151 lines
5.3 KiB
Python
151 lines
5.3 KiB
Python
from django.core.exceptions import ImproperlyConfigured
|
|
from django.utils.importlib import import_module
|
|
|
|
# Cache of actual callables.
|
|
_standard_context_processors = None
|
|
# We need the CSRF processor no matter what the user has in their settings,
|
|
# because otherwise it is a security vulnerability, and we can't afford to leave
|
|
# this to human error or failure to read migration instructions.
|
|
_builtin_context_processors = ('django.core.context_processors.csrf',)
|
|
|
|
class ContextPopException(Exception):
|
|
"pop() has been called more times than push()"
|
|
pass
|
|
|
|
class BaseContext(object):
|
|
def __init__(self, dict_=None):
|
|
dict_ = dict_ or {}
|
|
self.dicts = [dict_]
|
|
|
|
def __repr__(self):
|
|
return repr(self.dicts)
|
|
|
|
def __iter__(self):
|
|
for d in reversed(self.dicts):
|
|
yield d
|
|
|
|
def push(self):
|
|
d = {}
|
|
self.dicts.append(d)
|
|
return d
|
|
|
|
def pop(self):
|
|
if len(self.dicts) == 1:
|
|
raise ContextPopException
|
|
return self.dicts.pop()
|
|
|
|
def __setitem__(self, key, value):
|
|
"Set a variable in the current context"
|
|
self.dicts[-1][key] = value
|
|
|
|
def __getitem__(self, key):
|
|
"Get a variable's value, starting at the current context and going upward"
|
|
for d in reversed(self.dicts):
|
|
if key in d:
|
|
return d[key]
|
|
raise KeyError(key)
|
|
|
|
def __delitem__(self, key):
|
|
"Delete a variable from the current context"
|
|
del self.dicts[-1][key]
|
|
|
|
def has_key(self, key):
|
|
for d in self.dicts:
|
|
if key in d:
|
|
return True
|
|
return False
|
|
|
|
def __contains__(self, key):
|
|
return self.has_key(key)
|
|
|
|
def get(self, key, otherwise=None):
|
|
for d in reversed(self.dicts):
|
|
if key in d:
|
|
return d[key]
|
|
return otherwise
|
|
|
|
class Context(BaseContext):
|
|
"A stack container for variable context"
|
|
def __init__(self, dict_=None, autoescape=True, current_app=None, use_l10n=None):
|
|
self.autoescape = autoescape
|
|
self.use_l10n = use_l10n
|
|
self.current_app = current_app
|
|
self.render_context = RenderContext()
|
|
super(Context, self).__init__(dict_)
|
|
|
|
def update(self, other_dict):
|
|
"Like dict.update(). Pushes an entire dictionary's keys and values onto the context."
|
|
if not hasattr(other_dict, '__getitem__'):
|
|
raise TypeError('other_dict must be a mapping (dictionary-like) object.')
|
|
self.dicts.append(other_dict)
|
|
return other_dict
|
|
|
|
class RenderContext(BaseContext):
|
|
"""
|
|
A stack container for storing Template state.
|
|
|
|
RenderContext simplifies the implementation of template Nodes by providing a
|
|
safe place to store state between invocations of a node's `render` method.
|
|
|
|
The RenderContext also provides scoping rules that are more sensible for
|
|
'template local' variables. The render context stack is pushed before each
|
|
template is rendered, creating a fresh scope with nothing in it. Name
|
|
resolution fails if a variable is not found at the top of the RequestContext
|
|
stack. Thus, variables are local to a specific template and don't affect the
|
|
rendering of other templates as they would if they were stored in the normal
|
|
template context.
|
|
"""
|
|
def __iter__(self):
|
|
for d in self.dicts[-1]:
|
|
yield d
|
|
|
|
def has_key(self, key):
|
|
return key in self.dicts[-1]
|
|
|
|
def get(self, key, otherwise=None):
|
|
d = self.dicts[-1]
|
|
if key in d:
|
|
return d[key]
|
|
return otherwise
|
|
|
|
# This is a function rather than module-level procedural code because we only
|
|
# want it to execute if somebody uses RequestContext.
|
|
def get_standard_processors():
|
|
from django.conf import settings
|
|
global _standard_context_processors
|
|
if _standard_context_processors is None:
|
|
processors = []
|
|
collect = []
|
|
collect.extend(_builtin_context_processors)
|
|
collect.extend(settings.TEMPLATE_CONTEXT_PROCESSORS)
|
|
for path in collect:
|
|
i = path.rfind('.')
|
|
module, attr = path[:i], path[i+1:]
|
|
try:
|
|
mod = import_module(module)
|
|
except ImportError, e:
|
|
raise ImproperlyConfigured('Error importing request processor module %s: "%s"' % (module, e))
|
|
try:
|
|
func = getattr(mod, attr)
|
|
except AttributeError:
|
|
raise ImproperlyConfigured('Module "%s" does not define a "%s" callable request processor' % (module, attr))
|
|
processors.append(func)
|
|
_standard_context_processors = tuple(processors)
|
|
return _standard_context_processors
|
|
|
|
class RequestContext(Context):
|
|
"""
|
|
This subclass of template.Context automatically populates itself using
|
|
the processors defined in TEMPLATE_CONTEXT_PROCESSORS.
|
|
Additional processors can be specified as a list of callables
|
|
using the "processors" keyword argument.
|
|
"""
|
|
def __init__(self, request, dict=None, processors=None, current_app=None, use_l10n=None):
|
|
Context.__init__(self, dict, current_app=current_app, use_l10n=use_l10n)
|
|
if processors is None:
|
|
processors = ()
|
|
else:
|
|
processors = tuple(processors)
|
|
for processor in get_standard_processors() + processors:
|
|
self.update(processor(request))
|