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 Run the test script repeatedly in multiple parallel jobs until 200 one of them fails. Useful for reproducing rare failures in 201 flaky tests. The number of parallel jobs is, in order of 202 precedence: the value of the GIT_TEST_STRESS_LOAD 203 environment variable, or twice the number of available 204 processors (as shown by the 'getconf' utility), or 8. 205 Implies `--verbose -x --immediate` to get the most information 206 about the failure. Note that the verbose output of each test 207 job is saved to 't/test-results/$TEST_NAME.stress-<nr>.out', 208 and only the output of the failed test job is shown on the 209 terminal. The names of the trash directories get a 210 '.stress-<nr>' suffix, and the trash directory of the failed 211 test job is renamed to end with a '.stress-failed' suffix. 212 213--stress-jobs=<N>:: 214 Override the number of parallel jobs. Implies `--stress`. 215 216--stress-limit=<N>:: 217 When combined with --stress run the test script repeatedly 218 this many times in each of the parallel jobs or until one of 219 them fails, whichever comes first. Implies `--stress`. 220 221You can also set the GIT_TEST_INSTALLED environment variable to 222the bindir of an existing git installation to test that installation. 223You still need to have built this git sandbox, from which various 224test-* support programs, templates, and perl libraries are used. 225If your installed git is incomplete, it will silently test parts of 226your built version instead. 227 228When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to 229override the location of the dashed-form subcommands (what 230GIT_EXEC_PATH would be used for during normal operation). 231GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`. 232 233 234Skipping Tests 235-------------- 236 237In some environments, certain tests have no way of succeeding 238due to platform limitation, such as lack of 'unzip' program, or 239filesystem that do not allow arbitrary sequence of non-NUL bytes 240as pathnames. 241 242You should be able to say something like 243 244 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh 245 246and even: 247 248 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make 249 250to omit such tests. The value of the environment variable is a 251SP separated list of patterns that tells which tests to skip, 252and either can match the "t[0-9]{4}" part to skip the whole 253test, or t[0-9]{4} followed by ".$number" to say which 254particular test to skip. 255 256For an individual test suite --run could be used to specify that 257only some tests should be run or that some tests should be 258excluded from a run. 259 260The argument for --run is a list of individual test numbers or 261ranges with an optional negation prefix that define what tests in 262a test suite to include in the run. A range is two numbers 263separated with a dash and matches a range of tests with both ends 264been included. You may omit the first or the second number to 265mean "from the first test" or "up to the very last test" 266respectively. 267 268Optional prefix of '!' means that the test or a range of tests 269should be excluded from the run. 270 271If --run starts with an unprefixed number or range the initial 272set of tests to run is empty. If the first item starts with '!' 273all the tests are added to the initial set. After initial set is 274determined every test number or range is added or excluded from 275the set one by one, from left to right. 276 277Individual numbers or ranges could be separated either by a space 278or a comma. 279 280For example, to run only tests up to a specific test (21), one 281could do this: 282 283 $ sh ./t9200-git-cvsexport-commit.sh --run='1-21' 284 285or this: 286 287 $ sh ./t9200-git-cvsexport-commit.sh --run='-21' 288 289Common case is to run several setup tests (1, 2, 3) and then a 290specific test (21) that relies on that setup: 291 292 $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21' 293 294or: 295 296 $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21 297 298or: 299 300 $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21' 301 302As noted above, the test set is built by going through the items 303from left to right, so this: 304 305 $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3' 306 307will run tests 1, 2, and 4. Items that come later have higher 308precedence. It means that this: 309 310 $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4' 311 312would just run tests from 1 to 4, including 3. 313 314You may use negation with ranges. The following will run all 315test in the test suite except from 7 up to 11: 316 317 $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11' 318 319Some tests in a test suite rely on the previous tests performing 320certain actions, specifically some tests are designated as 321"setup" test, so you cannot _arbitrarily_ disable one test and 322expect the rest to function correctly. 323 324--run is mostly useful when you want to focus on a specific test 325and know what setup is needed for it. Or when you want to run 326everything up to a certain test. 327 328 329Running tests with special setups 330--------------------------------- 331 332The whole test suite could be run to test some special features 333that cannot be easily covered by a few specific test cases. These 334could be enabled by running the test suite with correct GIT_TEST_ 335environment set. 336 337GIT_TEST_FAIL_PREREQS<non-empty?> fails all prerequisites. This is 338useful for discovering issues with the tests where say a later test 339implicitly depends on an optional earlier test. 340 341There's a "FAIL_PREREQS" prerequisite that can be used to test for 342whether this mode is active, and e.g. skip some tests that are hard to 343refactor to deal with it. The "SYMLINKS" prerequisite is currently 344excluded as so much relies on it, but this might change in the future. 345 346GIT_TEST_GETTEXT_POISON=<boolean> turns all strings marked for 347translation into gibberish if true. Used for 348spotting those tests that need to be marked with a C_LOCALE_OUTPUT 349prerequisite when adding more strings for translation. See "Testing 350marked strings" in po/README for details. 351 352GIT_TEST_SPLIT_INDEX=<boolean> forces split-index mode on the whole 353test suite. Accept any boolean values that are accepted by git-config. 354 355GIT_TEST_PROTOCOL_VERSION=<n>, when set, overrides the 356'protocol.version' setting to n if it is less than n. 357 358GIT_TEST_FULL_IN_PACK_ARRAY=<boolean> exercises the uncommon 359pack-objects code path where there are more than 1024 packs even if 360the actual number of packs in repository is below this limit. Accept 361any boolean values that are accepted by git-config. 362 363GIT_TEST_OE_SIZE=<n> exercises the uncommon pack-objects code path 364where we do not cache object size in memory and read it from existing 365packs on demand. This normally only happens when the object size is 366over 2GB. This variable forces the code path on any object larger than 367<n> bytes. 368 369GIT_TEST_OE_DELTA_SIZE=<n> exercises the uncommon pack-objects code 370path where deltas larger than this limit require extra memory 371allocation for bookkeeping. 372 373GIT_TEST_VALIDATE_INDEX_CACHE_ENTRIES=<boolean> checks that cache-tree 374records are valid when the index is written out or after a merge. This 375is mostly to catch missing invalidation. Default is true. 376 377GIT_TEST_COMMIT_GRAPH=<boolean>, when true, forces the commit-graph to 378be written after every 'git commit' command, and overrides the 379'core.commitGraph' setting to true. 380 381GIT_TEST_FSMONITOR=$PWD/t7519/fsmonitor-all exercises the fsmonitor 382code path for utilizing a file system monitor to speed up detecting 383new or changed files. 384 385GIT_TEST_INDEX_VERSION=<n> exercises the index read/write code path 386for the index version specified. Can be set to any valid version 387(currently 2, 3, or 4). 388 389GIT_TEST_PACK_SPARSE=<boolean> if enabled will default the pack-objects 390builtin to use the sparse object walk. This can still be overridden by 391the --no-sparse command-line argument. 392 393GIT_TEST_PRELOAD_INDEX=<boolean> exercises the preload-index code path 394by overriding the minimum number of cache entries required per thread. 395 396GIT_TEST_STASH_USE_BUILTIN=<boolean>, when false, disables the 397built-in version of git-stash. See 'stash.useBuiltin' in 398git-config(1). 399 400GIT_TEST_INDEX_THREADS=<n> enables exercising the multi-threaded loading 401of the index for the whole test suite by bypassing the default number of 402cache entries and thread minimums. Setting this to 1 will make the 403index loading single threaded. 404 405GIT_TEST_MULTI_PACK_INDEX=<boolean>, when true, forces the multi-pack- 406index to be written after every 'git repack' command, and overrides the 407'core.multiPackIndex' setting to true. 408 409GIT_TEST_SIDEBAND_ALL=<boolean>, when true, overrides the 410'uploadpack.allowSidebandAll' setting to true, and when false, forces 411fetch-pack to not request sideband-all (even if the server advertises 412sideband-all). 413 414GIT_TEST_DISALLOW_ABBREVIATED_OPTIONS=<boolean>, when true (which is 415the default when running tests), errors out when an abbreviated option 416is used. 417 418Naming Tests 419------------ 420 421The test files are named as: 422 423 tNNNN-commandname-details.sh 424 425where N is a decimal digit. 426 427First digit tells the family: 428 429 0 - the absolute basics and global stuff 430 1 - the basic commands concerning database 431 2 - the basic commands concerning the working tree 432 3 - the other basic commands (e.g. ls-files) 433 4 - the diff commands 434 5 - the pull and exporting commands 435 6 - the revision tree commands (even e.g. merge-base) 436 7 - the porcelainish commands concerning the working tree 437 8 - the porcelainish commands concerning forensics 438 9 - the git tools 439 440Second digit tells the particular command we are testing. 441 442Third digit (optionally) tells the particular switch or group of switches 443we are testing. 444 445If you create files under t/ directory (i.e. here) that is not 446the top-level test script, never name the file to match the above 447pattern. The Makefile here considers all such files as the 448top-level test script and tries to run all of them. Care is 449especially needed if you are creating a common test library 450file, similar to test-lib.sh, because such a library file may 451not be suitable for standalone execution. 452 453 454Writing Tests 455------------- 456 457The test script is written as a shell script. It should start 458with the standard "#!/bin/sh", and an 459assignment to variable 'test_description', like this: 460 461 #!/bin/sh 462 463 test_description='xxx test (option --frotz) 464 465 This test registers the following structure in the cache 466 and tries to run git-ls-files with option --frotz.' 467 468 469Source 'test-lib.sh' 470-------------------- 471 472After assigning test_description, the test script should source 473test-lib.sh like this: 474 475 . ./test-lib.sh 476 477This test harness library does the following things: 478 479 - If the script is invoked with command line argument --help 480 (or -h), it shows the test_description and exits. 481 482 - Creates an empty test directory with an empty .git/objects database 483 and chdir(2) into it. This directory is 't/trash 484 directory.$test_name_without_dotsh', with t/ subject to change by 485 the --root option documented above, and a '.stress-<N>' suffix 486 appended by the --stress option. 487 488 - Defines standard test helper functions for your scripts to 489 use. These functions are designed to make all scripts behave 490 consistently when command line arguments --verbose (or -v), 491 --debug (or -d), and --immediate (or -i) is given. 492 493Do's & don'ts 494------------- 495 496Here are a few examples of things you probably should and shouldn't do 497when writing tests. 498 499Here are the "do's:" 500 501 - Put all code inside test_expect_success and other assertions. 502 503 Even code that isn't a test per se, but merely some setup code 504 should be inside a test assertion. 505 506 - Chain your test assertions 507 508 Write test code like this: 509 510 git merge foo && 511 git push bar && 512 test ... 513 514 Instead of: 515 516 git merge hla 517 git push gh 518 test ... 519 520 That way all of the commands in your tests will succeed or fail. If 521 you must ignore the return value of something, consider using a 522 helper function (e.g. use sane_unset instead of unset, in order 523 to avoid unportable return value for unsetting a variable that was 524 already unset), or prepending the command with test_might_fail or 525 test_must_fail. 526 527 - Check the test coverage for your tests. See the "Test coverage" 528 below. 529 530 Don't blindly follow test coverage metrics; if a new function you added 531 doesn't have any coverage, then you're probably doing something wrong, 532 but having 100% coverage doesn't necessarily mean that you tested 533 everything. 534 535 Tests that are likely to smoke out future regressions are better 536 than tests that just inflate the coverage metrics. 537 538 - When a test checks for an absolute path that a git command generated, 539 construct the expected value using $(pwd) rather than $PWD, 540 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on 541 Windows, where the shell (MSYS bash) mangles absolute path names. 542 For details, see the commit message of 4114156ae9. 543 544 - Remember that inside the <script> part, the standard output and 545 standard error streams are discarded, and the test harness only 546 reports "ok" or "not ok" to the end user running the tests. Under 547 --verbose, they are shown to help debug the tests. 548 549And here are the "don'ts:" 550 551 - Don't exit() within a <script> part. 552 553 The harness will catch this as a programming error of the test. 554 Use test_done instead if you need to stop the tests early (see 555 "Skipping tests" below). 556 557 - Don't use '! git cmd' when you want to make sure the git command 558 exits with failure in a controlled way by calling "die()". Instead, 559 use 'test_must_fail git cmd'. This will signal a failure if git 560 dies in an unexpected way (e.g. segfault). 561 562 On the other hand, don't use test_must_fail for running regular 563 platform commands; just use '! cmd'. We are not in the business 564 of verifying that the world given to us sanely works. 565 566 - Don't feed the output of a git command to a pipe, as in: 567 568 git -C repo ls-files | 569 xargs -n 1 basename | 570 grep foo 571 572 which will discard git's exit code and may mask a crash. In the 573 above example, all exit codes are ignored except grep's. 574 575 Instead, write the output of that command to a temporary 576 file with ">" or assign it to a variable with "x=$(git ...)" rather 577 than pipe it. 578 579 - Don't use command substitution in a way that discards git's exit 580 code. When assigning to a variable, the exit code is not discarded, 581 e.g.: 582 583 x=$(git cat-file -p $sha) && 584 ... 585 586 is OK because a crash in "git cat-file" will cause the "&&" chain 587 to fail, but: 588 589 test "refs/heads/foo" = "$(git symbolic-ref HEAD)" 590 591 is not OK and a crash in git could go undetected. 592 593 - Don't use perl without spelling it as "$PERL_PATH". This is to help 594 our friends on Windows where the platform Perl often adds CR before 595 the end of line, and they bundle Git with a version of Perl that 596 does not do so, whose path is specified with $PERL_PATH. Note that we 597 provide a "perl" function which uses $PERL_PATH under the hood, so 598 you do not need to worry when simply running perl in the test scripts 599 (but you do, for example, on a shebang line or in a sub script 600 created via "write_script"). 601 602 - Don't use sh without spelling it as "$SHELL_PATH", when the script 603 can be misinterpreted by broken platform shell (e.g. Solaris). 604 605 - Don't chdir around in tests. It is not sufficient to chdir to 606 somewhere and then chdir back to the original location later in 607 the test, as any intermediate step can fail and abort the test, 608 causing the next test to start in an unexpected directory. Do so 609 inside a subshell if necessary. 610 611 - Don't save and verify the standard error of compound commands, i.e. 612 group commands, subshells, and shell functions (except test helper 613 functions like 'test_must_fail') like this: 614 615 ( cd dir && git cmd ) 2>error && 616 test_cmp expect error 617 618 When running the test with '-x' tracing, then the trace of commands 619 executed in the compound command will be included in standard error 620 as well, quite possibly throwing off the subsequent checks examining 621 the output. Instead, save only the relevant git command's standard 622 error: 623 624 ( cd dir && git cmd 2>../error ) && 625 test_cmp expect error 626 627 - Don't break the TAP output 628 629 The raw output from your test may be interpreted by a TAP harness. TAP 630 harnesses will ignore everything they don't know about, but don't step 631 on their toes in these areas: 632 633 - Don't print lines like "$x..$y" where $x and $y are integers. 634 635 - Don't print lines that begin with "ok" or "not ok". 636 637 TAP harnesses expect a line that begins with either "ok" and "not 638 ok" to signal a test passed or failed (and our harness already 639 produces such lines), so your script shouldn't emit such lines to 640 their output. 641 642 You can glean some further possible issues from the TAP grammar 643 (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR) 644 but the best indication is to just run the tests with prove(1), 645 it'll complain if anything is amiss. 646 647 648Skipping tests 649-------------- 650 651If you need to skip tests you should do so by using the three-arg form 652of the test_* functions (see the "Test harness library" section 653below), e.g.: 654 655 test_expect_success PERL 'I need Perl' ' 656 perl -e "hlagh() if unf_unf()" 657 ' 658 659The advantage of skipping tests like this is that platforms that don't 660have the PERL and other optional dependencies get an indication of how 661many tests they're missing. 662 663If the test code is too hairy for that (i.e. does a lot of setup work 664outside test assertions) you can also skip all remaining tests by 665setting skip_all and immediately call test_done: 666 667 if ! test_have_prereq PERL 668 then 669 skip_all='skipping perl interface tests, perl not available' 670 test_done 671 fi 672 673The string you give to skip_all will be used as an explanation for why 674the test was skipped. 675 676End with test_done 677------------------ 678 679Your script will be a sequence of tests, using helper functions 680from the test harness library. At the end of the script, call 681'test_done'. 682 683 684Test harness library 685-------------------- 686 687There are a handful helper functions defined in the test harness 688library for your script to use. 689 690 - test_expect_success [<prereq>] <message> <script> 691 692 Usually takes two strings as parameters, and evaluates the 693 <script>. If it yields success, test is considered 694 successful. <message> should state what it is testing. 695 696 Example: 697 698 test_expect_success \ 699 'git-write-tree should be able to write an empty tree.' \ 700 'tree=$(git-write-tree)' 701 702 If you supply three parameters the first will be taken to be a 703 prerequisite; see the test_set_prereq and test_have_prereq 704 documentation below: 705 706 test_expect_success TTY 'git --paginate rev-list uses a pager' \ 707 ' ... ' 708 709 You can also supply a comma-separated list of prerequisites, in the 710 rare case where your test depends on more than one: 711 712 test_expect_success PERL,PYTHON 'yo dawg' \ 713 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" ' 714 715 - test_expect_failure [<prereq>] <message> <script> 716 717 This is NOT the opposite of test_expect_success, but is used 718 to mark a test that demonstrates a known breakage. Unlike 719 the usual test_expect_success tests, which say "ok" on 720 success and "FAIL" on failure, this will say "FIXED" on 721 success and "still broken" on failure. Failures from these 722 tests won't cause -i (immediate) to stop. 723 724 Like test_expect_success this function can optionally use a three 725 argument invocation with a prerequisite as the first argument. 726 727 - test_debug <script> 728 729 This takes a single argument, <script>, and evaluates it only 730 when the test script is started with --debug command line 731 argument. This is primarily meant for use during the 732 development of a new test script. 733 734 - debug <git-command> 735 736 Run a git command inside a debugger. This is primarily meant for 737 use when debugging a failing test script. 738 739 - test_done 740 741 Your test script must have test_done at the end. Its purpose 742 is to summarize successes and failures in the test script and 743 exit with an appropriate error code. 744 745 - test_tick 746 747 Make commit and tag names consistent by setting the author and 748 committer times to defined state. Subsequent calls will 749 advance the times by a fixed amount. 750 751 - test_commit <message> [<filename> [<contents>]] 752 753 Creates a commit with the given message, committing the given 754 file with the given contents (default for both is to reuse the 755 message string), and adds a tag (again reusing the message 756 string as name). Calls test_tick to make the SHA-1s 757 reproducible. 758 759 - test_merge <message> <commit-or-tag> 760 761 Merges the given rev using the given message. Like test_commit, 762 creates a tag and calls test_tick before committing. 763 764 - test_set_prereq <prereq> 765 766 Set a test prerequisite to be used later with test_have_prereq. The 767 test-lib will set some prerequisites for you, see the 768 "Prerequisites" section below for a full list of these. 769 770 Others you can set yourself and use later with either 771 test_have_prereq directly, or the three argument invocation of 772 test_expect_success and test_expect_failure. 773 774 - test_have_prereq <prereq> 775 776 Check if we have a prerequisite previously set with test_set_prereq. 777 The most common way to use this explicitly (as opposed to the 778 implicit use when an argument is passed to test_expect_*) is to skip 779 all the tests at the start of the test script if we don't have some 780 essential prerequisite: 781 782 if ! test_have_prereq PERL 783 then 784 skip_all='skipping perl interface tests, perl not available' 785 test_done 786 fi 787 788 - test_external [<prereq>] <message> <external> <script> 789 790 Execute a <script> with an <external> interpreter (like perl). This 791 was added for tests like t9700-perl-git.sh which do most of their 792 work in an external test script. 793 794 test_external \ 795 'GitwebCache::*FileCache*' \ 796 perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl 797 798 If the test is outputting its own TAP you should set the 799 test_external_has_tap variable somewhere before calling the first 800 test_external* function. See t9700-perl-git.sh for an example. 801 802 # The external test will outputs its own plan 803 test_external_has_tap=1 804 805 - test_external_without_stderr [<prereq>] <message> <external> <script> 806 807 Like test_external but fail if there's any output on stderr, 808 instead of checking the exit code. 809 810 test_external_without_stderr \ 811 'Perl API' \ 812 perl "$TEST_DIRECTORY"/t9700/test.pl 813 814 - test_expect_code <exit-code> <command> 815 816 Run a command and ensure that it exits with the given exit code. 817 For example: 818 819 test_expect_success 'Merge with d/f conflicts' ' 820 test_expect_code 1 git merge "merge msg" B master 821 ' 822 823 - test_must_fail [<options>] <git-command> 824 825 Run a git command and ensure it fails in a controlled way. Use 826 this instead of "! <git-command>". When git-command dies due to a 827 segfault, test_must_fail diagnoses it as an error; "! <git-command>" 828 treats it as just another expected failure, which would let such a 829 bug go unnoticed. 830 831 Accepts the following options: 832 833 ok=<signal-name>[,<...>]: 834 Don't treat an exit caused by the given signal as error. 835 Multiple signals can be specified as a comma separated list. 836 Currently recognized signal names are: sigpipe, success. 837 (Don't use 'success', use 'test_might_fail' instead.) 838 839 - test_might_fail [<options>] <git-command> 840 841 Similar to test_must_fail, but tolerate success, too. Use this 842 instead of "<git-command> || :" to catch failures due to segv. 843 844 Accepts the same options as test_must_fail. 845 846 - test_cmp <expected> <actual> 847 848 Check whether the content of the <actual> file matches the 849 <expected> file. This behaves like "cmp" but produces more 850 helpful output when the test is run with "-v" option. 851 852 - test_cmp_rev <expected> <actual> 853 854 Check whether the <expected> rev points to the same commit as the 855 <actual> rev. 856 857 - test_line_count (= | -lt | -ge | ...) <length> <file> 858 859 Check whether a file has the length it is expected to. 860 861 - test_path_is_file <path> [<diagnosis>] 862 test_path_is_dir <path> [<diagnosis>] 863 test_path_is_missing <path> [<diagnosis>] 864 865 Check if the named path is a file, if the named path is a 866 directory, or if the named path does not exist, respectively, 867 and fail otherwise, showing the <diagnosis> text. 868 869 - test_when_finished <script> 870 871 Prepend <script> to a list of commands to run to clean up 872 at the end of the current test. If some clean-up command 873 fails, the test will not pass. 874 875 Example: 876 877 test_expect_success 'branch pointing to non-commit' ' 878 git rev-parse HEAD^{tree} >.git/refs/heads/invalid && 879 test_when_finished "git update-ref -d refs/heads/invalid" && 880 ... 881 ' 882 883 - test_atexit <script> 884 885 Prepend <script> to a list of commands to run unconditionally to 886 clean up before the test script exits, e.g. to stop a daemon: 887 888 test_expect_success 'test git daemon' ' 889 git daemon & 890 daemon_pid=$! && 891 test_atexit 'kill $daemon_pid' && 892 hello world 893 ' 894 895 The commands will be executed before the trash directory is removed, 896 i.e. the atexit commands will still be able to access any pidfiles or 897 socket files. 898 899 Note that these commands will be run even when a test script run 900 with '--immediate' fails. Be careful with your atexit commands to 901 minimize any changes to the failed state. 902 903 - test_write_lines <lines> 904 905 Write <lines> on standard output, one line per argument. 906 Useful to prepare multi-line files in a compact form. 907 908 Example: 909 910 test_write_lines a b c d e f g >foo 911 912 Is a more compact equivalent of: 913 cat >foo <<-EOF 914 a 915 b 916 c 917 d 918 e 919 f 920 g 921 EOF 922 923 924 - test_pause 925 926 This command is useful for writing and debugging tests and must be 927 removed before submitting. It halts the execution of the test and 928 spawns a shell in the trash directory. Exit the shell to continue 929 the test. Example: 930 931 test_expect_success 'test' ' 932 git do-something >actual && 933 test_pause && 934 test_cmp expected actual 935 ' 936 937 - test_ln_s_add <path1> <path2> 938 939 This function helps systems whose filesystem does not support symbolic 940 links. Use it to add a symbolic link entry to the index when it is not 941 important that the file system entry is a symbolic link, i.e., instead 942 of the sequence 943 944 ln -s foo bar && 945 git add bar 946 947 Sometimes it is possible to split a test in a part that does not need 948 the symbolic link in the file system and a part that does; then only 949 the latter part need be protected by a SYMLINKS prerequisite (see below). 950 951 - test_oid_init 952 953 This function loads facts and useful object IDs related to the hash 954 algorithm(s) in use from the files in t/oid-info. 955 956 - test_oid_cache 957 958 This function reads per-hash algorithm information from standard 959 input (usually a heredoc) in the format described in 960 t/oid-info/README. This is useful for test-specific values, such as 961 object IDs, which must vary based on the hash algorithm. 962 963 Certain fixed values, such as hash sizes and common placeholder 964 object IDs, can be loaded with test_oid_init (described above). 965 966 - test_oid <key> 967 968 This function looks up a value for the hash algorithm in use, based 969 on the key given. The value must have been loaded using 970 test_oid_init or test_oid_cache. Providing an unknown key is an 971 error. 972 973 - yes [<string>] 974 975 This is often seen in modern UNIX but some platforms lack it, so 976 the test harness overrides the platform implementation with a 977 more limited one. Use this only when feeding a handful lines of 978 output to the downstream---unlike the real version, it generates 979 only up to 99 lines. 980 981 982Prerequisites 983------------- 984 985These are the prerequisites that the test library predefines with 986test_have_prereq. 987 988See the prereq argument to the test_* functions in the "Test harness 989library" section above and the "test_have_prereq" function for how to 990use these, and "test_set_prereq" for how to define your own. 991 992 - PYTHON 993 994 Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that 995 need Python with this. 996 997 - PERL 998 999 Git wasn't compiled with NO_PERL=YesPlease.10001001 Even without the PERL prerequisite, tests can assume there is a1002 usable perl interpreter at $PERL_PATH, though it need not be1003 particularly modern.10041005 - POSIXPERM10061007 The filesystem supports POSIX style permission bits.10081009 - BSLASHPSPEC10101011 Backslashes in pathspec are not directory separators. This is not1012 set on Windows. See 6fd1106a for details.10131014 - EXECKEEPSPID10151016 The process retains the same pid across exec(2). See fb9a2bea for1017 details.10181019 - PIPE10201021 The filesystem we're on supports creation of FIFOs (named pipes)1022 via mkfifo(1).10231024 - SYMLINKS10251026 The filesystem we're on supports symbolic links. E.g. a FAT1027 filesystem doesn't support these. See 704a3143 for details.10281029 - SANITY10301031 Test is not run by root user, and an attempt to write to an1032 unwritable file is expected to fail correctly.10331034 - PCRE10351036 Git was compiled with support for PCRE. Wrap any tests1037 that use git-grep --perl-regexp or git-grep -P in these.10381039 - LIBPCRE110401041 Git was compiled with PCRE v1 support via1042 USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some1043 reason need v1 of the PCRE library instead of v2 in these.10441045 - LIBPCRE210461047 Git was compiled with PCRE v2 support via1048 USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some1049 reason need v2 of the PCRE library instead of v1 in these.10501051 - CASE_INSENSITIVE_FS10521053 Test is run on a case insensitive file system.10541055 - UTF8_NFD_TO_NFC10561057 Test is run on a filesystem which converts decomposed utf-8 (nfd)1058 to precomposed utf-8 (nfc).10591060 - PTHREADS10611062 Git wasn't compiled with NO_PTHREADS=YesPlease.10631064Tips for Writing Tests1065----------------------10661067As with any programming projects, existing programs are the best1068source of the information. However, do _not_ emulate1069t0000-basic.sh when writing your tests. The test is special in1070that it tries to validate the very core of GIT. For example, it1071knows that there will be 256 subdirectories under .git/objects/,1072and it knows that the object ID of an empty tree is a certain107340-byte string. This is deliberately done so in t0000-basic.sh1074because the things the very basic core test tries to achieve is1075to serve as a basis for people who are changing the GIT internal1076drastically. For these people, after making certain changes,1077not seeing failures from the basic test _is_ a failure. And1078such drastic changes to the core GIT that even changes these1079otherwise supposedly stable object IDs should be accompanied by1080an update to t0000-basic.sh.10811082However, other tests that simply rely on basic parts of the core1083GIT working properly should not have that level of intimate1084knowledge of the core GIT internals. If all the test scripts1085hardcoded the object IDs like t0000-basic.sh does, that defeats1086the purpose of t0000-basic.sh, which is to isolate that level of1087validation in one place. Your test also ends up needing1088updating when such a change to the internal happens, so do _not_1089do it and leave the low level of validation to t0000-basic.sh.10901091Test coverage1092-------------10931094You can use the coverage tests to find code paths that are not being1095used or properly exercised yet.10961097To do that, run the coverage target at the top-level (not in the t/1098directory):10991100 make coverage11011102That'll compile Git with GCC's coverage arguments, and generate a test1103report with gcov after the tests finish. Running the coverage tests1104can take a while, since running the tests in parallel is incompatible1105with GCC's coverage mode.11061107After the tests have run you can generate a list of untested1108functions:11091110 make coverage-untested-functions11111112You can also generate a detailed per-file HTML report using the1113Devel::Cover module. To install it do:11141115 # On Debian or Ubuntu:1116 sudo aptitude install libdevel-cover-perl11171118 # From the CPAN with cpanminus1119 curl -L http://cpanmin.us | perl - --sudo --self-upgrade1120 cpanm --sudo Devel::Cover11211122Then, at the top-level:11231124 make cover_db_html11251126That'll generate a detailed cover report in the "cover_db_html"1127directory, which you can then copy to a webserver, or inspect locally1128in a browser.