test-lib: tests skipped by GIT_SKIP_TESTS say so
[git/git.git] / t / README
CommitLineData
986aa7f1
JH
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 ***
5099b99d
ÆAB
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
986aa7f1 24 ...
5099b99d
ÆAB
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
85b0b34e 36be run with any TAP harness. Here's an example of parallel testing
5099b99d
ÆAB
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
986aa7f1 52
28d836c8
MG
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
5099b99d 59You can also run each test individually from command line, like this:
986aa7f1 60
5099b99d
ÆAB
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
986aa7f1
JH
69
70You can pass --verbose (or -v), --debug (or -d), and --immediate
4e1be63c
JS
71(or -i) command line argument to the test, or by setting GIT_TEST_OPTS
72appropriately before running "make".
986aa7f1 73
5e3b4fce 74-v::
986aa7f1
JH
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
ff09af3f
TR
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
5e3b4fce 85-d::
986aa7f1
JH
86--debug::
87 This may help the person who is developing a new test.
88 It causes the command defined with test_debug to run.
0986de94
PK
89 The "trash" directory (used to store all temporary data
90 during testing) is not deleted even if there are no
91 failed tests so that you can inspect its contents after
92 the test finished.
986aa7f1 93
5e3b4fce 94-i::
986aa7f1
JH
95--immediate::
96 This causes the test to immediately exit upon the first
13cb3bb7
SR
97 failed test. Cleanup commands requested with
98 test_when_finished are not executed if the test failed,
99 in order to keep the state for inspection by the tester
100 to diagnose the bug.
986aa7f1 101
5e3b4fce 102-l::
5e2c08c6
LW
103--long-tests::
104 This causes additional long-running tests to be run (where
105 available), for more exhaustive testing.
106
952af351
TR
107--valgrind=<tool>::
108 Execute all Git binaries under valgrind tool <tool> and exit
109 with status 126 on errors (just like regular tests, this will
110 only stop the test script when running under -i).
986aa7f1 111
3da93652
JS
112 Since it makes no sense to run the tests with --valgrind and
113 not see any output, this option implies --verbose. For
114 convenience, it also implies --tee.
115
952af351
TR
116 <tool> defaults to 'memcheck', just like valgrind itself.
117 Other particularly useful choices include 'helgrind' and
118 'drd', but you may use any tool recognized by your valgrind
119 installation.
120
95d9d5ec
TR
121 As a special case, <tool> can be 'memcheck-fast', which uses
122 memcheck but disables --track-origins. Use this if you are
123 running tests in bulk, to see if there are _any_ memory
124 issues.
125
952af351 126 Note that memcheck is run with the option --leak-check=no,
9aec68d3
CMN
127 as the git process is short-lived and some errors are not
128 interesting. In order to run a single command under the same
129 conditions manually, you should set GIT_VALGRIND to point to
130 the 't/valgrind/' directory and use the commands under
131 't/valgrind/bin/'.
132
5dfc368f
TR
133--valgrind-only=<pattern>::
134 Like --valgrind, but the effect is limited to tests with
135 numbers matching <pattern>. The number matched against is
136 simply the running count of the test within the file.
137
44138559
JS
138--tee::
139 In addition to printing the test output to the terminal,
140 write it to files named 't/test-results/$TEST_NAME.out'.
141 As the names depend on the tests' file names, it is safe to
142 run the tests with this option in parallel.
143
e4597aae
MO
144--with-dashes::
145 By default tests are run without dashed forms of
146 commands (like git-commit) in the PATH (it only uses
147 wrappers from ../bin-wrappers). Use this option to include
148 the build directory (..) in the PATH, which contains all
149 the dashed forms of commands. This option is currently
150 implied by other options like --valgrind and
151 GIT_TEST_INSTALLED.
152
0d4dbcd3
TR
153--root=<directory>::
154 Create "trash" directories used to store all temporary data during
155 testing under <directory>, instead of the t/ directory.
156 Using this option with a RAM-based filesystem (such as tmpfs)
157 can massively speed up the test suite.
158
e160da7f
MO
159You can also set the GIT_TEST_INSTALLED environment variable to
160the bindir of an existing git installation to test that installation.
161You still need to have built this git sandbox, from which various
162test-* support programs, templates, and perl libraries are used.
163If your installed git is incomplete, it will silently test parts of
164your built version instead.
165
166When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
167override the location of the dashed-form subcommands (what
168GIT_EXEC_PATH would be used for during normal operation).
169GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
170
171
fbd458a3
JN
172Skipping Tests
173--------------
174
175In some environments, certain tests have no way of succeeding
176due to platform limitation, such as lack of 'unzip' program, or
177filesystem that do not allow arbitrary sequence of non-NUL bytes
178as pathnames.
179
180You should be able to say something like
181
182 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
183
184and even:
185
186 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
187
188to omit such tests. The value of the environment variable is a
189SP separated list of patterns that tells which tests to skip,
190and either can match the "t[0-9]{4}" part to skip the whole
191test, or t[0-9]{4} followed by ".$number" to say which
192particular test to skip.
193
194Note that some tests in the existing test suite rely on previous
195test item, so you cannot arbitrarily disable one and expect the
196remainder of test to check what the test originally was intended
197to check.
198
199
f50c9f76
PB
200Naming Tests
201------------
202
203The test files are named as:
204
205 tNNNN-commandname-details.sh
206
207where N is a decimal digit.
208
209First digit tells the family:
210
211 0 - the absolute basics and global stuff
212 1 - the basic commands concerning database
213 2 - the basic commands concerning the working tree
214 3 - the other basic commands (e.g. ls-files)
215 4 - the diff commands
216 5 - the pull and exporting commands
217 6 - the revision tree commands (even e.g. merge-base)
8f4a9b62 218 7 - the porcelainish commands concerning the working tree
8757749e
JN
219 8 - the porcelainish commands concerning forensics
220 9 - the git tools
f50c9f76
PB
221
222Second digit tells the particular command we are testing.
223
224Third digit (optionally) tells the particular switch or group of switches
225we are testing.
226
77656600
JH
227If you create files under t/ directory (i.e. here) that is not
228the top-level test script, never name the file to match the above
229pattern. The Makefile here considers all such files as the
63d32945 230top-level test script and tries to run all of them. Care is
77656600
JH
231especially needed if you are creating a common test library
232file, similar to test-lib.sh, because such a library file may
233not be suitable for standalone execution.
234
f50c9f76 235
986aa7f1
JH
236Writing Tests
237-------------
238
239The test script is written as a shell script. It should start
240with the standard "#!/bin/sh" with copyright notices, and an
241assignment to variable 'test_description', like this:
242
243 #!/bin/sh
244 #
245 # Copyright (c) 2005 Junio C Hamano
246 #
247
14cd1ff3 248 test_description='xxx test (option --frotz)
986aa7f1
JH
249
250 This test registers the following structure in the cache
251 and tries to run git-ls-files with option --frotz.'
252
f50c9f76 253
986aa7f1
JH
254Source 'test-lib.sh'
255--------------------
256
257After assigning test_description, the test script should source
258test-lib.sh like this:
259
260 . ./test-lib.sh
261
262This test harness library does the following things:
263
264 - If the script is invoked with command line argument --help
265 (or -h), it shows the test_description and exits.
266
e1ca1c9d
ÆAB
267 - Creates an empty test directory with an empty .git/objects database
268 and chdir(2) into it. This directory is 't/trash
269 directory.$test_name_without_dotsh', with t/ subject to change by
270 the --root option documented above.
986aa7f1
JH
271
272 - Defines standard test helper functions for your scripts to
273 use. These functions are designed to make all scripts behave
274 consistently when command line arguments --verbose (or -v),
275 --debug (or -d), and --immediate (or -i) is given.
276
20873f45
ÆAB
277Do's, don'ts & things to keep in mind
278-------------------------------------
279
6fd45295 280Here are a few examples of things you probably should and shouldn't do
20873f45
ÆAB
281when writing tests.
282
283Do:
284
6fd45295 285 - Put all code inside test_expect_success and other assertions.
20873f45
ÆAB
286
287 Even code that isn't a test per se, but merely some setup code
6fd45295 288 should be inside a test assertion.
20873f45
ÆAB
289
290 - Chain your test assertions
291
292 Write test code like this:
293
294 git merge foo &&
295 git push bar &&
296 test ...
297
298 Instead of:
299
300 git merge hla
301 git push gh
302 test ...
303
304 That way all of the commands in your tests will succeed or fail. If
00648ba0
EN
305 you must ignore the return value of something, consider using a
306 helper function (e.g. use sane_unset instead of unset, in order
307 to avoid unportable return value for unsetting a variable that was
308 already unset), or prepending the command with test_might_fail or
309 test_must_fail.
20873f45 310
0c357544
ÆAB
311 - Check the test coverage for your tests. See the "Test coverage"
312 below.
313
63d32945
MW
314 Don't blindly follow test coverage metrics; if a new function you added
315 doesn't have any coverage, then you're probably doing something wrong,
e8b55f5c
ÆAB
316 but having 100% coverage doesn't necessarily mean that you tested
317 everything.
318
319 Tests that are likely to smoke out future regressions are better
320 than tests that just inflate the coverage metrics.
321
95b104c8
JS
322 - When a test checks for an absolute path that a git command generated,
323 construct the expected value using $(pwd) rather than $PWD,
324 $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
325 Windows, where the shell (MSYS bash) mangles absolute path names.
326 For details, see the commit message of 4114156ae9.
327
20873f45
ÆAB
328Don't:
329
330 - exit() within a <script> part.
331
332 The harness will catch this as a programming error of the test.
333 Use test_done instead if you need to stop the tests early (see
334 "Skipping tests" below).
335
ad78585e
JH
336 - use '! git cmd' when you want to make sure the git command exits
337 with failure in a controlled way by calling "die()". Instead,
338 use 'test_must_fail git cmd'. This will signal a failure if git
339 dies in an unexpected way (e.g. segfault).
340
f445500e
JH
341 On the other hand, don't use test_must_fail for running regular
342 platform commands; just use '! cmd'.
343
ad78585e
JH
344 - use perl without spelling it as "$PERL_PATH". This is to help our
345 friends on Windows where the platform Perl often adds CR before
346 the end of line, and they bundle Git with a version of Perl that
a0e0ec9f
JK
347 does not do so, whose path is specified with $PERL_PATH. Note that we
348 provide a "perl" function which uses $PERL_PATH under the hood, so
349 you do not need to worry when simply running perl in the test scripts
350 (but you do, for example, on a shebang line or in a sub script
351 created via "write_script").
ad78585e
JH
352
353 - use sh without spelling it as "$SHELL_PATH", when the script can
354 be misinterpreted by broken platform shell (e.g. Solaris).
355
356 - chdir around in tests. It is not sufficient to chdir to
357 somewhere and then chdir back to the original location later in
358 the test, as any intermediate step can fail and abort the test,
359 causing the next test to start in an unexpected directory. Do so
360 inside a subshell if necessary.
361
20873f45
ÆAB
362 - Break the TAP output
363
6fd45295
JH
364 The raw output from your test may be interpreted by a TAP harness. TAP
365 harnesses will ignore everything they don't know about, but don't step
366 on their toes in these areas:
20873f45
ÆAB
367
368 - Don't print lines like "$x..$y" where $x and $y are integers.
369
370 - Don't print lines that begin with "ok" or "not ok".
371
6fd45295 372 TAP harnesses expect a line that begins with either "ok" and "not
20873f45
ÆAB
373 ok" to signal a test passed or failed (and our harness already
374 produces such lines), so your script shouldn't emit such lines to
375 their output.
376
377 You can glean some further possible issues from the TAP grammar
378 (see http://search.cpan.org/perldoc?TAP::Parser::Grammar#TAP_Grammar)
379 but the best indication is to just run the tests with prove(1),
380 it'll complain if anything is amiss.
381
382Keep in mind:
383
6fd45295 384 - Inside <script> part, the standard output and standard error
20873f45
ÆAB
385 streams are discarded, and the test harness only reports "ok" or
386 "not ok" to the end user running the tests. Under --verbose, they
387 are shown to help debugging the tests.
388
389
b5500d16
ÆAB
390Skipping tests
391--------------
392
681186ae 393If you need to skip tests you should do so by using the three-arg form
99d9050d
ÆAB
394of the test_* functions (see the "Test harness library" section
395below), e.g.:
396
ad78585e 397 test_expect_success PERL 'I need Perl' '
a0e0ec9f 398 perl -e "hlagh() if unf_unf()"
ad78585e 399 '
99d9050d
ÆAB
400
401The advantage of skipping tests like this is that platforms that don't
402have the PERL and other optional dependencies get an indication of how
403many tests they're missing.
404
405If the test code is too hairy for that (i.e. does a lot of setup work
406outside test assertions) you can also skip all remaining tests by
407setting skip_all and immediately call test_done:
b5500d16
ÆAB
408
409 if ! test_have_prereq PERL
410 then
411 skip_all='skipping perl interface tests, perl not available'
412 test_done
413 fi
14cd1ff3 414
99d9050d
ÆAB
415The string you give to skip_all will be used as an explanation for why
416the test was skipped.
417
986aa7f1
JH
418End with test_done
419------------------
420
421Your script will be a sequence of tests, using helper functions
422from the test harness library. At the end of the script, call
423'test_done'.
424
425
426Test harness library
427--------------------
428
429There are a handful helper functions defined in the test harness
430library for your script to use.
431
9a897893 432 - test_expect_success [<prereq>] <message> <script>
986aa7f1 433
72942a61 434 Usually takes two strings as parameters, and evaluates the
986aa7f1
JH
435 <script>. If it yields success, test is considered
436 successful. <message> should state what it is testing.
437
438 Example:
439
440 test_expect_success \
441 'git-write-tree should be able to write an empty tree.' \
442 'tree=$(git-write-tree)'
443
9a897893 444 If you supply three parameters the first will be taken to be a
72942a61 445 prerequisite; see the test_set_prereq and test_have_prereq
9a897893
ÆAB
446 documentation below:
447
448 test_expect_success TTY 'git --paginate rev-list uses a pager' \
449 ' ... '
450
93a57246
ÆAB
451 You can also supply a comma-separated list of prerequisites, in the
452 rare case where your test depends on more than one:
453
454 test_expect_success PERL,PYTHON 'yo dawg' \
455 ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" '
456
9a897893 457 - test_expect_failure [<prereq>] <message> <script>
986aa7f1 458
41ac414e
JH
459 This is NOT the opposite of test_expect_success, but is used
460 to mark a test that demonstrates a known breakage. Unlike
461 the usual test_expect_success tests, which say "ok" on
462 success and "FAIL" on failure, this will say "FIXED" on
463 success and "still broken" on failure. Failures from these
464 tests won't cause -i (immediate) to stop.
986aa7f1 465
9a897893
ÆAB
466 Like test_expect_success this function can optionally use a three
467 argument invocation with a prerequisite as the first argument.
468
986aa7f1
JH
469 - test_debug <script>
470
471 This takes a single argument, <script>, and evaluates it only
472 when the test script is started with --debug command line
473 argument. This is primarily meant for use during the
474 development of a new test script.
475
476 - test_done
477
478 Your test script must have test_done at the end. Its purpose
479 is to summarize successes and failures in the test script and
480 exit with an appropriate error code.
481
00884968
JS
482 - test_tick
483
484 Make commit and tag names consistent by setting the author and
63d32945 485 committer times to defined state. Subsequent calls will
00884968
JS
486 advance the times by a fixed amount.
487
488 - test_commit <message> [<filename> [<contents>]]
489
490 Creates a commit with the given message, committing the given
491 file with the given contents (default for both is to reuse the
492 message string), and adds a tag (again reusing the message
493 string as name). Calls test_tick to make the SHA-1s
494 reproducible.
495
496 - test_merge <message> <commit-or-tag>
497
498 Merges the given rev using the given message. Like test_commit,
499 creates a tag and calls test_tick before committing.
986aa7f1 500
72942a61 501 - test_set_prereq <prereq>
9a897893
ÆAB
502
503 Set a test prerequisite to be used later with test_have_prereq. The
be53deef
ÆAB
504 test-lib will set some prerequisites for you, see the
505 "Prerequisites" section below for a full list of these.
506
507 Others you can set yourself and use later with either
508 test_have_prereq directly, or the three argument invocation of
509 test_expect_success and test_expect_failure.
9a897893 510
72942a61 511 - test_have_prereq <prereq>
9a897893
ÆAB
512
513 Check if we have a prerequisite previously set with
514 test_set_prereq. The most common use of this directly is to skip
515 all the tests if we don't have some essential prerequisite:
516
517 if ! test_have_prereq PERL
518 then
519 skip_all='skipping perl interface tests, perl not available'
520 test_done
521 fi
522
2fac6a4b
ÆAB
523 - test_external [<prereq>] <message> <external> <script>
524
525 Execute a <script> with an <external> interpreter (like perl). This
526 was added for tests like t9700-perl-git.sh which do most of their
527 work in an external test script.
528
529 test_external \
530 'GitwebCache::*FileCache*' \
a0e0ec9f 531 perl "$TEST_DIRECTORY"/t9503/test_cache_interface.pl
2fac6a4b
ÆAB
532
533 If the test is outputting its own TAP you should set the
534 test_external_has_tap variable somewhere before calling the first
535 test_external* function. See t9700-perl-git.sh for an example.
536
537 # The external test will outputs its own plan
538 test_external_has_tap=1
539
540 - test_external_without_stderr [<prereq>] <message> <external> <script>
541
542 Like test_external but fail if there's any output on stderr,
543 instead of checking the exit code.
544
545 test_external_without_stderr \
546 'Perl API' \
a0e0ec9f 547 perl "$TEST_DIRECTORY"/t9700/test.pl
2fac6a4b 548
892e6f7e
ÆAB
549 - test_expect_code <exit-code> <command>
550
551 Run a command and ensure that it exits with the given exit code.
552 For example:
553
554 test_expect_success 'Merge with d/f conflicts' '
555 test_expect_code 1 git merge "merge msg" B master
556 '
557
c9667456
JN
558 - test_must_fail <git-command>
559
560 Run a git command and ensure it fails in a controlled way. Use
971ecbd1
BC
561 this instead of "! <git-command>". When git-command dies due to a
562 segfault, test_must_fail diagnoses it as an error; "! <git-command>"
563 treats it as just another expected failure, which would let such a
564 bug go unnoticed.
c9667456
JN
565
566 - test_might_fail <git-command>
567
568 Similar to test_must_fail, but tolerate success, too. Use this
569 instead of "<git-command> || :" to catch failures due to segv.
570
571 - test_cmp <expected> <actual>
572
573 Check whether the content of the <actual> file matches the
574 <expected> file. This behaves like "cmp" but produces more
575 helpful output when the test is run with "-v" option.
576
fb3340a6
JN
577 - test_line_count (= | -lt | -ge | ...) <length> <file>
578
579 Check whether a file has the length it is expected to.
580
72942a61
ML
581 - test_path_is_file <path> [<diagnosis>]
582 test_path_is_dir <path> [<diagnosis>]
2caf20c5
MM
583 test_path_is_missing <path> [<diagnosis>]
584
72942a61
ML
585 Check if the named path is a file, if the named path is a
586 directory, or if the named path does not exist, respectively,
587 and fail otherwise, showing the <diagnosis> text.
2caf20c5 588
c9667456
JN
589 - test_when_finished <script>
590
591 Prepend <script> to a list of commands to run to clean up
592 at the end of the current test. If some clean-up command
593 fails, the test will not pass.
594
595 Example:
596
597 test_expect_success 'branch pointing to non-commit' '
598 git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
599 test_when_finished "git update-ref -d refs/heads/invalid" &&
600 ...
601 '
602
c4d2539a
JL
603 - test_pause
604
605 This command is useful for writing and debugging tests and must be
606 removed before submitting. It halts the execution of the test and
607 spawns a shell in the trash directory. Exit the shell to continue
608 the test. Example:
609
610 test_expect_success 'test' '
611 git do-something >actual &&
612 test_pause &&
613 test_cmp expected actual
614 '
615
9ce415d9
JS
616 - test_ln_s_add <path1> <path2>
617
618 This function helps systems whose filesystem does not support symbolic
619 links. Use it to add a symbolic link entry to the index when it is not
620 important that the file system entry is a symbolic link, i.e., instead
621 of the sequence
622
623 ln -s foo bar &&
624 git add bar
625
626 Sometimes it is possible to split a test in a part that does not need
627 the symbolic link in the file system and a part that does; then only
628 the latter part need be protected by a SYMLINKS prerequisite (see below).
629
be53deef
ÆAB
630Prerequisites
631-------------
632
633These are the prerequisites that the test library predefines with
634test_have_prereq.
635
636See the prereq argument to the test_* functions in the "Test harness
637library" section above and the "test_have_prereq" function for how to
638use these, and "test_set_prereq" for how to define your own.
639
f8fc0ee3 640 - PYTHON
be53deef 641
f8fc0ee3
JN
642 Git wasn't compiled with NO_PYTHON=YesPlease. Wrap any tests that
643 need Python with this.
644
645 - PERL
646
647 Git wasn't compiled with NO_PERL=YesPlease.
648
649 Even without the PERL prerequisite, tests can assume there is a
650 usable perl interpreter at $PERL_PATH, though it need not be
651 particularly modern.
be53deef
ÆAB
652
653 - POSIXPERM
654
655 The filesystem supports POSIX style permission bits.
656
657 - BSLASHPSPEC
658
659 Backslashes in pathspec are not directory separators. This is not
660 set on Windows. See 6fd1106a for details.
661
662 - EXECKEEPSPID
663
664 The process retains the same pid across exec(2). See fb9a2bea for
665 details.
666
20073274
AS
667 - PIPE
668
669 The filesystem we're on supports creation of FIFOs (named pipes)
670 via mkfifo(1).
671
be53deef
ÆAB
672 - SYMLINKS
673
674 The filesystem we're on supports symbolic links. E.g. a FAT
675 filesystem doesn't support these. See 704a3143 for details.
2fac6a4b 676
c91cfd19
ÆAB
677 - SANITY
678
679 Test is not run by root user, and an attempt to write to an
680 unwritable file is expected to fail correctly.
2fac6a4b 681
8f852ce6
MK
682 - LIBPCRE
683
684 Git was compiled with USE_LIBPCRE=YesPlease. Wrap any tests
685 that use git-grep --perl-regexp or git-grep -P in these.
686
ac39aa61
MG
687 - CASE_INSENSITIVE_FS
688
689 Test is run on a case insensitive file system.
690
5b0b5dd4
MG
691 - UTF8_NFD_TO_NFC
692
693 Test is run on a filesystem which converts decomposed utf-8 (nfd)
694 to precomposed utf-8 (nfc).
695
986aa7f1
JH
696Tips for Writing Tests
697----------------------
698
699As with any programming projects, existing programs are the best
700source of the information. However, do _not_ emulate
701t0000-basic.sh when writing your tests. The test is special in
702that it tries to validate the very core of GIT. For example, it
703knows that there will be 256 subdirectories under .git/objects/,
704and it knows that the object ID of an empty tree is a certain
70540-byte string. This is deliberately done so in t0000-basic.sh
706because the things the very basic core test tries to achieve is
707to serve as a basis for people who are changing the GIT internal
708drastically. For these people, after making certain changes,
709not seeing failures from the basic test _is_ a failure. And
710such drastic changes to the core GIT that even changes these
711otherwise supposedly stable object IDs should be accompanied by
712an update to t0000-basic.sh.
713
714However, other tests that simply rely on basic parts of the core
715GIT working properly should not have that level of intimate
716knowledge of the core GIT internals. If all the test scripts
717hardcoded the object IDs like t0000-basic.sh does, that defeats
718the purpose of t0000-basic.sh, which is to isolate that level of
719validation in one place. Your test also ends up needing
720updating when such a change to the internal happens, so do _not_
721do it and leave the low level of validation to t0000-basic.sh.
d15e9ebc 722
0c357544
ÆAB
723Test coverage
724-------------
725
726You can use the coverage tests to find code paths that are not being
727used or properly exercised yet.
728
729To do that, run the coverage target at the top-level (not in the t/
730directory):
731
732 make coverage
733
734That'll compile Git with GCC's coverage arguments, and generate a test
735report with gcov after the tests finish. Running the coverage tests
736can take a while, since running the tests in parallel is incompatible
737with GCC's coverage mode.
738
739After the tests have run you can generate a list of untested
740functions:
741
742 make coverage-untested-functions
743
744You can also generate a detailed per-file HTML report using the
745Devel::Cover module. To install it do:
746
747 # On Debian or Ubuntu:
748 sudo aptitude install libdevel-cover-perl
749
750 # From the CPAN with cpanminus
751 curl -L http://cpanmin.us | perl - --sudo --self-upgrade
752 cpanm --sudo Devel::Cover
753
754Then, at the top-level:
755
756 make cover_db_html
757
758That'll generate a detailed cover report in the "cover_db_html"
759directory, which you can then copy to a webserver, or inspect locally
760in a browser.