| OLD | NEW |
| 1 # Copyright 2014 The Chromium Authors. All rights reserved. | 1 # Copyright 2014 The Chromium Authors. All rights reserved. |
| 2 # Use of this source code is governed by a BSD-style license that can be | 2 # Use of this source code is governed by a BSD-style license that can be |
| 3 # found in the LICENSE file. | 3 # found in the LICENSE file. |
| 4 | 4 |
| 5 from collections import defaultdict |
| 6 import copy |
| 5 from datetime import datetime | 7 from datetime import datetime |
| 6 import os | 8 import os |
| 7 | 9 |
| 8 from google.appengine.api import users | 10 from google.appengine.api import users |
| 9 | 11 |
| 10 from base_handler import BaseHandler | 12 from base_handler import BaseHandler |
| 11 from base_handler import Permission | 13 from base_handler import Permission |
| 14 from handlers import handlers_util |
| 15 from handlers import result_status |
| 16 from handlers.result_status import NO_TRY_JOB_REASON_MAP |
| 17 from model import wf_analysis_status |
| 12 from model.wf_analysis import WfAnalysis | 18 from model.wf_analysis import WfAnalysis |
| 13 from model.wf_analysis_result_status import RESULT_STATUS_TO_DESCRIPTION | 19 from model.wf_analysis_result_status import RESULT_STATUS_TO_DESCRIPTION |
| 14 from waterfall import build_failure_analysis_pipelines | 20 from waterfall import build_failure_analysis_pipelines |
| 15 from waterfall import buildbot | 21 from waterfall import buildbot |
| 16 from waterfall import waterfall_config | 22 from waterfall import waterfall_config |
| 17 | 23 |
| 18 | 24 |
| 19 BUILD_FAILURE_ANALYSIS_TASKQUEUE = 'build-failure-analysis-queue' | 25 BUILD_FAILURE_ANALYSIS_TASKQUEUE = 'build-failure-analysis-queue' |
| 26 NON_SWARMING = object() |
| 20 | 27 |
| 21 | 28 |
| 22 def _FormatDatetime(dt): | 29 def _FormatDatetime(dt): |
| 23 if not dt: | 30 if not dt: |
| 24 return None | 31 return None |
| 25 else: | 32 else: |
| 26 return dt.strftime('%Y-%m-%d %H:%M:%S UTC') | 33 return dt.strftime('%Y-%m-%d %H:%M:%S UTC') |
| 27 | 34 |
| 28 | 35 |
| 29 def _GetTriageHistory(analysis): | 36 def _GetTriageHistory(analysis): |
| 30 if (not users.is_current_user_admin() or | 37 if (not users.is_current_user_admin() or |
| 31 not analysis.completed or | 38 not analysis.completed or |
| 32 not analysis.triage_history): | 39 not analysis.triage_history): |
| 33 return None | 40 return None |
| 34 | 41 |
| 35 triage_history = [] | 42 triage_history = [] |
| 36 for triage_record in analysis.triage_history: | 43 for triage_record in analysis.triage_history: |
| 37 triage_history.append({ | 44 triage_history.append({ |
| 38 'triage_time': _FormatDatetime( | 45 'triage_time': _FormatDatetime( |
| 39 datetime.utcfromtimestamp(triage_record['triage_timestamp'])), | 46 datetime.utcfromtimestamp(triage_record['triage_timestamp'])), |
| 40 'user_name': triage_record['user_name'], | 47 'user_name': triage_record['user_name'], |
| 41 'result_status': RESULT_STATUS_TO_DESCRIPTION.get( | 48 'result_status': RESULT_STATUS_TO_DESCRIPTION.get( |
| 42 triage_record['result_status']), | 49 triage_record['result_status']), |
| 43 'version': triage_record.get('version'), | 50 'version': triage_record.get('version'), |
| 44 }) | 51 }) |
| 45 | 52 |
| 46 return triage_history | 53 return triage_history |
| 47 | 54 |
| 48 | 55 |
| 56 def _GetOrganizedAnalysisResultBySuspectedCL(analysis_result): |
| 57 """Group tests it they have the same suspected CLs.""" |
| 58 organized_results = defaultdict(list) |
| 59 |
| 60 if not analysis_result: |
| 61 return organized_results |
| 62 |
| 63 for step_failure in analysis_result.get('failures', []): |
| 64 step_name = step_failure['step_name'] |
| 65 supported = step_failure.get('supported', True) |
| 66 step_revisions_index = {} |
| 67 organized_suspected_cls = organized_results[step_name] |
| 68 |
| 69 if not step_failure.get('tests'): |
| 70 # Non swarming, just group the whole step together. |
| 71 shared_result = { |
| 72 'first_failure': step_failure['first_failure'], |
| 73 'last_pass': step_failure.get('last_pass'), |
| 74 'supported': supported, |
| 75 'tests': [], |
| 76 'suspected_cls': step_failure['suspected_cls'] |
| 77 } |
| 78 organized_suspected_cls.append(shared_result) |
| 79 continue |
| 80 |
| 81 # Swarming tests. |
| 82 for index, cl in enumerate(step_failure['suspected_cls']): |
| 83 step_revisions_index[cl['revision']] = index |
| 84 |
| 85 # Groups tests by suspected CLs' revision. |
| 86 # Keys are the indices of each test in the test list. |
| 87 # Format is as below: |
| 88 # { |
| 89 # 1: { |
| 90 # 'tests': ['test1', 'test2'], |
| 91 # 'revisions': ['rev1'], |
| 92 # 'suspected_cls': [ |
| 93 # # suspected cl info for rev1 at step level. |
| 94 # ] |
| 95 # }, |
| 96 # 3: { |
| 97 # 'tests': ['test3'], |
| 98 # 'revisions': ['rev3', 'rev2'], |
| 99 # 'suspected_cls': [ |
| 100 # # suspected cl info for rev2, rev3 at step level. |
| 101 # ] |
| 102 # } |
| 103 # } |
| 104 tests_group = defaultdict(list) |
| 105 for index, test in enumerate(step_failure['tests']): |
| 106 # Get all revisions for this test and check if there is |
| 107 # any other test has the same culprit(represented by revision) set. |
| 108 test_name = test['test_name'] |
| 109 found_group = False |
| 110 revisions = set() |
| 111 for cl in test['suspected_cls']: |
| 112 revisions.add(cl['revision']) |
| 113 for group in tests_group.values(): |
| 114 # Found tests that have the same culprit(represented by revision), |
| 115 # add current test to group. |
| 116 if revisions == set(group['revisions']): |
| 117 group['tests'].append(test_name) |
| 118 found_group = True |
| 119 break |
| 120 if not found_group: |
| 121 # First test with that revision set, add a new group. |
| 122 group_suspected_cls = [] |
| 123 for revision in revisions: |
| 124 group_suspected_cls.append( |
| 125 step_failure['suspected_cls'][step_revisions_index[revision]]) |
| 126 tests_group[index] = { |
| 127 'tests': [test_name], |
| 128 'revisions': list(revisions), |
| 129 'suspected_cls': group_suspected_cls |
| 130 } |
| 131 |
| 132 for index, group in tests_group.iteritems(): |
| 133 # Reorganize heuristic results by culprits. |
| 134 test_result = step_failure['tests'][index] |
| 135 shared_result = { |
| 136 'first_failure': test_result['first_failure'], |
| 137 'last_pass': test_result.get('last_pass'), |
| 138 'supported': supported, |
| 139 'tests': group['tests'], |
| 140 'suspected_cls': group['suspected_cls'] |
| 141 } |
| 142 organized_suspected_cls.append(shared_result) |
| 143 |
| 144 return organized_results |
| 145 |
| 146 |
| 147 def _GetAnalysisResultWithTryJobInfo( |
| 148 organized_results, master_name, builder_name, build_number): |
| 149 """Reorganizes analysis result and try job result by step_name and culprit. |
| 150 |
| 151 Returns: |
| 152 update_result (dict): A dict of classified results. |
| 153 |
| 154 The format for those dicts are as below: |
| 155 { |
| 156 # A dict of results that contains both |
| 157 # heuristic analysis results and try job results. |
| 158 'reliable_failures': { |
| 159 'step1': { |
| 160 'results': [ |
| 161 { |
| 162 'try_job':{ |
| 163 'ref_name': 'step1', |
| 164 'try_job_key': 'm/b/119', |
| 165 'status': wf_analysis_status.ANALYZED, |
| 166 'try_job_url': 'url/121', |
| 167 'try_job_build_number': 121, |
| 168 }, |
| 169 'heuristic_analysis': { |
| 170 'suspected_cls': [ |
| 171 { |
| 172 'build_number': 98, |
| 173 'repo_name': 'chromium', |
| 174 'revision': 'r98_1', |
| 175 'commit_position': None, |
| 176 'url': None, |
| 177 'score': 5, |
| 178 'hints': { |
| 179 'added f98.cc (and it was in log)': 5, |
| 180 }, |
| 181 } |
| 182 ] |
| 183 } |
| 184 'tests': ['test1', 'test2'], |
| 185 'first_failure': 98, |
| 186 'last_pass': 97, |
| 187 'supported': True |
| 188 } |
| 189 ] |
| 190 } |
| 191 }, |
| 192 # A dict of result for flaky tests. |
| 193 'flaky_failures': {...}, |
| 194 # A dict of results for all the other conditions, |
| 195 # such as non-swarming tests or swarming rerun failed. |
| 196 'unclassified_failures': {...} |
| 197 } |
| 198 """ |
| 199 updated_results = defaultdict(lambda: defaultdict(lambda: defaultdict(list))) |
| 200 |
| 201 if not organized_results: |
| 202 return updated_results |
| 203 |
| 204 try_job_info = handlers_util.GetAllTryJobResults( |
| 205 master_name, builder_name, build_number) |
| 206 if not try_job_info: |
| 207 return updated_results |
| 208 |
| 209 for step_name, try_jobs in try_job_info.iteritems(): |
| 210 try_jobs = try_jobs['try_jobs'] |
| 211 step_heuristic_results = organized_results[step_name] |
| 212 step_updated_results = updated_results[step_name]['results'] |
| 213 |
| 214 # Finds out try job result index and heuristic result index for each test. |
| 215 test_result_map = defaultdict(lambda: defaultdict(int)) |
| 216 |
| 217 for index, try_job in enumerate(try_jobs): |
| 218 if not try_job.get('tests'): # Compile or non-swarming. |
| 219 test_result_map[NON_SWARMING]['try_job_index'] = index |
| 220 continue |
| 221 for test_name in try_job['tests']: |
| 222 test_result_map[test_name]['try_job_index'] = index |
| 223 |
| 224 for index, heuristic_result in enumerate(step_heuristic_results): |
| 225 if not heuristic_result.get('tests'): # Compile or non-swarming. |
| 226 test_result_map[NON_SWARMING]['heuristic_index'] = index |
| 227 continue |
| 228 for test_name in heuristic_result['tests']: |
| 229 test_result_map[test_name]['heuristic_index'] = index |
| 230 |
| 231 # Group tests based on indices. |
| 232 indices_test_map = defaultdict(list) |
| 233 for test_name, indices in test_result_map.iteritems(): |
| 234 indices_test_map[ |
| 235 (indices['try_job_index'], indices['heuristic_index'])].append( |
| 236 test_name) |
| 237 |
| 238 for (try_job_index, heuristic_index), tests in indices_test_map.iteritems(): |
| 239 try_job_result = try_jobs[try_job_index] |
| 240 heuristic_result = step_heuristic_results[heuristic_index] |
| 241 |
| 242 final_result = { |
| 243 'try_job': try_job_result, |
| 244 'heuristic_analysis': { |
| 245 'suspected_cls': heuristic_result['suspected_cls'] |
| 246 }, |
| 247 'tests': tests if tests != [NON_SWARMING] else [], |
| 248 'first_failure': heuristic_result['first_failure'], |
| 249 'last_pass': heuristic_result['last_pass'], |
| 250 'supported': heuristic_result['supported'] |
| 251 } |
| 252 |
| 253 if try_job_result['status'] == result_status.FLAKY: |
| 254 step_updated_results['flaky_failures'].append(final_result) |
| 255 elif try_job_result['status'] in NO_TRY_JOB_REASON_MAP.values(): |
| 256 # There is no try job info but only heuristic result. |
| 257 step_updated_results['unclassified_failures'].append(final_result) |
| 258 else: |
| 259 step_updated_results['reliable_failures'].append(final_result) |
| 260 |
| 261 return updated_results |
| 262 |
| 263 |
| 49 class BuildFailure(BaseHandler): | 264 class BuildFailure(BaseHandler): |
| 50 PERMISSION_LEVEL = Permission.ANYONE | 265 PERMISSION_LEVEL = Permission.ANYONE |
| 51 | 266 |
| 52 def _ShowDebugInfo(self): | 267 def _ShowDebugInfo(self): |
| 53 # Show debug info only if the app is run locally during development, if the | 268 # Show debug info only if the app is run locally during development, if the |
| 54 # currently logged-in user is an admin, or if it is explicitly requested | 269 # currently logged-in user is an admin, or if it is explicitly requested |
| 55 # with parameter 'debug=1'. | 270 # with parameter 'debug=1'. |
| 56 return (os.environ['SERVER_SOFTWARE'].startswith('Development') or | 271 return (os.environ['SERVER_SOFTWARE'].startswith('Development') or |
| 57 users.is_current_user_admin() or self.request.get('debug') == '1') | 272 users.is_current_user_admin() or self.request.get('debug') == '1') |
| 58 | 273 |
| (...skipping 29 matching lines...) Expand all Loading... |
| 88 # Only allow admin to force a re-run and set the build_completed. | 303 # Only allow admin to force a re-run and set the build_completed. |
| 89 force = (users.is_current_user_admin() and | 304 force = (users.is_current_user_admin() and |
| 90 self.request.get('force') == '1') | 305 self.request.get('force') == '1') |
| 91 build_completed = (users.is_current_user_admin() and | 306 build_completed = (users.is_current_user_admin() and |
| 92 self.request.get('build_completed') == '1') | 307 self.request.get('build_completed') == '1') |
| 93 analysis = build_failure_analysis_pipelines.ScheduleAnalysisIfNeeded( | 308 analysis = build_failure_analysis_pipelines.ScheduleAnalysisIfNeeded( |
| 94 master_name, builder_name, build_number, | 309 master_name, builder_name, build_number, |
| 95 build_completed=build_completed, force=force, | 310 build_completed=build_completed, force=force, |
| 96 queue_name=BUILD_FAILURE_ANALYSIS_TASKQUEUE) | 311 queue_name=BUILD_FAILURE_ANALYSIS_TASKQUEUE) |
| 97 | 312 |
| 313 organized_results = _GetOrganizedAnalysisResultBySuspectedCL( |
| 314 analysis.result) |
| 315 analysis_result = _GetAnalysisResultWithTryJobInfo( |
| 316 organized_results, *build_info) |
| 317 |
| 98 data = { | 318 data = { |
| 99 'master_name': analysis.master_name, | 319 'master_name': analysis.master_name, |
| 100 'builder_name': analysis.builder_name, | 320 'builder_name': analysis.builder_name, |
| 101 'build_number': analysis.build_number, | 321 'build_number': analysis.build_number, |
| 102 'pipeline_status_path': analysis.pipeline_status_path, | 322 'pipeline_status_path': analysis.pipeline_status_path, |
| 103 'show_debug_info': self._ShowDebugInfo(), | 323 'show_debug_info': self._ShowDebugInfo(), |
| 104 'analysis_request_time': _FormatDatetime(analysis.request_time), | 324 'analysis_request_time': _FormatDatetime(analysis.request_time), |
| 105 'analysis_start_time': _FormatDatetime(analysis.start_time), | 325 'analysis_start_time': _FormatDatetime(analysis.start_time), |
| 106 'analysis_end_time': _FormatDatetime(analysis.end_time), | 326 'analysis_end_time': _FormatDatetime(analysis.end_time), |
| 107 'analysis_duration': analysis.duration, | 327 'analysis_duration': analysis.duration, |
| 108 'analysis_update_time': _FormatDatetime(analysis.updated_time), | 328 'analysis_update_time': _FormatDatetime(analysis.updated_time), |
| 109 'analysis_completed': analysis.completed, | 329 'analysis_completed': analysis.completed, |
| 110 'analysis_failed': analysis.failed, | 330 'analysis_failed': analysis.failed, |
| 111 'analysis_result': analysis.result, | 331 'analysis_result': analysis_result, |
| 112 'analysis_correct': analysis.correct, | 332 'analysis_correct': analysis.correct, |
| 113 'triage_history': _GetTriageHistory(analysis), | 333 'triage_history': _GetTriageHistory(analysis), |
| 114 'show_triage_help_button': self._ShowTriageHelpButton(), | 334 'show_triage_help_button': self._ShowTriageHelpButton(), |
| 335 'status_message_map': result_status.STATUS_MESSAGE_MAP |
| 115 } | 336 } |
| 116 | 337 |
| 117 return {'template': 'build_failure.html', 'data': data} | 338 return {'template': 'build_failure.html', 'data': data} |
| 118 | 339 |
| 119 def HandlePost(self): # pragma: no cover | 340 def HandlePost(self): # pragma: no cover |
| 120 return self.HandleGet() | 341 return self.HandleGet() |
| OLD | NEW |