4 This directory holds many test scripts for core GIT tools. The
5 first part of this short document describes how to run the tests
8 When fixing the tools or adding enhancements, you are strongly
9 encouraged to add tests in this directory to cover what you are
10 trying to fix or enhance. The later part of this short document
11 describes how your test scripts should be organized.
17 The easiest way to run tests is to say "make". This runs all
20 *** t0000-basic.sh ***
21 ok 1 - .git/objects should be empty after git init in an empty repo.
22 ok 2 - .git/objects should have 3 subdirectories.
23 ok 3 - success is reported like this
25 ok 43 - very long name in the index handled sanely
26 # fixed 1 known breakage(s)
27 # still have 1 known breakage(s)
28 # passed all remaining 42 test(s)
32 ok 2 - plain with GIT_WORK_TREE
35 Since the tests all output TAP (see http://testanything.org) they can
36 be run with any TAP harness. Here's an example of parallel testing
37 powered by a recent version of prove(1):
39 $ prove --timer --jobs 15 ./t[0-9]*.sh
40 [19:17:33] ./t0005-signals.sh ................................... ok 36 ms
41 [19:17:33] ./t0022-crlf-rename.sh ............................... ok 69 ms
42 [19:17:33] ./t0024-crlf-archive.sh .............................. ok 154 ms
43 [19:17:33] ./t0004-unwritable.sh ................................ ok 289 ms
44 [19:17:33] ./t0002-gitfile.sh ................................... ok 480 ms
45 ===( 102;0 25/? 6/? 5/? 16/? 1/? 4/? 2/? 1/? 3/? 1... )===
47 prove and other harnesses come with a lot of useful options. The
48 --state option in particular is very useful:
50 # Repeat until no more failures
51 $ prove -j 15 --state=failed,save ./t[0-9]*.sh
53 You can give DEFAULT_TEST_TARGET=prove on the make command (or define it
54 in config.mak) to cause "make test" to run tests under prove.
55 GIT_PROVE_OPTS can be used to pass additional options, e.g.
57 $ make DEFAULT_TEST_TARGET=prove GIT_PROVE_OPTS='--timer --jobs 16' test
59 You can also run each test individually from command line, like this:
61 $ sh ./t3010-ls-files-killed-modified.sh
62 ok 1 - git update-index --add to add various paths.
63 ok 2 - git ls-files -k to show killed files.
64 ok 3 - validate git ls-files -k output.
65 ok 4 - git ls-files -m to show modified files.
66 ok 5 - validate git ls-files -m output.
67 # passed all 5 test(s)
70 You can pass --verbose (or -v), --debug (or -d), and --immediate
71 (or -i) command line argument to the test, or by setting GIT_TEST_OPTS
72 appropriately before running "make".
76 This makes the test more verbose. Specifically, the
77 command being run and their output if any are also
80 --verbose-only=<pattern>::
81 Like --verbose, but the effect is limited to tests with
82 numbers matching <pattern>. The number matched against is
83 simply the running count of the test within the file.
86 Turn on shell tracing (i.e., `set -x`) during the tests
87 themselves. Implies `--verbose`.
88 Ignored in test scripts that set the variable 'test_untraceable'
89 to a non-empty value, unless it's run with a Bash version
90 supporting BASH_XTRACEFD, i.e. v4.1 or later.
94 This may help the person who is developing a new test.
95 It causes the command defined with test_debug to run.
96 The "trash" directory (used to store all temporary data
97 during testing) is not deleted even if there are no
98 failed tests so that you can inspect its contents after
103 This causes the test to immediately exit upon the first
104 failed test. Cleanup commands requested with
105 test_when_finished are not executed if the test failed,
106 in order to keep the state for inspection by the tester
111 This causes additional long-running tests to be run (where
112 available), for more exhaustive testing.
115 --run=<test-selector>::
116 Run only the subset of tests indicated by
117 <test-selector>. See section "Skipping Tests" below for
118 <test-selector> syntax.
121 Execute all Git binaries under valgrind tool <tool> and exit
122 with status 126 on errors (just like regular tests, this will
123 only stop the test script when running under -i).
125 Since it makes no sense to run the tests with --valgrind and
126 not see any output, this option implies --verbose. For
127 convenience, it also implies --tee.
129 <tool> defaults to 'memcheck', just like valgrind itself.
130 Other particularly useful choices include 'helgrind' and
131 'drd', but you may use any tool recognized by your valgrind
134 As a special case, <tool> can be 'memcheck-fast', which uses
135 memcheck but disables --track-origins. Use this if you are
136 running tests in bulk, to see if there are _any_ memory
139 Note that memcheck is run with the option --leak-check=no,
140 as the git process is short-lived and some errors are not
141 interesting. In order to run a single command under the same
142 conditions manually, you should set GIT_VALGRIND to point to
143 the 't/valgrind/' directory and use the commands under
146 --valgrind-only=<pattern>::
147 Like --valgrind, but the effect is limited to tests with
148 numbers matching <pattern>. The number matched against is
149 simply the running count of the test within the file.
152 In addition to printing the test output to the terminal,
153 write it to files named 't/test-results/$TEST_NAME.out'.
154 As the names depend on the tests' file names, it is safe to
155 run the tests with this option in parallel.
159 Write verbose output to the same logfile as `--tee`, but do
160 _not_ write it to stdout. Unlike `--tee --verbose`, this option
161 is safe to use when stdout is being consumed by a TAP parser
162 like `prove`. Implies `--tee` and `--verbose`.
165 By default tests are run without dashed forms of
166 commands (like git-commit) in the PATH (it only uses
167 wrappers from ../bin-wrappers). Use this option to include
168 the build directory (..) in the PATH, which contains all
169 the dashed forms of commands. This option is currently
170 implied by other options like --valgrind and
174 By default, the test suite uses the wrappers in
175 `../bin-wrappers/` to execute `git` and friends. With this option,
176 `../git` and friends are run directly. This is not recommended
177 in general, as the wrappers contain safeguards to ensure that no
178 files from an installed Git are used, but can speed up test runs
179 especially on platforms where running shell scripts is expensive
180 (most notably, Windows).
183 Create "trash" directories used to store all temporary data during
184 testing under <directory>, instead of the t/ directory.
185 Using this option with a RAM-based filesystem (such as tmpfs)
186 can massively speed up the test suite.
190 If --chain-lint is enabled, the test harness will check each
191 test to make sure that it properly "&&-chains" all commands (so
192 that a failure in the middle does not go unnoticed by the final
193 exit code of the test). This check is performed in addition to
194 running the tests themselves. You may also enable or disable
195 this feature by setting the GIT_TEST_CHAIN_LINT environment
196 variable to "1" or "0", respectively.
199 Run the test script repeatedly in multiple parallel jobs until
200 one of them fails. Useful for reproducing rare failures in
201 flaky tests. The number of parallel jobs is, in order of
202 precedence: the value of the GIT_TEST_STRESS_LOAD
203 environment variable, or twice the number of available
204 processors (as shown by the 'getconf' utility), or 8.
205 Implies `--verbose -x --immediate` to get the most information
206 about the failure. Note that the verbose output of each test
207 job is saved to 't/test-results/$TEST_NAME.stress-<nr>.out',
208 and only the output of the failed test job is shown on the
209 terminal. The names of the trash directories get a
210 '.stress-<nr>' suffix, and the trash directory of the failed
211 test job is renamed to end with a '.stress-failed' suffix.
214 Override the number of parallel jobs. Implies `--stress`.
217 When combined with --stress run the test script repeatedly
218 this many times in each of the parallel jobs or until one of
219 them fails, whichever comes first. Implies `--stress`.
221 You can also set the GIT_TEST_INSTALLED environment variable to
222 the bindir of an existing git installation to test that installation.
223 You still need to have built this git sandbox, from which various
224 test-* support programs, templates, and perl libraries are used.
225 If your installed git is incomplete, it will silently test parts of
226 your built version instead.
228 When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
229 override the location of the dashed-form subcommands (what
230 GIT_EXEC_PATH would be used for during normal operation).
231 GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
237 In some environments, certain tests have no way of succeeding
238 due to platform limitation, such as lack of 'unzip' program, or
239 filesystem that do not allow arbitrary sequence of non-NUL bytes
242 You should be able to say something like
244 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
248 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
250 to omit such tests. The value of the environment variable is a
251 SP separated list of patterns that tells which tests to skip,
252 and either can match the "t[0-9]{4}" part to skip the whole
253 test, or t[0-9]{4} followed by ".$number" to say which
254 particular test to skip.
256 For an individual test suite --run could be used to specify that
257 only some tests should be run or that some tests should be
260 The argument for --run is a list of individual test numbers or
261 ranges with an optional negation prefix that define what tests in
262 a test suite to include in the run. A range is two numbers
263 separated with a dash and matches a range of tests with both ends
264 been included. You may omit the first or the second number to
265 mean "from the first test" or "up to the very last test"
268 Optional prefix of '!' means that the test or a range of tests
269 should be excluded from the run.
271 If --run starts with an unprefixed number or range the initial
272 set of tests to run is empty. If the first item starts with '!'
273 all the tests are added to the initial set. After initial set is
274 determined every test number or range is added or excluded from
275 the set one by one, from left to right.
277 Individual numbers or ranges could be separated either by a space
280 For example, to run only tests up to a specific test (21), one
283 $ sh ./t9200-git-cvsexport-commit.sh --run='1-21'
287 $ sh ./t9200-git-cvsexport-commit.sh --run='-21'
289 Common case is to run several setup tests (1, 2, 3) and then a
290 specific test (21) that relies on that setup:
292 $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21'
296 $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21
300 $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21'
302 As noted above, the test set is built by going through the items
303 from left to right, so this:
305 $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3'
307 will run tests 1, 2, and 4. Items that come later have higher
308 precedence. It means that this:
310 $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4'
312 would just run tests from 1 to 4, including 3.
314 You may use negation with ranges. The following will run all
315 test in the test suite except from 7 up to 11:
317 $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11'
319 Some tests in a test suite rely on the previous tests performing
320 certain actions, specifically some tests are designated as
321 "setup" test, so you cannot _arbitrarily_ disable one test and
322 expect the rest to function correctly.
324 --run is mostly useful when you want to focus on a specific test
325 and know what setup is needed for it. Or when you want to run
326 everything up to a certain test.
329 Running tests with special setups
330 ---------------------------------
332 The whole test suite could be run to test some special features
333 that cannot be easily covered by a few specific test cases. These
334 could be enabled by running the test suite with correct GIT_TEST_
337 GIT_TEST_FAIL_PREREQS=<boolean> fails all prerequisites. This is
338 useful for discovering issues with the tests where say a later test
339 implicitly depends on an optional earlier test.
341 There's a "FAIL_PREREQS" prerequisite that can be used to test for
342 whether this mode is active, and e.g. skip some tests that are hard to
343 refactor to deal with it. The "SYMLINKS" prerequisite is currently
344 excluded as so much relies on it, but this might change in the future.
346 GIT_TEST_GETTEXT_POISON=<boolean> turns all strings marked for
347 translation into gibberish if true. Used for spotting those tests that
348 need to be marked with a C_LOCALE_OUTPUT prerequisite when adding more
349 strings for translation. See "Testing marked strings" in po/README for
352 GIT_TEST_SPLIT_INDEX=<boolean> forces split-index mode on the whole
353 test suite. Accept any boolean values that are accepted by git-config.
355 GIT_TEST_PROTOCOL_VERSION=<n>, when set, makes 'protocol.version'
358 GIT_TEST_FULL_IN_PACK_ARRAY=<boolean> exercises the uncommon
359 pack-objects code path where there are more than 1024 packs even if
360 the actual number of packs in repository is below this limit. Accept
361 any boolean values that are accepted by git-config.
363 GIT_TEST_OE_SIZE=<n> exercises the uncommon pack-objects code path
364 where we do not cache object size in memory and read it from existing
365 packs on demand. This normally only happens when the object size is
366 over 2GB. This variable forces the code path on any object larger than
369 GIT_TEST_OE_DELTA_SIZE=<n> exercises the uncommon pack-objects code
370 path where deltas larger than this limit require extra memory
371 allocation for bookkeeping.
373 GIT_TEST_VALIDATE_INDEX_CACHE_ENTRIES=<boolean> checks that cache-tree
374 records are valid when the index is written out or after a merge. This
375 is mostly to catch missing invalidation. Default is true.
377 GIT_TEST_COMMIT_GRAPH=<boolean>, when true, forces the commit-graph to
378 be written after every 'git commit' command, and overrides the
379 'core.commitGraph' setting to true.
381 GIT_TEST_FSMONITOR=$PWD/t7519/fsmonitor-all exercises the fsmonitor
382 code path for utilizing a file system monitor to speed up detecting
383 new or changed files.
385 GIT_TEST_INDEX_VERSION=<n> exercises the index read/write code path
386 for the index version specified. Can be set to any valid version
387 (currently 2, 3, or 4).
389 GIT_TEST_PACK_SPARSE=<boolean> if enabled will default the pack-objects
390 builtin to use the sparse object walk. This can still be overridden by
391 the --no-sparse command-line argument.
393 GIT_TEST_PRELOAD_INDEX=<boolean> exercises the preload-index code path
394 by overriding the minimum number of cache entries required per thread.
396 GIT_TEST_STASH_USE_BUILTIN=<boolean>, when false, disables the
397 built-in version of git-stash. See 'stash.useBuiltin' in
400 GIT_TEST_ADD_I_USE_BUILTIN=<boolean>, when true, enables the
401 built-in version of git add -i. See 'add.interactive.useBuiltin' in
404 GIT_TEST_INDEX_THREADS=<n> enables exercising the multi-threaded loading
405 of the index for the whole test suite by bypassing the default number of
406 cache entries and thread minimums. Setting this to 1 will make the
407 index loading single threaded.
409 GIT_TEST_MULTI_PACK_INDEX=<boolean>, when true, forces the multi-pack-
410 index to be written after every 'git repack' command, and overrides the
411 'core.multiPackIndex' setting to true.
413 GIT_TEST_SIDEBAND_ALL=<boolean>, when true, overrides the
414 'uploadpack.allowSidebandAll' setting to true, and when false, forces
415 fetch-pack to not request sideband-all (even if the server advertises
418 GIT_TEST_DISALLOW_ABBREVIATED_OPTIONS=<boolean>, when true (which is
419 the default when running tests), errors out when an abbreviated option
425 The test files are named as:
427 tNNNN-commandname-details.sh
429 where N is a decimal digit.
431 First digit tells the family:
433 0 - the absolute basics and global stuff
434 1 - the basic commands concerning database
435 2 - the basic commands concerning the working tree
436 3 - the other basic commands (e.g. ls-files)
437 4 - the diff commands
438 5 - the pull and exporting commands
439 6 - the revision tree commands (even e.g. merge-base)
440 7 - the porcelainish commands concerning the working tree
441 8 - the porcelainish commands concerning forensics
444 Second digit tells the particular command we are testing.
446 Third digit (optionally) tells the particular switch or group of switches
449 If you create files under t/ directory (i.e. here) that is not
450 the top-level test script, never name the file to match the above
451 pattern. The Makefile here considers all such files as the
452 top-level test script and tries to run all of them. Care is
453 especially needed if you are creating a common test library
454 file, similar to test-lib.sh, because such a library file may
455 not be suitable for standalone execution.
461 The test script is written as a shell script. It should start
462 with the standard "#!/bin/sh", and an
463 assignment to variable 'test_description', like this:
467 test_description='xxx test (option --frotz)
469 This test registers the following structure in the cache
470 and tries to run git-ls-files with option --frotz.'
476 After assigning test_description, the test script should source
477 test-lib.sh like this:
481 This test harness library does the following things:
483 - If the script is invoked with command line argument --help
484 (or -h), it shows the test_description and exits.
486 - Creates an empty test directory with an empty .git/objects database
487 and chdir(2) into it. This directory is 't/trash
488 directory.$test_name_without_dotsh', with t/ subject to change by
489 the --root option documented above, and a '.stress-<N>' suffix
490 appended by the --stress option.
492 - Defines standard test helper functions for your scripts to
493 use. These functions are designed to make all scripts behave
494 consistently when command line arguments --verbose (or -v),
495 --debug (or -d), and --immediate (or -i) is given.
500 Here are a few examples of things you probably should and shouldn't do
505 - Put all code inside test_expect_success and other assertions.
507 Even code that isn't a test per se, but merely some setup code
508 should be inside a test assertion.
510 - Chain your test assertions
512 Write test code like this:
524 That way all of the commands in your tests will succeed or fail. If
525 you must ignore the return value of something, consider using a
526 helper function (e.g. use sane_unset instead of unset, in order
527 to avoid unportable return value for unsetting a variable that was
528 already unset), or prepending the command with test_might_fail or
531 - Check the test coverage for your tests. See the "Test coverage"
534 Don't blindly follow test coverage metrics; if a new function you added
535 doesn't have any coverage, then you're probably doing something wrong,
536 but having 100% coverage doesn't necessarily mean that you tested
539 Tests that are likely to smoke out future regressions are better
540 than tests that just inflate the coverage metrics.
542 - When a test checks for an absolute path that a git command generated,
543 construct the expected value using $(pwd) rather than $PWD,
544 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
545 Windows, where the shell (MSYS bash) mangles absolute path names.
546 For details, see the commit message of 4114156ae9.
548 - Remember that inside the <script> part, the standard output and
549 standard error streams are discarded, and the test harness only
550 reports "ok" or "not ok" to the end user running the tests. Under
551 --verbose, they are shown to help debug the tests.
553 - Be careful when you loop
555 You may need to verify multiple things in a loop, but the
556 following does not work correctly:
558 test_expect_success 'test three things' '
559 for i in one two three
566 Because the status of the loop itself is the exit status of the
567 test_something in the last round, the loop does not fail when
568 "test_something" for "one" or "two" fails. This is not what you
571 Instead, you can break out of the loop immediately when you see a
572 failure. Because all test_expect_* snippets are executed inside
573 a function, "return 1" can be used to fail the test immediately
576 test_expect_success 'test three things' '
577 for i in one two three
579 test_something "$i" || return 1
584 Note that we still &&-chain the loop to propagate failures from
588 And here are the "don'ts:"
590 - Don't exit() within a <script> part.
592 The harness will catch this as a programming error of the test.
593 Use test_done instead if you need to stop the tests early (see
594 "Skipping tests" below).
596 - Don't use '! git cmd' when you want to make sure the git command
597 exits with failure in a controlled way by calling "die()". Instead,
598 use 'test_must_fail git cmd'. This will signal a failure if git
599 dies in an unexpected way (e.g. segfault).
601 On the other hand, don't use test_must_fail for running regular
602 platform commands; just use '! cmd'. We are not in the business
603 of verifying that the world given to us sanely works.
605 - Don't feed the output of a git command to a pipe, as in:
607 git -C repo ls-files |
608 xargs -n 1 basename |
611 which will discard git's exit code and may mask a crash. In the
612 above example, all exit codes are ignored except grep's.
614 Instead, write the output of that command to a temporary
615 file with ">" or assign it to a variable with "x=$(git ...)" rather
618 - Don't use command substitution in a way that discards git's exit
619 code. When assigning to a variable, the exit code is not discarded,
622 x=$(git cat-file -p $sha) &&
625 is OK because a crash in "git cat-file" will cause the "&&" chain
628 test "refs/heads/foo" = "$(git symbolic-ref HEAD)"
630 is not OK and a crash in git could go undetected.
632 - Don't use perl without spelling it as "$PERL_PATH". This is to help
633 our friends on Windows where the platform Perl often adds CR before
634 the end of line, and they bundle Git with a version of Perl that
635 does not do so, whose path is specified with $PERL_PATH. Note that we
636 provide a "perl" function which uses $PERL_PATH under the hood, so
637 you do not need to worry when simply running perl in the test scripts
638 (but you do, for example, on a shebang line or in a sub script
639 created via "write_script").
641 - Don't use sh without spelling it as "$SHELL_PATH", when the script
642 can be misinterpreted by broken platform shell (e.g. Solaris).
644 - Don't chdir around in tests. It is not sufficient to chdir to
645 somewhere and then chdir back to the original location later in
646 the test, as any intermediate step can fail and abort the test,
647 causing the next test to start in an unexpected directory. Do so
648 inside a subshell if necessary.
650 - Don't save and verify the standard error of compound commands, i.e.
651 group commands, subshells, and shell functions (except test helper
652 functions like 'test_must_fail') like this:
654 ( cd dir && git cmd ) 2>error &&
655 test_cmp expect error
657 When running the test with '-x' tracing, then the trace of commands
658 executed in the compound command will be included in standard error
659 as well, quite possibly throwing off the subsequent checks examining
660 the output. Instead, save only the relevant git command's standard
663 ( cd dir && git cmd 2>../error ) &&
664 test_cmp expect error
666 - Don't break the TAP output
668 The raw output from your test may be interpreted by a TAP harness. TAP
669 harnesses will ignore everything they don't know about, but don't step
670 on their toes in these areas:
672 - Don't print lines like "$x..$y" where $x and $y are integers.
674 - Don't print lines that begin with "ok" or "not ok".
676 TAP harnesses expect a line that begins with either "ok" and "not
677 ok" to signal a test passed or failed (and our harness already
678 produces such lines), so your script shouldn't emit such lines to
681 You can glean some further possible issues from the TAP grammar
682 (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR)
683 but the best indication is to just run the tests with prove(1),
684 it'll complain if anything is amiss.
690 If you need to skip tests you should do so by using the three-arg form
691 of the test_* functions (see the "Test harness library" section
694 test_expect_success PERL 'I need Perl' '
695 perl -e "hlagh() if unf_unf()"
698 The advantage of skipping tests like this is that platforms that don't
699 have the PERL and other optional dependencies get an indication of how
700 many tests they're missing.
702 If the test code is too hairy for that (i.e. does a lot of setup work
703 outside test assertions) you can also skip all remaining tests by
704 setting skip_all and immediately call test_done:
706 if ! test_have_prereq PERL
708 skip_all='skipping perl interface tests, perl not available'
712 The string you give to skip_all will be used as an explanation for why
713 the test was skipped.
718 Your script will be a sequence of tests, using helper functions
719 from the test harness library. At the end of the script, call
726 There are a handful helper functions defined in the test harness
727 library for your script to use.
729 - test_expect_success [<prereq>] <message> <script>
731 Usually takes two strings as parameters, and evaluates the
732 <script>. If it yields success, test is considered
733 successful. <message> should state what it is testing.
737 test_expect_success \
738 'git-write-tree should be able to write an empty tree.' \
739 'tree=$(git-write-tree)'
741 If you supply three parameters the first will be taken to be a
742 prerequisite; see the test_set_prereq and test_have_prereq
745 test_expect_success TTY 'git --paginate rev-list uses a pager' \
748 You can also supply a comma-separated list of prerequisites, in the
749 rare case where your test depends on more than one:
751 test_expect_success PERL,PYTHON 'yo dawg' \
752 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" '
754 - test_expect_failure [<prereq>] <message> <script>
756 This is NOT the opposite of test_expect_success, but is used
757 to mark a test that demonstrates a known breakage. Unlike
758 the usual test_expect_success tests, which say "ok" on
759 success and "FAIL" on failure, this will say "FIXED" on
760 success and "still broken" on failure. Failures from these
761 tests won't cause -i (immediate) to stop.
763 Like test_expect_success this function can optionally use a three
764 argument invocation with a prerequisite as the first argument.
766 - test_debug <script>
768 This takes a single argument, <script>, and evaluates it only
769 when the test script is started with --debug command line
770 argument. This is primarily meant for use during the
771 development of a new test script.
773 - debug <git-command>
775 Run a git command inside a debugger. This is primarily meant for
776 use when debugging a failing test script.
780 Your test script must have test_done at the end. Its purpose
781 is to summarize successes and failures in the test script and
782 exit with an appropriate error code.
786 Make commit and tag names consistent by setting the author and
787 committer times to defined state. Subsequent calls will
788 advance the times by a fixed amount.
790 - test_commit <message> [<filename> [<contents>]]
792 Creates a commit with the given message, committing the given
793 file with the given contents (default for both is to reuse the
794 message string), and adds a tag (again reusing the message
795 string as name). Calls test_tick to make the SHA-1s
798 - test_merge <message> <commit-or-tag>
800 Merges the given rev using the given message. Like test_commit,
801 creates a tag and calls test_tick before committing.
803 - test_set_prereq <prereq>
805 Set a test prerequisite to be used later with test_have_prereq. The
806 test-lib will set some prerequisites for you, see the
807 "Prerequisites" section below for a full list of these.
809 Others you can set yourself and use later with either
810 test_have_prereq directly, or the three argument invocation of
811 test_expect_success and test_expect_failure.
813 - test_have_prereq <prereq>
815 Check if we have a prerequisite previously set with test_set_prereq.
816 The most common way to use this explicitly (as opposed to the
817 implicit use when an argument is passed to test_expect_*) is to skip
818 all the tests at the start of the test script if we don't have some
819 essential prerequisite:
821 if ! test_have_prereq PERL
823 skip_all='skipping perl interface tests, perl not available'
827 - test_external [<prereq>] <message> <external> <script>
829 Execute a <script> with an <external> interpreter (like perl). This
830 was added for tests like t9700-perl-git.sh which do most of their
831 work in an external test script.
834 'GitwebCache::*FileCache*' \
835 perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl
837 If the test is outputting its own TAP you should set the
838 test_external_has_tap variable somewhere before calling the first
839 test_external* function. See t9700-perl-git.sh for an example.
841 # The external test will outputs its own plan
842 test_external_has_tap=1
844 - test_external_without_stderr [<prereq>] <message> <external> <script>
846 Like test_external but fail if there's any output on stderr,
847 instead of checking the exit code.
849 test_external_without_stderr \
851 perl "$TEST_DIRECTORY"/t9700/test.pl
853 - test_expect_code <exit-code> <command>
855 Run a command and ensure that it exits with the given exit code.
858 test_expect_success 'Merge with d/f conflicts' '
859 test_expect_code 1 git merge "merge msg" B master
862 - test_must_fail [<options>] <git-command>
864 Run a git command and ensure it fails in a controlled way. Use
865 this instead of "! <git-command>". When git-command dies due to a
866 segfault, test_must_fail diagnoses it as an error; "! <git-command>"
867 treats it as just another expected failure, which would let such a
870 Accepts the following options:
872 ok=<signal-name>[,<...>]:
873 Don't treat an exit caused by the given signal as error.
874 Multiple signals can be specified as a comma separated list.
875 Currently recognized signal names are: sigpipe, success.
876 (Don't use 'success', use 'test_might_fail' instead.)
878 - test_might_fail [<options>] <git-command>
880 Similar to test_must_fail, but tolerate success, too. Use this
881 instead of "<git-command> || :" to catch failures due to segv.
883 Accepts the same options as test_must_fail.
885 - test_cmp <expected> <actual>
887 Check whether the content of the <actual> file matches the
888 <expected> file. This behaves like "cmp" but produces more
889 helpful output when the test is run with "-v" option.
891 - test_cmp_rev <expected> <actual>
893 Check whether the <expected> rev points to the same commit as the
896 - test_line_count (= | -lt | -ge | ...) <length> <file>
898 Check whether a file has the length it is expected to.
900 - test_path_is_file <path> [<diagnosis>]
901 test_path_is_dir <path> [<diagnosis>]
902 test_path_is_missing <path> [<diagnosis>]
904 Check if the named path is a file, if the named path is a
905 directory, or if the named path does not exist, respectively,
906 and fail otherwise, showing the <diagnosis> text.
908 - test_when_finished <script>
910 Prepend <script> to a list of commands to run to clean up
911 at the end of the current test. If some clean-up command
912 fails, the test will not pass.
916 test_expect_success 'branch pointing to non-commit' '
917 git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
918 test_when_finished "git update-ref -d refs/heads/invalid" &&
922 - test_atexit <script>
924 Prepend <script> to a list of commands to run unconditionally to
925 clean up before the test script exits, e.g. to stop a daemon:
927 test_expect_success 'test git daemon' '
930 test_atexit 'kill $daemon_pid' &&
934 The commands will be executed before the trash directory is removed,
935 i.e. the atexit commands will still be able to access any pidfiles or
938 Note that these commands will be run even when a test script run
939 with '--immediate' fails. Be careful with your atexit commands to
940 minimize any changes to the failed state.
942 - test_write_lines <lines>
944 Write <lines> on standard output, one line per argument.
945 Useful to prepare multi-line files in a compact form.
949 test_write_lines a b c d e f g >foo
951 Is a more compact equivalent of:
965 This command is useful for writing and debugging tests and must be
966 removed before submitting. It halts the execution of the test and
967 spawns a shell in the trash directory. Exit the shell to continue
970 test_expect_success 'test' '
971 git do-something >actual &&
973 test_cmp expected actual
976 - test_ln_s_add <path1> <path2>
978 This function helps systems whose filesystem does not support symbolic
979 links. Use it to add a symbolic link entry to the index when it is not
980 important that the file system entry is a symbolic link, i.e., instead
986 Sometimes it is possible to split a test in a part that does not need
987 the symbolic link in the file system and a part that does; then only
988 the latter part need be protected by a SYMLINKS prerequisite (see below).
992 This function loads facts and useful object IDs related to the hash
993 algorithm(s) in use from the files in t/oid-info.
997 This function reads per-hash algorithm information from standard
998 input (usually a heredoc) in the format described in
999 t/oid-info/README. This is useful for test-specific values, such as
1000 object IDs, which must vary based on the hash algorithm.
1002 Certain fixed values, such as hash sizes and common placeholder
1003 object IDs, can be loaded with test_oid_init (described above).
1007 This function looks up a value for the hash algorithm in use, based
1008 on the key given. The value must have been loaded using
1009 test_oid_init or test_oid_cache. Providing an unknown key is an
1014 This is often seen in modern UNIX but some platforms lack it, so
1015 the test harness overrides the platform implementation with a
1016 more limited one. Use this only when feeding a handful lines of
1017 output to the downstream---unlike the real version, it generates
1018 only up to 99 lines.
1020 - test_bool_env <env-variable-name> <default-value>
1022 Given the name of an environment variable with a bool value,
1023 normalize its value to a 0 (true) or 1 (false or empty string)
1024 return code. Return with code corresponding to the given default
1025 value if the variable is unset.
1026 Abort the test script if either the value of the variable or the
1027 default are not valid bool values.
1033 These are the prerequisites that the test library predefines with
1036 See the prereq argument to the test_* functions in the "Test harness
1037 library" section above and the "test_have_prereq" function for how to
1038 use these, and "test_set_prereq" for how to define your own.
1042 Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that
1043 need Python with this.
1047 Git wasn't compiled with NO_PERL=YesPlease.
1049 Even without the PERL prerequisite, tests can assume there is a
1050 usable perl interpreter at $PERL_PATH, though it need not be
1051 particularly modern.
1055 The filesystem supports POSIX style permission bits.
1059 Backslashes in pathspec are not directory separators. This is not
1060 set on Windows. See 6fd1106a for details.
1064 The process retains the same pid across exec(2). See fb9a2bea for
1069 The filesystem we're on supports creation of FIFOs (named pipes)
1074 The filesystem we're on supports symbolic links. E.g. a FAT
1075 filesystem doesn't support these. See 704a3143 for details.
1079 Test is not run by root user, and an attempt to write to an
1080 unwritable file is expected to fail correctly.
1084 Git was compiled with support for PCRE. Wrap any tests
1085 that use git-grep --perl-regexp or git-grep -P in these.
1089 Git was compiled with PCRE v1 support via
1090 USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some
1091 reason need v1 of the PCRE library instead of v2 in these.
1095 Git was compiled with PCRE v2 support via
1096 USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some
1097 reason need v2 of the PCRE library instead of v1 in these.
1099 - CASE_INSENSITIVE_FS
1101 Test is run on a case insensitive file system.
1105 Test is run on a filesystem which converts decomposed utf-8 (nfd)
1106 to precomposed utf-8 (nfc).
1110 Git wasn't compiled with NO_PTHREADS=YesPlease.
1112 Tips for Writing Tests
1113 ----------------------
1115 As with any programming projects, existing programs are the best
1116 source of the information. However, do _not_ emulate
1117 t0000-basic.sh when writing your tests. The test is special in
1118 that it tries to validate the very core of GIT. For example, it
1119 knows that there will be 256 subdirectories under .git/objects/,
1120 and it knows that the object ID of an empty tree is a certain
1121 40-byte string. This is deliberately done so in t0000-basic.sh
1122 because the things the very basic core test tries to achieve is
1123 to serve as a basis for people who are changing the GIT internal
1124 drastically. For these people, after making certain changes,
1125 not seeing failures from the basic test _is_ a failure. And
1126 such drastic changes to the core GIT that even changes these
1127 otherwise supposedly stable object IDs should be accompanied by
1128 an update to t0000-basic.sh.
1130 However, other tests that simply rely on basic parts of the core
1131 GIT working properly should not have that level of intimate
1132 knowledge of the core GIT internals. If all the test scripts
1133 hardcoded the object IDs like t0000-basic.sh does, that defeats
1134 the purpose of t0000-basic.sh, which is to isolate that level of
1135 validation in one place. Your test also ends up needing
1136 updating when such a change to the internal happens, so do _not_
1137 do it and leave the low level of validation to t0000-basic.sh.
1142 You can use the coverage tests to find code paths that are not being
1143 used or properly exercised yet.
1145 To do that, run the coverage target at the top-level (not in the t/
1150 That'll compile Git with GCC's coverage arguments, and generate a test
1151 report with gcov after the tests finish. Running the coverage tests
1152 can take a while, since running the tests in parallel is incompatible
1153 with GCC's coverage mode.
1155 After the tests have run you can generate a list of untested
1158 make coverage-untested-functions
1160 You can also generate a detailed per-file HTML report using the
1161 Devel::Cover module. To install it do:
1163 # On Debian or Ubuntu:
1164 sudo aptitude install libdevel-cover-perl
1166 # From the CPAN with cpanminus
1167 curl -L http://cpanmin.us | perl - --sudo --self-upgrade
1168 cpanm --sudo Devel::Cover
1170 Then, at the top-level:
1174 That'll generate a detailed cover report in the "cover_db_html"
1175 directory, which you can then copy to a webserver, or inspect locally