Show More
@@ -1,286 +1,288 b'' | |||
|
1 | 1 | # httpconnection.py - urllib2 handler for new http support |
|
2 | 2 | # |
|
3 | 3 | # Copyright 2005, 2006, 2007, 2008 Matt Mackall <mpm@selenic.com> |
|
4 | 4 | # Copyright 2006, 2007 Alexis S. L. Carvalho <alexis@cecm.usp.br> |
|
5 | 5 | # Copyright 2006 Vadim Gelfer <vadim.gelfer@gmail.com> |
|
6 | 6 | # Copyright 2011 Google, Inc. |
|
7 | 7 | # |
|
8 | 8 | # This software may be used and distributed according to the terms of the |
|
9 | 9 | # GNU General Public License version 2 or any later version. |
|
10 | 10 | import logging |
|
11 | 11 | import socket |
|
12 | 12 | import urllib |
|
13 | 13 | import urllib2 |
|
14 | 14 | import os |
|
15 | 15 | |
|
16 | 16 | from mercurial import httpclient |
|
17 | 17 | from mercurial import sslutil |
|
18 | 18 | from mercurial import util |
|
19 | 19 | from mercurial.i18n import _ |
|
20 | 20 | |
|
21 | 21 | # moved here from url.py to avoid a cycle |
|
22 | 22 | class httpsendfile(object): |
|
23 | 23 | """This is a wrapper around the objects returned by python's "open". |
|
24 | 24 | |
|
25 | 25 | Its purpose is to send file-like objects via HTTP. |
|
26 | 26 | It do however not define a __len__ attribute because the length |
|
27 | 27 | might be more than Py_ssize_t can handle. |
|
28 | 28 | """ |
|
29 | 29 | |
|
30 | 30 | def __init__(self, ui, *args, **kwargs): |
|
31 | 31 | # We can't just "self._data = open(*args, **kwargs)" here because there |
|
32 | 32 | # is an "open" function defined in this module that shadows the global |
|
33 | 33 | # one |
|
34 | 34 | self.ui = ui |
|
35 | 35 | self._data = open(*args, **kwargs) |
|
36 | 36 | self.seek = self._data.seek |
|
37 | 37 | self.close = self._data.close |
|
38 | 38 | self.write = self._data.write |
|
39 | 39 | self.length = os.fstat(self._data.fileno()).st_size |
|
40 | 40 | self._pos = 0 |
|
41 | 41 | self._total = self.length // 1024 * 2 |
|
42 | 42 | |
|
43 | 43 | def read(self, *args, **kwargs): |
|
44 | 44 | try: |
|
45 | 45 | ret = self._data.read(*args, **kwargs) |
|
46 | 46 | except EOFError: |
|
47 | 47 | self.ui.progress(_('sending'), None) |
|
48 | 48 | self._pos += len(ret) |
|
49 | 49 | # We pass double the max for total because we currently have |
|
50 | 50 | # to send the bundle twice in the case of a server that |
|
51 | 51 | # requires authentication. Since we can't know until we try |
|
52 | 52 | # once whether authentication will be required, just lie to |
|
53 | 53 | # the user and maybe the push succeeds suddenly at 50%. |
|
54 | 54 | self.ui.progress(_('sending'), self._pos // 1024, |
|
55 | 55 | unit=_('kb'), total=self._total) |
|
56 | 56 | return ret |
|
57 | 57 | |
|
58 | 58 | # moved here from url.py to avoid a cycle |
|
59 | 59 | def readauthforuri(ui, uri, user): |
|
60 | 60 | # Read configuration |
|
61 | 61 | config = dict() |
|
62 | 62 | for key, val in ui.configitems('auth'): |
|
63 | 63 | if '.' not in key: |
|
64 | 64 | ui.warn(_("ignoring invalid [auth] key '%s'\n") % key) |
|
65 | 65 | continue |
|
66 | 66 | group, setting = key.rsplit('.', 1) |
|
67 | 67 | gdict = config.setdefault(group, dict()) |
|
68 | 68 | if setting in ('username', 'cert', 'key'): |
|
69 | 69 | val = util.expandpath(val) |
|
70 | 70 | gdict[setting] = val |
|
71 | 71 | |
|
72 | 72 | # Find the best match |
|
73 | 73 | if '://' in uri: |
|
74 | 74 | scheme, hostpath = uri.split('://', 1) |
|
75 | 75 | else: |
|
76 | 76 | # Python 2.4.1 doesn't provide the full URI |
|
77 | 77 | scheme, hostpath = 'http', uri |
|
78 | 78 | bestuser = None |
|
79 | 79 | bestlen = 0 |
|
80 | 80 | bestauth = None |
|
81 | 81 | for group, auth in config.iteritems(): |
|
82 | 82 | if user and user != auth.get('username', user): |
|
83 | 83 | # If a username was set in the URI, the entry username |
|
84 | 84 | # must either match it or be unset |
|
85 | 85 | continue |
|
86 | 86 | prefix = auth.get('prefix') |
|
87 | 87 | if not prefix: |
|
88 | 88 | continue |
|
89 | 89 | p = prefix.split('://', 1) |
|
90 | 90 | if len(p) > 1: |
|
91 | 91 | schemes, prefix = [p[0]], p[1] |
|
92 | 92 | else: |
|
93 | 93 | schemes = (auth.get('schemes') or 'https').split() |
|
94 | 94 | if (prefix == '*' or hostpath.startswith(prefix)) and \ |
|
95 | 95 | (len(prefix) > bestlen or (len(prefix) == bestlen and \ |
|
96 | 96 | not bestuser and 'username' in auth)) \ |
|
97 | 97 | and scheme in schemes: |
|
98 | 98 | bestlen = len(prefix) |
|
99 | 99 | bestauth = group, auth |
|
100 | 100 | bestuser = auth.get('username') |
|
101 | 101 | if user and not bestuser: |
|
102 | 102 | auth['username'] = user |
|
103 | 103 | return bestauth |
|
104 | 104 | |
|
105 | 105 | # Mercurial (at least until we can remove the old codepath) requires |
|
106 | 106 | # that the http response object be sufficiently file-like, so we |
|
107 | 107 | # provide a close() method here. |
|
108 | 108 | class HTTPResponse(httpclient.HTTPResponse): |
|
109 | 109 | def close(self): |
|
110 | 110 | pass |
|
111 | 111 | |
|
112 | 112 | class HTTPConnection(httpclient.HTTPConnection): |
|
113 | 113 | response_class = HTTPResponse |
|
114 | 114 | def request(self, method, uri, body=None, headers={}): |
|
115 | 115 | if isinstance(body, httpsendfile): |
|
116 | 116 | body.seek(0) |
|
117 | 117 | httpclient.HTTPConnection.request(self, method, uri, body=body, |
|
118 | 118 | headers=headers) |
|
119 | 119 | |
|
120 | 120 | |
|
121 | 121 | _configuredlogging = False |
|
122 | 122 | LOGFMT = '%(levelname)s:%(name)s:%(lineno)d:%(message)s' |
|
123 | 123 | # Subclass BOTH of these because otherwise urllib2 "helpfully" |
|
124 | 124 | # reinserts them since it notices we don't include any subclasses of |
|
125 | 125 | # them. |
|
126 | 126 | class http2handler(urllib2.HTTPHandler, urllib2.HTTPSHandler): |
|
127 | 127 | def __init__(self, ui, pwmgr): |
|
128 | 128 | global _configuredlogging |
|
129 | 129 | urllib2.AbstractHTTPHandler.__init__(self) |
|
130 | 130 | self.ui = ui |
|
131 | 131 | self.pwmgr = pwmgr |
|
132 | 132 | self._connections = {} |
|
133 | 133 | loglevel = ui.config('ui', 'http2debuglevel', default=None) |
|
134 | 134 | if loglevel and not _configuredlogging: |
|
135 | 135 | _configuredlogging = True |
|
136 | 136 | logger = logging.getLogger('mercurial.httpclient') |
|
137 | 137 | logger.setLevel(getattr(logging, loglevel.upper())) |
|
138 | 138 | handler = logging.StreamHandler() |
|
139 | 139 | handler.setFormatter(logging.Formatter(LOGFMT)) |
|
140 | 140 | logger.addHandler(handler) |
|
141 | 141 | |
|
142 | 142 | def close_all(self): |
|
143 | 143 | """Close and remove all connection objects being kept for reuse.""" |
|
144 | 144 | for openconns in self._connections.values(): |
|
145 | 145 | for conn in openconns: |
|
146 | 146 | conn.close() |
|
147 | 147 | self._connections = {} |
|
148 | 148 | |
|
149 | 149 | # shamelessly borrowed from urllib2.AbstractHTTPHandler |
|
150 | 150 | def do_open(self, http_class, req, use_ssl): |
|
151 | 151 | """Return an addinfourl object for the request, using http_class. |
|
152 | 152 | |
|
153 | 153 | http_class must implement the HTTPConnection API from httplib. |
|
154 | 154 | The addinfourl return value is a file-like object. It also |
|
155 | 155 | has methods and attributes including: |
|
156 | 156 | - info(): return a mimetools.Message object for the headers |
|
157 | 157 | - geturl(): return the original request URL |
|
158 | 158 | - code: HTTP status code |
|
159 | 159 | """ |
|
160 | 160 | # If using a proxy, the host returned by get_host() is |
|
161 | 161 | # actually the proxy. On Python 2.6.1, the real destination |
|
162 | 162 | # hostname is encoded in the URI in the urllib2 request |
|
163 | 163 | # object. On Python 2.6.5, it's stored in the _tunnel_host |
|
164 | 164 | # attribute which has no accessor. |
|
165 | 165 | tunhost = getattr(req, '_tunnel_host', None) |
|
166 | 166 | host = req.get_host() |
|
167 | 167 | if tunhost: |
|
168 | 168 | proxyhost = host |
|
169 | 169 | host = tunhost |
|
170 | 170 | elif req.has_proxy(): |
|
171 | 171 | proxyhost = req.get_host() |
|
172 | 172 | host = req.get_selector().split('://', 1)[1].split('/', 1)[0] |
|
173 | 173 | else: |
|
174 | 174 | proxyhost = None |
|
175 | 175 | |
|
176 | 176 | if proxyhost: |
|
177 | 177 | if ':' in proxyhost: |
|
178 | 178 | # Note: this means we'll explode if we try and use an |
|
179 | 179 | # IPv6 http proxy. This isn't a regression, so we |
|
180 | 180 | # won't worry about it for now. |
|
181 | 181 | proxyhost, proxyport = proxyhost.rsplit(':', 1) |
|
182 | 182 | else: |
|
183 | 183 | proxyport = 3128 # squid default |
|
184 | 184 | proxy = (proxyhost, proxyport) |
|
185 | 185 | else: |
|
186 | 186 | proxy = None |
|
187 | 187 | |
|
188 | 188 | if not host: |
|
189 | 189 | raise urllib2.URLError('no host given') |
|
190 | 190 | |
|
191 | 191 | connkey = use_ssl, host, proxy |
|
192 | 192 | allconns = self._connections.get(connkey, []) |
|
193 | 193 | conns = [c for c in allconns if not c.busy()] |
|
194 | 194 | if conns: |
|
195 | 195 | h = conns[0] |
|
196 | 196 | else: |
|
197 | 197 | if allconns: |
|
198 | 198 | self.ui.debug('all connections for %s busy, making a new ' |
|
199 | 199 | 'one\n' % host) |
|
200 | 200 | timeout = None |
|
201 | 201 | if req.timeout is not socket._GLOBAL_DEFAULT_TIMEOUT: |
|
202 | 202 | timeout = req.timeout |
|
203 | 203 | h = http_class(host, timeout=timeout, proxy_hostport=proxy) |
|
204 | 204 | self._connections.setdefault(connkey, []).append(h) |
|
205 | 205 | |
|
206 | 206 | headers = dict(req.headers) |
|
207 | 207 | headers.update(req.unredirected_hdrs) |
|
208 | 208 | headers = dict( |
|
209 | 209 | (name.title(), val) for name, val in headers.items()) |
|
210 | 210 | try: |
|
211 | 211 | path = req.get_selector() |
|
212 | 212 | if '://' in path: |
|
213 | 213 | path = path.split('://', 1)[1].split('/', 1)[1] |
|
214 | 214 | if path[0] != '/': |
|
215 | 215 | path = '/' + path |
|
216 | 216 | h.request(req.get_method(), path, req.data, headers) |
|
217 | 217 | r = h.getresponse() |
|
218 | 218 | except socket.error, err: # XXX what error? |
|
219 | 219 | raise urllib2.URLError(err) |
|
220 | 220 | |
|
221 | 221 | # Pick apart the HTTPResponse object to get the addinfourl |
|
222 | 222 | # object initialized properly. |
|
223 | 223 | r.recv = r.read |
|
224 | 224 | |
|
225 | 225 | resp = urllib.addinfourl(r, r.headers, req.get_full_url()) |
|
226 | 226 | resp.code = r.status |
|
227 | 227 | resp.msg = r.reason |
|
228 | 228 | return resp |
|
229 | 229 | |
|
230 | 230 | # httplib always uses the given host/port as the socket connect |
|
231 | 231 | # target, and then allows full URIs in the request path, which it |
|
232 | 232 | # then observes and treats as a signal to do proxying instead. |
|
233 | 233 | def http_open(self, req): |
|
234 | 234 | if req.get_full_url().startswith('https'): |
|
235 | 235 | return self.https_open(req) |
|
236 | 236 | def makehttpcon(*args, **kwargs): |
|
237 | return HTTPConnection(*args, use_ssl=False, **kwargs) | |
|
237 | k2 = dict(kwargs) | |
|
238 | k2['use_ssl'] = False | |
|
239 | return HTTPConnection(*args, **k2) | |
|
238 | 240 | return self.do_open(makehttpcon, req, False) |
|
239 | 241 | |
|
240 | 242 | def https_open(self, req): |
|
241 | 243 | # req.get_full_url() does not contain credentials and we may |
|
242 | 244 | # need them to match the certificates. |
|
243 | 245 | url = req.get_full_url() |
|
244 | 246 | user, password = self.pwmgr.find_stored_password(url) |
|
245 | 247 | res = readauthforuri(self.ui, url, user) |
|
246 | 248 | if res: |
|
247 | 249 | group, auth = res |
|
248 | 250 | self.auth = auth |
|
249 | 251 | self.ui.debug("using auth.%s.* for authentication\n" % group) |
|
250 | 252 | else: |
|
251 | 253 | self.auth = None |
|
252 | 254 | return self.do_open(self._makesslconnection, req, True) |
|
253 | 255 | |
|
254 | 256 | def _makesslconnection(self, host, port=443, *args, **kwargs): |
|
255 | 257 | keyfile = None |
|
256 | 258 | certfile = None |
|
257 | 259 | |
|
258 | 260 | if args: # key_file |
|
259 | 261 | keyfile = args.pop(0) |
|
260 | 262 | if args: # cert_file |
|
261 | 263 | certfile = args.pop(0) |
|
262 | 264 | |
|
263 | 265 | # if the user has specified different key/cert files in |
|
264 | 266 | # hgrc, we prefer these |
|
265 | 267 | if self.auth and 'key' in self.auth and 'cert' in self.auth: |
|
266 | 268 | keyfile = self.auth['key'] |
|
267 | 269 | certfile = self.auth['cert'] |
|
268 | 270 | |
|
269 | 271 | # let host port take precedence |
|
270 | 272 | if ':' in host and '[' not in host or ']:' in host: |
|
271 | 273 | host, port = host.rsplit(':', 1) |
|
272 | 274 | port = int(port) |
|
273 | 275 | if '[' in host: |
|
274 | 276 | host = host[1:-1] |
|
275 | 277 | |
|
276 | 278 | if keyfile: |
|
277 | 279 | kwargs['keyfile'] = keyfile |
|
278 | 280 | if certfile: |
|
279 | 281 | kwargs['certfile'] = certfile |
|
280 | 282 | |
|
281 | 283 | kwargs.update(sslutil.sslkwargs(self.ui, host)) |
|
282 | 284 | |
|
283 | 285 | con = HTTPConnection(host, port, use_ssl=True, |
|
284 | 286 | ssl_validator=sslutil.validator(self.ui, host), |
|
285 | 287 | **kwargs) |
|
286 | 288 | return con |
General Comments 0
You need to be logged in to leave comments.
Login now