1
# Copyright (C) 2006-2010, 2012, 2013, 2016 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
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
41
import http.client as http_client
42
except ImportError: # python < 3
43
import httplib as http_client
49
from ..sixish import (
52
from ..transport.http import (
56
from .file_utils import (
61
class ReadSocket(object):
62
"""A socket-like object that can be given a predefined content."""
64
def __init__(self, data):
65
self.readfile = BytesIO(data)
67
def makefile(self, mode='r', bufsize=None):
71
class FakeHTTPConnection(_urllib2_wrappers.HTTPConnection):
73
def __init__(self, sock):
74
_urllib2_wrappers.HTTPConnection.__init__(self, 'localhost')
75
# Set the socket to bypass the connection
79
"""Ignores the writes on the socket."""
83
class TestResponseFileIter(tests.TestCase):
85
def test_iter_empty(self):
86
f = response.ResponseFile('empty', BytesIO())
87
self.assertEqual([], list(f))
89
def test_iter_many(self):
90
f = response.ResponseFile('many', BytesIO(b'0\n1\nboo!\n'))
91
self.assertEqual([b'0\n', b'1\n', b'boo!\n'], list(f))
94
class TestHTTPConnection(tests.TestCase):
96
def test_cleanup_pipe(self):
97
sock = ReadSocket(b"""HTTP/1.1 200 OK\r
98
Content-Type: text/plain; charset=UTF-8\r
103
conn = FakeHTTPConnection(sock)
104
# Simulate the request sending so that the connection will be able to
106
conn.putrequest('GET', 'http://localhost/fictious')
108
# Now, get the response
109
resp = conn.getresponse()
110
# Read part of the response
111
self.assertEqual(b'0123456789\n', resp.read(11))
112
# Override the thresold to force the warning emission
113
conn._range_warning_thresold = 6 # There are 7 bytes pending
115
self.assertContainsRe(self.get_log(), 'Got a 200 response when asking')
118
class TestRangeFileMixin(object):
119
"""Tests for accessing the first range in a RangeFile."""
121
# A simple string used to represent a file part (also called a range), in
122
# which offsets are easy to calculate for test writers. It's used as a
123
# building block with slight variations but basically 'a' is the first char
124
# of the range and 'z' is the last.
125
alpha = b'abcdefghijklmnopqrstuvwxyz'
127
def test_can_read_at_first_access(self):
128
"""Test that the just created file can be read."""
129
self.assertEqual(self.alpha, self._file.read())
131
def test_seek_read(self):
132
"""Test seek/read inside the range."""
134
start = self.first_range_start
135
# Before any use, tell() should be at the range start
136
self.assertEqual(start, f.tell())
137
cur = start # For an overall offset assertion
140
self.assertEqual(b'def', f.read(3))
144
self.assertEqual(b'klmn', f.read(4))
146
# read(0) in the middle of a range
147
self.assertEqual(b'', f.read(0))
151
self.assertEqual(here, f.tell())
152
self.assertEqual(cur, f.tell())
154
def test_read_zero(self):
156
self.assertEqual(b'', f.read(0))
158
self.assertEqual(b'', f.read(0))
160
def test_seek_at_range_end(self):
164
def test_read_at_range_end(self):
165
"""Test read behaviour at range end."""
167
self.assertEqual(self.alpha, f.read())
168
self.assertEqual(b'', f.read(0))
169
self.assertRaises(errors.InvalidRange, f.read, 1)
171
def test_unbounded_read_after_seek(self):
174
# Should not cross ranges
175
self.assertEqual(b'yz', f.read())
177
def test_seek_backwards(self):
179
start = self.first_range_start
182
self.assertRaises(errors.InvalidRange, f.seek, start + 5)
184
def test_seek_outside_single_range(self):
186
if f._size == -1 or f._boundary is not None:
187
raise tests.TestNotApplicable('Needs a fully defined range')
188
# Will seek past the range and then errors out
189
self.assertRaises(errors.InvalidRange,
190
f.seek, self.first_range_start + 27)
192
def test_read_past_end_of_range(self):
195
raise tests.TestNotApplicable("Can't check an unknown size")
196
start = self.first_range_start
198
self.assertRaises(errors.InvalidRange, f.read, 10)
200
def test_seek_from_end(self):
201
"""Test seeking from the end of the file.
203
The semantic is unclear in case of multiple ranges. Seeking from end
204
exists only for the http transports, cannot be used if the file size is
205
unknown and is not used in breezy itself. This test must be (and is)
206
overridden by daughter classes.
208
Reading from end makes sense only when a range has been requested from
209
the end of the file (see HttpTransportBase._get() when using the
210
'tail_amount' parameter). The HTTP response can only be a whole file or
215
self.assertEqual(b'yz', f.read())
218
class TestRangeFileSizeUnknown(tests.TestCase, TestRangeFileMixin):
219
"""Test a RangeFile for a whole file whose size is not known."""
222
super(TestRangeFileSizeUnknown, self).setUp()
223
self._file = response.RangeFile('Whole_file_size_known',
225
# We define no range, relying on RangeFile to provide default values
226
self.first_range_start = 0 # It's the whole file
228
def test_seek_from_end(self):
229
"""See TestRangeFileMixin.test_seek_from_end.
231
The end of the file can't be determined since the size is unknown.
233
self.assertRaises(errors.InvalidRange, self._file.seek, -1, 2)
235
def test_read_at_range_end(self):
236
"""Test read behaviour at range end."""
238
self.assertEqual(self.alpha, f.read())
239
self.assertEqual(b'', f.read(0))
240
self.assertEqual(b'', f.read(1))
243
class TestRangeFileSizeKnown(tests.TestCase, TestRangeFileMixin):
244
"""Test a RangeFile for a whole file whose size is known."""
247
super(TestRangeFileSizeKnown, self).setUp()
248
self._file = response.RangeFile('Whole_file_size_known',
250
self._file.set_range(0, len(self.alpha))
251
self.first_range_start = 0 # It's the whole file
254
class TestRangeFileSingleRange(tests.TestCase, TestRangeFileMixin):
255
"""Test a RangeFile for a single range."""
258
super(TestRangeFileSingleRange, self).setUp()
259
self._file = response.RangeFile('Single_range_file',
261
self.first_range_start = 15
262
self._file.set_range(self.first_range_start, len(self.alpha))
265
def test_read_before_range(self):
266
# This can't occur under normal circumstances, we have to force it
268
f._pos = 0 # Force an invalid pos
269
self.assertRaises(errors.InvalidRange, f.read, 2)
272
class TestRangeFileMultipleRanges(tests.TestCase, TestRangeFileMixin):
273
"""Test a RangeFile for multiple ranges.
275
The RangeFile used for the tests contains three ranges:
277
- at offset 25: alpha
278
- at offset 100: alpha
279
- at offset 126: alpha.upper()
281
The two last ranges are contiguous. This only rarely occurs (should not in
282
fact) in real uses but may lead to hard to track bugs.
285
# The following is used to represent the boundary paramter defined
286
# in HTTP response headers and the boundary lines that separate
289
boundary = b"separation"
292
super(TestRangeFileMultipleRanges, self).setUp()
294
boundary = self.boundary
297
self.first_range_start = 25
298
file_size = 200 # big enough to encompass all ranges
299
for (start, part) in [(self.first_range_start, self.alpha),
300
# Two contiguous ranges
302
(126, self.alpha.upper())]:
303
content += self._multipart_byterange(part, start, boundary,
306
content += self._boundary_line()
308
self._file = response.RangeFile('Multiple_ranges_file',
310
self.set_file_boundary()
312
def _boundary_line(self):
313
"""Helper to build the formatted boundary line."""
314
return b'--' + self.boundary + b'\r\n'
316
def set_file_boundary(self):
317
# Ranges are set by decoding the range headers, the RangeFile user is
318
# supposed to call the following before using seek or read since it
319
# requires knowing the *response* headers (in that case the boundary
320
# which is part of the Content-Type header).
321
self._file.set_boundary(self.boundary)
323
def _multipart_byterange(self, data, offset, boundary, file_size=b'*'):
324
"""Encode a part of a file as a multipart/byterange MIME type.
326
When a range request is issued, the HTTP response body can be
327
decomposed in parts, each one representing a range (start, size) in a
330
:param data: The payload.
331
:param offset: where data starts in the file
332
:param boundary: used to separate the parts
333
:param file_size: the size of the file containing the range (default to
336
:return: a string containing the data encoded as it will appear in the
339
bline = self._boundary_line()
340
# Each range begins with a boundary line
342
# A range is described by a set of headers, but only 'Content-Range' is
343
# required for our implementation (TestHandleResponse below will
344
# exercise ranges with multiple or missing headers')
345
if isinstance(file_size, int):
346
file_size = b'%d' % file_size
347
range += b'Content-Range: bytes %d-%d/%s\r\n' % (offset,
351
# Finally the raw bytes
355
def test_read_all_ranges(self):
357
self.assertEqual(self.alpha, f.read()) # Read first range
358
f.seek(100) # Trigger the second range recognition
359
self.assertEqual(self.alpha, f.read()) # Read second range
360
self.assertEqual(126, f.tell())
361
f.seek(126) # Start of third range which is also the current pos !
362
self.assertEqual(b'A', f.read(1))
364
self.assertEqual(b'LMN', f.read(3))
366
def test_seek_from_end(self):
367
"""See TestRangeFileMixin.test_seek_from_end."""
368
# The actual implementation will seek from end for the first range only
369
# and then fail. Since seeking from end is intended to be used for a
370
# single range only anyway, this test just document the actual
374
self.assertEqual(b'yz', f.read())
375
self.assertRaises(errors.InvalidRange, f.seek, -2, 2)
377
def test_seek_into_void(self):
379
start = self.first_range_start
381
# Seeking to a point between two ranges is possible (only once) but
382
# reading there is forbidden
384
# We crossed a range boundary, so now the file is positioned at the
385
# start of the new range (i.e. trying to seek below 100 will error out)
389
def test_seek_across_ranges(self):
391
f.seek(126) # skip the two first ranges
392
self.assertEqual(b'AB', f.read(2))
394
def test_checked_read_dont_overflow_buffers(self):
396
# We force a very low value to exercise all code paths in _checked_read
397
f._discarded_buf_size = 8
398
f.seek(126) # skip the two first ranges
399
self.assertEqual(b'AB', f.read(2))
401
def test_seek_twice_between_ranges(self):
403
start = self.first_range_start
404
f.seek(start + 40) # Past the first range but before the second
405
# Now the file is positioned at the second range start (100)
406
self.assertRaises(errors.InvalidRange, f.seek, start + 41)
408
def test_seek_at_range_end(self):
409
"""Test seek behavior at range end."""
415
def test_read_at_range_end(self):
417
self.assertEqual(self.alpha, f.read())
418
self.assertEqual(self.alpha, f.read())
419
self.assertEqual(self.alpha.upper(), f.read())
420
self.assertRaises(errors.InvalidHttpResponse, f.read, 1)
423
class TestRangeFileMultipleRangesQuotedBoundaries(TestRangeFileMultipleRanges):
424
"""Perform the same tests as TestRangeFileMultipleRanges, but uses
425
an angle-bracket quoted boundary string like IIS 6.0 and 7.0
426
(but not IIS 5, which breaks the RFC in a different way
427
by using square brackets, not angle brackets)
429
This reveals a bug caused by
431
- The bad implementation of RFC 822 unquoting in Python (angles are not
432
quotes), coupled with
434
- The bad implementation of RFC 2046 in IIS (angles are not permitted chars
438
# The boundary as it appears in boundary lines
439
# IIS 6 and 7 use this value
440
_boundary_trimmed = b"q1w2e3r4t5y6u7i8o9p0zaxscdvfbgnhmjklkl"
441
boundary = b'<' + _boundary_trimmed + b'>'
443
def set_file_boundary(self):
444
# Emulate broken rfc822.unquote() here by removing angles
445
self._file.set_boundary(self._boundary_trimmed)
448
class TestRangeFileVarious(tests.TestCase):
449
"""Tests RangeFile aspects not covered elsewhere."""
451
def test_seek_whence(self):
452
"""Test the seek whence parameter values."""
453
f = response.RangeFile('foo', BytesIO(b'abc'))
458
self.assertRaises(ValueError, f.seek, 0, 14)
460
def test_range_syntax(self):
461
"""Test the Content-Range scanning."""
463
f = response.RangeFile('foo', BytesIO())
465
def ok(expected, header_value):
466
f.set_range_from_header(header_value)
467
# Slightly peek under the covers to get the size
468
self.assertEqual(expected, (f.tell(), f._size))
470
ok((1, 10), 'bytes 1-10/11')
471
ok((1, 10), 'bytes 1-10/*')
472
ok((12, 2), '\tbytes 12-13/*')
473
ok((28, 1), ' bytes 28-28/*')
474
ok((2123, 2120), 'bytes 2123-4242/12310')
475
ok((1, 10), 'bytes 1-10/ttt') # We don't check total (ttt)
477
def nok(header_value):
478
self.assertRaises(errors.InvalidHttpRange,
479
f.set_range_from_header, header_value)
483
nok('bytes xx-yyy/zzz')
484
nok('bytes xx-12/zzz')
485
nok('bytes 11-yy/zzz')
489
# Taken from real request responses
490
_full_text_response = (200, b"""HTTP/1.1 200 OK\r
491
Date: Tue, 11 Jul 2006 04:32:56 GMT\r
492
Server: Apache/2.0.54 (Fedora)\r
493
Last-Modified: Sun, 23 Apr 2006 19:35:20 GMT\r
494
ETag: "56691-23-38e9ae00"\r
495
Accept-Ranges: bytes\r
498
Content-Type: text/plain; charset=UTF-8\r
500
""", b"""Bazaar-NG meta directory, format 1
504
_single_range_response = (206, b"""HTTP/1.1 206 Partial Content\r
505
Date: Tue, 11 Jul 2006 04:45:22 GMT\r
506
Server: Apache/2.0.54 (Fedora)\r
507
Last-Modified: Thu, 06 Jul 2006 20:22:05 GMT\r
508
ETag: "238a3c-16ec2-805c5540"\r
509
Accept-Ranges: bytes\r
510
Content-Length: 100\r
511
Content-Range: bytes 100-199/93890\r
513
Content-Type: text/plain; charset=UTF-8\r
515
""", b"""mbp@sourcefrog.net-20050309040815-13242001617e4a06
516
mbp@sourcefrog.net-20050309040929-eee0eb3e6d1e762""")
519
_single_range_no_content_type = (206, b"""HTTP/1.1 206 Partial Content\r
520
Date: Tue, 11 Jul 2006 04:45:22 GMT\r
521
Server: Apache/2.0.54 (Fedora)\r
522
Last-Modified: Thu, 06 Jul 2006 20:22:05 GMT\r
523
ETag: "238a3c-16ec2-805c5540"\r
524
Accept-Ranges: bytes\r
525
Content-Length: 100\r
526
Content-Range: bytes 100-199/93890\r
529
""", b"""mbp@sourcefrog.net-20050309040815-13242001617e4a06
530
mbp@sourcefrog.net-20050309040929-eee0eb3e6d1e762""")
533
_multipart_range_response = (206, b"""HTTP/1.1 206 Partial Content\r
534
Date: Tue, 11 Jul 2006 04:49:48 GMT\r
535
Server: Apache/2.0.54 (Fedora)\r
536
Last-Modified: Thu, 06 Jul 2006 20:22:05 GMT\r
537
ETag: "238a3c-16ec2-805c5540"\r
538
Accept-Ranges: bytes\r
539
Content-Length: 1534\r
541
Content-Type: multipart/byteranges; boundary=418470f848b63279b\r
543
\r""", b"""--418470f848b63279b\r
544
Content-type: text/plain; charset=UTF-8\r
545
Content-range: bytes 0-254/93890\r
547
mbp@sourcefrog.net-20050309040815-13242001617e4a06
548
mbp@sourcefrog.net-20050309040929-eee0eb3e6d1e7627
549
mbp@sourcefrog.net-20050309040957-6cad07f466bb0bb8
550
mbp@sourcefrog.net-20050309041501-c840e09071de3b67
551
mbp@sourcefrog.net-20050309044615-c24a3250be83220a
553
--418470f848b63279b\r
554
Content-type: text/plain; charset=UTF-8\r
555
Content-range: bytes 1000-2049/93890\r
558
mbp@sourcefrog.net-20050311063625-07858525021f270b
559
mbp@sourcefrog.net-20050311231934-aa3776aff5200bb9
560
mbp@sourcefrog.net-20050311231953-73aeb3a131c3699a
561
mbp@sourcefrog.net-20050311232353-f5e33da490872c6a
562
mbp@sourcefrog.net-20050312071639-0a8f59a34a024ff0
563
mbp@sourcefrog.net-20050312073432-b2c16a55e0d6e9fb
564
mbp@sourcefrog.net-20050312073831-a47c3335ece1920f
565
mbp@sourcefrog.net-20050312085412-13373aa129ccbad3
566
mbp@sourcefrog.net-20050313052251-2bf004cb96b39933
567
mbp@sourcefrog.net-20050313052856-3edd84094687cb11
568
mbp@sourcefrog.net-20050313053233-e30a4f28aef48f9d
569
mbp@sourcefrog.net-20050313053853-7c64085594ff3072
570
mbp@sourcefrog.net-20050313054757-a86c3f5871069e22
571
mbp@sourcefrog.net-20050313061422-418f1f73b94879b9
572
mbp@sourcefrog.net-20050313120651-497bd231b19df600
573
mbp@sourcefrog.net-20050314024931-eae0170ef25a5d1a
574
mbp@sourcefrog.net-20050314025438-d52099f915fe65fc
575
mbp@sourcefrog.net-20050314025539-637a636692c055cf
576
mbp@sourcefrog.net-20050314025737-55eb441f430ab4ba
577
mbp@sourcefrog.net-20050314025901-d74aa93bb7ee8f62
579
--418470f848b63279b--\r
583
_multipart_squid_range_response = (206, b"""HTTP/1.0 206 Partial Content\r
584
Date: Thu, 31 Aug 2006 21:16:22 GMT\r
585
Server: Apache/2.2.2 (Unix) DAV/2\r
586
Last-Modified: Thu, 31 Aug 2006 17:57:06 GMT\r
587
Accept-Ranges: bytes\r
588
Content-Type: multipart/byteranges; boundary="squid/2.5.STABLE12:C99323425AD4FE26F726261FA6C24196"\r
589
Content-Length: 598\r
590
X-Cache: MISS from localhost.localdomain\r
591
X-Cache-Lookup: HIT from localhost.localdomain:3128\r
592
Proxy-Connection: keep-alive\r
596
--squid/2.5.STABLE12:C99323425AD4FE26F726261FA6C24196\r
597
Content-Type: text/plain\r
598
Content-Range: bytes 0-99/18672\r
602
scott@netsplit.com-20050708230047-47c7868f276b939f fulltext 0 863 :
604
--squid/2.5.STABLE12:C99323425AD4FE26F726261FA6C24196\r
605
Content-Type: text/plain\r
606
Content-Range: bytes 300-499/18672\r
608
com-20050708231537-2b124b835395399a :
609
scott@netsplit.com-20050820234126-551311dbb7435b51 line-delta 1803 479 .scott@netsplit.com-20050820232911-dc4322a084eadf7e :
610
scott@netsplit.com-20050821213706-c86\r
611
--squid/2.5.STABLE12:C99323425AD4FE26F726261FA6C24196--\r
616
_full_text_response_no_content_type = (200, b"""HTTP/1.1 200 OK\r
617
Date: Tue, 11 Jul 2006 04:32:56 GMT\r
618
Server: Apache/2.0.54 (Fedora)\r
619
Last-Modified: Sun, 23 Apr 2006 19:35:20 GMT\r
620
ETag: "56691-23-38e9ae00"\r
621
Accept-Ranges: bytes\r
625
""", b"""Bazaar-NG meta directory, format 1
629
_full_text_response_no_content_length = (200, b"""HTTP/1.1 200 OK\r
630
Date: Tue, 11 Jul 2006 04:32:56 GMT\r
631
Server: Apache/2.0.54 (Fedora)\r
632
Last-Modified: Sun, 23 Apr 2006 19:35:20 GMT\r
633
ETag: "56691-23-38e9ae00"\r
634
Accept-Ranges: bytes\r
636
Content-Type: text/plain; charset=UTF-8\r
638
""", b"""Bazaar-NG meta directory, format 1
642
_single_range_no_content_range = (206, b"""HTTP/1.1 206 Partial Content\r
643
Date: Tue, 11 Jul 2006 04:45:22 GMT\r
644
Server: Apache/2.0.54 (Fedora)\r
645
Last-Modified: Thu, 06 Jul 2006 20:22:05 GMT\r
646
ETag: "238a3c-16ec2-805c5540"\r
647
Accept-Ranges: bytes\r
648
Content-Length: 100\r
651
""", b"""mbp@sourcefrog.net-20050309040815-13242001617e4a06
652
mbp@sourcefrog.net-20050309040929-eee0eb3e6d1e762""")
655
_single_range_response_truncated = (206, b"""HTTP/1.1 206 Partial Content\r
656
Date: Tue, 11 Jul 2006 04:45:22 GMT\r
657
Server: Apache/2.0.54 (Fedora)\r
658
Last-Modified: Thu, 06 Jul 2006 20:22:05 GMT\r
659
ETag: "238a3c-16ec2-805c5540"\r
660
Accept-Ranges: bytes\r
661
Content-Length: 100\r
662
Content-Range: bytes 100-199/93890\r
664
Content-Type: text/plain; charset=UTF-8\r
666
""", b"""mbp@sourcefrog.net-20050309040815-13242001617e4a06""")
669
_invalid_response = (444, b"""HTTP/1.1 444 Bad Response\r
670
Date: Tue, 11 Jul 2006 04:32:56 GMT\r
672
Content-Type: text/html; charset=iso-8859-1\r
674
""", b"""<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
676
<title>404 Not Found</title>
679
<p>I don't know what I'm doing</p>
685
_multipart_no_content_range = (206, b"""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
699
_multipart_no_boundary = (206, b"""HTTP/1.0 206 Partial Content\r
700
Content-Type: multipart/byteranges; boundary=THIS_SEPARATES\r
701
Content-Length: 598\r
706
Content-Type: text/plain\r
707
Content-Range: bytes 0-18/18672\r
711
The range ended at the line above, this text is garbage instead of a boundary
716
class TestHandleResponse(tests.TestCase):
718
def _build_HTTPMessage(self, raw_headers):
719
status_and_headers = BytesIO(raw_headers)
720
# Get rid of the status line
721
status_and_headers.readline()
722
msg = http_client.HTTPMessage(status_and_headers)
725
def get_response(self, a_response):
726
"""Process a supplied response, and return the result."""
727
code, raw_headers, body = a_response
728
msg = self._build_HTTPMessage(raw_headers)
729
return response.handle_response('http://foo', code, msg,
730
BytesIO(a_response[2]))
732
def test_full_text(self):
733
out = self.get_response(_full_text_response)
734
# It is a BytesIO from the original data
735
self.assertEqual(_full_text_response[2], out.read())
737
def test_single_range(self):
738
out = self.get_response(_single_range_response)
741
self.assertEqual(_single_range_response[2], out.read(100))
743
def test_single_range_no_content(self):
744
out = self.get_response(_single_range_no_content_type)
747
self.assertEqual(_single_range_no_content_type[2], out.read(100))
749
def test_single_range_truncated(self):
750
out = self.get_response(_single_range_response_truncated)
751
# Content-Range declares 100 but only 51 present
752
self.assertRaises(errors.ShortReadvError, out.seek, out.tell() + 51)
754
def test_multi_range(self):
755
out = self.get_response(_multipart_range_response)
757
# Just make sure we can read the right contents
764
def test_multi_squid_range(self):
765
out = self.get_response(_multipart_squid_range_response)
767
# Just make sure we can read the right contents
774
def test_invalid_response(self):
775
self.assertRaises(errors.InvalidHttpResponse,
776
self.get_response, _invalid_response)
778
def test_full_text_no_content_type(self):
779
# We should not require Content-Type for a full response
780
code, raw_headers, body = _full_text_response_no_content_type
781
msg = self._build_HTTPMessage(raw_headers)
782
out = response.handle_response('http://foo', code, msg, BytesIO(body))
783
self.assertEqual(body, out.read())
785
def test_full_text_no_content_length(self):
786
code, raw_headers, body = _full_text_response_no_content_length
787
msg = self._build_HTTPMessage(raw_headers)
788
out = response.handle_response('http://foo', code, msg, BytesIO(body))
789
self.assertEqual(body, out.read())
791
def test_missing_content_range(self):
792
code, raw_headers, body = _single_range_no_content_range
793
msg = self._build_HTTPMessage(raw_headers)
794
self.assertRaises(errors.InvalidHttpResponse,
795
response.handle_response,
796
'http://bogus', code, msg, BytesIO(body))
798
def test_multipart_no_content_range(self):
799
code, raw_headers, body = _multipart_no_content_range
800
msg = self._build_HTTPMessage(raw_headers)
801
self.assertRaises(errors.InvalidHttpResponse,
802
response.handle_response,
803
'http://bogus', code, msg, BytesIO(body))
805
def test_multipart_no_boundary(self):
806
out = self.get_response(_multipart_no_boundary)
807
out.read() # Read the whole range
808
# Fail to find the boundary line
809
self.assertRaises(errors.InvalidHttpResponse, out.seek, 1, 1)
812
class TestRangeFileSizeReadLimited(tests.TestCase):
813
"""Test RangeFile _max_read_size functionality which limits the size of
814
read blocks to prevent MemoryError messages in socket.recv.
818
super(TestRangeFileSizeReadLimited, self).setUp()
819
# create a test datablock larger than _max_read_size.
820
chunk_size = response.RangeFile._max_read_size
821
test_pattern = b'0123456789ABCDEF'
822
self.test_data = test_pattern * (3 * chunk_size // len(test_pattern))
823
self.test_data_len = len(self.test_data)
825
def test_max_read_size(self):
826
"""Read data in blocks and verify that the reads are not larger than
827
the maximum read size.
829
# retrieve data in large blocks from response.RangeFile object
830
mock_read_file = FakeReadFile(self.test_data)
831
range_file = response.RangeFile('test_max_read_size', mock_read_file)
832
response_data = range_file.read(self.test_data_len)
834
# verify read size was equal to the maximum read size
835
self.assertTrue(mock_read_file.get_max_read_size() > 0)
836
self.assertEqual(mock_read_file.get_max_read_size(),
837
response.RangeFile._max_read_size)
838
self.assertEqual(mock_read_file.get_read_count(), 3)
840
# report error if the data wasn't equal (we only report the size due
841
# to the length of the data)
842
if response_data != self.test_data:
843
message = "Data not equal. Expected %d bytes, received %d."
844
self.fail(message % (len(response_data), self.test_data_len))