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. 188 189--stress:: 190--stress=<N>:: 191 Run the test script repeatedly in multiple parallel jobs until 192 one of them fails. Useful for reproducing rare failures in 193 flaky tests. The number of parallel jobs is, in order of 194 precedence: <N>, or the value of the GIT_TEST_STRESS_LOAD 195 environment variable, or twice the number of available 196 processors (as shown by the 'getconf' utility), or 8. 197 Implies `--verbose -x --immediate` to get the most information 198 about the failure. Note that the verbose output of each test 199 job is saved to 't/test-results/$TEST_NAME.stress-<nr>.out', 200 and only the output of the failed test job is shown on the 201 terminal. The names of the trash directories get a 202 '.stress-<nr>' suffix, and the trash directory of the failed 203 test job is renamed to end with a '.stress-failed' suffix. 204 205--stress-limit=<N>:: 206 When combined with --stress run the test script repeatedly 207 this many times in each of the parallel jobs or until one of 208 them fails, whichever comes first. 209 210You can also set the GIT_TEST_INSTALLED environment variable to 211the bindir of an existing git installation to test that installation. 212You still need to have built this git sandbox, from which various 213test-* support programs, templates, and perl libraries are used. 214If your installed git is incomplete, it will silently test parts of 215your built version instead. 216 217When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to 218override the location of the dashed-form subcommands (what 219GIT_EXEC_PATH would be used for during normal operation). 220GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`. 221 222 223Skipping Tests 224-------------- 225 226In some environments, certain tests have no way of succeeding 227due to platform limitation, such as lack of 'unzip' program, or 228filesystem that do not allow arbitrary sequence of non-NUL bytes 229as pathnames. 230 231You should be able to say something like 232 233 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh 234 235and even: 236 237 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make 238 239to omit such tests. The value of the environment variable is a 240SP separated list of patterns that tells which tests to skip, 241and either can match the "t[0-9]{4}" part to skip the whole 242test, or t[0-9]{4} followed by ".$number" to say which 243particular test to skip. 244 245For an individual test suite --run could be used to specify that 246only some tests should be run or that some tests should be 247excluded from a run. 248 249The argument for --run is a list of individual test numbers or 250ranges with an optional negation prefix that define what tests in 251a test suite to include in the run. A range is two numbers 252separated with a dash and matches a range of tests with both ends 253been included. You may omit the first or the second number to 254mean "from the first test" or "up to the very last test" 255respectively. 256 257Optional prefix of '!' means that the test or a range of tests 258should be excluded from the run. 259 260If --run starts with an unprefixed number or range the initial 261set of tests to run is empty. If the first item starts with '!' 262all the tests are added to the initial set. After initial set is 263determined every test number or range is added or excluded from 264the set one by one, from left to right. 265 266Individual numbers or ranges could be separated either by a space 267or a comma. 268 269For example, to run only tests up to a specific test (21), one 270could do this: 271 272 $ sh ./t9200-git-cvsexport-commit.sh --run='1-21' 273 274or this: 275 276 $ sh ./t9200-git-cvsexport-commit.sh --run='-21' 277 278Common case is to run several setup tests (1, 2, 3) and then a 279specific test (21) that relies on that setup: 280 281 $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21' 282 283or: 284 285 $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21 286 287or: 288 289 $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21' 290 291As noted above, the test set is built by going through the items 292from left to right, so this: 293 294 $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3' 295 296will run tests 1, 2, and 4. Items that come later have higher 297precedence. It means that this: 298 299 $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4' 300 301would just run tests from 1 to 4, including 3. 302 303You may use negation with ranges. The following will run all 304test in the test suite except from 7 up to 11: 305 306 $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11' 307 308Some tests in a test suite rely on the previous tests performing 309certain actions, specifically some tests are designated as 310"setup" test, so you cannot _arbitrarily_ disable one test and 311expect the rest to function correctly. 312 313--run is mostly useful when you want to focus on a specific test 314and know what setup is needed for it. Or when you want to run 315everything up to a certain test. 316 317 318Running tests with special setups 319--------------------------------- 320 321The whole test suite could be run to test some special features 322that cannot be easily covered by a few specific test cases. These 323could be enabled by running the test suite with correct GIT_TEST_ 324environment set. 325 326GIT_TEST_GETTEXT_POISON=<non-empty?> turns all strings marked for 327translation into gibberish if non-empty (think "test -n"). Used for 328spotting those tests that need to be marked with a C_LOCALE_OUTPUT 329prerequisite when adding more strings for translation. See "Testing 330marked strings" in po/README for details. 331 332GIT_TEST_SPLIT_INDEX=<boolean> forces split-index mode on the whole 333test suite. Accept any boolean values that are accepted by git-config. 334 335GIT_TEST_FULL_IN_PACK_ARRAY=<boolean> exercises the uncommon 336pack-objects code path where there are more than 1024 packs even if 337the actual number of packs in repository is below this limit. Accept 338any boolean values that are accepted by git-config. 339 340GIT_TEST_OE_SIZE=<n> exercises the uncommon pack-objects code path 341where we do not cache object size in memory and read it from existing 342packs on demand. This normally only happens when the object size is 343over 2GB. This variable forces the code path on any object larger than 344<n> bytes. 345 346GIT_TEST_OE_DELTA_SIZE=<n> exercises the uncommon pack-objects code 347path where deltas larger than this limit require extra memory 348allocation for bookkeeping. 349 350GIT_TEST_VALIDATE_INDEX_CACHE_ENTRIES=<boolean> checks that cache-tree 351records are valid when the index is written out or after a merge. This 352is mostly to catch missing invalidation. Default is true. 353 354GIT_TEST_COMMIT_GRAPH=<boolean>, when true, forces the commit-graph to 355be written after every 'git commit' command, and overrides the 356'core.commitGraph' setting to true. 357 358GIT_TEST_FSMONITOR=$PWD/t7519/fsmonitor-all exercises the fsmonitor 359code path for utilizing a file system monitor to speed up detecting 360new or changed files. 361 362GIT_TEST_INDEX_VERSION=<n> exercises the index read/write code path 363for the index version specified. Can be set to any valid version 364(currently 2, 3, or 4). 365 366GIT_TEST_PRELOAD_INDEX=<boolean> exercises the preload-index code path 367by overriding the minimum number of cache entries required per thread. 368 369GIT_TEST_REBASE_USE_BUILTIN=<boolean>, when false, disables the 370builtin version of git-rebase. See 'rebase.useBuiltin' in 371git-config(1). 372 373GIT_TEST_INDEX_THREADS=<n> enables exercising the multi-threaded loading 374of the index for the whole test suite by bypassing the default number of 375cache entries and thread minimums. Setting this to 1 will make the 376index loading single threaded. 377 378GIT_TEST_MULTI_PACK_INDEX=<boolean>, when true, forces the multi-pack- 379index to be written after every 'git repack' command, and overrides the 380'core.multiPackIndex' setting to true. 381 382Naming Tests 383------------ 384 385The test files are named as: 386 387 tNNNN-commandname-details.sh 388 389where N is a decimal digit. 390 391First digit tells the family: 392 393 0 - the absolute basics and global stuff 394 1 - the basic commands concerning database 395 2 - the basic commands concerning the working tree 396 3 - the other basic commands (e.g. ls-files) 397 4 - the diff commands 398 5 - the pull and exporting commands 399 6 - the revision tree commands (even e.g. merge-base) 400 7 - the porcelainish commands concerning the working tree 401 8 - the porcelainish commands concerning forensics 402 9 - the git tools 403 404Second digit tells the particular command we are testing. 405 406Third digit (optionally) tells the particular switch or group of switches 407we are testing. 408 409If you create files under t/ directory (i.e. here) that is not 410the top-level test script, never name the file to match the above 411pattern. The Makefile here considers all such files as the 412top-level test script and tries to run all of them. Care is 413especially needed if you are creating a common test library 414file, similar to test-lib.sh, because such a library file may 415not be suitable for standalone execution. 416 417 418Writing Tests 419------------- 420 421The test script is written as a shell script. It should start 422with the standard "#!/bin/sh", and an 423assignment to variable 'test_description', like this: 424 425 #!/bin/sh 426 427 test_description='xxx test (option --frotz) 428 429 This test registers the following structure in the cache 430 and tries to run git-ls-files with option --frotz.' 431 432 433Source 'test-lib.sh' 434-------------------- 435 436After assigning test_description, the test script should source 437test-lib.sh like this: 438 439 . ./test-lib.sh 440 441This test harness library does the following things: 442 443 - If the script is invoked with command line argument --help 444 (or -h), it shows the test_description and exits. 445 446 - Creates an empty test directory with an empty .git/objects database 447 and chdir(2) into it. This directory is 't/trash 448 directory.$test_name_without_dotsh', with t/ subject to change by 449 the --root option documented above, and a '.stress-<N>' suffix 450 appended by the --stress option. 451 452 - Defines standard test helper functions for your scripts to 453 use. These functions are designed to make all scripts behave 454 consistently when command line arguments --verbose (or -v), 455 --debug (or -d), and --immediate (or -i) is given. 456 457Do's & don'ts 458------------- 459 460Here are a few examples of things you probably should and shouldn't do 461when writing tests. 462 463Here are the "do's:" 464 465 - Put all code inside test_expect_success and other assertions. 466 467 Even code that isn't a test per se, but merely some setup code 468 should be inside a test assertion. 469 470 - Chain your test assertions 471 472 Write test code like this: 473 474 git merge foo && 475 git push bar && 476 test ... 477 478 Instead of: 479 480 git merge hla 481 git push gh 482 test ... 483 484 That way all of the commands in your tests will succeed or fail. If 485 you must ignore the return value of something, consider using a 486 helper function (e.g. use sane_unset instead of unset, in order 487 to avoid unportable return value for unsetting a variable that was 488 already unset), or prepending the command with test_might_fail or 489 test_must_fail. 490 491 - Check the test coverage for your tests. See the "Test coverage" 492 below. 493 494 Don't blindly follow test coverage metrics; if a new function you added 495 doesn't have any coverage, then you're probably doing something wrong, 496 but having 100% coverage doesn't necessarily mean that you tested 497 everything. 498 499 Tests that are likely to smoke out future regressions are better 500 than tests that just inflate the coverage metrics. 501 502 - When a test checks for an absolute path that a git command generated, 503 construct the expected value using $(pwd) rather than $PWD, 504 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on 505 Windows, where the shell (MSYS bash) mangles absolute path names. 506 For details, see the commit message of 4114156ae9. 507 508 - Remember that inside the <script> part, the standard output and 509 standard error streams are discarded, and the test harness only 510 reports "ok" or "not ok" to the end user running the tests. Under 511 --verbose, they are shown to help debug the tests. 512 513And here are the "don'ts:" 514 515 - Don't exit() within a <script> part. 516 517 The harness will catch this as a programming error of the test. 518 Use test_done instead if you need to stop the tests early (see 519 "Skipping tests" below). 520 521 - Don't use '! git cmd' when you want to make sure the git command 522 exits with failure in a controlled way by calling "die()". Instead, 523 use 'test_must_fail git cmd'. This will signal a failure if git 524 dies in an unexpected way (e.g. segfault). 525 526 On the other hand, don't use test_must_fail for running regular 527 platform commands; just use '! cmd'. We are not in the business 528 of verifying that the world given to us sanely works. 529 530 - Don't feed the output of a git command to a pipe, as in: 531 532 git -C repo ls-files | 533 xargs -n 1 basename | 534 grep foo 535 536 which will discard git's exit code and may mask a crash. In the 537 above example, all exit codes are ignored except grep's. 538 539 Instead, write the output of that command to a temporary 540 file with ">" or assign it to a variable with "x=$(git ...)" rather 541 than pipe it. 542 543 - Don't use command substitution in a way that discards git's exit 544 code. When assigning to a variable, the exit code is not discarded, 545 e.g.: 546 547 x=$(git cat-file -p $sha) && 548 ... 549 550 is OK because a crash in "git cat-file" will cause the "&&" chain 551 to fail, but: 552 553 test "refs/heads/foo" = "$(git symbolic-ref HEAD)" 554 555 is not OK and a crash in git could go undetected. 556 557 - Don't use perl without spelling it as "$PERL_PATH". This is to help 558 our friends on Windows where the platform Perl often adds CR before 559 the end of line, and they bundle Git with a version of Perl that 560 does not do so, whose path is specified with $PERL_PATH. Note that we 561 provide a "perl" function which uses $PERL_PATH under the hood, so 562 you do not need to worry when simply running perl in the test scripts 563 (but you do, for example, on a shebang line or in a sub script 564 created via "write_script"). 565 566 - Don't use sh without spelling it as "$SHELL_PATH", when the script 567 can be misinterpreted by broken platform shell (e.g. Solaris). 568 569 - Don't chdir around in tests. It is not sufficient to chdir to 570 somewhere and then chdir back to the original location later in 571 the test, as any intermediate step can fail and abort the test, 572 causing the next test to start in an unexpected directory. Do so 573 inside a subshell if necessary. 574 575 - Don't save and verify the standard error of compound commands, i.e. 576 group commands, subshells, and shell functions (except test helper 577 functions like 'test_must_fail') like this: 578 579 ( cd dir && git cmd ) 2>error && 580 test_cmp expect error 581 582 When running the test with '-x' tracing, then the trace of commands 583 executed in the compound command will be included in standard error 584 as well, quite possibly throwing off the subsequent checks examining 585 the output. Instead, save only the relevant git command's standard 586 error: 587 588 ( cd dir && git cmd 2>../error ) && 589 test_cmp expect error 590 591 - Don't break the TAP output 592 593 The raw output from your test may be interpreted by a TAP harness. TAP 594 harnesses will ignore everything they don't know about, but don't step 595 on their toes in these areas: 596 597 - Don't print lines like "$x..$y" where $x and $y are integers. 598 599 - Don't print lines that begin with "ok" or "not ok". 600 601 TAP harnesses expect a line that begins with either "ok" and "not 602 ok" to signal a test passed or failed (and our harness already 603 produces such lines), so your script shouldn't emit such lines to 604 their output. 605 606 You can glean some further possible issues from the TAP grammar 607 (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR) 608 but the best indication is to just run the tests with prove(1), 609 it'll complain if anything is amiss. 610 611 612Skipping tests 613-------------- 614 615If you need to skip tests you should do so by using the three-arg form 616of the test_* functions (see the "Test harness library" section 617below), e.g.: 618 619 test_expect_success PERL 'I need Perl' ' 620 perl -e "hlagh() if unf_unf()" 621 ' 622 623The advantage of skipping tests like this is that platforms that don't 624have the PERL and other optional dependencies get an indication of how 625many tests they're missing. 626 627If the test code is too hairy for that (i.e. does a lot of setup work 628outside test assertions) you can also skip all remaining tests by 629setting skip_all and immediately call test_done: 630 631 if ! test_have_prereq PERL 632 then 633 skip_all='skipping perl interface tests, perl not available' 634 test_done 635 fi 636 637The string you give to skip_all will be used as an explanation for why 638the test was skipped. 639 640End with test_done 641------------------ 642 643Your script will be a sequence of tests, using helper functions 644from the test harness library. At the end of the script, call 645'test_done'. 646 647 648Test harness library 649-------------------- 650 651There are a handful helper functions defined in the test harness 652library for your script to use. 653 654 - test_expect_success [<prereq>] <message> <script> 655 656 Usually takes two strings as parameters, and evaluates the 657 <script>. If it yields success, test is considered 658 successful. <message> should state what it is testing. 659 660 Example: 661 662 test_expect_success \ 663 'git-write-tree should be able to write an empty tree.' \ 664 'tree=$(git-write-tree)' 665 666 If you supply three parameters the first will be taken to be a 667 prerequisite; see the test_set_prereq and test_have_prereq 668 documentation below: 669 670 test_expect_success TTY 'git --paginate rev-list uses a pager' \ 671 ' ... ' 672 673 You can also supply a comma-separated list of prerequisites, in the 674 rare case where your test depends on more than one: 675 676 test_expect_success PERL,PYTHON 'yo dawg' \ 677 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" ' 678 679 - test_expect_failure [<prereq>] <message> <script> 680 681 This is NOT the opposite of test_expect_success, but is used 682 to mark a test that demonstrates a known breakage. Unlike 683 the usual test_expect_success tests, which say "ok" on 684 success and "FAIL" on failure, this will say "FIXED" on 685 success and "still broken" on failure. Failures from these 686 tests won't cause -i (immediate) to stop. 687 688 Like test_expect_success this function can optionally use a three 689 argument invocation with a prerequisite as the first argument. 690 691 - test_debug <script> 692 693 This takes a single argument, <script>, and evaluates it only 694 when the test script is started with --debug command line 695 argument. This is primarily meant for use during the 696 development of a new test script. 697 698 - debug <git-command> 699 700 Run a git command inside a debugger. This is primarily meant for 701 use when debugging a failing test script. 702 703 - test_done 704 705 Your test script must have test_done at the end. Its purpose 706 is to summarize successes and failures in the test script and 707 exit with an appropriate error code. 708 709 - test_tick 710 711 Make commit and tag names consistent by setting the author and 712 committer times to defined state. Subsequent calls will 713 advance the times by a fixed amount. 714 715 - test_commit <message> [<filename> [<contents>]] 716 717 Creates a commit with the given message, committing the given 718 file with the given contents (default for both is to reuse the 719 message string), and adds a tag (again reusing the message 720 string as name). Calls test_tick to make the SHA-1s 721 reproducible. 722 723 - test_merge <message> <commit-or-tag> 724 725 Merges the given rev using the given message. Like test_commit, 726 creates a tag and calls test_tick before committing. 727 728 - test_set_prereq <prereq> 729 730 Set a test prerequisite to be used later with test_have_prereq. The 731 test-lib will set some prerequisites for you, see the 732 "Prerequisites" section below for a full list of these. 733 734 Others you can set yourself and use later with either 735 test_have_prereq directly, or the three argument invocation of 736 test_expect_success and test_expect_failure. 737 738 - test_have_prereq <prereq> 739 740 Check if we have a prerequisite previously set with test_set_prereq. 741 The most common way to use this explicitly (as opposed to the 742 implicit use when an argument is passed to test_expect_*) is to skip 743 all the tests at the start of the test script if we don't have some 744 essential prerequisite: 745 746 if ! test_have_prereq PERL 747 then 748 skip_all='skipping perl interface tests, perl not available' 749 test_done 750 fi 751 752 - test_external [<prereq>] <message> <external> <script> 753 754 Execute a <script> with an <external> interpreter (like perl). This 755 was added for tests like t9700-perl-git.sh which do most of their 756 work in an external test script. 757 758 test_external \ 759 'GitwebCache::*FileCache*' \ 760 perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl 761 762 If the test is outputting its own TAP you should set the 763 test_external_has_tap variable somewhere before calling the first 764 test_external* function. See t9700-perl-git.sh for an example. 765 766 # The external test will outputs its own plan 767 test_external_has_tap=1 768 769 - test_external_without_stderr [<prereq>] <message> <external> <script> 770 771 Like test_external but fail if there's any output on stderr, 772 instead of checking the exit code. 773 774 test_external_without_stderr \ 775 'Perl API' \ 776 perl "$TEST_DIRECTORY"/t9700/test.pl 777 778 - test_expect_code <exit-code> <command> 779 780 Run a command and ensure that it exits with the given exit code. 781 For example: 782 783 test_expect_success 'Merge with d/f conflicts' ' 784 test_expect_code 1 git merge "merge msg" B master 785 ' 786 787 - test_must_fail [<options>] <git-command> 788 789 Run a git command and ensure it fails in a controlled way. Use 790 this instead of "! <git-command>". When git-command dies due to a 791 segfault, test_must_fail diagnoses it as an error; "! <git-command>" 792 treats it as just another expected failure, which would let such a 793 bug go unnoticed. 794 795 Accepts the following options: 796 797 ok=<signal-name>[,<...>]: 798 Don't treat an exit caused by the given signal as error. 799 Multiple signals can be specified as a comma separated list. 800 Currently recognized signal names are: sigpipe, success. 801 (Don't use 'success', use 'test_might_fail' instead.) 802 803 - test_might_fail [<options>] <git-command> 804 805 Similar to test_must_fail, but tolerate success, too. Use this 806 instead of "<git-command> || :" to catch failures due to segv. 807 808 Accepts the same options as test_must_fail. 809 810 - test_cmp <expected> <actual> 811 812 Check whether the content of the <actual> file matches the 813 <expected> file. This behaves like "cmp" but produces more 814 helpful output when the test is run with "-v" option. 815 816 - test_cmp_rev <expected> <actual> 817 818 Check whether the <expected> rev points to the same commit as the 819 <actual> rev. 820 821 - test_line_count (= | -lt | -ge | ...) <length> <file> 822 823 Check whether a file has the length it is expected to. 824 825 - test_path_is_file <path> [<diagnosis>] 826 test_path_is_dir <path> [<diagnosis>] 827 test_path_is_missing <path> [<diagnosis>] 828 829 Check if the named path is a file, if the named path is a 830 directory, or if the named path does not exist, respectively, 831 and fail otherwise, showing the <diagnosis> text. 832 833 - test_when_finished <script> 834 835 Prepend <script> to a list of commands to run to clean up 836 at the end of the current test. If some clean-up command 837 fails, the test will not pass. 838 839 Example: 840 841 test_expect_success 'branch pointing to non-commit' ' 842 git rev-parse HEAD^{tree} >.git/refs/heads/invalid && 843 test_when_finished "git update-ref -d refs/heads/invalid" && 844 ... 845 ' 846 847 - test_write_lines <lines> 848 849 Write <lines> on standard output, one line per argument. 850 Useful to prepare multi-line files in a compact form. 851 852 Example: 853 854 test_write_lines a b c d e f g >foo 855 856 Is a more compact equivalent of: 857 cat >foo <<-EOF 858 a 859 b 860 c 861 d 862 e 863 f 864 g 865 EOF 866 867 868 - test_pause 869 870 This command is useful for writing and debugging tests and must be 871 removed before submitting. It halts the execution of the test and 872 spawns a shell in the trash directory. Exit the shell to continue 873 the test. Example: 874 875 test_expect_success 'test' ' 876 git do-something >actual && 877 test_pause && 878 test_cmp expected actual 879 ' 880 881 - test_ln_s_add <path1> <path2> 882 883 This function helps systems whose filesystem does not support symbolic 884 links. Use it to add a symbolic link entry to the index when it is not 885 important that the file system entry is a symbolic link, i.e., instead 886 of the sequence 887 888 ln -s foo bar && 889 git add bar 890 891 Sometimes it is possible to split a test in a part that does not need 892 the symbolic link in the file system and a part that does; then only 893 the latter part need be protected by a SYMLINKS prerequisite (see below). 894 895 - test_oid_init 896 897 This function loads facts and useful object IDs related to the hash 898 algorithm(s) in use from the files in t/oid-info. 899 900 - test_oid_cache 901 902 This function reads per-hash algorithm information from standard 903 input (usually a heredoc) in the format described in 904 t/oid-info/README. This is useful for test-specific values, such as 905 object IDs, which must vary based on the hash algorithm. 906 907 Certain fixed values, such as hash sizes and common placeholder 908 object IDs, can be loaded with test_oid_init (described above). 909 910 - test_oid <key> 911 912 This function looks up a value for the hash algorithm in use, based 913 on the key given. The value must have been loaded using 914 test_oid_init or test_oid_cache. Providing an unknown key is an 915 error. 916 917Prerequisites 918------------- 919 920These are the prerequisites that the test library predefines with 921test_have_prereq. 922 923See the prereq argument to the test_* functions in the "Test harness 924library" section above and the "test_have_prereq" function for how to 925use these, and "test_set_prereq" for how to define your own. 926 927 - PYTHON 928 929 Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that 930 need Python with this. 931 932 - PERL 933 934 Git wasn't compiled with NO_PERL=YesPlease. 935 936 Even without the PERL prerequisite, tests can assume there is a 937 usable perl interpreter at $PERL_PATH, though it need not be 938 particularly modern. 939 940 - POSIXPERM 941 942 The filesystem supports POSIX style permission bits. 943 944 - BSLASHPSPEC 945 946 Backslashes in pathspec are not directory separators. This is not 947 set on Windows. See 6fd1106a for details. 948 949 - EXECKEEPSPID 950 951 The process retains the same pid across exec(2). See fb9a2bea for 952 details. 953 954 - PIPE 955 956 The filesystem we're on supports creation of FIFOs (named pipes) 957 via mkfifo(1). 958 959 - SYMLINKS 960 961 The filesystem we're on supports symbolic links. E.g. a FAT 962 filesystem doesn't support these. See 704a3143 for details. 963 964 - SANITY 965 966 Test is not run by root user, and an attempt to write to an 967 unwritable file is expected to fail correctly. 968 969 - PCRE 970 971 Git was compiled with support for PCRE. Wrap any tests 972 that use git-grep --perl-regexp or git-grep -P in these. 973 974 - LIBPCRE1 975 976 Git was compiled with PCRE v1 support via 977 USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some 978 reason need v1 of the PCRE library instead of v2 in these. 979 980 - LIBPCRE2 981 982 Git was compiled with PCRE v2 support via 983 USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some 984 reason need v2 of the PCRE library instead of v1 in these. 985 986 - CASE_INSENSITIVE_FS 987 988 Test is run on a case insensitive file system. 989 990 - UTF8_NFD_TO_NFC 991 992 Test is run on a filesystem which converts decomposed utf-8 (nfd) 993 to precomposed utf-8 (nfc). 994 995 - PTHREADS 996 997 Git wasn't compiled with NO_PTHREADS=YesPlease. 998 999Tips for Writing Tests1000----------------------10011002As with any programming projects, existing programs are the best1003source of the information. However, do _not_ emulate1004t0000-basic.sh when writing your tests. The test is special in1005that it tries to validate the very core of GIT. For example, it1006knows that there will be 256 subdirectories under .git/objects/,1007and it knows that the object ID of an empty tree is a certain100840-byte string. This is deliberately done so in t0000-basic.sh1009because the things the very basic core test tries to achieve is1010to serve as a basis for people who are changing the GIT internal1011drastically. For these people, after making certain changes,1012not seeing failures from the basic test _is_ a failure. And1013such drastic changes to the core GIT that even changes these1014otherwise supposedly stable object IDs should be accompanied by1015an update to t0000-basic.sh.10161017However, other tests that simply rely on basic parts of the core1018GIT working properly should not have that level of intimate1019knowledge of the core GIT internals. If all the test scripts1020hardcoded the object IDs like t0000-basic.sh does, that defeats1021the purpose of t0000-basic.sh, which is to isolate that level of1022validation in one place. Your test also ends up needing1023updating when such a change to the internal happens, so do _not_1024do it and leave the low level of validation to t0000-basic.sh.10251026Test coverage1027-------------10281029You can use the coverage tests to find code paths that are not being1030used or properly exercised yet.10311032To do that, run the coverage target at the top-level (not in the t/1033directory):10341035 make coverage10361037That'll compile Git with GCC's coverage arguments, and generate a test1038report with gcov after the tests finish. Running the coverage tests1039can take a while, since running the tests in parallel is incompatible1040with GCC's coverage mode.10411042After the tests have run you can generate a list of untested1043functions:10441045 make coverage-untested-functions10461047You can also generate a detailed per-file HTML report using the1048Devel::Cover module. To install it do:10491050 # On Debian or Ubuntu:1051 sudo aptitude install libdevel-cover-perl10521053 # From the CPAN with cpanminus1054 curl -L http://cpanmin.us | perl - --sudo --self-upgrade1055 cpanm --sudo Devel::Cover10561057Then, at the top-level:10581059 make cover_db_html10601061That'll generate a detailed cover report in the "cover_db_html"1062directory, which you can then copy to a webserver, or inspect locally1063in a browser.