|
1 """WSGI request adapter for cubicweb |
|
2 |
|
3 NOTE: each docstring tagged with ``COME FROM DJANGO`` means that |
|
4 the code has been taken (or adapted) from Djanco source code : |
|
5 http://www.djangoproject.com/ |
|
6 |
|
7 :organization: Logilab |
|
8 :copyright: 2008 LOGILAB S.A. (Paris, FRANCE), all rights reserved. |
|
9 :contact: http://www.logilab.fr/ -- mailto:contact@logilab.fr |
|
10 """ |
|
11 |
|
12 __docformat__ = "restructuredtext en" |
|
13 |
|
14 from StringIO import StringIO |
|
15 from urllib import quote |
|
16 |
|
17 from logilab.common.decorators import cached |
|
18 |
|
19 from cubicweb.web.request import CubicWebRequestBase |
|
20 from cubicweb.wsgi import (pformat, qs2dict, safe_copyfileobj, parse_file_upload, |
|
21 normalize_header) |
|
22 |
|
23 |
|
24 |
|
25 class CubicWebWsgiRequest(CubicWebRequestBase): |
|
26 """most of this code COMES FROM DJANO |
|
27 """ |
|
28 |
|
29 def __init__(self, environ, vreg, base_url=None): |
|
30 self.environ = environ |
|
31 self.path = environ['PATH_INFO'] |
|
32 self.method = environ['REQUEST_METHOD'].upper() |
|
33 self._headers = dict([(normalize_header(k[5:]), v) for k, v in self.environ.items() |
|
34 if k.startswith('HTTP_')]) |
|
35 https = environ.get("HTTPS") in ('yes', 'on', '1') |
|
36 self._base_url = base_url or self.application_uri() |
|
37 post, files = self.get_posted_data() |
|
38 super(CubicWebWsgiRequest, self).__init__(vreg, https, post) |
|
39 if files is not None: |
|
40 for fdef in files.itervalues(): |
|
41 fdef[0] = unicode(fdef[0], self.encoding) |
|
42 self.form.update(files) |
|
43 # prepare output headers |
|
44 self.headers_out = {} |
|
45 |
|
46 def __repr__(self): |
|
47 # Since this is called as part of error handling, we need to be very |
|
48 # robust against potentially malformed input. |
|
49 form = pformat(self.form) |
|
50 meta = pformat(self.environ) |
|
51 return '<CubicWebWsgiRequest\FORM:%s,\nMETA:%s>' % \ |
|
52 (form, meta) |
|
53 |
|
54 ## cubicweb request interface ################################################ |
|
55 |
|
56 def base_url(self): |
|
57 return self._base_url |
|
58 |
|
59 def http_method(self): |
|
60 """returns 'POST', 'GET', 'HEAD', etc.""" |
|
61 return self.method |
|
62 |
|
63 def relative_path(self, includeparams=True): |
|
64 """return the normalized path of the request (ie at least relative |
|
65 to the application's root, but some other normalization may be needed |
|
66 so that the returned path may be used to compare to generated urls |
|
67 |
|
68 :param includeparams: |
|
69 boolean indicating if GET form parameters should be kept in the path |
|
70 """ |
|
71 path = self.environ['PATH_INFO'] |
|
72 path = path[1:] # remove leading '/' |
|
73 if includeparams: |
|
74 qs = self.environ.get('QUERY_STRING') |
|
75 if qs: |
|
76 return '%s?%s' % (path, qs) |
|
77 |
|
78 return path |
|
79 |
|
80 def get_header(self, header, default=None): |
|
81 """return the value associated with the given input HTTP header, |
|
82 raise KeyError if the header is not set |
|
83 """ |
|
84 return self._headers.get(normalize_header(header), default) |
|
85 |
|
86 def set_header(self, header, value, raw=True): |
|
87 """set an output HTTP header""" |
|
88 assert raw, "don't know anything about non-raw headers for wsgi requests" |
|
89 self.headers_out[header] = value |
|
90 |
|
91 def add_header(self, header, value): |
|
92 """add an output HTTP header""" |
|
93 self.headers_out[header] = value |
|
94 |
|
95 def remove_header(self, header): |
|
96 """remove an output HTTP header""" |
|
97 self.headers_out.pop(header, None) |
|
98 |
|
99 def header_if_modified_since(self): |
|
100 """If the HTTP header If-modified-since is set, return the equivalent |
|
101 mx date time value (GMT), else return None |
|
102 """ |
|
103 return None |
|
104 |
|
105 ## wsgi request helpers ################################################### |
|
106 |
|
107 def application_uri(self): |
|
108 """Return the application's base URI (no PATH_INFO or QUERY_STRING) |
|
109 |
|
110 see python2.5's wsgiref.util.application_uri code |
|
111 """ |
|
112 environ = self.environ |
|
113 url = environ['wsgi.url_scheme'] + '://' |
|
114 if environ.get('HTTP_HOST'): |
|
115 url += environ['HTTP_HOST'] |
|
116 else: |
|
117 url += environ['SERVER_NAME'] |
|
118 if environ['wsgi.url_scheme'] == 'https': |
|
119 if environ['SERVER_PORT'] != '443': |
|
120 url += ':' + environ['SERVER_PORT'] |
|
121 else: |
|
122 if environ['SERVER_PORT'] != '80': |
|
123 url += ':' + environ['SERVER_PORT'] |
|
124 url += quote(environ.get('SCRIPT_NAME') or '/') |
|
125 return url |
|
126 |
|
127 def get_full_path(self): |
|
128 return '%s%s' % (self.path, self.environ.get('QUERY_STRING', '') and ('?' + self.environ.get('QUERY_STRING', '')) or '') |
|
129 |
|
130 def is_secure(self): |
|
131 return 'wsgi.url_scheme' in self.environ \ |
|
132 and self.environ['wsgi.url_scheme'] == 'https' |
|
133 |
|
134 def get_posted_data(self): |
|
135 files = None |
|
136 if self.method == 'POST': |
|
137 if self.environ.get('CONTENT_TYPE', '').startswith('multipart'): |
|
138 header_dict = dict((normalize_header(k[5:]), v) |
|
139 for k, v in self.environ.items() |
|
140 if k.startswith('HTTP_')) |
|
141 header_dict['Content-Type'] = self.environ.get('CONTENT_TYPE', '') |
|
142 post, files = parse_file_upload(header_dict, self.raw_post_data) |
|
143 else: |
|
144 post = qs2dict(self.raw_post_data) |
|
145 else: |
|
146 # The WSGI spec says 'QUERY_STRING' may be absent. |
|
147 post = qs2dict(self.environ.get('QUERY_STRING', '')) |
|
148 return post, files |
|
149 |
|
150 @property |
|
151 @cached |
|
152 def raw_post_data(self): |
|
153 buf = StringIO() |
|
154 try: |
|
155 # CONTENT_LENGTH might be absent if POST doesn't have content at all (lighttpd) |
|
156 content_length = int(self.environ.get('CONTENT_LENGTH', 0)) |
|
157 except ValueError: # if CONTENT_LENGTH was empty string or not an integer |
|
158 content_length = 0 |
|
159 if content_length > 0: |
|
160 safe_copyfileobj(self.environ['wsgi.input'], buf, |
|
161 size=content_length) |
|
162 postdata = buf.getvalue() |
|
163 buf.close() |
|
164 return postdata |
|
165 |
|
166 def _validate_cache(self): |
|
167 """raise a `DirectResponse` exception if a cached page along the way |
|
168 exists and is still usable |
|
169 """ |
|
170 # XXX |
|
171 # if self.get_header('Cache-Control') in ('max-age=0', 'no-cache'): |
|
172 # # Expires header seems to be required by IE7 |
|
173 # self.add_header('Expires', 'Sat, 01 Jan 2000 00:00:00 GMT') |
|
174 # return |
|
175 # try: |
|
176 # http.checkPreconditions(self._twreq, _PreResponse(self)) |
|
177 # except http.HTTPError, ex: |
|
178 # self.info('valid http cache, no actual rendering') |
|
179 # raise DirectResponse(ex.response) |
|
180 # Expires header seems to be required by IE7 |
|
181 self.add_header('Expires', 'Sat, 01 Jan 2000 00:00:00 GMT') |