summaryrefslogtreecommitdiffstats
path: root/tools/valgrind/scan-build.py
blob: b58b6cc447e52871011ab3208fdd27c0fa9da766 (plain)
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
#!/usr/bin/env python
# Copyright (c) 2013 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.

import argparse
import errno
import os
import re
import sys
import urllib
import urllib2

# Where all the data lives.
ROOT_URL = "http://build.chromium.org/p/chromium.memory.fyi/builders"

# TODO(groby) - support multi-line search from the command line. Useful when
# scanning for classes of failures, see below.
SEARCH_STRING = """<p class=\"failure result\">
Failed memory test: content
</p>"""

# Location of the log cache.
CACHE_DIR = "buildlogs.tmp"

# If we don't find anything after searching |CUTOFF| logs, we're probably done.
CUTOFF = 100

def EnsurePath(path):
  """Makes sure |path| does exist, tries to create it if it doesn't."""
  try:
    os.makedirs(path)
  except OSError as exception:
    if exception.errno != errno.EEXIST:
      raise


class Cache(object):
  def __init__(self, root_dir):
    self._root_dir = os.path.abspath(root_dir)

  def _LocalName(self, name):
    """If name is a relative path, treat it as relative to cache root.
       If it is absolute and under cache root, pass it through.
       Otherwise, raise error.
    """
    if os.path.isabs(name):
      assert os.path.commonprefix([name, self._root_dir]) == self._root_dir
    else:
      name = os.path.join(self._root_dir, name)
    return name

  def _FetchLocal(self, local_name):
    local_name = self._LocalName(local_name)
    EnsurePath(os.path.dirname(local_name))
    if os.path.exists(local_name):
      f = open(local_name, 'r')
      return f.readlines();
    return None

  def _FetchRemote(self, remote_name):
    try:
      response = urllib2.urlopen(remote_name)
    except:
      print "Could not fetch", remote_name
      raise
    return response.read()

  def Update(self, local_name, remote_name):
    local_name = self._LocalName(local_name)
    EnsurePath(os.path.dirname(local_name))
    blob = self._FetchRemote(remote_name)
    f = open(local_name, "w")
    f.write(blob)
    return blob.splitlines()

  def FetchData(self, local_name, remote_name):
    result = self._FetchLocal(local_name)
    if result:
      return result
    # If we get here, the local cache does not exist yet. Fetch, and store.
    return self.Update(local_name, remote_name)


class Builder(object):
  def __init__(self, waterfall, name):
    self._name = name
    self._waterfall = waterfall

  def Name(self):
    return self._name

  def LatestBuild(self):
    return self._waterfall.GetLatestBuild(self._name)

  def GetBuildPath(self, build_num):
    return "%s/%s/builds/%d" % (
        self._waterfall._root_url, urllib.quote(self._name), build_num)

  def _FetchBuildLog(self, build_num):
    local_build_path = "builds/%s" % self._name
    local_build_file = os.path.join(local_build_path, "%d.log" % build_num)
    return self._waterfall._cache.FetchData(local_build_file,
                                            self.GetBuildPath(build_num))

  def _CheckLog(self, build_num, tester):
    log_lines = self._FetchBuildLog(build_num)
    return any(tester(line) for line in log_lines)

  def ScanLogs(self, tester):
    occurrences = []
    build = self.LatestBuild()
    no_results = 0
    while build != 0 and no_results < CUTOFF:
      if self._CheckLog(build, tester):
        occurrences.append(build)
      else:
        no_results = no_results + 1
      build = build - 1
    return occurrences


class Waterfall(object):
  def __init__(self, root_url, cache_dir):
    self._root_url = root_url
    self._builders = {}
    self._top_revision = {}
    self._cache = Cache(cache_dir)

  def Builders(self):
    return self._builders.values()

  def Update(self):
    self._cache.Update("builders", self._root_url)
    self.FetchInfo()

  def FetchInfo(self):
    if self._top_revision:
      return

    html = self._cache.FetchData("builders", self._root_url)

    """ Search for both builders and latest build number in HTML
    <td class="box"><a href="builders/<builder-name>"> identifies a builder
    <a href="builders/<builder-name>/builds/<build-num>"> is the latest build.
    """
    box_matcher = re.compile('.*a href[^>]*>([^<]*)\<')
    build_matcher = re.compile('.*a href=\"builders/(.*)/builds/([0-9]+)\".*')
    last_builder = ""
    for line in html:
      if 'a href="builders/' in line:
        if 'td class="box"' in line:
          last_builder = box_matcher.match(line).group(1)
          self._builders[last_builder] = Builder(self, last_builder)
        else:
          result = build_matcher.match(line)
          builder = result.group(1)
          assert builder == urllib.quote(last_builder)
          self._top_revision[last_builder] = int(result.group(2))

  def GetLatestBuild(self, name):
    self.FetchInfo()
    assert self._top_revision
    return self._top_revision[name]


class MultiLineChange(object):
  def __init__(self, lines):
    self._tracked_lines = lines
    self._current = 0

  def __call__(self, line):
    """ Test a single line against multi-line change.

    If it matches the currently active line, advance one line.
    If the current line is the last line, report a match.
    """
    if self._tracked_lines[self._current] in line:
      self._current = self._current + 1
      if self._current == len(self._tracked_lines):
        self._current = 0
        return True
    else:
      self._current = 0
    return False


def main(argv):
  # Create argument parser.
  parser = argparse.ArgumentParser()
  commands = parser.add_mutually_exclusive_group(required=True)
  commands.add_argument("--update", action='store_true')
  commands.add_argument("--find", metavar='search term')
  args = parser.parse_args()

  path = os.path.abspath(os.path.dirname(argv[0]))
  cache_path = os.path.join(path, CACHE_DIR)

  fyi = Waterfall(ROOT_URL, cache_path)

  if args.update:
    fyi.Update()
    for builder in fyi.Builders():
      print "Updating", builder.Name()
      builder.ScanLogs(lambda x:False)

  if args.find:
    tester = MultiLineChange(args.find.splitlines())
    fyi.FetchInfo()

    print "SCANNING FOR ", args.find
    for builder in fyi.Builders():
      print "Scanning", builder.Name()
      occurrences = builder.ScanLogs(tester)
      if occurrences:
        min_build = min(occurrences)
        path = builder.GetBuildPath(min_build)
        print "Earliest occurrence in build %d" % min_build
        print "Latest occurrence in build %d" % max(occurrences)
        print "Latest build: %d" % builder.LatestBuild()
        print path
        print "%d total" % len(occurrences)


if __name__ == "__main__":
  sys.exit(main(sys.argv))