summaryrefslogtreecommitdiffstats
path: root/Lib/wsgiref/validate.py
blob: 1a1853cd63a0d2aedcdb93e5f692040e48c9092e (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
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
# (c) 2005 Ian Bicking and contributors; written for Paste (http://pythonpaste.org)
# Licensed under the MIT license: https://opensource.org/licenses/mit-license.php
# Also licenced under the Apache License, 2.0: https://opensource.org/licenses/apache2.0.php
# Licensed to PSF under a Contributor Agreement
"""
Middleware to check for obedience to the WSGI specification.

Some of the things this checks:

* Signature of the application and start_response (including that
  keyword arguments are not used).

* Environment checks:

  - Environment is a dictionary (and not a subclass).

  - That all the required keys are in the environment: REQUEST_METHOD,
    SERVER_NAME, SERVER_PORT, wsgi.version, wsgi.input, wsgi.errors,
    wsgi.multithread, wsgi.multiprocess, wsgi.run_once

  - That HTTP_CONTENT_TYPE and HTTP_CONTENT_LENGTH are not in the
    environment (these headers should appear as CONTENT_LENGTH and
    CONTENT_TYPE).

  - Warns if QUERY_STRING is missing, as the cgi module acts
    unpredictably in that case.

  - That CGI-style variables (that don't contain a .) have
    (non-unicode) string values

  - That wsgi.version is a tuple

  - That wsgi.url_scheme is 'http' or 'https' (@@: is this too
    restrictive?)

  - Warns if the REQUEST_METHOD is not known (@@: probably too
    restrictive).

  - That SCRIPT_NAME and PATH_INFO are empty or start with /

  - That at least one of SCRIPT_NAME or PATH_INFO are set.

  - That CONTENT_LENGTH is a positive integer.

  - That SCRIPT_NAME is not '/' (it should be '', and PATH_INFO should
    be '/').

  - That wsgi.input has the methods read, readline, readlines, and
    __iter__

  - That wsgi.errors has the methods flush, write, writelines

* The status is a string, contains a space, starts with an integer,
  and that integer is in range (> 100).

* That the headers is a list (not a subclass, not another kind of
  sequence).

* That the items of the headers are tuples of strings.

* That there is no 'status' header (that is used in CGI, but not in
  WSGI).

* That the headers don't contain newlines or colons, end in _ or -, or
  contain characters codes below 037.

* That Content-Type is given if there is content (CGI often has a
  default content type, but WSGI does not).

* That no Content-Type is given when there is no content (@@: is this
  too restrictive?)

* That the exc_info argument to start_response is a tuple or None.

* That all calls to the writer are with strings, and no other methods
  on the writer are accessed.

* That wsgi.input is used properly:

  - .read() is called with exactly one argument

  - That it returns a string

  - That readline, readlines, and __iter__ return strings

  - That .close() is not called

  - No other methods are provided

* That wsgi.errors is used properly:

  - .write() and .writelines() is called with a string

  - That .close() is not called, and no other methods are provided.

* The response iterator:

  - That it is not a string (it should be a list of a single string; a
    string will work, but perform horribly).

  - That .__next__() returns a string

  - That the iterator is not iterated over until start_response has
    been called (that can signal either a server or application
    error).

  - That .close() is called (doesn't raise exception, only prints to
    sys.stderr, because we only know it isn't called when the object
    is garbage collected).
"""
__all__ = ['validator']


import re
import sys
import warnings

header_re = re.compile(r'^[a-zA-Z][a-zA-Z0-9\-_]*$')
bad_header_value_re = re.compile(r'[\000-\037]')

class WSGIWarning(Warning):
    """
    Raised in response to WSGI-spec-related warnings
    """

def assert_(cond, *args):
    if not cond:
        raise AssertionError(*args)

def check_string_type(value, title):
    if type (value) is str:
        return value
    raise AssertionError(
        "{0} must be of type str (got {1})".format(title, repr(value)))

def validator(application):

    """
    When applied between a WSGI server and a WSGI application, this
    middleware will check for WSGI compliance on a number of levels.
    This middleware does not modify the request or response in any
    way, but will raise an AssertionError if anything seems off
    (except for a failure to close the application iterator, which
    will be printed to stderr -- there's no way to raise an exception
    at that point).
    """

    def lint_app(*args, **kw):
        assert_(len(args) == 2, "Two arguments required")
        assert_(not kw, "No keyword arguments allowed")
        environ, start_response = args

        check_environ(environ)

        # We use this to check if the application returns without
        # calling start_response:
        start_response_started = []

        def start_response_wrapper(*args, **kw):
            assert_(len(args) == 2 or len(args) == 3, (
                "Invalid number of arguments: %s" % (args,)))
            assert_(not kw, "No keyword arguments allowed")
            status = args[0]
            headers = args[1]
            if len(args) == 3:
                exc_info = args[2]
            else:
                exc_info = None

            check_status(status)
            check_headers(headers)
            check_content_type(status, headers)
            check_exc_info(exc_info)

            start_response_started.append(None)
            return WriteWrapper(start_response(*args))

        environ['wsgi.input'] = InputWrapper(environ['wsgi.input'])
        environ['wsgi.errors'] = ErrorWrapper(environ['wsgi.errors'])

        iterator = application(environ, start_response_wrapper)
        assert_(iterator is not None and iterator != False,
            "The application must return an iterator, if only an empty list")

        check_iterator(iterator)

        return IteratorWrapper(iterator, start_response_started)

    return lint_app

class InputWrapper:

    def __init__(self, wsgi_input):
        self.input = wsgi_input

    def read(self, *args):
        assert_(len(args) == 1)
        v = self.input.read(*args)
        assert_(type(v) is bytes)
        return v

    def readline(self, *args):
        assert_(len(args) <= 1)
        v = self.input.readline(*args)
        assert_(type(v) is bytes)
        return v

    def readlines(self, *args):
        assert_(len(args) <= 1)
        lines = self.input.readlines(*args)
        assert_(type(lines) is list)
        for line in lines:
            assert_(type(line) is bytes)
        return lines

    def __iter__(self):
        while line := self.readline():
            yield line

    def close(self):
        assert_(0, "input.close() must not be called")

class ErrorWrapper:

    def __init__(self, wsgi_errors):
        self.errors = wsgi_errors

    def write(self, s):
        assert_(type(s) is str)
        self.errors.write(s)

    def flush(self):
        self.errors.flush()

    def writelines(self, seq):
        for line in seq:
            self.write(line)

    def close(self):
        assert_(0, "errors.close() must not be called")

class WriteWrapper:

    def __init__(self, wsgi_writer):
        self.writer = wsgi_writer

    def __call__(self, s):
        assert_(type(s) is bytes)
        self.writer(s)

class PartialIteratorWrapper:

    def __init__(self, wsgi_iterator):
        self.iterator = wsgi_iterator

    def __iter__(self):
        # We want to make sure __iter__ is called
        return IteratorWrapper(self.iterator, None)

class IteratorWrapper:

    def __init__(self, wsgi_iterator, check_start_response):
        self.original_iterator = wsgi_iterator
        self.iterator = iter(wsgi_iterator)
        self.closed = False
        self.check_start_response = check_start_response

    def __iter__(self):
        return self

    def __next__(self):
        assert_(not self.closed,
            "Iterator read after closed")
        v = next(self.iterator)
        if type(v) is not bytes:
            assert_(False, "Iterator yielded non-bytestring (%r)" % (v,))
        if self.check_start_response is not None:
            assert_(self.check_start_response,
                "The application returns and we started iterating over its body, but start_response has not yet been called")
            self.check_start_response = None
        return v

    def close(self):
        self.closed = True
        if hasattr(self.original_iterator, 'close'):
            self.original_iterator.close()

    def __del__(self):
        if not self.closed:
            sys.stderr.write(
                "Iterator garbage collected without being closed")
        assert_(self.closed,
            "Iterator garbage collected without being closed")

def check_environ(environ):
    assert_(type(environ) is dict,
        "Environment is not of the right type: %r (environment: %r)"
        % (type(environ), environ))

    for key in ['REQUEST_METHOD', 'SERVER_NAME', 'SERVER_PORT',
                'wsgi.version', 'wsgi.input', 'wsgi.errors',
                'wsgi.multithread', 'wsgi.multiprocess',
                'wsgi.run_once']:
        assert_(key in environ,
            "Environment missing required key: %r" % (key,))

    for key in ['HTTP_CONTENT_TYPE', 'HTTP_CONTENT_LENGTH']:
        assert_(key not in environ,
            "Environment should not have the key: %s "
            "(use %s instead)" % (key, key[5:]))

    if 'QUERY_STRING' not in environ:
        warnings.warn(
            'QUERY_STRING is not in the WSGI environment; the cgi '
            'module will use sys.argv when this variable is missing, '
            'so application errors are more likely',
            WSGIWarning)

    for key in environ.keys():
        if '.' in key:
            # Extension, we don't care about its type
            continue
        assert_(type(environ[key]) is str,
            "Environmental variable %s is not a string: %r (value: %r)"
            % (key, type(environ[key]), environ[key]))

    assert_(type(environ['wsgi.version']) is tuple,
        "wsgi.version should be a tuple (%r)" % (environ['wsgi.version'],))
    assert_(environ['wsgi.url_scheme'] in ('http', 'https'),
        "wsgi.url_scheme unknown: %r" % environ['wsgi.url_scheme'])

    check_input(environ['wsgi.input'])
    check_errors(environ['wsgi.errors'])

    # @@: these need filling out:
    if environ['REQUEST_METHOD'] not in (
        'GET', 'HEAD', 'POST', 'OPTIONS', 'PATCH', 'PUT', 'DELETE', 'TRACE'):
        warnings.warn(
            "Unknown REQUEST_METHOD: %r" % environ['REQUEST_METHOD'],
            WSGIWarning)

    assert_(not environ.get('SCRIPT_NAME')
            or environ['SCRIPT_NAME'].startswith('/'),
        "SCRIPT_NAME doesn't start with /: %r" % environ['SCRIPT_NAME'])
    assert_(not environ.get('PATH_INFO')
            or environ['PATH_INFO'].startswith('/'),
        "PATH_INFO doesn't start with /: %r" % environ['PATH_INFO'])
    if environ.get('CONTENT_LENGTH'):
        assert_(int(environ['CONTENT_LENGTH']) >= 0,
            "Invalid CONTENT_LENGTH: %r" % environ['CONTENT_LENGTH'])

    if not environ.get('SCRIPT_NAME'):
        assert_('PATH_INFO' in environ,
            "One of SCRIPT_NAME or PATH_INFO are required (PATH_INFO "
            "should at least be '/' if SCRIPT_NAME is empty)")
    assert_(environ.get('SCRIPT_NAME') != '/',
        "SCRIPT_NAME cannot be '/'; it should instead be '', and "
        "PATH_INFO should be '/'")

def check_input(wsgi_input):
    for attr in ['read', 'readline', 'readlines', '__iter__']:
        assert_(hasattr(wsgi_input, attr),
            "wsgi.input (%r) doesn't have the attribute %s"
            % (wsgi_input, attr))

def check_errors(wsgi_errors):
    for attr in ['flush', 'write', 'writelines']:
        assert_(hasattr(wsgi_errors, attr),
            "wsgi.errors (%r) doesn't have the attribute %s"
            % (wsgi_errors, attr))

def check_status(status):
    status = check_string_type(status, "Status")
    # Implicitly check that we can turn it into an integer:
    status_code = status.split(None, 1)[0]
    assert_(len(status_code) == 3,
        "Status codes must be three characters: %r" % status_code)
    status_int = int(status_code)
    assert_(status_int >= 100, "Status code is invalid: %r" % status_int)
    if len(status) < 4 or status[3] != ' ':
        warnings.warn(
            "The status string (%r) should be a three-digit integer "
            "followed by a single space and a status explanation"
            % status, WSGIWarning)

def check_headers(headers):
    assert_(type(headers) is list,
        "Headers (%r) must be of type list: %r"
        % (headers, type(headers)))
    for item in headers:
        assert_(type(item) is tuple,
            "Individual headers (%r) must be of type tuple: %r"
            % (item, type(item)))
        assert_(len(item) == 2)
        name, value = item
        name = check_string_type(name, "Header name")
        value = check_string_type(value, "Header value")
        assert_(name.lower() != 'status',
            "The Status header cannot be used; it conflicts with CGI "
            "script, and HTTP status is not given through headers "
            "(value: %r)." % value)
        assert_('\n' not in name and ':' not in name,
            "Header names may not contain ':' or '\\n': %r" % name)
        assert_(header_re.search(name), "Bad header name: %r" % name)
        assert_(not name.endswith('-') and not name.endswith('_'),
            "Names may not end in '-' or '_': %r" % name)
        if bad_header_value_re.search(value):
            assert_(0, "Bad header value: %r (bad char: %r)"
            % (value, bad_header_value_re.search(value).group(0)))

def check_content_type(status, headers):
    status = check_string_type(status, "Status")
    code = int(status.split(None, 1)[0])
    # @@: need one more person to verify this interpretation of RFC 2616
    #     http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
    NO_MESSAGE_BODY = (204, 304)
    for name, value in headers:
        name = check_string_type(name, "Header name")
        if name.lower() == 'content-type':
            if code not in NO_MESSAGE_BODY:
                return
            assert_(0, ("Content-Type header found in a %s response, "
                        "which must not return content.") % code)
    if code not in NO_MESSAGE_BODY:
        assert_(0, "No Content-Type header found in headers (%s)" % headers)

def check_exc_info(exc_info):
    assert_(exc_info is None or type(exc_info) is tuple,
        "exc_info (%r) is not a tuple: %r" % (exc_info, type(exc_info)))
    # More exc_info checks?

def check_iterator(iterator):
    # Technically a bytestring is legal, which is why it's a really bad
    # idea, because it may cause the response to be returned
    # character-by-character
    assert_(not isinstance(iterator, (str, bytes)),
        "You should not return a string as your application iterator, "
        "instead return a single-item list containing a bytestring.")