aboutsummaryrefslogtreecommitdiffstats
path: root/netlib/http.py
blob: 0f2caa5a39b7fa741b5e92054cef76e1da16d714 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
import string, urlparse, binascii
import odict, utils

class HttpError(Exception):
    def __init__(self, code, msg):
        self.code, self.msg = code, msg

    def __str__(self):
        return "HttpError(%s, %s)"%(self.code, self.msg)


class HttpErrorConnClosed(HttpError): pass


def _is_valid_port(port):
    if not 0 <= port <= 65535:
        return False
    return True


def _is_valid_host(host):
    try:
        host.decode("idna")
    except ValueError:
        return False
    if "\0" in host:
        return None
    return True


def parse_url(url):
    """
        Returns a (scheme, host, port, path) tuple, or None on error.

        Checks that:
            port is an integer 0-65535
            host is a valid IDNA-encoded hostname with no null-bytes
            path is valid ASCII
    """
    try:
        scheme, netloc, path, params, query, fragment = urlparse.urlparse(url)
    except ValueError:
        return None
    if not scheme:
        return None
    if ':' in netloc:
        host, port = string.rsplit(netloc, ':', maxsplit=1)
        try:
            port = int(port)
        except ValueError:
            return None
    else:
        host = netloc
        if scheme == "https":
            port = 443
        else:
            port = 80
    path = urlparse.urlunparse(('', '', path, params, query, fragment))
    if not path.startswith("/"):
        path = "/" + path
    if not _is_valid_host(host):
        return None
    if not utils.isascii(path):
        return None
    if not _is_valid_port(port):
        return None
    return scheme, host, port, path


def read_headers(fp):
    """
        Read a set of headers from a file pointer. Stop once a blank line is
        reached. Return a ODictCaseless object, or None if headers are invalid.
    """
    ret = []
    name = ''
    while 1:
        line = fp.readline()
        if not line or line == '\r\n' or line == '\n':
            break
        if line[0] in ' \t':
            if not ret:
                return None
            # continued header
            ret[-1][1] = ret[-1][1] + '\r\n ' + line.strip()
        else:
            i = line.find(':')
            # We're being liberal in what we accept, here.
            if i > 0:
                name = line[:i]
                value = line[i+1:].strip()
                ret.append([name, value])
            else:
                return None
    return odict.ODictCaseless(ret)


def read_chunked(code, fp, limit):
    """
        Read a chunked HTTP body.

        May raise HttpError.
    """
    content = ""
    total = 0
    while 1:
        line = fp.readline(128)
        if line == "":
            raise HttpErrorConnClosed(code, "Connection closed prematurely")
        if line != '\r\n' and line != '\n':
            try:
                length = int(line, 16)
            except ValueError:
                # FIXME: Not strictly correct - this could be from the server, in which
                # case we should send a 502.
                raise HttpError(code, "Invalid chunked encoding length: %s"%line)
            if not length:
                break
            total += length
            if limit is not None and total > limit:
                msg = "HTTP Body too large."\
                      " Limit is %s, chunked content length was at least %s"%(limit, total)
                raise HttpError(code, msg)
            content += fp.read(length)
            line = fp.readline(5)
            if line != '\r\n':
                raise HttpError(code, "Malformed chunked body")
    while 1:
        line = fp.readline()
        if line == "":
            raise HttpErrorConnClosed(code, "Connection closed prematurely")
        if line == '\r\n' or line == '\n':
            break
    return content


def get_header_tokens(headers, key):
    """
        Retrieve all tokens for a header key. A number of different headers
        follow a pattern where each header line can containe comma-separated
        tokens, and headers can be set multiple times.
    """
    toks = []
    for i in headers[key]:
        for j in i.split(","):
            toks.append(j.strip())
    return toks


def has_chunked_encoding(headers):
    return "chunked" in [i.lower() for i in get_header_tokens(headers, "transfer-encoding")]


def read_http_body(code, rfile, headers, all, limit):
    """
        Read an HTTP body:

            code: The HTTP error code to be used when raising HttpError
            rfile: A file descriptor to read from
            headers: An ODictCaseless object
            all: Should we read all data?
            limit: Size limit.
    """
    if has_chunked_encoding(headers):
        content = read_chunked(code, rfile, limit)
    elif "content-length" in headers:
        try:
            l = int(headers["content-length"][0])
        except ValueError:
            # FIXME: Not strictly correct - this could be from the server, in which
            # case we should send a 502.
            raise HttpError(code, "Invalid content-length header: %s"%headers["content-length"])
        if limit is not None and l > limit:
            raise HttpError(code, "HTTP Body too large. Limit is %s, content-length was %s"%(limit, l))
        content = rfile.read(l)
    elif all:
        content = rfile.read(limit if limit else -1)
    else:
        content = ""
    return content


def parse_http_protocol(s):
    """
        Parse an HTTP protocol declaration. Returns a (major, minor) tuple, or
        None.
    """
    if not s.startswith("HTTP/"):
        return None
    _, version = s.split('/', 1)
    if "." not in version:
        return None
    major, minor = version.split('.', 1)
    try:
        major = int(major)
        minor = int(minor)
    except ValueError:
        return None
    return major, minor


def parse_http_basic_auth(s):
    words = s.split()
    if len(words) != 2:
        return None
    scheme = words[0]
    try:
        user = binascii.a2b_base64(words[1])
    except binascii.Error:
        return None
    parts = user.split(':')
    if len(parts) != 2:
        return None
    return scheme, parts[0], parts[1]


def assemble_http_basic_auth(scheme, username, password):
    v = binascii.b2a_base64(username + ":" + password)
    return scheme + " " + v


def parse_init(line):
    try:
        method, url, protocol = string.split(line)
    except ValueError:
        return None
    httpversion = parse_http_protocol(protocol)
    if not httpversion:
        return None
    return method, url, httpversion


def parse_init_connect(line):
    v = parse_init(line)
    if not v:
        return None
    method, url, httpversion = v

    if method.upper() != 'CONNECT':
        return None
    try:
        host, port = url.split(":")
    except ValueError:
        return None
    try:
        port = int(port)
    except ValueError:
        return None
    if not _is_valid_port(port):
        return None
    if not _is_valid_host(host):
        return None
    return host, port, httpversion


def parse_init_proxy(line):
    v = parse_init(line)
    if not v:
        return None
    method, url, httpversion = v

    parts = parse_url(url)
    if not parts:
        return None
    scheme, host, port, path = parts
    return method, scheme, host, port, path, httpversion


def parse_init_http(line):
    """
        Returns (method, url, httpversion)
    """
    v = parse_init(line)
    if not v:
        return None
    method, url, httpversion = v
    if not utils.isascii(url):
        return None
    if not (url.startswith("/") or url == "*"):
        return None
    return method, url, httpversion


def request_connection_close(httpversion, headers):
    """
        Checks the request to see if the client connection should be closed.
    """
    if "connection" in headers:
        toks = get_header_tokens(headers, "connection")
        if "close" in toks:
            return True
        elif "keep-alive" in toks:
            return False
    # HTTP 1.1 connections are assumed to be persistent
    if httpversion == (1, 1):
        return False
    return True


def response_connection_close(httpversion, headers):
    """
        Checks the response to see if the client connection should be closed.
    """
    if request_connection_close(httpversion, headers):
        return True
    elif (not has_chunked_encoding(headers)) and "content-length" in headers:
        return False
    return True


def read_http_body_request(rfile, wfile, headers, httpversion, limit):
    """
        Read the HTTP body from a client request.
    """
    if "expect" in headers:
        # FIXME: Should be forwarded upstream
        if "100-continue" in headers['expect'] and httpversion >= (1, 1):
            wfile.write('HTTP/1.1 100 Continue\r\n')
            wfile.write('\r\n')
            del headers['expect']
    return read_http_body(400, rfile, headers, False, limit)


def read_http_body_response(rfile, headers, limit):
    """
        Read the HTTP body from a server response.
    """
    all = "close" in get_header_tokens(headers, "connection")
    return read_http_body(500, rfile, headers, all, limit)


def parse_response_line(line):
    parts = line.strip().split(" ", 2)
    if len(parts) == 2: # handle missing message gracefully
        parts.append("")
    if len(parts) != 3:
        return None
    proto, code, msg = parts
    try:
        code = int(code)
    except ValueError:
        return None
    return (proto, code, msg)


def read_response(rfile, method, body_size_limit):
    """
        Return an (httpversion, code, msg, headers, content) tuple.
    """
    line = rfile.readline()
    if line == "\r\n" or line == "\n": # Possible leftover from previous message
        line = rfile.readline()
    if not line:
        raise HttpErrorConnClosed(502, "Server disconnect.")
    parts = parse_response_line(line)
    if not parts:
        raise HttpError(502, "Invalid server response: %s"%repr(line))
    proto, code, msg = parts
    httpversion = parse_http_protocol(proto)
    if httpversion is None:
        raise HttpError(502, "Invalid HTTP version in line: %s"%repr(proto))
    headers = read_headers(rfile)
    if headers is None:
        raise HttpError(502, "Invalid headers.")
    if code >= 100 and code <= 199:
        return read_response(rfile, method, body_size_limit)
    if method == "HEAD" or code == 204 or code == 304:
        content = ""
    else:
        content = read_http_body_response(rfile, headers, body_size_limit)
    return httpversion, code, msg, headers, content