14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Tests from HTTP response parsing."""
17
"""Tests from HTTP response parsing.
19
The handle_response method read the response body of a GET request an returns
20
the corresponding RangeFile.
22
There are four different kinds of RangeFile:
23
- a whole file whose size is unknown, seen as a simple byte stream,
24
- a whole file whose size is known, we can't read past its end,
25
- a single range file, a part of a file with a start and a size,
26
- a multiple range file, several consecutive parts with known start offset
29
Some properties are common to all kinds:
30
- seek can only be forward (its really a socket underneath),
31
- read can't cross ranges,
32
- successive ranges are taken into account transparently,
34
- the expected pattern of use is either seek(offset)+read(size) or a single
35
read with no size specified. For multiple range files, multiple read() will
36
return the corresponding ranges, trying to read further will raise
19
40
from cStringIO import StringIO
22
from bzrlib import errors
23
from bzrlib.transport import http
24
from bzrlib.transport.http import response
25
from bzrlib.tests import TestCase
28
class TestResponseRange(TestCase):
29
"""Test the ResponseRange class."""
32
RR = response.ResponseRange
35
self.assertTrue(r1 < r2)
36
self.assertFalse(r1 > r2)
37
self.assertTrue(r1 < 5)
38
self.assertFalse(r2 < 5)
40
self.assertEqual(RR(0, 10, 5), RR(0, 10, 5))
41
self.assertNotEqual(RR(0, 10, 5), RR(0, 8, 5))
42
self.assertNotEqual(RR(0, 10, 5), RR(0, 10, 6))
44
def test_sort_list(self):
45
"""Ensure longer ranges are sorted after shorter ones"""
46
RR = response.ResponseRange
47
lst = [RR(3, 8, 0), 5, RR(3, 7, 0), 6]
49
self.assertEqual([RR(3,7,0), RR(3,8,0), 5, 6], lst)
52
class TestRangeFile(TestCase):
56
content = "abcdefghijklmnopqrstuvwxyz"
57
self.fp = response.RangeFile('foo', StringIO(content))
58
self.fp._add_range(0, 9, 0)
59
self.fp._add_range(20, 29, 10)
60
self.fp._add_range(30, 39, 15)
62
def test_valid_accesses(self):
63
"""Test so that valid accesses work to the file."""
65
self.assertEquals(self.fp.read(3), 'abc')
66
self.assertEquals(self.fp.read(3), 'def')
67
self.assertEquals(self.fp.tell(), 6)
69
self.assertEquals(self.fp.read(3), 'klm')
70
self.assertEquals(self.fp.read(2), 'no')
71
self.assertEquals(self.fp.tell(), 25)
72
# should wrap over to 30-39 entity
73
self.assertEquals(self.fp.read(3), 'pqr')
75
self.assertEquals(self.fp.read(3), 'def')
76
self.assertEquals(self.fp.tell(), 6)
78
def test_invalid_accesses(self):
79
"""Test so that invalid accesses trigger errors."""
81
self.assertRaises(errors.InvalidRange, self.fp.read, 2)
83
self.assertRaises(errors.InvalidRange, self.fp.read, 2)
85
self.assertRaises(errors.InvalidRange, self.fp.read, 2)
87
def test__finish_ranges(self):
88
"""Test that after RangeFile._finish_ranges the list is sorted."""
89
self.fp._add_range(1, 2, 3)
90
self.fp._add_range(8, 9, 10)
91
self.fp._add_range(3, 4, 5)
93
# TODO: jam 20060706 If we switch to inserting
94
# in sorted order, remove this test
95
self.assertNotEqual(self.fp._ranges, sorted(self.fp._ranges))
97
self.fp._finish_ranges()
98
self.assertEqual(self.fp._ranges, sorted(self.fp._ranges))
100
def test_seek_and_tell(self):
101
# Check for seeking before start
103
self.assertEqual(0, self.fp.tell())
106
self.assertEqual(5, self.fp.tell())
109
self.assertEqual(3, self.fp.tell())
111
# TODO: jam 20060706 following tests will fail if this
112
# is not true, and would be difficult to debug
113
# but it is a layering violation
114
self.assertEqual(39, self.fp._len)
117
self.assertEqual(39, self.fp.tell())
120
self.assertEqual(29, self.fp.tell())
122
self.assertRaises(ValueError, self.fp.seek, 0, 4)
123
self.assertRaises(ValueError, self.fp.seek, 0, -1)
126
class TestRegexes(TestCase):
128
def assertRegexMatches(self, groups, text):
129
"""Check that the regex matches and returns the right values"""
130
m = self.regex.match(text)
131
self.assertNotEqual(None, m, "text %s did not match regex" % (text,))
133
self.assertEqual(groups, m.groups())
135
def test_range_re(self):
136
"""Test that we match valid ranges."""
137
self.regex = response.HttpRangeResponse._CONTENT_RANGE_RE
138
self.assertRegexMatches(('bytes', '1', '10', '11'),
140
self.assertRegexMatches(('bytes', '1', '10', '11'),
142
self.assertRegexMatches(('bytes', '2123', '4242', '1231'),
143
'\tbytes 2123-4242/1231 ')
144
self.assertRegexMatches(('chars', '1', '2', '3'),
147
def test_content_type_re(self):
148
self.regex = response.HttpMultipartRangeResponse._CONTENT_TYPE_RE
149
self.assertRegexMatches(('', 'xxyyzz'),
150
'multipart/byteranges; boundary = xxyyzz')
151
self.assertRegexMatches(('', 'xxyyzz'),
152
'multipart/byteranges;boundary=xxyyzz')
153
self.assertRegexMatches(('', 'xx yy zz'),
154
' multipart/byteranges ; boundary= xx yy zz ')
155
self.assertRegexMatches(('"', 'xx yy zz'),
156
' multipart/byteranges ; boundary= "xx yy zz" ')
157
self.assertEqual(None,
159
' multipart/byteranges ; boundary= "xx yy zz '))
160
self.assertEqual(None,
162
' multipart/byteranges ; boundary= xx yy zz" '))
163
self.assertEqual(None,
164
self.regex.match('multipart byteranges;boundary=xx'))
170
Content-range: bytes 1-10/20\r
174
Content-Range: bytes 21-30/20\r
179
content-range: bytes 41-50/20\r
183
content-range: bytes 51-60/20\r
189
class TestHelpers(TestCase):
190
"""Test the helper functions"""
192
def test__parse_range(self):
193
"""Test that _parse_range acts reasonably."""
194
content = StringIO('')
195
parse_range = response.HttpRangeResponse._parse_range
196
self.assertEqual((1,2), parse_range('bytes 1-2/3'))
197
self.assertEqual((10,20), parse_range('bytes 10-20/2'))
199
self.assertRaises(errors.InvalidHttpRange, parse_range, 'char 1-3/2')
200
self.assertRaises(errors.InvalidHttpRange, parse_range, 'bytes a-3/2')
203
parse_range('bytes x-10/3', path='http://foo/bar')
204
except errors.InvalidHttpRange, e:
205
self.assertContainsRe(str(e), 'http://foo/bar')
206
self.assertContainsRe(str(e), 'bytes x-10/3')
208
self.fail('Did not raise InvalidHttpRange')
210
def test__parse_boundary_simple(self):
211
"""Test that _parse_boundary handles Content-type properly"""
212
parse_boundary = response.HttpMultipartRangeResponse._parse_boundary
213
m = parse_boundary(' multipart/byteranges; boundary=xxyyzz')
214
self.assertNotEqual(None, m)
215
# Check that the returned regex is capable of splitting simple_data
216
matches = list(m.finditer(simple_data))
217
self.assertEqual(4, len(matches))
219
# match.group() should be the content-range entry
220
# and match.end() should be the start of the content
221
self.assertEqual(' bytes 1-10/20', matches[0].group(1))
222
self.assertEqual(simple_data.find('1234567890'), matches[0].end())
223
self.assertEqual(' bytes 21-30/20', matches[1].group(1))
224
self.assertEqual(simple_data.find('abcdefghij'), matches[1].end())
225
self.assertEqual(' bytes 41-50/20', matches[2].group(1))
226
self.assertEqual(simple_data.find('zyxwvutsrq'), matches[2].end())
227
self.assertEqual(' bytes 51-60/20', matches[3].group(1))
228
self.assertEqual(simple_data.find('xxyyzz fbd'), matches[3].end())
230
def test__parse_boundary_invalid(self):
231
parse_boundary = response.HttpMultipartRangeResponse._parse_boundary
233
parse_boundary(' multipart/bytes;boundary=xxyyzz',
234
path='http://foo/bar')
235
except errors.InvalidHttpContentType, e:
236
self.assertContainsRe(str(e), 'http://foo/bar')
237
self.assertContainsRe(str(e), 'multipart/bytes;boundary=xxyyzz')
239
self.fail('Did not raise InvalidHttpContentType')
242
class TestHttpRangeResponse(TestCase):
244
def test_smoketest(self):
245
"""A basic test that HttpRangeResponse is reasonable."""
246
content = StringIO('0123456789')
247
f = response.HttpRangeResponse('http://foo', 'bytes 1-10/9', content)
248
self.assertEqual([response.ResponseRange(1,10,0)], f._ranges)
47
from bzrlib.transport.http import (
51
from bzrlib.tests.file_utils import (
56
class ReadSocket(object):
57
"""A socket-like object that can be given a predefined content."""
59
def __init__(self, data):
60
self.readfile = StringIO(data)
62
def makefile(self, mode='r', bufsize=None):
66
class FakeHTTPConnection(_urllib2_wrappers.HTTPConnection):
68
def __init__(self, sock):
69
_urllib2_wrappers.HTTPConnection.__init__(self, 'localhost')
70
# Set the socket to bypass the connection
74
"""Ignores the writes on the socket."""
78
class TestHTTPConnection(tests.TestCase):
80
def test_cleanup_pipe(self):
81
sock = ReadSocket("""HTTP/1.1 200 OK\r
82
Content-Type: text/plain; charset=UTF-8\r
87
conn = FakeHTTPConnection(sock)
88
# Simulate the request sending so that the connection will be able to
90
conn.putrequest('GET', 'http://localhost/fictious')
92
# Now, get the response
93
resp = conn.getresponse()
94
# Read part of the response
95
self.assertEquals('0123456789\n', resp.read(11))
96
# Override the thresold to force the warning emission
97
conn._range_warning_thresold = 6 # There are 7 bytes pending
99
self.assertContainsRe(self._get_log(keep_log_file=True),
100
'Got a 200 response when asking')
103
class TestRangeFileMixin(object):
104
"""Tests for accessing the first range in a RangeFile."""
106
# A simple string used to represent a file part (also called a range), in
107
# which offsets are easy to calculate for test writers. It's used as a
108
# building block with slight variations but basically 'a' is the first char
109
# of the range and 'z' is the last.
110
alpha = 'abcdefghijklmnopqrstuvwxyz'
112
def test_can_read_at_first_access(self):
113
"""Test that the just created file can be read."""
114
self.assertEquals(self.alpha, self._file.read())
116
def test_seek_read(self):
117
"""Test seek/read inside the range."""
119
start = self.first_range_start
120
# Before any use, tell() should be at the range start
121
self.assertEquals(start, f.tell())
122
cur = start # For an overall offset assertion
125
self.assertEquals('def', f.read(3))
129
self.assertEquals('klmn', f.read(4))
131
# read(0) in the middle of a range
132
self.assertEquals('', f.read(0))
136
self.assertEquals(here, f.tell())
137
self.assertEquals(cur, f.tell())
139
def test_read_zero(self):
141
start = self.first_range_start
142
self.assertEquals('', f.read(0))
144
self.assertEquals('', f.read(0))
146
def test_seek_at_range_end(self):
150
def test_read_at_range_end(self):
151
"""Test read behaviour at range end."""
153
self.assertEquals(self.alpha, f.read())
154
self.assertEquals('', f.read(0))
155
self.assertRaises(errors.InvalidRange, f.read, 1)
157
def test_unbounded_read_after_seek(self):
160
# Should not cross ranges
161
self.assertEquals('yz', f.read())
163
def test_seek_backwards(self):
165
start = self.first_range_start
168
self.assertRaises(errors.InvalidRange, f.seek, start + 5)
170
def test_seek_outside_single_range(self):
172
if f._size == -1 or f._boundary is not None:
173
raise tests.TestNotApplicable('Needs a fully defined range')
174
# Will seek past the range and then errors out
175
self.assertRaises(errors.InvalidRange,
176
f.seek, self.first_range_start + 27)
178
def test_read_past_end_of_range(self):
181
raise tests.TestNotApplicable("Can't check an unknown size")
182
start = self.first_range_start
184
self.assertRaises(errors.InvalidRange, f.read, 10)
186
def test_seek_from_end(self):
187
"""Test seeking from the end of the file.
189
The semantic is unclear in case of multiple ranges. Seeking from end
190
exists only for the http transports, cannot be used if the file size is
191
unknown and is not used in bzrlib itself. This test must be (and is)
192
overridden by daughter classes.
194
Reading from end makes sense only when a range has been requested from
195
the end of the file (see HttpTransportBase._get() when using the
196
'tail_amount' parameter). The HTTP response can only be a whole file or
201
self.assertEquals('yz', f.read())
204
class TestRangeFileSizeUnknown(tests.TestCase, TestRangeFileMixin):
205
"""Test a RangeFile for a whole file whose size is not known."""
208
super(TestRangeFileSizeUnknown, self).setUp()
209
self._file = response.RangeFile('Whole_file_size_known',
210
StringIO(self.alpha))
211
# We define no range, relying on RangeFile to provide default values
212
self.first_range_start = 0 # It's the whole file
214
def test_seek_from_end(self):
215
"""See TestRangeFileMixin.test_seek_from_end.
217
The end of the file can't be determined since the size is unknown.
219
self.assertRaises(errors.InvalidRange, self._file.seek, -1, 2)
221
def test_read_at_range_end(self):
222
"""Test read behaviour at range end."""
224
self.assertEquals(self.alpha, f.read())
225
self.assertEquals('', f.read(0))
226
self.assertEquals('', f.read(1))
229
class TestRangeFileSizeKnown(tests.TestCase, TestRangeFileMixin):
230
"""Test a RangeFile for a whole file whose size is known."""
233
super(TestRangeFileSizeKnown, self).setUp()
234
self._file = response.RangeFile('Whole_file_size_known',
235
StringIO(self.alpha))
236
self._file.set_range(0, len(self.alpha))
237
self.first_range_start = 0 # It's the whole file
240
class TestRangeFileSingleRange(tests.TestCase, TestRangeFileMixin):
241
"""Test a RangeFile for a single range."""
244
super(TestRangeFileSingleRange, self).setUp()
245
self._file = response.RangeFile('Single_range_file',
246
StringIO(self.alpha))
247
self.first_range_start = 15
248
self._file.set_range(self.first_range_start, len(self.alpha))
251
def test_read_before_range(self):
252
# This can't occur under normal circumstances, we have to force it
254
f._pos = 0 # Force an invalid pos
255
self.assertRaises(errors.InvalidRange, f.read, 2)
258
class TestRangeFileMultipleRanges(tests.TestCase, TestRangeFileMixin):
259
"""Test a RangeFile for multiple ranges.
261
The RangeFile used for the tests contains three ranges:
263
- at offset 25: alpha
264
- at offset 100: alpha
265
- at offset 126: alpha.upper()
267
The two last ranges are contiguous. This only rarely occurs (should not in
268
fact) in real uses but may lead to hard to track bugs.
271
# The following is used to represent the boundary paramter defined
272
# in HTTP response headers and the boundary lines that separate
275
boundary = "separation"
278
super(TestRangeFileMultipleRanges, self).setUp()
280
boundary = self.boundary
283
self.first_range_start = 25
284
file_size = 200 # big enough to encompass all ranges
285
for (start, part) in [(self.first_range_start, self.alpha),
286
# Two contiguous ranges
288
(126, self.alpha.upper())]:
289
content += self._multipart_byterange(part, start, boundary,
292
content += self._boundary_line()
294
self._file = response.RangeFile('Multiple_ranges_file',
296
self.set_file_boundary()
298
def _boundary_line(self):
299
"""Helper to build the formatted boundary line."""
300
return '--' + self.boundary + '\r\n'
302
def set_file_boundary(self):
303
# Ranges are set by decoding the range headers, the RangeFile user is
304
# supposed to call the following before using seek or read since it
305
# requires knowing the *response* headers (in that case the boundary
306
# which is part of the Content-Type header).
307
self._file.set_boundary(self.boundary)
309
def _multipart_byterange(self, data, offset, boundary, file_size='*'):
310
"""Encode a part of a file as a multipart/byterange MIME type.
312
When a range request is issued, the HTTP response body can be
313
decomposed in parts, each one representing a range (start, size) in a
316
:param data: The payload.
317
:param offset: where data starts in the file
318
:param boundary: used to separate the parts
319
:param file_size: the size of the file containing the range (default to
322
:return: a string containing the data encoded as it will appear in the
325
bline = self._boundary_line()
326
# Each range begins with a boundary line
328
# A range is described by a set of headers, but only 'Content-Range' is
329
# required for our implementation (TestHandleResponse below will
330
# exercise ranges with multiple or missing headers')
331
range += 'Content-Range: bytes %d-%d/%d\r\n' % (offset,
335
# Finally the raw bytes
339
def test_read_all_ranges(self):
341
self.assertEquals(self.alpha, f.read()) # Read first range
342
f.seek(100) # Trigger the second range recognition
343
self.assertEquals(self.alpha, f.read()) # Read second range
344
self.assertEquals(126, f.tell())
345
f.seek(126) # Start of third range which is also the current pos !
346
self.assertEquals('A', f.read(1))
348
self.assertEquals('LMN', f.read(3))
350
def test_seek_from_end(self):
351
"""See TestRangeFileMixin.test_seek_from_end."""
352
# The actual implementation will seek from end for the first range only
353
# and then fail. Since seeking from end is intended to be used for a
354
# single range only anyway, this test just document the actual
358
self.assertEquals('yz', f.read())
359
self.assertRaises(errors.InvalidRange, f.seek, -2, 2)
361
def test_seek_into_void(self):
363
start = self.first_range_start
365
# Seeking to a point between two ranges is possible (only once) but
366
# reading there is forbidden
368
# We crossed a range boundary, so now the file is positioned at the
369
# start of the new range (i.e. trying to seek below 100 will error out)
373
def test_seek_across_ranges(self):
375
start = self.first_range_start
376
f.seek(126) # skip the two first ranges
377
self.assertEquals('AB', f.read(2))
379
def test_checked_read_dont_overflow_buffers(self):
381
start = self.first_range_start
382
# We force a very low value to exercise all code paths in _checked_read
383
f._discarded_buf_size = 8
384
f.seek(126) # skip the two first ranges
385
self.assertEquals('AB', f.read(2))
387
def test_seek_twice_between_ranges(self):
389
start = self.first_range_start
390
f.seek(start + 40) # Past the first range but before the second
391
# Now the file is positioned at the second range start (100)
392
self.assertRaises(errors.InvalidRange, f.seek, start + 41)
394
def test_seek_at_range_end(self):
395
"""Test seek behavior at range end."""
401
def test_read_at_range_end(self):
403
self.assertEquals(self.alpha, f.read())
404
self.assertEquals(self.alpha, f.read())
405
self.assertEquals(self.alpha.upper(), f.read())
406
self.assertRaises(errors.InvalidHttpResponse, f.read, 1)
409
class TestRangeFileMultipleRangesQuotedBoundaries(TestRangeFileMultipleRanges):
410
"""Perform the same tests as TestRangeFileMultipleRanges, but uses
411
an angle-bracket quoted boundary string like IIS 6.0 and 7.0
412
(but not IIS 5, which breaks the RFC in a different way
413
by using square brackets, not angle brackets)
415
This reveals a bug caused by
417
- The bad implementation of RFC 822 unquoting in Python (angles are not
418
quotes), coupled with
420
- The bad implementation of RFC 2046 in IIS (angles are not permitted chars
424
# The boundary as it appears in boundary lines
425
# IIS 6 and 7 use this value
426
_boundary_trimmed = "q1w2e3r4t5y6u7i8o9p0zaxscdvfbgnhmjklkl"
427
boundary = '<' + _boundary_trimmed + '>'
429
def set_file_boundary(self):
430
# Emulate broken rfc822.unquote() here by removing angles
431
self._file.set_boundary(self._boundary_trimmed)
434
class TestRangeFileVarious(tests.TestCase):
435
"""Tests RangeFile aspects not covered elsewhere."""
437
def test_seek_whence(self):
438
"""Test the seek whence parameter values."""
439
f = response.RangeFile('foo', StringIO('abc'))
251
self.assertRaises(errors.InvalidRange, f.read, 2)
253
self.assertEqual('012345', f.read(6))
255
def test_invalid(self):
257
f = response.HttpRangeResponse('http://foo', 'bytes x-10/9',
258
StringIO('0123456789'))
259
except errors.InvalidHttpRange, e:
260
self.assertContainsRe(str(e), 'http://foo')
261
self.assertContainsRe(str(e), 'bytes x-10/9')
263
self.fail('Failed to raise InvalidHttpRange')
266
class TestHttpMultipartRangeResponse(TestCase):
267
"""Test the handling of multipart range responses"""
269
def test_simple(self):
270
content = StringIO(simple_data)
271
multi = response.HttpMultipartRangeResponse('http://foo',
272
'multipart/byteranges; boundary = xxyyzz', content)
274
self.assertEqual(4, len(multi._ranges))
277
self.assertEqual('1234567890', multi.read(10))
279
self.assertEqual('abcdefghij', multi.read(10))
281
self.assertEqual('zyxwvutsrq', multi.read(10))
283
self.assertEqual('xxyyzz fbd', multi.read(10))
284
# TODO: jam 20060706 Currently RangeFile does not support
285
# reading across ranges. Consider adding it.
287
# self.assertEqual('zyxwvutsrqxxyyzz fbd', multi.read(20))
288
self.assertRaises(errors.InvalidRange, multi.read, 20)
291
self.assertRaises(errors.InvalidRange, multi.read, 11)
293
self.assertRaises(errors.InvalidRange, multi.read, 10)
295
def test_invalid(self):
296
content = StringIO('')
298
response.HttpMultipartRangeResponse('http://foo',
299
'multipart/byte;boundary=invalid', content)
300
except errors.InvalidHttpContentType, e:
301
self.assertContainsRe(str(e), 'http://foo')
302
self.assertContainsRe(str(e), 'multipart/byte;')
444
self.assertRaises(ValueError, f.seek, 0, 14)
446
def test_range_syntax(self):
447
"""Test the Content-Range scanning."""
449
f = response.RangeFile('foo', StringIO())
451
def ok(expected, header_value):
452
f.set_range_from_header(header_value)
453
# Slightly peek under the covers to get the size
454
self.assertEquals(expected, (f.tell(), f._size))
456
ok((1, 10), 'bytes 1-10/11')
457
ok((1, 10), 'bytes 1-10/*')
458
ok((12, 2), '\tbytes 12-13/*')
459
ok((28, 1), ' bytes 28-28/*')
460
ok((2123, 2120), 'bytes 2123-4242/12310')
461
ok((1, 10), 'bytes 1-10/ttt') # We don't check total (ttt)
463
def nok(header_value):
464
self.assertRaises(errors.InvalidHttpRange,
465
f.set_range_from_header, header_value)
469
nok('bytes xx-yyy/zzz')
470
nok('bytes xx-12/zzz')
471
nok('bytes 11-yy/zzz')
305
475
# Taken from real request responses
479
# This should be in test_http.py, but the headers we
480
# want to parse are here
481
class TestExtractHeader(TestCase):
483
def use_response(self, response):
484
self.headers = http._extract_headers(response[1], 'http://foo')
486
def check_header(self, header, value):
487
self.assertEqual(value, self.headers[header])
489
def test_full_text(self):
490
self.use_response(_full_text_response)
492
self.check_header('Date', 'Tue, 11 Jul 2006 04:32:56 GMT')
493
self.check_header('date', 'Tue, 11 Jul 2006 04:32:56 GMT')
494
self.check_header('Content-Length', '35')
495
self.check_header('Content-Type', 'text/plain; charset=UTF-8')
496
self.check_header('content-type', 'text/plain; charset=UTF-8')
498
def test_missing_response(self):
499
self.use_response(_missing_response)
501
self.check_header('Content-Length', '336')
502
self.check_header('Content-Type', 'text/html; charset=iso-8859-1')
504
def test_single_range(self):
505
self.use_response(_single_range_response)
507
self.check_header('Content-Length', '100')
508
self.check_header('Content-Range', 'bytes 100-199/93890')
509
self.check_header('Content-Type', 'text/plain; charset=UTF-8')
511
def test_multi_range(self):
512
self.use_response(_multipart_range_response)
514
self.check_header('Content-Length', '1534')
515
self.check_header('Content-Type',
516
'multipart/byteranges; boundary=418470f848b63279b')
518
def test_multi_squid_range(self):
519
self.use_response(_multipart_squid_range_response)
521
self.check_header('Content-Length', '598')
522
self.check_header('Content-Type',
523
'multipart/byteranges; '\
524
'boundary="squid/2.5.STABLE12:C99323425AD4FE26F726261FA6C24196"')
526
def test_redirect(self):
527
"""We default to returning the last group of headers in the file."""
528
self.use_response(_redirect_response)
529
self.check_header('Content-Range', 'bytes 8623075-8623499/8623500')
530
self.check_header('Content-Type', 'text/plain; charset=UTF-8')
532
def test_empty(self):
533
self.assertRaises(errors.InvalidHttpResponse,
534
http._extract_headers, '', 'bad url')
536
def test_no_opening_http(self):
537
# Remove the HTTP line from the header
538
first, txt = _full_text_response[1].split('\r\n', 1)
539
self.assertRaises(errors.InvalidHttpResponse,
540
http._extract_headers, txt, 'missing HTTTP')
542
def test_trailing_whitespace(self):
543
# Test that we ignore bogus whitespace on the end
544
code, txt, body = _full_text_response
545
txt += '\r\n\n\n\n\n'
546
self.use_response((code, txt, body))
548
self.check_header('Date', 'Tue, 11 Jul 2006 04:32:56 GMT')
549
self.check_header('Content-Length', '35')
550
self.check_header('Content-Type', 'text/plain; charset=UTF-8')
552
def test_trailing_non_http(self):
553
# Test that we ignore bogus stuff on the end
554
code, txt, body = _full_text_response
555
txt = txt + 'Foo: Bar\r\nBaz: Bling\r\n\r\n'
556
self.use_response((code, txt, body))
558
self.check_header('Date', 'Tue, 11 Jul 2006 04:32:56 GMT')
559
self.check_header('Content-Length', '35')
560
self.check_header('Content-Type', 'text/plain; charset=UTF-8')
561
self.assertRaises(KeyError, self.headers.__getitem__, 'Foo')
563
def test_extra_whitespace(self):
564
# Test that we read an HTTP response, even with extra whitespace
565
code, txt, body = _redirect_response
566
# Find the second HTTP location
567
loc = txt.find('HTTP', 5)
568
txt = txt[:loc] + '\r\n\n' + txt[loc:]
569
self.use_response((code, txt, body))
570
self.check_header('Content-Range', 'bytes 8623075-8623499/8623500')
571
self.check_header('Content-Type', 'text/plain; charset=UTF-8')
574
class TestHandleResponse(TestCase):
671
_multipart_no_content_range = (206, """HTTP/1.0 206 Partial Content\r
672
Content-Type: multipart/byteranges; boundary=THIS_SEPARATES\r
673
Content-Length: 598\r
678
Content-Type: text/plain\r
685
_multipart_no_boundary = (206, """HTTP/1.0 206 Partial Content\r
686
Content-Type: multipart/byteranges; boundary=THIS_SEPARATES\r
687
Content-Length: 598\r
692
Content-Type: text/plain\r
693
Content-Range: bytes 0-18/18672\r
697
The range ended at the line above, this text is garbage instead of a boundary
702
class TestHandleResponse(tests.TestCase):
704
def _build_HTTPMessage(self, raw_headers):
705
status_and_headers = StringIO(raw_headers)
706
# Get rid of the status line
707
status_and_headers.readline()
708
msg = httplib.HTTPMessage(status_and_headers)
576
711
def get_response(self, a_response):
577
712
"""Process a supplied response, and return the result."""
578
headers = http._extract_headers(a_response[1], 'http://foo')
579
return response.handle_response('http://foo', a_response[0], headers,
713
code, raw_headers, body = a_response
714
msg = self._build_HTTPMessage(raw_headers)
715
return response.handle_response('http://foo', code, msg,
580
716
StringIO(a_response[2]))
582
718
def test_full_text(self):
622
760
def test_invalid_response(self):
623
761
self.assertRaises(errors.InvalidHttpResponse,
624
self.get_response, _invalid_response)
762
self.get_response, _invalid_response)
626
764
def test_full_text_no_content_type(self):
627
765
# We should not require Content-Type for a full response
628
a_response = _full_text_response
629
headers = http._extract_headers(a_response[1], 'http://foo')
630
del headers['Content-Type']
631
out = response.handle_response('http://foo', a_response[0], headers,
632
StringIO(a_response[2]))
633
self.assertEqual(_full_text_response[2], out.read())
635
def test_missing_no_content_type(self):
636
# Without Content-Type we should still raise NoSuchFile on a 404
637
a_response = _missing_response
638
headers = http._extract_headers(a_response[1], 'http://missing')
639
del headers['Content-Type']
640
self.assertRaises(errors.NoSuchFile,
641
response.handle_response, 'http://missing', a_response[0], headers,
642
StringIO(a_response[2]))
644
def test_missing_content_type(self):
645
a_response = _single_range_response
646
headers = http._extract_headers(a_response[1], 'http://nocontent')
647
del headers['Content-Type']
648
self.assertRaises(errors.InvalidHttpContentType,
649
response.handle_response, 'http://nocontent', a_response[0],
650
headers, StringIO(a_response[2]))
766
code, raw_headers, body = _full_text_response_no_content_type
767
msg = self._build_HTTPMessage(raw_headers)
768
out = response.handle_response('http://foo', code, msg, StringIO(body))
769
self.assertEqual(body, out.read())
771
def test_full_text_no_content_length(self):
772
code, raw_headers, body = _full_text_response_no_content_length
773
msg = self._build_HTTPMessage(raw_headers)
774
out = response.handle_response('http://foo', code, msg, StringIO(body))
775
self.assertEqual(body, out.read())
652
777
def test_missing_content_range(self):
653
a_response = _single_range_response
654
headers = http._extract_headers(a_response[1], 'http://nocontent')
655
del headers['Content-Range']
656
self.assertRaises(errors.InvalidHttpResponse,
657
response.handle_response, 'http://nocontent', a_response[0],
658
headers, StringIO(a_response[2]))
778
code, raw_headers, body = _single_range_no_content_range
779
msg = self._build_HTTPMessage(raw_headers)
780
self.assertRaises(errors.InvalidHttpResponse,
781
response.handle_response,
782
'http://bogus', code, msg, StringIO(body))
784
def test_multipart_no_content_range(self):
785
code, raw_headers, body = _multipart_no_content_range
786
msg = self._build_HTTPMessage(raw_headers)
787
self.assertRaises(errors.InvalidHttpResponse,
788
response.handle_response,
789
'http://bogus', code, msg, StringIO(body))
791
def test_multipart_no_boundary(self):
792
out = self.get_response(_multipart_no_boundary)
793
out.read() # Read the whole range
794
# Fail to find the boundary line
795
self.assertRaises(errors.InvalidHttpResponse, out.seek, 1, 1)
798
class TestRangeFileSizeReadLimited(tests.TestCase):
799
"""Test RangeFile _max_read_size functionality which limits the size of
800
read blocks to prevent MemoryError messages in socket.recv.
804
# create a test datablock larger than _max_read_size.
805
chunk_size = response.RangeFile._max_read_size
806
test_pattern = '0123456789ABCDEF'
807
self.test_data = test_pattern * (3 * chunk_size / len(test_pattern))
808
self.test_data_len = len(self.test_data)
810
def test_max_read_size(self):
811
"""Read data in blocks and verify that the reads are not larger than
812
the maximum read size.
814
# retrieve data in large blocks from response.RangeFile object
815
mock_read_file = FakeReadFile(self.test_data)
816
range_file = response.RangeFile('test_max_read_size', mock_read_file)
817
response_data = range_file.read(self.test_data_len)
819
# verify read size was equal to the maximum read size
820
self.assertTrue(mock_read_file.get_max_read_size() > 0)
821
self.assertEqual(mock_read_file.get_max_read_size(),
822
response.RangeFile._max_read_size)
823
self.assertEqual(mock_read_file.get_read_count(), 3)
825
# report error if the data wasn't equal (we only report the size due
826
# to the length of the data)
827
if response_data != self.test_data:
828
message = "Data not equal. Expected %d bytes, received %d."
829
self.fail(message % (len(response_data), self.test_data_len))
added
removed
bzrlib/tests/test_http_response.py
