1 Writing Python Regression Tests
2 -------------------------------
9 If you add a new module to Python or modify the functionality of an existing
10 module, you should write one or more test cases to exercise that new
11 functionality. The mechanics of how the test system operates are fairly
12 straightforward. When a test case is run, the output is compared with the
13 expected output that is stored in .../Lib/test/output. If the test runs to
14 completion and the actual and expected outputs match, the test succeeds, if
15 not, it fails. If an ImportError or test_support.TestSkipped error is
16 raised, the test is not run.
18 You will be writing unit tests (isolated tests of functions and objects
19 defined by the module) using white box techniques. Unlike black box
20 testing, where you only have the external interfaces to guide your test case
21 writing, in white box testing you can see the code being tested and tailor
22 your test cases to exercise it more completely. In particular, you will be
23 able to refer to the C and Python code in the CVS repository when writing
24 your regression test cases.
29 If you are writing test cases for module spam, you need to create a file
30 in .../Lib/test named test_spam.py and an expected output file in
31 .../Lib/test/output named test_spam ("..." represents the top-level
32 directory in the Python source tree, the directory containing the configure
33 script). From the top-level directory, generate the initial version of the
34 test output file by executing:
36 ./python Lib/test/regrtest.py -g test_spam.py
38 Any time you modify test_spam.py you need to generate a new expected
39 output file. Don't forget to desk check the generated output to make sure
40 it's really what you expected to find! To run a single test after modifying
41 a module, simply run regrtest.py without the -g flag:
43 ./python Lib/test/regrtest.py test_spam.py
45 While debugging a regression test, you can of course execute it
46 independently of the regression testing framework and see what it prints:
48 ./python Lib/test/test_spam.py
50 To run the entire test suite, make the "test" target at the top level:
54 On non-Unix platforms where make may not be available, you can simply
55 execute the two runs of regrtest (optimized and non-optimized) directly:
57 ./python Lib/test/regrtest.py
58 ./python -O Lib/test/regrtest.py
61 Test cases generate output based upon values computed by the test code.
62 When executed, regrtest.py compares the actual output generated by executing
63 the test case with the expected output and reports success or failure. It
64 stands to reason that if the actual and expected outputs are to match, they
65 must not contain any machine dependencies. This means your test cases
66 should not print out absolute machine addresses (e.g. the return value of
67 the id() builtin function) or floating point numbers with large numbers of
68 significant digits (unless you understand what you are doing!).
71 Test Case Writing Tips
73 Writing good test cases is a skilled task and is too complex to discuss in
74 detail in this short document. Many books have been written on the subject.
75 I'll show my age by suggesting that Glenford Myers' "The Art of Software
76 Testing", published in 1979, is still the best introduction to the subject
77 available. It is short (177 pages), easy to read, and discusses the major
78 elements of software testing, though its publication predates the
79 object-oriented software revolution, so doesn't cover that subject at all.
80 Unfortunately, it is very expensive (about $100 new). If you can borrow it
81 or find it used (around $20), I strongly urge you to pick up a copy.
83 The most important goal when writing test cases is to break things. A test
84 case that doesn't uncover a bug is much less valuable than one that does.
85 In designing test cases you should pay attention to the following:
87 * Your test cases should exercise all the functions and objects defined
88 in the module, not just the ones meant to be called by users of your
89 module. This may require you to write test code that uses the module
90 in ways you don't expect (explicitly calling internal functions, for
91 example - see test_atexit.py).
93 * You should consider any boundary values that may tickle exceptional
94 conditions (e.g. if you were writing regression tests for division,
95 you might well want to generate tests with numerators and denominators
96 at the limits of floating point and integer numbers on the machine
97 performing the tests as well as a denominator of zero).
99 * You should exercise as many paths through the code as possible. This
100 may not always be possible, but is a goal to strive for. In
101 particular, when considering if statements (or their equivalent), you
102 want to create test cases that exercise both the true and false
103 branches. For loops, you should create test cases that exercise the
104 loop zero, one and multiple times.
106 * You should test with obviously invalid input. If you know that a
107 function requires an integer input, try calling it with other types of
108 objects to see how it responds.
110 * You should test with obviously out-of-range input. If the domain of a
111 function is only defined for positive integers, try calling it with a
114 * If you are going to fix a bug that wasn't uncovered by an existing
115 test, try to write a test case that exposes the bug (preferably before
118 * If you need to create a temporary file, you can use the filename in
119 test_support.TESTFN to do so. It is important to remove the file
120 when done; other tests should be able to use the name without cleaning
124 Regression Test Writing Rules
126 Each test case is different. There is no "standard" form for a Python
127 regression test case, though there are some general rules:
129 * If your test case detects a failure, raise TestFailed (found in
132 * Import everything you'll need as early as possible.
134 * If you'll be importing objects from a module that is at least
135 partially platform-dependent, only import those objects you need for
136 the current test case to avoid spurious ImportError exceptions that
137 prevent the test from running to completion.
139 * Print all your test case results using the print statement. For
140 non-fatal errors, print an error message (or omit a successful
141 completion print) to indicate the failure, but proceed instead of
144 * Use "assert" sparingly, if at all. It's usually better to just print
145 what you got, and rely on regrtest's got-vs-expected comparison to
146 catch deviations from what you expect. assert statements aren't
147 executed at all when regrtest is run in -O mode; and, because they
148 cause the test to stop immediately, can lead to a long & tedious
149 test-fix, test-fix, test-fix, ... cycle when things are badly broken
150 (and note that "badly broken" often includes running the test suite
151 for the first time on new platforms or under new implementations of
157 There is a test_support module you can import from your test case. It
158 provides the following useful objects:
160 * TestFailed - raise this exception when your regression test detects a
163 * TestSkipped - raise this if the test could not be run because the
164 platform doesn't offer all the required facilities (like large
165 file support), even if all the required modules are available.
167 * findfile(file) - you can call this function to locate a file somewhere
168 along sys.path or in the Lib/test tree - see test_linuxaudiodev.py for
169 an example of its use.
171 * verbose - you can use this variable to control print output. Many
172 modules use it. Search for "verbose" in the test_*.py files to see
175 * use_large_resources - true iff tests requiring large time or space
178 * fcmp(x,y) - you can call this function to compare two floating point
179 numbers when you expect them to only be approximately equal withing a
180 fuzz factor (test_support.FUZZ, which defaults to 1e-6).
182 NOTE: Always import something from test_support like so:
184 from test_support import verbose
189 ... use test_support.verbose in the code ...
191 Never import anything from test_support like this:
193 from test.test_support import verbose
195 "test" is a package already, so can refer to modules it contains without
196 "test." qualification. If you do an explicit "test.xxx" qualification, that
197 can fool Python into believing test.xxx is a module distinct from the xxx
198 in the current package, and you can end up importing two distinct copies of
199 xxx. This is especially bad if xxx=test_support, as regrtest.py can (and
200 routinely does) overwrite its "verbose" and "use_large_resources"
201 attributes: if you get a second copy of test_support loaded, it may not
202 have the same values for those as regrtest intended.
205 Python and C statement coverage results are currently available at
207 http://www.musi-cal.com/~skip/python/Python/dist/src/
209 As of this writing (July, 2000) these results are being generated nightly.
210 You can refer to the summaries and the test coverage output files to see
211 where coverage is adequate or lacking and write test cases to beef up the