
pytest-2.3: reasoning for fixture/funcarg evolution
***************************************************

**Target audience**: Reading this document requires basic knowledge of
python testing, xUnit setup methods and the (previous) basic pytest
funcarg mechanism, see http://pytest.org/2.2.4/funcargs.html If you
are new to pytest, then you can simply ignore this section and read
the other sections.


Shortcomings of the previous "pytest_funcarg__" mechanism
=========================================================

The pre pytest-2.3 funcarg mechanism calls a factory each time a
funcarg for a test function is required.  If a factory wants to re-use
a resource across different scopes, it often used the
"request.cached_setup()" helper to manage caching of resources.  Here
is a basic example how we could implement a per-session Database
object:

   # content of conftest.py
   class Database:
       def __init__(self):
           print ("database instance created")
       def destroy(self):
           print ("database instance destroyed")

   def pytest_funcarg__db(request):
       return request.cached_setup(setup=DataBase,
                                   teardown=lambda db: db.destroy,
                                   scope="session")

There are several limitations and difficulties with this approach:

1. Scoping funcarg resource creation is not straight forward, instead
   one must understand the intricate cached_setup() method mechanics.

2. parametrizing the "db" resource is not straight forward: you need
   to apply a "parametrize" decorator or implement a
   "pytest_generate_tests()" hook calling "parametrize()" which
   performs parametrization at the places where the resource is used.
   Moreover, you need to modify the factory to use an "extrakey"
   parameter containing "request.param" to the "cached_setup()" call.

3. Multiple parametrized session-scoped resources will be active at
   the same time, making it hard for them to affect global state of
   the application under test.

4. there is no way how you can make use of funcarg factories in xUnit
   setup methods.

5. A non-parametrized fixture function cannot use a parametrized
   funcarg resource if it isn't stated in the test function signature.

All of these limitations are addressed with pytest-2.3 and its
improved *fixture mechanism*.


Direct scoping of fixture/funcarg factories
===========================================

Instead of calling cached_setup() with a cache scope, you can use the
*@pytest.fixture* decorator and directly state the scope:

   @pytest.fixture(scope="session")
   def db(request):
       # factory will only be invoked once per session -
       db = DataBase()
       request.addfinalizer(db.destroy)  # destroy when session is finished
       return db

This factory implementation does not need to call "cached_setup()"
anymore because it will only be invoked once per session.  Moreover,
the "request.addfinalizer()" registers a finalizer according to the
specified resource scope on which the factory function is operating.


Direct parametrization of funcarg resource factories
====================================================

Previously, funcarg factories could not directly cause
parametrization. You needed to specify a "@parametrize" decorator on
your test function or implement a "pytest_generate_tests" hook to
perform parametrization, i.e. calling a test multiple times with
different value sets.  pytest-2.3 introduces a decorator for use on
the factory itself:

   @pytest.fixture(params=["mysql", "pg"])
   def db(request):
       ... # use request.param

Here the factory will be invoked twice (with the respective "mysql"
and "pg" values set as "request.param" attributes) and and all of the
tests requiring "db" will run twice as well.  The "mysql" and "pg"
values will also be used for reporting the test-invocation variants.

This new way of parametrizing funcarg factories should in many cases
allow to re-use already written factories because effectively
"request.param" was already used when test functions/classes were
parametrized via "parametrize(indirect=True)()" calls.

Of course it's perfectly fine to combine parametrization and scoping:

   @pytest.fixture(scope="session", params=["mysql", "pg"])
   def db(request):
       if request.param == "mysql":
           db = MySQL()
       elif request.param == "pg":
           db = PG()
       request.addfinalizer(db.destroy)  # destroy when session is finished
       return db

This would execute all tests requiring the per-session "db" resource
twice, receiving the values created by the two respective invocations
to the factory function.


No "pytest_funcarg__" prefix when using @fixture decorator
==========================================================

When using the "@fixture" decorator the name of the function denotes
the name under which the resource can be accessed as a function
argument:

   @pytest.fixture()
   def db(request):
       ...

The name under which the funcarg resource can be requested is "db".

You can still use the "old" non-decorator way of specifying funcarg
factories aka:

   def pytest_funcarg__db(request):
       ...

But it is then not possible to define scoping and parametrization. It
is thus recommended to use the factory decorator.


solving per-session setup / autouse fixtures
============================================

pytest for a long time offered a pytest_configure and a
pytest_sessionstart hook which are often used to setup global
resources.  This suffers from several problems:

1. in distributed testing the master process would setup test
   resources that are never needed because it only co-ordinates the
   test run activities of the slave processes.

2. if you only perform a collection (with "--collectonly") resource-
   setup will still be executed.

3. If a pytest_sessionstart is contained in some subdirectories
   conftest.py file, it will not be called.  This stems from the fact
   that this hook is actually used for reporting, in particular the
   test-header with platform/custom information.

Moreover, it was not easy to define a scoped setup from plugins or
conftest files other than to implement a "pytest_runtest_setup()" hook
and caring for scoping/caching yourself.  And it's virtually
impossible to do this with parametrization as "pytest_runtest_setup()"
is called during test execution and parametrization happens at
collection time.

It follows that pytest_configure/session/runtest_setup are often not
appropriate for implementing common fixture needs.  Therefore,
pytest-2.3 introduces *autouse fixtures (xUnit setup on steroids)*
which fully integrate with the generic *fixture mechanism* and
obsolete many prior uses of pytest hooks.


funcargs/fixture discovery now happens at collection time
=========================================================

pytest-2.3 takes care to discover fixture/funcarg factories at
collection time.  This is more efficient especially for large test
suites. Moreover, a call to "py.test --collectonly" should be able to
in the future show a lot of setup-information and thus presents a nice
method to get an overview of fixture management in your project.


Conclusion and compatibility notes
==================================

**funcargs** were originally introduced to pytest-2.0.  In pytest-2.3
the mechanism was extended and refined and is now described as
fixtures:

* previously funcarg factories were specified with a special
  "pytest_funcarg__NAME" prefix instead of using the "@pytest.fixture"
  decorator.

* Factories received a "request" object which managed caching through
  "request.cached_setup()" calls and allowed using other funcargs via
  "request.getfuncargvalue()" calls.  These intricate APIs made it
  hard to do proper parametrization and implement resource caching.
  The new "pytest.fixture`()" decorator allows to declare the scope
  and let pytest figure things out for you.

* if you used parametrization and funcarg factories which made use of
  "request.cached_setup()" it is recommeneded to invest a few minutes
  and simplify your fixture function code to use the *Fixtures as
  Function arguments (funcargs)* decorator instead.  This will also
  allow to take advantage of the automatic per-resource grouping of
  tests.
