#!/usr/bin/env python3
# Copyright 2017 The Chromium Authors
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Makes sure OWNERS files have consistent TEAM and COMPONENT tags."""
from __future__ import print_function
import json
import logging
import argparse
import os
import re
import sys
import urllib.request
from collections import defaultdict
from owners_file_tags import parse, uniform_path_format
DEFAULT_MAPPING_URL = \
'https://storage.googleapis.com/chromium-owners/component_map.json'
# Represent tags that may be found in OWNERS files.
_COMPONENT = 'COMPONENT'
_TEAM = 'TEAM'
# Represents the pattern of a line that contains a TEAM or COMPONENT tag. Note
# that matching lines may not have valid formatting. For example, if {} is
# replaced with TEAM, the line ' #TEAMS :[email protected]' matches the
# pattern; however, the correct formatting is '# TEAM: [email protected]'.
LINE_PATTERN = '.*#\s*{}.*:.*'
def _get_entries(lines, tag):
"""Returns a line for the given tag, e.g. '# TEAM: [email protected]'.
Also, validates the tag's prefix format.
Args:
lines: A list of lines from which to extract information, e.g.
['[email protected]', '# COMPONENT: UI', '# TEAM: [email protected]'].
tag: The string tag denoting the entry for which to extract information,
e.g. 'COMPONENT'.
Raises:
ValueError: Raised if (A) a line matches the specified line pattern, but
does not match the the prefix pattern. This can happen, e.g., if a line is
'#TEAM: [email protected]', which is missing a space between # and TEAM.
Also, raised if (B) there are multiple lines matching the line pattern.
This can happen if there are, e.g. two team lines.
"""
line_pattern = re.compile(LINE_PATTERN.format(tag))
entries = [line for line in lines if line_pattern.match(line)]
if not entries:
return None
if len(entries) == 1:
if re.match('^# {}: .*$'.format(tag), entries[0]):
return entries[0]
raise ValueError(
'{} has an invalid prefix. Use \'# {}: \''.format(entries[0], tag))
raise ValueError('Contains more than one {} per directory'.format(tag))
def rel_and_full_paths(root, owners_path):
if root:
full_path = os.path.join(root, owners_path)
rel_path = owners_path
else:
full_path = os.path.abspath(owners_path)
rel_path = os.path.relpath(owners_path)
return rel_path, full_path
def validate_mappings(arguments, owners_files):
"""Ensure team/component mapping remains consistent after patch.
The main purpose of this check is to notify the user if any edited (or added)
team tag makes a component map to multiple teams.
Args:
arguments: Command line arguments from argparse
owners_files: List of paths to affected OWNERS files
Returns:
A string containing the details of any multi-team per component.
"""
mappings_file = json.load(
urllib.request.urlopen(arguments.current_mapping_url))
new_dir_to_component = mappings_file.get('dir-to-component', {})
new_dir_to_team = mappings_file.get('dir-to-team', {})
affected = {}
deleted = []
affected_components = set()
# Parse affected OWNERS files
for f in owners_files:
rel, full = rel_and_full_paths(arguments.root, f)
if os.path.exists(full):
affected[os.path.dirname(rel)] = parse(full)
else:
deleted.append(os.path.dirname(rel))
# Update component mapping with current changes.
for rel_path_native, tags in affected.items():
# Make the path use forward slashes always.
rel_path = uniform_path_format(rel_path_native)
component = tags.get('component')
team = tags.get('team')
os_tag = tags.get('os')
if component:
if os_tag:
component = '%s(%s)' % (component, os_tag)
new_dir_to_component[rel_path] = component
affected_components.add(component)
elif rel_path in new_dir_to_component:
del new_dir_to_component[rel_path]
if team:
new_dir_to_team[rel_path] = team
elif rel_path in new_dir_to_team:
del new_dir_to_team[rel_path]
for deleted_dir in deleted:
if deleted_dir in new_dir_to_component:
del new_dir_to_component[deleted_dir]
if deleted_dir in new_dir_to_team:
del new_dir_to_team[deleted_dir]
# For the components affected by this patch, compute the directories that map
# to it.
affected_component_to_dirs = defaultdict(list)
for d, component in new_dir_to_component.items():
if component in affected_components:
affected_component_to_dirs[component].append(d)
# Convert component->[dirs], dir->team to component->[teams].
affected_component_to_teams = {
component:
list({new_dir_to_team[d]
for d in dirs if d in new_dir_to_team})
for component, dirs in affected_component_to_dirs.items()
}
# Perform cardinality check.
warnings = ''
for component, teams in affected_component_to_teams.items():
if len(teams) > 1:
warnings += ('\nThe set of all OWNERS files with COMPONENT: %s list '
"multiple TEAM's: %s") % (component, ', '.join(teams))
if warnings:
warnings = ('Are you sure these are correct? After landing this patch:%s'
% warnings)
return warnings
def check_owners(rel_path, full_path):
"""Component and Team check in OWNERS files. crbug.com/667954"""
def result_dict(error):
return {
'error': error,
'full_path': full_path,
'rel_path': rel_path,
}
if not os.path.exists(full_path):
return None
with open(full_path) as f:
owners_file_lines = f.readlines()
try:
component_line = _get_entries(owners_file_lines, _COMPONENT)
team_line = _get_entries(owners_file_lines, _TEAM)
except ValueError as error:
return result_dict(str(error))
if component_line:
component = component_line.split(':')[1]
if not component:
return result_dict('Has COMPONENT line but no component name')
# Check for either of the following formats:
# component1, component2, ...
# component1,component2,...
# component1 component2 ...
component_count = max(
len(component.strip().split()),
len(component.strip().split(',')))
if component_count > 1:
return result_dict('Has more than one component name')
# TODO(robertocn): Check against a static list of valid components,
# perhaps obtained from monorail at the beginning of presubmit.
if team_line:
team_entry_parts = team_line.split('@')
if len(team_entry_parts) != 2:
return result_dict('Has TEAM line, but not exactly 1 team email')
# TODO(robertocn): Raise a warning if only one of (COMPONENT, TEAM) is
# present.
def main():
usage = """Usage: python %prog [--root <dir>] <owners_file1> <owners_file2>...
owners_fileX specifies the path to the file to check, these are expected
to be relative to the root directory if --root is used.
Examples:
python %prog --root /home/<user>/chromium/src/ tools/OWNERS v8/OWNERS
python %prog /home/<user>/chromium/src/tools/OWNERS
python %prog ./OWNERS
"""
parser = argparse.ArgumentParser(usage=usage)
parser.add_argument('--root', help='Specifies the repository root.')
parser.add_argument('-v',
'--verbose',
action='count',
default=0,
help='Print debug logging')
parser.add_argument('--bare',
action='store_true',
default=False,
help='Prints the bare filename triggering the checks')
parser.add_argument(
'--current_mapping_url',
default=DEFAULT_MAPPING_URL,
help='URL for existing dir/component and component/team mapping')
parser.add_argument('--json', help='Path to JSON output file')
args, owners_files = parser.parse_known_args()
levels = [logging.ERROR, logging.INFO, logging.DEBUG]
logging.basicConfig(level=levels[min(len(levels) - 1, args.verbose)])
errors = list(
filter(None, (check_owners(*rel_and_full_paths(args.root, f))
for f in owners_files)))
warnings = None
if not errors:
warnings = validate_mappings(args, owners_files)
if args.json:
with open(args.json, 'w') as f:
json.dump(errors, f)
if errors:
if args.bare:
print('\n'.join(e['full_path'] for e in errors))
else:
print('\nFAILED\n')
print('\n'.join('%s: %s' % (e['full_path'], e['error']) for e in errors))
return 1
if not args.bare:
if warnings:
print(warnings)
return 0
if '__main__' == __name__:
sys.exit(main())
Codebase Browser
chromium