t / READMEon commit Merge branch 'maint-1.8.2' into maint (f2b4626)
   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--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        The "trash" directory (used to store all temporary data
  83        during testing) is not deleted even if there are no
  84        failed tests so that you can inspect its contents after
  85        the test finished.
  86
  87--immediate::
  88        This causes the test to immediately exit upon the first
  89        failed test. Cleanup commands requested with
  90        test_when_finished are not executed if the test failed,
  91        in order to keep the state for inspection by the tester
  92        to diagnose the bug.
  93
  94--long-tests::
  95        This causes additional long-running tests to be run (where
  96        available), for more exhaustive testing.
  97
  98--valgrind=<tool>::
  99        Execute all Git binaries under valgrind tool <tool> and exit
 100        with status 126 on errors (just like regular tests, this will
 101        only stop the test script when running under -i).
 102
 103        Since it makes no sense to run the tests with --valgrind and
 104        not see any output, this option implies --verbose.  For
 105        convenience, it also implies --tee.
 106
 107        <tool> defaults to 'memcheck', just like valgrind itself.
 108        Other particularly useful choices include 'helgrind' and
 109        'drd', but you may use any tool recognized by your valgrind
 110        installation.
 111
 112        As a special case, <tool> can be 'memcheck-fast', which uses
 113        memcheck but disables --track-origins.  Use this if you are
 114        running tests in bulk, to see if there are _any_ memory
 115        issues.
 116
 117        Note that memcheck is run with the option --leak-check=no,
 118        as the git process is short-lived and some errors are not
 119        interesting. In order to run a single command under the same
 120        conditions manually, you should set GIT_VALGRIND to point to
 121        the 't/valgrind/' directory and use the commands under
 122        't/valgrind/bin/'.
 123
 124--tee::
 125        In addition to printing the test output to the terminal,
 126        write it to files named 't/test-results/$TEST_NAME.out'.
 127        As the names depend on the tests' file names, it is safe to
 128        run the tests with this option in parallel.
 129
 130--with-dashes::
 131        By default tests are run without dashed forms of
 132        commands (like git-commit) in the PATH (it only uses
 133        wrappers from ../bin-wrappers).  Use this option to include
 134        the build directory (..) in the PATH, which contains all
 135        the dashed forms of commands.  This option is currently
 136        implied by other options like --valgrind and
 137        GIT_TEST_INSTALLED.
 138
 139--root=<directory>::
 140        Create "trash" directories used to store all temporary data during
 141        testing under <directory>, instead of the t/ directory.
 142        Using this option with a RAM-based filesystem (such as tmpfs)
 143        can massively speed up the test suite.
 144
 145You can also set the GIT_TEST_INSTALLED environment variable to
 146the bindir of an existing git installation to test that installation.
 147You still need to have built this git sandbox, from which various
 148test-* support programs, templates, and perl libraries are used.
 149If your installed git is incomplete, it will silently test parts of
 150your built version instead.
 151
 152When using GIT_TEST_INSTALLED, you can also set GIT_TEST_EXEC_PATH to
 153override the location of the dashed-form subcommands (what
 154GIT_EXEC_PATH would be used for during normal operation).
 155GIT_TEST_EXEC_PATH defaults to `$GIT_TEST_INSTALLED/git --exec-path`.
 156
 157
 158Skipping Tests
 159--------------
 160
 161In some environments, certain tests have no way of succeeding
 162due to platform limitation, such as lack of 'unzip' program, or
 163filesystem that do not allow arbitrary sequence of non-NUL bytes
 164as pathnames.
 165
 166You should be able to say something like
 167
 168    $ GIT_SKIP_TESTS=t9200.8 sh ./t9200-git-cvsexport-commit.sh
 169
 170and even:
 171
 172    $ GIT_SKIP_TESTS='t[0-4]??? t91?? t9200.8' make
 173
 174to omit such tests.  The value of the environment variable is a
 175SP separated list of patterns that tells which tests to skip,
 176and either can match the "t[0-9]{4}" part to skip the whole
 177test, or t[0-9]{4} followed by ".$number" to say which
 178particular test to skip.
 179
 180Note that some tests in the existing test suite rely on previous
 181test item, so you cannot arbitrarily disable one and expect the
 182remainder of test to check what the test originally was intended
 183to check.
 184
 185
 186Naming Tests
 187------------
 188
 189The test files are named as:
 190
 191        tNNNN-commandname-details.sh
 192
 193where N is a decimal digit.
 194
 195First digit tells the family:
 196
 197        0 - the absolute basics and global stuff
 198        1 - the basic commands concerning database
 199        2 - the basic commands concerning the working tree
 200        3 - the other basic commands (e.g. ls-files)
 201        4 - the diff commands
 202        5 - the pull and exporting commands
 203        6 - the revision tree commands (even e.g. merge-base)
 204        7 - the porcelainish commands concerning the working tree
 205        8 - the porcelainish commands concerning forensics
 206        9 - the git tools
 207
 208Second digit tells the particular command we are testing.
 209
 210Third digit (optionally) tells the particular switch or group of switches
 211we are testing.
 212
 213If you create files under t/ directory (i.e. here) that is not
 214the top-level test script, never name the file to match the above
 215pattern.  The Makefile here considers all such files as the
 216top-level test script and tries to run all of them.  Care is
 217especially needed if you are creating a common test library
 218file, similar to test-lib.sh, because such a library file may
 219not be suitable for standalone execution.
 220
 221
 222Writing Tests
 223-------------
 224
 225The test script is written as a shell script.  It should start
 226with the standard "#!/bin/sh" with copyright notices, and an
 227assignment to variable 'test_description', like this:
 228
 229        #!/bin/sh
 230        #
 231        # Copyright (c) 2005 Junio C Hamano
 232        #
 233
 234        test_description='xxx test (option --frotz)
 235
 236        This test registers the following structure in the cache
 237        and tries to run git-ls-files with option --frotz.'
 238
 239
 240Source 'test-lib.sh'
 241--------------------
 242
 243After assigning test_description, the test script should source
 244test-lib.sh like this:
 245
 246        . ./test-lib.sh
 247
 248This test harness library does the following things:
 249
 250 - If the script is invoked with command line argument --help
 251   (or -h), it shows the test_description and exits.
 252
 253 - Creates an empty test directory with an empty .git/objects database
 254   and chdir(2) into it.  This directory is 't/trash
 255   directory.$test_name_without_dotsh', with t/ subject to change by
 256   the --root option documented above.
 257
 258 - Defines standard test helper functions for your scripts to
 259   use.  These functions are designed to make all scripts behave
 260   consistently when command line arguments --verbose (or -v),
 261   --debug (or -d), and --immediate (or -i) is given.
 262
 263Do's, don'ts & things to keep in mind
 264-------------------------------------
 265
 266Here are a few examples of things you probably should and shouldn't do
 267when writing tests.
 268
 269Do:
 270
 271 - Put all code inside test_expect_success and other assertions.
 272
 273   Even code that isn't a test per se, but merely some setup code
 274   should be inside a test assertion.
 275
 276 - Chain your test assertions
 277
 278   Write test code like this:
 279
 280        git merge foo &&
 281        git push bar &&
 282        test ...
 283
 284   Instead of:
 285
 286        git merge hla
 287        git push gh
 288        test ...
 289
 290   That way all of the commands in your tests will succeed or fail. If
 291   you must ignore the return value of something, consider using a
 292   helper function (e.g. use sane_unset instead of unset, in order
 293   to avoid unportable return value for unsetting a variable that was
 294   already unset), or prepending the command with test_might_fail or
 295   test_must_fail.
 296
 297 - Check the test coverage for your tests. See the "Test coverage"
 298   below.
 299
 300   Don't blindly follow test coverage metrics; if a new function you added
 301   doesn't have any coverage, then you're probably doing something wrong,
 302   but having 100% coverage doesn't necessarily mean that you tested
 303   everything.
 304
 305   Tests that are likely to smoke out future regressions are better
 306   than tests that just inflate the coverage metrics.
 307
 308 - When a test checks for an absolute path that a git command generated,
 309   construct the expected value using $(pwd) rather than $PWD,
 310   $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
 311   Windows, where the shell (MSYS bash) mangles absolute path names.
 312   For details, see the commit message of 4114156ae9.
 313
 314Don't:
 315
 316 - exit() within a <script> part.
 317
 318   The harness will catch this as a programming error of the test.
 319   Use test_done instead if you need to stop the tests early (see
 320   "Skipping tests" below).
 321
 322 - use '! git cmd' when you want to make sure the git command exits
 323   with failure in a controlled way by calling "die()".  Instead,
 324   use 'test_must_fail git cmd'.  This will signal a failure if git
 325   dies in an unexpected way (e.g. segfault).
 326
 327   On the other hand, don't use test_must_fail for running regular
 328   platform commands; just use '! cmd'.
 329
 330 - use perl without spelling it as "$PERL_PATH". This is to help our
 331   friends on Windows where the platform Perl often adds CR before
 332   the end of line, and they bundle Git with a version of Perl that
 333   does not do so, whose path is specified with $PERL_PATH.
 334
 335 - use sh without spelling it as "$SHELL_PATH", when the script can
 336   be misinterpreted by broken platform shell (e.g. Solaris).
 337
 338 - chdir around in tests.  It is not sufficient to chdir to
 339   somewhere and then chdir back to the original location later in
 340   the test, as any intermediate step can fail and abort the test,
 341   causing the next test to start in an unexpected directory.  Do so
 342   inside a subshell if necessary.
 343
 344 - Break the TAP output
 345
 346   The raw output from your test may be interpreted by a TAP harness. TAP
 347   harnesses will ignore everything they don't know about, but don't step
 348   on their toes in these areas:
 349
 350   - Don't print lines like "$x..$y" where $x and $y are integers.
 351
 352   - Don't print lines that begin with "ok" or "not ok".
 353
 354   TAP harnesses expect a line that begins with either "ok" and "not
 355   ok" to signal a test passed or failed (and our harness already
 356   produces such lines), so your script shouldn't emit such lines to
 357   their output.
 358
 359   You can glean some further possible issues from the TAP grammar
 360   (see http://search.cpan.org/perldoc?TAP::Parser::Grammar#TAP_Grammar)
 361   but the best indication is to just run the tests with prove(1),
 362   it'll complain if anything is amiss.
 363
 364Keep in mind:
 365
 366 - Inside <script> part, the standard output and standard error
 367   streams are discarded, and the test harness only reports "ok" or
 368   "not ok" to the end user running the tests. Under --verbose, they
 369   are shown to help debugging the tests.
 370
 371
 372Skipping tests
 373--------------
 374
 375If you need to skip tests you should do so by using the three-arg form
 376of the test_* functions (see the "Test harness library" section
 377below), e.g.:
 378
 379    test_expect_success PERL 'I need Perl' '
 380        "$PERL_PATH" -e "hlagh() if unf_unf()"
 381    '
 382
 383The advantage of skipping tests like this is that platforms that don't
 384have the PERL and other optional dependencies get an indication of how
 385many tests they're missing.
 386
 387If the test code is too hairy for that (i.e. does a lot of setup work
 388outside test assertions) you can also skip all remaining tests by
 389setting skip_all and immediately call test_done:
 390
 391        if ! test_have_prereq PERL
 392        then
 393            skip_all='skipping perl interface tests, perl not available'
 394            test_done
 395        fi
 396
 397The string you give to skip_all will be used as an explanation for why
 398the test was skipped.
 399
 400End with test_done
 401------------------
 402
 403Your script will be a sequence of tests, using helper functions
 404from the test harness library.  At the end of the script, call
 405'test_done'.
 406
 407
 408Test harness library
 409--------------------
 410
 411There are a handful helper functions defined in the test harness
 412library for your script to use.
 413
 414 - test_expect_success [<prereq>] <message> <script>
 415
 416   Usually takes two strings as parameters, and evaluates the
 417   <script>.  If it yields success, test is considered
 418   successful.  <message> should state what it is testing.
 419
 420   Example:
 421
 422        test_expect_success \
 423            'git-write-tree should be able to write an empty tree.' \
 424            'tree=$(git-write-tree)'
 425
 426   If you supply three parameters the first will be taken to be a
 427   prerequisite; see the test_set_prereq and test_have_prereq
 428   documentation below:
 429
 430        test_expect_success TTY 'git --paginate rev-list uses a pager' \
 431            ' ... '
 432
 433   You can also supply a comma-separated list of prerequisites, in the
 434   rare case where your test depends on more than one:
 435
 436        test_expect_success PERL,PYTHON 'yo dawg' \
 437            ' test $(perl -E 'print eval "1 +" . qx[python -c "print 2"]') == "4" '
 438
 439 - test_expect_failure [<prereq>] <message> <script>
 440
 441   This is NOT the opposite of test_expect_success, but is used
 442   to mark a test that demonstrates a known breakage.  Unlike
 443   the usual test_expect_success tests, which say "ok" on
 444   success and "FAIL" on failure, this will say "FIXED" on
 445   success and "still broken" on failure.  Failures from these
 446   tests won't cause -i (immediate) to stop.
 447
 448   Like test_expect_success this function can optionally use a three
 449   argument invocation with a prerequisite as the first argument.
 450
 451 - test_debug <script>
 452
 453   This takes a single argument, <script>, and evaluates it only
 454   when the test script is started with --debug command line
 455   argument.  This is primarily meant for use during the
 456   development of a new test script.
 457
 458 - test_done
 459
 460   Your test script must have test_done at the end.  Its purpose
 461   is to summarize successes and failures in the test script and
 462   exit with an appropriate error code.
 463
 464 - test_tick
 465
 466   Make commit and tag names consistent by setting the author and
 467   committer times to defined state.  Subsequent calls will
 468   advance the times by a fixed amount.
 469
 470 - test_commit <message> [<filename> [<contents>]]
 471
 472   Creates a commit with the given message, committing the given
 473   file with the given contents (default for both is to reuse the
 474   message string), and adds a tag (again reusing the message
 475   string as name).  Calls test_tick to make the SHA-1s
 476   reproducible.
 477
 478 - test_merge <message> <commit-or-tag>
 479
 480   Merges the given rev using the given message.  Like test_commit,
 481   creates a tag and calls test_tick before committing.
 482
 483 - test_set_prereq <prereq>
 484
 485   Set a test prerequisite to be used later with test_have_prereq. The
 486   test-lib will set some prerequisites for you, see the
 487   "Prerequisites" section below for a full list of these.
 488
 489   Others you can set yourself and use later with either
 490   test_have_prereq directly, or the three argument invocation of
 491   test_expect_success and test_expect_failure.
 492
 493 - test_have_prereq <prereq>
 494
 495   Check if we have a prerequisite previously set with
 496   test_set_prereq. The most common use of this directly is to skip
 497   all the tests if we don't have some essential prerequisite:
 498
 499        if ! test_have_prereq PERL
 500        then
 501            skip_all='skipping perl interface tests, perl not available'
 502            test_done
 503        fi
 504
 505 - test_external [<prereq>] <message> <external> <script>
 506
 507   Execute a <script> with an <external> interpreter (like perl). This
 508   was added for tests like t9700-perl-git.sh which do most of their
 509   work in an external test script.
 510
 511        test_external \
 512            'GitwebCache::*FileCache*' \
 513            "$PERL_PATH" "$TEST_DIRECTORY"/t9503/test_cache_interface.pl
 514
 515   If the test is outputting its own TAP you should set the
 516   test_external_has_tap variable somewhere before calling the first
 517   test_external* function. See t9700-perl-git.sh for an example.
 518
 519        # The external test will outputs its own plan
 520        test_external_has_tap=1
 521
 522 - test_external_without_stderr [<prereq>] <message> <external> <script>
 523
 524   Like test_external but fail if there's any output on stderr,
 525   instead of checking the exit code.
 526
 527        test_external_without_stderr \
 528            'Perl API' \
 529            "$PERL_PATH" "$TEST_DIRECTORY"/t9700/test.pl
 530
 531 - test_expect_code <exit-code> <command>
 532
 533   Run a command and ensure that it exits with the given exit code.
 534   For example:
 535
 536        test_expect_success 'Merge with d/f conflicts' '
 537                test_expect_code 1 git merge "merge msg" B master
 538        '
 539
 540 - test_must_fail <git-command>
 541
 542   Run a git command and ensure it fails in a controlled way.  Use
 543   this instead of "! <git-command>".  When git-command dies due to a
 544   segfault, test_must_fail diagnoses it as an error; "! <git-command>"
 545   treats it as just another expected failure, which would let such a
 546   bug go unnoticed.
 547
 548 - test_might_fail <git-command>
 549
 550   Similar to test_must_fail, but tolerate success, too.  Use this
 551   instead of "<git-command> || :" to catch failures due to segv.
 552
 553 - test_cmp <expected> <actual>
 554
 555   Check whether the content of the <actual> file matches the
 556   <expected> file.  This behaves like "cmp" but produces more
 557   helpful output when the test is run with "-v" option.
 558
 559 - test_line_count (= | -lt | -ge | ...) <length> <file>
 560
 561   Check whether a file has the length it is expected to.
 562
 563 - test_path_is_file <path> [<diagnosis>]
 564   test_path_is_dir <path> [<diagnosis>]
 565   test_path_is_missing <path> [<diagnosis>]
 566
 567   Check if the named path is a file, if the named path is a
 568   directory, or if the named path does not exist, respectively,
 569   and fail otherwise, showing the <diagnosis> text.
 570
 571 - test_when_finished <script>
 572
 573   Prepend <script> to a list of commands to run to clean up
 574   at the end of the current test.  If some clean-up command
 575   fails, the test will not pass.
 576
 577   Example:
 578
 579        test_expect_success 'branch pointing to non-commit' '
 580                git rev-parse HEAD^{tree} >.git/refs/heads/invalid &&
 581                test_when_finished "git update-ref -d refs/heads/invalid" &&
 582                ...
 583        '
 584
 585 - test_pause
 586
 587        This command is useful for writing and debugging tests and must be
 588        removed before submitting. It halts the execution of the test and
 589        spawns a shell in the trash directory. Exit the shell to continue
 590        the test. Example:
 591
 592        test_expect_success 'test' '
 593                git do-something >actual &&
 594                test_pause &&
 595                test_cmp expected actual
 596        '
 597
 598Prerequisites
 599-------------
 600
 601These are the prerequisites that the test library predefines with
 602test_have_prereq.
 603
 604See the prereq argument to the test_* functions in the "Test harness
 605library" section above and the "test_have_prereq" function for how to
 606use these, and "test_set_prereq" for how to define your own.
 607
 608 - PERL & PYTHON
 609
 610   Git wasn't compiled with NO_PERL=YesPlease or
 611   NO_PYTHON=YesPlease. Wrap any tests that need Perl or Python in
 612   these.
 613
 614 - POSIXPERM
 615
 616   The filesystem supports POSIX style permission bits.
 617
 618 - BSLASHPSPEC
 619
 620   Backslashes in pathspec are not directory separators. This is not
 621   set on Windows. See 6fd1106a for details.
 622
 623 - EXECKEEPSPID
 624
 625   The process retains the same pid across exec(2). See fb9a2bea for
 626   details.
 627
 628 - PIPE
 629
 630   The filesystem we're on supports creation of FIFOs (named pipes)
 631   via mkfifo(1).
 632
 633 - SYMLINKS
 634
 635   The filesystem we're on supports symbolic links. E.g. a FAT
 636   filesystem doesn't support these. See 704a3143 for details.
 637
 638 - SANITY
 639
 640   Test is not run by root user, and an attempt to write to an
 641   unwritable file is expected to fail correctly.
 642
 643 - LIBPCRE
 644
 645   Git was compiled with USE_LIBPCRE=YesPlease. Wrap any tests
 646   that use git-grep --perl-regexp or git-grep -P in these.
 647
 648 - CASE_INSENSITIVE_FS
 649
 650   Test is run on a case insensitive file system.
 651
 652 - UTF8_NFD_TO_NFC
 653
 654   Test is run on a filesystem which converts decomposed utf-8 (nfd)
 655   to precomposed utf-8 (nfc).
 656
 657Tips for Writing Tests
 658----------------------
 659
 660As with any programming projects, existing programs are the best
 661source of the information.  However, do _not_ emulate
 662t0000-basic.sh when writing your tests.  The test is special in
 663that it tries to validate the very core of GIT.  For example, it
 664knows that there will be 256 subdirectories under .git/objects/,
 665and it knows that the object ID of an empty tree is a certain
 66640-byte string.  This is deliberately done so in t0000-basic.sh
 667because the things the very basic core test tries to achieve is
 668to serve as a basis for people who are changing the GIT internal
 669drastically.  For these people, after making certain changes,
 670not seeing failures from the basic test _is_ a failure.  And
 671such drastic changes to the core GIT that even changes these
 672otherwise supposedly stable object IDs should be accompanied by
 673an update to t0000-basic.sh.
 674
 675However, other tests that simply rely on basic parts of the core
 676GIT working properly should not have that level of intimate
 677knowledge of the core GIT internals.  If all the test scripts
 678hardcoded the object IDs like t0000-basic.sh does, that defeats
 679the purpose of t0000-basic.sh, which is to isolate that level of
 680validation in one place.  Your test also ends up needing
 681updating when such a change to the internal happens, so do _not_
 682do it and leave the low level of validation to t0000-basic.sh.
 683
 684Test coverage
 685-------------
 686
 687You can use the coverage tests to find code paths that are not being
 688used or properly exercised yet.
 689
 690To do that, run the coverage target at the top-level (not in the t/
 691directory):
 692
 693    make coverage
 694
 695That'll compile Git with GCC's coverage arguments, and generate a test
 696report with gcov after the tests finish. Running the coverage tests
 697can take a while, since running the tests in parallel is incompatible
 698with GCC's coverage mode.
 699
 700After the tests have run you can generate a list of untested
 701functions:
 702
 703    make coverage-untested-functions
 704
 705You can also generate a detailed per-file HTML report using the
 706Devel::Cover module. To install it do:
 707
 708   # On Debian or Ubuntu:
 709   sudo aptitude install libdevel-cover-perl
 710
 711   # From the CPAN with cpanminus
 712   curl -L http://cpanmin.us | perl - --sudo --self-upgrade
 713   cpanm --sudo Devel::Cover
 714
 715Then, at the top-level:
 716
 717    make cover_db_html
 718
 719That'll generate a detailed cover report in the "cover_db_html"
 720directory, which you can then copy to a webserver, or inspect locally
 721in a browser.