Files
git/t/unit-tests/clar/README.md
Patrick Steinhardt 9b7caa2809 t: import the clar unit testing framework
Our unit testing framework is a homegrown solution. While it supports
most of our needs, it is likely that the volume of unit tests will grow
quite a bit in the future such that we can exercise low-level subsystems
directly. This surfaces several shortcomings that the current solution
has:

  - There is no way to run only one specific tests. While some of our
    unit tests wire this up manually, others don't. In general, it
    requires quite a bit of boilerplate to get this set up correctly.

  - Failures do not cause a test to stop execution directly. Instead,
    the test author needs to return manually whenever an assertion
    fails. This is rather verbose and is not done correctly in most of
    our unit tests.

  - Wiring up a new testcase requires both implementing the test
    function and calling it in the respective test suite's main
    function, which is creating code duplication.

We can of course fix all of these issues ourselves, but that feels
rather pointless when there are already so many unit testing frameworks
out there that have those features.

We line out some requirements for any unit testing framework in
"Documentation/technical/unit-tests.txt". The "clar" unit testing
framework, which isn't listed in that table yet, ticks many of the
boxes:

  - It is licensed under ISC, which is compatible.

  - It is easily vendorable because it is rather tiny at around 1200
    lines of code.

  - It is easily hackable due to the same reason.

  - It has TAP support.

  - It has skippable tests.

  - It preprocesses test files in order to extract test functions, which
    then get wired up automatically.

While it's not perfect, the fact that clar originates from the libgit2
project means that it should be rather easy for us to collaborate with
upstream to plug any gaps.

Import the clar unit testing framework at commit 1516124 (Merge pull
request #97 from pks-t/pks-whitespace-fixes, 2024-08-15). The framework
will be wired up in subsequent commits.

Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2024-09-04 08:41:36 -07:00

330 lines
10 KiB
Markdown

Come out and Clar
=================
In Catalan, "clar" means clear, easy to perceive. Using clar will make it
easy to test and make clear the quality of your code.
> _Historical note_
>
> Originally the clar project was named "clay" because the word "test" has its
> roots in the latin word *"testum"*, meaning "earthen pot", and *"testa"*,
> meaning "piece of burned clay"?
>
> This is because historically, testing implied melting metal in a pot to
> check its quality. Clay is what tests are made of.
## Quick Usage Overview
Clar is a minimal C unit testing framework. It's been written to replace the
old framework in [libgit2][libgit2], but it's both very versatile and
straightforward to use.
Can you count to funk?
- **Zero: Initialize test directory**
~~~~ sh
$ mkdir tests
$ cp -r $CLAR_ROOT/clar* tests
$ cp $CLAR_ROOT/test/clar_test.h tests
$ cp $CLAR_ROOT/test/main.c.sample tests/main.c
~~~~
- **One: Write some tests**
File: tests/adding.c:
~~~~ c
/* adding.c for the "Adding" suite */
#include "clar.h"
static int *answer;
void test_adding__initialize(void)
{
answer = malloc(sizeof(int));
cl_assert_(answer != NULL, "No memory left?");
*answer = 42;
}
void test_adding__cleanup(void)
{
free(answer);
}
void test_adding__make_sure_math_still_works(void)
{
cl_assert_(5 > 3, "Five should probably be greater than three");
cl_assert_(-5 < 2, "Negative numbers are small, I think");
cl_assert_(*answer == 42, "The universe is doing OK. And the initializer too.");
}
~~~~~
- **Two: Build the test executable**
~~~~ sh
$ cd tests
$ $CLAR_PATH/generate.py .
Written `clar.suite` (1 suites)
$ gcc -I. clar.c main.c adding.c -o testit
~~~~
- **Funk: Funk it.**
~~~~ sh
$ ./testit
~~~~
## The Clar Test Suite
Writing a test suite is pretty straightforward. Each test suite is a `*.c`
file with a descriptive name: this encourages modularity.
Each test suite has optional initialize and cleanup methods. These methods
will be called before and after running **each** test in the suite, even if
such test fails. As a rule of thumb, if a test needs a different initializer
or cleanup method than another test in the same module, that means it
doesn't belong in that module. Keep that in mind when grouping tests
together.
The `initialize` and `cleanup` methods have the following syntax, with
`suitename` being the current suite name, e.g. `adding` for the `adding.c`
suite.
~~~~ c
void test_suitename__initialize(void)
{
/* init */
}
void test_suitename__cleanup(void)
{
/* cleanup */
}
~~~~
These methods are encouraged to use static, global variables to store the state
that will be used by all tests inside the suite.
~~~~ c
static git_repository *_repository;
void test_status__initialize(void)
{
create_tmp_repo(STATUS_REPO);
git_repository_open(_repository, STATUS_REPO);
}
void test_status__cleanup(void)
{
git_repository_close(_repository);
git_path_rm(STATUS_REPO);
}
void test_status__simple_test(void)
{
/* do something with _repository */
}
~~~~
Writing the actual tests is just as straightforward. Tests have the
`void test_suitename__test_name(void)` signature, and they should **not**
be static. Clar will automatically detect and list them.
Tests are run as they appear on their original suites: they have no return
value. A test is considered "passed" if it doesn't raise any errors. Check
the "Clar API" section to see the various helper functions to check and
raise errors during test execution.
__Caution:__ If you use assertions inside of `test_suitename__initialize`,
make sure that you do not rely on `__initialize` being completely run
inside your `test_suitename__cleanup` function. Otherwise you might
encounter ressource cleanup twice.
## How does Clar work?
To use Clar:
1. copy the Clar boilerplate to your test directory
2. copy (and probably modify) the sample `main.c` (from
`$CLAR_PATH/test/main.c.sample`)
3. run the Clar mixer (a.k.a. `generate.py`) to scan your test directory and
write out the test suite metadata.
4. compile your test files and the Clar boilerplate into a single test
executable
5. run the executable to test!
The Clar boilerplate gives you a set of useful test assertions and features
(like accessing or making sandbox copies of fixture data). It consists of
the `clar.c` and `clar.h` files, plus the code in the `clar/` subdirectory.
You should not need to edit these files.
The sample `main.c` (i.e. `$CLAR_PATH/test/main.c.sample`) file invokes
`clar_test(argc, argv)` to run the tests. Usually, you will edit this file
to perform any framework specific initialization and teardown that you need.
The Clar mixer (`generate.py`) recursively scans your test directory for
any `.c` files, parses them, and writes the `clar.suite` file with all of
the metadata about your tests. When you build, the `clar.suite` file is
included into `clar.c`.
The mixer can be run with **Python 2.5, 2.6, 2.7, 3.0, 3.1, 3.2 and PyPy 1.6**.
Commandline usage of the mixer is as follows:
$ ./generate.py .
Where `.` is the folder where all the test suites can be found. The mixer
will automatically locate all the relevant source files and build the
testing metadata. The metadata will be written to `clar.suite`, in the same
folder as all the test suites. This file is included by `clar.c` and so
must be accessible via `#include` when building the test executable.
$ gcc -I. clar.c main.c suite1.c test2.c -o run_tests
**Note that the Clar mixer only needs to be ran when adding new tests to a
suite, in order to regenerate the metadata**. As a result, the `clar.suite`
file can be checked into version control if you wish to be able to build
your test suite without having to re-run the mixer.
This is handy when e.g. generating tests in a local computer, and then
building and testing them on an embedded device or a platform where Python
is not available.
### Fixtures
Clar can create sandboxed fixtures for you to use in your test. You'll need to compile *clar.c* with an additional `CFLAG`, `-DCLAR_FIXTURE_PATH`. This should be an absolute path to your fixtures directory.
Once that's done, you can use the fixture API as defined below.
## The Clar API
Clar makes the following methods available from all functions in a test
suite.
- `cl_must_pass(call)`, `cl_must_pass_(call, message)`: Verify that the given
function call passes, in the POSIX sense (returns a value greater or equal
to 0).
- `cl_must_fail(call)`, `cl_must_fail_(call, message)`: Verify that the given
function call fails, in the POSIX sense (returns a value less than 0).
- `cl_assert(expr)`, `cl_assert_(expr, message)`: Verify that `expr` is true.
- `cl_check_pass(call)`, `cl_check_pass_(call, message)`: Verify that the
given function call passes, in the POSIX sense (returns a value greater or
equal to 0). If the function call doesn't succeed, a test failure will be
logged but the test's execution will continue.
- `cl_check_fail(call)`, `cl_check_fail_(call, message)`: Verify that the
given function call fails, in the POSIX sense (returns a value less than
0). If the function call doesn't fail, a test failure will be logged but
the test's execution will continue.
- `cl_check(expr)`: Verify that `expr` is true. If `expr` is not
true, a test failure will be logged but the test's execution will continue.
- `cl_fail(message)`: Fail the current test with the given message.
- `cl_warning(message)`: Issue a warning. This warning will be
logged as a test failure but the test's execution will continue.
- `cl_set_cleanup(void (*cleanup)(void *), void *opaque)`: Set the cleanup
method for a single test. This method will be called with `opaque` as its
argument before the test returns (even if the test has failed).
If a global cleanup method is also available, the local cleanup will be
called first, and then the global.
- `cl_assert_equal_i(int,int)`: Verify that two integer values are equal.
The advantage of this over a simple `cl_assert` is that it will format
a much nicer error report if the values are not equal.
- `cl_assert_equal_s(const char *,const char *)`: Verify that two strings
are equal. The expected value can also be NULL and this will correctly
test for that.
- `cl_fixture_sandbox(const char *)`: Sets up a sandbox for a fixture
so that you can mutate the file directly.
- `cl_fixture_cleanup(const char *)`: Tears down the previous fixture
sandbox.
- `cl_fixture(const char *)`: Gets the full path to a fixture file.
Please do note that these methods are *always* available whilst running a
test, even when calling auxiliary/static functions inside the same file.
It's strongly encouraged to perform test assertions in auxiliary methods,
instead of returning error values. This is considered good Clar style.
Style Example:
~~~~ c
/*
* Bad style: auxiliary functions return an error code
*/
static int check_string(const char *str)
{
const char *aux = process_string(str);
if (aux == NULL)
return -1;
return strcmp(my_function(aux), str) == 0 ? 0 : -1;
}
void test_example__a_test_with_auxiliary_methods(void)
{
cl_must_pass_(
check_string("foo"),
"String differs after processing"
);
cl_must_pass_(
check_string("bar"),
"String differs after processing"
);
}
~~~~
~~~~ c
/*
* Good style: auxiliary functions perform assertions
*/
static void check_string(const char *str)
{
const char *aux = process_string(str);
cl_assert_(
aux != NULL,
"String processing failed"
);
cl_assert_(
strcmp(my_function(aux), str) == 0,
"String differs after processing"
);
}
void test_example__a_test_with_auxiliary_methods(void)
{
check_string("foo");
check_string("bar");
}
~~~~
About Clar
==========
Clar has been written from scratch by [Vicent Martí](https://github.com/vmg),
to replace the old testing framework in [libgit2][libgit2].
Do you know what languages are *in* on the SF startup scene? Node.js *and*
Latin. Follow [@vmg](https://www.twitter.com/vmg) on Twitter to
receive more lessons on word etymology. You can be hip too.
[libgit2]: https://github.com/libgit2/libgit2