Django 源码小剖: 初探中间件(middleware)

因为考虑到文章的长度, 所以 BaseHandler 的展开被推迟了. 在 BaseHandler 中隐藏着中间件的信息, 较常见的 SessionMiddleware 就已经默认安装.  BaseHandler 的展开主要是以代码为主, 但已经加入了注释; 文章的最后附一张美图 .

最后, 祝程序员们节日快乐, 别太宅了 ;)

BaseHandler 详解

BaseHandler 在 django.core.handlers.base.py 中定义, 有两个核心的成员方法不得不提, 里面就涉及了中间件的信息, 照抄如下(有点长, 但已经加入注释):

# 好经典的 handler
class BaseHandler(object):
    # Changes that are always applied to a response (in this order).
    response_fixes = [
        http.fix_location_header,
        http.conditional_content_removal,
        http.fix_IE_for_attach,
        http.fix_IE_for_vary,
    ]

    初始化函数, 初始化请求中间件, 视图中间件, 模版中间件, 响应中间件和异常中间件.
    def __init__(self):
        self._request_middleware = self._view_middleware =
            self._template_response_middleware =
            self._response_middleware =
            self._exception_middleware = None  视图, 模版相应, 相应, 异常中间件, 请求中间件

    根据 mysite.settings.py 中的 `MIDDLEWARE_CLASSES` 添加所有的中间件.
    def load_middleware(self):
        """
        Populate middleware lists from settings.MIDDLEWARE_CLASSES.

        从 settings 中加载各种中间件

        Must be called after the environment is fixed (see __call__ in subclasses).
        """
        # 初始化四种中间件
        self._view_middleware = []
        self._template_response_middleware = []
        self._response_middleware = []
        self._exception_middleware = []

        # 临时的请求中间件, 因为在加入中间件的过程中, 可能会出现异常, 而出现异常都导致加载中间件的不成功, 因此将 self._request_middleware 的赋值放在最后, 表示已经成功.
        request_middleware = []

        # settings.MIDDLEWARE_CLASSES 设置项指定需要预装的中间件
        for middleware_path in settings.MIDDLEWARE_CLASSES:
            try:
                mw_module, mw_classname = middleware_path.rsplit('.', 1)
            except ValueError:
                raise exceptions.ImproperlyConfigured('%s isn\'t a middleware module' % middleware_path)

            try:
                尝试导入中间件所在模块.
                mod = import_module(mw_module)
            except ImportError as e:
                raise exceptions.ImproperlyConfigured('Error importing middleware %s: "%s"' % (mw_module, e))

            try:
                尝试得到某种中间件类
                mw_class = getattr(mod, mw_classname)
            except AttributeError:
                raise exceptions.ImproperlyConfigured('Middleware module "%s" does not define a "%s" class' % (mw_module, mw_classname))

            try:
                尝试实例化
                mw_instance = mw_class()
            except exceptions.MiddlewareNotUsed:
                continue

            和 urllib 的处理方法类似: 请求预处理, 视图处理?, 模版处理, 相应处理, 错误处理(详见我的 urllib 源码剖析)
            if hasattr(mw_instance, 'process_request'):
                # 这里 request_middleware 用的是 append(), 这里是有讲究的:
                # django 规定, 多个请求中间件调用的次序是其出现的次序, 下同
                request_middleware.append(mw_instance.process_request)

            if hasattr(mw_instance, 'process_view'):
                self._view_middleware.append(mw_instance.process_view)

            if hasattr(mw_instance, 'process_template_response'):
                # 这里 _template_response_middleware 用的是 insert() 头插法, 这里是有讲究的:
                # django 规定, 多个模版相应中间件调用的次序是其出现次序的逆序, 下同
                self._template_response_middleware.insert(0, mw_instance.process_template_response)

            if hasattr(mw_instance, 'process_response'):
                self._response_middleware.insert(0, mw_instance.process_response)

            if hasattr(mw_instance, 'process_exception'):
                self._exception_middleware.insert(0, mw_instance.process_exception)

        # We only assign to this when initialization is complete as it is used
        # as a flag for initialization being complete.
        # 结束的标识, 表明中间件加载成功
        self._request_middleware = request_middleware

    # 处理请求的函数, 并返回 response
    def get_response(self, request):
        "Returns an HttpResponse object for the given HttpRequest"
        根据请求, 得到响应

        try:
            为该线程提供默认的 url 处理器
            # Setup default url resolver for this thread, this code is outside
            # the try/except so we don't get a spurious "unbound local
            # variable" exception in the event an exception is raised before
            # resolver is set

            #ROOT_URLCONF = 'mysite.urls'
            urlconf = settings.ROOT_URLCONF

            # set_urlconf() 会设置 url 配置即 settings.ROOT_URLCONF
            urlresolvers.set_urlconf(urlconf)

            # 实例化 RegexURLResolver, 暂且将其理解为一个 url 的匹配处理器, 下节展开
            resolver = urlresolvers.RegexURLResolver(r'^/', urlconf)

            try:
                response = None

                # Apply request middleware 调用请求中间件
                for middleware_method in self._request_middleware:
                    response = middleware_method(request)

                    # 如果此 response 有效, 即不走下面的逻辑
                    if response:
                        break

                # 如果没有结果
                if response is None:
                    # 尝试 request 中是否有 urlconf, 一般没有, 可以忽略此段代码!!!
                    if hasattr(request, 'urlconf'):
                        # Reset url resolver with a custom urlconf. 自定义的 urlconf
                        urlconf = request.urlconf
                        urlresolvers.set_urlconf(urlconf)
                        resolver = urlresolvers.RegexURLResolver(r'^/', urlconf)
                    # 调用 RegexURLResolver.resolve(), 可以理解为启动匹配的函数; 返回 ResolverMatch 实例
                    resolver_match = resolver.resolve(request.path_info)

                    # resolver_match 对象中存储了有用的信息, 譬如 callback 就是我们在 views.py 中定义的函数.
                    callback, callback_args, callback_kwargs = resolver_match

                    # 将返回的 resolver_match 挂钩到 request
                    request.resolver_match = resolver_match

                    # Apply view middleware 调用视图中间件
                    for middleware_method in self._view_middleware:
                        response = middleware_method(request, callback, callback_args, callback_kwargs)

                        # 如果此 response 有效, 即不走下面的逻辑
                        if response:
                            break

                # response 还是为空
                if response is None:
                    try:
                        # 这里调用的是真正的处理函数, 我们一般在 view.py 中定义这些函数
                        response = callback(request, *callback_args, **callback_kwargs)

                    except Exception as e:
                        # If the view raised an exception, run it through exception
                        # middleware, and if the exception middleware returns a
                        # response, use that. Otherwise, reraise the exception.

                        # 出现异常, 调用异常中间件
                        for middleware_method in self._exception_middleware:
                            response = middleware_method(request, e)

                            # 如果此 response 有效, 即不走下面的逻辑
                            if response:
                                break

                        if response is None:
                            raise

                # response 还是为空, 可能就要异常了
                # Complain if the view returned None (a common error).
                if response is None:
                    if isinstance(callback, types.FunctionType):    # FBV
                        view_name = callback.__name__
                    else:                                           # CBV
                        view_name = callback.__class__.__name__ + '.__call__'
                    raise ValueError("The view %s.%s didn't return an HttpResponse object." % (callback.__module__, view_name))

                # If the response supports deferred rendering, apply template
                # response middleware and the render the response 如果 response 实现了 render, 那么渲染返回.
                if hasattr(response, 'render') and callable(response.render):
                    for middleware_method in self._template_response_middleware:
                        response = middleware_method(request, response)
                    response = response.render()

            except http.Http404 as e:
                logger.warning('Not Found: %s', request.path,
                            extra={
                                'status_code': 404,
                                'request': request
                            })

                # 如果是调试下, 直接要返回 404 页面
                if settings.DEBUG:
                    response = debug.technical_404_response(request, e)
                else:
                    try:
                        # 非调试模式下, 获取 url 处理器的默认 404 处理
                        callback, param_dict = resolver.resolve404()
                        response = callback(request, **param_dict)
                    except:
                        signals.got_request_exception.send(sender=self.__class__, request=request)
                        response = self.handle_uncaught_exception(request, resolver, sys.exc_info())

            # 访问拒绝
            except exceptions.PermissionDenied:
                logger.warning(
                    'Forbidden (Permission denied): %s', request.path,
                    extra={
                        'status_code': 403,
                        'request': request
                    })
                try:
                    callback, param_dict = resolver.resolve403()
                    response = callback(request, **param_dict)
                except:
                    signals.got_request_exception.send(
                            sender=self.__class__, request=request)
                    response = self.handle_uncaught_exception(request,
                            resolver, sys.exc_info())

            except SystemExit:
                # Allow sys.exit() to actually exit. See tickets #1023 and #4701
                raise

            except: # Handle everything else, including SuspiciousOperation, etc.
                # Get the exception info now, in case another exception is thrown later.
                signals.got_request_exception.send(sender=self.__class__, request=request)
                response = self.handle_uncaught_exception(request, resolver, sys.exc_info())
        finally:
            # Reset URLconf for this thread on the way out for complete
            # isolation of request.urlconf 重置, 因为前面有两种 url resolver 的可能, 拒绝混淆
            urlresolvers.set_urlconf(None)

        try:
            # Apply response middleware, regardless of the response 调用响应中间件
            for middleware_method in self._response_middleware:
                response = middleware_method(request, response)

            response = self.apply_response_fixes(request, response)

        except: # Any exception should be gathered and handled
            signals.got_request_exception.send(sender=self.__class__, request=request)
            response = self.handle_uncaught_exception(request, resolver, sys.exc_info())

        return response

    def handle_uncaught_exception(self, request, resolver, exc_info):
        """
        处理未能捕捉的错误

        Processing for any otherwise uncaught exceptions (those that will
        generate HTTP 500 responses). Can be overridden by subclasses who want
        customised 500 handling. 子类中可以重写 500 状态的处理

        Be *very* careful when overriding this because the error could be
        caused by anything, so assuming something like the database is always
        available would be an error.
        """
        if settings.DEBUG_PROPAGATE_EXCEPTIONS:
            raise

        logger.error('Internal Server Error: %s', request.path,
            exc_info=exc_info,
            extra={
                'status_code': 500,
                'request': request
            }
        )

        调试模式特殊处理
        if settings.DEBUG:
            return debug.technical_500_response(request, *exc_info)

        # If Http500 handler is not installed, re-raise last exception 如果http500 处理器都没有安装, 可能会崩溃
        if resolver.urlconf_module is None:
            six.reraise(*exc_info)

        # Return an HttpResponse that displays a friendly error message.
        #这是自定义的 500 处理器
        callback, param_dict = resolver.resolve500()
        return callback(request, **param_dict)

    def apply_response_fixes(self, request, response):
        """
        Applies each of the functions in self.response_fixes to the request and
        response, modifying the response in the process. Returns the new
        response.
        """
        for func in self.response_fixes:
            response = func(request, response)
        return response

故此总结

load_middleware() 函数会根据 mysite.settings.py 中的 MIDDLEWARE_CLASSES 导入所有的中间件. 在 eclipse + pydev 创建 django 的默认设置当中就有默认的中间件:

MIDDLEWARE_CLASSES = (
    'django.middleware.common.CommonMiddleware',
    'django.contrib.sessions.middleware.SessionMiddleware',
    'django.middleware.csrf.CsrfViewMiddleware',
    'django.contrib.auth.middleware.AuthenticationMiddleware',
    'django.contrib.messages.middleware.MessageMiddleware',
    # Uncomment the next line for simple clickjacking protection:
    # 'django.middleware.clickjacking.XFrameOptionsMiddleware',
)

每一个中间件都是一个类, 其内部会实现 process_request(),process_view(),process_template_response(),process_response() 或者 process_exception() 方法. 不一定都实现, 看需求. 而这些方法如果存在, 都会被保存响应的函数列表中, 待将来调用.

get_response() 方法, 中间件调用执行的顺序是请求中间件, 视图中间件, 模版中间件, 异常中间件(可选), 响应中间件. 习惯上, 我把这些简称为请求预处理和响应善后处理.get_response() 返回了 response, 但一长串的 url 是如何匹配的, 并且自己在 views.py 中的函数是在什么时候调用的?

我已经在 github 备份了 Django 源码的注释: Decode-Django, 有兴趣的童鞋 fork 吧.

图解中间件:

django 处理流程图

捣乱 2013-9-14

http://daoluan.net

13 September 2013

disqus 可能被墙~