From 8fd62207ed7404e5f1231392ed33624f42451c7f Mon Sep 17 00:00:00 2001
From: Ray Strode <halfline@gmail.com>
Date: Mon, 19 Oct 2026 13:30:12 -0400
Subject: [PATCH v2 1/2] utils: Find maintainers from Git history
From: Ray Strode <rstrode@redhat.com>
This commit adds get_maintainer.py to find maintainers and reviewers from
local Git history. Contributors can supply patches or repository paths
instead of searching commit logs manually for people familiar with an area.
The utility ranks review and acknowledgment credits ahead of authorship,
normalizes identities through mailmap, and limits contacts per path so
busy areas do not crowd out quiet ones. Each result includes roles and
commit counts to explain the match.
History ends at the configured upstream by default to exclude local
unsubmitted work. Kernel-style options control the history window,
selection thresholds, and annotations for use with git send-email.
Signed-off-by: Ray Strode <rstrode@redhat.com>
---
utils/get_maintainer.py | 316 ++++++++++++++++++++++++++++++++++++++++
1 file changed, 316 insertions(+)
create mode 100755 utils/get_maintainer.py
new file mode 100755
@@ -0,0 +1,316 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: GPL-2.0-or-later
+# Copyright (C) 2026, Red Hat Inc.
+#
+# Author: Ray Strode <rstrode@redhat.com>
+#
+# Find libcamera maintainers and reviewers from local Git history.
+
+import argparse
+import subprocess
+import sys
+from dataclasses import dataclass, field
+from datetime import datetime, timezone
+from email.headerregistry import Address
+from email.utils import parseaddr
+from pathlib import Path, PurePosixPath
+
+
+MAILING_LIST = 'libcamera-devel@lists.libcamera.org'
+
+
+def git(directory, *args, input=None):
+ """Run a read-only Git command, keeping diagnostics off standard output."""
+ result = subprocess.run(['git', '--no-pager', '--literal-pathspecs',
+ '-C', str(directory), *args],
+ input=input, stdout=subprocess.PIPE,
+ stderr=subprocess.PIPE)
+ if result.returncode:
+ raise ValueError(result.stderr.decode(errors='replace').strip())
+ return result.stdout.decode('utf-8', errors='surrogateescape')
+
+
+def repository():
+ try:
+ root = git(Path.cwd(), 'rev-parse', '--show-toplevel')
+ except ValueError:
+ root = git(Path(__file__).resolve().parent, 'rev-parse',
+ '--show-toplevel')
+ return Path(root.strip()).resolve()
+
+
+def default_revision(root):
+ try:
+ return git(root, 'rev-parse', '--symbolic-full-name', '@{upstream}').strip()
+ except ValueError:
+ return 'HEAD'
+
+
+def file_path(root, name, revision):
+ # Resolve the parent so that a tracked symlink is treated as a file, even
+ # when its target is outside the repository.
+ path = Path(name).absolute()
+ path = path.parent.resolve() / path.name
+ try:
+ relative = path.relative_to(root).as_posix()
+ except ValueError:
+ raise ValueError(f'{name}: path is outside the repository') from None
+
+ if not path.exists() and not git(root, 'ls-tree', revision, '--', relative):
+ raise ValueError(f'{name}: no such file in the worktree or history tip')
+ return relative or '.'
+
+
+def patch_paths(root, data):
+ """Let Git parse filenames, including quoted names, renames and deletions."""
+ paths = set()
+ # Forward statistics name a rename's destination; reverse statistics name
+ # its source. Neither command applies the patch or changes the worktree.
+ for direction in ((), ('--reverse',)):
+ output = git(root, 'apply', '--numstat', '-z', *direction, input=data)
+ for record in output.split('\0'):
+ if not record:
+ continue
+ name = record.split('\t', 2)[2]
+ path = PurePosixPath(name)
+ if path.is_absolute() or '..' in path.parts:
+ raise ValueError(f'{name}: patch path is outside the repository')
+ paths.add(path.as_posix())
+ if not paths:
+ raise ValueError('patch contains no changed paths')
+ return paths
+
+
+@dataclass
+class Commit:
+ oid: str
+ timestamp: int
+ author: str
+ reviewers: set
+ ackers: set
+
+
+def history(root, revision, since, path):
+ # Ask Git to parse trailers; prose containing "Reviewed-by" is not a credit.
+ # NUL separators keep the variable-length fields unambiguous.
+ fmt = ('%H%x00%ct%x00%an <%ae>%x00'
+ '%(trailers:key=Reviewed-by,valueonly,unfold)%x00'
+ '%(trailers:key=Acked-by,valueonly,unfold)')
+ output = git(root, 'log', '--no-merges', '--no-notes',
+ '--no-show-signature', '-z', f'--format={fmt}',
+ f'--since-as-filter={since}', revision, '--', path)
+ fields = output.split('\0')[:-1]
+ commits = []
+ for offset in range(0, len(fields), 5):
+ oid, timestamp, author, reviews, acks = fields[offset:offset + 5]
+ commits.append(Commit(oid, int(timestamp), author,
+ set(reviews.splitlines()), set(acks.splitlines())))
+ return commits
+
+
+def identities(root, commits, use_mailmap):
+ """Canonicalize authors and trailer identities together, once per address."""
+ names = set()
+ for commit in commits:
+ names.update({commit.author} | commit.reviewers | commit.ackers)
+ names = sorted(names)
+ if use_mailmap and names:
+ data = ('\n'.join(names) + '\n').encode('utf-8', errors='surrogateescape')
+ canonical = git(root, 'check-mailmap', '--stdin', input=data).splitlines()
+ else:
+ canonical = names
+
+ result = {}
+ for original, mapped in zip(names, canonical):
+ name, address = parseaddr(mapped)
+ if '@' in address:
+ result[original] = (name, address.lower())
+ return result
+
+
+@dataclass
+class Contact:
+ name: str
+ email: str
+ reviews: set = field(default_factory=set)
+ acks: set = field(default_factory=set)
+ authorship: set = field(default_factory=set)
+ latest: int = 0
+
+ def rank(self):
+ # Review experience comes first. Authorship supplies additional contacts
+ # and breaks ties; the date is the matching commit's date, not a review
+ # timestamp. Each commit contributes at most once to the review count.
+ return (-len(self.reviews | self.acks), -len(self.reviews),
+ -len(self.authorship), -self.latest, self.email)
+
+ def description(self, total, stats):
+ roles = []
+ for role, commits in (('reviewer', self.reviews), ('acker', self.acks),
+ ('author', self.authorship)):
+ if not commits:
+ continue
+ if stats:
+ roles.append(f'{role}:{len(commits)}/{total}'
+ f'={len(commits) * 100 / total:.0f}%')
+ else:
+ roles.append(role)
+ if stats:
+ date = datetime.fromtimestamp(self.latest, timezone.utc).date()
+ roles.append(f'last-commit:{date}')
+ return ', '.join(roles)
+
+
+def contacts(commits, names):
+ people = {}
+ for commit in commits:
+ for members, role in (({commit.author}, 'authorship'),
+ (commit.reviewers, 'reviews'),
+ (commit.ackers, 'acks')):
+ for identity in members:
+ if identity not in names:
+ continue
+ name, email = names[identity]
+ person = people.setdefault(email, Contact(name, email))
+ if commit.timestamp > person.latest:
+ person.name = name
+ person.latest = commit.timestamp
+ getattr(person, role).add(commit.oid)
+ return people
+
+
+def select_contacts(people, total, args):
+ selected = []
+ for person in sorted(people.values(), key=Contact.rank):
+ count = len(person.reviews | person.acks | person.authorship)
+ if count < args.git_min_signatures:
+ continue
+ if count * 100 < args.git_min_percent * total:
+ continue
+ selected.append(person.email)
+ if len(selected) == args.git_max_maintainers:
+ break
+ return selected
+
+
+def add_boolean_argument(parser, name, default, help_text):
+ group = parser.add_mutually_exclusive_group()
+ group.add_argument(f'--{name}', dest=name, action='store_true', help=help_text)
+ group.add_argument(f'--no{name}', f'--no-{name}', dest=name,
+ action='store_false', help=f'disable --{name}')
+ parser.set_defaults(**{name: default})
+
+
+def parse_arguments(argv):
+ parser = argparse.ArgumentParser(
+ description='Find maintainers and reviewers from local Git history.',
+ usage='%(prog)s [options] patchfile ...\n'
+ ' %(prog)s [options] -f file-or-directory ...',
+ epilog='Reads a patch from stdin when no patchfile is given; use - for '
+ 'stdin explicitly. History is enabled by default. Contacts are '
+ 'selected using Reviewed-by, Acked-by and authorship.')
+ parser.add_argument('inputs', nargs='*', help='patch files, or paths with -f')
+ parser.add_argument('-f', '--file', action='store_true',
+ help='interpret positional arguments as repository paths')
+ parser.add_argument('--git', action='store_true',
+ help='use Git history (enabled by default)')
+ parser.add_argument('--email', action='store_true',
+ help='print email addresses (enabled by default)')
+ parser.add_argument('--git-since', default='1-year-ago', metavar='DATE',
+ help='Git date limiting history (default: 1-year-ago)')
+ parser.add_argument('--git-revision', metavar='REF',
+ help='history tip (default: branch upstream, or HEAD)')
+ parser.add_argument('--git-min-signatures', type=int, default=1, metavar='N',
+ help='minimum distinct matching commits (default: 1)')
+ parser.add_argument('--git-min-percent', type=float, default=5, metavar='PCT',
+ help='minimum share of matching commits (default: 5)')
+ parser.add_argument('--git-max-maintainers', type=int, default=5, metavar='N',
+ help='maximum suggestions per input path (default: 5)')
+ add_boolean_argument(parser, 'mailmap', True, 'normalize identities with Git mailmap')
+ add_boolean_argument(parser, 'roles', True, 'annotate each address with its roles')
+ add_boolean_argument(parser, 'rolestats', True, 'include commit counts and last commit date')
+ add_boolean_argument(parser, 'n', True, 'include names with email addresses')
+ add_boolean_argument(parser, 'l', True, 'include the libcamera-devel mailing list')
+ args = parser.parse_intermixed_args(argv[1:])
+ if args.file and not args.inputs:
+ parser.error('-f requires at least one file or directory')
+ if args.git_min_signatures < 1 or args.git_max_maintainers < 1:
+ parser.error('commit and contact limits must be positive')
+ if not 0 <= args.git_min_percent <= 100:
+ parser.error('--git-min-percent must be between 0 and 100')
+ if not args.inputs and sys.stdin.isatty():
+ parser.error('provide a patchfile, stdin, or -f file-or-directory')
+ return args
+
+
+def main(argv):
+ args = parse_arguments(argv)
+ try:
+ root = repository()
+ ref = args.git_revision or default_revision(root)
+ revision = git(root, 'rev-parse', '--verify', '--end-of-options',
+ f'{ref}^{{commit}}').strip()
+ if args.file:
+ paths = {file_path(root, name, revision) for name in args.inputs}
+ else:
+ paths = set()
+ for name in args.inputs or ['-']:
+ if name == '-':
+ data = sys.stdin.buffer.read()
+ else:
+ data = Path(name).read_bytes()
+ paths.update(patch_paths(root, data))
+
+ if git(root, 'rev-parse', '--is-shallow-repository').strip() == 'true':
+ print('warning: shallow repository; reviewer history is incomplete',
+ file=sys.stderr)
+
+ scopes = []
+ for path in sorted(paths):
+ commits = history(root, revision, args.git_since, path)
+ if not commits:
+ parent = PurePosixPath(path).parent.as_posix()
+ # An immediate parent can supply contacts for a new file. Never
+ # widen a query to the entire repository implicitly.
+ if parent != '.':
+ print(f'No recent history for {path}; checking {parent}/',
+ file=sys.stderr)
+ commits = history(root, revision, args.git_since, parent)
+ if not commits:
+ print(f'No recent history for {path}', file=sys.stderr)
+ scopes.append(commits)
+
+ all_commits = {commit.oid: commit for scope in scopes for commit in scope}
+ names = identities(root, all_commits.values(), args.mailmap)
+ selected = set()
+ for commits in scopes:
+ selected.update(select_contacts(contacts(commits, names),
+ len(commits), args))
+
+ # Select per path so a busy subsystem cannot crowd out a quiet one.
+ # Merge evidence by commit ID before printing, so overlapping paths
+ # cannot inflate a person's counts.
+ people = contacts(all_commits.values(), names)
+ for person in sorted(people.values(), key=Contact.rank):
+ if person.email not in selected or (args.l and person.email == MAILING_LIST):
+ continue
+ if args.n:
+ address = str(Address(person.name, addr_spec=person.email))
+ else:
+ address = person.email
+ if args.roles or args.rolestats:
+ address += f' ({person.description(len(all_commits), args.rolestats)})'
+ print(address)
+ if args.l:
+ address = MAILING_LIST
+ if args.roles or args.rolestats:
+ address += ' (open list:libcamera-devel)'
+ print(address)
+ return 0
+ except (OSError, ValueError) as error:
+ print(f'{Path(argv[0]).name}: {error}', file=sys.stderr)
+ return 1
+
+
+if __name__ == '__main__':
+ sys.exit(main(sys.argv))
base-commit: 7278a59eba64358bbf3d51f557506dd5b0bce5c4
--
2.56.0