t / READMEon commit Merge branch 'sr/http-sslverify-config-doc' (593fdcc)
   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`. Note that in non-bash shells,
  88        this can cause failures in some tests which redirect and test
  89        the output of shell functions. Use with caution.
  90
  91-d::
  92--debug::
  93        This may help the person who is developing a new test.
  94        It causes the command defined with test_debug to run.
  95        The "trash" directory (used to store all temporary data
  96        during testing) is not deleted even if there are no
  97        failed tests so that you can inspect its contents after
  98        the test finished.
  99
 100-i::
 101--immediate::
 102        This causes the test to immediately exit upon the first
 103        failed test. Cleanup commands requested with
 104        test_when_finished are not executed if the test failed,
 105        in order to keep the state for inspection by the tester
 106        to diagnose the bug.
 107
 108-l::
 109--long-tests::
 110        This causes additional long-running tests to be run (where
 111        available), for more exhaustive testing.
 112
 113-r::
 114--run=<test-selector>::
 115        Run only the subset of tests indicated by
 116        <test-selector>.  See section "Skipping Tests" below for
 117        <test-selector> syntax.
 118
 119--valgrind=<tool>::
 120        Execute all Git binaries under valgrind tool <tool> and exit
 121        with status 126 on errors (just like regular tests, this will
 122        only stop the test script when running under -i).
 123
 124        Since it makes no sense to run the tests with --valgrind and
 125        not see any output, this option implies --verbose.  For
 126        convenience, it also implies --tee.
 127
 128        <tool> defaults to 'memcheck', just like valgrind itself.
 129        Other particularly useful choices include 'helgrind' and
 130        'drd', but you may use any tool recognized by your valgrind
 131        installation.
 132
 133        As a special case, <tool> can be 'memcheck-fast', which uses
 134        memcheck but disables --track-origins.  Use this if you are
 135        running tests in bulk, to see if there are _any_ memory
 136        issues.
 137
 138        Note that memcheck is run with the option --leak-check=no,
 139        as the git process is short-lived and some errors are not
 140        interesting. In order to run a single command under the same
 141        conditions manually, you should set GIT_VALGRIND to point to
 142        the 't/valgrind/' directory and use the commands under
 143        't/valgrind/bin/'.
 144
 145--valgrind-only=<pattern>::
 146        Like --valgrind, but the effect is limited to tests with
 147        numbers matching <pattern>.  The number matched against is
 148        simply the running count of the test within the file.
 149
 150--tee::
 151        In addition to printing the test output to the terminal,
 152        write it to files named 't/test-results/$TEST_NAME.out'.
 153        As the names depend on the tests' file names, it is safe to
 154        run the tests with this option in parallel.
 155
 156--verbose-log::
 157        Write verbose output to the same logfile as `--tee`, but do
 158        _not_ write it to stdout. Unlike `--tee --verbose`, this option
 159        is safe to use when stdout is being consumed by a TAP parser
 160        like `prove`. Implies `--tee` and `--verbose`.
 161
 162--with-dashes::
 163        By default tests are run without dashed forms of
 164        commands (like git-commit) in the PATH (it only uses
 165        wrappers from ../bin-wrappers).  Use this option to include
 166        the build directory (..) in the PATH, which contains all
 167        the dashed forms of commands.  This option is currently
 168        implied by other options like --valgrind and
 169        GIT_TEST_INSTALLED.
 170
 171--root=<directory>::
 172        Create "trash" directories used to store all temporary data during
 173        testing under <directory>, instead of the t/ directory.
 174        Using this option with a RAM-based filesystem (such as tmpfs)
 175        can massively speed up the test suite.
 176
 177--chain-lint::
 178--no-chain-lint::
 179        If --chain-lint is enabled, the test harness will check each
 180        test to make sure that it properly "&&-chains" all commands (so
 181        that a failure in the middle does not go unnoticed by the final
 182        exit code of the test). This check is performed in addition to
 183        running the tests themselves. You may also enable or disable
 184        this feature by setting the GIT_TEST_CHAIN_LINT environment
 185        variable to "1" or "0", respectively.
 186
 187You can also set the GIT_TEST_INSTALLED environment variable to
 188the bindir of an existing git installation to test that installation.
 189You still need to have built this git sandbox, from which various
 190test-* support programs, templates, and perl libraries are used.
 191If your installed git is incomplete, it will silently test parts of
 192your built version instead.
 193
 194When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
 195override the location of the dashed-form subcommands (what
 196GIT_EXEC_PATH would be used for during normal operation).
 197GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
 198
 199
 200Skipping Tests
 201--------------
 202
 203In some environments, certain tests have no way of succeeding
 204due to platform limitation, such as lack of 'unzip' program, or
 205filesystem that do not allow arbitrary sequence of non-NUL bytes
 206as pathnames.
 207
 208You should be able to say something like
 209
 210    $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
 211
 212and even:
 213
 214    $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
 215
 216to omit such tests.  The value of the environment variable is a
 217SP separated list of patterns that tells which tests to skip,
 218and either can match the "t[0-9]{4}" part to skip the whole
 219test, or t[0-9]{4} followed by ".$number" to say which
 220particular test to skip.
 221
 222For an individual test suite --run could be used to specify that
 223only some tests should be run or that some tests should be
 224excluded from a run.
 225
 226The argument for --run is a list of individual test numbers or
 227ranges with an optional negation prefix that define what tests in
 228a test suite to include in the run.  A range is two numbers
 229separated with a dash and matches a range of tests with both ends
 230been included.  You may omit the first or the second number to
 231mean "from the first test" or "up to the very last test"
 232respectively.
 233
 234Optional prefix of '!' means that the test or a range of tests
 235should be excluded from the run.
 236
 237If --run starts with an unprefixed number or range the initial
 238set of tests to run is empty. If the first item starts with '!'
 239all the tests are added to the initial set.  After initial set is
 240determined every test number or range is added or excluded from
 241the set one by one, from left to right.
 242
 243Individual numbers or ranges could be separated either by a space
 244or a comma.
 245
 246For example, to run only tests up to a specific test (21), one
 247could do this:
 248
 249    $ sh ./t9200-git-cvsexport-commit.sh --run='1-21'
 250
 251or this:
 252
 253    $ sh ./t9200-git-cvsexport-commit.sh --run='-21'
 254
 255Common case is to run several setup tests (1, 2, 3) and then a
 256specific test (21) that relies on that setup:
 257
 258    $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21'
 259
 260or:
 261
 262    $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21
 263
 264or:
 265
 266    $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21'
 267
 268As noted above, the test set is built by going through the items
 269from left to right, so this:
 270
 271    $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3'
 272
 273will run tests 1, 2, and 4.  Items that come later have higher
 274precedence.  It means that this:
 275
 276    $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4'
 277
 278would just run tests from 1 to 4, including 3.
 279
 280You may use negation with ranges.  The following will run all
 281test in the test suite except from 7 up to 11:
 282
 283    $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11'
 284
 285Some tests in a test suite rely on the previous tests performing
 286certain actions, specifically some tests are designated as
 287"setup" test, so you cannot _arbitrarily_ disable one test and
 288expect the rest to function correctly.
 289
 290--run is mostly useful when you want to focus on a specific test
 291and know what setup is needed for it.  Or when you want to run
 292everything up to a certain test.
 293
 294
 295Naming Tests
 296------------
 297
 298The test files are named as:
 299
 300        tNNNN-commandname-details.sh
 301
 302where N is a decimal digit.
 303
 304First digit tells the family:
 305
 306        0 - the absolute basics and global stuff
 307        1 - the basic commands concerning database
 308        2 - the basic commands concerning the working tree
 309        3 - the other basic commands (e.g. ls-files)
 310        4 - the diff commands
 311        5 - the pull and exporting commands
 312        6 - the revision tree commands (even e.g. merge-base)
 313        7 - the porcelainish commands concerning the working tree
 314        8 - the porcelainish commands concerning forensics
 315        9 - the git tools
 316
 317Second digit tells the particular command we are testing.
 318
 319Third digit (optionally) tells the particular switch or group of switches
 320we are testing.
 321
 322If you create files under t/ directory (i.e. here) that is not
 323the top-level test script, never name the file to match the above
 324pattern.  The Makefile here considers all such files as the
 325top-level test script and tries to run all of them.  Care is
 326especially needed if you are creating a common test library
 327file, similar to test-lib.sh, because such a library file may
 328not be suitable for standalone execution.
 329
 330
 331Writing Tests
 332-------------
 333
 334The test script is written as a shell script.  It should start
 335with the standard "#!/bin/sh", and an
 336assignment to variable 'test_description', like this:
 337
 338        #!/bin/sh
 339
 340        test_description='xxx test (option --frotz)
 341
 342        This test registers the following structure in the cache
 343        and tries to run git-ls-files with option --frotz.'
 344
 345
 346Source 'test-lib.sh'
 347--------------------
 348
 349After assigning test_description, the test script should source
 350test-lib.sh like this:
 351
 352        . ./test-lib.sh
 353
 354This test harness library does the following things:
 355
 356 - If the script is invoked with command line argument --help
 357   (or -h), it shows the test_description and exits.
 358
 359 - Creates an empty test directory with an empty .git/objects database
 360   and chdir(2) into it.  This directory is 't/trash
 361   directory.$test_name_without_dotsh', with t/ subject to change by
 362   the --root option documented above.
 363
 364 - Defines standard test helper functions for your scripts to
 365   use.  These functions are designed to make all scripts behave
 366   consistently when command line arguments --verbose (or -v),
 367   --debug (or -d), and --immediate (or -i) is given.
 368
 369Do's, don'ts & things to keep in mind
 370-------------------------------------
 371
 372Here are a few examples of things you probably should and shouldn't do
 373when writing tests.
 374
 375Do:
 376
 377 - Put all code inside test_expect_success and other assertions.
 378
 379   Even code that isn't a test per se, but merely some setup code
 380   should be inside a test assertion.
 381
 382 - Chain your test assertions
 383
 384   Write test code like this:
 385
 386        git merge foo &&
 387        git push bar &&
 388        test ...
 389
 390   Instead of:
 391
 392        git merge hla
 393        git push gh
 394        test ...
 395
 396   That way all of the commands in your tests will succeed or fail. If
 397   you must ignore the return value of something, consider using a
 398   helper function (e.g. use sane_unset instead of unset, in order
 399   to avoid unportable return value for unsetting a variable that was
 400   already unset), or prepending the command with test_might_fail or
 401   test_must_fail.
 402
 403 - Check the test coverage for your tests. See the "Test coverage"
 404   below.
 405
 406   Don't blindly follow test coverage metrics; if a new function you added
 407   doesn't have any coverage, then you're probably doing something wrong,
 408   but having 100% coverage doesn't necessarily mean that you tested
 409   everything.
 410
 411   Tests that are likely to smoke out future regressions are better
 412   than tests that just inflate the coverage metrics.
 413
 414 - When a test checks for an absolute path that a git command generated,
 415   construct the expected value using $(pwd) rather than $PWD,
 416   $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
 417   Windows, where the shell (MSYS bash) mangles absolute path names.
 418   For details, see the commit message of 4114156ae9.
 419
 420Don't:
 421
 422 - exit() within a <script> part.
 423
 424   The harness will catch this as a programming error of the test.
 425   Use test_done instead if you need to stop the tests early (see
 426   "Skipping tests" below).
 427
 428 - use '! git cmd' when you want to make sure the git command exits
 429   with failure in a controlled way by calling "die()".  Instead,
 430   use 'test_must_fail git cmd'.  This will signal a failure if git
 431   dies in an unexpected way (e.g. segfault).
 432
 433   On the other hand, don't use test_must_fail for running regular
 434   platform commands; just use '! cmd'.  We are not in the business
 435   of verifying that the world given to us sanely works.
 436
 437 - use perl without spelling it as "$PERL_PATH". This is to help our
 438   friends on Windows where the platform Perl often adds CR before
 439   the end of line, and they bundle Git with a version of Perl that
 440   does not do so, whose path is specified with $PERL_PATH. Note that we
 441   provide a "perl" function which uses $PERL_PATH under the hood, so
 442   you do not need to worry when simply running perl in the test scripts
 443   (but you do, for example, on a shebang line or in a sub script
 444   created via "write_script").
 445
 446 - use sh without spelling it as "$SHELL_PATH", when the script can
 447   be misinterpreted by broken platform shell (e.g. Solaris).
 448
 449 - chdir around in tests.  It is not sufficient to chdir to
 450   somewhere and then chdir back to the original location later in
 451   the test, as any intermediate step can fail and abort the test,
 452   causing the next test to start in an unexpected directory.  Do so
 453   inside a subshell if necessary.
 454
 455 - Break the TAP output
 456
 457   The raw output from your test may be interpreted by a TAP harness. TAP
 458   harnesses will ignore everything they don't know about, but don't step
 459   on their toes in these areas:
 460
 461   - Don't print lines like "$x..$y" where $x and $y are integers.
 462
 463   - Don't print lines that begin with "ok" or "not ok".
 464
 465   TAP harnesses expect a line that begins with either "ok" and "not
 466   ok" to signal a test passed or failed (and our harness already
 467   produces such lines), so your script shouldn't emit such lines to
 468   their output.
 469
 470   You can glean some further possible issues from the TAP grammar
 471   (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR)
 472   but the best indication is to just run the tests with prove(1),
 473   it'll complain if anything is amiss.
 474
 475Keep in mind:
 476
 477 - Inside the <script> part, the standard output and standard error
 478   streams are discarded, and the test harness only reports "ok" or
 479   "not ok" to the end user running the tests. Under --verbose, they
 480   are shown to help debugging the tests.
 481
 482
 483Skipping tests
 484--------------
 485
 486If you need to skip tests you should do so by using the three-arg form
 487of the test_* functions (see the "Test harness library" section
 488below), e.g.:
 489
 490    test_expect_success PERL 'I need Perl' '
 491        perl -e "hlagh() if unf_unf()"
 492    '
 493
 494The advantage of skipping tests like this is that platforms that don't
 495have the PERL and other optional dependencies get an indication of how
 496many tests they're missing.
 497
 498If the test code is too hairy for that (i.e. does a lot of setup work
 499outside test assertions) you can also skip all remaining tests by
 500setting skip_all and immediately call test_done:
 501
 502        if ! test_have_prereq PERL
 503        then
 504            skip_all='skipping perl interface tests, perl not available'
 505            test_done
 506        fi
 507
 508The string you give to skip_all will be used as an explanation for why
 509the test was skipped.
 510
 511End with test_done
 512------------------
 513
 514Your script will be a sequence of tests, using helper functions
 515from the test harness library.  At the end of the script, call
 516'test_done'.
 517
 518
 519Test harness library
 520--------------------
 521
 522There are a handful helper functions defined in the test harness
 523library for your script to use.
 524
 525 - test_expect_success [<prereq>] <message> <script>
 526
 527   Usually takes two strings as parameters, and evaluates the
 528   <script>.  If it yields success, test is considered
 529   successful.  <message> should state what it is testing.
 530
 531   Example:
 532
 533        test_expect_success \
 534            'git-write-tree should be able to write an empty tree.' \
 535            'tree=$(git-write-tree)'
 536
 537   If you supply three parameters the first will be taken to be a
 538   prerequisite; see the test_set_prereq and test_have_prereq
 539   documentation below:
 540
 541        test_expect_success TTY 'git --paginate rev-list uses a pager' \
 542            ' ... '
 543
 544   You can also supply a comma-separated list of prerequisites, in the
 545   rare case where your test depends on more than one:
 546
 547        test_expect_success PERL,PYTHON 'yo dawg' \
 548            ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" '
 549
 550 - test_expect_failure [<prereq>] <message> <script>
 551
 552   This is NOT the opposite of test_expect_success, but is used
 553   to mark a test that demonstrates a known breakage.  Unlike
 554   the usual test_expect_success tests, which say "ok" on
 555   success and "FAIL" on failure, this will say "FIXED" on
 556   success and "still broken" on failure.  Failures from these
 557   tests won't cause -i (immediate) to stop.
 558
 559   Like test_expect_success this function can optionally use a three
 560   argument invocation with a prerequisite as the first argument.
 561
 562 - test_debug <script>
 563
 564   This takes a single argument, <script>, and evaluates it only
 565   when the test script is started with --debug command line
 566   argument.  This is primarily meant for use during the
 567   development of a new test script.
 568
 569 - debug <git-command>
 570
 571   Run a git command inside a debugger. This is primarily meant for
 572   use when debugging a failing test script.
 573
 574 - test_done
 575
 576   Your test script must have test_done at the end.  Its purpose
 577   is to summarize successes and failures in the test script and
 578   exit with an appropriate error code.
 579
 580 - test_tick
 581
 582   Make commit and tag names consistent by setting the author and
 583   committer times to defined state.  Subsequent calls will
 584   advance the times by a fixed amount.
 585
 586 - test_commit <message> [<filename> [<contents>]]
 587
 588   Creates a commit with the given message, committing the given
 589   file with the given contents (default for both is to reuse the
 590   message string), and adds a tag (again reusing the message
 591   string as name).  Calls test_tick to make the SHA-1s
 592   reproducible.
 593
 594 - test_merge <message> <commit-or-tag>
 595
 596   Merges the given rev using the given message.  Like test_commit,
 597   creates a tag and calls test_tick before committing.
 598
 599 - test_set_prereq <prereq>
 600
 601   Set a test prerequisite to be used later with test_have_prereq. The
 602   test-lib will set some prerequisites for you, see the
 603   "Prerequisites" section below for a full list of these.
 604
 605   Others you can set yourself and use later with either
 606   test_have_prereq directly, or the three argument invocation of
 607   test_expect_success and test_expect_failure.
 608
 609 - test_have_prereq <prereq>
 610
 611   Check if we have a prerequisite previously set with test_set_prereq.
 612   The most common way to use this explicitly (as opposed to the
 613   implicit use when an argument is passed to test_expect_*) is to skip
 614   all the tests at the start of the test script if we don't have some
 615   essential prerequisite:
 616
 617        if ! test_have_prereq PERL
 618        then
 619            skip_all='skipping perl interface tests, perl not available'
 620            test_done
 621        fi
 622
 623 - test_external [<prereq>] <message> <external> <script>
 624
 625   Execute a <script> with an <external> interpreter (like perl). This
 626   was added for tests like t9700-perl-git.sh which do most of their
 627   work in an external test script.
 628
 629        test_external \
 630            'GitwebCache::*FileCache*' \
 631            perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl
 632
 633   If the test is outputting its own TAP you should set the
 634   test_external_has_tap variable somewhere before calling the first
 635   test_external* function. See t9700-perl-git.sh for an example.
 636
 637        # The external test will outputs its own plan
 638        test_external_has_tap=1
 639
 640 - test_external_without_stderr [<prereq>] <message> <external> <script>
 641
 642   Like test_external but fail if there's any output on stderr,
 643   instead of checking the exit code.
 644
 645        test_external_without_stderr \
 646            'Perl API' \
 647            perl "$TEST_DIRECTORY"/t9700/test.pl
 648
 649 - test_expect_code <exit-code> <command>
 650
 651   Run a command and ensure that it exits with the given exit code.
 652   For example:
 653
 654        test_expect_success 'Merge with d/f conflicts' '
 655                test_expect_code 1 git merge "merge msg" B master
 656        '
 657
 658 - test_must_fail <git-command>
 659
 660   Run a git command and ensure it fails in a controlled way.  Use
 661   this instead of "! <git-command>".  When git-command dies due to a
 662   segfault, test_must_fail diagnoses it as an error; "! <git-command>"
 663   treats it as just another expected failure, which would let such a
 664   bug go unnoticed.
 665
 666 - test_might_fail <git-command>
 667
 668   Similar to test_must_fail, but tolerate success, too.  Use this
 669   instead of "<git-command> || :" to catch failures due to segv.
 670
 671 - test_cmp <expected> <actual>
 672
 673   Check whether the content of the <actual> file matches the
 674   <expected> file.  This behaves like "cmp" but produces more
 675   helpful output when the test is run with "-v" option.
 676
 677 - test_cmp_rev <expected> <actual>
 678
 679   Check whether the <expected> rev points to the same commit as the
 680   <actual> rev.
 681
 682 - test_line_count (= | -lt | -ge | ...) <length> <file>
 683
 684   Check whether a file has the length it is expected to.
 685
 686 - test_path_is_file <path> [<diagnosis>]
 687   test_path_is_dir <path> [<diagnosis>]
 688   test_path_is_missing <path> [<diagnosis>]
 689
 690   Check if the named path is a file, if the named path is a
 691   directory, or if the named path does not exist, respectively,
 692   and fail otherwise, showing the <diagnosis> text.
 693
 694 - test_when_finished <script>
 695
 696   Prepend <script> to a list of commands to run to clean up
 697   at the end of the current test.  If some clean-up command
 698   fails, the test will not pass.
 699
 700   Example:
 701
 702        test_expect_success 'branch pointing to non-commit' '
 703                git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
 704                test_when_finished "git update-ref -d refs/heads/invalid" &&
 705                ...
 706        '
 707
 708 - test_write_lines <lines>
 709
 710   Write <lines> on standard output, one line per argument.
 711   Useful to prepare multi-line files in a compact form.
 712
 713   Example:
 714
 715        test_write_lines a b c d e f g >foo
 716
 717   Is a more compact equivalent of:
 718        cat >foo <<-EOF
 719        a
 720        b
 721        c
 722        d
 723        e
 724        f
 725        g
 726        EOF
 727
 728
 729 - test_pause
 730
 731        This command is useful for writing and debugging tests and must be
 732        removed before submitting. It halts the execution of the test and
 733        spawns a shell in the trash directory. Exit the shell to continue
 734        the test. Example:
 735
 736        test_expect_success 'test' '
 737                git do-something >actual &&
 738                test_pause &&
 739                test_cmp expected actual
 740        '
 741
 742 - test_ln_s_add <path1> <path2>
 743
 744   This function helps systems whose filesystem does not support symbolic
 745   links. Use it to add a symbolic link entry to the index when it is not
 746   important that the file system entry is a symbolic link, i.e., instead
 747   of the sequence
 748
 749        ln -s foo bar &&
 750        git add bar
 751
 752   Sometimes it is possible to split a test in a part that does not need
 753   the symbolic link in the file system and a part that does; then only
 754   the latter part need be protected by a SYMLINKS prerequisite (see below).
 755
 756Prerequisites
 757-------------
 758
 759These are the prerequisites that the test library predefines with
 760test_have_prereq.
 761
 762See the prereq argument to the test_* functions in the "Test harness
 763library" section above and the "test_have_prereq" function for how to
 764use these, and "test_set_prereq" for how to define your own.
 765
 766 - PYTHON
 767
 768   Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that
 769   need Python with this.
 770
 771 - PERL
 772
 773   Git wasn't compiled with NO_PERL=YesPlease.
 774
 775   Even without the PERL prerequisite, tests can assume there is a
 776   usable perl interpreter at $PERL_PATH, though it need not be
 777   particularly modern.
 778
 779 - POSIXPERM
 780
 781   The filesystem supports POSIX style permission bits.
 782
 783 - BSLASHPSPEC
 784
 785   Backslashes in pathspec are not directory separators. This is not
 786   set on Windows. See 6fd1106a for details.
 787
 788 - EXECKEEPSPID
 789
 790   The process retains the same pid across exec(2). See fb9a2bea for
 791   details.
 792
 793 - PIPE
 794
 795   The filesystem we're on supports creation of FIFOs (named pipes)
 796   via mkfifo(1).
 797
 798 - SYMLINKS
 799
 800   The filesystem we're on supports symbolic links. E.g. a FAT
 801   filesystem doesn't support these. See 704a3143 for details.
 802
 803 - SANITY
 804
 805   Test is not run by root user, and an attempt to write to an
 806   unwritable file is expected to fail correctly.
 807
 808 - PCRE
 809
 810   Git was compiled with support for PCRE. Wrap any tests
 811   that use git-grep --perl-regexp or git-grep -P in these.
 812
 813 - LIBPCRE1
 814
 815   Git was compiled with PCRE v1 support via
 816   USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some
 817   reason need v1 of the PCRE library instead of v2 in these.
 818
 819 - LIBPCRE2
 820
 821   Git was compiled with PCRE v2 support via
 822   USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some
 823   reason need v2 of the PCRE library instead of v1 in these.
 824
 825 - CASE_INSENSITIVE_FS
 826
 827   Test is run on a case insensitive file system.
 828
 829 - UTF8_NFD_TO_NFC
 830
 831   Test is run on a filesystem which converts decomposed utf-8 (nfd)
 832   to precomposed utf-8 (nfc).
 833
 834 - PTHREADS
 835
 836   Git wasn't compiled with NO_PTHREADS=YesPlease.
 837
 838Tips for Writing Tests
 839----------------------
 840
 841As with any programming projects, existing programs are the best
 842source of the information.  However, do _not_ emulate
 843t0000-basic.sh when writing your tests.  The test is special in
 844that it tries to validate the very core of GIT.  For example, it
 845knows that there will be 256 subdirectories under .git/objects/,
 846and it knows that the object ID of an empty tree is a certain
 84740-byte string.  This is deliberately done so in t0000-basic.sh
 848because the things the very basic core test tries to achieve is
 849to serve as a basis for people who are changing the GIT internal
 850drastically.  For these people, after making certain changes,
 851not seeing failures from the basic test _is_ a failure.  And
 852such drastic changes to the core GIT that even changes these
 853otherwise supposedly stable object IDs should be accompanied by
 854an update to t0000-basic.sh.
 855
 856However, other tests that simply rely on basic parts of the core
 857GIT working properly should not have that level of intimate
 858knowledge of the core GIT internals.  If all the test scripts
 859hardcoded the object IDs like t0000-basic.sh does, that defeats
 860the purpose of t0000-basic.sh, which is to isolate that level of
 861validation in one place.  Your test also ends up needing
 862updating when such a change to the internal happens, so do _not_
 863do it and leave the low level of validation to t0000-basic.sh.
 864
 865Test coverage
 866-------------
 867
 868You can use the coverage tests to find code paths that are not being
 869used or properly exercised yet.
 870
 871To do that, run the coverage target at the top-level (not in the t/
 872directory):
 873
 874    make coverage
 875
 876That'll compile Git with GCC's coverage arguments, and generate a test
 877report with gcov after the tests finish. Running the coverage tests
 878can take a while, since running the tests in parallel is incompatible
 879with GCC's coverage mode.
 880
 881After the tests have run you can generate a list of untested
 882functions:
 883
 884    make coverage-untested-functions
 885
 886You can also generate a detailed per-file HTML report using the
 887Devel::Cover module. To install it do:
 888
 889   # On Debian or Ubuntu:
 890   sudo aptitude install libdevel-cover-perl
 891
 892   # From the CPAN with cpanminus
 893   curl -L http://cpanmin.us | perl - --sudo --self-upgrade
 894   cpanm --sudo Devel::Cover
 895
 896Then, at the top-level:
 897
 898    make cover_db_html
 899
 900That'll generate a detailed cover report in the "cover_db_html"
 901directory, which you can then copy to a webserver, or inspect locally
 902in a browser.