t / READMEon commit Merge branch 'ss/msvc-strcasecmp' (c0e286a)
   1Core GIT Tests
   2==============
   3
   4This directory holds many test scripts for core GIT tools.  The
   5first part of this short document describes how to run the tests
   6and read their output.
   7
   8When fixing the tools or adding enhancements, you are strongly
   9encouraged to add tests in this directory to cover what you are
  10trying to fix or enhance.  The later part of this short document
  11describes how your test scripts should be organized.
  12
  13
  14Running Tests
  15-------------
  16
  17The easiest way to run tests is to say "make".  This runs all
  18the tests.
  19
  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
  24    ...
  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)
  29    1..43
  30    *** t0001-init.sh ***
  31    ok 1 - plain
  32    ok 2 - plain with GIT_WORK_TREE
  33    ok 3 - plain bare
  34
  35Since the tests all output TAP (see http://testanything.org) they can
  36be run with any TAP harness. Here's an example of parallel testing
  37powered by a recent version of prove(1):
  38
  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... )===
  46
  47prove and other harnesses come with a lot of useful options. The
  48--state option in particular is very useful:
  49
  50    # Repeat until no more failures
  51    $ prove -j 15 --state=failed,save ./t[0-9]*.sh
  52
  53You can give DEFAULT_TEST_TARGET=prove on the make command (or define it
  54in config.mak) to cause "make test" to run tests under prove.
  55GIT_PROVE_OPTS can be used to pass additional options, e.g.
  56
  57    $ make DEFAULT_TEST_TARGET=prove GIT_PROVE_OPTS='--timer --jobs 16' test
  58
  59You can also run each test individually from command line, like this:
  60
  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)
  68    1..5
  69
  70You 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
  72appropriately before running "make".
  73
  74-v::
  75--verbose::
  76        This makes the test more verbose.  Specifically, the
  77        command being run and their output if any are also
  78        output.
  79
  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.
  84
  85-x::
  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.
  91
  92-d::
  93--debug::
  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
  99        the test finished.
 100
 101-i::
 102--immediate::
 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
 107        to diagnose the bug.
 108
 109-l::
 110--long-tests::
 111        This causes additional long-running tests to be run (where
 112        available), for more exhaustive testing.
 113
 114-r::
 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.
 119
 120--valgrind=<tool>::
 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).
 124
 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.
 128
 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
 132        installation.
 133
 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
 137        issues.
 138
 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
 144        't/valgrind/bin/'.
 145
 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.
 150
 151--tee::
 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.
 156
 157-V::
 158--verbose-log::
 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`.
 163
 164--with-dashes::
 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
 171        GIT_TEST_INSTALLED.
 172
 173--root=<directory>::
 174        Create "trash" directories used to store all temporary data during
 175        testing under <directory>, instead of the t/ directory.
 176        Using this option with a RAM-based filesystem (such as tmpfs)
 177        can massively speed up the test suite.
 178
 179--chain-lint::
 180--no-chain-lint::
 181        If --chain-lint is enabled, the test harness will check each
 182        test to make sure that it properly "&&-chains" all commands (so
 183        that a failure in the middle does not go unnoticed by the final
 184        exit code of the test). This check is performed in addition to
 185        running the tests themselves. You may also enable or disable
 186        this feature by setting the GIT_TEST_CHAIN_LINT environment
 187        variable to "1" or "0", respectively.
 188
 189You can also set the GIT_TEST_INSTALLED environment variable to
 190the bindir of an existing git installation to test that installation.
 191You still need to have built this git sandbox, from which various
 192test-* support programs, templates, and perl libraries are used.
 193If your installed git is incomplete, it will silently test parts of
 194your built version instead.
 195
 196When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
 197override the location of the dashed-form subcommands (what
 198GIT_EXEC_PATH would be used for during normal operation).
 199GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
 200
 201
 202Skipping Tests
 203--------------
 204
 205In some environments, certain tests have no way of succeeding
 206due to platform limitation, such as lack of 'unzip' program, or
 207filesystem that do not allow arbitrary sequence of non-NUL bytes
 208as pathnames.
 209
 210You should be able to say something like
 211
 212    $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
 213
 214and even:
 215
 216    $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
 217
 218to omit such tests.  The value of the environment variable is a
 219SP separated list of patterns that tells which tests to skip,
 220and either can match the "t[0-9]{4}" part to skip the whole
 221test, or t[0-9]{4} followed by ".$number" to say which
 222particular test to skip.
 223
 224For an individual test suite --run could be used to specify that
 225only some tests should be run or that some tests should be
 226excluded from a run.
 227
 228The argument for --run is a list of individual test numbers or
 229ranges with an optional negation prefix that define what tests in
 230a test suite to include in the run.  A range is two numbers
 231separated with a dash and matches a range of tests with both ends
 232been included.  You may omit the first or the second number to
 233mean "from the first test" or "up to the very last test"
 234respectively.
 235
 236Optional prefix of '!' means that the test or a range of tests
 237should be excluded from the run.
 238
 239If --run starts with an unprefixed number or range the initial
 240set of tests to run is empty. If the first item starts with '!'
 241all the tests are added to the initial set.  After initial set is
 242determined every test number or range is added or excluded from
 243the set one by one, from left to right.
 244
 245Individual numbers or ranges could be separated either by a space
 246or a comma.
 247
 248For example, to run only tests up to a specific test (21), one
 249could do this:
 250
 251    $ sh ./t9200-git-cvsexport-commit.sh --run='1-21'
 252
 253or this:
 254
 255    $ sh ./t9200-git-cvsexport-commit.sh --run='-21'
 256
 257Common case is to run several setup tests (1, 2, 3) and then a
 258specific test (21) that relies on that setup:
 259
 260    $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21'
 261
 262or:
 263
 264    $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21
 265
 266or:
 267
 268    $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21'
 269
 270As noted above, the test set is built by going through the items
 271from left to right, so this:
 272
 273    $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3'
 274
 275will run tests 1, 2, and 4.  Items that come later have higher
 276precedence.  It means that this:
 277
 278    $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4'
 279
 280would just run tests from 1 to 4, including 3.
 281
 282You may use negation with ranges.  The following will run all
 283test in the test suite except from 7 up to 11:
 284
 285    $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11'
 286
 287Some tests in a test suite rely on the previous tests performing
 288certain actions, specifically some tests are designated as
 289"setup" test, so you cannot _arbitrarily_ disable one test and
 290expect the rest to function correctly.
 291
 292--run is mostly useful when you want to focus on a specific test
 293and know what setup is needed for it.  Or when you want to run
 294everything up to a certain test.
 295
 296
 297Running tests with special setups
 298---------------------------------
 299
 300The whole test suite could be run to test some special features
 301that cannot be easily covered by a few specific test cases. These
 302could be enabled by running the test suite with correct GIT_TEST_
 303environment set.
 304
 305GIT_TEST_GETTEXT_POISON=<non-empty?> turns all strings marked for
 306translation into gibberish if non-empty (think "test -n"). Used for
 307spotting those tests that need to be marked with a C_LOCALE_OUTPUT
 308prerequisite when adding more strings for translation. See "Testing
 309marked strings" in po/README for details.
 310
 311GIT_TEST_SPLIT_INDEX=<boolean> forces split-index mode on the whole
 312test suite. Accept any boolean values that are accepted by git-config.
 313
 314GIT_TEST_FULL_IN_PACK_ARRAY=<boolean> exercises the uncommon
 315pack-objects code path where there are more than 1024 packs even if
 316the actual number of packs in repository is below this limit. Accept
 317any boolean values that are accepted by git-config.
 318
 319GIT_TEST_OE_SIZE=<n> exercises the uncommon pack-objects code path
 320where we do not cache object size in memory and read it from existing
 321packs on demand. This normally only happens when the object size is
 322over 2GB. This variable forces the code path on any object larger than
 323<n> bytes.
 324
 325GIT_TEST_OE_DELTA_SIZE=<n> exercises the uncommon pack-objects code
 326path where deltas larger than this limit require extra memory
 327allocation for bookkeeping.
 328
 329GIT_TEST_VALIDATE_INDEX_CACHE_ENTRIES=<boolean> checks that cache-tree
 330records are valid when the index is written out or after a merge. This
 331is mostly to catch missing invalidation. Default is true.
 332
 333GIT_TEST_COMMIT_GRAPH=<boolean>, when true, forces the commit-graph to
 334be written after every 'git commit' command, and overrides the
 335'core.commitGraph' setting to true.
 336
 337GIT_TEST_FSMONITOR=$PWD/t7519/fsmonitor-all exercises the fsmonitor
 338code path for utilizing a file system monitor to speed up detecting
 339new or changed files.
 340
 341GIT_TEST_INDEX_VERSION=<n> exercises the index read/write code path
 342for the index version specified.  Can be set to any valid version
 343(currently 2, 3, or 4).
 344
 345GIT_TEST_PRELOAD_INDEX=<boolean> exercises the preload-index code path
 346by overriding the minimum number of cache entries required per thread.
 347
 348GIT_TEST_REBASE_USE_BUILTIN=<boolean>, when false, disables the
 349builtin version of git-rebase. See 'rebase.useBuiltin' in
 350git-config(1).
 351
 352GIT_TEST_INDEX_THREADS=<n> enables exercising the multi-threaded loading
 353of the index for the whole test suite by bypassing the default number of
 354cache entries and thread minimums. Setting this to 1 will make the
 355index loading single threaded.
 356
 357GIT_TEST_MULTI_PACK_INDEX=<boolean>, when true, forces the multi-pack-
 358index to be written after every 'git repack' command, and overrides the
 359'core.multiPackIndex' setting to true.
 360
 361Naming Tests
 362------------
 363
 364The test files are named as:
 365
 366        tNNNN-commandname-details.sh
 367
 368where N is a decimal digit.
 369
 370First digit tells the family:
 371
 372        0 - the absolute basics and global stuff
 373        1 - the basic commands concerning database
 374        2 - the basic commands concerning the working tree
 375        3 - the other basic commands (e.g. ls-files)
 376        4 - the diff commands
 377        5 - the pull and exporting commands
 378        6 - the revision tree commands (even e.g. merge-base)
 379        7 - the porcelainish commands concerning the working tree
 380        8 - the porcelainish commands concerning forensics
 381        9 - the git tools
 382
 383Second digit tells the particular command we are testing.
 384
 385Third digit (optionally) tells the particular switch or group of switches
 386we are testing.
 387
 388If you create files under t/ directory (i.e. here) that is not
 389the top-level test script, never name the file to match the above
 390pattern.  The Makefile here considers all such files as the
 391top-level test script and tries to run all of them.  Care is
 392especially needed if you are creating a common test library
 393file, similar to test-lib.sh, because such a library file may
 394not be suitable for standalone execution.
 395
 396
 397Writing Tests
 398-------------
 399
 400The test script is written as a shell script.  It should start
 401with the standard "#!/bin/sh", and an
 402assignment to variable 'test_description', like this:
 403
 404        #!/bin/sh
 405
 406        test_description='xxx test (option --frotz)
 407
 408        This test registers the following structure in the cache
 409        and tries to run git-ls-files with option --frotz.'
 410
 411
 412Source 'test-lib.sh'
 413--------------------
 414
 415After assigning test_description, the test script should source
 416test-lib.sh like this:
 417
 418        . ./test-lib.sh
 419
 420This test harness library does the following things:
 421
 422 - If the script is invoked with command line argument --help
 423   (or -h), it shows the test_description and exits.
 424
 425 - Creates an empty test directory with an empty .git/objects database
 426   and chdir(2) into it.  This directory is 't/trash
 427   directory.$test_name_without_dotsh', with t/ subject to change by
 428   the --root option documented above.
 429
 430 - Defines standard test helper functions for your scripts to
 431   use.  These functions are designed to make all scripts behave
 432   consistently when command line arguments --verbose (or -v),
 433   --debug (or -d), and --immediate (or -i) is given.
 434
 435Do's & don'ts
 436-------------
 437
 438Here are a few examples of things you probably should and shouldn't do
 439when writing tests.
 440
 441Here are the "do's:"
 442
 443 - Put all code inside test_expect_success and other assertions.
 444
 445   Even code that isn't a test per se, but merely some setup code
 446   should be inside a test assertion.
 447
 448 - Chain your test assertions
 449
 450   Write test code like this:
 451
 452        git merge foo &&
 453        git push bar &&
 454        test ...
 455
 456   Instead of:
 457
 458        git merge hla
 459        git push gh
 460        test ...
 461
 462   That way all of the commands in your tests will succeed or fail. If
 463   you must ignore the return value of something, consider using a
 464   helper function (e.g. use sane_unset instead of unset, in order
 465   to avoid unportable return value for unsetting a variable that was
 466   already unset), or prepending the command with test_might_fail or
 467   test_must_fail.
 468
 469 - Check the test coverage for your tests. See the "Test coverage"
 470   below.
 471
 472   Don't blindly follow test coverage metrics; if a new function you added
 473   doesn't have any coverage, then you're probably doing something wrong,
 474   but having 100% coverage doesn't necessarily mean that you tested
 475   everything.
 476
 477   Tests that are likely to smoke out future regressions are better
 478   than tests that just inflate the coverage metrics.
 479
 480 - When a test checks for an absolute path that a git command generated,
 481   construct the expected value using $(pwd) rather than $PWD,
 482   $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
 483   Windows, where the shell (MSYS bash) mangles absolute path names.
 484   For details, see the commit message of 4114156ae9.
 485
 486 - Remember that inside the <script> part, the standard output and
 487   standard error streams are discarded, and the test harness only
 488   reports "ok" or "not ok" to the end user running the tests. Under
 489   --verbose, they are shown to help debug the tests.
 490
 491And here are the "don'ts:"
 492
 493 - Don't exit() within a <script> part.
 494
 495   The harness will catch this as a programming error of the test.
 496   Use test_done instead if you need to stop the tests early (see
 497   "Skipping tests" below).
 498
 499 - Don't use '! git cmd' when you want to make sure the git command
 500   exits with failure in a controlled way by calling "die()".  Instead,
 501   use 'test_must_fail git cmd'.  This will signal a failure if git
 502   dies in an unexpected way (e.g. segfault).
 503
 504   On the other hand, don't use test_must_fail for running regular
 505   platform commands; just use '! cmd'.  We are not in the business
 506   of verifying that the world given to us sanely works.
 507
 508 - Don't feed the output of a git command to a pipe, as in:
 509
 510     git -C repo ls-files |
 511     xargs -n 1 basename |
 512     grep foo
 513
 514   which will discard git's exit code and may mask a crash. In the
 515   above example, all exit codes are ignored except grep's.
 516
 517   Instead, write the output of that command to a temporary
 518   file with ">" or assign it to a variable with "x=$(git ...)" rather
 519   than pipe it.
 520
 521 - Don't use command substitution in a way that discards git's exit
 522   code. When assigning to a variable, the exit code is not discarded,
 523   e.g.:
 524
 525     x=$(git cat-file -p $sha) &&
 526     ...
 527
 528   is OK because a crash in "git cat-file" will cause the "&&" chain
 529   to fail, but:
 530
 531     test "refs/heads/foo" = "$(git symbolic-ref HEAD)"
 532
 533   is not OK and a crash in git could go undetected.
 534
 535 - Don't use perl without spelling it as "$PERL_PATH". This is to help
 536   our friends on Windows where the platform Perl often adds CR before
 537   the end of line, and they bundle Git with a version of Perl that
 538   does not do so, whose path is specified with $PERL_PATH. Note that we
 539   provide a "perl" function which uses $PERL_PATH under the hood, so
 540   you do not need to worry when simply running perl in the test scripts
 541   (but you do, for example, on a shebang line or in a sub script
 542   created via "write_script").
 543
 544 - Don't use sh without spelling it as "$SHELL_PATH", when the script
 545   can be misinterpreted by broken platform shell (e.g. Solaris).
 546
 547 - Don't chdir around in tests.  It is not sufficient to chdir to
 548   somewhere and then chdir back to the original location later in
 549   the test, as any intermediate step can fail and abort the test,
 550   causing the next test to start in an unexpected directory.  Do so
 551   inside a subshell if necessary.
 552
 553 - Don't save and verify the standard error of compound commands, i.e.
 554   group commands, subshells, and shell functions (except test helper
 555   functions like 'test_must_fail') like this:
 556
 557     ( cd dir && git cmd ) 2>error &&
 558     test_cmp expect error
 559
 560   When running the test with '-x' tracing, then the trace of commands
 561   executed in the compound command will be included in standard error
 562   as well, quite possibly throwing off the subsequent checks examining
 563   the output.  Instead, save only the relevant git command's standard
 564   error:
 565
 566     ( cd dir && git cmd 2>../error ) &&
 567     test_cmp expect error
 568
 569 - Don't break the TAP output
 570
 571   The raw output from your test may be interpreted by a TAP harness. TAP
 572   harnesses will ignore everything they don't know about, but don't step
 573   on their toes in these areas:
 574
 575   - Don't print lines like "$x..$y" where $x and $y are integers.
 576
 577   - Don't print lines that begin with "ok" or "not ok".
 578
 579   TAP harnesses expect a line that begins with either "ok" and "not
 580   ok" to signal a test passed or failed (and our harness already
 581   produces such lines), so your script shouldn't emit such lines to
 582   their output.
 583
 584   You can glean some further possible issues from the TAP grammar
 585   (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR)
 586   but the best indication is to just run the tests with prove(1),
 587   it'll complain if anything is amiss.
 588
 589
 590Skipping tests
 591--------------
 592
 593If you need to skip tests you should do so by using the three-arg form
 594of the test_* functions (see the "Test harness library" section
 595below), e.g.:
 596
 597    test_expect_success PERL 'I need Perl' '
 598        perl -e "hlagh() if unf_unf()"
 599    '
 600
 601The advantage of skipping tests like this is that platforms that don't
 602have the PERL and other optional dependencies get an indication of how
 603many tests they're missing.
 604
 605If the test code is too hairy for that (i.e. does a lot of setup work
 606outside test assertions) you can also skip all remaining tests by
 607setting skip_all and immediately call test_done:
 608
 609        if ! test_have_prereq PERL
 610        then
 611            skip_all='skipping perl interface tests, perl not available'
 612            test_done
 613        fi
 614
 615The string you give to skip_all will be used as an explanation for why
 616the test was skipped.
 617
 618End with test_done
 619------------------
 620
 621Your script will be a sequence of tests, using helper functions
 622from the test harness library.  At the end of the script, call
 623'test_done'.
 624
 625
 626Test harness library
 627--------------------
 628
 629There are a handful helper functions defined in the test harness
 630library for your script to use.
 631
 632 - test_expect_success [<prereq>] <message> <script>
 633
 634   Usually takes two strings as parameters, and evaluates the
 635   <script>.  If it yields success, test is considered
 636   successful.  <message> should state what it is testing.
 637
 638   Example:
 639
 640        test_expect_success \
 641            'git-write-tree should be able to write an empty tree.' \
 642            'tree=$(git-write-tree)'
 643
 644   If you supply three parameters the first will be taken to be a
 645   prerequisite; see the test_set_prereq and test_have_prereq
 646   documentation below:
 647
 648        test_expect_success TTY 'git --paginate rev-list uses a pager' \
 649            ' ... '
 650
 651   You can also supply a comma-separated list of prerequisites, in the
 652   rare case where your test depends on more than one:
 653
 654        test_expect_success PERL,PYTHON 'yo dawg' \
 655            ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" '
 656
 657 - test_expect_failure [<prereq>] <message> <script>
 658
 659   This is NOT the opposite of test_expect_success, but is used
 660   to mark a test that demonstrates a known breakage.  Unlike
 661   the usual test_expect_success tests, which say "ok" on
 662   success and "FAIL" on failure, this will say "FIXED" on
 663   success and "still broken" on failure.  Failures from these
 664   tests won't cause -i (immediate) to stop.
 665
 666   Like test_expect_success this function can optionally use a three
 667   argument invocation with a prerequisite as the first argument.
 668
 669 - test_debug <script>
 670
 671   This takes a single argument, <script>, and evaluates it only
 672   when the test script is started with --debug command line
 673   argument.  This is primarily meant for use during the
 674   development of a new test script.
 675
 676 - debug <git-command>
 677
 678   Run a git command inside a debugger. This is primarily meant for
 679   use when debugging a failing test script.
 680
 681 - test_done
 682
 683   Your test script must have test_done at the end.  Its purpose
 684   is to summarize successes and failures in the test script and
 685   exit with an appropriate error code.
 686
 687 - test_tick
 688
 689   Make commit and tag names consistent by setting the author and
 690   committer times to defined state.  Subsequent calls will
 691   advance the times by a fixed amount.
 692
 693 - test_commit <message> [<filename> [<contents>]]
 694
 695   Creates a commit with the given message, committing the given
 696   file with the given contents (default for both is to reuse the
 697   message string), and adds a tag (again reusing the message
 698   string as name).  Calls test_tick to make the SHA-1s
 699   reproducible.
 700
 701 - test_merge <message> <commit-or-tag>
 702
 703   Merges the given rev using the given message.  Like test_commit,
 704   creates a tag and calls test_tick before committing.
 705
 706 - test_set_prereq <prereq>
 707
 708   Set a test prerequisite to be used later with test_have_prereq. The
 709   test-lib will set some prerequisites for you, see the
 710   "Prerequisites" section below for a full list of these.
 711
 712   Others you can set yourself and use later with either
 713   test_have_prereq directly, or the three argument invocation of
 714   test_expect_success and test_expect_failure.
 715
 716 - test_have_prereq <prereq>
 717
 718   Check if we have a prerequisite previously set with test_set_prereq.
 719   The most common way to use this explicitly (as opposed to the
 720   implicit use when an argument is passed to test_expect_*) is to skip
 721   all the tests at the start of the test script if we don't have some
 722   essential prerequisite:
 723
 724        if ! test_have_prereq PERL
 725        then
 726            skip_all='skipping perl interface tests, perl not available'
 727            test_done
 728        fi
 729
 730 - test_external [<prereq>] <message> <external> <script>
 731
 732   Execute a <script> with an <external> interpreter (like perl). This
 733   was added for tests like t9700-perl-git.sh which do most of their
 734   work in an external test script.
 735
 736        test_external \
 737            'GitwebCache::*FileCache*' \
 738            perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl
 739
 740   If the test is outputting its own TAP you should set the
 741   test_external_has_tap variable somewhere before calling the first
 742   test_external* function. See t9700-perl-git.sh for an example.
 743
 744        # The external test will outputs its own plan
 745        test_external_has_tap=1
 746
 747 - test_external_without_stderr [<prereq>] <message> <external> <script>
 748
 749   Like test_external but fail if there's any output on stderr,
 750   instead of checking the exit code.
 751
 752        test_external_without_stderr \
 753            'Perl API' \
 754            perl "$TEST_DIRECTORY"/t9700/test.pl
 755
 756 - test_expect_code <exit-code> <command>
 757
 758   Run a command and ensure that it exits with the given exit code.
 759   For example:
 760
 761        test_expect_success 'Merge with d/f conflicts' '
 762                test_expect_code 1 git merge "merge msg" B master
 763        '
 764
 765 - test_must_fail [<options>] <git-command>
 766
 767   Run a git command and ensure it fails in a controlled way.  Use
 768   this instead of "! <git-command>".  When git-command dies due to a
 769   segfault, test_must_fail diagnoses it as an error; "! <git-command>"
 770   treats it as just another expected failure, which would let such a
 771   bug go unnoticed.
 772
 773   Accepts the following options:
 774
 775     ok=<signal-name>[,<...>]:
 776       Don't treat an exit caused by the given signal as error.
 777       Multiple signals can be specified as a comma separated list.
 778       Currently recognized signal names are: sigpipe, success.
 779       (Don't use 'success', use 'test_might_fail' instead.)
 780
 781 - test_might_fail [<options>] <git-command>
 782
 783   Similar to test_must_fail, but tolerate success, too.  Use this
 784   instead of "<git-command> || :" to catch failures due to segv.
 785
 786   Accepts the same options as test_must_fail.
 787
 788 - test_cmp <expected> <actual>
 789
 790   Check whether the content of the <actual> file matches the
 791   <expected> file.  This behaves like "cmp" but produces more
 792   helpful output when the test is run with "-v" option.
 793
 794 - test_cmp_rev <expected> <actual>
 795
 796   Check whether the <expected> rev points to the same commit as the
 797   <actual> rev.
 798
 799 - test_line_count (= | -lt | -ge | ...) <length> <file>
 800
 801   Check whether a file has the length it is expected to.
 802
 803 - test_path_is_file <path> [<diagnosis>]
 804   test_path_is_dir <path> [<diagnosis>]
 805   test_path_is_missing <path> [<diagnosis>]
 806
 807   Check if the named path is a file, if the named path is a
 808   directory, or if the named path does not exist, respectively,
 809   and fail otherwise, showing the <diagnosis> text.
 810
 811 - test_when_finished <script>
 812
 813   Prepend <script> to a list of commands to run to clean up
 814   at the end of the current test.  If some clean-up command
 815   fails, the test will not pass.
 816
 817   Example:
 818
 819        test_expect_success 'branch pointing to non-commit' '
 820                git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
 821                test_when_finished "git update-ref -d refs/heads/invalid" &&
 822                ...
 823        '
 824
 825 - test_write_lines <lines>
 826
 827   Write <lines> on standard output, one line per argument.
 828   Useful to prepare multi-line files in a compact form.
 829
 830   Example:
 831
 832        test_write_lines a b c d e f g >foo
 833
 834   Is a more compact equivalent of:
 835        cat >foo <<-EOF
 836        a
 837        b
 838        c
 839        d
 840        e
 841        f
 842        g
 843        EOF
 844
 845
 846 - test_pause
 847
 848        This command is useful for writing and debugging tests and must be
 849        removed before submitting. It halts the execution of the test and
 850        spawns a shell in the trash directory. Exit the shell to continue
 851        the test. Example:
 852
 853        test_expect_success 'test' '
 854                git do-something >actual &&
 855                test_pause &&
 856                test_cmp expected actual
 857        '
 858
 859 - test_ln_s_add <path1> <path2>
 860
 861   This function helps systems whose filesystem does not support symbolic
 862   links. Use it to add a symbolic link entry to the index when it is not
 863   important that the file system entry is a symbolic link, i.e., instead
 864   of the sequence
 865
 866        ln -s foo bar &&
 867        git add bar
 868
 869   Sometimes it is possible to split a test in a part that does not need
 870   the symbolic link in the file system and a part that does; then only
 871   the latter part need be protected by a SYMLINKS prerequisite (see below).
 872
 873 - test_oid_init
 874
 875   This function loads facts and useful object IDs related to the hash
 876   algorithm(s) in use from the files in t/oid-info.
 877
 878 - test_oid_cache
 879
 880   This function reads per-hash algorithm information from standard
 881   input (usually a heredoc) in the format described in
 882   t/oid-info/README.  This is useful for test-specific values, such as
 883   object IDs, which must vary based on the hash algorithm.
 884
 885   Certain fixed values, such as hash sizes and common placeholder
 886   object IDs, can be loaded with test_oid_init (described above).
 887
 888 - test_oid <key>
 889
 890   This function looks up a value for the hash algorithm in use, based
 891   on the key given.  The value must have been loaded using
 892   test_oid_init or test_oid_cache.  Providing an unknown key is an
 893   error.
 894
 895Prerequisites
 896-------------
 897
 898These are the prerequisites that the test library predefines with
 899test_have_prereq.
 900
 901See the prereq argument to the test_* functions in the "Test harness
 902library" section above and the "test_have_prereq" function for how to
 903use these, and "test_set_prereq" for how to define your own.
 904
 905 - PYTHON
 906
 907   Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that
 908   need Python with this.
 909
 910 - PERL
 911
 912   Git wasn't compiled with NO_PERL=YesPlease.
 913
 914   Even without the PERL prerequisite, tests can assume there is a
 915   usable perl interpreter at $PERL_PATH, though it need not be
 916   particularly modern.
 917
 918 - POSIXPERM
 919
 920   The filesystem supports POSIX style permission bits.
 921
 922 - BSLASHPSPEC
 923
 924   Backslashes in pathspec are not directory separators. This is not
 925   set on Windows. See 6fd1106a for details.
 926
 927 - EXECKEEPSPID
 928
 929   The process retains the same pid across exec(2). See fb9a2bea for
 930   details.
 931
 932 - PIPE
 933
 934   The filesystem we're on supports creation of FIFOs (named pipes)
 935   via mkfifo(1).
 936
 937 - SYMLINKS
 938
 939   The filesystem we're on supports symbolic links. E.g. a FAT
 940   filesystem doesn't support these. See 704a3143 for details.
 941
 942 - SANITY
 943
 944   Test is not run by root user, and an attempt to write to an
 945   unwritable file is expected to fail correctly.
 946
 947 - PCRE
 948
 949   Git was compiled with support for PCRE. Wrap any tests
 950   that use git-grep --perl-regexp or git-grep -P in these.
 951
 952 - LIBPCRE1
 953
 954   Git was compiled with PCRE v1 support via
 955   USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some
 956   reason need v1 of the PCRE library instead of v2 in these.
 957
 958 - LIBPCRE2
 959
 960   Git was compiled with PCRE v2 support via
 961   USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some
 962   reason need v2 of the PCRE library instead of v1 in these.
 963
 964 - CASE_INSENSITIVE_FS
 965
 966   Test is run on a case insensitive file system.
 967
 968 - UTF8_NFD_TO_NFC
 969
 970   Test is run on a filesystem which converts decomposed utf-8 (nfd)
 971   to precomposed utf-8 (nfc).
 972
 973 - PTHREADS
 974
 975   Git wasn't compiled with NO_PTHREADS=YesPlease.
 976
 977Tips for Writing Tests
 978----------------------
 979
 980As with any programming projects, existing programs are the best
 981source of the information.  However, do _not_ emulate
 982t0000-basic.sh when writing your tests.  The test is special in
 983that it tries to validate the very core of GIT.  For example, it
 984knows that there will be 256 subdirectories under .git/objects/,
 985and it knows that the object ID of an empty tree is a certain
 98640-byte string.  This is deliberately done so in t0000-basic.sh
 987because the things the very basic core test tries to achieve is
 988to serve as a basis for people who are changing the GIT internal
 989drastically.  For these people, after making certain changes,
 990not seeing failures from the basic test _is_ a failure.  And
 991such drastic changes to the core GIT that even changes these
 992otherwise supposedly stable object IDs should be accompanied by
 993an update to t0000-basic.sh.
 994
 995However, other tests that simply rely on basic parts of the core
 996GIT working properly should not have that level of intimate
 997knowledge of the core GIT internals.  If all the test scripts
 998hardcoded the object IDs like t0000-basic.sh does, that defeats
 999the purpose of t0000-basic.sh, which is to isolate that level of
1000validation in one place.  Your test also ends up needing
1001updating when such a change to the internal happens, so do _not_
1002do it and leave the low level of validation to t0000-basic.sh.
1003
1004Test coverage
1005-------------
1006
1007You can use the coverage tests to find code paths that are not being
1008used or properly exercised yet.
1009
1010To do that, run the coverage target at the top-level (not in the t/
1011directory):
1012
1013    make coverage
1014
1015That'll compile Git with GCC's coverage arguments, and generate a test
1016report with gcov after the tests finish. Running the coverage tests
1017can take a while, since running the tests in parallel is incompatible
1018with GCC's coverage mode.
1019
1020After the tests have run you can generate a list of untested
1021functions:
1022
1023    make coverage-untested-functions
1024
1025You can also generate a detailed per-file HTML report using the
1026Devel::Cover module. To install it do:
1027
1028   # On Debian or Ubuntu:
1029   sudo aptitude install libdevel-cover-perl
1030
1031   # From the CPAN with cpanminus
1032   curl -L http://cpanmin.us | perl - --sudo --self-upgrade
1033   cpanm --sudo Devel::Cover
1034
1035Then, at the top-level:
1036
1037    make cover_db_html
1038
1039That'll generate a detailed cover report in the "cover_db_html"
1040directory, which you can then copy to a webserver, or inspect locally
1041in a browser.