1lit - LLVM Integrated Tester 2============================ 3 4SYNOPSIS 5-------- 6 7:program:`lit` [*options*] [*tests*] 8 9DESCRIPTION 10----------- 11 12:program:`lit` is a portable tool for executing LLVM and Clang style test 13suites, summarizing their results, and providing indication of failures. 14:program:`lit` is designed to be a lightweight testing tool with as simple a 15user interface as possible. 16 17:program:`lit` should be run with one or more *tests* to run specified on the 18command line. Tests can be either individual test files or directories to 19search for tests (see :ref:`test-discovery`). 20 21Each specified test will be executed (potentially in parallel) and once all 22tests have been run :program:`lit` will print summary information on the number 23of tests which passed or failed (see :ref:`test-status-results`). The 24:program:`lit` program will execute with a non-zero exit code if any tests 25fail. 26 27By default :program:`lit` will use a succinct progress display and will only 28print summary information for test failures. See :ref:`output-options` for 29options controlling the :program:`lit` progress display and output. 30 31:program:`lit` also includes a number of options for controlling how tests are 32executed (specific features may depend on the particular test format). See 33:ref:`execution-options` for more information. 34 35Finally, :program:`lit` also supports additional options for only running a 36subset of the options specified on the command line, see 37:ref:`selection-options` for more information. 38 39Users interested in the :program:`lit` architecture or designing a 40:program:`lit` testing implementation should see :ref:`lit-infrastructure`. 41 42GENERAL OPTIONS 43--------------- 44 45.. option:: -h, --help 46 47 Show the :program:`lit` help message. 48 49.. option:: -j N, --threads=N 50 51 Run ``N`` tests in parallel. By default, this is automatically chosen to 52 match the number of detected available CPUs. 53 54.. option:: --config-prefix=NAME 55 56 Search for :file:`{NAME}.cfg` and :file:`{NAME}.site.cfg` when searching for 57 test suites, instead of :file:`lit.cfg` and :file:`lit.site.cfg`. 58 59.. option:: --param NAME, --param NAME=VALUE 60 61 Add a user defined parameter ``NAME`` with the given ``VALUE`` (or the empty 62 string if not given). The meaning and use of these parameters is test suite 63 dependent. 64 65.. _output-options: 66 67OUTPUT OPTIONS 68-------------- 69 70.. option:: -q, --quiet 71 72 Suppress any output except for test failures. 73 74.. option:: -s, --succinct 75 76 Show less output, for example don't show information on tests that pass. 77 78.. option:: -v, --verbose 79 80 Show more information on test failures, for example the entire test output 81 instead of just the test result. 82 83.. option:: --no-progress-bar 84 85 Do not use curses based progress bar. 86 87.. _execution-options: 88 89EXECUTION OPTIONS 90----------------- 91 92.. option:: --path=PATH 93 94 Specify an additional ``PATH`` to use when searching for executables in tests. 95 96.. option:: --vg 97 98 Run individual tests under valgrind (using the memcheck tool). The 99 ``--error-exitcode`` argument for valgrind is used so that valgrind failures 100 will cause the program to exit with a non-zero status. 101 102 When this option is enabled, :program:`lit` will also automatically provide a 103 "``valgrind``" feature that can be used to conditionally disable (or expect 104 failure in) certain tests. 105 106.. option:: --vg-arg=ARG 107 108 When :option:`--vg` is used, specify an additional argument to pass to 109 :program:`valgrind` itself. 110 111.. option:: --vg-leak 112 113 When :option:`--vg` is used, enable memory leak checks. When this option is 114 enabled, :program:`lit` will also automatically provide a "``vg_leak``" 115 feature that can be used to conditionally disable (or expect failure in) 116 certain tests. 117 118.. option:: --time-tests 119 120 Track the wall time individual tests take to execute and includes the results 121 in the summary output. This is useful for determining which tests in a test 122 suite take the most time to execute. Note that this option is most useful 123 with ``-j 1``. 124 125.. _selection-options: 126 127SELECTION OPTIONS 128----------------- 129 130.. option:: --max-tests=N 131 132 Run at most ``N`` tests and then terminate. 133 134.. option:: --max-time=N 135 136 Spend at most ``N`` seconds (approximately) running tests and then terminate. 137 138.. option:: --shuffle 139 140 Run the tests in a random order. 141 142ADDITIONAL OPTIONS 143------------------ 144 145.. option:: --debug 146 147 Run :program:`lit` in debug mode, for debugging configuration issues and 148 :program:`lit` itself. 149 150.. option:: --show-suites 151 152 List the discovered test suites and exit. 153 154.. option:: --show-tests 155 156 List all of the the discovered tests and exit. 157 158EXIT STATUS 159----------- 160 161:program:`lit` will exit with an exit code of 1 if there are any FAIL or XPASS 162results. Otherwise, it will exit with the status 0. Other exit codes are used 163for non-test related failures (for example a user error or an internal program 164error). 165 166.. _test-discovery: 167 168TEST DISCOVERY 169-------------- 170 171The inputs passed to :program:`lit` can be either individual tests, or entire 172directories or hierarchies of tests to run. When :program:`lit` starts up, the 173first thing it does is convert the inputs into a complete list of tests to run 174as part of *test discovery*. 175 176In the :program:`lit` model, every test must exist inside some *test suite*. 177:program:`lit` resolves the inputs specified on the command line to test suites 178by searching upwards from the input path until it finds a :file:`lit.cfg` or 179:file:`lit.site.cfg` file. These files serve as both a marker of test suites 180and as configuration files which :program:`lit` loads in order to understand 181how to find and run the tests inside the test suite. 182 183Once :program:`lit` has mapped the inputs into test suites it traverses the 184list of inputs adding tests for individual files and recursively searching for 185tests in directories. 186 187This behavior makes it easy to specify a subset of tests to run, while still 188allowing the test suite configuration to control exactly how tests are 189interpreted. In addition, :program:`lit` always identifies tests by the test 190suite they are in, and their relative path inside the test suite. For 191appropriately configured projects, this allows :program:`lit` to provide 192convenient and flexible support for out-of-tree builds. 193 194.. _test-status-results: 195 196TEST STATUS RESULTS 197------------------- 198 199Each test ultimately produces one of the following six results: 200 201**PASS** 202 203 The test succeeded. 204 205**XFAIL** 206 207 The test failed, but that is expected. This is used for test formats which allow 208 specifying that a test does not currently work, but wish to leave it in the test 209 suite. 210 211**XPASS** 212 213 The test succeeded, but it was expected to fail. This is used for tests which 214 were specified as expected to fail, but are now succeeding (generally because 215 the feature they test was broken and has been fixed). 216 217**FAIL** 218 219 The test failed. 220 221**UNRESOLVED** 222 223 The test result could not be determined. For example, this occurs when the test 224 could not be run, the test itself is invalid, or the test was interrupted. 225 226**UNSUPPORTED** 227 228 The test is not supported in this environment. This is used by test formats 229 which can report unsupported tests. 230 231Depending on the test format tests may produce additional information about 232their status (generally only for failures). See the :ref:`output-options` 233section for more information. 234 235.. _lit-infrastructure: 236 237LIT INFRASTRUCTURE 238------------------ 239 240This section describes the :program:`lit` testing architecture for users interested in 241creating a new :program:`lit` testing implementation, or extending an existing one. 242 243:program:`lit` proper is primarily an infrastructure for discovering and running 244arbitrary tests, and to expose a single convenient interface to these 245tests. :program:`lit` itself doesn't know how to run tests, rather this logic is 246defined by *test suites*. 247 248TEST SUITES 249~~~~~~~~~~~ 250 251As described in :ref:`test-discovery`, tests are always located inside a *test 252suite*. Test suites serve to define the format of the tests they contain, the 253logic for finding those tests, and any additional information to run the tests. 254 255:program:`lit` identifies test suites as directories containing ``lit.cfg`` or 256``lit.site.cfg`` files (see also :option:`--config-prefix`). Test suites are 257initially discovered by recursively searching up the directory hierarchy for 258all the input files passed on the command line. You can use 259:option:`--show-suites` to display the discovered test suites at startup. 260 261Once a test suite is discovered, its config file is loaded. Config files 262themselves are Python modules which will be executed. When the config file is 263executed, two important global variables are predefined: 264 265**lit** 266 267 The global **lit** configuration object (a *LitConfig* instance), which defines 268 the builtin test formats, global configuration parameters, and other helper 269 routines for implementing test configurations. 270 271**config** 272 273 This is the config object (a *TestingConfig* instance) for the test suite, 274 which the config file is expected to populate. The following variables are also 275 available on the *config* object, some of which must be set by the config and 276 others are optional or predefined: 277 278 **name** *[required]* The name of the test suite, for use in reports and 279 diagnostics. 280 281 **test_format** *[required]* The test format object which will be used to 282 discover and run tests in the test suite. Generally this will be a builtin test 283 format available from the *lit.formats* module. 284 285 **test_source_root** The filesystem path to the test suite root. For out-of-dir 286 builds this is the directory that will be scanned for tests. 287 288 **test_exec_root** For out-of-dir builds, the path to the test suite root inside 289 the object directory. This is where tests will be run and temporary output files 290 placed. 291 292 **environment** A dictionary representing the environment to use when executing 293 tests in the suite. 294 295 **suffixes** For **lit** test formats which scan directories for tests, this 296 variable is a list of suffixes to identify test files. Used by: *ShTest*. 297 298 **substitutions** For **lit** test formats which substitute variables into a test 299 script, the list of substitutions to perform. Used by: *ShTest*. 300 301 **unsupported** Mark an unsupported directory, all tests within it will be 302 reported as unsupported. Used by: *ShTest*. 303 304 **parent** The parent configuration, this is the config object for the directory 305 containing the test suite, or None. 306 307 **root** The root configuration. This is the top-most :program:`lit` configuration in 308 the project. 309 310 **on_clone** The config is actually cloned for every subdirectory inside a test 311 suite, to allow local configuration on a per-directory basis. The *on_clone* 312 variable can be set to a Python function which will be called whenever a 313 configuration is cloned (for a subdirectory). The function should takes three 314 arguments: (1) the parent configuration, (2) the new configuration (which the 315 *on_clone* function will generally modify), and (3) the test path to the new 316 directory being scanned. 317 318 **pipefail** Normally a test using a shell pipe fails if any of the commands 319 on the pipe fail. If this is not desired, setting this variable to false 320 makes the test fail only if the last command in the pipe fails. 321 322TEST DISCOVERY 323~~~~~~~~~~~~~~ 324 325Once test suites are located, :program:`lit` recursively traverses the source 326directory (following *test_source_root*) looking for tests. When :program:`lit` 327enters a sub-directory, it first checks to see if a nested test suite is 328defined in that directory. If so, it loads that test suite recursively, 329otherwise it instantiates a local test config for the directory (see 330:ref:`local-configuration-files`). 331 332Tests are identified by the test suite they are contained within, and the 333relative path inside that suite. Note that the relative path may not refer to 334an actual file on disk; some test formats (such as *GoogleTest*) define 335"virtual tests" which have a path that contains both the path to the actual 336test file and a subpath to identify the virtual test. 337 338.. _local-configuration-files: 339 340LOCAL CONFIGURATION FILES 341~~~~~~~~~~~~~~~~~~~~~~~~~ 342 343When :program:`lit` loads a subdirectory in a test suite, it instantiates a 344local test configuration by cloning the configuration for the parent direction 345--- the root of this configuration chain will always be a test suite. Once the 346test configuration is cloned :program:`lit` checks for a *lit.local.cfg* file 347in the subdirectory. If present, this file will be loaded and can be used to 348specialize the configuration for each individual directory. This facility can 349be used to define subdirectories of optional tests, or to change other 350configuration parameters --- for example, to change the test format, or the 351suffixes which identify test files. 352 353TEST RUN OUTPUT FORMAT 354~~~~~~~~~~~~~~~~~~~~~~ 355 356The :program:`lit` output for a test run conforms to the following schema, in 357both short and verbose modes (although in short mode no PASS lines will be 358shown). This schema has been chosen to be relatively easy to reliably parse by 359a machine (for example in buildbot log scraping), and for other tools to 360generate. 361 362Each test result is expected to appear on a line that matches: 363 364.. code-block:: none 365 366 <result code>: <test name> (<progress info>) 367 368where ``<result-code>`` is a standard test result such as PASS, FAIL, XFAIL, 369XPASS, UNRESOLVED, or UNSUPPORTED. The performance result codes of IMPROVED and 370REGRESSED are also allowed. 371 372The ``<test name>`` field can consist of an arbitrary string containing no 373newline. 374 375The ``<progress info>`` field can be used to report progress information such 376as (1/300) or can be empty, but even when empty the parentheses are required. 377 378Each test result may include additional (multiline) log information in the 379following format: 380 381.. code-block:: none 382 383 <log delineator> TEST '(<test name>)' <trailing delineator> 384 ... log message ... 385 <log delineator> 386 387where ``<test name>`` should be the name of a preceding reported test, ``<log 388delineator>`` is a string of "*" characters *at least* four characters long 389(the recommended length is 20), and ``<trailing delineator>`` is an arbitrary 390(unparsed) string. 391 392The following is an example of a test run output which consists of four tests A, 393B, C, and D, and a log message for the failing test C: 394 395.. code-block:: none 396 397 PASS: A (1 of 4) 398 PASS: B (2 of 4) 399 FAIL: C (3 of 4) 400 ******************** TEST 'C' FAILED ******************** 401 Test 'C' failed as a result of exit code 1. 402 ******************** 403 PASS: D (4 of 4) 404 405LIT EXAMPLE TESTS 406~~~~~~~~~~~~~~~~~ 407 408The :program:`lit` distribution contains several example implementations of 409test suites in the *ExampleTests* directory. 410 411SEE ALSO 412-------- 413 414valgrind(1) 415