1# Test hashlib module
2#
3# $Id$
4#
5#  Copyright (C) 2005-2010   Gregory P. Smith (greg@krypto.org)
6#  Licensed to PSF under a Contributor Agreement.
7#
8
9import array
10import hashlib
11import itertools
12import sys
13try:
14    import threading
15except ImportError:
16    threading = None
17import unittest
18import warnings
19from test import test_support
20from test.test_support import _4G, precisionbigmemtest
21
22# Were we compiled --with-pydebug or with #define Py_DEBUG?
23COMPILED_WITH_PYDEBUG = hasattr(sys, 'gettotalrefcount')
24
25
26def hexstr(s):
27    import string
28    h = string.hexdigits
29    r = ''
30    for c in s:
31        i = ord(c)
32        r = r + h[(i >> 4) & 0xF] + h[i & 0xF]
33    return r
34
35
36class HashLibTestCase(unittest.TestCase):
37    supported_hash_names = ( 'md5', 'MD5', 'sha1', 'SHA1',
38                             'sha224', 'SHA224', 'sha256', 'SHA256',
39                             'sha384', 'SHA384', 'sha512', 'SHA512' )
40
41    _warn_on_extension_import = COMPILED_WITH_PYDEBUG
42
43    def _conditional_import_module(self, module_name):
44        """Import a module and return a reference to it or None on failure."""
45        try:
46            exec('import '+module_name)
47        except ImportError, error:
48            if self._warn_on_extension_import:
49                warnings.warn('Did a C extension fail to compile? %s' % error)
50        return locals().get(module_name)
51
52    def __init__(self, *args, **kwargs):
53        algorithms = set()
54        for algorithm in self.supported_hash_names:
55            algorithms.add(algorithm.lower())
56        self.constructors_to_test = {}
57        for algorithm in algorithms:
58            self.constructors_to_test[algorithm] = set()
59
60        # For each algorithm, test the direct constructor and the use
61        # of hashlib.new given the algorithm name.
62        for algorithm, constructors in self.constructors_to_test.items():
63            constructors.add(getattr(hashlib, algorithm))
64            def _test_algorithm_via_hashlib_new(data=None, _alg=algorithm):
65                if data is None:
66                    return hashlib.new(_alg)
67                return hashlib.new(_alg, data)
68            constructors.add(_test_algorithm_via_hashlib_new)
69
70        _hashlib = self._conditional_import_module('_hashlib')
71        if _hashlib:
72            # These two algorithms should always be present when this module
73            # is compiled.  If not, something was compiled wrong.
74            assert hasattr(_hashlib, 'openssl_md5')
75            assert hasattr(_hashlib, 'openssl_sha1')
76            for algorithm, constructors in self.constructors_to_test.items():
77                constructor = getattr(_hashlib, 'openssl_'+algorithm, None)
78                if constructor:
79                    constructors.add(constructor)
80
81        _md5 = self._conditional_import_module('_md5')
82        if _md5:
83            self.constructors_to_test['md5'].add(_md5.new)
84        _sha = self._conditional_import_module('_sha')
85        if _sha:
86            self.constructors_to_test['sha1'].add(_sha.new)
87        _sha256 = self._conditional_import_module('_sha256')
88        if _sha256:
89            self.constructors_to_test['sha224'].add(_sha256.sha224)
90            self.constructors_to_test['sha256'].add(_sha256.sha256)
91        _sha512 = self._conditional_import_module('_sha512')
92        if _sha512:
93            self.constructors_to_test['sha384'].add(_sha512.sha384)
94            self.constructors_to_test['sha512'].add(_sha512.sha512)
95
96        super(HashLibTestCase, self).__init__(*args, **kwargs)
97
98    def test_hash_array(self):
99        a = array.array("b", range(10))
100        constructors = self.constructors_to_test.itervalues()
101        for cons in itertools.chain.from_iterable(constructors):
102            c = cons(a)
103            c.hexdigest()
104
105    def test_algorithms_attribute(self):
106        self.assertEqual(hashlib.algorithms,
107            tuple([_algo for _algo in self.supported_hash_names if
108                                                _algo.islower()]))
109
110    def test_unknown_hash(self):
111        self.assertRaises(ValueError, hashlib.new, 'spam spam spam spam spam')
112        self.assertRaises(TypeError, hashlib.new, 1)
113
114    def test_get_builtin_constructor(self):
115        get_builtin_constructor = hashlib.__dict__[
116                '__get_builtin_constructor']
117        self.assertRaises(ValueError, get_builtin_constructor, 'test')
118        try:
119            import _md5
120        except ImportError:
121            pass
122        # This forces an ImportError for "import _md5" statements
123        sys.modules['_md5'] = None
124        try:
125            self.assertRaises(ValueError, get_builtin_constructor, 'md5')
126        finally:
127            if '_md5' in locals():
128                sys.modules['_md5'] = _md5
129            else:
130                del sys.modules['_md5']
131        self.assertRaises(TypeError, get_builtin_constructor, 3)
132
133    def test_hexdigest(self):
134        for name in self.supported_hash_names:
135            h = hashlib.new(name)
136            self.assertTrue(hexstr(h.digest()) == h.hexdigest())
137
138    def test_large_update(self):
139        aas = 'a' * 128
140        bees = 'b' * 127
141        cees = 'c' * 126
142        abcs = aas + bees + cees
143
144        for name in self.supported_hash_names:
145            m1 = hashlib.new(name)
146            m1.update(aas)
147            m1.update(bees)
148            m1.update(cees)
149
150            m2 = hashlib.new(name)
151            m2.update(abcs)
152            self.assertEqual(m1.digest(), m2.digest(), name+' update problem.')
153
154            m3 = hashlib.new(name, abcs)
155            self.assertEqual(m1.digest(), m3.digest(), name+' new problem.')
156
157    def check(self, name, data, digest):
158        constructors = self.constructors_to_test[name]
159        # 2 is for hashlib.name(...) and hashlib.new(name, ...)
160        self.assertGreaterEqual(len(constructors), 2)
161        for hash_object_constructor in constructors:
162            computed = hash_object_constructor(data).hexdigest()
163            self.assertEqual(
164                    computed, digest,
165                    "Hash algorithm %s constructed using %s returned hexdigest"
166                    " %r for %d byte input data that should have hashed to %r."
167                    % (name, hash_object_constructor,
168                       computed, len(data), digest))
169
170    def check_update(self, name, data, digest):
171        constructors = self.constructors_to_test[name]
172        # 2 is for hashlib.name(...) and hashlib.new(name, ...)
173        self.assertGreaterEqual(len(constructors), 2)
174        for hash_object_constructor in constructors:
175            h = hash_object_constructor()
176            h.update(data)
177            computed = h.hexdigest()
178            self.assertEqual(
179                    computed, digest,
180                    "Hash algorithm %s using %s when updated returned hexdigest"
181                    " %r for %d byte input data that should have hashed to %r."
182                    % (name, hash_object_constructor,
183                       computed, len(data), digest))
184
185    def check_unicode(self, algorithm_name):
186        # Unicode objects are not allowed as input.
187        expected = hashlib.new(algorithm_name, str(u'spam')).hexdigest()
188        self.check(algorithm_name, u'spam', expected)
189
190    def test_unicode(self):
191        # In python 2.x unicode is auto-encoded to the system default encoding
192        # when passed to hashlib functions.
193        self.check_unicode('md5')
194        self.check_unicode('sha1')
195        self.check_unicode('sha224')
196        self.check_unicode('sha256')
197        self.check_unicode('sha384')
198        self.check_unicode('sha512')
199
200    def test_case_md5_0(self):
201        self.check('md5', '', 'd41d8cd98f00b204e9800998ecf8427e')
202
203    def test_case_md5_1(self):
204        self.check('md5', 'abc', '900150983cd24fb0d6963f7d28e17f72')
205
206    def test_case_md5_2(self):
207        self.check('md5', 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789',
208                   'd174ab98d277d9f5a5611c2c9f419d9f')
209
210    @precisionbigmemtest(size=_4G + 5, memuse=1)
211    def test_case_md5_huge(self, size):
212        if size == _4G + 5:
213            try:
214                self.check('md5', 'A'*size, 'c9af2dff37468ce5dfee8f2cfc0a9c6d')
215            except OverflowError:
216                pass # 32-bit arch
217
218    @precisionbigmemtest(size=_4G + 5, memuse=1)
219    def test_case_md5_huge_update(self, size):
220        if size == _4G + 5:
221            try:
222                self.check_update('md5', 'A'*size,
223                        'c9af2dff37468ce5dfee8f2cfc0a9c6d')
224            except OverflowError:
225                pass # 32-bit arch
226
227    @precisionbigmemtest(size=_4G - 1, memuse=1)
228    def test_case_md5_uintmax(self, size):
229        if size == _4G - 1:
230            try:
231                self.check('md5', 'A'*size, '28138d306ff1b8281f1a9067e1a1a2b3')
232            except OverflowError:
233                pass # 32-bit arch
234
235    # use the three examples from Federal Information Processing Standards
236    # Publication 180-1, Secure Hash Standard,  1995 April 17
237    # http://www.itl.nist.gov/div897/pubs/fip180-1.htm
238
239    def test_case_sha1_0(self):
240        self.check('sha1', "",
241                   "da39a3ee5e6b4b0d3255bfef95601890afd80709")
242
243    def test_case_sha1_1(self):
244        self.check('sha1', "abc",
245                   "a9993e364706816aba3e25717850c26c9cd0d89d")
246
247    def test_case_sha1_2(self):
248        self.check('sha1', "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
249                   "84983e441c3bd26ebaae4aa1f95129e5e54670f1")
250
251    def test_case_sha1_3(self):
252        self.check('sha1', "a" * 1000000,
253                   "34aa973cd4c4daa4f61eeb2bdbad27316534016f")
254
255    @precisionbigmemtest(size=_4G + 5, memuse=1)
256    def test_case_sha1_huge(self, size):
257        if size == _4G + 5:
258            try:
259                self.check('sha1', 'A'*size,
260                        '87d745c50e6b2879ffa0fb2c930e9fbfe0dc9a5b')
261            except OverflowError:
262                pass # 32-bit arch
263
264    @precisionbigmemtest(size=_4G + 5, memuse=1)
265    def test_case_sha1_huge_update(self, size):
266        if size == _4G + 5:
267            try:
268                self.check_update('sha1', 'A'*size,
269                        '87d745c50e6b2879ffa0fb2c930e9fbfe0dc9a5b')
270            except OverflowError:
271                pass # 32-bit arch
272
273    # use the examples from Federal Information Processing Standards
274    # Publication 180-2, Secure Hash Standard,  2002 August 1
275    # http://csrc.nist.gov/publications/fips/fips180-2/fips180-2.pdf
276
277    def test_case_sha224_0(self):
278        self.check('sha224', "",
279          "d14a028c2a3a2bc9476102bb288234c415a2b01f828ea62ac5b3e42f")
280
281    def test_case_sha224_1(self):
282        self.check('sha224', "abc",
283          "23097d223405d8228642a477bda255b32aadbce4bda0b3f7e36c9da7")
284
285    def test_case_sha224_2(self):
286        self.check('sha224',
287          "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
288          "75388b16512776cc5dba5da1fd890150b0c6455cb4f58b1952522525")
289
290    def test_case_sha224_3(self):
291        self.check('sha224', "a" * 1000000,
292          "20794655980c91d8bbb4c1ea97618a4bf03f42581948b2ee4ee7ad67")
293
294
295    def test_case_sha256_0(self):
296        self.check('sha256', "",
297          "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855")
298
299    def test_case_sha256_1(self):
300        self.check('sha256', "abc",
301          "ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad")
302
303    def test_case_sha256_2(self):
304        self.check('sha256',
305          "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
306          "248d6a61d20638b8e5c026930c3e6039a33ce45964ff2167f6ecedd419db06c1")
307
308    def test_case_sha256_3(self):
309        self.check('sha256', "a" * 1000000,
310          "cdc76e5c9914fb9281a1c7e284d73e67f1809a48a497200e046d39ccc7112cd0")
311
312
313    def test_case_sha384_0(self):
314        self.check('sha384', "",
315          "38b060a751ac96384cd9327eb1b1e36a21fdb71114be07434c0cc7bf63f6e1da"+
316          "274edebfe76f65fbd51ad2f14898b95b")
317
318    def test_case_sha384_1(self):
319        self.check('sha384', "abc",
320          "cb00753f45a35e8bb5a03d699ac65007272c32ab0eded1631a8b605a43ff5bed"+
321          "8086072ba1e7cc2358baeca134c825a7")
322
323    def test_case_sha384_2(self):
324        self.check('sha384',
325                   "abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmn"+
326                   "hijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu",
327          "09330c33f71147e83d192fc782cd1b4753111b173b3b05d22fa08086e3b0f712"+
328          "fcc7c71a557e2db966c3e9fa91746039")
329
330    def test_case_sha384_3(self):
331        self.check('sha384', "a" * 1000000,
332          "9d0e1809716474cb086e834e310a4a1ced149e9c00f248527972cec5704c2a5b"+
333          "07b8b3dc38ecc4ebae97ddd87f3d8985")
334
335
336    def test_case_sha512_0(self):
337        self.check('sha512', "",
338          "cf83e1357eefb8bdf1542850d66d8007d620e4050b5715dc83f4a921d36ce9ce"+
339          "47d0d13c5d85f2b0ff8318d2877eec2f63b931bd47417a81a538327af927da3e")
340
341    def test_case_sha512_1(self):
342        self.check('sha512', "abc",
343          "ddaf35a193617abacc417349ae20413112e6fa4e89a97ea20a9eeee64b55d39a"+
344          "2192992a274fc1a836ba3c23a3feebbd454d4423643ce80e2a9ac94fa54ca49f")
345
346    def test_case_sha512_2(self):
347        self.check('sha512',
348                   "abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmn"+
349                   "hijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu",
350          "8e959b75dae313da8cf4f72814fc143f8f7779c6eb9f7fa17299aeadb6889018"+
351          "501d289e4900f7e4331b99dec4b5433ac7d329eeb6dd26545e96e55b874be909")
352
353    def test_case_sha512_3(self):
354        self.check('sha512', "a" * 1000000,
355          "e718483d0ce769644e2e42c7bc15b4638e1f98b13b2044285632a803afa973eb"+
356          "de0ff244877ea60a4cb0432ce577c31beb009c5c2c49aa2e4eadb217ad8cc09b")
357
358    @unittest.skipUnless(threading, 'Threading required for this test.')
359    @test_support.reap_threads
360    def test_threaded_hashing(self):
361        # Updating the same hash object from several threads at once
362        # using data chunk sizes containing the same byte sequences.
363        #
364        # If the internal locks are working to prevent multiple
365        # updates on the same object from running at once, the resulting
366        # hash will be the same as doing it single threaded upfront.
367        hasher = hashlib.sha1()
368        num_threads = 5
369        smallest_data = 'swineflu'
370        data = smallest_data*200000
371        expected_hash = hashlib.sha1(data*num_threads).hexdigest()
372
373        def hash_in_chunks(chunk_size, event):
374            index = 0
375            while index < len(data):
376                hasher.update(data[index:index+chunk_size])
377                index += chunk_size
378            event.set()
379
380        events = []
381        for threadnum in xrange(num_threads):
382            chunk_size = len(data) // (10**threadnum)
383            assert chunk_size > 0
384            assert chunk_size % len(smallest_data) == 0
385            event = threading.Event()
386            events.append(event)
387            threading.Thread(target=hash_in_chunks,
388                             args=(chunk_size, event)).start()
389
390        for event in events:
391            event.wait()
392
393        self.assertEqual(expected_hash, hasher.hexdigest())
394
395def test_main():
396    test_support.run_unittest(HashLibTestCase)
397
398if __name__ == "__main__":
399    test_main()
400