1Core GIT Tests 2============== 3 4This directory holds many test scripts for core GIT tools. The 5first part of this short document describes how to run the tests 6and read their output. 7 8When fixing the tools or adding enhancements, you are strongly 9encouraged to add tests in this directory to cover what you are 10trying to fix or enhance. The later part of this short document 11describes how your test scripts should be organized. 12 13 14Running Tests 15------------- 16 17The easiest way to run tests is to say "make". This runs all 18the tests. 19 20 *** t0000-basic.sh *** 21 ok 1 - .git/objects should be empty after git init in an empty repo. 22 ok 2 - .git/objects should have 3 subdirectories. 23 ok 3 - success is reported like this 24 ... 25 ok 43 - very long name in the index handled sanely 26 # fixed 1 known breakage(s) 27 # still have 1 known breakage(s) 28 # passed all remaining 42 test(s) 29 1..43 30 *** t0001-init.sh *** 31 ok 1 - plain 32 ok 2 - plain with GIT_WORK_TREE 33 ok 3 - plain bare 34 35Since the tests all output TAP (see http://testanything.org) they can 36be run with any TAP harness. Here's an example of parallel testing 37powered by a recent version of prove(1): 38 39 $ prove --timer --jobs 15 ./t[0-9]*.sh 40 [19:17:33] ./t0005-signals.sh ................................... ok 36 ms 41 [19:17:33] ./t0022-crlf-rename.sh ............................... ok 69 ms 42 [19:17:33] ./t0024-crlf-archive.sh .............................. ok 154 ms 43 [19:17:33] ./t0004-unwritable.sh ................................ ok 289 ms 44 [19:17:33] ./t0002-gitfile.sh ................................... ok 480 ms 45 ===( 102;0 25/? 6/? 5/? 16/? 1/? 4/? 2/? 1/? 3/? 1... )=== 46 47prove and other harnesses come with a lot of useful options. The 48--state option in particular is very useful: 49 50 # Repeat until no more failures 51 $ prove -j 15 --state=failed,save ./t[0-9]*.sh 52 53You can give DEFAULT_TEST_TARGET=prove on the make command (or define it 54in config.mak) to cause "make test" to run tests under prove. 55GIT_PROVE_OPTS can be used to pass additional options, e.g. 56 57 $ make DEFAULT_TEST_TARGET=prove GIT_PROVE_OPTS='--timer --jobs 16' test 58 59You can also run each test individually from command line, like this: 60 61 $ sh ./t3010-ls-files-killed-modified.sh 62 ok 1 - git update-index --add to add various paths. 63 ok 2 - git ls-files -k to show killed files. 64 ok 3 - validate git ls-files -k output. 65 ok 4 - git ls-files -m to show modified files. 66 ok 5 - validate git ls-files -m output. 67 # passed all 5 test(s) 68 1..5 69 70You can pass --verbose (or -v), --debug (or -d), and --immediate 71(or -i) command line argument to the test, or by setting GIT_TEST_OPTS 72appropriately before running "make". 73 74-v:: 75--verbose:: 76 This makes the test more verbose. Specifically, the 77 command being run and their output if any are also 78 output. 79 80--verbose-only=<pattern>:: 81 Like --verbose, but the effect is limited to tests with 82 numbers matching <pattern>. The number matched against is 83 simply the running count of the test within the file. 84 85-x:: 86 Turn on shell tracing (i.e., `set -x`) during the tests 87 themselves. Implies `--verbose`. Note that in non-bash shells, 88 this can cause failures in some tests which redirect and test 89 the output of shell functions. Use with caution. 90 Ignored in test scripts that set the variable 'test_untraceable' 91 to a non-empty value, unless it's run with a Bash version 92 supporting BASH_XTRACEFD, i.e. v4.1 or later. 93 94-d:: 95--debug:: 96 This may help the person who is developing a new test. 97 It causes the command defined with test_debug to run. 98 The "trash" directory (used to store all temporary data 99 during testing) is not deleted even if there are no 100 failed tests so that you can inspect its contents after 101 the test finished. 102 103-i:: 104--immediate:: 105 This causes the test to immediately exit upon the first 106 failed test. Cleanup commands requested with 107 test_when_finished are not executed if the test failed, 108 in order to keep the state for inspection by the tester 109 to diagnose the bug. 110 111-l:: 112--long-tests:: 113 This causes additional long-running tests to be run (where 114 available), for more exhaustive testing. 115 116-r:: 117--run=<test-selector>:: 118 Run only the subset of tests indicated by 119 <test-selector>. See section "Skipping Tests" below for 120 <test-selector> syntax. 121 122--valgrind=<tool>:: 123 Execute all Git binaries under valgrind tool <tool> and exit 124 with status 126 on errors (just like regular tests, this will 125 only stop the test script when running under -i). 126 127 Since it makes no sense to run the tests with --valgrind and 128 not see any output, this option implies --verbose. For 129 convenience, it also implies --tee. 130 131 <tool> defaults to 'memcheck', just like valgrind itself. 132 Other particularly useful choices include 'helgrind' and 133 'drd', but you may use any tool recognized by your valgrind 134 installation. 135 136 As a special case, <tool> can be 'memcheck-fast', which uses 137 memcheck but disables --track-origins. Use this if you are 138 running tests in bulk, to see if there are _any_ memory 139 issues. 140 141 Note that memcheck is run with the option --leak-check=no, 142 as the git process is short-lived and some errors are not 143 interesting. In order to run a single command under the same 144 conditions manually, you should set GIT_VALGRIND to point to 145 the 't/valgrind/' directory and use the commands under 146 't/valgrind/bin/'. 147 148--valgrind-only=<pattern>:: 149 Like --valgrind, but the effect is limited to tests with 150 numbers matching <pattern>. The number matched against is 151 simply the running count of the test within the file. 152 153--tee:: 154 In addition to printing the test output to the terminal, 155 write it to files named 't/test-results/$TEST_NAME.out'. 156 As the names depend on the tests' file names, it is safe to 157 run the tests with this option in parallel. 158 159--verbose-log:: 160 Write verbose output to the same logfile as `--tee`, but do 161 _not_ write it to stdout. Unlike `--tee --verbose`, this option 162 is safe to use when stdout is being consumed by a TAP parser 163 like `prove`. Implies `--tee` and `--verbose`. 164 165--with-dashes:: 166 By default tests are run without dashed forms of 167 commands (like git-commit) in the PATH (it only uses 168 wrappers from ../bin-wrappers). Use this option to include 169 the build directory (..) in the PATH, which contains all 170 the dashed forms of commands. This option is currently 171 implied by other options like --valgrind and 172 GIT_TEST_INSTALLED. 173 174--root=<directory>:: 175 Create "trash" directories used to store all temporary data during 176 testing under <directory>, instead of the t/ directory. 177 Using this option with a RAM-based filesystem (such as tmpfs) 178 can massively speed up the test suite. 179 180--chain-lint:: 181--no-chain-lint:: 182 If --chain-lint is enabled, the test harness will check each 183 test to make sure that it properly "&&-chains" all commands (so 184 that a failure in the middle does not go unnoticed by the final 185 exit code of the test). This check is performed in addition to 186 running the tests themselves. You may also enable or disable 187 this feature by setting the GIT_TEST_CHAIN_LINT environment 188 variable to "1" or "0", respectively. 189 190You can also set the GIT_TEST_INSTALLED environment variable to 191the bindir of an existing git installation to test that installation. 192You still need to have built this git sandbox, from which various 193test-* support programs, templates, and perl libraries are used. 194If your installed git is incomplete, it will silently test parts of 195your built version instead. 196 197When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to 198override the location of the dashed-form subcommands (what 199GIT_EXEC_PATH would be used for during normal operation). 200GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`. 201 202 203Skipping Tests 204-------------- 205 206In some environments, certain tests have no way of succeeding 207due to platform limitation, such as lack of 'unzip' program, or 208filesystem that do not allow arbitrary sequence of non-NUL bytes 209as pathnames. 210 211You should be able to say something like 212 213 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh 214 215and even: 216 217 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make 218 219to omit such tests. The value of the environment variable is a 220SP separated list of patterns that tells which tests to skip, 221and either can match the "t[0-9]{4}" part to skip the whole 222test, or t[0-9]{4} followed by ".$number" to say which 223particular test to skip. 224 225For an individual test suite --run could be used to specify that 226only some tests should be run or that some tests should be 227excluded from a run. 228 229The argument for --run is a list of individual test numbers or 230ranges with an optional negation prefix that define what tests in 231a test suite to include in the run. A range is two numbers 232separated with a dash and matches a range of tests with both ends 233been included. You may omit the first or the second number to 234mean "from the first test" or "up to the very last test" 235respectively. 236 237Optional prefix of '!' means that the test or a range of tests 238should be excluded from the run. 239 240If --run starts with an unprefixed number or range the initial 241set of tests to run is empty. If the first item starts with '!' 242all the tests are added to the initial set. After initial set is 243determined every test number or range is added or excluded from 244the set one by one, from left to right. 245 246Individual numbers or ranges could be separated either by a space 247or a comma. 248 249For example, to run only tests up to a specific test (21), one 250could do this: 251 252 $ sh ./t9200-git-cvsexport-commit.sh --run='1-21' 253 254or this: 255 256 $ sh ./t9200-git-cvsexport-commit.sh --run='-21' 257 258Common case is to run several setup tests (1, 2, 3) and then a 259specific test (21) that relies on that setup: 260 261 $ sh ./t9200-git-cvsexport-commit.sh --run='1 2 3 21' 262 263or: 264 265 $ sh ./t9200-git-cvsexport-commit.sh --run=1,2,3,21 266 267or: 268 269 $ sh ./t9200-git-cvsexport-commit.sh --run='-3 21' 270 271As noted above, the test set is built by going through the items 272from left to right, so this: 273 274 $ sh ./t9200-git-cvsexport-commit.sh --run='1-4 !3' 275 276will run tests 1, 2, and 4. Items that come later have higher 277precedence. It means that this: 278 279 $ sh ./t9200-git-cvsexport-commit.sh --run='!3 1-4' 280 281would just run tests from 1 to 4, including 3. 282 283You may use negation with ranges. The following will run all 284test in the test suite except from 7 up to 11: 285 286 $ sh ./t9200-git-cvsexport-commit.sh --run='!7-11' 287 288Some tests in a test suite rely on the previous tests performing 289certain actions, specifically some tests are designated as 290"setup" test, so you cannot _arbitrarily_ disable one test and 291expect the rest to function correctly. 292 293--run is mostly useful when you want to focus on a specific test 294and know what setup is needed for it. Or when you want to run 295everything up to a certain test. 296 297 298Naming Tests 299------------ 300 301The test files are named as: 302 303 tNNNN-commandname-details.sh 304 305where N is a decimal digit. 306 307First digit tells the family: 308 309 0 - the absolute basics and global stuff 310 1 - the basic commands concerning database 311 2 - the basic commands concerning the working tree 312 3 - the other basic commands (e.g. ls-files) 313 4 - the diff commands 314 5 - the pull and exporting commands 315 6 - the revision tree commands (even e.g. merge-base) 316 7 - the porcelainish commands concerning the working tree 317 8 - the porcelainish commands concerning forensics 318 9 - the git tools 319 320Second digit tells the particular command we are testing. 321 322Third digit (optionally) tells the particular switch or group of switches 323we are testing. 324 325If you create files under t/ directory (i.e. here) that is not 326the top-level test script, never name the file to match the above 327pattern. The Makefile here considers all such files as the 328top-level test script and tries to run all of them. Care is 329especially needed if you are creating a common test library 330file, similar to test-lib.sh, because such a library file may 331not be suitable for standalone execution. 332 333 334Writing Tests 335------------- 336 337The test script is written as a shell script. It should start 338with the standard "#!/bin/sh", and an 339assignment to variable 'test_description', like this: 340 341 #!/bin/sh 342 343 test_description='xxx test (option --frotz) 344 345 This test registers the following structure in the cache 346 and tries to run git-ls-files with option --frotz.' 347 348 349Source 'test-lib.sh' 350-------------------- 351 352After assigning test_description, the test script should source 353test-lib.sh like this: 354 355 . ./test-lib.sh 356 357This test harness library does the following things: 358 359 - If the script is invoked with command line argument --help 360 (or -h), it shows the test_description and exits. 361 362 - Creates an empty test directory with an empty .git/objects database 363 and chdir(2) into it. This directory is 't/trash 364 directory.$test_name_without_dotsh', with t/ subject to change by 365 the --root option documented above. 366 367 - Defines standard test helper functions for your scripts to 368 use. These functions are designed to make all scripts behave 369 consistently when command line arguments --verbose (or -v), 370 --debug (or -d), and --immediate (or -i) is given. 371 372Do's, don'ts & things to keep in mind 373------------------------------------- 374 375Here are a few examples of things you probably should and shouldn't do 376when writing tests. 377 378Do: 379 380 - Put all code inside test_expect_success and other assertions. 381 382 Even code that isn't a test per se, but merely some setup code 383 should be inside a test assertion. 384 385 - Chain your test assertions 386 387 Write test code like this: 388 389 git merge foo && 390 git push bar && 391 test ... 392 393 Instead of: 394 395 git merge hla 396 git push gh 397 test ... 398 399 That way all of the commands in your tests will succeed or fail. If 400 you must ignore the return value of something, consider using a 401 helper function (e.g. use sane_unset instead of unset, in order 402 to avoid unportable return value for unsetting a variable that was 403 already unset), or prepending the command with test_might_fail or 404 test_must_fail. 405 406 - Check the test coverage for your tests. See the "Test coverage" 407 below. 408 409 Don't blindly follow test coverage metrics; if a new function you added 410 doesn't have any coverage, then you're probably doing something wrong, 411 but having 100% coverage doesn't necessarily mean that you tested 412 everything. 413 414 Tests that are likely to smoke out future regressions are better 415 than tests that just inflate the coverage metrics. 416 417 - When a test checks for an absolute path that a git command generated, 418 construct the expected value using $(pwd) rather than $PWD, 419 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on 420 Windows, where the shell (MSYS bash) mangles absolute path names. 421 For details, see the commit message of 4114156ae9. 422 423Don't: 424 425 - exit() within a <script> part. 426 427 The harness will catch this as a programming error of the test. 428 Use test_done instead if you need to stop the tests early (see 429 "Skipping tests" below). 430 431 - use '! git cmd' when you want to make sure the git command exits 432 with failure in a controlled way by calling "die()". Instead, 433 use 'test_must_fail git cmd'. This will signal a failure if git 434 dies in an unexpected way (e.g. segfault). 435 436 On the other hand, don't use test_must_fail for running regular 437 platform commands; just use '! cmd'. We are not in the business 438 of verifying that the world given to us sanely works. 439 440 - use perl without spelling it as "$PERL_PATH". This is to help our 441 friends on Windows where the platform Perl often adds CR before 442 the end of line, and they bundle Git with a version of Perl that 443 does not do so, whose path is specified with $PERL_PATH. Note that we 444 provide a "perl" function which uses $PERL_PATH under the hood, so 445 you do not need to worry when simply running perl in the test scripts 446 (but you do, for example, on a shebang line or in a sub script 447 created via "write_script"). 448 449 - use sh without spelling it as "$SHELL_PATH", when the script can 450 be misinterpreted by broken platform shell (e.g. Solaris). 451 452 - chdir around in tests. It is not sufficient to chdir to 453 somewhere and then chdir back to the original location later in 454 the test, as any intermediate step can fail and abort the test, 455 causing the next test to start in an unexpected directory. Do so 456 inside a subshell if necessary. 457 458 - Break the TAP output 459 460 The raw output from your test may be interpreted by a TAP harness. TAP 461 harnesses will ignore everything they don't know about, but don't step 462 on their toes in these areas: 463 464 - Don't print lines like "$x..$y" where $x and $y are integers. 465 466 - Don't print lines that begin with "ok" or "not ok". 467 468 TAP harnesses expect a line that begins with either "ok" and "not 469 ok" to signal a test passed or failed (and our harness already 470 produces such lines), so your script shouldn't emit such lines to 471 their output. 472 473 You can glean some further possible issues from the TAP grammar 474 (see https://metacpan.org/pod/TAP::Parser::Grammar#TAP-GRAMMAR) 475 but the best indication is to just run the tests with prove(1), 476 it'll complain if anything is amiss. 477 478Keep in mind: 479 480 - Inside the <script> part, the standard output and standard error 481 streams are discarded, and the test harness only reports "ok" or 482 "not ok" to the end user running the tests. Under --verbose, they 483 are shown to help debugging the tests. 484 485 486Skipping tests 487-------------- 488 489If you need to skip tests you should do so by using the three-arg form 490of the test_* functions (see the "Test harness library" section 491below), e.g.: 492 493 test_expect_success PERL 'I need Perl' ' 494 perl -e "hlagh() if unf_unf()" 495 ' 496 497The advantage of skipping tests like this is that platforms that don't 498have the PERL and other optional dependencies get an indication of how 499many tests they're missing. 500 501If the test code is too hairy for that (i.e. does a lot of setup work 502outside test assertions) you can also skip all remaining tests by 503setting skip_all and immediately call test_done: 504 505 if ! test_have_prereq PERL 506 then 507 skip_all='skipping perl interface tests, perl not available' 508 test_done 509 fi 510 511The string you give to skip_all will be used as an explanation for why 512the test was skipped. 513 514End with test_done 515------------------ 516 517Your script will be a sequence of tests, using helper functions 518from the test harness library. At the end of the script, call 519'test_done'. 520 521 522Test harness library 523-------------------- 524 525There are a handful helper functions defined in the test harness 526library for your script to use. 527 528 - test_expect_success [<prereq>] <message> <script> 529 530 Usually takes two strings as parameters, and evaluates the 531 <script>. If it yields success, test is considered 532 successful. <message> should state what it is testing. 533 534 Example: 535 536 test_expect_success \ 537 'git-write-tree should be able to write an empty tree.' \ 538 'tree=$(git-write-tree)' 539 540 If you supply three parameters the first will be taken to be a 541 prerequisite; see the test_set_prereq and test_have_prereq 542 documentation below: 543 544 test_expect_success TTY 'git --paginate rev-list uses a pager' \ 545 ' ... ' 546 547 You can also supply a comma-separated list of prerequisites, in the 548 rare case where your test depends on more than one: 549 550 test_expect_success PERL,PYTHON 'yo dawg' \ 551 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" ' 552 553 - test_expect_failure [<prereq>] <message> <script> 554 555 This is NOT the opposite of test_expect_success, but is used 556 to mark a test that demonstrates a known breakage. Unlike 557 the usual test_expect_success tests, which say "ok" on 558 success and "FAIL" on failure, this will say "FIXED" on 559 success and "still broken" on failure. Failures from these 560 tests won't cause -i (immediate) to stop. 561 562 Like test_expect_success this function can optionally use a three 563 argument invocation with a prerequisite as the first argument. 564 565 - test_debug <script> 566 567 This takes a single argument, <script>, and evaluates it only 568 when the test script is started with --debug command line 569 argument. This is primarily meant for use during the 570 development of a new test script. 571 572 - debug <git-command> 573 574 Run a git command inside a debugger. This is primarily meant for 575 use when debugging a failing test script. 576 577 - test_done 578 579 Your test script must have test_done at the end. Its purpose 580 is to summarize successes and failures in the test script and 581 exit with an appropriate error code. 582 583 - test_tick 584 585 Make commit and tag names consistent by setting the author and 586 committer times to defined state. Subsequent calls will 587 advance the times by a fixed amount. 588 589 - test_commit <message> [<filename> [<contents>]] 590 591 Creates a commit with the given message, committing the given 592 file with the given contents (default for both is to reuse the 593 message string), and adds a tag (again reusing the message 594 string as name). Calls test_tick to make the SHA-1s 595 reproducible. 596 597 - test_merge <message> <commit-or-tag> 598 599 Merges the given rev using the given message. Like test_commit, 600 creates a tag and calls test_tick before committing. 601 602 - test_set_prereq <prereq> 603 604 Set a test prerequisite to be used later with test_have_prereq. The 605 test-lib will set some prerequisites for you, see the 606 "Prerequisites" section below for a full list of these. 607 608 Others you can set yourself and use later with either 609 test_have_prereq directly, or the three argument invocation of 610 test_expect_success and test_expect_failure. 611 612 - test_have_prereq <prereq> 613 614 Check if we have a prerequisite previously set with test_set_prereq. 615 The most common way to use this explicitly (as opposed to the 616 implicit use when an argument is passed to test_expect_*) is to skip 617 all the tests at the start of the test script if we don't have some 618 essential prerequisite: 619 620 if ! test_have_prereq PERL 621 then 622 skip_all='skipping perl interface tests, perl not available' 623 test_done 624 fi 625 626 - test_external [<prereq>] <message> <external> <script> 627 628 Execute a <script> with an <external> interpreter (like perl). This 629 was added for tests like t9700-perl-git.sh which do most of their 630 work in an external test script. 631 632 test_external \ 633 'GitwebCache::*FileCache*' \ 634 perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl 635 636 If the test is outputting its own TAP you should set the 637 test_external_has_tap variable somewhere before calling the first 638 test_external* function. See t9700-perl-git.sh for an example. 639 640 # The external test will outputs its own plan 641 test_external_has_tap=1 642 643 - test_external_without_stderr [<prereq>] <message> <external> <script> 644 645 Like test_external but fail if there's any output on stderr, 646 instead of checking the exit code. 647 648 test_external_without_stderr \ 649 'Perl API' \ 650 perl "$TEST_DIRECTORY"/t9700/test.pl 651 652 - test_expect_code <exit-code> <command> 653 654 Run a command and ensure that it exits with the given exit code. 655 For example: 656 657 test_expect_success 'Merge with d/f conflicts' ' 658 test_expect_code 1 git merge "merge msg" B master 659 ' 660 661 - test_must_fail <git-command> 662 663 Run a git command and ensure it fails in a controlled way. Use 664 this instead of "! <git-command>". When git-command dies due to a 665 segfault, test_must_fail diagnoses it as an error; "! <git-command>" 666 treats it as just another expected failure, which would let such a 667 bug go unnoticed. 668 669 - test_might_fail <git-command> 670 671 Similar to test_must_fail, but tolerate success, too. Use this 672 instead of "<git-command> || :" to catch failures due to segv. 673 674 - test_cmp <expected> <actual> 675 676 Check whether the content of the <actual> file matches the 677 <expected> file. This behaves like "cmp" but produces more 678 helpful output when the test is run with "-v" option. 679 680 - test_cmp_rev <expected> <actual> 681 682 Check whether the <expected> rev points to the same commit as the 683 <actual> rev. 684 685 - test_line_count (= | -lt | -ge | ...) <length> <file> 686 687 Check whether a file has the length it is expected to. 688 689 - test_path_is_file <path> [<diagnosis>] 690 test_path_is_dir <path> [<diagnosis>] 691 test_path_is_missing <path> [<diagnosis>] 692 693 Check if the named path is a file, if the named path is a 694 directory, or if the named path does not exist, respectively, 695 and fail otherwise, showing the <diagnosis> text. 696 697 - test_when_finished <script> 698 699 Prepend <script> to a list of commands to run to clean up 700 at the end of the current test. If some clean-up command 701 fails, the test will not pass. 702 703 Example: 704 705 test_expect_success 'branch pointing to non-commit' ' 706 git rev-parse HEAD^{tree} >.git/refs/heads/invalid && 707 test_when_finished "git update-ref -d refs/heads/invalid" && 708 ... 709 ' 710 711 - test_write_lines <lines> 712 713 Write <lines> on standard output, one line per argument. 714 Useful to prepare multi-line files in a compact form. 715 716 Example: 717 718 test_write_lines a b c d e f g >foo 719 720 Is a more compact equivalent of: 721 cat >foo <<-EOF 722 a 723 b 724 c 725 d 726 e 727 f 728 g 729 EOF 730 731 732 - test_pause 733 734 This command is useful for writing and debugging tests and must be 735 removed before submitting. It halts the execution of the test and 736 spawns a shell in the trash directory. Exit the shell to continue 737 the test. Example: 738 739 test_expect_success 'test' ' 740 git do-something >actual && 741 test_pause && 742 test_cmp expected actual 743 ' 744 745 - test_ln_s_add <path1> <path2> 746 747 This function helps systems whose filesystem does not support symbolic 748 links. Use it to add a symbolic link entry to the index when it is not 749 important that the file system entry is a symbolic link, i.e., instead 750 of the sequence 751 752 ln -s foo bar && 753 git add bar 754 755 Sometimes it is possible to split a test in a part that does not need 756 the symbolic link in the file system and a part that does; then only 757 the latter part need be protected by a SYMLINKS prerequisite (see below). 758 759Prerequisites 760------------- 761 762These are the prerequisites that the test library predefines with 763test_have_prereq. 764 765See the prereq argument to the test_* functions in the "Test harness 766library" section above and the "test_have_prereq" function for how to 767use these, and "test_set_prereq" for how to define your own. 768 769 - PYTHON 770 771 Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that 772 need Python with this. 773 774 - PERL 775 776 Git wasn't compiled with NO_PERL=YesPlease. 777 778 Even without the PERL prerequisite, tests can assume there is a 779 usable perl interpreter at $PERL_PATH, though it need not be 780 particularly modern. 781 782 - POSIXPERM 783 784 The filesystem supports POSIX style permission bits. 785 786 - BSLASHPSPEC 787 788 Backslashes in pathspec are not directory separators. This is not 789 set on Windows. See 6fd1106a for details. 790 791 - EXECKEEPSPID 792 793 The process retains the same pid across exec(2). See fb9a2bea for 794 details. 795 796 - PIPE 797 798 The filesystem we're on supports creation of FIFOs (named pipes) 799 via mkfifo(1). 800 801 - SYMLINKS 802 803 The filesystem we're on supports symbolic links. E.g. a FAT 804 filesystem doesn't support these. See 704a3143 for details. 805 806 - SANITY 807 808 Test is not run by root user, and an attempt to write to an 809 unwritable file is expected to fail correctly. 810 811 - PCRE 812 813 Git was compiled with support for PCRE. Wrap any tests 814 that use git-grep --perl-regexp or git-grep -P in these. 815 816 - LIBPCRE1 817 818 Git was compiled with PCRE v1 support via 819 USE_LIBPCRE1=YesPlease. Wrap any PCRE using tests that for some 820 reason need v1 of the PCRE library instead of v2 in these. 821 822 - LIBPCRE2 823 824 Git was compiled with PCRE v2 support via 825 USE_LIBPCRE2=YesPlease. Wrap any PCRE using tests that for some 826 reason need v2 of the PCRE library instead of v1 in these. 827 828 - CASE_INSENSITIVE_FS 829 830 Test is run on a case insensitive file system. 831 832 - UTF8_NFD_TO_NFC 833 834 Test is run on a filesystem which converts decomposed utf-8 (nfd) 835 to precomposed utf-8 (nfc). 836 837 - PTHREADS 838 839 Git wasn't compiled with NO_PTHREADS=YesPlease. 840 841Tips for Writing Tests 842---------------------- 843 844As with any programming projects, existing programs are the best 845source of the information. However, do _not_ emulate 846t0000-basic.sh when writing your tests. The test is special in 847that it tries to validate the very core of GIT. For example, it 848knows that there will be 256 subdirectories under .git/objects/, 849and it knows that the object ID of an empty tree is a certain 85040-byte string. This is deliberately done so in t0000-basic.sh 851because the things the very basic core test tries to achieve is 852to serve as a basis for people who are changing the GIT internal 853drastically. For these people, after making certain changes, 854not seeing failures from the basic test _is_ a failure. And 855such drastic changes to the core GIT that even changes these 856otherwise supposedly stable object IDs should be accompanied by 857an update to t0000-basic.sh. 858 859However, other tests that simply rely on basic parts of the core 860GIT working properly should not have that level of intimate 861knowledge of the core GIT internals. If all the test scripts 862hardcoded the object IDs like t0000-basic.sh does, that defeats 863the purpose of t0000-basic.sh, which is to isolate that level of 864validation in one place. Your test also ends up needing 865updating when such a change to the internal happens, so do _not_ 866do it and leave the low level of validation to t0000-basic.sh. 867 868Test coverage 869------------- 870 871You can use the coverage tests to find code paths that are not being 872used or properly exercised yet. 873 874To do that, run the coverage target at the top-level (not in the t/ 875directory): 876 877 make coverage 878 879That'll compile Git with GCC's coverage arguments, and generate a test 880report with gcov after the tests finish. Running the coverage tests 881can take a while, since running the tests in parallel is incompatible 882with GCC's coverage mode. 883 884After the tests have run you can generate a list of untested 885functions: 886 887 make coverage-untested-functions 888 889You can also generate a detailed per-file HTML report using the 890Devel::Cover module. To install it do: 891 892 # On Debian or Ubuntu: 893 sudo aptitude install libdevel-cover-perl 894 895 # From the CPAN with cpanminus 896 curl -L http://cpanmin.us | perl - --sudo --self-upgrade 897 cpanm --sudo Devel::Cover 898 899Then, at the top-level: 900 901 make cover_db_html 902 903That'll generate a detailed cover report in the "cover_db_html" 904directory, which you can then copy to a webserver, or inspect locally 905in a browser.