Show More
@@ -1,623 +1,623 b'' | |||||
1 | # hgweb/request.py - An http request from either CGI or the standalone server. |
|
1 | # hgweb/request.py - An http request from either CGI or the standalone server. | |
2 | # |
|
2 | # | |
3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
3 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> | |
4 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> |
|
4 | # Copyright 2005, 2006 Matt Mackall <mpm@selenic.com> | |
5 | # |
|
5 | # | |
6 | # This software may be used and distributed according to the terms of the |
|
6 | # This software may be used and distributed according to the terms of the | |
7 | # GNU General Public License version 2 or any later version. |
|
7 | # GNU General Public License version 2 or any later version. | |
8 |
|
8 | |||
9 | from __future__ import absolute_import |
|
9 | from __future__ import absolute_import | |
10 |
|
10 | |||
11 | # import wsgiref.validate |
|
11 | # import wsgiref.validate | |
12 |
|
12 | |||
13 | from ..thirdparty import attr |
|
13 | from ..thirdparty import attr | |
14 | from .. import ( |
|
14 | from .. import ( | |
15 | error, |
|
15 | error, | |
16 | pycompat, |
|
16 | pycompat, | |
17 | util, |
|
17 | util, | |
18 | ) |
|
18 | ) | |
19 |
|
19 | |||
20 |
|
20 | |||
21 | class multidict(object): |
|
21 | class multidict(object): | |
22 | """A dict like object that can store multiple values for a key. |
|
22 | """A dict like object that can store multiple values for a key. | |
23 |
|
23 | |||
24 | Used to store parsed request parameters. |
|
24 | Used to store parsed request parameters. | |
25 |
|
25 | |||
26 | This is inspired by WebOb's class of the same name. |
|
26 | This is inspired by WebOb's class of the same name. | |
27 | """ |
|
27 | """ | |
28 |
|
28 | |||
29 | def __init__(self): |
|
29 | def __init__(self): | |
30 | self._items = {} |
|
30 | self._items = {} | |
31 |
|
31 | |||
32 | def __getitem__(self, key): |
|
32 | def __getitem__(self, key): | |
33 | """Returns the last set value for a key.""" |
|
33 | """Returns the last set value for a key.""" | |
34 | return self._items[key][-1] |
|
34 | return self._items[key][-1] | |
35 |
|
35 | |||
36 | def __setitem__(self, key, value): |
|
36 | def __setitem__(self, key, value): | |
37 | """Replace a values for a key with a new value.""" |
|
37 | """Replace a values for a key with a new value.""" | |
38 | self._items[key] = [value] |
|
38 | self._items[key] = [value] | |
39 |
|
39 | |||
40 | def __delitem__(self, key): |
|
40 | def __delitem__(self, key): | |
41 | """Delete all values for a key.""" |
|
41 | """Delete all values for a key.""" | |
42 | del self._items[key] |
|
42 | del self._items[key] | |
43 |
|
43 | |||
44 | def __contains__(self, key): |
|
44 | def __contains__(self, key): | |
45 | return key in self._items |
|
45 | return key in self._items | |
46 |
|
46 | |||
47 | def __len__(self): |
|
47 | def __len__(self): | |
48 | return len(self._items) |
|
48 | return len(self._items) | |
49 |
|
49 | |||
50 | def get(self, key, default=None): |
|
50 | def get(self, key, default=None): | |
51 | try: |
|
51 | try: | |
52 | return self.__getitem__(key) |
|
52 | return self.__getitem__(key) | |
53 | except KeyError: |
|
53 | except KeyError: | |
54 | return default |
|
54 | return default | |
55 |
|
55 | |||
56 | def add(self, key, value): |
|
56 | def add(self, key, value): | |
57 | """Add a new value for a key. Does not replace existing values.""" |
|
57 | """Add a new value for a key. Does not replace existing values.""" | |
58 | self._items.setdefault(key, []).append(value) |
|
58 | self._items.setdefault(key, []).append(value) | |
59 |
|
59 | |||
60 | def getall(self, key): |
|
60 | def getall(self, key): | |
61 | """Obtains all values for a key.""" |
|
61 | """Obtains all values for a key.""" | |
62 | return self._items.get(key, []) |
|
62 | return self._items.get(key, []) | |
63 |
|
63 | |||
64 | def getone(self, key): |
|
64 | def getone(self, key): | |
65 | """Obtain a single value for a key. |
|
65 | """Obtain a single value for a key. | |
66 |
|
66 | |||
67 | Raises KeyError if key not defined or it has multiple values set. |
|
67 | Raises KeyError if key not defined or it has multiple values set. | |
68 | """ |
|
68 | """ | |
69 | vals = self._items[key] |
|
69 | vals = self._items[key] | |
70 |
|
70 | |||
71 | if len(vals) > 1: |
|
71 | if len(vals) > 1: | |
72 | raise KeyError(b'multiple values for %r' % key) |
|
72 | raise KeyError(b'multiple values for %r' % key) | |
73 |
|
73 | |||
74 | return vals[0] |
|
74 | return vals[0] | |
75 |
|
75 | |||
76 | def asdictoflists(self): |
|
76 | def asdictoflists(self): | |
77 | return {k: list(v) for k, v in pycompat.iteritems(self._items)} |
|
77 | return {k: list(v) for k, v in pycompat.iteritems(self._items)} | |
78 |
|
78 | |||
79 |
|
79 | |||
80 | @attr.s(frozen=True) |
|
80 | @attr.s(frozen=True) | |
81 | class parsedrequest(object): |
|
81 | class parsedrequest(object): | |
82 | """Represents a parsed WSGI request. |
|
82 | """Represents a parsed WSGI request. | |
83 |
|
83 | |||
84 | Contains both parsed parameters as well as a handle on the input stream. |
|
84 | Contains both parsed parameters as well as a handle on the input stream. | |
85 | """ |
|
85 | """ | |
86 |
|
86 | |||
87 | # Request method. |
|
87 | # Request method. | |
88 | method = attr.ib() |
|
88 | method = attr.ib() | |
89 | # Full URL for this request. |
|
89 | # Full URL for this request. | |
90 | url = attr.ib() |
|
90 | url = attr.ib() | |
91 | # URL without any path components. Just <proto>://<host><port>. |
|
91 | # URL without any path components. Just <proto>://<host><port>. | |
92 | baseurl = attr.ib() |
|
92 | baseurl = attr.ib() | |
93 | # Advertised URL. Like ``url`` and ``baseurl`` but uses SERVER_NAME instead |
|
93 | # Advertised URL. Like ``url`` and ``baseurl`` but uses SERVER_NAME instead | |
94 | # of HTTP: Host header for hostname. This is likely what clients used. |
|
94 | # of HTTP: Host header for hostname. This is likely what clients used. | |
95 | advertisedurl = attr.ib() |
|
95 | advertisedurl = attr.ib() | |
96 | advertisedbaseurl = attr.ib() |
|
96 | advertisedbaseurl = attr.ib() | |
97 | # URL scheme (part before ``://``). e.g. ``http`` or ``https``. |
|
97 | # URL scheme (part before ``://``). e.g. ``http`` or ``https``. | |
98 | urlscheme = attr.ib() |
|
98 | urlscheme = attr.ib() | |
99 | # Value of REMOTE_USER, if set, or None. |
|
99 | # Value of REMOTE_USER, if set, or None. | |
100 | remoteuser = attr.ib() |
|
100 | remoteuser = attr.ib() | |
101 | # Value of REMOTE_HOST, if set, or None. |
|
101 | # Value of REMOTE_HOST, if set, or None. | |
102 | remotehost = attr.ib() |
|
102 | remotehost = attr.ib() | |
103 | # Relative WSGI application path. If defined, will begin with a |
|
103 | # Relative WSGI application path. If defined, will begin with a | |
104 | # ``/``. |
|
104 | # ``/``. | |
105 | apppath = attr.ib() |
|
105 | apppath = attr.ib() | |
106 | # List of path parts to be used for dispatch. |
|
106 | # List of path parts to be used for dispatch. | |
107 | dispatchparts = attr.ib() |
|
107 | dispatchparts = attr.ib() | |
108 | # URL path component (no query string) used for dispatch. Can be |
|
108 | # URL path component (no query string) used for dispatch. Can be | |
109 | # ``None`` to signal no path component given to the request, an |
|
109 | # ``None`` to signal no path component given to the request, an | |
110 | # empty string to signal a request to the application's root URL, |
|
110 | # empty string to signal a request to the application's root URL, | |
111 | # or a string not beginning with ``/`` containing the requested |
|
111 | # or a string not beginning with ``/`` containing the requested | |
112 | # path under the application. |
|
112 | # path under the application. | |
113 | dispatchpath = attr.ib() |
|
113 | dispatchpath = attr.ib() | |
114 | # The name of the repository being accessed. |
|
114 | # The name of the repository being accessed. | |
115 | reponame = attr.ib() |
|
115 | reponame = attr.ib() | |
116 | # Raw query string (part after "?" in URL). |
|
116 | # Raw query string (part after "?" in URL). | |
117 | querystring = attr.ib() |
|
117 | querystring = attr.ib() | |
118 | # multidict of query string parameters. |
|
118 | # multidict of query string parameters. | |
119 | qsparams = attr.ib() |
|
119 | qsparams = attr.ib() | |
120 | # wsgiref.headers.Headers instance. Operates like a dict with case |
|
120 | # wsgiref.headers.Headers instance. Operates like a dict with case | |
121 | # insensitive keys. |
|
121 | # insensitive keys. | |
122 | headers = attr.ib() |
|
122 | headers = attr.ib() | |
123 | # Request body input stream. |
|
123 | # Request body input stream. | |
124 | bodyfh = attr.ib() |
|
124 | bodyfh = attr.ib() | |
125 | # WSGI environment dict, unmodified. |
|
125 | # WSGI environment dict, unmodified. | |
126 | rawenv = attr.ib() |
|
126 | rawenv = attr.ib() | |
127 |
|
127 | |||
128 |
|
128 | |||
129 | def parserequestfromenv(env, reponame=None, altbaseurl=None, bodyfh=None): |
|
129 | def parserequestfromenv(env, reponame=None, altbaseurl=None, bodyfh=None): | |
130 | """Parse URL components from environment variables. |
|
130 | """Parse URL components from environment variables. | |
131 |
|
131 | |||
132 | WSGI defines request attributes via environment variables. This function |
|
132 | WSGI defines request attributes via environment variables. This function | |
133 | parses the environment variables into a data structure. |
|
133 | parses the environment variables into a data structure. | |
134 |
|
134 | |||
135 | If ``reponame`` is defined, the leading path components matching that |
|
135 | If ``reponame`` is defined, the leading path components matching that | |
136 | string are effectively shifted from ``PATH_INFO`` to ``SCRIPT_NAME``. |
|
136 | string are effectively shifted from ``PATH_INFO`` to ``SCRIPT_NAME``. | |
137 | This simulates the world view of a WSGI application that processes |
|
137 | This simulates the world view of a WSGI application that processes | |
138 | requests from the base URL of a repo. |
|
138 | requests from the base URL of a repo. | |
139 |
|
139 | |||
140 | If ``altbaseurl`` (typically comes from ``web.baseurl`` config option) |
|
140 | If ``altbaseurl`` (typically comes from ``web.baseurl`` config option) | |
141 | is defined, it is used - instead of the WSGI environment variables - for |
|
141 | is defined, it is used - instead of the WSGI environment variables - for | |
142 | constructing URL components up to and including the WSGI application path. |
|
142 | constructing URL components up to and including the WSGI application path. | |
143 | For example, if the current WSGI application is at ``/repo`` and a request |
|
143 | For example, if the current WSGI application is at ``/repo`` and a request | |
144 | is made to ``/rev/@`` with this argument set to |
|
144 | is made to ``/rev/@`` with this argument set to | |
145 | ``http://myserver:9000/prefix``, the URL and path components will resolve as |
|
145 | ``http://myserver:9000/prefix``, the URL and path components will resolve as | |
146 | if the request were to ``http://myserver:9000/prefix/rev/@``. In other |
|
146 | if the request were to ``http://myserver:9000/prefix/rev/@``. In other | |
147 | words, ``wsgi.url_scheme``, ``SERVER_NAME``, ``SERVER_PORT``, and |
|
147 | words, ``wsgi.url_scheme``, ``SERVER_NAME``, ``SERVER_PORT``, and | |
148 | ``SCRIPT_NAME`` are all effectively replaced by components from this URL. |
|
148 | ``SCRIPT_NAME`` are all effectively replaced by components from this URL. | |
149 |
|
149 | |||
150 | ``bodyfh`` can be used to specify a file object to read the request body |
|
150 | ``bodyfh`` can be used to specify a file object to read the request body | |
151 | from. If not defined, ``wsgi.input`` from the environment dict is used. |
|
151 | from. If not defined, ``wsgi.input`` from the environment dict is used. | |
152 | """ |
|
152 | """ | |
153 | # PEP 3333 defines the WSGI spec and is a useful reference for this code. |
|
153 | # PEP 3333 defines the WSGI spec and is a useful reference for this code. | |
154 |
|
154 | |||
155 | # We first validate that the incoming object conforms with the WSGI spec. |
|
155 | # We first validate that the incoming object conforms with the WSGI spec. | |
156 | # We only want to be dealing with spec-conforming WSGI implementations. |
|
156 | # We only want to be dealing with spec-conforming WSGI implementations. | |
157 | # TODO enable this once we fix internal violations. |
|
157 | # TODO enable this once we fix internal violations. | |
158 | # wsgiref.validate.check_environ(env) |
|
158 | # wsgiref.validate.check_environ(env) | |
159 |
|
159 | |||
160 | # PEP-0333 states that environment keys and values are native strings |
|
160 | # PEP-0333 states that environment keys and values are native strings | |
161 | # (bytes on Python 2 and str on Python 3). The code points for the Unicode |
|
161 | # (bytes on Python 2 and str on Python 3). The code points for the Unicode | |
162 | # strings on Python 3 must be between \00000-\000FF. We deal with bytes |
|
162 | # strings on Python 3 must be between \00000-\000FF. We deal with bytes | |
163 | # in Mercurial, so mass convert string keys and values to bytes. |
|
163 | # in Mercurial, so mass convert string keys and values to bytes. | |
164 | if pycompat.ispy3: |
|
164 | if pycompat.ispy3: | |
165 | env = {k.encode('latin-1'): v for k, v in pycompat.iteritems(env)} |
|
165 | def tobytes(s): | |
166 | env = { |
|
166 | if not isinstance(s, str): | |
167 | k: v.encode('latin-1') if isinstance(v, str) else v |
|
167 | return s | |
168 | for k, v in pycompat.iteritems(env) |
|
168 | return s.encode('latin-1') | |
169 | } |
|
169 | env = {tobytes(k): tobytes(v) for k, v in pycompat.iteritems(env)} | |
170 |
|
170 | |||
171 | # Some hosting solutions are emulating hgwebdir, and dispatching directly |
|
171 | # Some hosting solutions are emulating hgwebdir, and dispatching directly | |
172 | # to an hgweb instance using this environment variable. This was always |
|
172 | # to an hgweb instance using this environment variable. This was always | |
173 | # checked prior to d7fd203e36cc; keep doing so to avoid breaking them. |
|
173 | # checked prior to d7fd203e36cc; keep doing so to avoid breaking them. | |
174 | if not reponame: |
|
174 | if not reponame: | |
175 | reponame = env.get(b'REPO_NAME') |
|
175 | reponame = env.get(b'REPO_NAME') | |
176 |
|
176 | |||
177 | if altbaseurl: |
|
177 | if altbaseurl: | |
178 | altbaseurl = util.url(altbaseurl) |
|
178 | altbaseurl = util.url(altbaseurl) | |
179 |
|
179 | |||
180 | # https://www.python.org/dev/peps/pep-0333/#environ-variables defines |
|
180 | # https://www.python.org/dev/peps/pep-0333/#environ-variables defines | |
181 | # the environment variables. |
|
181 | # the environment variables. | |
182 | # https://www.python.org/dev/peps/pep-0333/#url-reconstruction defines |
|
182 | # https://www.python.org/dev/peps/pep-0333/#url-reconstruction defines | |
183 | # how URLs are reconstructed. |
|
183 | # how URLs are reconstructed. | |
184 | fullurl = env[b'wsgi.url_scheme'] + b'://' |
|
184 | fullurl = env[b'wsgi.url_scheme'] + b'://' | |
185 |
|
185 | |||
186 | if altbaseurl and altbaseurl.scheme: |
|
186 | if altbaseurl and altbaseurl.scheme: | |
187 | advertisedfullurl = altbaseurl.scheme + b'://' |
|
187 | advertisedfullurl = altbaseurl.scheme + b'://' | |
188 | else: |
|
188 | else: | |
189 | advertisedfullurl = fullurl |
|
189 | advertisedfullurl = fullurl | |
190 |
|
190 | |||
191 | def addport(s, port): |
|
191 | def addport(s, port): | |
192 | if s.startswith(b'https://'): |
|
192 | if s.startswith(b'https://'): | |
193 | if port != b'443': |
|
193 | if port != b'443': | |
194 | s += b':' + port |
|
194 | s += b':' + port | |
195 | else: |
|
195 | else: | |
196 | if port != b'80': |
|
196 | if port != b'80': | |
197 | s += b':' + port |
|
197 | s += b':' + port | |
198 |
|
198 | |||
199 | return s |
|
199 | return s | |
200 |
|
200 | |||
201 | if env.get(b'HTTP_HOST'): |
|
201 | if env.get(b'HTTP_HOST'): | |
202 | fullurl += env[b'HTTP_HOST'] |
|
202 | fullurl += env[b'HTTP_HOST'] | |
203 | else: |
|
203 | else: | |
204 | fullurl += env[b'SERVER_NAME'] |
|
204 | fullurl += env[b'SERVER_NAME'] | |
205 | fullurl = addport(fullurl, env[b'SERVER_PORT']) |
|
205 | fullurl = addport(fullurl, env[b'SERVER_PORT']) | |
206 |
|
206 | |||
207 | if altbaseurl and altbaseurl.host: |
|
207 | if altbaseurl and altbaseurl.host: | |
208 | advertisedfullurl += altbaseurl.host |
|
208 | advertisedfullurl += altbaseurl.host | |
209 |
|
209 | |||
210 | if altbaseurl.port: |
|
210 | if altbaseurl.port: | |
211 | port = altbaseurl.port |
|
211 | port = altbaseurl.port | |
212 | elif altbaseurl.scheme == b'http' and not altbaseurl.port: |
|
212 | elif altbaseurl.scheme == b'http' and not altbaseurl.port: | |
213 | port = b'80' |
|
213 | port = b'80' | |
214 | elif altbaseurl.scheme == b'https' and not altbaseurl.port: |
|
214 | elif altbaseurl.scheme == b'https' and not altbaseurl.port: | |
215 | port = b'443' |
|
215 | port = b'443' | |
216 | else: |
|
216 | else: | |
217 | port = env[b'SERVER_PORT'] |
|
217 | port = env[b'SERVER_PORT'] | |
218 |
|
218 | |||
219 | advertisedfullurl = addport(advertisedfullurl, port) |
|
219 | advertisedfullurl = addport(advertisedfullurl, port) | |
220 | else: |
|
220 | else: | |
221 | advertisedfullurl += env[b'SERVER_NAME'] |
|
221 | advertisedfullurl += env[b'SERVER_NAME'] | |
222 | advertisedfullurl = addport(advertisedfullurl, env[b'SERVER_PORT']) |
|
222 | advertisedfullurl = addport(advertisedfullurl, env[b'SERVER_PORT']) | |
223 |
|
223 | |||
224 | baseurl = fullurl |
|
224 | baseurl = fullurl | |
225 | advertisedbaseurl = advertisedfullurl |
|
225 | advertisedbaseurl = advertisedfullurl | |
226 |
|
226 | |||
227 | fullurl += util.urlreq.quote(env.get(b'SCRIPT_NAME', b'')) |
|
227 | fullurl += util.urlreq.quote(env.get(b'SCRIPT_NAME', b'')) | |
228 | fullurl += util.urlreq.quote(env.get(b'PATH_INFO', b'')) |
|
228 | fullurl += util.urlreq.quote(env.get(b'PATH_INFO', b'')) | |
229 |
|
229 | |||
230 | if altbaseurl: |
|
230 | if altbaseurl: | |
231 | path = altbaseurl.path or b'' |
|
231 | path = altbaseurl.path or b'' | |
232 | if path and not path.startswith(b'/'): |
|
232 | if path and not path.startswith(b'/'): | |
233 | path = b'/' + path |
|
233 | path = b'/' + path | |
234 | advertisedfullurl += util.urlreq.quote(path) |
|
234 | advertisedfullurl += util.urlreq.quote(path) | |
235 | else: |
|
235 | else: | |
236 | advertisedfullurl += util.urlreq.quote(env.get(b'SCRIPT_NAME', b'')) |
|
236 | advertisedfullurl += util.urlreq.quote(env.get(b'SCRIPT_NAME', b'')) | |
237 |
|
237 | |||
238 | advertisedfullurl += util.urlreq.quote(env.get(b'PATH_INFO', b'')) |
|
238 | advertisedfullurl += util.urlreq.quote(env.get(b'PATH_INFO', b'')) | |
239 |
|
239 | |||
240 | if env.get(b'QUERY_STRING'): |
|
240 | if env.get(b'QUERY_STRING'): | |
241 | fullurl += b'?' + env[b'QUERY_STRING'] |
|
241 | fullurl += b'?' + env[b'QUERY_STRING'] | |
242 | advertisedfullurl += b'?' + env[b'QUERY_STRING'] |
|
242 | advertisedfullurl += b'?' + env[b'QUERY_STRING'] | |
243 |
|
243 | |||
244 | # If ``reponame`` is defined, that must be a prefix on PATH_INFO |
|
244 | # If ``reponame`` is defined, that must be a prefix on PATH_INFO | |
245 | # that represents the repository being dispatched to. When computing |
|
245 | # that represents the repository being dispatched to. When computing | |
246 | # the dispatch info, we ignore these leading path components. |
|
246 | # the dispatch info, we ignore these leading path components. | |
247 |
|
247 | |||
248 | if altbaseurl: |
|
248 | if altbaseurl: | |
249 | apppath = altbaseurl.path or b'' |
|
249 | apppath = altbaseurl.path or b'' | |
250 | if apppath and not apppath.startswith(b'/'): |
|
250 | if apppath and not apppath.startswith(b'/'): | |
251 | apppath = b'/' + apppath |
|
251 | apppath = b'/' + apppath | |
252 | else: |
|
252 | else: | |
253 | apppath = env.get(b'SCRIPT_NAME', b'') |
|
253 | apppath = env.get(b'SCRIPT_NAME', b'') | |
254 |
|
254 | |||
255 | if reponame: |
|
255 | if reponame: | |
256 | repoprefix = b'/' + reponame.strip(b'/') |
|
256 | repoprefix = b'/' + reponame.strip(b'/') | |
257 |
|
257 | |||
258 | if not env.get(b'PATH_INFO'): |
|
258 | if not env.get(b'PATH_INFO'): | |
259 | raise error.ProgrammingError(b'reponame requires PATH_INFO') |
|
259 | raise error.ProgrammingError(b'reponame requires PATH_INFO') | |
260 |
|
260 | |||
261 | if not env[b'PATH_INFO'].startswith(repoprefix): |
|
261 | if not env[b'PATH_INFO'].startswith(repoprefix): | |
262 | raise error.ProgrammingError( |
|
262 | raise error.ProgrammingError( | |
263 | b'PATH_INFO does not begin with repo ' |
|
263 | b'PATH_INFO does not begin with repo ' | |
264 | b'name: %s (%s)' % (env[b'PATH_INFO'], reponame) |
|
264 | b'name: %s (%s)' % (env[b'PATH_INFO'], reponame) | |
265 | ) |
|
265 | ) | |
266 |
|
266 | |||
267 | dispatchpath = env[b'PATH_INFO'][len(repoprefix) :] |
|
267 | dispatchpath = env[b'PATH_INFO'][len(repoprefix) :] | |
268 |
|
268 | |||
269 | if dispatchpath and not dispatchpath.startswith(b'/'): |
|
269 | if dispatchpath and not dispatchpath.startswith(b'/'): | |
270 | raise error.ProgrammingError( |
|
270 | raise error.ProgrammingError( | |
271 | b'reponame prefix of PATH_INFO does ' |
|
271 | b'reponame prefix of PATH_INFO does ' | |
272 | b'not end at path delimiter: %s (%s)' |
|
272 | b'not end at path delimiter: %s (%s)' | |
273 | % (env[b'PATH_INFO'], reponame) |
|
273 | % (env[b'PATH_INFO'], reponame) | |
274 | ) |
|
274 | ) | |
275 |
|
275 | |||
276 | apppath = apppath.rstrip(b'/') + repoprefix |
|
276 | apppath = apppath.rstrip(b'/') + repoprefix | |
277 | dispatchparts = dispatchpath.strip(b'/').split(b'/') |
|
277 | dispatchparts = dispatchpath.strip(b'/').split(b'/') | |
278 | dispatchpath = b'/'.join(dispatchparts) |
|
278 | dispatchpath = b'/'.join(dispatchparts) | |
279 |
|
279 | |||
280 | elif b'PATH_INFO' in env: |
|
280 | elif b'PATH_INFO' in env: | |
281 | if env[b'PATH_INFO'].strip(b'/'): |
|
281 | if env[b'PATH_INFO'].strip(b'/'): | |
282 | dispatchparts = env[b'PATH_INFO'].strip(b'/').split(b'/') |
|
282 | dispatchparts = env[b'PATH_INFO'].strip(b'/').split(b'/') | |
283 | dispatchpath = b'/'.join(dispatchparts) |
|
283 | dispatchpath = b'/'.join(dispatchparts) | |
284 | else: |
|
284 | else: | |
285 | dispatchparts = [] |
|
285 | dispatchparts = [] | |
286 | dispatchpath = b'' |
|
286 | dispatchpath = b'' | |
287 | else: |
|
287 | else: | |
288 | dispatchparts = [] |
|
288 | dispatchparts = [] | |
289 | dispatchpath = None |
|
289 | dispatchpath = None | |
290 |
|
290 | |||
291 | querystring = env.get(b'QUERY_STRING', b'') |
|
291 | querystring = env.get(b'QUERY_STRING', b'') | |
292 |
|
292 | |||
293 | # We store as a list so we have ordering information. We also store as |
|
293 | # We store as a list so we have ordering information. We also store as | |
294 | # a dict to facilitate fast lookup. |
|
294 | # a dict to facilitate fast lookup. | |
295 | qsparams = multidict() |
|
295 | qsparams = multidict() | |
296 | for k, v in util.urlreq.parseqsl(querystring, keep_blank_values=True): |
|
296 | for k, v in util.urlreq.parseqsl(querystring, keep_blank_values=True): | |
297 | qsparams.add(k, v) |
|
297 | qsparams.add(k, v) | |
298 |
|
298 | |||
299 | # HTTP_* keys contain HTTP request headers. The Headers structure should |
|
299 | # HTTP_* keys contain HTTP request headers. The Headers structure should | |
300 | # perform case normalization for us. We just rewrite underscore to dash |
|
300 | # perform case normalization for us. We just rewrite underscore to dash | |
301 | # so keys match what likely went over the wire. |
|
301 | # so keys match what likely went over the wire. | |
302 | headers = [] |
|
302 | headers = [] | |
303 | for k, v in pycompat.iteritems(env): |
|
303 | for k, v in pycompat.iteritems(env): | |
304 | if k.startswith(b'HTTP_'): |
|
304 | if k.startswith(b'HTTP_'): | |
305 | headers.append((k[len(b'HTTP_') :].replace(b'_', b'-'), v)) |
|
305 | headers.append((k[len(b'HTTP_') :].replace(b'_', b'-'), v)) | |
306 |
|
306 | |||
307 | from . import wsgiheaders # avoid cycle |
|
307 | from . import wsgiheaders # avoid cycle | |
308 |
|
308 | |||
309 | headers = wsgiheaders.Headers(headers) |
|
309 | headers = wsgiheaders.Headers(headers) | |
310 |
|
310 | |||
311 | # This is kind of a lie because the HTTP header wasn't explicitly |
|
311 | # This is kind of a lie because the HTTP header wasn't explicitly | |
312 | # sent. But for all intents and purposes it should be OK to lie about |
|
312 | # sent. But for all intents and purposes it should be OK to lie about | |
313 | # this, since a consumer will either either value to determine how many |
|
313 | # this, since a consumer will either either value to determine how many | |
314 | # bytes are available to read. |
|
314 | # bytes are available to read. | |
315 | if b'CONTENT_LENGTH' in env and b'HTTP_CONTENT_LENGTH' not in env: |
|
315 | if b'CONTENT_LENGTH' in env and b'HTTP_CONTENT_LENGTH' not in env: | |
316 | headers[b'Content-Length'] = env[b'CONTENT_LENGTH'] |
|
316 | headers[b'Content-Length'] = env[b'CONTENT_LENGTH'] | |
317 |
|
317 | |||
318 | if b'CONTENT_TYPE' in env and b'HTTP_CONTENT_TYPE' not in env: |
|
318 | if b'CONTENT_TYPE' in env and b'HTTP_CONTENT_TYPE' not in env: | |
319 | headers[b'Content-Type'] = env[b'CONTENT_TYPE'] |
|
319 | headers[b'Content-Type'] = env[b'CONTENT_TYPE'] | |
320 |
|
320 | |||
321 | if bodyfh is None: |
|
321 | if bodyfh is None: | |
322 | bodyfh = env[b'wsgi.input'] |
|
322 | bodyfh = env[b'wsgi.input'] | |
323 | if b'Content-Length' in headers: |
|
323 | if b'Content-Length' in headers: | |
324 | bodyfh = util.cappedreader( |
|
324 | bodyfh = util.cappedreader( | |
325 | bodyfh, int(headers[b'Content-Length'] or b'0') |
|
325 | bodyfh, int(headers[b'Content-Length'] or b'0') | |
326 | ) |
|
326 | ) | |
327 |
|
327 | |||
328 | return parsedrequest( |
|
328 | return parsedrequest( | |
329 | method=env[b'REQUEST_METHOD'], |
|
329 | method=env[b'REQUEST_METHOD'], | |
330 | url=fullurl, |
|
330 | url=fullurl, | |
331 | baseurl=baseurl, |
|
331 | baseurl=baseurl, | |
332 | advertisedurl=advertisedfullurl, |
|
332 | advertisedurl=advertisedfullurl, | |
333 | advertisedbaseurl=advertisedbaseurl, |
|
333 | advertisedbaseurl=advertisedbaseurl, | |
334 | urlscheme=env[b'wsgi.url_scheme'], |
|
334 | urlscheme=env[b'wsgi.url_scheme'], | |
335 | remoteuser=env.get(b'REMOTE_USER'), |
|
335 | remoteuser=env.get(b'REMOTE_USER'), | |
336 | remotehost=env.get(b'REMOTE_HOST'), |
|
336 | remotehost=env.get(b'REMOTE_HOST'), | |
337 | apppath=apppath, |
|
337 | apppath=apppath, | |
338 | dispatchparts=dispatchparts, |
|
338 | dispatchparts=dispatchparts, | |
339 | dispatchpath=dispatchpath, |
|
339 | dispatchpath=dispatchpath, | |
340 | reponame=reponame, |
|
340 | reponame=reponame, | |
341 | querystring=querystring, |
|
341 | querystring=querystring, | |
342 | qsparams=qsparams, |
|
342 | qsparams=qsparams, | |
343 | headers=headers, |
|
343 | headers=headers, | |
344 | bodyfh=bodyfh, |
|
344 | bodyfh=bodyfh, | |
345 | rawenv=env, |
|
345 | rawenv=env, | |
346 | ) |
|
346 | ) | |
347 |
|
347 | |||
348 |
|
348 | |||
349 | class offsettrackingwriter(object): |
|
349 | class offsettrackingwriter(object): | |
350 | """A file object like object that is append only and tracks write count. |
|
350 | """A file object like object that is append only and tracks write count. | |
351 |
|
351 | |||
352 | Instances are bound to a callable. This callable is called with data |
|
352 | Instances are bound to a callable. This callable is called with data | |
353 | whenever a ``write()`` is attempted. |
|
353 | whenever a ``write()`` is attempted. | |
354 |
|
354 | |||
355 | Instances track the amount of written data so they can answer ``tell()`` |
|
355 | Instances track the amount of written data so they can answer ``tell()`` | |
356 | requests. |
|
356 | requests. | |
357 |
|
357 | |||
358 | The intent of this class is to wrap the ``write()`` function returned by |
|
358 | The intent of this class is to wrap the ``write()`` function returned by | |
359 | a WSGI ``start_response()`` function. Since ``write()`` is a callable and |
|
359 | a WSGI ``start_response()`` function. Since ``write()`` is a callable and | |
360 | not a file object, it doesn't implement other file object methods. |
|
360 | not a file object, it doesn't implement other file object methods. | |
361 | """ |
|
361 | """ | |
362 |
|
362 | |||
363 | def __init__(self, writefn): |
|
363 | def __init__(self, writefn): | |
364 | self._write = writefn |
|
364 | self._write = writefn | |
365 | self._offset = 0 |
|
365 | self._offset = 0 | |
366 |
|
366 | |||
367 | def write(self, s): |
|
367 | def write(self, s): | |
368 | res = self._write(s) |
|
368 | res = self._write(s) | |
369 | # Some Python objects don't report the number of bytes written. |
|
369 | # Some Python objects don't report the number of bytes written. | |
370 | if res is None: |
|
370 | if res is None: | |
371 | self._offset += len(s) |
|
371 | self._offset += len(s) | |
372 | else: |
|
372 | else: | |
373 | self._offset += res |
|
373 | self._offset += res | |
374 |
|
374 | |||
375 | def flush(self): |
|
375 | def flush(self): | |
376 | pass |
|
376 | pass | |
377 |
|
377 | |||
378 | def tell(self): |
|
378 | def tell(self): | |
379 | return self._offset |
|
379 | return self._offset | |
380 |
|
380 | |||
381 |
|
381 | |||
382 | class wsgiresponse(object): |
|
382 | class wsgiresponse(object): | |
383 | """Represents a response to a WSGI request. |
|
383 | """Represents a response to a WSGI request. | |
384 |
|
384 | |||
385 | A response consists of a status line, headers, and a body. |
|
385 | A response consists of a status line, headers, and a body. | |
386 |
|
386 | |||
387 | Consumers must populate the ``status`` and ``headers`` fields and |
|
387 | Consumers must populate the ``status`` and ``headers`` fields and | |
388 | make a call to a ``setbody*()`` method before the response can be |
|
388 | make a call to a ``setbody*()`` method before the response can be | |
389 | issued. |
|
389 | issued. | |
390 |
|
390 | |||
391 | When it is time to start sending the response over the wire, |
|
391 | When it is time to start sending the response over the wire, | |
392 | ``sendresponse()`` is called. It handles emitting the header portion |
|
392 | ``sendresponse()`` is called. It handles emitting the header portion | |
393 | of the response message. It then yields chunks of body data to be |
|
393 | of the response message. It then yields chunks of body data to be | |
394 | written to the peer. Typically, the WSGI application itself calls |
|
394 | written to the peer. Typically, the WSGI application itself calls | |
395 | and returns the value from ``sendresponse()``. |
|
395 | and returns the value from ``sendresponse()``. | |
396 | """ |
|
396 | """ | |
397 |
|
397 | |||
398 | def __init__(self, req, startresponse): |
|
398 | def __init__(self, req, startresponse): | |
399 | """Create an empty response tied to a specific request. |
|
399 | """Create an empty response tied to a specific request. | |
400 |
|
400 | |||
401 | ``req`` is a ``parsedrequest``. ``startresponse`` is the |
|
401 | ``req`` is a ``parsedrequest``. ``startresponse`` is the | |
402 | ``start_response`` function passed to the WSGI application. |
|
402 | ``start_response`` function passed to the WSGI application. | |
403 | """ |
|
403 | """ | |
404 | self._req = req |
|
404 | self._req = req | |
405 | self._startresponse = startresponse |
|
405 | self._startresponse = startresponse | |
406 |
|
406 | |||
407 | self.status = None |
|
407 | self.status = None | |
408 | from . import wsgiheaders # avoid cycle |
|
408 | from . import wsgiheaders # avoid cycle | |
409 |
|
409 | |||
410 | self.headers = wsgiheaders.Headers([]) |
|
410 | self.headers = wsgiheaders.Headers([]) | |
411 |
|
411 | |||
412 | self._bodybytes = None |
|
412 | self._bodybytes = None | |
413 | self._bodygen = None |
|
413 | self._bodygen = None | |
414 | self._bodywillwrite = False |
|
414 | self._bodywillwrite = False | |
415 | self._started = False |
|
415 | self._started = False | |
416 | self._bodywritefn = None |
|
416 | self._bodywritefn = None | |
417 |
|
417 | |||
418 | def _verifybody(self): |
|
418 | def _verifybody(self): | |
419 | if ( |
|
419 | if ( | |
420 | self._bodybytes is not None |
|
420 | self._bodybytes is not None | |
421 | or self._bodygen is not None |
|
421 | or self._bodygen is not None | |
422 | or self._bodywillwrite |
|
422 | or self._bodywillwrite | |
423 | ): |
|
423 | ): | |
424 | raise error.ProgrammingError(b'cannot define body multiple times') |
|
424 | raise error.ProgrammingError(b'cannot define body multiple times') | |
425 |
|
425 | |||
426 | def setbodybytes(self, b): |
|
426 | def setbodybytes(self, b): | |
427 | """Define the response body as static bytes. |
|
427 | """Define the response body as static bytes. | |
428 |
|
428 | |||
429 | The empty string signals that there is no response body. |
|
429 | The empty string signals that there is no response body. | |
430 | """ |
|
430 | """ | |
431 | self._verifybody() |
|
431 | self._verifybody() | |
432 | self._bodybytes = b |
|
432 | self._bodybytes = b | |
433 | self.headers[b'Content-Length'] = b'%d' % len(b) |
|
433 | self.headers[b'Content-Length'] = b'%d' % len(b) | |
434 |
|
434 | |||
435 | def setbodygen(self, gen): |
|
435 | def setbodygen(self, gen): | |
436 | """Define the response body as a generator of bytes.""" |
|
436 | """Define the response body as a generator of bytes.""" | |
437 | self._verifybody() |
|
437 | self._verifybody() | |
438 | self._bodygen = gen |
|
438 | self._bodygen = gen | |
439 |
|
439 | |||
440 | def setbodywillwrite(self): |
|
440 | def setbodywillwrite(self): | |
441 | """Signal an intent to use write() to emit the response body. |
|
441 | """Signal an intent to use write() to emit the response body. | |
442 |
|
442 | |||
443 | **This is the least preferred way to send a body.** |
|
443 | **This is the least preferred way to send a body.** | |
444 |
|
444 | |||
445 | It is preferred for WSGI applications to emit a generator of chunks |
|
445 | It is preferred for WSGI applications to emit a generator of chunks | |
446 | constituting the response body. However, some consumers can't emit |
|
446 | constituting the response body. However, some consumers can't emit | |
447 | data this way. So, WSGI provides a way to obtain a ``write(data)`` |
|
447 | data this way. So, WSGI provides a way to obtain a ``write(data)`` | |
448 | function that can be used to synchronously perform an unbuffered |
|
448 | function that can be used to synchronously perform an unbuffered | |
449 | write. |
|
449 | write. | |
450 |
|
450 | |||
451 | Calling this function signals an intent to produce the body in this |
|
451 | Calling this function signals an intent to produce the body in this | |
452 | manner. |
|
452 | manner. | |
453 | """ |
|
453 | """ | |
454 | self._verifybody() |
|
454 | self._verifybody() | |
455 | self._bodywillwrite = True |
|
455 | self._bodywillwrite = True | |
456 |
|
456 | |||
457 | def sendresponse(self): |
|
457 | def sendresponse(self): | |
458 | """Send the generated response to the client. |
|
458 | """Send the generated response to the client. | |
459 |
|
459 | |||
460 | Before this is called, ``status`` must be set and one of |
|
460 | Before this is called, ``status`` must be set and one of | |
461 | ``setbodybytes()`` or ``setbodygen()`` must be called. |
|
461 | ``setbodybytes()`` or ``setbodygen()`` must be called. | |
462 |
|
462 | |||
463 | Calling this method multiple times is not allowed. |
|
463 | Calling this method multiple times is not allowed. | |
464 | """ |
|
464 | """ | |
465 | if self._started: |
|
465 | if self._started: | |
466 | raise error.ProgrammingError( |
|
466 | raise error.ProgrammingError( | |
467 | b'sendresponse() called multiple times' |
|
467 | b'sendresponse() called multiple times' | |
468 | ) |
|
468 | ) | |
469 |
|
469 | |||
470 | self._started = True |
|
470 | self._started = True | |
471 |
|
471 | |||
472 | if not self.status: |
|
472 | if not self.status: | |
473 | raise error.ProgrammingError(b'status line not defined') |
|
473 | raise error.ProgrammingError(b'status line not defined') | |
474 |
|
474 | |||
475 | if ( |
|
475 | if ( | |
476 | self._bodybytes is None |
|
476 | self._bodybytes is None | |
477 | and self._bodygen is None |
|
477 | and self._bodygen is None | |
478 | and not self._bodywillwrite |
|
478 | and not self._bodywillwrite | |
479 | ): |
|
479 | ): | |
480 | raise error.ProgrammingError(b'response body not defined') |
|
480 | raise error.ProgrammingError(b'response body not defined') | |
481 |
|
481 | |||
482 | # RFC 7232 Section 4.1 states that a 304 MUST generate one of |
|
482 | # RFC 7232 Section 4.1 states that a 304 MUST generate one of | |
483 | # {Cache-Control, Content-Location, Date, ETag, Expires, Vary} |
|
483 | # {Cache-Control, Content-Location, Date, ETag, Expires, Vary} | |
484 | # and SHOULD NOT generate other headers unless they could be used |
|
484 | # and SHOULD NOT generate other headers unless they could be used | |
485 | # to guide cache updates. Furthermore, RFC 7230 Section 3.3.2 |
|
485 | # to guide cache updates. Furthermore, RFC 7230 Section 3.3.2 | |
486 | # states that no response body can be issued. Content-Length can |
|
486 | # states that no response body can be issued. Content-Length can | |
487 | # be sent. But if it is present, it should be the size of the response |
|
487 | # be sent. But if it is present, it should be the size of the response | |
488 | # that wasn't transferred. |
|
488 | # that wasn't transferred. | |
489 | if self.status.startswith(b'304 '): |
|
489 | if self.status.startswith(b'304 '): | |
490 | # setbodybytes('') will set C-L to 0. This doesn't conform with the |
|
490 | # setbodybytes('') will set C-L to 0. This doesn't conform with the | |
491 | # spec. So remove it. |
|
491 | # spec. So remove it. | |
492 | if self.headers.get(b'Content-Length') == b'0': |
|
492 | if self.headers.get(b'Content-Length') == b'0': | |
493 | del self.headers[b'Content-Length'] |
|
493 | del self.headers[b'Content-Length'] | |
494 |
|
494 | |||
495 | # Strictly speaking, this is too strict. But until it causes |
|
495 | # Strictly speaking, this is too strict. But until it causes | |
496 | # problems, let's be strict. |
|
496 | # problems, let's be strict. | |
497 | badheaders = { |
|
497 | badheaders = { | |
498 | k |
|
498 | k | |
499 | for k in self.headers.keys() |
|
499 | for k in self.headers.keys() | |
500 | if k.lower() |
|
500 | if k.lower() | |
501 | not in ( |
|
501 | not in ( | |
502 | b'date', |
|
502 | b'date', | |
503 | b'etag', |
|
503 | b'etag', | |
504 | b'expires', |
|
504 | b'expires', | |
505 | b'cache-control', |
|
505 | b'cache-control', | |
506 | b'content-location', |
|
506 | b'content-location', | |
507 | b'content-security-policy', |
|
507 | b'content-security-policy', | |
508 | b'vary', |
|
508 | b'vary', | |
509 | ) |
|
509 | ) | |
510 | } |
|
510 | } | |
511 | if badheaders: |
|
511 | if badheaders: | |
512 | raise error.ProgrammingError( |
|
512 | raise error.ProgrammingError( | |
513 | b'illegal header on 304 response: %s' |
|
513 | b'illegal header on 304 response: %s' | |
514 | % b', '.join(sorted(badheaders)) |
|
514 | % b', '.join(sorted(badheaders)) | |
515 | ) |
|
515 | ) | |
516 |
|
516 | |||
517 | if self._bodygen is not None or self._bodywillwrite: |
|
517 | if self._bodygen is not None or self._bodywillwrite: | |
518 | raise error.ProgrammingError( |
|
518 | raise error.ProgrammingError( | |
519 | b"must use setbodybytes('') with 304 responses" |
|
519 | b"must use setbodybytes('') with 304 responses" | |
520 | ) |
|
520 | ) | |
521 |
|
521 | |||
522 | # Various HTTP clients (notably httplib) won't read the HTTP response |
|
522 | # Various HTTP clients (notably httplib) won't read the HTTP response | |
523 | # until the HTTP request has been sent in full. If servers (us) send a |
|
523 | # until the HTTP request has been sent in full. If servers (us) send a | |
524 | # response before the HTTP request has been fully sent, the connection |
|
524 | # response before the HTTP request has been fully sent, the connection | |
525 | # may deadlock because neither end is reading. |
|
525 | # may deadlock because neither end is reading. | |
526 | # |
|
526 | # | |
527 | # We work around this by "draining" the request data before |
|
527 | # We work around this by "draining" the request data before | |
528 | # sending any response in some conditions. |
|
528 | # sending any response in some conditions. | |
529 | drain = False |
|
529 | drain = False | |
530 | close = False |
|
530 | close = False | |
531 |
|
531 | |||
532 | # If the client sent Expect: 100-continue, we assume it is smart enough |
|
532 | # If the client sent Expect: 100-continue, we assume it is smart enough | |
533 | # to deal with the server sending a response before reading the request. |
|
533 | # to deal with the server sending a response before reading the request. | |
534 | # (httplib doesn't do this.) |
|
534 | # (httplib doesn't do this.) | |
535 | if self._req.headers.get(b'Expect', b'').lower() == b'100-continue': |
|
535 | if self._req.headers.get(b'Expect', b'').lower() == b'100-continue': | |
536 | pass |
|
536 | pass | |
537 | # Only tend to request methods that have bodies. Strictly speaking, |
|
537 | # Only tend to request methods that have bodies. Strictly speaking, | |
538 | # we should sniff for a body. But this is fine for our existing |
|
538 | # we should sniff for a body. But this is fine for our existing | |
539 | # WSGI applications. |
|
539 | # WSGI applications. | |
540 | elif self._req.method not in (b'POST', b'PUT'): |
|
540 | elif self._req.method not in (b'POST', b'PUT'): | |
541 | pass |
|
541 | pass | |
542 | else: |
|
542 | else: | |
543 | # If we don't know how much data to read, there's no guarantee |
|
543 | # If we don't know how much data to read, there's no guarantee | |
544 | # that we can drain the request responsibly. The WSGI |
|
544 | # that we can drain the request responsibly. The WSGI | |
545 | # specification only says that servers *should* ensure the |
|
545 | # specification only says that servers *should* ensure the | |
546 | # input stream doesn't overrun the actual request. So there's |
|
546 | # input stream doesn't overrun the actual request. So there's | |
547 | # no guarantee that reading until EOF won't corrupt the stream |
|
547 | # no guarantee that reading until EOF won't corrupt the stream | |
548 | # state. |
|
548 | # state. | |
549 | if not isinstance(self._req.bodyfh, util.cappedreader): |
|
549 | if not isinstance(self._req.bodyfh, util.cappedreader): | |
550 | close = True |
|
550 | close = True | |
551 | else: |
|
551 | else: | |
552 | # We /could/ only drain certain HTTP response codes. But 200 and |
|
552 | # We /could/ only drain certain HTTP response codes. But 200 and | |
553 | # non-200 wire protocol responses both require draining. Since |
|
553 | # non-200 wire protocol responses both require draining. Since | |
554 | # we have a capped reader in place for all situations where we |
|
554 | # we have a capped reader in place for all situations where we | |
555 | # drain, it is safe to read from that stream. We'll either do |
|
555 | # drain, it is safe to read from that stream. We'll either do | |
556 | # a drain or no-op if we're already at EOF. |
|
556 | # a drain or no-op if we're already at EOF. | |
557 | drain = True |
|
557 | drain = True | |
558 |
|
558 | |||
559 | if close: |
|
559 | if close: | |
560 | self.headers[b'Connection'] = b'Close' |
|
560 | self.headers[b'Connection'] = b'Close' | |
561 |
|
561 | |||
562 | if drain: |
|
562 | if drain: | |
563 | assert isinstance(self._req.bodyfh, util.cappedreader) |
|
563 | assert isinstance(self._req.bodyfh, util.cappedreader) | |
564 | while True: |
|
564 | while True: | |
565 | chunk = self._req.bodyfh.read(32768) |
|
565 | chunk = self._req.bodyfh.read(32768) | |
566 | if not chunk: |
|
566 | if not chunk: | |
567 | break |
|
567 | break | |
568 |
|
568 | |||
569 | strheaders = [ |
|
569 | strheaders = [ | |
570 | (pycompat.strurl(k), pycompat.strurl(v)) |
|
570 | (pycompat.strurl(k), pycompat.strurl(v)) | |
571 | for k, v in self.headers.items() |
|
571 | for k, v in self.headers.items() | |
572 | ] |
|
572 | ] | |
573 | write = self._startresponse(pycompat.sysstr(self.status), strheaders) |
|
573 | write = self._startresponse(pycompat.sysstr(self.status), strheaders) | |
574 |
|
574 | |||
575 | if self._bodybytes: |
|
575 | if self._bodybytes: | |
576 | yield self._bodybytes |
|
576 | yield self._bodybytes | |
577 | elif self._bodygen: |
|
577 | elif self._bodygen: | |
578 | for chunk in self._bodygen: |
|
578 | for chunk in self._bodygen: | |
579 | # PEP-3333 says that output must be bytes. And some WSGI |
|
579 | # PEP-3333 says that output must be bytes. And some WSGI | |
580 | # implementations enforce this. We cast bytes-like types here |
|
580 | # implementations enforce this. We cast bytes-like types here | |
581 | # for convenience. |
|
581 | # for convenience. | |
582 | if isinstance(chunk, bytearray): |
|
582 | if isinstance(chunk, bytearray): | |
583 | chunk = bytes(chunk) |
|
583 | chunk = bytes(chunk) | |
584 |
|
584 | |||
585 | yield chunk |
|
585 | yield chunk | |
586 | elif self._bodywillwrite: |
|
586 | elif self._bodywillwrite: | |
587 | self._bodywritefn = write |
|
587 | self._bodywritefn = write | |
588 | else: |
|
588 | else: | |
589 | error.ProgrammingError(b'do not know how to send body') |
|
589 | error.ProgrammingError(b'do not know how to send body') | |
590 |
|
590 | |||
591 | def getbodyfile(self): |
|
591 | def getbodyfile(self): | |
592 | """Obtain a file object like object representing the response body. |
|
592 | """Obtain a file object like object representing the response body. | |
593 |
|
593 | |||
594 | For this to work, you must call ``setbodywillwrite()`` and then |
|
594 | For this to work, you must call ``setbodywillwrite()`` and then | |
595 | ``sendresponse()`` first. ``sendresponse()`` is a generator and the |
|
595 | ``sendresponse()`` first. ``sendresponse()`` is a generator and the | |
596 | function won't run to completion unless the generator is advanced. The |
|
596 | function won't run to completion unless the generator is advanced. The | |
597 | generator yields not items. The easiest way to consume it is with |
|
597 | generator yields not items. The easiest way to consume it is with | |
598 | ``list(res.sendresponse())``, which should resolve to an empty list - |
|
598 | ``list(res.sendresponse())``, which should resolve to an empty list - | |
599 | ``[]``. |
|
599 | ``[]``. | |
600 | """ |
|
600 | """ | |
601 | if not self._bodywillwrite: |
|
601 | if not self._bodywillwrite: | |
602 | raise error.ProgrammingError(b'must call setbodywillwrite() first') |
|
602 | raise error.ProgrammingError(b'must call setbodywillwrite() first') | |
603 |
|
603 | |||
604 | if not self._started: |
|
604 | if not self._started: | |
605 | raise error.ProgrammingError( |
|
605 | raise error.ProgrammingError( | |
606 | b'must call sendresponse() first; did ' |
|
606 | b'must call sendresponse() first; did ' | |
607 | b'you remember to consume it since it ' |
|
607 | b'you remember to consume it since it ' | |
608 | b'is a generator?' |
|
608 | b'is a generator?' | |
609 | ) |
|
609 | ) | |
610 |
|
610 | |||
611 | assert self._bodywritefn |
|
611 | assert self._bodywritefn | |
612 | return offsettrackingwriter(self._bodywritefn) |
|
612 | return offsettrackingwriter(self._bodywritefn) | |
613 |
|
613 | |||
614 |
|
614 | |||
615 | def wsgiapplication(app_maker): |
|
615 | def wsgiapplication(app_maker): | |
616 | '''For compatibility with old CGI scripts. A plain hgweb() or hgwebdir() |
|
616 | '''For compatibility with old CGI scripts. A plain hgweb() or hgwebdir() | |
617 | can and should now be used as a WSGI application.''' |
|
617 | can and should now be used as a WSGI application.''' | |
618 | application = app_maker() |
|
618 | application = app_maker() | |
619 |
|
619 | |||
620 | def run_wsgi(env, respond): |
|
620 | def run_wsgi(env, respond): | |
621 | return application(env, respond) |
|
621 | return application(env, respond) | |
622 |
|
622 | |||
623 | return run_wsgi |
|
623 | return run_wsgi |
General Comments 0
You need to be logged in to leave comments.
Login now