test suite: don't consider skipped individual tests as failing

[notmuch] / test / README

Notmuch test suite
==================
This directory contains the test suite for notmuch.

When fixing bugs or enhancing notmuch, you are strongly encouraged to
add tests in this directory to cover what you are trying to fix or
enhance.

Prerequisites
-------------
Some tests require external dependencies to run. Without them, they
will be skipped, or (rarely) marked failed. Please install these, so
that you know if you break anything.

  - dtach(1)
  - emacs(1)
  - emacsclient(1)
  - gdb(1)
  - gpg(1)
  - python(1)

Running Tests
-------------
The easiest way to run tests is to say "make test", (or simply run the
notmuch-test script). Either command will run all available tests.

Alternately, you can run a specific subset of tests by simply invoking
one of the executable scripts in this directory, (such as ./search,
./reply, etc). Note that you will probably want "make test-binaries"
before running individual tests.

The following command-line options are available when running tests:

--debug::
        This may help the person who is developing a new test.
        It causes the command defined with test_debug to run.

--immediate::
        This causes the test to immediately exit upon the first
        failed test.

--valgrind::
        Execute notmuch with valgrind and exit with status
        126 on errors (just like regular tests, this will only stop
        the test script when running under -i).  Valgrind errors
        go to stderr, so you might want to pass the -v option, too.

        Since it makes no sense to run the tests with --valgrind and
        not see any output, this option implies --verbose.  For
        convenience, it also implies --tee.

--tee::
        In addition to printing the test output to the terminal,
        write it to files named 't/test-results/$TEST_NAME.out'.
        As the names depend on the tests' file names, it is safe to
        run the tests with this option in parallel.

--root=<dir>::
        This runs the testsuites specified under a separate directory.
        However, caution is advised, as not all tests are maintained
        with this relocation in mind, so some tests may behave
        differently.

        Pointing this argument at a tmpfs filesystem can improve the
        speed of the test suite for some users.

Certain tests require precomputed databases to complete. You can fetch these
databases with

        make download-test-databases

If you do not download the test databases, the relevant tests will be
skipped.

When invoking the test suite via "make test" any of the above options
can be specified as follows:

        make test OPTIONS="--verbose"

You can choose an emacs binary (and corresponding emacsclient) to run
the tests in one of the following ways.

        TEST_EMACS=my-special-emacs TEST_EMACSCLIENT=my-emacsclient make test
        TEST_EMACS=my-special-emacs TEST_EMACSCLIENT=my-emacsclient ./emacs
        make test TEST_EMACS=my-special-emacs TEST_EMACSCLIENT=my-emacsclient

Some tests may require a c compiler. You can choose the name and flags similarly
to with emacs, e.g.

     make test TEST_CC=gcc TEST_CFLAGS="-g -O2"

Quiet Execution
---------------

Normally, when new script starts and when test PASSes you get a message
printed on screen. This printing can be disabled by setting the
NOTMUCH_TEST_QUIET variable to a non-null value. Message on test
failures and skips are still printed.

Skipping Tests
--------------
If, for any reason, you need to skip one or more tests, you can do so
by setting the NOTMUCH_SKIP_TESTS variable to the name of one or more
sections of tests.

For example:

    $ NOTMUCH_SKIP_TESTS="search reply" make test

Even more fine-grained skipping is possible by appending a test number
(or glob pattern) after the section name. For example, the first
search test and the second reply test could be skipped with:

    $ NOTMUCH_SKIP_TESTS="search.1 reply.2" make test

Note that some tests in the existing test suite rely on previous test
items, so you cannot arbitrarily skip any test and expect the
remaining tests to be unaffected.

Currently we do not consider skipped tests as build failures. For
maximum robustness, when setting up automated build processes, you
should explicitely skip tests, rather than relying on notmuch's
detection of missing prerequisites. In the future we may treat tests
unable to run because of missing prerequisites, but not explicitely
skipped by the user, as failures.

Writing Tests
-------------
The test script is written as a shell script.  It should start with
the standard "#!/usr/bin/env bash" with copyright notices, and an
assignment to variable 'test_description', like this:

        #!/usr/bin/env bash
        #
        # Copyright (c) 2005 Junio C Hamano
        #

        test_description='xxx test (option --frotz)

        This test exercises the "notmuch xxx" command when
        given the option --frotz.'

Source 'test-lib.sh'
--------------------
After assigning test_description, the test script should source
test-lib.sh like this:

        . ./test-lib.sh || exit 1

This test harness library does the following things:

 - If the script is invoked with command line argument --help
   (or -h), it shows the test_description and exits.

 - Creates a temporary directory with default notmuch-config and a
   mail store with a corpus of mail, (initially, 50 early messages
   sent to the notmuch list). This directory is
   test/tmp.<test-basename>. The path to notmuch-config is exported in
   NOTMUCH_CONFIG environment variable and mail store path is stored
   in MAIL_DIR variable.

 - Defines standard test helper functions for your scripts to
   use.  These functions are designed to make all scripts behave
   consistently when command line arguments --verbose (or -v),
   --debug (or -d), and --immediate (or -i) is given.

End with test_done
------------------
Your script will be a sequence of tests, using helper functions
from the test harness library.  At the end of the script, call
'test_done'.

Test harness library
--------------------
There are a handful helper functions defined in the test harness
library for your script to use.

 test_expect_success <message> <script>

   This takes two strings as parameter, and evaluates the
   <script>.  If it yields success, test is considered
   successful.  <message> should state what it is testing.

 test_begin_subtest <message>

   Set the test description message for a subsequent test_expect_equal
   invocation (see below).

 test_subtest_known_broken

   Mark the current test as broken.  Such tests are expected to fail.
   Unlike the normal tests, which say "PASS" on success and "FAIL" on
   failure, these will say "FIXED" on success and "BROKEN" on failure.
   Failures from these tests won't cause -i (immediate) to stop.  A
   test must call this before any test_expect_* function.

 test_expect_equal <output> <expected>

   This is an often-used convenience function built on top of
   test_expect_success. It uses the message from the last
   test_begin_subtest call, so call before calling
   test_expect_equal. This function generates a successful test if
   both the <output> and <expected> strings are identical. If not, it
   will generate a failure and print the difference of the two
   strings.

 test_expect_equal_file <file1> <file2>

   Identical to test_expect_equal, except that <file1> and <file2>
   are files instead of strings.  This is a much more robust method to
   compare formatted textual information, since it also notices
   whitespace and closing newline differences.

 test_expect_equal_json <output> <expected>

   Identical to test_expect_equal, except that the two strings are
   treated as JSON and canonicalized before equality testing.  This is
   useful to abstract away from whitespace differences in the expected
   output and that generated by running a notmuch command.

 test_debug <script>

   This takes a single argument, <script>, and evaluates it only
   when the test script is started with --debug command line
   argument.  This is primarily meant for use during the
   development of a new test script.

 test_emacs <emacs-lisp-expressions>

   This function executes the provided emacs lisp script within
   emacs. The script can be a sequence of emacs lisp expressions,
   (that is, they will be evaluated within a progn form). Emacs
   stdout and stderr is not available, the common way to get output
   is to save it to a file. There are some auxiliary functions
   useful in emacs tests provided in test-lib.el. Do not use `setq'
   for setting variables in Emacs tests because it affects other
   tests that may run in the same Emacs instance.  Use `let' instead
   so the scope of the changed variables is limited to a single test.

 test_emacs_expect_t <emacs-lisp-expressions>

  This function executes the provided emacs lisp script within
  emacs in a manner similar to 'test_emacs'. The expressions should
  return the value `t' to indicate that the test has passed. If the
  test does not return `t' then it is considered failed and all data
  returned by the test is reported to the tester.

 test_done

   Your test script must have test_done at the end.  Its purpose
   is to summarize successes and failures in the test script and
   exit with an appropriate error code.

There are also a number of notmuch-specific auxiliary functions and
variables which are useful in writing tests:

  generate_message

    Generates a message with an optional template. Most tests will
    actually prefer to call add_message. See below.

  add_message

    Generate a message and add it to the database (by calling "notmuch
    new"). It is sufficient to simply call add_message with no
    arguments if you don't care about the content of the message. If
    more control is needed, arguments can be provide to specify many
    different header values for the new message. See the documentation
    within test-lib.sh or refer to many example calls within existing
    tests.

  add_email_corpus

    This function should be called at the beginning of a test file
    when a test needs to operate on a non-empty body of messages. It
    will initialize the mail database to a known state of 50 sample
    messages, (culled from the early history of the notmuch mailing
    list).

  notmuch_counter_reset
  $notmuch_counter_command
  notmuch_counter_value

    These allow to count how many times notmuch binary is called.
    notmuch_counter_reset() function generates a script that counts
    how many times it is called and resets the counter to zero.  The
    function sets $notmuch_counter_command variable to the path to the
    generated script that should be called instead of notmuch to do
    the counting.  The notmuch_counter_value() function prints the
    current counter value.

There are also functions which remove various environment-dependent
values from notmuch output; these are useful to ensure that test
results remain consistent across different machines.

 notmuch_search_sanitize
 notmuch_show_sanitize
 notmuch_show_sanitize_all
 notmuch_json_show_sanitize

   All these functions should receive the text to be sanitized as the
   input of a pipe, e.g.
   output=`notmuch search "..." | notmuch_search_sanitize`