Typos: t/README
[git/git.git] / t / README
1 Core GIT Tests
2 ==============
3
4 This directory holds many test scripts for core GIT tools. The
5 first part of this short document describes how to run the tests
6 and read their output.
7
8 When fixing the tools or adding enhancements, you are strongly
9 encouraged to add tests in this directory to cover what you are
10 trying to fix or enhance. The later part of this short document
11 describes how your test scripts should be organized.
12
13
14 Running Tests
15 -------------
16
17 The easiest way to run tests is to say "make". This runs all
18 the 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
35 Since the tests all output TAP (see http://testanything.org) they can
36 be run with any TAP harness. Here's an example of parallel testing
37 powered 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
47 prove 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
53 You can give DEFAULT_TEST_TARGET=prove on the make command (or define it
54 in config.mak) to cause "make test" to run tests under prove.
55 GIT_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
59 You 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
70 You 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
72 appropriately before running "make".
73
74 --verbose::
75 This makes the test more verbose. Specifically, the
76 command being run and their output if any are also
77 output.
78
79 --debug::
80 This may help the person who is developing a new test.
81 It causes the command defined with test_debug to run.
82
83 --immediate::
84 This causes the test to immediately exit upon the first
85 failed test.
86
87 --long-tests::
88 This causes additional long-running tests to be run (where
89 available), for more exhaustive testing.
90
91 --valgrind::
92 Execute all Git binaries with valgrind and exit with status
93 126 on errors (just like regular tests, this will only stop
94 the test script when running under -i). Valgrind errors
95 go to stderr, so you might want to pass the -v option, too.
96
97 Since it makes no sense to run the tests with --valgrind and
98 not see any output, this option implies --verbose. For
99 convenience, it also implies --tee.
100
101 Note that valgrind is run with the option --leak-check=no,
102 as the git process is short-lived and some errors are not
103 interesting. In order to run a single command under the same
104 conditions manually, you should set GIT_VALGRIND to point to
105 the 't/valgrind/' directory and use the commands under
106 't/valgrind/bin/'.
107
108 --tee::
109 In addition to printing the test output to the terminal,
110 write it to files named 't/test-results/$TEST_NAME.out'.
111 As the names depend on the tests' file names, it is safe to
112 run the tests with this option in parallel.
113
114 --with-dashes::
115 By default tests are run without dashed forms of
116 commands (like git-commit) in the PATH (it only uses
117 wrappers from ../bin-wrappers). Use this option to include
118 the build directory (..) in the PATH, which contains all
119 the dashed forms of commands. This option is currently
120 implied by other options like --valgrind and
121 GIT_TEST_INSTALLED.
122
123 --root=<directory>::
124 Create "trash" directories used to store all temporary data during
125 testing under <directory>, instead of the t/ directory.
126 Using this option with a RAM-based filesystem (such as tmpfs)
127 can massively speed up the test suite.
128
129 You can also set the GIT_TEST_INSTALLED environment variable to
130 the bindir of an existing git installation to test that installation.
131 You still need to have built this git sandbox, from which various
132 test-* support programs, templates, and perl libraries are used.
133 If your installed git is incomplete, it will silently test parts of
134 your built version instead.
135
136 When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
137 override the location of the dashed-form subcommands (what
138 GIT_EXEC_PATH would be used for during normal operation).
139 GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
140
141
142 Skipping Tests
143 --------------
144
145 In some environments, certain tests have no way of succeeding
146 due to platform limitation, such as lack of 'unzip' program, or
147 filesystem that do not allow arbitrary sequence of non-NUL bytes
148 as pathnames.
149
150 You should be able to say something like
151
152 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
153
154 and even:
155
156 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
157
158 to omit such tests. The value of the environment variable is a
159 SP separated list of patterns that tells which tests to skip,
160 and either can match the "t[0-9]{4}" part to skip the whole
161 test, or t[0-9]{4} followed by ".$number" to say which
162 particular test to skip.
163
164 Note that some tests in the existing test suite rely on previous
165 test item, so you cannot arbitrarily disable one and expect the
166 remainder of test to check what the test originally was intended
167 to check.
168
169
170 Naming Tests
171 ------------
172
173 The test files are named as:
174
175 tNNNN-commandname-details.sh
176
177 where N is a decimal digit.
178
179 First digit tells the family:
180
181 0 - the absolute basics and global stuff
182 1 - the basic commands concerning database
183 2 - the basic commands concerning the working tree
184 3 - the other basic commands (e.g. ls-files)
185 4 - the diff commands
186 5 - the pull and exporting commands
187 6 - the revision tree commands (even e.g. merge-base)
188 7 - the porcelainish commands concerning the working tree
189 8 - the porcelainish commands concerning forensics
190 9 - the git tools
191
192 Second digit tells the particular command we are testing.
193
194 Third digit (optionally) tells the particular switch or group of switches
195 we are testing.
196
197 If you create files under t/ directory (i.e. here) that is not
198 the top-level test script, never name the file to match the above
199 pattern. The Makefile here considers all such files as the
200 top-level test script and tries to run all of them. Care is
201 especially needed if you are creating a common test library
202 file, similar to test-lib.sh, because such a library file may
203 not be suitable for standalone execution.
204
205
206 Writing Tests
207 -------------
208
209 The test script is written as a shell script. It should start
210 with the standard "#!/bin/sh" with copyright notices, and an
211 assignment to variable 'test_description', like this:
212
213 #!/bin/sh
214 #
215 # Copyright (c) 2005 Junio C Hamano
216 #
217
218 test_description='xxx test (option --frotz)
219
220 This test registers the following structure in the cache
221 and tries to run git-ls-files with option --frotz.'
222
223
224 Source 'test-lib.sh'
225 --------------------
226
227 After assigning test_description, the test script should source
228 test-lib.sh like this:
229
230 . ./test-lib.sh
231
232 This test harness library does the following things:
233
234 - If the script is invoked with command line argument --help
235 (or -h), it shows the test_description and exits.
236
237 - Creates an empty test directory with an empty .git/objects database
238 and chdir(2) into it. This directory is 't/trash
239 directory.$test_name_without_dotsh', with t/ subject to change by
240 the --root option documented above.
241
242 - Defines standard test helper functions for your scripts to
243 use. These functions are designed to make all scripts behave
244 consistently when command line arguments --verbose (or -v),
245 --debug (or -d), and --immediate (or -i) is given.
246
247 Do's, don'ts & things to keep in mind
248 -------------------------------------
249
250 Here are a few examples of things you probably should and shouldn't do
251 when writing tests.
252
253 Do:
254
255 - Put all code inside test_expect_success and other assertions.
256
257 Even code that isn't a test per se, but merely some setup code
258 should be inside a test assertion.
259
260 - Chain your test assertions
261
262 Write test code like this:
263
264 git merge foo &&
265 git push bar &&
266 test ...
267
268 Instead of:
269
270 git merge hla
271 git push gh
272 test ...
273
274 That way all of the commands in your tests will succeed or fail. If
275 you must ignore the return value of something, consider using a
276 helper function (e.g. use sane_unset instead of unset, in order
277 to avoid unportable return value for unsetting a variable that was
278 already unset), or prepending the command with test_might_fail or
279 test_must_fail.
280
281 - Check the test coverage for your tests. See the "Test coverage"
282 below.
283
284 Don't blindly follow test coverage metrics; if a new function you added
285 doesn't have any coverage, then you're probably doing something wrong,
286 but having 100% coverage doesn't necessarily mean that you tested
287 everything.
288
289 Tests that are likely to smoke out future regressions are better
290 than tests that just inflate the coverage metrics.
291
292 - When a test checks for an absolute path that a git command generated,
293 construct the expected value using $(pwd) rather than $PWD,
294 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
295 Windows, where the shell (MSYS bash) mangles absolute path names.
296 For details, see the commit message of 4114156ae9.
297
298 Don't:
299
300 - exit() within a <script> part.
301
302 The harness will catch this as a programming error of the test.
303 Use test_done instead if you need to stop the tests early (see
304 "Skipping tests" below).
305
306 - Break the TAP output
307
308 The raw output from your test may be interpreted by a TAP harness. TAP
309 harnesses will ignore everything they don't know about, but don't step
310 on their toes in these areas:
311
312 - Don't print lines like "$x..$y" where $x and $y are integers.
313
314 - Don't print lines that begin with "ok" or "not ok".
315
316 TAP harnesses expect a line that begins with either "ok" and "not
317 ok" to signal a test passed or failed (and our harness already
318 produces such lines), so your script shouldn't emit such lines to
319 their output.
320
321 You can glean some further possible issues from the TAP grammar
322 (see http://search.cpan.org/perldoc?TAP::Parser::Grammar#TAP_Grammar)
323 but the best indication is to just run the tests with prove(1),
324 it'll complain if anything is amiss.
325
326 Keep in mind:
327
328 - Inside <script> part, the standard output and standard error
329 streams are discarded, and the test harness only reports "ok" or
330 "not ok" to the end user running the tests. Under --verbose, they
331 are shown to help debugging the tests.
332
333
334 Skipping tests
335 --------------
336
337 If you need to skip tests you should do so by using the three-arg form
338 of the test_* functions (see the "Test harness library" section
339 below), e.g.:
340
341 test_expect_success PERL 'I need Perl' "
342 '$PERL_PATH' -e 'hlagh() if unf_unf()'
343 "
344
345 The advantage of skipping tests like this is that platforms that don't
346 have the PERL and other optional dependencies get an indication of how
347 many tests they're missing.
348
349 If the test code is too hairy for that (i.e. does a lot of setup work
350 outside test assertions) you can also skip all remaining tests by
351 setting skip_all and immediately call test_done:
352
353 if ! test_have_prereq PERL
354 then
355 skip_all='skipping perl interface tests, perl not available'
356 test_done
357 fi
358
359 The string you give to skip_all will be used as an explanation for why
360 the test was skipped.
361
362 End with test_done
363 ------------------
364
365 Your script will be a sequence of tests, using helper functions
366 from the test harness library. At the end of the script, call
367 'test_done'.
368
369
370 Test harness library
371 --------------------
372
373 There are a handful helper functions defined in the test harness
374 library for your script to use.
375
376 - test_expect_success [<prereq>] <message> <script>
377
378 Usually takes two strings as parameter, and evaluates the
379 <script>. If it yields success, test is considered
380 successful. <message> should state what it is testing.
381
382 Example:
383
384 test_expect_success \
385 'git-write-tree should be able to write an empty tree.' \
386 'tree=$(git-write-tree)'
387
388 If you supply three parameters the first will be taken to be a
389 prerequisite, see the test_set_prereq and test_have_prereq
390 documentation below:
391
392 test_expect_success TTY 'git --paginate rev-list uses a pager' \
393 ' ... '
394
395 You can also supply a comma-separated list of prerequisites, in the
396 rare case where your test depends on more than one:
397
398 test_expect_success PERL,PYTHON 'yo dawg' \
399 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" '
400
401 - test_expect_failure [<prereq>] <message> <script>
402
403 This is NOT the opposite of test_expect_success, but is used
404 to mark a test that demonstrates a known breakage. Unlike
405 the usual test_expect_success tests, which say "ok" on
406 success and "FAIL" on failure, this will say "FIXED" on
407 success and "still broken" on failure. Failures from these
408 tests won't cause -i (immediate) to stop.
409
410 Like test_expect_success this function can optionally use a three
411 argument invocation with a prerequisite as the first argument.
412
413 - test_debug <script>
414
415 This takes a single argument, <script>, and evaluates it only
416 when the test script is started with --debug command line
417 argument. This is primarily meant for use during the
418 development of a new test script.
419
420 - test_done
421
422 Your test script must have test_done at the end. Its purpose
423 is to summarize successes and failures in the test script and
424 exit with an appropriate error code.
425
426 - test_tick
427
428 Make commit and tag names consistent by setting the author and
429 committer times to defined state. Subsequent calls will
430 advance the times by a fixed amount.
431
432 - test_commit <message> [<filename> [<contents>]]
433
434 Creates a commit with the given message, committing the given
435 file with the given contents (default for both is to reuse the
436 message string), and adds a tag (again reusing the message
437 string as name). Calls test_tick to make the SHA-1s
438 reproducible.
439
440 - test_merge <message> <commit-or-tag>
441
442 Merges the given rev using the given message. Like test_commit,
443 creates a tag and calls test_tick before committing.
444
445 - test_set_prereq SOME_PREREQ
446
447 Set a test prerequisite to be used later with test_have_prereq. The
448 test-lib will set some prerequisites for you, see the
449 "Prerequisites" section below for a full list of these.
450
451 Others you can set yourself and use later with either
452 test_have_prereq directly, or the three argument invocation of
453 test_expect_success and test_expect_failure.
454
455 - test_have_prereq SOME PREREQ
456
457 Check if we have a prerequisite previously set with
458 test_set_prereq. The most common use of this directly is to skip
459 all the tests if we don't have some essential prerequisite:
460
461 if ! test_have_prereq PERL
462 then
463 skip_all='skipping perl interface tests, perl not available'
464 test_done
465 fi
466
467 - test_external [<prereq>] <message> <external> <script>
468
469 Execute a <script> with an <external> interpreter (like perl). This
470 was added for tests like t9700-perl-git.sh which do most of their
471 work in an external test script.
472
473 test_external \
474 'GitwebCache::*FileCache*' \
475 "$PERL_PATH" "$TEST_DIRECTORY"/t9503/test_cache_interface.pl
476
477 If the test is outputting its own TAP you should set the
478 test_external_has_tap variable somewhere before calling the first
479 test_external* function. See t9700-perl-git.sh for an example.
480
481 # The external test will outputs its own plan
482 test_external_has_tap=1
483
484 - test_external_without_stderr [<prereq>] <message> <external> <script>
485
486 Like test_external but fail if there's any output on stderr,
487 instead of checking the exit code.
488
489 test_external_without_stderr \
490 'Perl API' \
491 "$PERL_PATH" "$TEST_DIRECTORY"/t9700/test.pl
492
493 - test_expect_code <exit-code> <command>
494
495 Run a command and ensure that it exits with the given exit code.
496 For example:
497
498 test_expect_success 'Merge with d/f conflicts' '
499 test_expect_code 1 git merge "merge msg" B master
500 '
501
502 - test_must_fail <git-command>
503
504 Run a git command and ensure it fails in a controlled way. Use
505 this instead of "! <git-command>". When git-command dies due to a
506 segfault, test_must_fail diagnoses it as an error; "! <git-command>"
507 treats it as just another expected failure, which would let such a
508 bug go unnoticed.
509
510 - test_might_fail <git-command>
511
512 Similar to test_must_fail, but tolerate success, too. Use this
513 instead of "<git-command> || :" to catch failures due to segv.
514
515 - test_cmp <expected> <actual>
516
517 Check whether the content of the <actual> file matches the
518 <expected> file. This behaves like "cmp" but produces more
519 helpful output when the test is run with "-v" option.
520
521 - test_line_count (= | -lt | -ge | ...) <length> <file>
522
523 Check whether a file has the length it is expected to.
524
525 - test_path_is_file <file> [<diagnosis>]
526 test_path_is_dir <dir> [<diagnosis>]
527 test_path_is_missing <path> [<diagnosis>]
528
529 Check whether a file/directory exists or doesn't. <diagnosis> will
530 be displayed if the test fails.
531
532 - test_when_finished <script>
533
534 Prepend <script> to a list of commands to run to clean up
535 at the end of the current test. If some clean-up command
536 fails, the test will not pass.
537
538 Example:
539
540 test_expect_success 'branch pointing to non-commit' '
541 git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
542 test_when_finished "git update-ref -d refs/heads/invalid" &&
543 ...
544 '
545
546 Prerequisites
547 -------------
548
549 These are the prerequisites that the test library predefines with
550 test_have_prereq.
551
552 See the prereq argument to the test_* functions in the "Test harness
553 library" section above and the "test_have_prereq" function for how to
554 use these, and "test_set_prereq" for how to define your own.
555
556 - PERL & PYTHON
557
558 Git wasn't compiled with NO_PERL=YesPlease or
559 NO_PYTHON=YesPlease. Wrap any tests that need Perl or Python in
560 these.
561
562 - POSIXPERM
563
564 The filesystem supports POSIX style permission bits.
565
566 - BSLASHPSPEC
567
568 Backslashes in pathspec are not directory separators. This is not
569 set on Windows. See 6fd1106a for details.
570
571 - EXECKEEPSPID
572
573 The process retains the same pid across exec(2). See fb9a2bea for
574 details.
575
576 - SYMLINKS
577
578 The filesystem we're on supports symbolic links. E.g. a FAT
579 filesystem doesn't support these. See 704a3143 for details.
580
581 - SANITY
582
583 Test is not run by root user, and an attempt to write to an
584 unwritable file is expected to fail correctly.
585
586 Tips for Writing Tests
587 ----------------------
588
589 As with any programming projects, existing programs are the best
590 source of the information. However, do _not_ emulate
591 t0000-basic.sh when writing your tests. The test is special in
592 that it tries to validate the very core of GIT. For example, it
593 knows that there will be 256 subdirectories under .git/objects/,
594 and it knows that the object ID of an empty tree is a certain
595 40-byte string. This is deliberately done so in t0000-basic.sh
596 because the things the very basic core test tries to achieve is
597 to serve as a basis for people who are changing the GIT internal
598 drastically. For these people, after making certain changes,
599 not seeing failures from the basic test _is_ a failure. And
600 such drastic changes to the core GIT that even changes these
601 otherwise supposedly stable object IDs should be accompanied by
602 an update to t0000-basic.sh.
603
604 However, other tests that simply rely on basic parts of the core
605 GIT working properly should not have that level of intimate
606 knowledge of the core GIT internals. If all the test scripts
607 hardcoded the object IDs like t0000-basic.sh does, that defeats
608 the purpose of t0000-basic.sh, which is to isolate that level of
609 validation in one place. Your test also ends up needing
610 updating when such a change to the internal happens, so do _not_
611 do it and leave the low level of validation to t0000-basic.sh.
612
613 Test coverage
614 -------------
615
616 You can use the coverage tests to find code paths that are not being
617 used or properly exercised yet.
618
619 To do that, run the coverage target at the top-level (not in the t/
620 directory):
621
622 make coverage
623
624 That'll compile Git with GCC's coverage arguments, and generate a test
625 report with gcov after the tests finish. Running the coverage tests
626 can take a while, since running the tests in parallel is incompatible
627 with GCC's coverage mode.
628
629 After the tests have run you can generate a list of untested
630 functions:
631
632 make coverage-untested-functions
633
634 You can also generate a detailed per-file HTML report using the
635 Devel::Cover module. To install it do:
636
637 # On Debian or Ubuntu:
638 sudo aptitude install libdevel-cover-perl
639
640 # From the CPAN with cpanminus
641 curl -L http://cpanmin.us | perl - --sudo --self-upgrade
642 cpanm --sudo Devel::Cover
643
644 Then, at the top-level:
645
646 make cover_db_html
647
648 That'll generate a detailed cover report in the "cover_db_html"
649 directory, which you can then copy to a webserver, or inspect locally
650 in a browser.
651
652 Smoke testing
653 -------------
654
655 The Git test suite has support for smoke testing. Smoke testing is
656 when you submit the results of a test run to a central server for
657 analysis and aggregation.
658
659 Running a smoke tester is an easy and valuable way of contributing to
660 Git development, particularly if you have access to an uncommon OS on
661 obscure hardware.
662
663 After building Git you can generate a smoke report like this in the
664 "t" directory:
665
666 make clean smoke
667
668 You can also pass arguments via the environment. This should make it
669 faster:
670
671 GIT_TEST_OPTS='--root=/dev/shm' TEST_JOBS=10 make clean smoke
672
673 The "smoke" target will run the Git test suite with Perl's
674 "TAP::Harness" module, and package up the results in a .tar.gz archive
675 with "TAP::Harness::Archive". The former is included with Perl v5.10.1
676 or later, but you'll need to install the latter from the CPAN. See the
677 "Test coverage" section above for how you might do that.
678
679 Once the "smoke" target finishes you'll see a message like this:
680
681 TAP Archive created at <path to git>/t/test-results/git-smoke.tar.gz
682
683 To upload the smoke report you need to have curl(1) installed, then
684 do:
685
686 make smoke_report
687
688 To upload the report anonymously. Hopefully that'll return something
689 like "Reported #7 added.".
690
691 If you're going to be uploading reports frequently please request a
692 user account by E-Mailing gitsmoke@v.nix.is. Once you have a username
693 and password you'll be able to do:
694
695 SMOKE_USERNAME=<username> SMOKE_PASSWORD=<password> make smoke_report
696
697 You can also add an additional comment to attach to the report, and/or
698 a comma separated list of tags:
699
700 SMOKE_USERNAME=<username> SMOKE_PASSWORD=<password> \
701 SMOKE_COMMENT=<comment> SMOKE_TAGS=<tags> \
702 make smoke_report
703
704 Once the report is uploaded it'll be made available at
705 http://smoke.git.nix.is, here's an overview of Recent Smoke Reports
706 for Git:
707
708 http://smoke.git.nix.is/app/projects/smoke_reports/1
709
710 The reports will also be mirrored to GitHub every few hours:
711
712 http://github.com/gitsmoke/smoke-reports
713
714 The Smolder SQLite database is also mirrored and made available for
715 download:
716
717 http://github.com/gitsmoke/smoke-database
718
719 Note that the database includes hashed (with crypt()) user passwords
720 and E-Mail addresses. Don't use a valuable password for the smoke
721 service if you have an account, or an E-Mail address you don't want to
722 be publicly known. The user accounts are just meant to be convenient
723 labels, they're not meant to be secure.