forked from chromium/chromium
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbisect-builds.py
executable file
·1277 lines (1138 loc) · 49 KB
/
bisect-builds.py
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
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#!/usr/bin/env python
# Copyright (c) 2012 The Chromium Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Snapshot Build Bisect Tool
This script bisects a snapshot archive using binary search. It starts at
a bad revision (it will try to guess HEAD) and asks for a last known-good
revision. It will then binary search across this revision range by downloading,
unzipping, and opening Chromium for you. After testing the specific revision,
it will ask you whether it is good or bad before continuing the search.
"""
# The base URL for stored build archives.
CHROMIUM_BASE_URL = ('http://commondatastorage.googleapis.com'
'/chromium-browser-snapshots')
WEBKIT_BASE_URL = ('http://commondatastorage.googleapis.com'
'/chromium-webkit-snapshots')
ASAN_BASE_URL = ('http://commondatastorage.googleapis.com'
'/chromium-browser-asan')
# GS bucket name.
GS_BUCKET_NAME = 'chrome-unsigned/desktop-W15K3Y'
# Base URL for downloading official builds.
GOOGLE_APIS_URL = 'commondatastorage.googleapis.com'
# The base URL for official builds.
OFFICIAL_BASE_URL = 'http://%s/%s' % (GOOGLE_APIS_URL, GS_BUCKET_NAME)
# URL template for viewing changelogs between revisions.
CHANGELOG_URL = ('https://chromium.googlesource.com/chromium/src/+log/%s..%s')
# URL to convert SVN revision to git hash.
CRREV_URL = ('https://cr-rev.appspot.com/_ah/api/crrev/v1/redirect/')
# URL template for viewing changelogs between official versions.
OFFICIAL_CHANGELOG_URL = ('https://chromium.googlesource.com/chromium/'
'src/+log/%s..%s?pretty=full')
# DEPS file URL.
DEPS_FILE_OLD = ('http://src.chromium.org/viewvc/chrome/trunk/src/'
'DEPS?revision=%d')
DEPS_FILE_NEW = ('https://chromium.googlesource.com/chromium/src/+/%s/DEPS')
# Blink changelogs URL.
BLINK_CHANGELOG_URL = ('http://build.chromium.org'
'/f/chromium/perf/dashboard/ui/changelog_blink.html'
'?url=/trunk&range=%d%%3A%d')
DONE_MESSAGE_GOOD_MIN = ('You are probably looking for a change made after %s ('
'known good), but no later than %s (first known bad).')
DONE_MESSAGE_GOOD_MAX = ('You are probably looking for a change made after %s ('
'known bad), but no later than %s (first known good).')
CHROMIUM_GITHASH_TO_SVN_URL = (
'https://chromium.googlesource.com/chromium/src/+/%s?format=json')
BLINK_GITHASH_TO_SVN_URL = (
'https://chromium.googlesource.com/chromium/blink/+/%s?format=json')
GITHASH_TO_SVN_URL = {
'chromium': CHROMIUM_GITHASH_TO_SVN_URL,
'blink': BLINK_GITHASH_TO_SVN_URL,
}
# Search pattern to be matched in the JSON output from
# CHROMIUM_GITHASH_TO_SVN_URL to get the chromium revision (svn revision).
CHROMIUM_SEARCH_PATTERN_OLD = (
r'.*git-svn-id: svn://svn.chromium.org/chrome/trunk/src@(\d+) ')
CHROMIUM_SEARCH_PATTERN = (
r'Cr-Commit-Position: refs/heads/master@{#(\d+)}')
# Search pattern to be matched in the json output from
# BLINK_GITHASH_TO_SVN_URL to get the blink revision (svn revision).
BLINK_SEARCH_PATTERN = (
r'.*git-svn-id: svn://svn.chromium.org/blink/trunk@(\d+) ')
SEARCH_PATTERN = {
'chromium': CHROMIUM_SEARCH_PATTERN,
'blink': BLINK_SEARCH_PATTERN,
}
CREDENTIAL_ERROR_MESSAGE = ('You are attempting to access protected data with '
'no configured credentials')
###############################################################################
import httplib
import json
import optparse
import os
import re
import shlex
import shutil
import subprocess
import sys
import tempfile
import threading
import urllib
from distutils.version import LooseVersion
from xml.etree import ElementTree
import zipfile
class PathContext(object):
"""A PathContext is used to carry the information used to construct URLs and
paths when dealing with the storage server and archives."""
def __init__(self, base_url, platform, good_revision, bad_revision,
is_official, is_asan, use_local_cache, flash_path = None):
super(PathContext, self).__init__()
# Store off the input parameters.
self.base_url = base_url
self.platform = platform # What's passed in to the '-a/--archive' option.
self.good_revision = good_revision
self.bad_revision = bad_revision
self.is_official = is_official
self.is_asan = is_asan
self.build_type = 'release'
self.flash_path = flash_path
# Dictionary which stores svn revision number as key and it's
# corresponding git hash as value. This data is populated in
# _FetchAndParse and used later in GetDownloadURL while downloading
# the build.
self.githash_svn_dict = {}
# The name of the ZIP file in a revision directory on the server.
self.archive_name = None
# Whether to cache and use the list of known revisions in a local file to
# speed up the initialization of the script at the next run.
self.use_local_cache = use_local_cache
# Locate the local checkout to speed up the script by using locally stored
# metadata.
abs_file_path = os.path.abspath(os.path.realpath(__file__))
local_src_path = os.path.join(os.path.dirname(abs_file_path), '..')
if abs_file_path.endswith(os.path.join('tools', 'bisect-builds.py')) and\
os.path.exists(os.path.join(local_src_path, '.git')):
self.local_src_path = os.path.normpath(local_src_path)
else:
self.local_src_path = None
# Set some internal members:
# _listing_platform_dir = Directory that holds revisions. Ends with a '/'.
# _archive_extract_dir = Uncompressed directory in the archive_name file.
# _binary_name = The name of the executable to run.
if self.platform in ('linux', 'linux64', 'linux-arm', 'chromeos'):
self._binary_name = 'chrome'
elif self.platform in ('mac', 'mac64'):
self.archive_name = 'chrome-mac.zip'
self._archive_extract_dir = 'chrome-mac'
elif self.platform in ('win', 'win64'):
self.archive_name = 'chrome-win32.zip'
self._archive_extract_dir = 'chrome-win32'
self._binary_name = 'chrome.exe'
else:
raise Exception('Invalid platform: %s' % self.platform)
if is_official:
if self.platform == 'linux':
self._listing_platform_dir = 'precise32/'
self.archive_name = 'chrome-precise32.zip'
self._archive_extract_dir = 'chrome-precise32'
elif self.platform == 'linux64':
self._listing_platform_dir = 'precise64/'
self.archive_name = 'chrome-precise64.zip'
self._archive_extract_dir = 'chrome-precise64'
elif self.platform == 'mac':
self._listing_platform_dir = 'mac/'
self._binary_name = 'Google Chrome.app/Contents/MacOS/Google Chrome'
elif self.platform == 'mac64':
self._listing_platform_dir = 'mac64/'
self._binary_name = 'Google Chrome.app/Contents/MacOS/Google Chrome'
elif self.platform == 'win':
self._listing_platform_dir = 'win/'
self.archive_name = 'chrome-win.zip'
self._archive_extract_dir = 'chrome-win'
elif self.platform == 'win64':
self._listing_platform_dir = 'win64/'
self.archive_name = 'chrome-win64.zip'
self._archive_extract_dir = 'chrome-win64'
else:
if self.platform in ('linux', 'linux64', 'linux-arm', 'chromeos'):
self.archive_name = 'chrome-linux.zip'
self._archive_extract_dir = 'chrome-linux'
if self.platform == 'linux':
self._listing_platform_dir = 'Linux/'
elif self.platform == 'linux64':
self._listing_platform_dir = 'Linux_x64/'
elif self.platform == 'linux-arm':
self._listing_platform_dir = 'Linux_ARM_Cross-Compile/'
elif self.platform == 'chromeos':
self._listing_platform_dir = 'Linux_ChromiumOS_Full/'
# There is no 64-bit distinction for non-official mac builds.
elif self.platform in ('mac', 'mac64'):
self._listing_platform_dir = 'Mac/'
self._binary_name = 'Chromium.app/Contents/MacOS/Chromium'
elif self.platform == 'win':
self._listing_platform_dir = 'Win/'
def GetASANPlatformDir(self):
"""ASAN builds are in directories like "linux-release", or have filenames
like "asan-win32-release-277079.zip". This aligns to our platform names
except in the case of Windows where they use "win32" instead of "win"."""
if self.platform == 'win':
return 'win32'
else:
return self.platform
def GetListingURL(self, marker=None):
"""Returns the URL for a directory listing, with an optional marker."""
marker_param = ''
if marker:
marker_param = '&marker=' + str(marker)
if self.is_asan:
prefix = '%s-%s' % (self.GetASANPlatformDir(), self.build_type)
return self.base_url + '/?delimiter=&prefix=' + prefix + marker_param
else:
return (self.base_url + '/?delimiter=/&prefix=' +
self._listing_platform_dir + marker_param)
def GetDownloadURL(self, revision):
"""Gets the download URL for a build archive of a specific revision."""
if self.is_asan:
return '%s/%s-%s/%s-%d.zip' % (
ASAN_BASE_URL, self.GetASANPlatformDir(), self.build_type,
self.GetASANBaseName(), revision)
if self.is_official:
return '%s/%s/%s%s' % (
OFFICIAL_BASE_URL, revision, self._listing_platform_dir,
self.archive_name)
else:
if str(revision) in self.githash_svn_dict:
revision = self.githash_svn_dict[str(revision)]
return '%s/%s%s/%s' % (self.base_url, self._listing_platform_dir,
revision, self.archive_name)
def GetLastChangeURL(self):
"""Returns a URL to the LAST_CHANGE file."""
return self.base_url + '/' + self._listing_platform_dir + 'LAST_CHANGE'
def GetASANBaseName(self):
"""Returns the base name of the ASAN zip file."""
if 'linux' in self.platform:
return 'asan-symbolized-%s-%s' % (self.GetASANPlatformDir(),
self.build_type)
else:
return 'asan-%s-%s' % (self.GetASANPlatformDir(), self.build_type)
def GetLaunchPath(self, revision):
"""Returns a relative path (presumably from the archive extraction location)
that is used to run the executable."""
if self.is_asan:
extract_dir = '%s-%d' % (self.GetASANBaseName(), revision)
else:
extract_dir = self._archive_extract_dir
return os.path.join(extract_dir, self._binary_name)
def ParseDirectoryIndex(self, last_known_rev):
"""Parses the Google Storage directory listing into a list of revision
numbers."""
def _GetMarkerForRev(revision):
if self.is_asan:
return '%s-%s/%s-%d.zip' % (
self.GetASANPlatformDir(), self.build_type,
self.GetASANBaseName(), revision)
return '%s%d' % (self._listing_platform_dir, revision)
def _FetchAndParse(url):
"""Fetches a URL and returns a 2-Tuple of ([revisions], next-marker). If
next-marker is not None, then the listing is a partial listing and another
fetch should be performed with next-marker being the marker= GET
parameter."""
handle = urllib.urlopen(url)
document = ElementTree.parse(handle)
# All nodes in the tree are namespaced. Get the root's tag name to extract
# the namespace. Etree does namespaces as |{namespace}tag|.
root_tag = document.getroot().tag
end_ns_pos = root_tag.find('}')
if end_ns_pos == -1:
raise Exception('Could not locate end namespace for directory index')
namespace = root_tag[:end_ns_pos + 1]
# Find the prefix (_listing_platform_dir) and whether or not the list is
# truncated.
prefix_len = len(document.find(namespace + 'Prefix').text)
next_marker = None
is_truncated = document.find(namespace + 'IsTruncated')
if is_truncated is not None and is_truncated.text.lower() == 'true':
next_marker = document.find(namespace + 'NextMarker').text
# Get a list of all the revisions.
revisions = []
githash_svn_dict = {}
if self.is_asan:
asan_regex = re.compile(r'.*%s-(\d+)\.zip$' % (self.GetASANBaseName()))
# Non ASAN builds are in a <revision> directory. The ASAN builds are
# flat
all_prefixes = document.findall(namespace + 'Contents/' +
namespace + 'Key')
for prefix in all_prefixes:
m = asan_regex.match(prefix.text)
if m:
try:
revisions.append(int(m.group(1)))
except ValueError:
pass
else:
all_prefixes = document.findall(namespace + 'CommonPrefixes/' +
namespace + 'Prefix')
# The <Prefix> nodes have content of the form of
# |_listing_platform_dir/revision/|. Strip off the platform dir and the
# trailing slash to just have a number.
for prefix in all_prefixes:
revnum = prefix.text[prefix_len:-1]
try:
if not revnum.isdigit():
# During the svn-git migration, some items were stored by hash.
# These items may appear anywhere in the list of items.
# If |last_known_rev| is set, assume that the full list has been
# retrieved before (including the hashes), so we can safely skip
# all git hashes and focus on the numeric revision numbers.
if last_known_rev:
revnum = None
else:
git_hash = revnum
revnum = self.GetSVNRevisionFromGitHash(git_hash)
githash_svn_dict[revnum] = git_hash
if revnum is not None:
revnum = int(revnum)
revisions.append(revnum)
except ValueError:
pass
return (revisions, next_marker, githash_svn_dict)
# Fetch the first list of revisions.
if last_known_rev:
revisions = []
# Optimization: Start paging at the last known revision (local cache).
next_marker = _GetMarkerForRev(last_known_rev)
# Optimization: Stop paging at the last known revision (remote).
last_change_rev = GetChromiumRevision(self, self.GetLastChangeURL())
if last_known_rev == last_change_rev:
return []
else:
(revisions, next_marker, new_dict) = _FetchAndParse(self.GetListingURL())
self.githash_svn_dict.update(new_dict)
last_change_rev = None
# If the result list was truncated, refetch with the next marker. Do this
# until an entire directory listing is done.
while next_marker:
sys.stdout.write('\rFetching revisions at marker %s' % next_marker)
sys.stdout.flush()
next_url = self.GetListingURL(next_marker)
(new_revisions, next_marker, new_dict) = _FetchAndParse(next_url)
revisions.extend(new_revisions)
self.githash_svn_dict.update(new_dict)
if last_change_rev and last_change_rev in new_revisions:
break
sys.stdout.write('\r')
sys.stdout.flush()
return revisions
def _GetSVNRevisionFromGitHashWithoutGitCheckout(self, git_sha1, depot):
json_url = GITHASH_TO_SVN_URL[depot] % git_sha1
response = urllib.urlopen(json_url)
if response.getcode() == 200:
try:
data = json.loads(response.read()[4:])
except ValueError:
print 'ValueError for JSON URL: %s' % json_url
raise ValueError
else:
raise ValueError
if 'message' in data:
message = data['message'].split('\n')
message = [line for line in message if line.strip()]
search_pattern = re.compile(SEARCH_PATTERN[depot])
result = search_pattern.search(message[len(message)-1])
if result:
return result.group(1)
else:
if depot == 'chromium':
result = re.search(CHROMIUM_SEARCH_PATTERN_OLD,
message[len(message)-1])
if result:
return result.group(1)
print 'Failed to get svn revision number for %s' % git_sha1
raise ValueError
def _GetSVNRevisionFromGitHashFromGitCheckout(self, git_sha1, depot):
def _RunGit(command, path):
command = ['git'] + command
shell = sys.platform.startswith('win')
proc = subprocess.Popen(command, shell=shell, stdout=subprocess.PIPE,
stderr=subprocess.PIPE, cwd=path)
(output, _) = proc.communicate()
return (output, proc.returncode)
path = self.local_src_path
if depot == 'blink':
path = os.path.join(self.local_src_path, 'third_party', 'WebKit')
revision = None
try:
command = ['svn', 'find-rev', git_sha1]
(git_output, return_code) = _RunGit(command, path)
if not return_code:
revision = git_output.strip('\n')
except ValueError:
pass
if not revision:
command = ['log', '-n1', '--format=%s', git_sha1]
(git_output, return_code) = _RunGit(command, path)
if not return_code:
revision = re.match('SVN changes up to revision ([0-9]+)', git_output)
revision = revision.group(1) if revision else None
if revision:
return revision
raise ValueError
def GetSVNRevisionFromGitHash(self, git_sha1, depot='chromium'):
if not self.local_src_path:
return self._GetSVNRevisionFromGitHashWithoutGitCheckout(git_sha1, depot)
else:
return self._GetSVNRevisionFromGitHashFromGitCheckout(git_sha1, depot)
def GetRevList(self):
"""Gets the list of revision numbers between self.good_revision and
self.bad_revision."""
cache = {}
# The cache is stored in the same directory as bisect-builds.py
cache_filename = os.path.join(
os.path.abspath(os.path.dirname(__file__)),
'.bisect-builds-cache.json')
cache_dict_key = self.GetListingURL()
def _LoadBucketFromCache():
if self.use_local_cache:
try:
with open(cache_filename) as cache_file:
for (key, value) in json.load(cache_file).items():
cache[key] = value
revisions = cache.get(cache_dict_key, [])
githash_svn_dict = cache.get('githash_svn_dict', {})
if revisions:
print 'Loaded revisions %d-%d from %s' % (revisions[0],
revisions[-1], cache_filename)
return (revisions, githash_svn_dict)
except (EnvironmentError, ValueError):
pass
return ([], {})
def _SaveBucketToCache():
"""Save the list of revisions and the git-svn mappings to a file.
The list of revisions is assumed to be sorted."""
if self.use_local_cache:
cache[cache_dict_key] = revlist_all
cache['githash_svn_dict'] = self.githash_svn_dict
try:
with open(cache_filename, 'w') as cache_file:
json.dump(cache, cache_file)
print 'Saved revisions %d-%d to %s' % (
revlist_all[0], revlist_all[-1], cache_filename)
except EnvironmentError:
pass
# Download the revlist and filter for just the range between good and bad.
minrev = min(self.good_revision, self.bad_revision)
maxrev = max(self.good_revision, self.bad_revision)
(revlist_all, self.githash_svn_dict) = _LoadBucketFromCache()
last_known_rev = revlist_all[-1] if revlist_all else 0
if last_known_rev < maxrev:
revlist_all.extend(map(int, self.ParseDirectoryIndex(last_known_rev)))
revlist_all = list(set(revlist_all))
revlist_all.sort()
_SaveBucketToCache()
revlist = [x for x in revlist_all if x >= int(minrev) and x <= int(maxrev)]
# Set good and bad revisions to be legit revisions.
if revlist:
if self.good_revision < self.bad_revision:
self.good_revision = revlist[0]
self.bad_revision = revlist[-1]
else:
self.bad_revision = revlist[0]
self.good_revision = revlist[-1]
# Fix chromium rev so that the deps blink revision matches REVISIONS file.
if self.base_url == WEBKIT_BASE_URL:
revlist_all.sort()
self.good_revision = FixChromiumRevForBlink(revlist,
revlist_all,
self,
self.good_revision)
self.bad_revision = FixChromiumRevForBlink(revlist,
revlist_all,
self,
self.bad_revision)
return revlist
def GetOfficialBuildsList(self):
"""Gets the list of official build numbers between self.good_revision and
self.bad_revision."""
def CheckDepotToolsInPath():
delimiter = ';' if sys.platform.startswith('win') else ':'
path_list = os.environ['PATH'].split(delimiter)
for path in path_list:
if path.rstrip(os.path.sep).endswith('depot_tools'):
return path
return None
def RunGsutilCommand(args):
gsutil_path = CheckDepotToolsInPath()
if gsutil_path is None:
print ('Follow the instructions in this document '
'http://dev.chromium.org/developers/how-tos/install-depot-tools'
' to install depot_tools and then try again.')
sys.exit(1)
gsutil_path = os.path.join(gsutil_path, 'third_party', 'gsutil', 'gsutil')
gsutil = subprocess.Popen([sys.executable, gsutil_path] + args,
stdout=subprocess.PIPE, stderr=subprocess.PIPE,
env=None)
stdout, stderr = gsutil.communicate()
if gsutil.returncode:
if (re.findall(r'status[ |=]40[1|3]', stderr) or
stderr.startswith(CREDENTIAL_ERROR_MESSAGE)):
print ('Follow these steps to configure your credentials and try'
' running the bisect-builds.py again.:\n'
' 1. Run "python %s config" and follow its instructions.\n'
' 2. If you have a @google.com account, use that account.\n'
' 3. For the project-id, just enter 0.' % gsutil_path)
sys.exit(1)
else:
raise Exception('Error running the gsutil command: %s' % stderr)
return stdout
def GsutilList(bucket):
query = 'gs://%s/' % bucket
stdout = RunGsutilCommand(['ls', query])
return [url[len(query):].strip('/') for url in stdout.splitlines()]
# Download the revlist and filter for just the range between good and bad.
minrev = min(self.good_revision, self.bad_revision)
maxrev = max(self.good_revision, self.bad_revision)
build_numbers = GsutilList(GS_BUCKET_NAME)
revision_re = re.compile(r'(\d\d\.\d\.\d{4}\.\d+)')
build_numbers = filter(lambda b: revision_re.search(b), build_numbers)
final_list = []
parsed_build_numbers = [LooseVersion(x) for x in build_numbers]
connection = httplib.HTTPConnection(GOOGLE_APIS_URL)
for build_number in sorted(parsed_build_numbers):
if build_number > maxrev:
break
if build_number < minrev:
continue
path = ('/' + GS_BUCKET_NAME + '/' + str(build_number) + '/' +
self._listing_platform_dir + self.archive_name)
connection.request('HEAD', path)
response = connection.getresponse()
if response.status == 200:
final_list.append(str(build_number))
response.read()
connection.close()
return final_list
def UnzipFilenameToDir(filename, directory):
"""Unzip |filename| to |directory|."""
cwd = os.getcwd()
if not os.path.isabs(filename):
filename = os.path.join(cwd, filename)
zf = zipfile.ZipFile(filename)
# Make base.
if not os.path.isdir(directory):
os.mkdir(directory)
os.chdir(directory)
# Extract files.
for info in zf.infolist():
name = info.filename
if name.endswith('/'): # dir
if not os.path.isdir(name):
os.makedirs(name)
else: # file
directory = os.path.dirname(name)
if not os.path.isdir(directory):
os.makedirs(directory)
out = open(name, 'wb')
out.write(zf.read(name))
out.close()
# Set permissions. Permission info in external_attr is shifted 16 bits.
os.chmod(name, info.external_attr >> 16L)
os.chdir(cwd)
def FetchRevision(context, rev, filename, quit_event=None, progress_event=None):
"""Downloads and unzips revision |rev|.
@param context A PathContext instance.
@param rev The Chromium revision number/tag to download.
@param filename The destination for the downloaded file.
@param quit_event A threading.Event which will be set by the master thread to
indicate that the download should be aborted.
@param progress_event A threading.Event which will be set by the master thread
to indicate that the progress of the download should be
displayed.
"""
def ReportHook(blocknum, blocksize, totalsize):
if quit_event and quit_event.isSet():
raise RuntimeError('Aborting download of revision %s' % str(rev))
if progress_event and progress_event.isSet():
size = blocknum * blocksize
if totalsize == -1: # Total size not known.
progress = 'Received %d bytes' % size
else:
size = min(totalsize, size)
progress = 'Received %d of %d bytes, %.2f%%' % (
size, totalsize, 100.0 * size / totalsize)
# Send a \r to let all progress messages use just one line of output.
sys.stdout.write('\r' + progress)
sys.stdout.flush()
download_url = context.GetDownloadURL(rev)
try:
urllib.urlretrieve(download_url, filename, ReportHook)
if progress_event and progress_event.isSet():
print
except RuntimeError:
pass
def RunRevision(context, revision, zip_file, profile, num_runs, command, args):
"""Given a zipped revision, unzip it and run the test."""
print 'Trying revision %s...' % str(revision)
# Create a temp directory and unzip the revision into it.
cwd = os.getcwd()
tempdir = tempfile.mkdtemp(prefix='bisect_tmp')
UnzipFilenameToDir(zip_file, tempdir)
# Hack: Chrome OS archives are missing icudtl.dat; try to copy it from
# the local directory.
if context.platform == 'chromeos':
icudtl_path = 'third_party/icu/source/data/in/icudtl.dat'
if not os.access(icudtl_path, os.F_OK):
print 'Couldn\'t find: ' + icudtl_path
sys.exit()
os.system('cp %s %s/chrome-linux/' % (icudtl_path, tempdir))
os.chdir(tempdir)
# Run the build as many times as specified.
testargs = ['--user-data-dir=%s' % profile] + args
# The sandbox must be run as root on Official Chrome, so bypass it.
if ((context.is_official or context.flash_path) and
context.platform.startswith('linux')):
testargs.append('--no-sandbox')
if context.flash_path:
testargs.append('--ppapi-flash-path=%s' % context.flash_path)
# We have to pass a large enough Flash version, which currently needs not
# be correct. Instead of requiring the user of the script to figure out and
# pass the correct version we just spoof it.
testargs.append('--ppapi-flash-version=99.9.999.999')
runcommand = []
for token in shlex.split(command):
if token == '%a':
runcommand.extend(testargs)
else:
runcommand.append(
token.replace('%p', os.path.abspath(context.GetLaunchPath(revision))).
replace('%s', ' '.join(testargs)))
results = []
for _ in range(num_runs):
subproc = subprocess.Popen(runcommand,
bufsize=-1,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
(stdout, stderr) = subproc.communicate()
results.append((subproc.returncode, stdout, stderr))
os.chdir(cwd)
try:
shutil.rmtree(tempdir, True)
except Exception:
pass
for (returncode, stdout, stderr) in results:
if returncode:
return (returncode, stdout, stderr)
return results[0]
# The arguments official_builds, status, stdout and stderr are unused.
# They are present here because this function is passed to Bisect which then
# calls it with 5 arguments.
# pylint: disable=W0613
def AskIsGoodBuild(rev, official_builds, status, stdout, stderr):
"""Asks the user whether build |rev| is good or bad."""
# Loop until we get a response that we can parse.
while True:
response = raw_input('Revision %s is '
'[(g)ood/(b)ad/(r)etry/(u)nknown/(s)tdout/(q)uit]: ' %
str(rev))
if response in ('g', 'b', 'r', 'u'):
return response
if response == 'q':
raise SystemExit()
if response == 's':
print stdout
print stderr
def IsGoodASANBuild(rev, official_builds, status, stdout, stderr):
"""Determine if an ASAN build |rev| is good or bad
Will examine stderr looking for the error message emitted by ASAN. If not
found then will fallback to asking the user."""
if stderr:
bad_count = 0
for line in stderr.splitlines():
print line
if line.find('ERROR: AddressSanitizer:') != -1:
bad_count += 1
if bad_count > 0:
print 'Revision %d determined to be bad.' % rev
return 'b'
return AskIsGoodBuild(rev, official_builds, status, stdout, stderr)
class DownloadJob(object):
"""DownloadJob represents a task to download a given Chromium revision."""
def __init__(self, context, name, rev, zip_file):
super(DownloadJob, self).__init__()
# Store off the input parameters.
self.context = context
self.name = name
self.rev = rev
self.zip_file = zip_file
self.quit_event = threading.Event()
self.progress_event = threading.Event()
self.thread = None
def Start(self):
"""Starts the download."""
fetchargs = (self.context,
self.rev,
self.zip_file,
self.quit_event,
self.progress_event)
self.thread = threading.Thread(target=FetchRevision,
name=self.name,
args=fetchargs)
self.thread.start()
def Stop(self):
"""Stops the download which must have been started previously."""
assert self.thread, 'DownloadJob must be started before Stop is called.'
self.quit_event.set()
self.thread.join()
os.unlink(self.zip_file)
def WaitFor(self):
"""Prints a message and waits for the download to complete. The download
must have been started previously."""
assert self.thread, 'DownloadJob must be started before WaitFor is called.'
print 'Downloading revision %s...' % str(self.rev)
self.progress_event.set() # Display progress of download.
try:
while self.thread.isAlive():
# The parameter to join is needed to keep the main thread responsive to
# signals. Without it, the program will not respond to interruptions.
self.thread.join(1)
except (KeyboardInterrupt, SystemExit):
self.Stop()
raise
def Bisect(context,
num_runs=1,
command='%p %a',
try_args=(),
profile=None,
interactive=True,
evaluate=AskIsGoodBuild):
"""Given known good and known bad revisions, run a binary search on all
archived revisions to determine the last known good revision.
@param context PathContext object initialized with user provided parameters.
@param num_runs Number of times to run each build for asking good/bad.
@param try_args A tuple of arguments to pass to the test application.
@param profile The name of the user profile to run with.
@param interactive If it is false, use command exit code for good or bad
judgment of the argument build.
@param evaluate A function which returns 'g' if the argument build is good,
'b' if it's bad or 'u' if unknown.
Threading is used to fetch Chromium revisions in the background, speeding up
the user's experience. For example, suppose the bounds of the search are
good_rev=0, bad_rev=100. The first revision to be checked is 50. Depending on
whether revision 50 is good or bad, the next revision to check will be either
25 or 75. So, while revision 50 is being checked, the script will download
revisions 25 and 75 in the background. Once the good/bad verdict on rev 50 is
known:
- If rev 50 is good, the download of rev 25 is cancelled, and the next test
is run on rev 75.
- If rev 50 is bad, the download of rev 75 is cancelled, and the next test
is run on rev 25.
"""
if not profile:
profile = 'profile'
good_rev = context.good_revision
bad_rev = context.bad_revision
cwd = os.getcwd()
print 'Downloading list of known revisions...',
if not context.use_local_cache and not context.is_official:
print '(use --use-local-cache to cache and re-use the list of revisions)'
else:
print
_GetDownloadPath = lambda rev: os.path.join(cwd,
'%s-%s' % (str(rev), context.archive_name))
if context.is_official:
revlist = context.GetOfficialBuildsList()
else:
revlist = context.GetRevList()
# Get a list of revisions to bisect across.
if len(revlist) < 2: # Don't have enough builds to bisect.
msg = 'We don\'t have enough builds to bisect. revlist: %s' % revlist
raise RuntimeError(msg)
# Figure out our bookends and first pivot point; fetch the pivot revision.
minrev = 0
maxrev = len(revlist) - 1
pivot = maxrev / 2
rev = revlist[pivot]
zip_file = _GetDownloadPath(rev)
fetch = DownloadJob(context, 'initial_fetch', rev, zip_file)
fetch.Start()
fetch.WaitFor()
# Binary search time!
while fetch and fetch.zip_file and maxrev - minrev > 1:
if bad_rev < good_rev:
min_str, max_str = 'bad', 'good'
else:
min_str, max_str = 'good', 'bad'
print 'Bisecting range [%s (%s), %s (%s)].' % (revlist[minrev], min_str,
revlist[maxrev], max_str)
# Pre-fetch next two possible pivots
# - down_pivot is the next revision to check if the current revision turns
# out to be bad.
# - up_pivot is the next revision to check if the current revision turns
# out to be good.
down_pivot = int((pivot - minrev) / 2) + minrev
down_fetch = None
if down_pivot != pivot and down_pivot != minrev:
down_rev = revlist[down_pivot]
down_fetch = DownloadJob(context, 'down_fetch', down_rev,
_GetDownloadPath(down_rev))
down_fetch.Start()
up_pivot = int((maxrev - pivot) / 2) + pivot
up_fetch = None
if up_pivot != pivot and up_pivot != maxrev:
up_rev = revlist[up_pivot]
up_fetch = DownloadJob(context, 'up_fetch', up_rev,
_GetDownloadPath(up_rev))
up_fetch.Start()
# Run test on the pivot revision.
status = None
stdout = None
stderr = None
try:
(status, stdout, stderr) = RunRevision(context,
rev,
fetch.zip_file,
profile,
num_runs,
command,
try_args)
except Exception, e:
print >> sys.stderr, e
# Call the evaluate function to see if the current revision is good or bad.
# On that basis, kill one of the background downloads and complete the
# other, as described in the comments above.
try:
if not interactive:
if status:
answer = 'b'
print 'Bad revision: %s' % rev
else:
answer = 'g'
print 'Good revision: %s' % rev
else:
answer = evaluate(rev, context.is_official, status, stdout, stderr)
if ((answer == 'g' and good_rev < bad_rev)
or (answer == 'b' and bad_rev < good_rev)):
fetch.Stop()
minrev = pivot
if down_fetch:
down_fetch.Stop() # Kill the download of the older revision.
fetch = None
if up_fetch:
up_fetch.WaitFor()
pivot = up_pivot
fetch = up_fetch
elif ((answer == 'b' and good_rev < bad_rev)
or (answer == 'g' and bad_rev < good_rev)):
fetch.Stop()
maxrev = pivot
if up_fetch:
up_fetch.Stop() # Kill the download of the newer revision.
fetch = None
if down_fetch:
down_fetch.WaitFor()
pivot = down_pivot
fetch = down_fetch
elif answer == 'r':
pass # Retry requires no changes.
elif answer == 'u':
# Nuke the revision from the revlist and choose a new pivot.
fetch.Stop()
revlist.pop(pivot)
maxrev -= 1 # Assumes maxrev >= pivot.
if maxrev - minrev > 1:
# Alternate between using down_pivot or up_pivot for the new pivot
# point, without affecting the range. Do this instead of setting the
# pivot to the midpoint of the new range because adjacent revisions
# are likely affected by the same issue that caused the (u)nknown
# response.
if up_fetch and down_fetch:
fetch = [up_fetch, down_fetch][len(revlist) % 2]
elif up_fetch:
fetch = up_fetch
else:
fetch = down_fetch
fetch.WaitFor()
if fetch == up_fetch:
pivot = up_pivot - 1 # Subtracts 1 because revlist was resized.
else:
pivot = down_pivot
zip_file = fetch.zip_file
if down_fetch and fetch != down_fetch:
down_fetch.Stop()
if up_fetch and fetch != up_fetch:
up_fetch.Stop()
else:
assert False, 'Unexpected return value from evaluate(): ' + answer
except SystemExit:
print 'Cleaning up...'
for f in [_GetDownloadPath(revlist[down_pivot]),
_GetDownloadPath(revlist[up_pivot])]:
try:
os.unlink(f)
except OSError:
pass
sys.exit(0)
rev = revlist[pivot]
return (revlist[minrev], revlist[maxrev], context)
def GetBlinkDEPSRevisionForChromiumRevision(self, rev):
"""Returns the blink revision that was in REVISIONS file at
chromium revision |rev|."""
def _GetBlinkRev(url, blink_re):
m = blink_re.search(url.read())
url.close()
if m:
return m.group(1)
url = urllib.urlopen(DEPS_FILE_OLD % rev)
if url.getcode() == 200:
# . doesn't match newlines without re.DOTALL, so this is safe.
blink_re = re.compile(r'webkit_revision\D*(\d+)')
return int(_GetBlinkRev(url, blink_re))
else:
url = urllib.urlopen(DEPS_FILE_NEW % GetGitHashFromSVNRevision(rev))
if url.getcode() == 200:
blink_re = re.compile(r'webkit_revision\D*\d+;\D*\d+;(\w+)')
blink_git_sha = _GetBlinkRev(url, blink_re)
return self.GetSVNRevisionFromGitHash(blink_git_sha, 'blink')