xref: /netbsd-src/external/apache2/llvm/dist/llvm/docs/CommandGuide/lit.rst (revision 82d56013d7b633d116a93943de88e08335357a7c)
1lit - LLVM Integrated Tester
2============================
3
4.. program:: lit
5
6SYNOPSIS
7--------
8
9:program:`lit` [*options*] [*tests*]
10
11DESCRIPTION
12-----------
13
14:program:`lit` is a portable tool for executing LLVM and Clang style test
15suites, summarizing their results, and providing indication of failures.
16:program:`lit` is designed to be a lightweight testing tool with as simple a
17user interface as possible.
18
19:program:`lit` should be run with one or more *tests* to run specified on the
20command line.  Tests can be either individual test files or directories to
21search for tests (see :ref:`test-discovery`).
22
23Each specified test will be executed (potentially concurrently) and once all
24tests have been run :program:`lit` will print summary information on the number
25of tests which passed or failed (see :ref:`test-status-results`).  The
26:program:`lit` program will execute with a non-zero exit code if any tests
27fail.
28
29By default :program:`lit` will use a succinct progress display and will only
30print summary information for test failures.  See :ref:`output-options` for
31options controlling the :program:`lit` progress display and output.
32
33:program:`lit` also includes a number of options for controlling how tests are
34executed (specific features may depend on the particular test format).  See
35:ref:`execution-options` for more information.
36
37Finally, :program:`lit` also supports additional options for only running a
38subset of the options specified on the command line, see
39:ref:`selection-options` for more information.
40
41:program:`lit` parses options from the environment variable ``LIT_OPTS`` after
42parsing options from the command line.  ``LIT_OPTS`` is primarily useful for
43supplementing or overriding the command-line options supplied to :program:`lit`
44by ``check`` targets defined by a project's build system.
45
46Users interested in the :program:`lit` architecture or designing a
47:program:`lit` testing implementation should see :ref:`lit-infrastructure`.
48
49GENERAL OPTIONS
50---------------
51
52.. option:: -h, --help
53
54 Show the :program:`lit` help message.
55
56.. option:: -j N, --workers=N
57
58 Run ``N`` tests in parallel.  By default, this is automatically chosen to
59 match the number of detected available CPUs.
60
61.. option:: --config-prefix=NAME
62
63 Search for :file:`{NAME}.cfg` and :file:`{NAME}.site.cfg` when searching for
64 test suites, instead of :file:`lit.cfg` and :file:`lit.site.cfg`.
65
66.. option:: -D NAME[=VALUE], --param NAME[=VALUE]
67
68 Add a user defined parameter ``NAME`` with the given ``VALUE`` (or the empty
69 string if not given).  The meaning and use of these parameters is test suite
70 dependent.
71
72.. _output-options:
73
74OUTPUT OPTIONS
75--------------
76
77.. option:: -q, --quiet
78
79 Suppress any output except for test failures.
80
81.. option:: -s, --succinct
82
83 Show less output, for example don't show information on tests that pass.
84 Also show a progress bar, unless ``--no-progress-bar`` is specified.
85
86.. option:: -v, --verbose
87
88 Show more information on test failures, for example the entire test output
89 instead of just the test result.
90
91.. option:: -vv, --echo-all-commands
92
93 Echo all commands to stdout, as they are being executed.
94 This can be valuable for debugging test failures, as the last echoed command
95 will be the one which has failed.
96 :program:`lit` normally inserts a no-op command (``:`` in the case of bash)
97 with argument ``'RUN: at line N'`` before each command pipeline, and this
98 option also causes those no-op commands to be echoed to stdout to help you
99 locate the source line of the failed command.
100 This option implies ``--verbose``.
101
102.. option:: -a, --show-all
103
104 Show more information about all tests, for example the entire test
105 commandline and output.
106
107.. option:: --no-progress-bar
108
109 Do not use curses based progress bar.
110
111.. option:: --show-unsupported
112
113 Show the names of unsupported tests.
114
115.. option:: --show-xfail
116
117 Show the names of tests that were expected to fail.
118
119.. _execution-options:
120
121EXECUTION OPTIONS
122-----------------
123
124.. option:: --path=PATH
125
126 Specify an additional ``PATH`` to use when searching for executables in tests.
127
128.. option:: --vg
129
130 Run individual tests under valgrind (using the memcheck tool).  The
131 ``--error-exitcode`` argument for valgrind is used so that valgrind failures
132 will cause the program to exit with a non-zero status.
133
134 When this option is enabled, :program:`lit` will also automatically provide a
135 "``valgrind``" feature that can be used to conditionally disable (or expect
136 failure in) certain tests.
137
138.. option:: --vg-arg=ARG
139
140 When :option:`--vg` is used, specify an additional argument to pass to
141 :program:`valgrind` itself.
142
143.. option:: --vg-leak
144
145 When :option:`--vg` is used, enable memory leak checks.  When this option is
146 enabled, :program:`lit` will also automatically provide a "``vg_leak``"
147 feature that can be used to conditionally disable (or expect failure in)
148 certain tests.
149
150.. option:: --time-tests
151
152 Track the wall time individual tests take to execute and includes the results
153 in the summary output.  This is useful for determining which tests in a test
154 suite take the most time to execute.
155
156.. option:: --ignore-fail
157
158 Exit with status zero even if some tests fail.
159
160.. option:: --no-indirectly-run-check
161
162 Do not error if a test would not be run if the user had specified the
163 containing directory instead of naming the test directly.
164
165.. _selection-options:
166
167SELECTION OPTIONS
168-----------------
169
170By default, `lit` will run failing tests first, then run tests in descending
171execution time order to optimize concurrency.
172
173The timing data is stored in the `test_exec_root` in a file named
174`.lit_test_times.txt`. If this file does not exist, then `lit` checks the
175`test_source_root` for the file to optionally accelerate clean builds.
176
177.. option:: --shuffle
178
179 Run the tests in a random order, not failing/slowest first.
180
181.. option:: --max-failures N
182
183 Stop execution after the given number ``N`` of failures.
184 An integer argument should be passed on the command line
185 prior to execution.
186
187.. option:: --max-tests=N
188
189 Run at most ``N`` tests and then terminate.
190
191.. option:: --max-time=N
192
193 Spend at most ``N`` seconds (approximately) running tests and then terminate.
194 Note that this is not an alias for :option:`--timeout`; the two are
195 different kinds of maximums.
196
197.. option:: --num-shards=M
198
199 Divide the set of selected tests into ``M`` equal-sized subsets or
200 "shards", and run only one of them.  Must be used with the
201 ``--run-shard=N`` option, which selects the shard to run. The environment
202 variable ``LIT_NUM_SHARDS`` can also be used in place of this
203 option. These two options provide a coarse mechanism for partitioning large
204 testsuites, for parallel execution on separate machines (say in a large
205 testing farm).
206
207.. option:: --run-shard=N
208
209 Select which shard to run, assuming the ``--num-shards=M`` option was
210 provided. The two options must be used together, and the value of ``N``
211 must be in the range ``1..M``. The environment variable
212 ``LIT_RUN_SHARD`` can also be used in place of this option.
213
214.. option:: --timeout=N
215
216 Spend at most ``N`` seconds (approximately) running each individual test.
217 ``0`` means no time limit, and ``0`` is the default. Note that this is not an
218 alias for :option:`--max-time`; the two are different kinds of maximums.
219
220.. option:: --filter=REGEXP
221
222  Run only those tests whose name matches the regular expression specified in
223  ``REGEXP``. The environment variable ``LIT_FILTER`` can be also used in place
224  of this option, which is especially useful in environments where the call
225  to ``lit`` is issued indirectly.
226
227.. option:: --filter-out=REGEXP
228
229  Filter out those tests whose name matches the regular expression specified in
230  ``REGEXP``. The environment variable ``LIT_FILTER_OUT`` can be also used in
231  place of this option, which is especially useful in environments where the
232  call to ``lit`` is issued indirectly.
233
234.. option:: --xfail=LIST
235
236  Treat those tests whose name is in the semicolon separated list ``LIST`` as
237  ``XFAIL``. This can be helpful when one does not want to modify the test
238  suite. The environment variable ``LIT_XFAIL`` can be also used in place of
239  this option, which is especially useful in environments where the call to
240  ``lit`` is issued indirectly.
241
242ADDITIONAL OPTIONS
243------------------
244
245.. option:: --debug
246
247 Run :program:`lit` in debug mode, for debugging configuration issues and
248 :program:`lit` itself.
249
250.. option:: --show-suites
251
252 List the discovered test suites and exit.
253
254.. option:: --show-tests
255
256 List all of the discovered tests and exit.
257
258EXIT STATUS
259-----------
260
261:program:`lit` will exit with an exit code of 1 if there are any FAIL or XPASS
262results.  Otherwise, it will exit with the status 0.  Other exit codes are used
263for non-test related failures (for example a user error or an internal program
264error).
265
266.. _test-discovery:
267
268TEST DISCOVERY
269--------------
270
271The inputs passed to :program:`lit` can be either individual tests, or entire
272directories or hierarchies of tests to run.  When :program:`lit` starts up, the
273first thing it does is convert the inputs into a complete list of tests to run
274as part of *test discovery*.
275
276In the :program:`lit` model, every test must exist inside some *test suite*.
277:program:`lit` resolves the inputs specified on the command line to test suites
278by searching upwards from the input path until it finds a :file:`lit.cfg` or
279:file:`lit.site.cfg` file.  These files serve as both a marker of test suites
280and as configuration files which :program:`lit` loads in order to understand
281how to find and run the tests inside the test suite.
282
283Once :program:`lit` has mapped the inputs into test suites it traverses the
284list of inputs adding tests for individual files and recursively searching for
285tests in directories.
286
287This behavior makes it easy to specify a subset of tests to run, while still
288allowing the test suite configuration to control exactly how tests are
289interpreted.  In addition, :program:`lit` always identifies tests by the test
290suite they are in, and their relative path inside the test suite.  For
291appropriately configured projects, this allows :program:`lit` to provide
292convenient and flexible support for out-of-tree builds.
293
294.. _test-status-results:
295
296TEST STATUS RESULTS
297-------------------
298
299Each test ultimately produces one of the following eight results:
300
301**PASS**
302
303 The test succeeded.
304
305**FLAKYPASS**
306
307 The test succeeded after being re-run more than once. This only applies to
308 tests containing an ``ALLOW_RETRIES:`` annotation.
309
310**XFAIL**
311
312 The test failed, but that is expected.  This is used for test formats which allow
313 specifying that a test does not currently work, but wish to leave it in the test
314 suite.
315
316**XPASS**
317
318 The test succeeded, but it was expected to fail.  This is used for tests which
319 were specified as expected to fail, but are now succeeding (generally because
320 the feature they test was broken and has been fixed).
321
322**FAIL**
323
324 The test failed.
325
326**UNRESOLVED**
327
328 The test result could not be determined.  For example, this occurs when the test
329 could not be run, the test itself is invalid, or the test was interrupted.
330
331**UNSUPPORTED**
332
333 The test is not supported in this environment.  This is used by test formats
334 which can report unsupported tests.
335
336**TIMEOUT**
337
338 The test was run, but it timed out before it was able to complete. This is
339 considered a failure.
340
341Depending on the test format tests may produce additional information about
342their status (generally only for failures).  See the :ref:`output-options`
343section for more information.
344
345.. _lit-infrastructure:
346
347LIT INFRASTRUCTURE
348------------------
349
350This section describes the :program:`lit` testing architecture for users interested in
351creating a new :program:`lit` testing implementation, or extending an existing one.
352
353:program:`lit` proper is primarily an infrastructure for discovering and running
354arbitrary tests, and to expose a single convenient interface to these
355tests. :program:`lit` itself doesn't know how to run tests, rather this logic is
356defined by *test suites*.
357
358TEST SUITES
359~~~~~~~~~~~
360
361As described in :ref:`test-discovery`, tests are always located inside a *test
362suite*.  Test suites serve to define the format of the tests they contain, the
363logic for finding those tests, and any additional information to run the tests.
364
365:program:`lit` identifies test suites as directories containing ``lit.cfg`` or
366``lit.site.cfg`` files (see also :option:`--config-prefix`).  Test suites are
367initially discovered by recursively searching up the directory hierarchy for
368all the input files passed on the command line.  You can use
369:option:`--show-suites` to display the discovered test suites at startup.
370
371Once a test suite is discovered, its config file is loaded.  Config files
372themselves are Python modules which will be executed.  When the config file is
373executed, two important global variables are predefined:
374
375**lit_config**
376
377 The global **lit** configuration object (a *LitConfig* instance), which defines
378 the builtin test formats, global configuration parameters, and other helper
379 routines for implementing test configurations.
380
381**config**
382
383 This is the config object (a *TestingConfig* instance) for the test suite,
384 which the config file is expected to populate.  The following variables are also
385 available on the *config* object, some of which must be set by the config and
386 others are optional or predefined:
387
388 **name** *[required]* The name of the test suite, for use in reports and
389 diagnostics.
390
391 **test_format** *[required]* The test format object which will be used to
392 discover and run tests in the test suite.  Generally this will be a builtin test
393 format available from the *lit.formats* module.
394
395 **test_source_root** The filesystem path to the test suite root.  For out-of-dir
396 builds this is the directory that will be scanned for tests.
397
398 **test_exec_root** For out-of-dir builds, the path to the test suite root inside
399 the object directory.  This is where tests will be run and temporary output files
400 placed.
401
402 **environment** A dictionary representing the environment to use when executing
403 tests in the suite.
404
405 **standalone_tests** When true, mark a directory with tests expected to be run
406 standalone. Test discovery is disabled for that directory and
407 *--no-indirectly-run-check* is in effect. *lit.suffixes* and *lit.excludes*
408 must be empty when this variable is true.
409
410 **suffixes** For **lit** test formats which scan directories for tests, this
411 variable is a list of suffixes to identify test files.  Used by: *ShTest*.
412
413 **substitutions** For **lit** test formats which substitute variables into a test
414 script, the list of substitutions to perform.  Used by: *ShTest*.
415
416 **unsupported** Mark an unsupported directory, all tests within it will be
417 reported as unsupported.  Used by: *ShTest*.
418
419 **parent** The parent configuration, this is the config object for the directory
420 containing the test suite, or None.
421
422 **root** The root configuration.  This is the top-most :program:`lit` configuration in
423 the project.
424
425 **pipefail** Normally a test using a shell pipe fails if any of the commands
426 on the pipe fail. If this is not desired, setting this variable to false
427 makes the test fail only if the last command in the pipe fails.
428
429 **available_features** A set of features that can be used in `XFAIL`,
430 `REQUIRES`, and `UNSUPPORTED` directives.
431
432TEST DISCOVERY
433~~~~~~~~~~~~~~
434
435Once test suites are located, :program:`lit` recursively traverses the source
436directory (following *test_source_root*) looking for tests.  When :program:`lit`
437enters a sub-directory, it first checks to see if a nested test suite is
438defined in that directory.  If so, it loads that test suite recursively,
439otherwise it instantiates a local test config for the directory (see
440:ref:`local-configuration-files`).
441
442Tests are identified by the test suite they are contained within, and the
443relative path inside that suite.  Note that the relative path may not refer to
444an actual file on disk; some test formats (such as *GoogleTest*) define
445"virtual tests" which have a path that contains both the path to the actual
446test file and a subpath to identify the virtual test.
447
448.. _local-configuration-files:
449
450LOCAL CONFIGURATION FILES
451~~~~~~~~~~~~~~~~~~~~~~~~~
452
453When :program:`lit` loads a subdirectory in a test suite, it instantiates a
454local test configuration by cloning the configuration for the parent directory
455--- the root of this configuration chain will always be a test suite.  Once the
456test configuration is cloned :program:`lit` checks for a *lit.local.cfg* file
457in the subdirectory.  If present, this file will be loaded and can be used to
458specialize the configuration for each individual directory.  This facility can
459be used to define subdirectories of optional tests, or to change other
460configuration parameters --- for example, to change the test format, or the
461suffixes which identify test files.
462
463SUBSTITUTIONS
464~~~~~~~~~~~~~
465
466:program:`lit` allows patterns to be substituted inside RUN commands. It also
467provides the following base set of substitutions, which are defined in
468TestRunner.py:
469
470 ======================= ==============
471  Macro                   Substitution
472 ======================= ==============
473 %s                      source path (path to the file currently being run)
474 %S                      source dir (directory of the file currently being run)
475 %p                      same as %S
476 %{pathsep}              path separator
477 %t                      temporary file name unique to the test
478 %basename_t             The last path component of %t but without the ``.tmp`` extension
479 %T                      parent directory of %t (not unique, deprecated, do not use)
480 %%                      %
481 %/s                     %s but ``\`` is replaced by ``/``
482 %/S                     %S but ``\`` is replaced by ``/``
483 %/p                     %p but ``\`` is replaced by ``/``
484 %/t                     %t but ``\`` is replaced by ``/``
485 %/T                     %T but ``\`` is replaced by ``/``
486 %{/s:regex_replacement} %/s but escaped for use in the replacement of a ``s@@@`` command in sed
487 %{/S:regex_replacement} %/S but escaped for use in the replacement of a ``s@@@`` command in sed
488 %{/p:regex_replacement} %/p but escaped for use in the replacement of a ``s@@@`` command in sed
489 %{/t:regex_replacement} %/t but escaped for use in the replacement of a ``s@@@`` command in sed
490 %{/T:regex_replacement} %/T but escaped for use in the replacement of a ``s@@@`` command in sed
491 %:s                     On Windows, %/s but a ``:`` is removed if its the second character.
492                         Otherwise, %s but with a single leading ``/`` removed.
493 %:S                     On Windows, %/S but a ``:`` is removed if its the second character.
494                         Otherwise, %S but with a single leading ``/`` removed.
495 %:p                     On Windows, %/p but a ``:`` is removed if its the second character.
496                         Otherwise, %p but with a single leading ``/`` removed.
497 %:t                     On Windows, %/t but a ``:`` is removed if its the second character.
498                         Otherwise, %t but with a single leading ``/`` removed.
499 %:T                     On Windows, %/T but a ``:`` is removed if its the second character.
500                         Otherwise, %T but with a single leading ``/`` removed.
501 ======================= ==============
502
503Other substitutions are provided that are variations on this base set and
504further substitution patterns can be defined by each test module. See the
505modules :ref:`local-configuration-files`.
506
507By default, substitutions are expanded exactly once, so that if e.g. a
508substitution ``%build`` is defined in top of another substitution ``%cxx``,
509``%build`` will expand to ``%cxx`` textually, not to what ``%cxx`` expands to.
510However, if the ``recursiveExpansionLimit`` property of the ``TestingConfig``
511is set to a non-negative integer, substitutions will be expanded recursively
512until that limit is reached. It is an error if the limit is reached and
513expanding substitutions again would yield a different result.
514
515More detailed information on substitutions can be found in the
516:doc:`../TestingGuide`.
517
518TEST RUN OUTPUT FORMAT
519~~~~~~~~~~~~~~~~~~~~~~
520
521The :program:`lit` output for a test run conforms to the following schema, in
522both short and verbose modes (although in short mode no PASS lines will be
523shown).  This schema has been chosen to be relatively easy to reliably parse by
524a machine (for example in buildbot log scraping), and for other tools to
525generate.
526
527Each test result is expected to appear on a line that matches:
528
529.. code-block:: none
530
531  <result code>: <test name> (<progress info>)
532
533where ``<result-code>`` is a standard test result such as PASS, FAIL, XFAIL,
534XPASS, UNRESOLVED, or UNSUPPORTED.  The performance result codes of IMPROVED and
535REGRESSED are also allowed.
536
537The ``<test name>`` field can consist of an arbitrary string containing no
538newline.
539
540The ``<progress info>`` field can be used to report progress information such
541as (1/300) or can be empty, but even when empty the parentheses are required.
542
543Each test result may include additional (multiline) log information in the
544following format:
545
546.. code-block:: none
547
548  <log delineator> TEST '(<test name>)' <trailing delineator>
549  ... log message ...
550  <log delineator>
551
552where ``<test name>`` should be the name of a preceding reported test, ``<log
553delineator>`` is a string of "*" characters *at least* four characters long
554(the recommended length is 20), and ``<trailing delineator>`` is an arbitrary
555(unparsed) string.
556
557The following is an example of a test run output which consists of four tests A,
558B, C, and D, and a log message for the failing test C:
559
560.. code-block:: none
561
562  PASS: A (1 of 4)
563  PASS: B (2 of 4)
564  FAIL: C (3 of 4)
565  ******************** TEST 'C' FAILED ********************
566  Test 'C' failed as a result of exit code 1.
567  ********************
568  PASS: D (4 of 4)
569
570LIT EXAMPLE TESTS
571~~~~~~~~~~~~~~~~~
572
573The :program:`lit` distribution contains several example implementations of
574test suites in the *ExampleTests* directory.
575
576SEE ALSO
577--------
578
579valgrind(1)
580