Lib/wsgiref/validate.py

   1 # (c) 2005 Ian Bicking and contributors; written for Paste (http://pythonpaste.org)
   2 # Licensed under the MIT license: http://www.opensource.org/licenses/mit-license.php
   3 # Also licenced under the Apache License, 2.0: http://opensource.org/licenses/apache2.0.php
   4 # Licensed to PSF under a Contributor Agreement
   5 """
   6 Middleware to check for obedience to the WSGI specification.
   7
   8 Some of the things this checks:
   9
  10 * Signature of the application and start_response (including that
  11   keyword arguments are not used).
  12
  13 * Environment checks:
  14
  15   - Environment is a dictionary (and not a subclass).
  16
  17   - That all the required keys are in the environment: REQUEST_METHOD,
  18     SERVER_NAME, SERVER_PORT, wsgi.version, wsgi.input, wsgi.errors,
  19     wsgi.multithread, wsgi.multiprocess, wsgi.run_once
  20
  21   - That HTTP_CONTENT_TYPE and HTTP_CONTENT_LENGTH are not in the
  22     environment (these headers should appear as CONTENT_LENGTH and
  23     CONTENT_TYPE).
  24
  25   - Warns if QUERY_STRING is missing, as the cgi module acts
  26     unpredictably in that case.
  27
  28   - That CGI-style variables (that don't contain a .) have
  29     (non-unicode) string values
  30
  31   - That wsgi.version is a tuple
  32
  33   - That wsgi.url_scheme is 'http' or 'https' (@@: is this too
  34     restrictive?)
  35
  36   - Warns if the REQUEST_METHOD is not known (@@: probably too
  37     restrictive).
  38
  39   - That SCRIPT_NAME and PATH_INFO are empty or start with /
  40
  41   - That at least one of SCRIPT_NAME or PATH_INFO are set.
  42
  43   - That CONTENT_LENGTH is a positive integer.
  44
  45   - That SCRIPT_NAME is not '/' (it should be '', and PATH_INFO should
  46     be '/').
  47
  48   - That wsgi.input has the methods read, readline, readlines, and
  49     __iter__
  50
  51   - That wsgi.errors has the methods flush, write, writelines
  52
  53 * The status is a string, contains a space, starts with an integer,
  54   and that integer is in range (> 100).
  55
  56 * That the headers is a list (not a subclass, not another kind of
  57   sequence).
  58
  59 * That the items of the headers are tuples of strings.
  60
  61 * That there is no 'status' header (that is used in CGI, but not in
  62   WSGI).
  63
  64 * That the headers don't contain newlines or colons, end in _ or -, or
  65   contain characters codes below 037.
  66
  67 * That Content-Type is given if there is content (CGI often has a
  68   default content type, but WSGI does not).
  69
  70 * That no Content-Type is given when there is no content (@@: is this
  71   too restrictive?)
  72
  73 * That the exc_info argument to start_response is a tuple or None.
  74
  75 * That all calls to the writer are with strings, and no other methods
  76   on the writer are accessed.
  77
  78 * That wsgi.input is used properly:
  79
  80   - .read() is called with zero or one argument
  81
  82   - That it returns a string
  83
  84   - That readline, readlines, and __iter__ return strings
  85
  86   - That .close() is not called
  87
  88   - No other methods are provided
  89
  90 * That wsgi.errors is used properly:
  91
  92   - .write() and .writelines() is called with a string
  93
  94   - That .close() is not called, and no other methods are provided.
  95
  96 * The response iterator:
  97
  98   - That it is not a string (it should be a list of a single string; a
  99     string will work, but perform horribly).
 100
 101   - That .next() returns a string
 102
 103   - That the iterator is not iterated over until start_response has
 104     been called (that can signal either a server or application
 105     error).
 106
 107   - That .close() is called (doesn't raise exception, only prints to
 108     sys.stderr, because we only know it isn't called when the object
 109     is garbage collected).
 110 """
 111 __all__ = ['validator']
 112
 113
 114 import re
 115 import sys
 116 from types import DictType, StringType, TupleType, ListType
 117 import warnings
 118
 119 header_re = re.compile(r'^[a-zA-Z][a-zA-Z0-9\-_]*$')
 120 bad_header_value_re = re.compile(r'[\000-\037]')
 121
 122 class WSGIWarning(Warning):
 123     """
 124     Raised in response to WSGI-spec-related warnings
 125     """
 126
 127 def assert_(cond, *args):
 128     if not cond:
 129         raise AssertionError(*args)
 130
 131 def validator(application):
 132
 133     """
 134     When applied between a WSGI server and a WSGI application, this
 135     middleware will check for WSGI compliancy on a number of levels.
 136     This middleware does not modify the request or response in any
 137     way, but will throw an AssertionError if anything seems off
 138     (except for a failure to close the application iterator, which
 139     will be printed to stderr -- there's no way to throw an exception
 140     at that point).
 141     """
 142
 143     def lint_app(*args, **kw):
 144         assert_(len(args) == 2, "Two arguments required")
 145         assert_(not kw, "No keyword arguments allowed")
 146         environ, start_response = args
 147
 148         check_environ(environ)
 149
 150         # We use this to check if the application returns without
 151         # calling start_response:
 152         start_response_started = []
 153
 154         def start_response_wrapper(*args, **kw):
 155             assert_(len(args) == 2 or len(args) == 3, (
 156                 "Invalid number of arguments: %s" % (args,)))
 157             assert_(not kw, "No keyword arguments allowed")
 158             status = args[0]
 159             headers = args[1]
 160             if len(args) == 3:
 161                 exc_info = args[2]
 162             else:
 163                 exc_info = None
 164
 165             check_status(status)
 166             check_headers(headers)
 167             check_content_type(status, headers)
 168             check_exc_info(exc_info)
 169
 170             start_response_started.append(None)
 171             return WriteWrapper(start_response(*args))
 172
 173         environ['wsgi.input'] = InputWrapper(environ['wsgi.input'])
 174         environ['wsgi.errors'] = ErrorWrapper(environ['wsgi.errors'])
 175
 176         iterator = application(environ, start_response_wrapper)
 177         assert_(iterator is not None and iterator != False,
 178             "The application must return an iterator, if only an empty list")
 179
 180         check_iterator(iterator)
 181
 182         return IteratorWrapper(iterator, start_response_started)
 183
 184     return lint_app
 185
 186 class InputWrapper:
 187
 188     def __init__(self, wsgi_input):
 189         self.input = wsgi_input
 190
 191     def read(self, *args):
 192         assert_(len(args) <= 1)
 193         v = self.input.read(*args)
 194         assert_(type(v) is type(""))
 195         return v
 196
 197     def readline(self):
 198         v = self.input.readline()
 199         assert_(type(v) is type(""))
 200         return v
 201
 202     def readlines(self, *args):
 203         assert_(len(args) <= 1)
 204         lines = self.input.readlines(*args)
 205         assert_(type(lines) is type([]))
 206         for line in lines:
 207             assert_(type(line) is type(""))
 208         return lines
 209
 210     def __iter__(self):
 211         while 1:
 212             line = self.readline()
 213             if not line:
 214                 return
 215             yield line
 216
 217     def close(self):
 218         assert_(0, "input.close() must not be called")
 219
 220 class ErrorWrapper:
 221
 222     def __init__(self, wsgi_errors):
 223         self.errors = wsgi_errors
 224
 225     def write(self, s):
 226         assert_(type(s) is type(""))
 227         self.errors.write(s)
 228
 229     def flush(self):
 230         self.errors.flush()
 231
 232     def writelines(self, seq):
 233         for line in seq:
 234             self.write(line)
 235
 236     def close(self):
 237         assert_(0, "errors.close() must not be called")
 238
 239 class WriteWrapper:
 240
 241     def __init__(self, wsgi_writer):
 242         self.writer = wsgi_writer
 243
 244     def __call__(self, s):
 245         assert_(type(s) is type(""))
 246         self.writer(s)
 247
 248 class PartialIteratorWrapper:
 249
 250     def __init__(self, wsgi_iterator):
 251         self.iterator = wsgi_iterator
 252
 253     def __iter__(self):
 254         # We want to make sure __iter__ is called
 255         return IteratorWrapper(self.iterator, None)
 256
 257 class IteratorWrapper:
 258
 259     def __init__(self, wsgi_iterator, check_start_response):
 260         self.original_iterator = wsgi_iterator
 261         self.iterator = iter(wsgi_iterator)
 262         self.closed = False
 263         self.check_start_response = check_start_response
 264
 265     def __iter__(self):
 266         return self
 267
 268     def next(self):
 269         assert_(not self.closed,
 270             "Iterator read after closed")
 271         v = self.iterator.next()
 272         if self.check_start_response is not None:
 273             assert_(self.check_start_response,
 274                 "The application returns and we started iterating over its body, but start_response has not yet been called")
 275             self.check_start_response = None
 276         return v
 277
 278     def close(self):
 279         self.closed = True
 280         if hasattr(self.original_iterator, 'close'):
 281             self.original_iterator.close()
 282
 283     def __del__(self):
 284         if not self.closed:
 285             sys.stderr.write(
 286                 "Iterator garbage collected without being closed")
 287         assert_(self.closed,
 288             "Iterator garbage collected without being closed")
 289
 290 def check_environ(environ):
 291     assert_(type(environ) is DictType,
 292         "Environment is not of the right type: %r (environment: %r)"
 293         % (type(environ), environ))
 294
 295     for key in ['REQUEST_METHOD', 'SERVER_NAME', 'SERVER_PORT',
 296                 'wsgi.version', 'wsgi.input', 'wsgi.errors',
 297                 'wsgi.multithread', 'wsgi.multiprocess',
 298                 'wsgi.run_once']:
 299         assert_(key in environ,
 300             "Environment missing required key: %r" % (key,))
 301
 302     for key in ['HTTP_CONTENT_TYPE', 'HTTP_CONTENT_LENGTH']:
 303         assert_(key not in environ,
 304             "Environment should not have the key: %s "
 305             "(use %s instead)" % (key, key[5:]))
 306
 307     if 'QUERY_STRING' not in environ:
 308         warnings.warn(
 309             'QUERY_STRING is not in the WSGI environment; the cgi '
 310             'module will use sys.argv when this variable is missing, '
 311             'so application errors are more likely',
 312             WSGIWarning)
 313
 314     for key in environ.keys():
 315         if '.' in key:
 316             # Extension, we don't care about its type
 317             continue
 318         assert_(type(environ[key]) is StringType,
 319             "Environmental variable %s is not a string: %r (value: %r)"
 320             % (key, type(environ[key]), environ[key]))
 321
 322     assert_(type(environ['wsgi.version']) is TupleType,
 323         "wsgi.version should be a tuple (%r)" % (environ['wsgi.version'],))
 324     assert_(environ['wsgi.url_scheme'] in ('http', 'https'),
 325         "wsgi.url_scheme unknown: %r" % environ['wsgi.url_scheme'])
 326
 327     check_input(environ['wsgi.input'])
 328     check_errors(environ['wsgi.errors'])
 329
 330     # @@: these need filling out:
 331     if environ['REQUEST_METHOD'] not in (
 332         'GET', 'HEAD', 'POST', 'OPTIONS','PUT','DELETE','TRACE'):
 333         warnings.warn(
 334             "Unknown REQUEST_METHOD: %r" % environ['REQUEST_METHOD'],
 335             WSGIWarning)
 336
 337     assert_(not environ.get('SCRIPT_NAME')
 338             or environ['SCRIPT_NAME'].startswith('/'),
 339         "SCRIPT_NAME doesn't start with /: %r" % environ['SCRIPT_NAME'])
 340     assert_(not environ.get('PATH_INFO')
 341             or environ['PATH_INFO'].startswith('/'),
 342         "PATH_INFO doesn't start with /: %r" % environ['PATH_INFO'])
 343     if environ.get('CONTENT_LENGTH'):
 344         assert_(int(environ['CONTENT_LENGTH']) >= 0,
 345             "Invalid CONTENT_LENGTH: %r" % environ['CONTENT_LENGTH'])
 346
 347     if not environ.get('SCRIPT_NAME'):
 348         assert_('PATH_INFO' in environ,
 349             "One of SCRIPT_NAME or PATH_INFO are required (PATH_INFO "
 350             "should at least be '/' if SCRIPT_NAME is empty)")
 351     assert_(environ.get('SCRIPT_NAME') != '/',
 352         "SCRIPT_NAME cannot be '/'; it should instead be '', and "
 353         "PATH_INFO should be '/'")
 354
 355 def check_input(wsgi_input):
 356     for attr in ['read', 'readline', 'readlines', '__iter__']:
 357         assert_(hasattr(wsgi_input, attr),
 358             "wsgi.input (%r) doesn't have the attribute %s"
 359             % (wsgi_input, attr))
 360
 361 def check_errors(wsgi_errors):
 362     for attr in ['flush', 'write', 'writelines']:
 363         assert_(hasattr(wsgi_errors, attr),
 364             "wsgi.errors (%r) doesn't have the attribute %s"
 365             % (wsgi_errors, attr))
 366
 367 def check_status(status):
 368     assert_(type(status) is StringType,
 369         "Status must be a string (not %r)" % status)
 370     # Implicitly check that we can turn it into an integer:
 371     status_code = status.split(None, 1)[0]
 372     assert_(len(status_code) == 3,
 373         "Status codes must be three characters: %r" % status_code)
 374     status_int = int(status_code)
 375     assert_(status_int >= 100, "Status code is invalid: %r" % status_int)
 376     if len(status) < 4 or status[3] != ' ':
 377         warnings.warn(
 378             "The status string (%r) should be a three-digit integer "
 379             "followed by a single space and a status explanation"
 380             % status, WSGIWarning)
 381
 382 def check_headers(headers):
 383     assert_(type(headers) is ListType,
 384         "Headers (%r) must be of type list: %r"
 385         % (headers, type(headers)))
 386     header_names = {}
 387     for item in headers:
 388         assert_(type(item) is TupleType,
 389             "Individual headers (%r) must be of type tuple: %r"
 390             % (item, type(item)))
 391         assert_(len(item) == 2)
 392         name, value = item
 393         assert_(name.lower() != 'status',
 394             "The Status header cannot be used; it conflicts with CGI "
 395             "script, and HTTP status is not given through headers "
 396             "(value: %r)." % value)
 397         header_names[name.lower()] = None
 398         assert_('\n' not in name and ':' not in name,
 399             "Header names may not contain ':' or '\\n': %r" % name)
 400         assert_(header_re.search(name), "Bad header name: %r" % name)
 401         assert_(not name.endswith('-') and not name.endswith('_'),
 402             "Names may not end in '-' or '_': %r" % name)
 403         if bad_header_value_re.search(value):
 404             assert_(0, "Bad header value: %r (bad char: %r)"
 405             % (value, bad_header_value_re.search(value).group(0)))
 406
 407 def check_content_type(status, headers):
 408     code = int(status.split(None, 1)[0])
 409     # @@: need one more person to verify this interpretation of RFC 2616
 410     #     http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
 411     NO_MESSAGE_BODY = (204, 304)
 412     for name, value in headers:
 413         if name.lower() == 'content-type':
 414             if code not in NO_MESSAGE_BODY:
 415                 return
 416             assert_(0, ("Content-Type header found in a %s response, "
 417                         "which must not return content.") % code)
 418     if code not in NO_MESSAGE_BODY:
 419         assert_(0, "No Content-Type header found in headers (%s)" % headers)
 420
 421 def check_exc_info(exc_info):
 422     assert_(exc_info is None or type(exc_info) is type(()),
 423         "exc_info (%r) is not a tuple: %r" % (exc_info, type(exc_info)))
 424     # More exc_info checks?
 425
 426 def check_iterator(iterator):
 427     # Technically a string is legal, which is why it's a really bad
 428     # idea, because it may cause the response to be returned
 429     # character-by-character
 430     assert_(not isinstance(iterator, str),
 431         "You should not return a string as your application iterator, "
 432         "instead return a single-item list containing that string.")