = run_cmd(base_app, 'history -r')
+    out, err = run_cmd(base_app, 'history -r')
     # this should generate an error, but we don't currently have a way to
     # capture stderr in these tests. So we assume that if we got nothing on
     # standard out, that the error occurred because if the command executed
@@ -283,9 +281,9 @@ def test_history_run_all_commands(base_app):
     assert out == []
 
 def test_history_run_one_command(base_app):
-    expected = run_cmd(base_app, 'help')
-    output = run_cmd(base_app, 'history -r 1')
-    assert output == expected
+    out1, err1 = run_cmd(base_app, 'help')
+    out2, err2 = run_cmd(base_app, 'history -r 1')
+    assert out1 == out2
 
 def test_history_clear(base_app):
     # Add commands to history
@@ -293,72 +291,72 @@ def test_history_clear(base_app):
     run_cmd(base_app, 'alias')
 
     # Make sure history has items
-    out = run_cmd(base_app, 'history')
+    out, err = run_cmd(base_app, 'history')
     assert out
 
     # Clear the history
     run_cmd(base_app, 'history --clear')
 
     # Make sure history is empty
-    out = run_cmd(base_app, 'history')
+    out, err = run_cmd(base_app, 'history')
     assert out == []
 
 def test_history_verbose_with_other_options(base_app):
     # make sure -v shows a usage error if any other options are present
     options_to_test = ['-r', '-e', '-o file', '-t file', '-c', '-x']
     for opt in options_to_test:
-        output = run_cmd(base_app, 'history -v ' + opt)
-        assert len(output) == 3
-        assert output[1].startswith('Usage:')
+        out, err = run_cmd(base_app, 'history -v ' + opt)
+        assert len(out) == 3
+        assert out[1].startswith('Usage:')
 
 def test_history_verbose(base_app):
     # validate function of -v option
     run_cmd(base_app, 'alias create s shortcuts')
     run_cmd(base_app, 's')
-    output = run_cmd(base_app, 'history -v')
-    assert len(output) == 3
+    out, err = run_cmd(base_app, 'history -v')
+    assert len(out) == 3
     # TODO test for basic formatting once we figure it out
 
 def test_history_script_with_invalid_options(base_app):
     # make sure -s shows a usage error if -c, -r, -e, -o, or -t are present
     options_to_test = ['-r', '-e', '-o file', '-t file', '-c']
     for opt in options_to_test:
-        output = run_cmd(base_app, 'history -s ' + opt)
-        assert len(output) == 3
-        assert output[1].startswith('Usage:')
+        out, err = run_cmd(base_app, 'history -s ' + opt)
+        assert len(out) == 3
+        assert out[1].startswith('Usage:')
 
 def test_history_script(base_app):
     cmds = ['alias create s shortcuts', 's']
     for cmd in cmds:
         run_cmd(base_app, cmd)
-    output = run_cmd(base_app, 'history -s')
-    assert output == cmds
+    out, err = run_cmd(base_app, 'history -s')
+    assert out == cmds
 
 def test_history_expanded_with_invalid_options(base_app):
     # make sure -x shows a usage error if -c, -r, -e, -o, or -t are present
     options_to_test = ['-r', '-e', '-o file', '-t file', '-c']
     for opt in options_to_test:
-        output = run_cmd(base_app, 'history -x ' + opt)
-        assert len(output) == 3
-        assert output[1].startswith('Usage:')
+        out, err = run_cmd(base_app, 'history -x ' + opt)
+        assert len(out) == 3
+        assert out[1].startswith('Usage:')
 
 def test_history_expanded(base_app):
     # validate function of -x option
     cmds = ['alias create s shortcuts', 's']
     for cmd in cmds:
         run_cmd(base_app, cmd)
-    output = run_cmd(base_app, 'history -x')
+    out, err = run_cmd(base_app, 'history -x')
     expected = ['    1  alias create s shortcuts', '    2  shortcuts']
-    assert output == expected
+    assert out == expected
 
 def test_history_script_expanded(base_app):
     # validate function of -s -x options together
     cmds = ['alias create s shortcuts', 's']
     for cmd in cmds:
         run_cmd(base_app, cmd)
-    output = run_cmd(base_app, 'history -sx')
+    out, err = run_cmd(base_app, 'history -sx')
     expected = ['alias create s shortcuts', 'shortcuts']
-    assert output == expected
+    assert out == expected
 
 
 #####
diff --git a/tests/test_pyscript.py b/tests/test_pyscript.py
index 78500185..6981980b 100644
--- a/tests/test_pyscript.py
+++ b/tests/test_pyscript.py
@@ -11,21 +11,15 @@ from .conftest import run_cmd
 def test_pyscript_help(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     python_script = os.path.join(test_dir, 'pyscript', 'help.py')
-    expected = run_cmd(base_app, 'help')
+    out1, err1 = run_cmd(base_app, 'help')
+    out2, err2 = run_cmd(base_app, 'pyscript {}'.format(python_script))
+    assert out1 and out1 == out2
 
-    assert len(expected) > 0
-    assert len(expected[0]) > 0
-    out = run_cmd(base_app, 'pyscript {}'.format(python_script))
-    assert len(out) > 0
-    assert out == expected
 
-
-def test_pyscript_dir(base_app, capsys, request):
+def test_pyscript_dir(base_app, request):
     test_dir = os.path.dirname(request.module.__file__)
     python_script = os.path.join(test_dir, 'pyscript', 'pyscript_dir.py')
 
-    run_cmd(base_app, 'pyscript {}'.format(python_script))
-    out, _ = capsys.readouterr()
-    out = out.strip()
-    assert len(out) > 0
-    assert out == "['cmd_echo']"
+    out, err = run_cmd(base_app, 'pyscript {}'.format(python_script))
+    assert out
+    assert out[0] == "['cmd_echo']"