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
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.
17 The easiest way to run tests is to say "make". This runs all
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 256 subdirectories.
23 * ok 3: git-update-index without --add should fail adding.
25 * ok 23: no diff after checkout and git-update-index --refresh.
26 * passed all 23 test(s)
27 *** t0100-environment-names.sh ***
28 * ok 1: using old names should issue warnings.
29 * ok 2: using old names but having new names should not issue warnings.
32 Or you can run each test individually from command line, like
35 $ sh ./t3001-ls-files-killed.sh
36 * ok 1: git-update-index --add to add various paths.
37 * ok 2: git-ls-files -k to show killed files.
38 * ok 3: validate git-ls-files -k output.
39 * passed all 3 test(s)
41 You can pass --verbose (or -v), --debug (or -d), and --immediate
42 (or -i) command line argument to the test, or by setting GIT_TEST_OPTS
43 appropriately before running "make".
46 This makes the test more verbose. Specifically, the
47 command being run and their output if any are also
51 This may help the person who is developing a new test.
52 It causes the command defined with test_debug to run.
55 This causes the test to immediately exit upon the first
59 This causes additional long-running tests to be run (where
60 available), for more exhaustive testing.
63 Execute all Git binaries with valgrind and exit with status
64 126 on errors (just like regular tests, this will only stop
65 the test script when running under -i). Valgrind errors
66 go to stderr, so you might want to pass the -v option, too.
68 Since it makes no sense to run the tests with --valgrind and
69 not see any output, this option implies --verbose. For
70 convenience, it also implies --tee.
73 In addition to printing the test output to the terminal,
74 write it to files named 't/test-results/$TEST_NAME.out'.
75 As the names depend on the tests' file names, it is safe to
76 run the tests with this option in parallel.
79 By default tests are run without dashed forms of
80 commands (like git-commit) in the PATH (it only uses
81 wrappers from ../bin-wrappers). Use this option to include
82 the build directory (..) in the PATH, which contains all
83 the dashed forms of commands. This option is currently
84 implied by other options like --valgrind and
87 You can also set the GIT_TEST_INSTALLED environment variable to
88 the bindir of an existing git installation to test that installation.
89 You still need to have built this git sandbox, from which various
90 test-* support programs, templates, and perl libraries are used.
91 If your installed git is incomplete, it will silently test parts of
92 your built version instead.
94 When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
95 override the location of the dashed-form subcommands (what
96 GIT_EXEC_PATH would be used for during normal operation).
97 GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
103 In some environments, certain tests have no way of succeeding
104 due to platform limitation, such as lack of 'unzip' program, or
105 filesystem that do not allow arbitrary sequence of non-NUL bytes
108 You should be able to say something like
110 $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
114 $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
116 to omit such tests. The value of the environment variable is a
117 SP separated list of patterns that tells which tests to skip,
118 and either can match the "t[0-9]{4}" part to skip the whole
119 test, or t[0-9]{4} followed by ".$number" to say which
120 particular test to skip.
122 Note that some tests in the existing test suite rely on previous
123 test item, so you cannot arbitrarily disable one and expect the
124 remainder of test to check what the test originally was intended
131 The test files are named as:
133 tNNNN-commandname-details.sh
135 where N is a decimal digit.
137 First digit tells the family:
139 0 - the absolute basics and global stuff
140 1 - the basic commands concerning database
141 2 - the basic commands concerning the working tree
142 3 - the other basic commands (e.g. ls-files)
143 4 - the diff commands
144 5 - the pull and exporting commands
145 6 - the revision tree commands (even e.g. merge-base)
146 7 - the porcelainish commands concerning the working tree
147 8 - the porcelainish commands concerning forensics
150 Second digit tells the particular command we are testing.
152 Third digit (optionally) tells the particular switch or group of switches
155 If you create files under t/ directory (i.e. here) that is not
156 the top-level test script, never name the file to match the above
157 pattern. The Makefile here considers all such files as the
158 top-level test script and tries to run all of them. A care is
159 especially needed if you are creating a common test library
160 file, similar to test-lib.sh, because such a library file may
161 not be suitable for standalone execution.
167 The test script is written as a shell script. It should start
168 with the standard "#!/bin/bash" with copyright notices, and an
169 assignment to variable 'test_description', like this:
173 # Copyright (c) 2005 Junio C Hamano
176 test_description='xxx test (option --frotz)
178 This test registers the following structure in the cache
179 and tries to run git-ls-files with option --frotz.'
185 After assigning test_description, the test script should source
186 test-lib.sh like this:
190 This test harness library does the following things:
192 - If the script is invoked with command line argument --help
193 (or -h), it shows the test_description and exits.
195 - Creates a test directory with default notmuch-config and empty mail
196 store. This directory is 't/trash directory.<test-basename>' (note
197 the space) if you must know, but I do not think you care. The path
198 to notmuch-config is exported in NOTMUCH_CONFIG environment
199 variable and mail store path is stored in MAIL_DIR variable.
201 - Defines standard test helper functions for your scripts to
202 use. These functions are designed to make all scripts behave
203 consistently when command line arguments --verbose (or -v),
204 --debug (or -d), and --immediate (or -i) is given.
210 Your script will be a sequence of tests, using helper functions
211 from the test harness library. At the end of the script, call
218 There are a handful helper functions defined in the test harness
219 library for your script to use.
221 - test_expect_success <message> <script>
223 This takes two strings as parameter, and evaluates the
224 <script>. If it yields success, test is considered
225 successful. <message> should state what it is testing.
229 test_expect_success \
230 'git-write-tree should be able to write an empty tree.' \
231 'tree=$(git-write-tree)'
233 - test_expect_failure <message> <script>
235 This is NOT the opposite of test_expect_success, but is used
236 to mark a test that demonstrates a known breakage. Unlike
237 the usual test_expect_success tests, which say "ok" on
238 success and "FAIL" on failure, this will say "FIXED" on
239 success and "still broken" on failure. Failures from these
240 tests won't cause -i (immediate) to stop.
242 - test_debug <script>
244 This takes a single argument, <script>, and evaluates it only
245 when the test script is started with --debug command line
246 argument. This is primarily meant for use during the
247 development of a new test script.
251 Your test script must have test_done at the end. Its purpose
252 is to summarize successes and failures in the test script and
253 exit with an appropriate error code.
257 Make commit and tag names consistent by setting the author and
258 committer times to defined stated. Subsequent calls will
259 advance the times by a fixed amount.
261 - test_commit <message> [<filename> [<contents>]]
263 Creates a commit with the given message, committing the given
264 file with the given contents (default for both is to reuse the
265 message string), and adds a tag (again reusing the message
266 string as name). Calls test_tick to make the SHA-1s
269 - test_merge <message> <commit-or-tag>
271 Merges the given rev using the given message. Like test_commit,
272 creates a tag and calls test_tick before committing.
274 Tips for Writing Tests
275 ----------------------
277 As with any programming projects, existing programs are the best
278 source of the information. However, do _not_ emulate
279 t0000-basic.sh when writing your tests. The test is special in
280 that it tries to validate the very core of GIT. For example, it
281 knows that there will be 256 subdirectories under .git/objects/,
282 and it knows that the object ID of an empty tree is a certain
283 40-byte string. This is deliberately done so in t0000-basic.sh
284 because the things the very basic core test tries to achieve is
285 to serve as a basis for people who are changing the GIT internal
286 drastically. For these people, after making certain changes,
287 not seeing failures from the basic test _is_ a failure. And
288 such drastic changes to the core GIT that even changes these
289 otherwise supposedly stable object IDs should be accompanied by
290 an update to t0000-basic.sh.
292 However, other tests that simply rely on basic parts of the core
293 GIT working properly should not have that level of intimate
294 knowledge of the core GIT internals. If all the test scripts
295 hardcoded the object IDs like t0000-basic.sh does, that defeats
296 the purpose of t0000-basic.sh, which is to isolate that level of
297 validation in one place. Your test also ends up needing
298 updating when such a change to the internal happens, so do _not_
299 do it and leave the low level of validation to t0000-basic.sh.