Add teardown_request decorator. Fixes issue #174

CHANGES

@@ -37,6 +37,8 @@ Release date to be announced, codename to be selected
   was incorrectly introduced in 0.6.
 - Added `create_jinja_loader` to override the loader creation process.
 - Implemented a silent flag for `config.from_pyfile`.
+- Added `teardown_request` decorator, for functions that should run at the end
+  of a request regardless of whether an exception occurred.
 
 Version 0.6.1
 -------------

docs/tutorial/dbcon.rst

@@ -8,8 +8,11 @@ but how can we elegantly do that for requests?  We will need the database
 connection in all our functions so it makes sense to initialize them
 before each request and shut them down afterwards.
 
-Flask allows us to do that with the :meth:`~flask.Flask.before_request` and
-:meth:`~flask.Flask.after_request` decorators::
+Flask allows us to do that with the :meth:`~flask.Flask.before_request`,
+:meth:`~flask.Flask.after_request` and :meth:`~flask.Flask.teardown_request`
+decorators. In debug mode, if an error is raised,
+:meth:`~flask.Flask.after_request` won't be run, and you'll have access to the
+db connection in the interactive debugger::
 
     @app.before_request
     def before_request():
@@ -20,13 +23,29 @@ Flask allows us to do that with the :meth:`~flask.Flask.before_request` and
         g.db.close()
         return response
 
+If you want to guarantee that the connection is always closed in debug mode, you
+can close it in a function decorated with :meth:`~flask.Flask.teardown_request`:
+
+    @app.before_request
+    def before_request():
+        g.db = connect_db()
+
+    @app.teardown_request
+    def teardown_request(exception):
+        g.db.close()
+
 Functions marked with :meth:`~flask.Flask.before_request` are called before
-a request and passed no arguments, functions marked with
+a request and passed no arguments.  Functions marked with
 :meth:`~flask.Flask.after_request` are called after a request and
 passed the response that will be sent to the client.  They have to return
 that response object or a different one.  In this case we just return it
 unchanged.
 
+Functions marked with :meth:`~flask.Flask.teardown_request` get called after the
+response has been constructed.  They are not allowed to modify the request, and
+their return values are ignored.  If an exception occurred while the request was
+being processed, it is passed to each function; otherwise, None is passed in.
+
 We store our current database connection on the special :data:`~flask.g`
 object that flask provides for us.  This object stores information for one
 request only and is available from within each function.  Never store such

flask/app.py

@@ -11,6 +11,7 @@
 
 from __future__ import with_statement
 
+import sys
 from threading import Lock
 from datetime import timedelta, datetime
 from itertools import chain
@@ -247,6 +248,18 @@ def __init__(self, import_name, static_path=None):
         #: :meth:`after_request` decorator.
         self.after_request_funcs = {}
 
+        #: A dictionary with lists of functions that are called after
+        #: each request, even if an exception has occurred. The key of the
+        #: dictionary is the name of the module this function is active for,
+        #: `None` for all requests. These functions are not allowed to modify
+        #: the request, and their return values are ignored. If an exception
+        #: occurred while processing the request, it gets passed to each
+        #: teardown_request function. To register a function here, use the
+        #: :meth:`teardown_request` decorator.
+        #:
+        #: .. versionadded:: 0.7
+        self.teardown_request_funcs = {}
+
         #: A dictionary with list of functions that are called without argument
         #: to populate the template context.  The key of the dictionary is the
         #: name of the module this function is active for, `None` for all
@@ -704,6 +717,11 @@ def after_request(self, f):
         self.after_request_funcs.setdefault(None, []).append(f)
         return f
 
+    def teardown_request(self, f):
+        """Register a function to be run at the end of each request, regardless of whether there was an exception or not."""
+        self.teardown_request_funcs.setdefault(None, []).append(f)
+        return f
+
     def context_processor(self, f):
         """Registers a template context processor function."""
         self.template_context_processors[None].append(f)
@@ -869,6 +887,20 @@ def process_response(self, response):
             response = handler(response)
         return response
 
+    def do_teardown_request(self):
+        """Called after the actual request dispatching and will
+        call every as :meth:`teardown_request` decorated function.
+        """
+        funcs = reversed(self.teardown_request_funcs.get(None, ()))
+        mod = request.module
+        if mod and mod in self.teardown_request_funcs:
+            funcs = chain(funcs, reversed(self.teardown_request_funcs[mod]))
+        exc = sys.exc_info()[1]
+        for func in funcs:
+            rv = func(exc)
+            if rv is not None:
+                return rv
+
     def request_context(self, environ):
         """Creates a request context from the given environment and binds
         it to the current context.  This must be used in combination with
@@ -947,6 +979,11 @@ def wsgi_app(self, environ, start_response):
            even if an exception happens database have the chance to
            properly close the connection.
 
+        .. versionchanged:: 0.7
+           The :meth:`teardown_request` functions get called at the very end of
+           processing the request. If an exception was thrown, it gets passed to
+           each teardown_request function.
+
         :param environ: a WSGI environment
         :param start_response: a callable accepting a status code,
                                a list of headers and an optional
@@ -965,6 +1002,8 @@ def wsgi_app(self, environ, start_response):
                 response = self.process_response(response)
             except Exception, e:
                 response = self.make_response(self.handle_exception(e))
+            finally:
+                self.do_teardown_request()
             request_finished.send(self, response=response)
             return response(environ, start_response)
 

tests/flask_tests.py

@@ -413,6 +413,72 @@ def fails():
         assert 'Internal Server Error' in rv.data
         assert len(called) == 1
 
+    def test_teardown_request_handler(self):
+        called = []
+        app = flask.Flask(__name__)
+        @app.teardown_request
+        def teardown_request(exc):
+            called.append(True)
+            return "Ignored"
+        @app.route('/')
+        def root():
+            return "Response"
+        rv = app.test_client().get('/')
+        assert rv.status_code == 200
+        assert 'Response' in rv.data
+        assert len(called) == 1
+
+    def test_teardown_request_handler_debug_mode(self):
+        called = []
+        app = flask.Flask(__name__)
+        app.debug = True
+        @app.teardown_request
+        def teardown_request(exc):
+            called.append(True)
+            return "Ignored"
+        @app.route('/')
+        def root():
+            return "Response"
+        rv = app.test_client().get('/')
+        assert rv.status_code == 200
+        assert 'Response' in rv.data
+        assert len(called) == 1
+
+
+    def test_teardown_request_handler_error(self):
+        called = []
+        app = flask.Flask(__name__)
+        @app.teardown_request
+        def teardown_request1(exc):
+            assert type(exc) == ZeroDivisionError
+            called.append(True)
+            # This raises a new error and blows away sys.exc_info(), so we can
+            # test that all teardown_requests get passed the same original
+            # exception.
+            try:
+                raise TypeError
+            except:
+                pass
+        @app.teardown_request
+        def teardown_request2(exc):
+            assert type(exc) == ZeroDivisionError
+            called.append(True)
+            # This raises a new error and blows away sys.exc_info(), so we can
+            # test that all teardown_requests get passed the same original
+            # exception.
+            try:
+                raise TypeError
+            except:
+                pass
+        @app.route('/')
+        def fails():
+            1/0
+        rv = app.test_client().get('/')
+        assert rv.status_code == 500
+        assert 'Internal Server Error' in rv.data
+        assert len(called) == 2
+
+
     def test_before_after_request_order(self):
         called = []
         app = flask.Flask(__name__)
@@ -430,12 +496,18 @@ def after1(response):
         def after2(response):
             called.append(3)
             return response
+        @app.teardown_request
+        def finish1(exc):
+            called.append(6)
+        @app.teardown_request
+        def finish2(exc):
+            called.append(5)
         @app.route('/')
         def index():
             return '42'
         rv = app.test_client().get('/')
         assert rv.data == '42'
-        assert called == [1, 2, 3, 4]
+        assert called == [1, 2, 3, 4, 5, 6]
 
     def test_error_handling(self):
         app = flask.Flask(__name__)