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--no-bin-wrappers:: 174 By default, the test suite uses the wrappers in 175 `../bin-wrappers/` to execute `git` and friends. With this option, 176 `../git` and friends are run directly. This is not recommended 177 in general, as the wrappers contain safeguards to ensure that no 178 files from an installed Git are used, but can speed up test runs 179 especially on platforms where running shell scripts is expensive 180 (most notably, Windows). 181 182--root=<directory>:: 183 Create "trash" directories used to store all temporary data during 184 testing under <directory>, instead of the t/ directory. 185 Using this option with a RAM-based filesystem (such as tmpfs) 186 can massively speed up the test suite. 187 188--chain-lint:: 189--no-chain-lint:: 190 If --chain-lint is enabled, the test harness will check each 191 test to make sure that it properly "&&-chains" all commands (so 192 that a failure in the middle does not go unnoticed by the final 193 exit code of the test). This check is performed in addition to 194 running the tests themselves. You may also enable or disable 195 this feature by setting the GIT_TEST_CHAIN_LINT environment 196 variable to "1" or "0", respectively. 197 198--stress:: 199--stress=<N>:: 200 Run the test script repeatedly in multiple parallel jobs until 201 one of them fails. Useful for reproducing rare failures in 202 flaky tests. The number of parallel jobs is, in order of 203 precedence: <N>, or the value of the GIT_TEST_STRESS_LOAD 204 environment variable, or twice the number of available 205 processors (as shown by the 'getconf' utility), or 8. 206 Implies `--verbose -x --immediate` to get the most information 207 about the failure. Note that the verbose output of each test 208 job is saved to 't/test-results/$TEST_NAME.stress-<nr>.out', 209 and only the output of the failed test job is shown on the 210 terminal. The names of the trash directories get a 211 '.stress-<nr>' suffix, and the trash directory of the failed 212 test job is renamed to end with a '.stress-failed' suffix. 213 214--stress-limit=<N>:: 215 When combined with --stress run the test script repeatedly 216 this many times in each of the parallel jobs or until one of 217 them fails, whichever comes first. 218 219You can also set the GIT_TEST_INSTALLED environment variable to 220the bindir of an existing git installation to test that installation. 221You still need to have built this git sandbox, from which various 222test-* support programs, templates, and perl libraries are used. 223If your installed git is incomplete, it will silently test parts of 224your built version instead. 225 226When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to 227override the location of the dashed-form subcommands (what 228GIT_EXEC_PATH would be used for during normal operation). 229GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`. 230 231 232Skipping Tests 233-------------- 234 235In some environments, certain tests have no way of succeeding 236due to platform limitation, such as lack of 'unzip' program, or 237filesystem that do not allow arbitrary sequence of non-NUL bytes 238as pathnames. 239 240You should be able to say something like 241 242 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh 243 244and even: 245 246 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make 247 248to omit such tests. The value of the environment variable is a 249SP separated list of patterns that tells which tests to skip, 250and either can match the "t[0-9]{4}" part to skip the whole 251test, or t[0-9]{4} followed by ".$number" to say which 252particular test to skip. 253 254For an individual test suite --run could be used to specify that 255only some tests should be run or that some tests should be 256excluded from a run. 257 258The argument for --run is a list of individual test numbers or 259ranges with an optional negation prefix that define what tests in 260a test suite to include in the run. A range is two numbers 261separated with a dash and matches a range of tests with both ends 262been included. You may omit the first or the second number to 263mean "from the first test" or "up to the very last test" 264respectively. 265 266Optional prefix of '!' means that the test or a range of tests 267should be excluded from the run. 268 269If --run starts with an unprefixed number or range the initial 270set of tests to run is empty. If the first item starts with '!' 271all the tests are added to the initial set. After initial set is 272determined every test number or range is added or excluded from 273the set one by one, from left to right. 274 275Individual numbers or ranges could be separated either by a space 276or a comma. 277 278For example, to run only tests up to a specific test (21), one 279could do this: 280 281 $ sh ./t9200-git-cvsexport-commit.sh --run='1-21' 282 283or this: 284 285 $ sh ./t9200-git-cvsexport-commit.sh --run='-21' 286 287Common case is to run several setup tests (1, 2, 3) and then a 288specific test (21) that relies on that setup: 289 290 $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21' 291 292or: 293 294 $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21 295 296or: 297 298 $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21' 299 300As noted above, the test set is built by going through the items 301from left to right, so this: 302 303 $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3' 304 305will run tests 1, 2, and 4. Items that come later have higher 306precedence. It means that this: 307 308 $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4' 309 310would just run tests from 1 to 4, including 3. 311 312You may use negation with ranges. The following will run all 313test in the test suite except from 7 up to 11: 314 315 $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11' 316 317Some tests in a test suite rely on the previous tests performing 318certain actions, specifically some tests are designated as 319"setup" test, so you cannot _arbitrarily_ disable one test and 320expect the rest to function correctly. 321 322--run is mostly useful when you want to focus on a specific test 323and know what setup is needed for it. Or when you want to run 324everything up to a certain test. 325 326 327Running tests with special setups 328--------------------------------- 329 330The whole test suite could be run to test some special features 331that cannot be easily covered by a few specific test cases. These 332could be enabled by running the test suite with correct GIT_TEST_ 333environment set. 334 335GIT_TEST_GETTEXT_POISON=<non-empty?> turns all strings marked for 336translation into gibberish if non-empty (think "test -n"). Used for 337spotting those tests that need to be marked with a C_LOCALE_OUTPUT 338prerequisite when adding more strings for translation. See "Testing 339marked strings" in po/README for details. 340 341GIT_TEST_SPLIT_INDEX=<boolean> forces split-index mode on the whole 342test suite. Accept any boolean values that are accepted by git-config. 343 344GIT_TEST_FULL_IN_PACK_ARRAY=<boolean> exercises the uncommon 345pack-objects code path where there are more than 1024 packs even if 346the actual number of packs in repository is below this limit. Accept 347any boolean values that are accepted by git-config. 348 349GIT_TEST_OE_SIZE=<n> exercises the uncommon pack-objects code path 350where we do not cache object size in memory and read it from existing 351packs on demand. This normally only happens when the object size is 352over 2GB. This variable forces the code path on any object larger than 353<n> bytes. 354 355GIT_TEST_OE_DELTA_SIZE=<n> exercises the uncommon pack-objects code 356path where deltas larger than this limit require extra memory 357allocation for bookkeeping. 358 359GIT_TEST_VALIDATE_INDEX_CACHE_ENTRIES=<boolean> checks that cache-tree 360records are valid when the index is written out or after a merge. This 361is mostly to catch missing invalidation. Default is true. 362 363GIT_TEST_COMMIT_GRAPH=<boolean>, when true, forces the commit-graph to 364be written after every 'git commit' command, and overrides the 365'core.commitGraph' setting to true. 366 367GIT_TEST_FSMONITOR=$PWD/t7519/fsmonitor-all exercises the fsmonitor 368code path for utilizing a file system monitor to speed up detecting 369new or changed files. 370 371GIT_TEST_INDEX_VERSION=<n> exercises the index read/write code path 372for the index version specified. Can be set to any valid version 373(currently 2, 3, or 4). 374 375GIT_TEST_PACK_SPARSE=<boolean> if enabled will default the pack-objects 376builtin to use the sparse object walk. This can still be overridden by 377the --no-sparse command-line argument. 378 379GIT_TEST_PRELOAD_INDEX=<boolean> exercises the preload-index code path 380by overriding the minimum number of cache entries required per thread. 381 382GIT_TEST_INDEX_THREADS=<n> enables exercising the multi-threaded loading 383of the index for the whole test suite by bypassing the default number of 384cache entries and thread minimums. Setting this to 1 will make the 385index loading single threaded. 386 387GIT_TEST_MULTI_PACK_INDEX=<boolean>, when true, forces the multi-pack- 388index to be written after every 'git repack' command, and overrides the 389'core.multiPackIndex' setting to true. 390 391GIT_TEST_SIDEBAND_ALL=<boolean>, when true, overrides the 392'uploadpack.allowSidebandAll' setting to true, and when false, forces 393fetch-pack to not request sideband-all (even if the server advertises 394sideband-all). 395 396Naming Tests 397------------ 398 399The test files are named as: 400 401 tNNNN-commandname-details.sh 402 403where N is a decimal digit. 404 405First digit tells the family: 406 407 0 - the absolute basics and global stuff 408 1 - the basic commands concerning database 409 2 - the basic commands concerning the working tree 410 3 - the other basic commands (e.g. ls-files) 411 4 - the diff commands 412 5 - the pull and exporting commands 413 6 - the revision tree commands (even e.g. merge-base) 414 7 - the porcelainish commands concerning the working tree 415 8 - the porcelainish commands concerning forensics 416 9 - the git tools 417 418Second digit tells the particular command we are testing. 419 420Third digit (optionally) tells the particular switch or group of switches 421we are testing. 422 423If you create files under t/ directory (i.e. here) that is not 424the top-level test script, never name the file to match the above 425pattern. The Makefile here considers all such files as the 426top-level test script and tries to run all of them. Care is 427especially needed if you are creating a common test library 428file, similar to test-lib.sh, because such a library file may 429not be suitable for standalone execution. 430 431 432Writing Tests 433------------- 434 435The test script is written as a shell script. It should start 436with the standard "#!/bin/sh", and an 437assignment to variable 'test_description', like this: 438 439 #!/bin/sh 440 441 test_description='xxx test (option --frotz) 442 443 This test registers the following structure in the cache 444 and tries to run git-ls-files with option --frotz.' 445 446 447Source 'test-lib.sh' 448-------------------- 449 450After assigning test_description, the test script should source 451test-lib.sh like this: 452 453 . ./test-lib.sh 454 455This test harness library does the following things: 456 457 - If the script is invoked with command line argument --help 458 (or -h), it shows the test_description and exits. 459 460 - Creates an empty test directory with an empty .git/objects database 461 and chdir(2) into it. This directory is 't/trash 462 directory.$test_name_without_dotsh', with t/ subject to change by 463 the --root option documented above, and a '.stress-<N>' suffix 464 appended by the --stress option. 465 466 - Defines standard test helper functions for your scripts to 467 use. These functions are designed to make all scripts behave 468 consistently when command line arguments --verbose (or -v), 469 --debug (or -d), and --immediate (or -i) is given. 470 471Do's & don'ts 472------------- 473 474Here are a few examples of things you probably should and shouldn't do 475when writing tests. 476 477Here are the "do's:" 478 479 - Put all code inside test_expect_success and other assertions. 480 481 Even code that isn't a test per se, but merely some setup code 482 should be inside a test assertion. 483 484 - Chain your test assertions 485 486 Write test code like this: 487 488 git merge foo && 489 git push bar && 490 test ... 491 492 Instead of: 493 494 git merge hla 495 git push gh 496 test ... 497 498 That way all of the commands in your tests will succeed or fail. If 499 you must ignore the return value of something, consider using a 500 helper function (e.g. use sane_unset instead of unset, in order 501 to avoid unportable return value for unsetting a variable that was 502 already unset), or prepending the command with test_might_fail or 503 test_must_fail. 504 505 - Check the test coverage for your tests. See the "Test coverage" 506 below. 507 508 Don't blindly follow test coverage metrics; if a new function you added 509 doesn't have any coverage, then you're probably doing something wrong, 510 but having 100% coverage doesn't necessarily mean that you tested 511 everything. 512 513 Tests that are likely to smoke out future regressions are better 514 than tests that just inflate the coverage metrics. 515 516 - When a test checks for an absolute path that a git command generated, 517 construct the expected value using $(pwd) rather than $PWD, 518 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on 519 Windows, where the shell (MSYS bash) mangles absolute path names. 520 For details, see the commit message of 4114156ae9. 521 522 - Remember that inside the <script> part, the standard output and 523 standard error streams are discarded, and the test harness only 524 reports "ok" or "not ok" to the end user running the tests. Under 525 --verbose, they are shown to help debug the tests. 526 527And here are the "don'ts:" 528 529 - Don't exit() within a <script> part. 530 531 The harness will catch this as a programming error of the test. 532 Use test_done instead if you need to stop the tests early (see 533 "Skipping tests" below). 534 535 - Don't use '! git cmd' when you want to make sure the git command 536 exits with failure in a controlled way by calling "die()". Instead, 537 use 'test_must_fail git cmd'. This will signal a failure if git 538 dies in an unexpected way (e.g. segfault). 539 540 On the other hand, don't use test_must_fail for running regular 541 platform commands; just use '! cmd'. We are not in the business 542 of verifying that the world given to us sanely works. 543 544 - Don't feed the output of a git command to a pipe, as in: 545 546 git -C repo ls-files | 547 xargs -n 1 basename | 548 grep foo 549 550 which will discard git's exit code and may mask a crash. In the 551 above example, all exit codes are ignored except grep's. 552 553 Instead, write the output of that command to a temporary 554 file with ">" or assign it to a variable with "x=$(git ...)" rather 555 than pipe it. 556 557 - Don't use command substitution in a way that discards git's exit 558 code. When assigning to a variable, the exit code is not discarded, 559 e.g.: 560 561 x=$(git cat-file -p $sha) && 562 ... 563 564 is OK because a crash in "git cat-file" will cause the "&&" chain 565 to fail, but: 566 567 test "refs/heads/foo" = "$(git symbolic-ref HEAD)" 568 569 is not OK and a crash in git could go undetected. 570 571 - Don't use perl without spelling it as "$PERL_PATH". This is to help 572 our friends on Windows where the platform Perl often adds CR before 573 the end of line, and they bundle Git with a version of Perl that 574 does not do so, whose path is specified with $PERL_PATH. Note that we 575 provide a "perl" function which uses $PERL_PATH under the hood, so 576 you do not need to worry when simply running perl in the test scripts 577 (but you do, for example, on a shebang line or in a sub script 578 created via "write_script"). 579 580 - Don't use sh without spelling it as "$SHELL_PATH", when the script 581 can be misinterpreted by broken platform shell (e.g. Solaris). 582 583 - Don't chdir around in tests. It is not sufficient to chdir to 584 somewhere and then chdir back to the original location later in 585 the test, as any intermediate step can fail and abort the test, 586 causing the next test to start in an unexpected directory. Do so 587 inside a subshell if necessary. 588 589 - Don't save and verify the standard error of compound commands, i.e. 590 group commands, subshells, and shell functions (except test helper 591 functions like 'test_must_fail') like this: 592 593 ( cd dir && git cmd ) 2>error && 594 test_cmp expect error 595 596 When running the test with '-x' tracing, then the trace of commands 597 executed in the compound command will be included in standard error 598 as well, quite possibly throwing off the subsequent checks examining 599 the output. Instead, save only the relevant git command's standard 600 error: 601 602 ( cd dir && git cmd 2>../error ) && 603 test_cmp expect error 604 605 - Don't break the TAP output 606 607 The raw output from your test may be interpreted by a TAP harness. TAP 608 harnesses will ignore everything they don't know about, but don't step 609 on their toes in these areas: 610 611 - Don't print lines like "$x..$y" where $x and $y are integers. 612 613 - Don't print lines that begin with "ok" or "not ok". 614 615 TAP harnesses expect a line that begins with either "ok" and "not 616 ok" to signal a test passed or failed (and our harness already 617 produces such lines), so your script shouldn't emit such lines to 618 their output. 619 620 You can glean some further possible issues from the TAP grammar 621 (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR) 622 but the best indication is to just run the tests with prove(1), 623 it'll complain if anything is amiss. 624 625 626Skipping tests 627-------------- 628 629If you need to skip tests you should do so by using the three-arg form 630of the test_* functions (see the "Test harness library" section 631below), e.g.: 632 633 test_expect_success PERL 'I need Perl' ' 634 perl -e "hlagh() if unf_unf()" 635 ' 636 637The advantage of skipping tests like this is that platforms that don't 638have the PERL and other optional dependencies get an indication of how 639many tests they're missing. 640 641If the test code is too hairy for that (i.e. does a lot of setup work 642outside test assertions) you can also skip all remaining tests by 643setting skip_all and immediately call test_done: 644 645 if ! test_have_prereq PERL 646 then 647 skip_all='skipping perl interface tests, perl not available' 648 test_done 649 fi 650 651The string you give to skip_all will be used as an explanation for why 652the test was skipped. 653 654End with test_done 655------------------ 656 657Your script will be a sequence of tests, using helper functions 658from the test harness library. At the end of the script, call 659'test_done'. 660 661 662Test harness library 663-------------------- 664 665There are a handful helper functions defined in the test harness 666library for your script to use. 667 668 - test_expect_success [<prereq>] <message> <script> 669 670 Usually takes two strings as parameters, and evaluates the 671 <script>. If it yields success, test is considered 672 successful. <message> should state what it is testing. 673 674 Example: 675 676 test_expect_success \ 677 'git-write-tree should be able to write an empty tree.' \ 678 'tree=$(git-write-tree)' 679 680 If you supply three parameters the first will be taken to be a 681 prerequisite; see the test_set_prereq and test_have_prereq 682 documentation below: 683 684 test_expect_success TTY 'git --paginate rev-list uses a pager' \ 685 ' ... ' 686 687 You can also supply a comma-separated list of prerequisites, in the 688 rare case where your test depends on more than one: 689 690 test_expect_success PERL,PYTHON 'yo dawg' \ 691 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" ' 692 693 - test_expect_failure [<prereq>] <message> <script> 694 695 This is NOT the opposite of test_expect_success, but is used 696 to mark a test that demonstrates a known breakage. Unlike 697 the usual test_expect_success tests, which say "ok" on 698 success and "FAIL" on failure, this will say "FIXED" on 699 success and "still broken" on failure. Failures from these 700 tests won't cause -i (immediate) to stop. 701 702 Like test_expect_success this function can optionally use a three 703 argument invocation with a prerequisite as the first argument. 704 705 - test_debug <script> 706 707 This takes a single argument, <script>, and evaluates it only 708 when the test script is started with --debug command line 709 argument. This is primarily meant for use during the 710 development of a new test script. 711 712 - debug <git-command> 713 714 Run a git command inside a debugger. This is primarily meant for 715 use when debugging a failing test script. 716 717 - test_done 718 719 Your test script must have test_done at the end. Its purpose 720 is to summarize successes and failures in the test script and 721 exit with an appropriate error code. 722 723 - test_tick 724 725 Make commit and tag names consistent by setting the author and 726 committer times to defined state. Subsequent calls will 727 advance the times by a fixed amount. 728 729 - test_commit <message> [<filename> [<contents>]] 730 731 Creates a commit with the given message, committing the given 732 file with the given contents (default for both is to reuse the 733 message string), and adds a tag (again reusing the message 734 string as name). Calls test_tick to make the SHA-1s 735 reproducible. 736 737 - test_merge <message> <commit-or-tag> 738 739 Merges the given rev using the given message. Like test_commit, 740 creates a tag and calls test_tick before committing. 741 742 - test_set_prereq <prereq> 743 744 Set a test prerequisite to be used later with test_have_prereq. The 745 test-lib will set some prerequisites for you, see the 746 "Prerequisites" section below for a full list of these. 747 748 Others you can set yourself and use later with either 749 test_have_prereq directly, or the three argument invocation of 750 test_expect_success and test_expect_failure. 751 752 - test_have_prereq <prereq> 753 754 Check if we have a prerequisite previously set with test_set_prereq. 755 The most common way to use this explicitly (as opposed to the 756 implicit use when an argument is passed to test_expect_*) is to skip 757 all the tests at the start of the test script if we don't have some 758 essential prerequisite: 759 760 if ! test_have_prereq PERL 761 then 762 skip_all='skipping perl interface tests, perl not available' 763 test_done 764 fi 765 766 - test_external [<prereq>] <message> <external> <script> 767 768 Execute a <script> with an <external> interpreter (like perl). This 769 was added for tests like t9700-perl-git.sh which do most of their 770 work in an external test script. 771 772 test_external \ 773 'GitwebCache::*FileCache*' \ 774 perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl 775 776 If the test is outputting its own TAP you should set the 777 test_external_has_tap variable somewhere before calling the first 778 test_external* function. See t9700-perl-git.sh for an example. 779 780 # The external test will outputs its own plan 781 test_external_has_tap=1 782 783 - test_external_without_stderr [<prereq>] <message> <external> <script> 784 785 Like test_external but fail if there's any output on stderr, 786 instead of checking the exit code. 787 788 test_external_without_stderr \ 789 'Perl API' \ 790 perl "$TEST_DIRECTORY"/t9700/test.pl 791 792 - test_expect_code <exit-code> <command> 793 794 Run a command and ensure that it exits with the given exit code. 795 For example: 796 797 test_expect_success 'Merge with d/f conflicts' ' 798 test_expect_code 1 git merge "merge msg" B master 799 ' 800 801 - test_must_fail [<options>] <git-command> 802 803 Run a git command and ensure it fails in a controlled way. Use 804 this instead of "! <git-command>". When git-command dies due to a 805 segfault, test_must_fail diagnoses it as an error; "! <git-command>" 806 treats it as just another expected failure, which would let such a 807 bug go unnoticed. 808 809 Accepts the following options: 810 811 ok=<signal-name>[,<...>]: 812 Don't treat an exit caused by the given signal as error. 813 Multiple signals can be specified as a comma separated list. 814 Currently recognized signal names are: sigpipe, success. 815 (Don't use 'success', use 'test_might_fail' instead.) 816 817 - test_might_fail [<options>] <git-command> 818 819 Similar to test_must_fail, but tolerate success, too. Use this 820 instead of "<git-command> || :" to catch failures due to segv. 821 822 Accepts the same options as test_must_fail. 823 824 - test_cmp <expected> <actual> 825 826 Check whether the content of the <actual> file matches the 827 <expected> file. This behaves like "cmp" but produces more 828 helpful output when the test is run with "-v" option. 829 830 - test_cmp_rev <expected> <actual> 831 832 Check whether the <expected> rev points to the same commit as the 833 <actual> rev. 834 835 - test_line_count (= | -lt | -ge | ...) <length> <file> 836 837 Check whether a file has the length it is expected to. 838 839 - test_path_is_file <path> [<diagnosis>] 840 test_path_is_dir <path> [<diagnosis>] 841 test_path_is_missing <path> [<diagnosis>] 842 843 Check if the named path is a file, if the named path is a 844 directory, or if the named path does not exist, respectively, 845 and fail otherwise, showing the <diagnosis> text. 846 847 - test_when_finished <script> 848 849 Prepend <script> to a list of commands to run to clean up 850 at the end of the current test. If some clean-up command 851 fails, the test will not pass. 852 853 Example: 854 855 test_expect_success 'branch pointing to non-commit' ' 856 git rev-parse HEAD^{tree} >.git/refs/heads/invalid && 857 test_when_finished "git update-ref -d refs/heads/invalid" && 858 ... 859 ' 860 861 - test_write_lines <lines> 862 863 Write <lines> on standard output, one line per argument. 864 Useful to prepare multi-line files in a compact form. 865 866 Example: 867 868 test_write_lines a b c d e f g >foo 869 870 Is a more compact equivalent of: 871 cat >foo <<-EOF 872 a 873 b 874 c 875 d 876 e 877 f 878 g 879 EOF 880 881 882 - test_pause 883 884 This command is useful for writing and debugging tests and must be 885 removed before submitting. It halts the execution of the test and 886 spawns a shell in the trash directory. Exit the shell to continue 887 the test. Example: 888 889 test_expect_success 'test' ' 890 git do-something >actual && 891 test_pause && 892 test_cmp expected actual 893 ' 894 895 - test_ln_s_add <path1> <path2> 896 897 This function helps systems whose filesystem does not support symbolic 898 links. Use it to add a symbolic link entry to the index when it is not 899 important that the file system entry is a symbolic link, i.e., instead 900 of the sequence 901 902 ln -s foo bar && 903 git add bar 904 905 Sometimes it is possible to split a test in a part that does not need 906 the symbolic link in the file system and a part that does; then only 907 the latter part need be protected by a SYMLINKS prerequisite (see below). 908 909 - test_oid_init 910 911 This function loads facts and useful object IDs related to the hash 912 algorithm(s) in use from the files in t/oid-info. 913 914 - test_oid_cache 915 916 This function reads per-hash algorithm information from standard 917 input (usually a heredoc) in the format described in 918 t/oid-info/README. This is useful for test-specific values, such as 919 object IDs, which must vary based on the hash algorithm. 920 921 Certain fixed values, such as hash sizes and common placeholder 922 object IDs, can be loaded with test_oid_init (described above). 923 924 - test_oid <key> 925 926 This function looks up a value for the hash algorithm in use, based 927 on the key given. The value must have been loaded using 928 test_oid_init or test_oid_cache. Providing an unknown key is an 929 error. 930 931 - yes [<string>] 932 933 This is often seen in modern UNIX but some platforms lack it, so 934 the test harness overrides the platform implementation with a 935 more limited one. Use this only when feeding a handful lines of 936 output to the downstream---unlike the real version, it generates 937 only up to 99 lines. 938 939 940Prerequisites 941------------- 942 943These are the prerequisites that the test library predefines with 944test_have_prereq. 945 946See the prereq argument to the test_* functions in the "Test harness 947library" section above and the "test_have_prereq" function for how to 948use these, and "test_set_prereq" for how to define your own. 949 950 - PYTHON 951 952 Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that 953 need Python with this. 954 955 - PERL 956 957 Git wasn't compiled with NO_PERL=YesPlease. 958 959 Even without the PERL prerequisite, tests can assume there is a 960 usable perl interpreter at $PERL_PATH, though it need not be 961 particularly modern. 962 963 - POSIXPERM 964 965 The filesystem supports POSIX style permission bits. 966 967 - BSLASHPSPEC 968 969 Backslashes in pathspec are not directory separators. This is not 970 set on Windows. See 6fd1106a for details. 971 972 - EXECKEEPSPID 973 974 The process retains the same pid across exec(2). See fb9a2bea for 975 details. 976 977 - PIPE 978 979 The filesystem we're on supports creation of FIFOs (named pipes) 980 via mkfifo(1). 981 982 - SYMLINKS 983 984 The filesystem we're on supports symbolic links. E.g. a FAT 985 filesystem doesn't support these. See 704a3143 for details. 986 987 - SANITY 988 989 Test is not run by root user, and an attempt to write to an 990 unwritable file is expected to fail correctly. 991 992 - PCRE 993 994 Git was compiled with support for PCRE. Wrap any tests 995 that use git-grep --perl-regexp or git-grep -P in these. 996 997 - LIBPCRE1 998 999 Git was compiled with PCRE v1 support via1000 USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some1001 reason need v1 of the PCRE library instead of v2 in these.10021003 - LIBPCRE210041005 Git was compiled with PCRE v2 support via1006 USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some1007 reason need v2 of the PCRE library instead of v1 in these.10081009 - CASE_INSENSITIVE_FS10101011 Test is run on a case insensitive file system.10121013 - UTF8_NFD_TO_NFC10141015 Test is run on a filesystem which converts decomposed utf-8 (nfd)1016 to precomposed utf-8 (nfc).10171018 - PTHREADS10191020 Git wasn't compiled with NO_PTHREADS=YesPlease.10211022Tips for Writing Tests1023----------------------10241025As with any programming projects, existing programs are the best1026source of the information. However, do _not_ emulate1027t0000-basic.sh when writing your tests. The test is special in1028that it tries to validate the very core of GIT. For example, it1029knows that there will be 256 subdirectories under .git/objects/,1030and it knows that the object ID of an empty tree is a certain103140-byte string. This is deliberately done so in t0000-basic.sh1032because the things the very basic core test tries to achieve is1033to serve as a basis for people who are changing the GIT internal1034drastically. For these people, after making certain changes,1035not seeing failures from the basic test _is_ a failure. And1036such drastic changes to the core GIT that even changes these1037otherwise supposedly stable object IDs should be accompanied by1038an update to t0000-basic.sh.10391040However, other tests that simply rely on basic parts of the core1041GIT working properly should not have that level of intimate1042knowledge of the core GIT internals. If all the test scripts1043hardcoded the object IDs like t0000-basic.sh does, that defeats1044the purpose of t0000-basic.sh, which is to isolate that level of1045validation in one place. Your test also ends up needing1046updating when such a change to the internal happens, so do _not_1047do it and leave the low level of validation to t0000-basic.sh.10481049Test coverage1050-------------10511052You can use the coverage tests to find code paths that are not being1053used or properly exercised yet.10541055To do that, run the coverage target at the top-level (not in the t/1056directory):10571058 make coverage10591060That'll compile Git with GCC's coverage arguments, and generate a test1061report with gcov after the tests finish. Running the coverage tests1062can take a while, since running the tests in parallel is incompatible1063with GCC's coverage mode.10641065After the tests have run you can generate a list of untested1066functions:10671068 make coverage-untested-functions10691070You can also generate a detailed per-file HTML report using the1071Devel::Cover module. To install it do:10721073 # On Debian or Ubuntu:1074 sudo aptitude install libdevel-cover-perl10751076 # From the CPAN with cpanminus1077 curl -L http://cpanmin.us | perl - --sudo --self-upgrade1078 cpanm --sudo Devel::Cover10791080Then, at the top-level:10811082 make cover_db_html10831084That'll generate a detailed cover report in the "cover_db_html"1085directory, which you can then copy to a webserver, or inspect locally1086in a browser.