analyze: send results to event_mon

scripts/slave/send_analyze_event.py

+#!/usr/bin/env python
+# Copyright 2016 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.
+
+"""Sends analyze results to event_mon.
+
+Needs to be run via tools/scripts/runit.py"""

[chrishall, 2016/10/07 06:36:26]

is it worth giving an example invocation?

[Paweł Hajdan Jr., 2016/10/07 10:50:54]

Recipe expectations contain some examples.

+
+
+import argparse
+import json
+import sys
+
+from infra_libs import event_mon
+
+
+def inner_main(args):
+  event = event_mon.Event(timestamp_kind='POINT')
+  analyze_event = event.proto.analyze_event
+
+  if args.master_name:
+    analyze_event.master_name = args.master_name
+  if args.builder_name:
+    analyze_event.builder_name = args.builder_name
+  if args.build_id:
+    analyze_event.build_id = args.build_id
+  if args.analyze_input:
+    analyze_input = json.load(args.analyze_input)
+    analyze_event.affected_files.extend(analyze_input['files'])
+    analyze_event.input_test_targets.extend(analyze_input['test_targets'])
+    analyze_event.input_compile_targets.extend(analyze_input[
+        'additional_compile_targets'])
+  if args.analyze_output:
+    analyze_output = json.load(args.analyze_output)
+
+    if 'test_targets' in analyze_output:
+      analyze_event.output_test_targets.extend(analyze_output['test_targets'])
+    if 'compile_targets' in analyze_output:
+      analyze_event.output_compile_targets.extend(
+          analyze_output['compile_targets'])
+
+    if 'error' in analyze_output:
+      analyze_event.result = analyze_event.AnalyzeResult.Value('ERROR')
+    elif 'invalid_targets' in analyze_output:
+      analyze_event.result = analyze_event.AnalyzeResult.Value(
+          'INVALID_TARGETS')
+      analyze_event.invalid_targets.extend(analyze_output['invalid_targets'])
+    elif analyze_output.get('status') == 'Found dependency':
+      analyze_event.result = analyze_event.AnalyzeResult.Value(
+          'FOUND_DEPENDENCY')
+    elif analyze_output.get('status') == 'Found dependency (all)':
+      analyze_event.result = analyze_event.AnalyzeResult.Value(
+          'FOUND_DEPENDENCY_ALL')
+    elif (analyze_output.get('status') ==
+          'No compile necessary (all files ignored)'):
+      analyze_event.result = analyze_event.AnalyzeResult.Value(
+          'ALL_FILES_IGNORED')
+    elif (analyze_output.get('status') ==
+          'Analyze disabled: matched exclusion'):
+      analyze_event.result = analyze_event.AnalyzeResult.Value(
+          'MATCHED_EXCLUSION')
+    elif (analyze_output.get('status') ==
+          'No dependency'):
+      analyze_event.result = analyze_event.AnalyzeResult.Value(
+          'NO_COMPILE_NECESSARY')
+
+  print event.proto

[chrishall, 2016/10/07 06:36:26]

should this print be here?

is this print statement the output being tested by all the json files?

[Paweł Hajdan Jr., 2016/10/07 10:50:54]

I find it useful, at least for now.
No, recipe expectations do not actually execute the script.

+
+  if event.send():
+    return 0
+
+  return 1
+
+
+def main(argv):
+  parser = argparse.ArgumentParser()
+  parser.add_argument('--master-name', help='Buildbot master name')
+  parser.add_argument('--builder-name', help='Buildbot builder name')
+  parser.add_argument('--build-id', help='Build ID (buildnumber)')
+  parser.add_argument('--analyze-input', help='JSON input passed to analyze',
+                      type=argparse.FileType('r'))
+  parser.add_argument('--analyze-output', help='JSON output from analyze',
+                      type=argparse.FileType('r'))
+  event_mon.add_argparse_options(parser)
+  args = parser.parse_args(argv)
+  event_mon.process_argparse_options(args)
+
+  try:
+    return inner_main(args)
+  finally:
+    event_mon.close()
+
+
+if __name__ == '__main__':
+  sys.exit(main(sys.argv[1:]))