blob: f6b6a21ad5aa8ccbce5fc10300cf29492853eb6e [file] [log] [blame]
README FOR IDLE TESTS IN IDLELIB.IDLE_TEST
1. Test Files
The idle directory, idlelib, has over 60 xyz.py files. The idle_test
subdirectory should contain a test_xyy.py for each. (For test modules, make
'xyz' lower case, and possibly shorten it.) Each file should start with the
something like the following template, with the blanks after after '.' and 'as',
and before and after '_' filled in.
---
import unittest
from test.support import requires
import idlelib. as
class _Test(unittest.TestCase):
def test_(self):
if __name__ == '__main__':
unittest.main(verbosity=2, exit=2)
---
Idle tests are run with unittest; do not use regrtest's test_main.
Once test_xyy is written, the following should go at the end of xyy.py,
with xyz (lowercased) added after 'test_'.
---
if __name__ == "__main__":
import unittest
unittest.main('idlelib.idle_test.test_', verbosity=2, exit=False)
---
2. Gui Tests
Gui tests need 'requires' from test.support (test.test_support in 2.7). A
test is a gui test if it creates a Tk root or master object either directly
or indirectly by instantiating a tkinter or idle class. For the benefit of
test processes that either have no graphical environment available or are not
allowed to use it, gui tests must be 'guarded' by "requires('gui')" in a
setUp function or method. This will typically be setUpClass.
To avoid interfering with other gui tests, all gui objects must be destroyed
and deleted by the end of the test. If a widget, such as a Tk root, is created
in a setUpX function, destroy it in the corresponding tearDownX. For module
and class attributes, also delete the widget.
---
@classmethod
def setUpClass(cls):
requires('gui')
cls.root = tk.Tk()
@classmethod
def tearDownClass(cls):
cls.root.destroy()
del cls.root
---
Support.requires('gui') causes the test(s) it guards to be skipped if any of
a few conditions are met:
- The tests are being run by regrtest.py, and it was started without
enabling the "gui" resource with the "-u" command line option.
- The tests are being run on Windows by a service that is not allowed to
interact with the graphical environment.
- The tests are being run on Mac OSX in a process that cannot make a window
manager connection.
- tkinter.Tk cannot be successfully instantiated for some reason.
- test.support.use_resources has been set by something other than
regrtest.py and does not contain "gui".
Since non-gui tests always run, but gui tests only sometimes, tests of non-gui
operations should best avoid needing a gui. Methods that make incidental use of
tkinter (tk) variables and messageboxes can do this by using the mock classes in
idle_test/mock_tk.py. There is also a mock text that will handle some uses of the
tk Text widget.
3. Running Tests
Assume that xyz.py and test_xyz.py end with the "if __name__" statements given
above. In Idle, pressing F5 in an editor window with either loaded will run all
tests in the test_xyz file with the version of Python running Idle. The test
report and any tracebacks will appear in the Shell window. The options in these
"if __name__" statements are appropriate for developers running (as opposed to
importing) either of the files during development: verbosity=2 lists all test
methods in the file; exit=False avoids a spurious sys.exit traceback that would
otherwise occur when running in Idle. The following command lines also run
all test methods, including gui tests, in test_xyz.py. (The exceptions are that
idlelib and idlelib.idle start Idle and idlelib.PyShell should (issue 18330).)
python -m idlelib.xyz # With the capitalization of the xyz module
python -m idlelib.idle_test.test_xyz
To run all idle_test/test_*.py tests, either interactively
('>>>', with unittest imported) or from a command line, use one of the
following. (Notes: in 2.7, 'test ' (with the space) is 'test.regrtest ';
where present, -v and -ugui can be omitted.)
>>> unittest.main('idlelib.idle_test', verbosity=2, exit=False)
python -m unittest -v idlelib.idle_test
python -m test -v -ugui test_idle
python -m test.test_idle
The idle tests are 'discovered' by idlelib.idle_test.__init__.load_tests,
which is also imported into test.test_idle. Normally, neither file should be
changed when working on individual test modules. The third command runs
unittest indirectly through regrtest. The same happens when the entire test
suite is run with 'python -m test'. So that command must work for buildbots
to stay green. Idle tests must not disturb the environment in a way that
makes other tests fail (issue 18081).
To run an individual Testcase or test method, extend the dotted name given to
unittest on the command line.
python -m unittest -v idlelib.idle_test.test_xyz.Test_case.test_meth