File 0001-dig-initial-b4-dig-implementation.patch of Package b4
From: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
Date: Fri, 10 Oct 2025 16:30:19 -0400
Subject: dig: initial b4 dig implementation
References: dig-support
Git-commit: 16329336c1c8faba853b11238a16249306742505
Patch-mainline: yes
For now, this only does `b4 dig -c [commitish]` and will do its best to
figure out which lore message the commit may have come from. We never
know for sure, so we always list multiple series when the patch exists
in multiple ones and let the end-user figure out which one they want.
Still requires testing and docs.
Signed-off-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
---
src/b4/__init__.py | 39 ++++++++--
src/b4/command.py | 11 +++
src/b4/dig.py | 188 +++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 230 insertions(+), 8 deletions(-)
create mode 100644 src/b4/dig.py
diff --git a/src/b4/__init__.py b/src/b4/__init__.py
index c9ed54c3ca1c..b6eab255103f 100644
--- a/src/b4/__init__.py
+++ b/src/b4/__init__.py
@@ -284,13 +284,25 @@ class LoreMailbox:
self.trailer_map[patchid] = list()
self.trailer_map[patchid] += fmsgs
+
+ def get_latest_revision(self) -> Optional[int]:
+ if not len(self.series):
+ return None
+ # Use the latest revision
+ revs = list(self.series.keys())
+ # sort by date of each series.submission_date
+ revs.sort(key=lambda r: self.series[r].submission_date or 0)
+ return revs[-1]
+
+
def get_series(self, revision: Optional[int] = None, sloppytrailers: bool = False,
reroll: bool = True, codereview_trailers: bool = True) -> Optional['LoreSeries']:
if revision is None:
if not len(self.series):
return None
- # Use the highest revision
- revision = max(self.series.keys())
+ revision = self.get_latest_revision()
+ if revision is None:
+ return None
elif revision not in self.series.keys():
return None
@@ -499,6 +511,7 @@ class LoreSeries:
change_id: Optional[str] = None
prereq_patch_ids: Optional[List[str]] = None
prereq_base_commit: Optional[str] = None
+ _submission_date: Optional[datetime.datetime] = None
def __init__(self, revision: int, expected: int) -> None:
self.revision = revision
@@ -783,8 +796,21 @@ class LoreSeries:
return msgs
+ @property
+ def submission_date(self) -> Optional[datetime.datetime]:
+ # Find the date of the first patch we have
+ if self._submission_date is not None:
+ return self._submission_date
+ for lmsg in self.patches:
+ if lmsg is None:
+ continue
+ self._submission_date = lmsg.date
+ break
+ return self._submission_date
+
def populate_indexes(self):
self.indexes = list()
+
seenfiles = set()
for lmsg in self.patches[1:]:
if lmsg is None or not lmsg.blob_indexes:
@@ -827,12 +853,9 @@ class LoreSeries:
def find_base(self, gitdir: str, branches: Optional[list] = None, maxdays: int = 30) -> Tuple[str, len, len]:
# Find the date of the first patch we have
- pdate = datetime.datetime.now()
- for lmsg in self.patches:
- if lmsg is None:
- continue
- pdate = lmsg.date
- break
+ pdate = self.submission_date
+ if not pdate:
+ pdate = datetime.datetime.now()
# Find the latest commit on that date
guntil = pdate.strftime('%Y-%m-%d')
diff --git a/src/b4/command.py b/src/b4/command.py
index 5d90b4f52b4f..c90705ce9c5e 100644
--- a/src/b4/command.py
+++ b/src/b4/command.py
@@ -118,6 +118,11 @@ def cmd_diff(cmdargs):
b4.diff.main(cmdargs)
+def cmd_dig(cmdargs: argparse.Namespace) -> None:
+ import b4.dig
+ b4.dig.main(cmdargs)
+
+
class ConfigOption(argparse.Action):
"""Action class for storing key=value arguments in a dict."""
def __call__(self, parser, namespace, keyval, option_string=None):
@@ -383,6 +388,12 @@ def setup_parser() -> argparse.ArgumentParser:
help='Submit the token received via verification email')
sp_send.set_defaults(func=cmd_send)
+ # b4 dig
+ sp_dig = subparsers.add_parser('dig', help='Dig into the details of a specific commit')
+ sp_dig.add_argument('-c', '--commit', dest='commit_id', metavar='COMMITISH',
+ help='Commit-ish object to dig into')
+ sp_dig.set_defaults(func=cmd_dig)
+
return parser
diff --git a/src/b4/dig.py b/src/b4/dig.py
new file mode 100644
index 000000000000..8b49b90c8e05
--- /dev/null
+++ b/src/b4/dig.py
@@ -0,0 +1,188 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+# SPDX-License-Identifier: GPL-2.0-or-later
+# Copyright (C) 2025 by the Linux Foundation
+#
+__author__ = 'Konstantin Ryabitsev <konstantin@linuxfoundation.org>'
+
+import os
+import sys
+import b4
+import argparse
+import email.parser
+
+from email.message import EmailMessage
+from typing import List, Set, Optional
+
+logger = b4.logger
+
+# Supported diff algorithms we will try to match
+try_diff_algos: List[str] = [
+ 'myers',
+ 'histogram',
+ 'patience',
+ 'minimal',
+]
+
+
+def dig_commit(cmdargs: argparse.Namespace) -> None:
+ config = b4.get_main_config()
+ cfg_llval = config.get('linkmask', '')
+ if isinstance(cfg_llval, str) and '%s' in cfg_llval:
+ linkmask = cfg_llval
+ else:
+ linkmask = f'{b4.LOREADDR}/%s/'
+ # Are we inside a git repo?
+ topdir = b4.git_get_toplevel()
+ if not topdir:
+ logger.error("Not inside a git repository.")
+ sys.exit(1)
+
+ # Can we resolve this commit to an object?
+ commit = b4.git_revparse_obj(cmdargs.commit_id, topdir)
+ if not commit:
+ logger.error('Cannot find a commit matching %s', cmdargs.commit_id)
+ sys.exit(1)
+
+ logger.info('Digging into commit %s', commit)
+ # Make sure it has exactly one parent (not a merge)
+ ecode, out = b4.git_run_command(
+ topdir, ['show', '--no-patch', '--format=%p', commit],
+ )
+ if ecode > 0:
+ logger.error('Could not get commit info for %s', commit)
+ sys.exit(1)
+ if out.strip().count(' ') != 0:
+ logger.error('Merge commit detected, please specify a single-parent commit.')
+ sys.exit(1)
+
+ # Find commit's author and subject from git
+ ecode, out = b4.git_run_command(
+ topdir, ['show', '--no-patch', '--format=%ae %s', commit],
+ )
+ if ecode > 0:
+ logger.error('Could not get commit info for %s', commit)
+ sys.exit(1)
+ fromeml, csubj = out.strip().split(maxsplit=1)
+ logger.debug('fromeml=%s, csubj=%s', fromeml, csubj)
+ logger.info('Attempting to match by exact patch-id...')
+ showargs = [
+ '--format=email',
+ '--binary',
+ '--encoding=utf-8',
+ '--find-renames',
+ ]
+ # Keep a record so we don't try git-patch-id on identical patches
+ bpatches: Set[bytes] = set()
+ lmbx: Optional[b4.LoreMailbox] = None
+ for algo in try_diff_algos:
+ logger.debug('Trying with diff-algorithm=%s', algo)
+ algoarg = f'--diff-algorithm={algo}'
+ logger.debug('showargs=%s', showargs + [algoarg])
+ ecode, bpatch = b4.git_run_command(
+ topdir, ['show'] + showargs + [algoarg] + [commit],
+ decode=False,
+ )
+ if ecode > 0:
+ logger.error('Could not get a patch out of %s', commit)
+ sys.exit(1)
+ if bpatch in bpatches:
+ logger.debug('Already saw this patch, skipping diff-algorithm=%s', algo)
+ continue
+ bpatches.add(bpatch)
+ gitargs = ['patch-id', '--stable']
+ ecode, out = b4.git_run_command(topdir, gitargs, stdin=bpatch)
+ if ecode > 0 or not len(out.strip()):
+ logger.error('Could not compute patch-id for commit %s', commit)
+ sys.exit(1)
+ patch_id = out.split(maxsplit=1)[0]
+ logger.debug('Patch-id for commit %s is %s', commit, patch_id)
+ logger.info('Trying to find matching series by patch-id %s', patch_id)
+ lmbx = b4.get_series_by_patch_id(patch_id)
+ if lmbx:
+ logger.info('Found matching series by patch-id')
+ break
+
+ if not lmbx:
+ logger.info('Attempting to match by author and subject...')
+ q = '(s:"%s" AND f:"%s")' % (csubj.replace('"', ''), fromeml)
+ msgs = b4.get_pi_search_results(q)
+ if msgs:
+ logger.info('Found %s matching messages', len(msgs))
+ lmbx = b4.LoreMailbox()
+ for msg in msgs:
+ lmbx.add_message(msg)
+ else:
+ logger.error('Could not find anything matching commit %s', commit)
+ # Look at the commit message and find any Link: trailers
+ ecode, out = b4.git_run_command(
+ topdir, ['show', '--no-patch', '--format=%B', commit],
+ )
+ if ecode > 0:
+ logger.error('Could not get commit message for %s', commit)
+ sys.exit(1)
+ trailers, _ = b4.LoreMessage.find_trailers(out)
+ ltrs = [t for t in trailers if t.name.lower() == 'link']
+ if ltrs:
+ logger.info('---')
+ logger.info('Try following these Link trailers:')
+ for ltr in ltrs:
+ logger.info(' %s', ltr.as_string())
+ sys.exit(1)
+
+ # Grab the latest series and see if we have a change_id
+ revs = list(lmbx.series.keys())
+ revs.sort(key=lambda r: lmbx.series[r].submission_date or 0)
+
+ change_id: Optional[str] = None
+ lser = lmbx.get_series(codereview_trailers=False)
+ for rev in revs:
+ change_id = lmbx.series[rev].change_id
+ if not change_id:
+ continue
+ logger.info('Backfilling any missing series by change-id')
+ logger.debug('change_id=%s', change_id)
+ # Fill in the rest of the series by change_id
+ q = f'nq:"change-id:{change_id}"'
+ q_msgs = b4.get_pi_search_results(q, full_threads=True)
+ if q_msgs:
+ for q_msg in q_msgs:
+ lmbx.add_message(q_msg)
+ break
+
+ logger.debug('Number of series in the mbox: %d', len(lmbx.series))
+ logger.info('---')
+ logger.info('This patch is present in the following series:')
+ logger.info('---')
+ firstmsg: Optional[b4.LoreMessage] = None
+ for rev in revs:
+ pref = f' v{rev}: '
+ lser = lmbx.series[rev]
+ lmsg: Optional[b4.LoreMessage] = None
+ if lser.has_cover:
+ firstmsg = lser.patches[0]
+ for lmsg in lser.patches[1:]:
+ if lmsg is None:
+ continue
+ if firstmsg is None:
+ firstmsg = lmsg
+ if lmsg.git_patch_id == patch_id:
+ logger.debug('Matched by exact patch-id')
+ break
+ if lmsg.subject == csubj:
+ logger.debug('Matched by subject')
+ break
+
+ if firstmsg is None:
+ logger.error('Internal error: no patches in the series?')
+ sys.exit(1)
+ if lmsg is None:
+ # Use the first patch in the series as a fallback
+ lmsg = firstmsg
+ logger.info('%s%s', pref, lmsg.full_subject)
+ logger.info('%s%s', ' ' * len(pref), linkmask % firstmsg.msgid)
+
+
+def main(cmdargs: argparse.Namespace) -> None:
+ if cmdargs.commit_id:
+ dig_commit(cmdargs)
--
2.51.0
