3 nitunit - executes the unit tests from Nit source files.
7 nitunit [*options*] FILE...
11 Unit testing in Nit can be achieved in two ways:
13 * using `DocUnits` in code comments or in markdown files
14 * using `TestSuites` with test unit files
16 `DocUnits` are executable pieces of code found in the documentation of groups, modules,
17 classes and properties.
18 They are used for documentation purpose, they should be kept simple and illustrative.
19 More advanced unit testing can be done using TestSuites.
21 `DocUnits` can also be used in any markdown files.
23 `TestSuites` are test files coupled to a tested module.
24 They contain a list of test methods called TestCase.
26 ## Working with `DocUnits`
28 DocUnits are blocks of executable code placed in comments of modules, classes and properties.
29 The execution can be verified using `assert`.
36 # assert foo.bar == 10
42 Everything used in the test must be declared.
43 To test a method you have to instantiate its class:
48 # assert foo.bar == 10
51 # assert foo.baz(1, 2) == 3
52 fun baz(a, b: Int) do return a + b
56 In a single piece of documentation, each docunit is considered a part of a single module, thus regrouped when
58 Therefore, it is possible (and recommended) to split docunits in small parts if it make the explanation easier.
61 # Some example of grouped docunits
63 # Declare and initialize a variable `a`.
67 # So the value of `a` can be used
71 # even in complex operations
77 Sometime, some blocks of code has to be included in documentation but not considered by `nitunit`.
78 Those blocks are distinguished by their tagged fences (untagged fences or fences tagged `nit` are considered to be docunits).
90 The special fence-tag `nitish` could also be used to indicate pseudo-nit that will be ignored by nitunit but highlighted by nitdoc.
91 Such `nitish` piece of code can be used to enclose examples that cannot compile or that one do not want to be automatically executed.
97 # var a: Int = someting
99 # if a == 1 then something else something-else
102 # Some code to not try to execute automatically
109 The `nitunit` command is used to test Nit files:
113 Groups (directories) can be given to test the documentation of the group and of all its Nit files:
117 Finally, standard markdown documents can be checked with:
121 When testing, the environment variable `NIT_TESTING` is set to `true`.
122 This flag can be used by libraries and program to prevent (or limit) the execution of dangerous pieces of code.
125 # NIT_TESTING is automatically set.
127 # assert "NIT_TESTING".environ == "true"
130 ## Working with `TestSuites`
132 TestSuites are Nit modules that define a set of TestCases.
134 A test suite is a module that uses the annotation `is test_suite`.
136 It is common that a test suite focuses on testing a single module.
137 In this case, the name of the test_suite is often `test_foo.nit` where `foo.nit` is the tested module.
139 The structure of a test suite is the following:
142 # test suite for module `foo`
143 module test_foo is test_suite
146 import foo # can be intrude to test private things
151 # test case for `foo::Foo::baz`
153 var subject = new Foo
154 assert subject.baz(1, 2) == 3
159 Test suite can be executed using the same `nitunit` command:
163 `nitunit` will execute a test for each method named `test_*` in a class
164 subclassing `TestSuite` so multiple tests can be executed for a single method:
171 var subject = new Foo
172 assert subject.baz(1, 2) == 3
175 var subject = new Foo
176 assert subject.baz(1, -2) == -1
183 Sometimes, it is easier to validate a `TestCase` by comparing its output with a text file containing the expected result.
185 For each TestCase `test_bar` of a TestSuite `test_mod.nit`, a corresponding file with the expected output is looked for:
187 * "test_mod.sav/test_bar.res". I.e. test-cases grouped by test-suites.
189 This is the default and is useful if there is a lot of test-suites and test-cases in a directory
191 * "sav/test_bar.res". I.e. all test-cases grouped in a common sub-directory.
193 Useful if there is a lot of test-suites OR test-cases in a directory.
195 * "test_bar.res" raw in the directory.
197 Useful is there is a few test-suites and test-cases in a directory.
199 All 3 are exclusive. If more than one exists, the test-case is failed.
201 If a corresponding file then the output of the test-case is compared with the file.
203 The `diff(1)` command is used to perform the comparison.
204 The test is failed if non-zero is returned by `diff`.
207 module test_mod is test_suite
218 Where `test_mod.sav/test_bar.res` contains
224 If no corresponding `.res` file exists, then the output of the TestCase is ignored.
226 To helps the management of the expected results, the option `--autosav` can be used to automatically create and update them.
229 ## Configuring TestSuites
231 `TestSuite`s also provide methods to configure the test run:
233 `before_test` and `after_test`: methods called before/after each test case.
234 They can be used to factorize repetitive tasks:
240 # Mandatory empty init
242 # Method executed before each test
247 assert subject.baz(1, 2) == 3
250 assert subject.baz(1, -2) == -1
255 When using custom test attributes, an empty `init` must be declared to allow automatic test running.
257 `before_module` and `after_module`: methods called before/after each test suite.
258 They have to be declared at top level:
261 module test_bdd_connector
263 # Testing the bdd_connector
265 # test cases using a server
267 # Method executed before testing the module
269 # start server before all test cases
271 # Method executed after testing the module
273 # stop server after all test cases
277 ## Generating test suites
279 Write test suites for big modules can be a repetitive and boring task...
280 To make it easier, `nitunit` can generate test skeletons for Nit modules:
282 $ nitunit --gen-suite foo.nit
284 This will generate the test suite `test_foo` containing test case stubs for all public
285 methods found in `foo.nit`.
291 Process also imported modules.
293 By default, only the modules indicated on the command line are tested.
295 With the `--full` option, all imported modules (even those in standard) are also precessed.
298 Output name (default is 'nitunit.xml').
300 `nitunit` produces a XML file compatible with JUnit.
303 Working directory (default is 'nitunit.out').
305 In order to execute the tests, nit files are generated then compiled and executed in the giver working directory.
307 In case of success, the directory is removed.
308 In case of failure, it is kept as is so files can be investigated.
311 nitc compiler to use.
313 By default, nitunit tries to locate the `nitc` program with the environment variable `NITC` or heuristics.
314 The option is used to indicate a specific nitc binary.
317 Does not compile and run tests.
319 ### `-p`, `--pattern`
320 Only run test case with name that match pattern.
322 Examples: `TestFoo`, `TestFoo*`, `TestFoo::test_foo`, `TestFoo::test_foo*`, `test_foo`, `test_foo*`
325 Automatically create/update .res files for black box testing.
327 If a black block test fails because a difference between the expected result and the current result then the expected result file is updated (and the test is passed).
329 If a test-case of a test-suite passes but that some output is generated, then an expected result file is created.
331 It is expected that the created/updated files are checked since the tests are considered passed.
332 A VCS like `git` is often a good tool to check the creation and modification of those files.
337 Generate test suite skeleton for a module.
340 Force test generation even if file exists.
342 Any existing test suite will be overwritten.
345 Also generate test case for private methods.
348 Only display the skeleton, do not write any file.
351 # ENVIRONMENT VARIABLES
355 Indicate the specific Nit compiler executable to use. See `--nitc`.
359 The environment variable `NIT_TESTING` is set to `true` during the execution of program tests.
360 Some libraries of programs can use it to produce specific reproducible results; or just to exit their executions.
362 Unit-tests may unset this environment variable to retrieve the original behavior of such piece of software.
366 In order to maximize reproducibility, `SRAND` is set to 0.
367 This make the pseudo-random generator no random at all.
368 See `Sys::srand` for details.
370 To retrieve the randomness, unit-tests may unset this environment variable then call `srand`.
374 Parallel executions can cause some race collisions on named resources (e.g. DB table names).
375 To solve this issue, `NIT_TESTING_ID` is initialized with a distinct integer identifier that can be used to give unique names to resources.
377 Note: `rand` is not a recommended way to get a distinct identifier because its randomness is disabled by default. See `SRAND`.
382 The Nit language documentation and the source code of its tools and libraries may be downloaded from <http://nitlanguage.org>