From a2f81bdad0c4a3cbc2dca4e78424030310219ba4 Mon Sep 17 00:00:00 2001 From: Konstantin Ryabitsev Date: Fri, 15 Jul 2022 16:15:00 -0400 Subject: Initial implementation of b4 submit This is the first rough implementation of "b4 submit". Currently implemented: - b4 submit --new : to start a new branch - b4 submit --edit-cover : to edit the cover message - b4 submit --update-trailers : to receive latest trailer updates from the mailing lists - b4 submit --send : sends the messages using existing git.sendemail configs For details, see "b4 submit --help". Signed-off-by: Konstantin Ryabitsev --- b4/__init__.py | 412 ++++++++++++++++++++++++---- b4/command.py | 52 ++++ b4/pr.py | 148 +++++----- b4/submit.py | 761 +++++++++++++++++++++++++++++++++++++++++++++++++++ b4/ty.py | 96 +------ misc/send-receive.py | 170 ++++++++++++ misc/test.sqlite | Bin 0 -> 24576 bytes patatt | 2 +- requirements.txt | 1 + 9 files changed, 1428 insertions(+), 214 deletions(-) create mode 100644 b4/submit.py create mode 100644 misc/send-receive.py create mode 100644 misc/test.sqlite diff --git a/b4/__init__.py b/b4/__init__.py index c07ee91..b5f2aa6 100644 --- a/b4/__init__.py +++ b/b4/__init__.py @@ -15,6 +15,7 @@ import email.generator import tempfile import pathlib import argparse +import smtplib import urllib.parse import datetime @@ -27,8 +28,9 @@ import pwd import requests +from pathlib import Path from contextlib import contextmanager -from typing import Optional, Tuple, Set, List, TextIO +from typing import Optional, Tuple, Set, List, TextIO, Union from email import charset charset.add_charset('utf-8', None) @@ -92,6 +94,7 @@ LOREADDR = 'https://lore.kernel.org' DEFAULT_CONFIG = { 'midmask': LOREADDR + '/all/%s', 'linkmask': LOREADDR + '/r/%s', + 'searchmask': LOREADDR + '/all/?x=m&t=1&q=%s', 'listid-preference': '*.feeds.kernel.org,*.linux.dev,*.kernel.org,*', 'save-maildirs': 'no', # off: do not bother checking attestation @@ -132,6 +135,8 @@ USER_CONFIG = None REQSESSION = None # Indicates that we've cleaned cache already _CACHE_CLEANED = False +# Used to track mailmap replacements +MAILMAP_INFO = dict() class LoreMailbox: @@ -341,13 +346,15 @@ class LoreMailbox: def add_message(self, msg): msgid = LoreMessage.get_clean_msgid(msg) - if msgid in self.msgid_map: + if msgid and msgid in self.msgid_map: logger.debug('Already have a message with this msgid, skipping %s', msgid) return lmsg = LoreMessage(msg) logger.debug('Looking at: %s', lmsg.full_subject) - self.msgid_map[lmsg.msgid] = lmsg + + if msgid: + self.msgid_map[lmsg.msgid] = lmsg if lmsg.reply: # We'll figure out where this belongs later @@ -479,6 +486,16 @@ class LoreSeries: return slug[:100] + def add_extra_trailers(self, trailers: tuple) -> None: + for lmsg in self.patches[1:]: + if lmsg is None: + continue + lmsg.followup_trailers += trailers + + def add_cover_trailers(self) -> None: + if self.patches[0] and self.patches[0].followup_trailers: # noqa + self.add_extra_trailers(self.patches[0].followup_trailers) # noqa + def get_am_ready(self, noaddtrailers=False, covertrailers=False, addmysob=False, addlink=False, linkmask=None, cherrypick=None, copyccs=False, allowbadchars=False) -> list: @@ -521,6 +538,9 @@ class LoreSeries: logger.debug('Attestation info is not the same') break + if covertrailers: + self.add_cover_trailers() + at = 1 msgs = list() logger.info('---') @@ -535,8 +555,6 @@ class LoreSeries: raise KeyError('Cherrypick not in series') if lmsg is not None: - if self.has_cover and covertrailers and self.patches[0].followup_trailers: # noqa - lmsg.followup_trailers += self.patches[0].followup_trailers # noqa if addlink: lmsg.followup_trailers.append(('Link', linkmask % lmsg.msgid, None, None)) if addmysob: @@ -837,6 +855,7 @@ class LoreMessage: # Body and body-based info self.body = None + self.message = None self.charset = 'utf-8' self.has_diff = False self.has_diffstat = False @@ -1302,6 +1321,14 @@ class LoreMessage: logger.debug('Picked duplicate from preferred source: %s', listid2) return msg2 + @staticmethod + def get_patch_id(diff: str) -> Optional[str]: + gitargs = ['patch-id', '--stable'] + ecode, out = git_run_command(None, gitargs, stdin=diff.encode()) + if ecode > 0: + return None + return out.split(maxsplit=1)[0] + @staticmethod def get_patchwork_hash(diff: str) -> str: """Generate a hash from a diff. Lifted verbatim from patchwork.""" @@ -1362,7 +1389,7 @@ class LoreMessage: return indexes @staticmethod - def find_trailers(body, followup=False): + def find_trailers(body: str, followup: bool = False) -> Tuple[List[Tuple], List[str]]: ignores = {'phone', 'email'} headers = {'subject', 'date', 'from'} nonperson = {'fixes', 'subject', 'date', 'link', 'buglink', 'obsoleted-by'} @@ -1418,7 +1445,9 @@ class LoreMessage: was_trailer = False others.append(line) - return trailers, others + # convert to tuples for ease of matching + ttrailers = [tuple(x) for x in trailers] + return ttrailers, others @staticmethod def get_body_parts(body): @@ -1484,7 +1513,7 @@ class LoreMessage: return githeaders, message, trailers, basement, signature - def fix_trailers(self, copyccs=False): + def fix_trailers(self, copyccs=False, signoff=None): config = get_main_config() attpolicy = config['attestation-policy'] @@ -1512,11 +1541,21 @@ class LoreMessage: trailers.append(('Cc', pair[1], None, None)) # noqa fixtrailers = list() + # If we received a signoff trailer: + # - if it's already present, we move it to the bottom + # - if not already present, we add it + new_signoff = True for trailer in trailers: if list(trailer[:3]) in fixtrailers: # Dupe continue + if signoff and tuple(trailer[:3]) == tuple(signoff): + # Skip it, we'll add it at the bottom + new_signoff = False + logger.debug(' . %s: %s', signoff[0], signoff[1]) + continue + fixtrailers.append(list(trailer[:3])) if trailer[:3] not in btrailers: extra = '' @@ -1537,6 +1576,12 @@ class LoreMessage: else: logger.debug(' . %s: %s', trailer[0], trailer[1]) + if signoff: + # Tack on our signoff at the bottom + fixtrailers.append(list(signoff)) + if new_signoff: + logger.info(' + %s: %s', signoff[0], signoff[1]) + # Reconstitute the message self.body = '' if bheaders: @@ -1545,16 +1590,21 @@ class LoreMessage: self.body += '%s: %s\n' % (bheader[0], bheader[1]) self.body += '\n' + newmessage = '' if len(message): - self.body += message.rstrip('\r\n') + '\n' + newmessage += message.rstrip('\r\n') + '\n' if len(fixtrailers): - self.body += '\n' + newmessage += '\n' if len(fixtrailers): for trailer in fixtrailers: - self.body += '%s: %s\n' % (trailer[0], trailer[1]) + newmessage += '%s: %s\n' % (trailer[0], trailer[1]) if trailer[2]: - self.body += '%s\n' % trailer[2] + newmessage += '%s\n' % trailer[2] + + self.message = self.subject + '\n\n' + newmessage + self.body += newmessage + if len(basement): self.body += '---\n' self.body += basement.rstrip('\r\n') + '\n\n' @@ -1702,9 +1752,9 @@ class LoreSubject: subject = re.sub(r'^\s*\[[^]]*]\s*', '', subject) self.subject = subject - def get_slug(self): - unsafe = '%04d_%s' % (self.counter, self.subject) - return re.sub(r'\W+', '_', unsafe).strip('_').lower() + def get_slug(self, sep='_'): + unsafe = '%04d%s%s' % (self.counter, sep, self.subject) + return re.sub(r'\W+', sep, unsafe).strip(sep).lower() def __repr__(self): out = list() @@ -1871,7 +1921,7 @@ def gpg_run_command(args: List[str], stdin: Optional[bytes] = None) -> Tuple[int def git_run_command(gitdir: Optional[str], args: List[str], stdin: Optional[bytes] = None, - logstderr: bool = False) -> Tuple[int, str]: + logstderr: bool = False, decode: bool = True) -> Tuple[int, Union[str, bytes]]: cmdargs = ['git', '--no-pager'] if gitdir: if os.path.exists(os.path.join(gitdir, '.git')): @@ -1881,10 +1931,12 @@ def git_run_command(gitdir: Optional[str], args: List[str], stdin: Optional[byte ecode, out, err = _run_command(cmdargs, stdin=stdin) - out = out.decode(errors='replace') + if decode: + out = out.decode(errors='replace') if logstderr and len(err.strip()): - err = err.decode(errors='replace') + if decode: + err = err.decode(errors='replace') logger.debug('Stderr: %s', err) out += err @@ -1903,6 +1955,11 @@ def git_get_command_lines(gitdir: Optional[str], args: list) -> List[str]: return lines +def git_get_repo_status(gitdir: Optional[str] = None) -> List[str]: + args = ['status', '--porcelain=v1'] + return git_get_command_lines(gitdir, args) + + @contextmanager def git_temp_worktree(gitdir=None, commitish=None): """Context manager that creates a temporary work tree and chdirs into it. The @@ -2229,32 +2286,48 @@ def mailsplit_bytes(bmbox: bytes, outdir: str) -> list: return msgs -def get_pi_thread_by_url(t_mbx_url, nocache=False): +def get_pi_search_results(query: str, nocache: bool = False): + config = get_main_config() + searchmask = config.get('searchmask') + if not searchmask: + logger.critical('b4.searchmask is not defined') + return None msgs = list() - cachedir = get_cache_file(t_mbx_url, 'pi.msgs') + query = urllib.parse.quote_plus(query) + query_url = searchmask % query + cachedir = get_cache_file(query_url, 'pi.msgs') if os.path.exists(cachedir) and not nocache: logger.debug('Using cached copy: %s', cachedir) for msg in os.listdir(cachedir): with open(os.path.join(cachedir, msg), 'rb') as fh: msgs.append(email.message_from_binary_file(fh)) - else: - logger.critical('Grabbing thread from %s', t_mbx_url.split('://')[1]) - session = get_requests_session() - resp = session.get(t_mbx_url) - if resp.status_code == 404: - logger.critical('That message-id is not known.') - return None - if resp.status_code != 200: - logger.critical('Server returned an error: %s', resp.status_code) - return None - t_mbox = gzip.decompress(resp.content) - resp.close() - if not len(t_mbox): - logger.critical('No messages found for that query') - return None - # Convert into individual files using git-mailsplit - with tempfile.TemporaryDirectory(suffix='-mailsplit') as tfd: - msgs = mailsplit_bytes(t_mbox, tfd) + return msgs + + loc = urllib.parse.urlparse(query_url) + logger.info('Grabbing search results from %s', loc.netloc) + session = get_requests_session() + # For the query to retrieve a mbox file, we need to send a POST request + resp = session.post(query_url, data='') + if resp.status_code == 404: + logger.info('Nothing matching that query.') + return None + if resp.status_code != 200: + logger.info('Server returned an error: %s', resp.status_code) + return None + t_mbox = gzip.decompress(resp.content) + resp.close() + if not len(t_mbox): + logger.critical('No messages found for that query') + return None + + return split_and_dedupe_pi_results(t_mbox, cachedir=cachedir) + + +def split_and_dedupe_pi_results(t_mbox: bytes, cachedir: Optional[str] = None) -> List[email.message.Message]: + # Convert into individual files using git-mailsplit + with tempfile.TemporaryDirectory(suffix='-mailsplit') as tfd: + msgs = mailsplit_bytes(t_mbox, tfd) + if cachedir: if os.path.exists(cachedir): shutil.rmtree(cachedir) shutil.copytree(tfd, cachedir) @@ -2269,6 +2342,34 @@ def get_pi_thread_by_url(t_mbx_url, nocache=False): return list(deduped.values()) +def get_pi_thread_by_url(t_mbx_url: str, nocache: bool = False): + msgs = list() + cachedir = get_cache_file(t_mbx_url, 'pi.msgs') + if os.path.exists(cachedir) and not nocache: + logger.debug('Using cached copy: %s', cachedir) + for msg in os.listdir(cachedir): + with open(os.path.join(cachedir, msg), 'rb') as fh: + msgs.append(email.message_from_binary_file(fh)) + return msgs + + logger.critical('Grabbing thread from %s', t_mbx_url.split('://')[1]) + session = get_requests_session() + resp = session.get(t_mbx_url) + if resp.status_code == 404: + logger.critical('That message-id is not known.') + return None + if resp.status_code != 200: + logger.critical('Server returned an error: %s', resp.status_code) + return None + t_mbox = gzip.decompress(resp.content) + resp.close() + if not len(t_mbox): + logger.critical('No messages found for that query') + return None + + return split_and_dedupe_pi_results(t_mbox, cachedir=cachedir) + + def get_pi_thread_by_msgid(msgid: str, useproject: Optional[str] = None, nocache: bool = False, onlymsgids: Optional[set] = None) -> Optional[list]: qmsgid = urllib.parse.quote_plus(msgid) @@ -2317,21 +2418,96 @@ def get_pi_thread_by_msgid(msgid: str, useproject: Optional[str] = None, nocache return strict -@contextmanager -def git_format_patches(gitdir, start, end, prefixes=None, extraopts=None): - with tempfile.TemporaryDirectory() as tmpd: - gitargs = ['format-patch', '--cover-letter', '-o', tmpd, '--signature', f'b4 {__VERSION__}'] - if prefixes is not None and len(prefixes): - gitargs += ['--subject-prefix', ' '.join(prefixes)] - if extraopts: - gitargs += extraopts - gitargs += ['%s..%s' % (start, end)] - ecode, out = git_run_command(gitdir, gitargs) +def git_range_to_patches(gitdir: Optional[str], start: str, end: str, + covermsg: Optional[email.message.EmailMessage] = None, + prefixes: Optional[List[str]] = None, + msgid_tpt: Optional[str] = None, + seriests: Optional[int] = None, + mailfrom: Optional[Tuple[str, str]] = None, + extrahdrs: Optional[List[Tuple[str, str]]] = None, + keepdate: bool = False) -> List[Tuple[str, email.message.Message]]: + patches = list() + commits = git_get_command_lines(gitdir, ['rev-list', f'{start}..{end}']) + if not commits: + raise RuntimeError(f'Could not run rev-list {start}..{end}') + for commit in commits: + ecode, out = git_run_command(gitdir, ['show', '--format=email', commit], decode=False) if ecode > 0: - logger.critical('ERROR: Could not convert pull request into patches') - logger.critical(out) - yield None - yield tmpd + raise RuntimeError(f'Could not get a patch out of {commit}') + msg = email.message_from_bytes(out) + logger.debug(' %s', msg.get('Subject')) + + patches.append((commit, msg)) + + startfrom = 1 + fullcount = len(patches) + patches.insert(0, (None, covermsg)) + if covermsg: + startfrom = 0 + + # Go through and apply any outstanding fixes + if prefixes: + prefixes = ' ' + ' '.join(prefixes) + else: + prefixes = '' + + for counter in range(startfrom, fullcount+1): + msg = patches[counter][1] + subject = msg.get('Subject') + csubject = re.sub(r'^\[PATCH]\s*', '', subject) + pline = '[PATCH%s %s/%s]' % (prefixes, str(counter).zfill(len(str(fullcount))), fullcount) + msg.replace_header('Subject', f'{pline} {csubject}') + inbodyhdrs = list() + if mailfrom: + # Move the original From and Date into the body + origfrom = msg.get('From') + if origfrom: + origpair = email.utils.parseaddr(origfrom) + if origpair[1] != mailfrom[1]: + msg.replace_header('From', format_addrs([mailfrom])) + inbodyhdrs.append(f'From: {origfrom}') + else: + msg.add_header('From', format_addrs([mailfrom])) + + if seriests: + patchts = seriests + counter + origdate = msg.get('Date') + if origdate: + if keepdate: + inbodyhdrs.append(f'Date: {origdate}') + msg.replace_header('Date', email.utils.formatdate(patchts, localtime=True)) + else: + msg.add_header('Date', email.utils.formatdate(patchts, localtime=True)) + + payload = msg.get_payload() + if inbodyhdrs: + payload = '\n'.join(inbodyhdrs) + '\n\n' + payload + if not payload.find('\n-- \n') > 0: + payload += f'\n-- \nb4 {__VERSION__}\n' + msg.set_payload(payload) + + if extrahdrs is None: + extrahdrs = list() + for hdrname, hdrval in extrahdrs: + try: + msg.replace_header(hdrname, hdrval) + except KeyError: + msg.add_header(hdrname, hdrval) + + if msgid_tpt: + msg.add_header('Message-Id', msgid_tpt % str(counter)) + refto = None + if counter > 0 and covermsg: + # Thread to the cover letter + refto = msgid_tpt % str(0) + if counter > 1 and not covermsg: + # Tread to the first patch + refto = msgid_tpt % str(1) + if refto: + msg.add_header('References', refto) + msg.add_header('In-Reply-To', refto) + + return patches def git_commit_exists(gitdir, commit_id): @@ -2529,8 +2705,8 @@ def read_template(tptfile): return tpt -def get_smtp(identity: Optional[str] = None): - import smtplib +def get_smtp(identity: Optional[str] = None, + dryrun: bool = False) -> Tuple[Union[smtplib.SMTP, smtplib.SMTP_SSL, None], str]: if identity: sconfig = get_config_from_git(rf'sendemail\.{identity}\..*') sectname = f'sendemail.{identity}' @@ -2550,6 +2726,9 @@ def get_smtp(identity: Optional[str] = None): raise smtplib.SMTPException('Invalid smtpport entry in %s' % sectname) encryption = sconfig.get('smtpencryption') + if dryrun: + return None, fromaddr + logger.info('Connecting to %s:%s', server, port) # We only authenticate if we have encryption if encryption: @@ -2647,3 +2826,128 @@ def patchwork_set_state(msgids: List[str], state: str) -> bool: logger.info(' -> %s : %s', state, title) except requests.exceptions.RequestException as ex: logger.debug('Patchwork REST error: %s', ex) + + +def send_smtp(smtp: Union[smtplib.SMTP, smtplib.SMTP_SSL, None], msg: email.message.Message, + fromaddr: str, destaddrs: Optional[Union[Tuple, Set]] = None, + patatt_sign: bool = False, dryrun: bool = False, + maxheaderlen: Optional[int] = None) -> bool: + if not msg.get('X-Mailer'): + msg.add_header('X-Mailer', f'b4 {__VERSION__}') + msg.set_charset('utf-8') + msg.replace_header('Content-Transfer-Encoding', '8bit') + msg.policy = email.policy.EmailPolicy(utf8=True, cte_type='8bit') + # Python's sendmail implementation seems to have some logic problems where 8-bit messages are involved. + # As far as I understand the difference between 8BITMIME (supported by nearly all smtp servers) and + # SMTPUTF8 (supported by very few), SMTPUTF8 is only required when the addresses specified in either + # "MAIL FROM" or "RCPT TO" lines of the _protocol exchange_ themselves have 8bit characters, not + # anything in the From: header of the DATA payload. Python's smtplib seems to always try to encode + # strings as ascii regardless of what was policy was specified. + # Work around this by getting the payload as string and then encoding to bytes ourselves. + if maxheaderlen is None: + if dryrun: + # Make it fit the terminal window, but no wider than 120 minus visual padding + ts = shutil.get_terminal_size((120, 20)) + maxheaderlen = ts.columns - 8 + if maxheaderlen > 112: + maxheaderlen = 112 + else: + # Use a sane-ish default (we don't need to stick to 80, but + # we need to make sure it's shorter than 255) + maxheaderlen = 120 + + emldata = msg.as_string(maxheaderlen=maxheaderlen) + # Force compliant eols + emldata = re.sub(r'\r\n|\n|\r(?!\n)', '\r\n', emldata) + bdata = emldata.encode() + if patatt_sign: + import patatt + # patatt.logger = logger + bdata = patatt.rfc2822_sign(bdata) + if dryrun or smtp is None: + logger.info(' --- DRYRUN: message follows ---') + logger.info(' | ' + bdata.decode().rstrip().replace('\n', '\n | ')) + logger.info(' --- DRYRUN: message ends ---') + return True + if not destaddrs: + alldests = email.utils.getaddresses([str(x) for x in msg.get_all('to', [])]) + alldests += email.utils.getaddresses([str(x) for x in msg.get_all('cc', [])]) + destaddrs = {x[1] for x in alldests} + smtp.sendmail(fromaddr, destaddrs, bdata) + # TODO: properly catch exceptions on sending + return True + + +def git_get_current_branch(gitdir: Optional[str] = None, short: bool = True) -> Optional[str]: + gitargs = ['symbolic-ref', '-q', 'HEAD'] + ecode, out = git_run_command(gitdir, gitargs) + if ecode > 0: + logger.critical('Not able to get current branch (git symbolic-ref HEAD)') + return None + mybranch = out.strip() + if short: + return re.sub(r'^refs/heads/', '', mybranch) + return mybranch + + +def get_excluded_addrs() -> Set[str]: + config = get_main_config() + excludes = set() + c_excludes = config.get('email-exclude') + if c_excludes: + for entry in c_excludes.split(','): + excludes.add(entry.strip()) + + return excludes + + +def cleanup_email_addrs(addresses: List[Tuple[str, str]], excludes: Set[str], + gitdir: Optional[str]) -> List[Tuple[str, str]]: + global MAILMAP_INFO + for entry in list(addresses): + # Check if it's in excludes + removed = False + for exclude in excludes: + if fnmatch.fnmatch(entry[1], exclude): + logger.debug('Removed %s due to matching %s', entry[1], exclude) + addresses.remove(entry) + removed = True + break + if removed: + continue + # Check if it's mailmap-replaced + if entry[1] in MAILMAP_INFO: + if MAILMAP_INFO[entry[1]]: + addresses.remove(entry) + addresses.append(MAILMAP_INFO[entry[1]]) + continue + logger.debug('Checking if %s is mailmap-replaced', entry[1]) + args = ['check-mailmap', f'<{entry[1]}>'] + ecode, out = git_run_command(gitdir, args) + if ecode != 0: + MAILMAP_INFO[entry[1]] = None + continue + replacement = email.utils.getaddresses([out.strip()]) + if len(replacement) == 1: + if entry[1] == replacement[0][1]: + MAILMAP_INFO[entry[1]] = None + continue + logger.debug('Replaced %s with mailmap-updated %s', entry[1], replacement[0][1]) + MAILMAP_INFO[entry[1]] = replacement[0] + addresses.remove(entry) + addresses.append(replacement[0]) + + return addresses + + +def get_email_signature() -> str: + usercfg = get_user_config() + # Do we have a .signature file? + sigfile = os.path.join(str(Path.home()), '.signature') + if os.path.exists(sigfile): + with open(sigfile, 'r', encoding='utf-8') as fh: + signature = fh.read() + else: + signature = '%s <%s>' % (usercfg['name'], usercfg['email']) + + return signature diff --git a/b4/command.py b/b4/command.py index 6a29f0a..49ce767 100644 --- a/b4/command.py +++ b/b4/command.py @@ -71,6 +71,11 @@ def cmd_kr(cmdargs): b4.kr.main(cmdargs) +def cmd_submit(cmdargs): + import b4.submit + b4.submit.main(cmdargs) + + def cmd_am(cmdargs): import b4.mbox b4.mbox.main(cmdargs) @@ -235,6 +240,53 @@ def cmd(): help='Show all developer keys found in a thread') sp_kr.set_defaults(func=cmd_kr) + # b4 submit + sp_submit = subparsers.add_parser('submit', help='Submit patches for review') + # xg_submit = sp_submit.add_mutually_exclusive_group() + # xg_submit.add_argument('--web-auth-new', action='store_true', default=False, + # help='Register a new email and pubkey with a web submission endpoint') + # xg_submit.add_argument('--web-auth-verify', + # help='Submit a response to a challenge received from a web submission endpoint') + sp_submit.add_argument('--edit-cover', action='store_true', default=False, + help='Edit the cover letter in your defined $EDITOR (or core.editor)') + sp_submit.add_argument('--reroll', action='store_true', default=False, + help='Increment revision and add changelog templates to the cover letter') + nn_submit = sp_submit.add_argument_group('New series', 'Set up a new work branch for a new patch series') + nn_submit.add_argument('-n', '--new', dest='new_series_name', + help='Create a new branch and start working on new series') + nn_submit.add_argument('-f', '--fork-point', dest='fork_point', + help='Create new branch at this fork point instead of HEAD') + ag_submit = sp_submit.add_argument_group('Sync trailers', 'Update series with latest received trailers') + ag_submit.add_argument('-u', '--update-trailers', action='store_true', default=False, + help='Update commits with latest received trailers') + ag_submit.add_argument('-s', '--signoff', action='store_true', default=False, + help='Add my Signed-off-by trailer, if not already present') + ag_submit.add_argument('-S', '--sloppy-trailers', dest='sloppytrailers', action='store_true', default=False, + help='Apply trailers without email address match checking') + ag_submit.add_argument('-F', '--trailers-from', dest='thread_msgid', + help='Look for new trailers in the thread with this msgid instead of using the change-id') + se_submit = sp_submit.add_argument_group('Send series', 'Submits your series for review') + se_submit.add_argument('--send', action='store_true', default=False, + help='Submit the series for review') + se_submit.add_argument('-d', '--dry-run', dest='dryrun', action='store_true', default=False, + help='Do not actually send, just dump out raw smtp messages to the stdout') + se_submit.add_argument('-o', '--output-dir', + help='Do not send, just write patches into this directory (git-format-patch mode)') + se_submit.add_argument('--prefixes', nargs='+', choices=['RFC', 'WIP', 'RESEND'], + help='Prefixes to add to PATCH (e.g. RFC, WIP, RESEND)') + se_submit.add_argument('--no-auto-to-cc', action='store_true', default=False, + help='Do not automatically collect To: and Cc: addresses') + se_submit.add_argument('--to', nargs='+', + help='Addresses to add to the automatically collected To: list') + se_submit.add_argument('--cc', nargs='+', + help='Addresses to add to the automatically collected Cc: list') + se_submit.add_argument('--not-me-too', action='store_true', default=False, + help='Remove yourself from the To: or Cc: list') + se_submit.add_argument('--no-sign', action='store_true', default=False, + help='Do not cryptographically sign your patches with patatt') + + sp_submit.set_defaults(func=cmd_submit) + cmdargs = parser.parse_args() logger.setLevel(logging.DEBUG) diff --git a/b4/pr.py b/b4/pr.py index 825391e..596d3ad 100644 --- a/b4/pr.py +++ b/b4/pr.py @@ -318,84 +318,78 @@ def explode(gitdir, lmsg, mailfrom=None, retrieve_links=True, fpopts=None): # of the archived threads. linked_ids.add(lmsg.msgid) - with b4.git_format_patches(gitdir, lmsg.pr_base_commit, 'FETCH_HEAD', prefixes=prefixes, extraopts=fpopts) as pdir: - if pdir is None: - raise RuntimeError('Could not run format-patches') - - for msgfile in sorted(os.listdir(pdir)): - with open(os.path.join(pdir, msgfile), 'rb') as fh: - msg = email.message_from_binary_file(fh) - - msubj = b4.LoreSubject(msg.get('subject', '')) - - # Is this the cover letter? - if msubj.counter == 0: - # We rebuild the message from scratch - # The cover letter body is the pull request body, plus a few trailers - body = '%s\n\nbase-commit: %s\nPR-Link: %s\n' % ( - lmsg.body.strip(), lmsg.pr_base_commit, config['linkmask'] % lmsg.msgid) - - # Make it a multipart if we're doing retrieve_links - if retrieve_links: - cmsg = MIMEMultipart() - cmsg.attach(MIMEText(body, 'plain')) - else: - cmsg = email.message.EmailMessage() - cmsg.set_payload(body) - - cmsg.add_header('From', mailfrom) - cmsg.add_header('Subject', '[' + ' '.join(msubj.prefixes) + '] ' + lmsg.subject) - cmsg.add_header('Date', lmsg.msg.get('Date')) - cmsg.set_charset('utf-8') - cmsg.replace_header('Content-Transfer-Encoding', '8bit') - - msg = cmsg - + # FIXME: This is currently broken due to changes to git_range_to_patches + msgs = b4.git_range_to_patches(gitdir, lmsg.pr_base_commit, 'FETCH_HEAD', with_cover=True, + prefixes=prefixes, extraopts=fpopts) + for msg in msgs: + msubj = b4.LoreSubject(msg.get('subject', '')) + + # Is this the cover letter? + if msubj.counter == 0: + # We rebuild the message from scratch + # The cover letter body is the pull request body, plus a few trailers + body = '%s\n\nbase-commit: %s\nPR-Link: %s\n' % ( + lmsg.body.strip(), lmsg.pr_base_commit, config['linkmask'] % lmsg.msgid) + + # Make it a multipart if we're doing retrieve_links + if retrieve_links: + cmsg = MIMEMultipart() + cmsg.attach(MIMEText(body, 'plain')) else: - # Move the original From and Date into the body - prepend = list() - if msg.get('From') != mailfrom: - cleanfrom = b4.LoreMessage.clean_header(msg['from']) - prepend.append('From: %s' % ''.join(cleanfrom)) - msg.replace_header('From', mailfrom) - - prepend.append('Date: %s' % msg['date']) - body = '%s\n\n%s' % ('\n'.join(prepend), msg.get_payload(decode=True).decode('utf-8')) - msg.set_payload(body) - msg.replace_header('Subject', msubj.full_subject) - - if retrieve_links: - matches = re.findall(r'^Link:\s+https?://.*/(\S+@\S+)[^/]', body, flags=re.M | re.I) - if matches: - linked_ids.update(matches) - matches = re.findall(r'^Message-ID:\s+(\S+@\S+)', body, flags=re.M | re.I) - if matches: - linked_ids.update(matches) - - # Add a number of seconds equalling the counter, in hopes it gets properly threaded - newdate = lmsg.date + timedelta(seconds=msubj.counter) - msg.replace_header('Date', utils.format_datetime(newdate)) - - # Thread it to the cover letter - msg.add_header('In-Reply-To', '' % lmsg.msgid) - msg.add_header('References', '' % lmsg.msgid) - - msg.add_header('To', format_addrs(allto)) - if allcc: - msg.add_header('Cc', format_addrs(allcc)) - - # Set the message-id based on the original pull request msgid - msg.add_header('Message-Id', '' % (msubj.counter, lmsg.msgid)) - - if mailfrom != lmsg.msg.get('From'): - msg.add_header('Reply-To', lmsg.msg.get('From')) - msg.add_header('X-Original-From', lmsg.msg.get('From')) - - if lmsg.msg['List-Id']: - msg.add_header('X-Original-List-Id', b4.LoreMessage.clean_header(lmsg.msg['List-Id'])) - logger.info(' %s', msg.get('Subject')) - msg.set_charset('utf-8') - msgs.append(msg) + cmsg = email.message.EmailMessage() + cmsg.set_payload(body) + + cmsg.add_header('From', mailfrom) + cmsg.add_header('Subject', '[' + ' '.join(msubj.prefixes) + '] ' + lmsg.subject) + cmsg.add_header('Date', lmsg.msg.get('Date')) + cmsg.set_charset('utf-8') + cmsg.replace_header('Content-Transfer-Encoding', '8bit') + + msg = cmsg + + else: + # Move the original From and Date into the body + prepend = list() + if msg.get('From') != mailfrom: + cleanfrom = b4.LoreMessage.clean_header(msg['from']) + prepend.append('From: %s' % ''.join(cleanfrom)) + msg.replace_header('From', mailfrom) + + prepend.append('Date: %s' % msg['date']) + body = '%s\n\n%s' % ('\n'.join(prepend), msg.get_payload(decode=True).decode('utf-8')) + msg.set_payload(body) + msg.replace_header('Subject', msubj.full_subject) + + if retrieve_links: + matches = re.findall(r'^Link:\s+https?://.*/(\S+@\S+)[^/]', body, flags=re.M | re.I) + if matches: + linked_ids.update(matches) + matches = re.findall(r'^Message-ID:\s+(\S+@\S+)', body, flags=re.M | re.I) + if matches: + linked_ids.update(matches) + + # Add a number of seconds equalling the counter, in hopes it gets properly threaded + newdate = lmsg.date + timedelta(seconds=msubj.counter) + msg.replace_header('Date', utils.format_datetime(newdate)) + + # Thread it to the cover letter + msg.add_header('In-Reply-To', '' % lmsg.msgid) + msg.add_header('References', '' % lmsg.msgid) + + msg.add_header('To', format_addrs(allto)) + if allcc: + msg.add_header('Cc', format_addrs(allcc)) + + # Set the message-id based on the original pull request msgid + msg.add_header('Message-Id', '' % (msubj.counter, lmsg.msgid)) + + if mailfrom != lmsg.msg.get('From'): + msg.add_header('Reply-To', lmsg.msg.get('From')) + msg.add_header('X-Original-From', lmsg.msg.get('From')) + + if lmsg.msg['List-Id']: + msg.add_header('X-Original-List-Id', b4.LoreMessage.clean_header(lmsg.msg['List-Id'])) + logger.info(' %s', msg.get('Subject')) logger.info('Exploded %s messages', len(msgs)) if retrieve_links and linked_ids: diff --git a/b4/submit.py b/b4/submit.py new file mode 100644 index 0000000..b82b022 --- /dev/null +++ b/b4/submit.py @@ -0,0 +1,761 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- +# SPDX-License-Identifier: GPL-2.0-or-later +# Copyright (C) 2020 by the Linux Foundation +# +__author__ = 'Konstantin Ryabitsev ' + +import email.message +import os +import sys +import b4 +import re +import argparse +import uuid +import time +import datetime +import json +import tempfile +import subprocess +import shlex +import email +import pathlib + +# from nacl.signing import SigningKey +# from nacl.encoding import Base64Encoder + +from typing import Optional, Tuple, List +from email import utils +from string import Template + +try: + import patatt + can_patatt = True +except ModuleNotFoundError: + can_patatt = False + +try: + import git_filter_repo as fr # noqa + can_gfr = True +except ModuleNotFoundError: + can_gfr = False + +logger = b4.logger + +MAGIC_MARKER = '--- b4-submit-tracking ---' + +DEFAULT_COVER_TEMPLATE = """ +${cover} + +--- +${shortlog} + +${diffstat} +--- +base-commit: ${base_commit} +change-id: ${change_id} + +Best regards, +-- +${signature} +""" + +DEFAULT_CHANGELOG_TEMPLATE = """ +Changes in v${newrev}: +- EDITME: describe what is new in this series revision. +- EDITME: use bulletpoints and terse descriptions. +- Link to v${oldrev}: ${oldrev_link} + +""" + +# def auth_new(cmdargs: argparse.Namespace) -> None: +# # Check if we have a patatt signingkey already defined +# endpoint, name, email, ptskey = get_configs() +# skey, pkey = get_patatt_ed25519keys(ptskey) +# logger.info('Will submit a new email authorization request to:') +# logger.info(' Endpoint: %s', endpoint) +# logger.info(' Name: %s', name) +# logger.info(' Email: %s', email) +# logger.info(' Key: %s (%s)', pkey, ptskey) +# logger.info('---') +# confirm = input('Confirm selection [y/N]: ') +# if confirm != 'y': +# logger.info('Exiting') +# sys.exit(0) +# req = { +# 'action': 'auth-new', +# 'name': name, +# 'email': email, +# 'key': pkey, +# } +# ses = b4.get_requests_session() +# res = ses.post(endpoint, json=req) +# logger.info('---') +# if res.status_code == 200: +# try: +# rdata = res.json() +# if rdata.get('result') == 'success': +# logger.info('Challenge generated and sent to %s', email) +# logger.info('Once you receive it, run b4 submit --web-auth-verify [challenge-string]') +# sys.exit(0) +# +# except Exception as ex: # noqa +# logger.critical('Odd response from the endpoint: %s', res.text) +# sys.exit(1) +# +# logger.critical('500 response from the endpoint: %s', res.text) +# sys.exit(1) +# +# +# def auth_verify(cmdargs: argparse.Namespace) -> None: +# endpoint, name, email, ptskey = get_configs() +# skey, pkey = get_patatt_ed25519keys(ptskey) +# challenge = cmdargs.auth_verify +# logger.info('Signing challenge using key %s', ptskey) +# sk = SigningKey(skey.encode(), encoder=Base64Encoder) +# bdata = sk.sign(challenge.encode(), encoder=Base64Encoder) +# req = { +# 'action': 'auth-verify', +# 'name': name, +# 'email': email, +# 'challenge': challenge, +# 'sigdata': bdata.decode(), +# } +# ses = b4.get_requests_session() +# res = ses.post(endpoint, json=req) +# logger.info('---') +# if res.status_code == 200: +# try: +# rdata = res.json() +# if rdata.get('result') == 'success': +# logger.info('Challenge successfully verified for %s', email) +# logger.info('You may now use this endpoint for submitting patches.') +# sys.exit(0) +# +# except Exception as ex: # noqa +# logger.critical('Odd response from the endpoint: %s', res.text) +# sys.exit(1) +# +# logger.critical('500 response from the endpoint: %s', res.text) +# sys.exit(1) + + +def start_new_series(cmdargs: argparse.Namespace) -> None: + status = b4.git_get_repo_status() + if len(status): + logger.critical('CRITICAL: Repository contains uncommitted changes.') + logger.critical(' Stash or commit them first.') + sys.exit(1) + + usercfg = b4.get_user_config() + if 'name' not in usercfg or 'email' not in usercfg: + logger.critical('CRITICAL: Unable to add your Signed-off-by: git returned no user.name or user.email') + sys.exit(1) + + if not cmdargs.fork_point: + cmdargs.fork_point = 'HEAD' + slug = re.sub(r'\W+', '-', cmdargs.new_series_name).strip('-').lower() + branchname = 'b4/%s' % slug + args = ['checkout', '-b', branchname, cmdargs.fork_point] + ecode, out = b4.git_run_command(None, args, logstderr=True) + if ecode > 0: + logger.critical('CRITICAL: Failed to create a new branch %s', branchname) + logger.critical(out) + sys.exit(ecode) + logger.info('Created new branch %s', branchname) + # create an empty commit containing basic cover letter details + msgdata = ('EDITME: cover title for %s' % cmdargs.new_series_name, + '', + '# Lines starting with # will be removed from the cover letter. You can use', + '# them to add notes or reminders to yourself.', + '', + 'EDITME: describe the purpose of this series. The information you put here', + 'will be used by the project maintainer to make a decision whether your', + 'patches should be reviewed, and in what priority order. Please be very', + 'detailed and link to any relevant discussions or sites that the maintainer', + 'can review to better understand your proposed changes.', + '', + 'Signed-off-by: %s <%s>' % (usercfg.get('name', ''), usercfg.get('email', '')), + '', + '# You can add other trailers to the cover letter. Any email addresses found in', + '# these trailers will be added to the addresses specified/generated during', + '# the --send stage.', + '', + '', + ) + # We don't need all the entropy of uuid, just some of it + changeid = '%s-%s-%s' % (datetime.date.today().strftime('%Y%m%d'), slug, uuid.uuid4().hex[:12]) + tracking = { + 'series': { + 'revision': 1, + 'change-id': changeid, + }, + } + message = '\n'.join(msgdata) + make_magic_json(tracking) + args = ['commit', '--allow-empty', '-F', '-'] + ecode, out = b4.git_run_command(None, args, stdin=message.encode(), logstderr=True) + if ecode > 0: + logger.critical('CRITICAL: Generating cover letter commit failed:') + logger.critical(out) + logger.info('Created empty commit with the cover letter.') + logger.info('You can prepare your commits now.') + + +def make_magic_json(data: dict) -> str: + mj = (f'{MAGIC_MARKER}\n' + '# This section is used internally by b4 submit for tracking purposes.\n') + return mj + json.dumps(data, indent=2) + + +def load_cover(cover_commit: str, strip_comments: bool = False) -> Tuple[str, dict]: + # Grab the cover contents + gitargs = ['show', '-s', '--format=%B', cover_commit] + ecode, out = b4.git_run_command(None, gitargs) + if ecode > 0: + logger.critical('CRITICAL: unable to load cover letter') + sys.exit(1) + # Split on MAGIC_MARKER + cover, magic_json = out.split(MAGIC_MARKER) + # drop everything until the first { + junk, mdata = magic_json.split('{', maxsplit=1) + jdata = json.loads('{' + mdata) + logger.debug('tracking data: %s', jdata) + if strip_comments: + cover = re.sub(r'^#.*$', '', cover, flags=re.M) + while '\n\n\n' in cover: + cover = cover.replace('\n\n\n', '\n\n') + return cover.strip(), jdata + + +def update_cover(commit: str, content: str, tracking: dict) -> None: + cover_message = content + '\n\n' + make_magic_json(tracking) + fred = FRCommitMessageEditor() + fred.add(commit, cover_message) + args = fr.FilteringOptions.parse_args(['--force', '--quiet', '--refs', f'{commit}~1..HEAD']) + args.refs = [f'{commit}~1..HEAD'] + frf = fr.RepoFilter(args, commit_callback=fred.callback) + logger.info('Invoking git-filter-repo to update the cover letter.') + frf.run() + + +def check_our_branch() -> bool: + mybranch = b4.git_get_current_branch() + if mybranch.startswith('b4/'): + return True + logger.info('CRITICAL: This does not look like a b4-managed branch.') + logger.info(' "%s" does not start with "b4/"', mybranch) + return False + + +def find_cover_commit() -> Optional[str]: + # Walk back commits until we find the cover letter + # Our covers always contain the MAGIC_MARKER line + logger.debug('Looking for the cover letter commit with magic marker "%s"', MAGIC_MARKER) + gitargs = ['log', '--grep', MAGIC_MARKER, '-F', '--pretty=oneline', '--max-count=1'] + lines = b4.git_get_command_lines(None, gitargs) + if not lines: + return None + found = lines[0].split()[0] + logger.debug('Cover commit found in %s', found) + return found + + +class FRCommitMessageEditor: + edit_map: dict + + def __init__(self, edit_map: Optional[dict] = None): + if edit_map: + self.edit_map = edit_map + else: + self.edit_map = dict() + + def add(self, commit: str, message: str): + self.edit_map[commit.encode()] = message.encode() + + def callback(self, commit, metadata): # noqa + if commit.original_id in self.edit_map: + commit.message = self.edit_map[commit.original_id] + + +def edit_cover(cover_commit: str) -> None: + cover, tracking = load_cover(cover_commit) + # What's our editor? And yes, the default is vi, bite me. + corecfg = b4.get_config_from_git(r'core\..*', {'editor': os.environ.get('EDITOR', 'vi')}) + editor = corecfg.get('editor') + logger.debug('editor=%s', editor) + # We give it a suffix .rst in hopes that editors autoload restructured-text rules + with tempfile.NamedTemporaryFile(suffix='.rst') as temp_cover: + temp_cover.write(cover.encode()) + temp_cover.seek(0) + sp = shlex.shlex(editor, posix=True) + sp.whitespace_split = True + cmdargs = list(sp) + [temp_cover.name] + logger.debug('Running %s' % ' '.join(cmdargs)) + sp = subprocess.Popen(cmdargs) + sp.wait() + new_cover = temp_cover.read().decode(errors='replace').strip() + + if new_cover == cover: + logger.info('Cover letter unchanged.') + return + if not len(new_cover.strip()): + logger.info('New cover letter blank, leaving current one unchanged.') + return + + update_cover(cover_commit, new_cover, tracking) + logger.info('Cover letter updated.') + + +def update_trailers(cover_commit: str, cmdargs: argparse.Namespace) -> None: + if cmdargs.signoff: + usercfg = b4.get_user_config() + if 'name' not in usercfg or 'email' not in usercfg: + logger.critical('CRITICAL: Unable to add your Signed-off-by: git returned no user.name or user.email') + sys.exit(1) + signoff = ('Signed-off-by', f"{usercfg['name']} <{usercfg['email']}>", None) + else: + signoff = None + + try: + patches = b4.git_range_to_patches(None, cover_commit, 'HEAD') + except RuntimeError as ex: + logger.critical('CRITICAL: Failed to convert range to patches: %s', ex) + sys.exit(1) + + logger.info('Calculating patch-ids from %s commits', len(patches)-1) + msg_map = dict() + commit_map = dict() + # Ignore the cover letter + for commit, msg in patches[1:]: + body = msg.get_payload() + patchid = b4.LoreMessage.get_patch_id(body) + msg_map[patchid] = msg + commit_map[patchid] = commit + + if cmdargs.thread_msgid: + cmdargs.msgid = cmdargs.thread_msgid + msgid = b4.get_msgid(cmdargs) + logger.info('Retrieving thread matching %s', msgid) + list_msgs = b4.get_pi_thread_by_msgid(msgid, nocache=True) + else: + cover, tracking = load_cover(cover_commit, strip_comments=True) + changeid = tracking['series'].get('change-id') + logger.info('Checking change-id "%s"', changeid) + query = f'"change-id: {changeid}"' + list_msgs = b4.get_pi_search_results(query, nocache=True) + + bbox = b4.LoreMailbox() + for list_msg in list_msgs: + bbox.add_message(list_msg) + + updates = dict() + lser = bbox.get_series(sloppytrailers=cmdargs.sloppytrailers) + mismatches = list(lser.trailer_mismatches) + for lmsg in lser.patches[1:]: + addtrailers = list(lmsg.followup_trailers) + if lser.has_cover and len(lser.patches[0].followup_trailers): + addtrailers += list(lser.patches[0].followup_trailers) + if not addtrailers: + logger.debug('No follow-up trailers received to the %s', lmsg.subject) + continue + patchid = b4.LoreMessage.get_patch_id(lmsg.body) + if patchid not in commit_map: + logger.debug('No match for patchid %s', patchid) + continue + parts = b4.LoreMessage.get_body_parts(msg_map[patchid].get_payload()) + if signoff and signoff not in parts[2]: + updates[patchid] = list() + for ftrailer in addtrailers: + if ftrailer[:3] not in parts[2]: + if patchid not in updates: + updates[patchid] = list() + updates[patchid].append(ftrailer) + # Check if we've applied mismatched trailers already + if not cmdargs.sloppytrailers and mismatches: + for mtrailer in list(mismatches): + check = (mtrailer[0], mtrailer[1], None) + if check in parts[2]: + logger.debug('Removing already-applied mismatch %s', check) + mismatches.remove(mtrailer) + + if not updates: + logger.info('No trailer updates found.') + return + + if len(mismatches): + logger.critical('---') + logger.critical('NOTE: some trailers ignored due to from/email mismatches:') + for tname, tvalue, fname, femail in lser.trailer_mismatches: + logger.critical(' ! Trailer: %s: %s', tname, tvalue) + logger.critical(' Msg From: %s <%s>', fname, femail) + logger.critical('NOTE: Rerun with -S to apply them anyway') + + logger.info('---') + # Create the map of new messages + fred = FRCommitMessageEditor() + for patchid, newtrailers in updates.items(): + # Make it a LoreMessage, so we can run attestation on received trailers + cmsg = b4.LoreMessage(msg_map[patchid]) + logger.info(' %s', cmsg.subject) + cmsg.followup_trailers = newtrailers + cmsg.fix_trailers(signoff=signoff) + fred.add(commit_map[patchid], cmsg.message) + logger.info('---') + args = fr.FilteringOptions.parse_args(['--force', '--quiet', '--refs', f'{cover_commit}..HEAD']) + args.refs = [f'{cover_commit}..HEAD'] + frf = fr.RepoFilter(args, commit_callback=fred.callback) + logger.info('Invoking git-filter-repo to update trailers.') + frf.run() + logger.info('Trailers updated.') + + +def get_addresses_from_cmd(cmdargs: List[str], msgbytes: bytes) -> List[Tuple[str, str]]: + ecode, out, err = b4._run_command(cmdargs, stdin=msgbytes) # noqa + if ecode > 0: + logger.critical('CRITICAL: Running %s failed:', ' '.join(cmdargs)) + logger.critical(err.decode()) + raise RuntimeError('Running command failed: %s' % ' '.join(cmdargs)) + addrs = out.strip().decode() + if not addrs: + return list() + return utils.getaddresses(addrs.split('\n')) + + +def get_series_details(cover_commit: str) -> Tuple[str, str, str]: + # Not sure if we can reasonably expect all automation to handle this correctly + # gitargs = ['describe', '--long', f'{cover_commit}~1'] + gitargs = ['rev-parse', f'{cover_commit}~1'] + lines = b4.git_get_command_lines(None, gitargs) + base_commit = lines[0] + gitargs = ['shortlog', f'{cover_commit}..'] + ecode, shortlog = b4.git_run_command(None, gitargs) + gitargs = ['diff', '--stat', f'{cover_commit}..'] + ecode, diffstat = b4.git_run_command(None, gitargs) + return base_commit, shortlog.rstrip(), diffstat.rstrip() + + +def send(cover_commit: str, cmdargs: argparse.Namespace) -> None: + # Check if the cover letter has 'EDITME' in it + cover, tracking = load_cover(cover_commit, strip_comments=True) + if 'EDITME' in cover: + logger.critical('CRITICAL: Looks like the cover letter needs to be edited first.') + logger.info('---') + logger.info(cover) + logger.info('---') + sys.exit(1) + + config = b4.get_main_config() + cover_template = DEFAULT_COVER_TEMPLATE + if config.get('submit-cover-template'): + # Try to load this template instead + try: + cover_template = b4.read_template(config['submit-cover-template']) + except FileNotFoundError: + logger.critical('ERROR: submit-cover-template says to use %s, but it does not exist', + config['submit-cover-template']) + sys.exit(2) + + # Generate the patches and collect all the addresses from trailers + parts = b4.LoreMessage.get_body_parts(cover) + trailers = set() + trailers.update(parts[2]) + + # Put together the cover letter + csubject, cbody = cover.split('\n', maxsplit=1) + base_commit, shortlog, diffstat = get_series_details(cover_commit) + change_id = tracking['series'].get('change-id') + revision = tracking['series'].get('revision') + tptvals = { + 'subject': csubject, + 'cover': cbody.strip(), + 'shortlog': shortlog, + 'diffstat': diffstat, + 'change_id': change_id, + 'base_commit': base_commit, + 'signature': b4.get_email_signature(), + } + body = Template(cover_template.lstrip()).safe_substitute(tptvals) + cmsg = email.message.EmailMessage() + cmsg.set_payload(body) + cmsg.add_header('Subject', csubject) + if cmdargs.prefixes: + prefixes = list(cmdargs.prefixes) + else: + prefixes = list() + + prefixes.append(f'v{revision}') + seriests = int(time.time()) + usercfg = b4.get_user_config() + myemail = usercfg.get('email') + myname = usercfg.get('name') + if myemail: + msgdomain = re.sub(r'^[^@]*@', '', myemail) + else: + # Use the hostname of the system + import platform + msgdomain = platform.node() + chunks = change_id.rsplit('-', maxsplit=1) + stablepart = chunks[0] + # Message-IDs must not be predictable to avoid stuffing attacks + randompart = uuid.uuid4().hex[:12] + msgid_tpt = f'<{stablepart}-v{revision}-%s-{randompart}@{msgdomain}>' + + try: + patches = b4.git_range_to_patches(None, cover_commit, 'HEAD', + covermsg=cmsg, prefixes=prefixes, + msgid_tpt=msgid_tpt, + seriests=seriests, + mailfrom=(myname, myemail)) + except RuntimeError as ex: + logger.critical('CRITICAL: Failed to convert range to patches: %s', ex) + sys.exit(1) + + logger.info('Converted the branch to %s patches', len(patches)-1) + seen = set() + todests = list() + if config.get('submit-to'): + for pair in utils.getaddresses([config.get('submit-to')]): + if pair[1] not in seen: + seen.add(pair[1]) + todests.append(pair) + ccdests = list() + if config.get('submit-cc'): + for pair in utils.getaddresses([config.get('submit-cc')]): + if pair[1] not in seen: + seen.add(pair[1]) + ccdests.append(pair) + excludes = set() + # These override config values + if cmdargs.to: + todests = [('', x) for x in cmdargs.to] + seen.update(set(cmdargs.to)) + if cmdargs.cc: + ccdests = [('', x) for x in cmdargs.cc] + seen.update(set(cmdargs.cc)) + + if not cmdargs.no_auto_to_cc: + logger.info('Populating the To: and Cc: fields with automatically collected addresses') + + # Use a sane tocmd and cccmd for the kernel + # TODO: make it definable in the config + tocmdstr = tocmd = None + cccmdstr = cccmd = None + topdir = b4.git_get_toplevel() + getm = os.path.join(topdir, 'scripts', 'get_maintainer.pl') + if os.access(getm, os.X_OK): + logger.debug('Using kernel get_maintainer.pl for to and cc list') + tocmdstr = f'{getm} --nogit --nogit-fallback --nogit-chief-penguins --norolestats --nol' + cccmdstr = f'{getm} --nogit --nogit-fallback --nogit-chief-penguins --norolestats --nom' + if tocmdstr: + sp = shlex.shlex(tocmdstr, posix=True) + sp.whitespace_split = True + tocmd = list(sp) + if cccmdstr: + sp = shlex.shlex(cccmdstr, posix=True) + sp.whitespace_split = True + cccmd = list(sp) + + seen = set() + # Go through them again to make to/cc headers + for commit, msg in patches: + if not msg: + continue + body = msg.get_payload() + parts = b4.LoreMessage.get_body_parts(body) + trailers.update(parts[2]) + msgbytes = msg.as_bytes() + if tocmd: + for pair in get_addresses_from_cmd(tocmd, msgbytes): + if pair[1] not in seen: + seen.add(pair[1]) + todests.append(pair) + if cccmd: + for pair in get_addresses_from_cmd(cccmd, msgbytes): + if pair[1] not in seen: + seen.add(pair[1]) + ccdests.append(pair) + + # add addresses seen in trailers + for trailer in trailers: + if '@' in trailer[1]: + for pair in utils.getaddresses([trailer[1]]): + if pair[1] not in seen: + seen.add(pair[1]) + ccdests.append(pair) + + excludes = b4.get_excluded_addrs() + if cmdargs.not_me_too: + excludes.add(myemail) + + allto = list() + allcc = list() + alldests = set() + + if todests: + allto = b4.cleanup_email_addrs(todests, excludes, None) + alldests.update(set([x[1] for x in allto])) + if ccdests: + allcc = b4.cleanup_email_addrs(ccdests, excludes, None) + alldests.update(set([x[1] for x in allcc])) + + if not len(allto): + # Move all cc's into the To field if there's nothing in "To" + allto = list(allcc) + allcc = list() + + if cmdargs.output_dir: + pathlib.Path(cmdargs.output_dir).mkdir(parents=True, exist_ok=True) + for commit, msg in patches: + if not msg: + continue + msg.add_header('To', b4.format_addrs(allto)) + if allcc: + msg.add_header('Cc', b4.format_addrs(allcc)) + msg.set_charset('utf-8') + msg.replace_header('Content-Transfer-Encoding', '8bit') + msg.policy = email.policy.EmailPolicy(utf8=True, cte_type='8bit') + subject = msg.get('Subject', '') + ls = b4.LoreSubject(subject) + filen = '%s.patch' % ls.get_slug(sep='-') + with open(os.path.join(cmdargs.output_dir, filen), 'w') as fh: + fh.write(msg.as_string(unixfrom=True, maxheaderlen=80)) + logger.info(' %s', filen) + return + + # And now we go through each message to set addressees and send them off + sign = True + if cmdargs.no_sign or config.get('submit-no-sign', '').lower() in {'yes', 'true', 'y'}: + sign = False + identity = config.get('sendemail-identity') + try: + smtp, fromaddr = b4.get_smtp(identity, dryrun=cmdargs.dryrun) + except Exception as ex: # noqa + logger.critical('Failed to configure the smtp connection:') + logger.critical(ex) + sys.exit(1) + + counter = 0 + cover_msgid = None + # TODO: Need to send obsoleted-by follow-ups, just need to figure out where. + for commit, msg in patches: + if not msg: + continue + if cover_msgid is None: + cover_msgid = b4.LoreMessage.get_clean_msgid(msg) + msg.add_header('To', b4.format_addrs(allto)) + if allcc: + msg.add_header('Cc', b4.format_addrs(allcc)) + logger.info(' %s', msg.get('Subject')) + if b4.send_smtp(smtp, msg, fromaddr=fromaddr, destaddrs=alldests, patatt_sign=sign, + dryrun=cmdargs.dryrun): + counter += 1 + + logger.info('---') + if cmdargs.dryrun: + logger.info('DRYRUN: Would have sent %s messages', counter) + return + else: + logger.info('Sent %s messages', counter) + + if not cover_msgid: + return + + logger.info('Recording series message-id in cover letter tracking') + cover, tracking = load_cover(cover_commit, strip_comments=False) + vrev = f'v{revision}' + if 'history' not in tracking['series']: + tracking['series']['history'] = dict() + if vrev not in tracking['series']['history']: + tracking['series']['history'][vrev] = list() + tracking['series']['history'][vrev].append(cover_msgid) + update_cover(cover_commit, cover, tracking) + + +def reroll(cover_commit: str, cmdargs: argparse.Namespace) -> None: + cover, tracking = load_cover(cover_commit, strip_comments=False) + oldrev = tracking['series']['revision'] + newrev = oldrev + 1 + tracking['series']['revision'] = newrev + sections = cover.split('---\n') + vrev = f'v{oldrev}' + if 'history' in tracking['series'] and vrev in tracking['series']['history']: + # Use the latest link we have + config = b4.get_main_config() + oldrev_link = config.get('linkmask') % tracking['series']['history'][vrev][-1] + else: + oldrev_link = 'EDITME (not found in tracking)' + tptvals = { + 'oldrev': oldrev, + 'newrev': newrev, + 'oldrev_link': oldrev_link, + } + prepend = Template(DEFAULT_CHANGELOG_TEMPLATE.lstrip()).safe_substitute(tptvals) + found = False + new_sections = list() + for section in sections: + if re.search(r'^changes in v\d+', section, flags=re.I | re.M): + # This is our section + new_sections.append(prepend + section) + found = True + else: + new_sections.append(section) + if found: + new_cover = '---\n'.join(new_sections) + else: + new_cover = cover + '\n\n---\n' + prepend + logger.info('Created new revision v%s', newrev) + logger.info('Updating cover letter with templated changelog entries.') + update_cover(cover_commit, new_cover, tracking) + logger.info('You may now edit the cover letter using "b4 submit --edit-cover"') + + +def main(cmdargs: argparse.Namespace) -> None: + if not can_gfr: + logger.critical('ERROR: b4 submit requires git-filter-repo. You should be able') + logger.critical(' to install it from your distro packages, or from pip.') + sys.exit(1) + + config = b4.get_main_config() + if 'submit-endpoint' not in config: + config['submit-endpoint'] = 'https://lkml.kernel.org/_b4_submit' + + if cmdargs.new_series_name: + start_new_series(cmdargs) + + if not check_our_branch(): + return + + cover_commit = find_cover_commit() + if not cover_commit: + logger.critical('CRITICAL: Unable to find cover letter commit') + sys.exit(1) + + if cmdargs.edit_cover: + edit_cover(cover_commit) + return + + elif cmdargs.update_trailers: + update_trailers(cover_commit, cmdargs) + return + + elif cmdargs.send: + send(cover_commit, cmdargs) + return + + elif cmdargs.reroll: + reroll(cover_commit, cmdargs) + return + + logger.critical('No action requested, please see "b4 submit --help"') + sys.exit(1) + + # if not can_patatt: + # logger.critical('ERROR: b4 submit requires patatt library. See:') + # logger.critical(' https://git.kernel.org/pub/scm/utils/patatt/patatt.git/about/') + # sys.exit(1) + + # if cmdargs.web_auth_new: + # auth_new(cmdargs) + # + # if cmdargs.web_auth_verify: + # auth_verify(cmdargs) diff --git a/b4/ty.py b/b4/ty.py index 3c607b0..dae5206 100644 --- a/b4/ty.py +++ b/b4/ty.py @@ -7,13 +7,13 @@ __author__ = 'Konstantin Ryabitsev ' import os import sys + import b4 import re import email import email.message import email.policy import json -import fnmatch from string import Template from email import utils @@ -51,8 +51,6 @@ ${signature} MY_COMMITS = None # Used to track additional branch info BRANCH_INFO = None -# Used to track mailmap replacements -MAILMAP_INFO = dict() def git_get_merge_id(gitdir, commit_id, branch=None): @@ -76,63 +74,21 @@ def git_get_commit_message(gitdir, rev): return b4.git_run_command(gitdir, args) -def fix_dests(addresses, excludes, gitdir): - global MAILMAP_INFO - for entry in list(addresses): - # Check if it's in excludes - removed = False - for exclude in excludes: - if fnmatch.fnmatch(entry[1], exclude): - logger.debug('Removed %s due to matching %s', entry[1], exclude) - addresses.remove(entry) - removed = True - break - if removed: - continue - # Check if it's mailmap-replaced - if entry[1] in MAILMAP_INFO: - if MAILMAP_INFO[entry[1]]: - addresses.remove(entry) - addresses.append(MAILMAP_INFO[entry[1]]) - continue - logger.debug('Checking if %s is mailmap-replaced', entry[1]) - args = ['check-mailmap', f'<{entry[1]}>'] - ecode, out = b4.git_run_command(gitdir, args) - if ecode != 0: - MAILMAP_INFO[entry[1]] = None - continue - replacement = utils.getaddresses([out.strip()]) - if len(replacement) == 1: - if entry == replacement[0]: - MAILMAP_INFO[entry[1]] = None - continue - logger.debug('Replaced %s with mailmap-updated %s', entry[1], replacement[0][1]) - MAILMAP_INFO[entry[1]] = replacement[0] - addresses.remove(entry) - addresses.append(replacement[0]) - - return addresses - - def make_reply(reply_template, jsondata, gitdir): body = Template(reply_template).safe_substitute(jsondata) # Conform to email standards body = body.replace('\n', '\r\n') msg = email.message_from_string(body) msg['From'] = '%s <%s>' % (jsondata['myname'], jsondata['myemail']) - config = b4.get_main_config() - # Remove ourselves and original sender from allto or allcc - excludes = [jsondata['myemail'], jsondata['fromemail']] - c_excludes = config.get('email-exclude') - if c_excludes: - for entry in c_excludes.split(','): - excludes.append(entry.strip()) - - allto = fix_dests(utils.getaddresses([jsondata['to']]), excludes, gitdir) - allcc = fix_dests(utils.getaddresses([jsondata['cc']]), excludes, gitdir) - - # Add original sender to the To - newto = fix_dests([(jsondata['fromname'], jsondata['fromemail'])], excludes[2:], gitdir) + excludes = b4.get_excluded_addrs() + newto = b4.cleanup_email_addrs([(jsondata['fromname'], jsondata['fromemail'])], excludes, gitdir) + + # Exclude ourselves and original sender from allto or allcc + excludes.add(jsondata['myemail']) + excludes.add(jsondata['fromemail']) + allto = b4.cleanup_email_addrs(utils.getaddresses([jsondata['to']]), excludes, gitdir) + allcc = b4.cleanup_email_addrs(utils.getaddresses([jsondata['cc']]), excludes, gitdir) + if newto: allto += newto @@ -428,6 +384,7 @@ def send_messages(listing, branch, cmdargs): gitdir = cmdargs.gitdir datadir = b4.get_data_dir() fromaddr = None + smtp = None if cmdargs.sendemail: # See if we have sendemail-identity set config = b4.get_main_config() @@ -445,13 +402,7 @@ def send_messages(listing, branch, cmdargs): os.mkdir(cmdargs.outdir) usercfg = b4.get_user_config() - # Do we have a .signature file? - sigfile = os.path.join(str(Path.home()), '.signature') - if os.path.exists(sigfile): - with open(sigfile, 'r', encoding='utf-8') as fh: - signature = fh.read() - else: - signature = '%s <%s>' % (usercfg['name'], usercfg['email']) + signature = b4.get_email_signature() outgoing = 0 msgids = list() @@ -477,29 +428,10 @@ def send_messages(listing, branch, cmdargs): msg.set_charset('utf-8') msg.replace_header('Content-Transfer-Encoding', '8bit') if cmdargs.sendemail: - msg.policy = email.policy.EmailPolicy(utf8=True, cte_type='8bit') - emldata = msg.as_string() if not fromaddr: fromaddr = jsondata['myemail'] - if cmdargs.dryrun: - logger.info('--- DRYRUN: message follows ---') - logger.info('\t' + emldata.replace('\n', '\n\t')) - logger.info('--- DRYRUN: message ends ---') - else: - alldests = email.utils.getaddresses([str(x) for x in msg.get_all('to', [])]) - alldests += email.utils.getaddresses([str(x) for x in msg.get_all('cc', [])]) - sendto = {x[1] for x in alldests} - logger.info(' Sending: %s', msg.get('subject')) - # Python's sendmail implementation seems to have some logic problems where 8-bit messages are involved. - # As far as I understand the difference between 8BITMIME (supported by nearly all smtp servers) and - # SMTPUTF8 (supported by very few), SMTPUTF8 is only required when the addresses specified in either - # "MAIL FROM" or "RCPT TO" lines of the _protocol exchange_ themselves have 8bit characters, not - # anything in the From: header of the DATA payload. Python's smtplib seems to always try to encode - # strings as ascii regardless of what was policy was specified. - # Work around this by getting the payload as string and then encoding to bytes ourselves. - # Force compliant eols - emldata = re.sub(r'(?:\r\n|\n|\r(?!\n))', '\r\n', emldata) # noqa - smtp.sendmail(fromaddr, sendto, emldata.encode()) # noqa + logger.info(' Sending: %s', msg.get('subject')) + b4.send_smtp(smtp, msg, fromaddr, dryrun=cmdargs.dryrun) else: slug_from = re.sub(r'\W', '_', jsondata['fromemail']) slug_subj = re.sub(r'\W', '_', jsondata['subject']) diff --git a/misc/send-receive.py b/misc/send-receive.py new file mode 100644 index 0000000..7b47798 --- /dev/null +++ b/misc/send-receive.py @@ -0,0 +1,170 @@ +#!/usr/bin/env python3 +# noinspection PyUnresolvedReferences +import falcon +import os +import logging +import json +import sqlalchemy as sa + +from nacl.signing import VerifyKey +from nacl.encoding import Base64Encoder +from nacl.exceptions import BadSignatureError + +DB_VERSION = 1 + +logger = logging.getLogger('b4-send-receive') + + +# noinspection PyBroadException, PyMethodMayBeStatic +class SendReceiveListener(object): + + def __init__(self, _engine): + self._engine = _engine + # You shouldn't use this in production + if self._engine.driver == 'pysqlite': + self._init_sa_db() + + def _init_sa_db(self): + logger.info('Creating tables') + conn = self._engine.connect() + md = sa.MetaData() + meta = sa.Table('meta', md, + sa.Column('version', sa.Integer()) + ) + auth = sa.Table('auth', md, + sa.Column('auth_id', sa.Integer(), primary_key=True), + sa.Column('created', sa.DateTime(), nullable=False, server_default=sa.sql.func.now()), + sa.Column('email', sa.Text(), nullable=False), + sa.Column('name', sa.Text(), nullable=False), + sa.Column('pubkey', sa.Text(), nullable=False), + ) + sa.Index('idx_email_pubkey', auth.c.pubkey, auth.c.email, unique=True) + challenge = sa.Table('challenge', md, + sa.Column('challenge_id', sa.Integer(), primary_key=True), + sa.Column('created', sa.DateTime(), nullable=False, server_default=sa.sql.func.now()), + sa.Column('pubkey', sa.Text(), nullable=False), + sa.Column('email', sa.Text(), nullable=False), + sa.Column('challenge', sa.Text(), nullable=False), + ) + sa.Index('idx_uniq_challenge', challenge.c.pubkey, challenge.c.email, challenge.c.challenge, unique=True) + md.create_all(self._engine) + q = sa.insert(meta).values(version=DB_VERSION) + conn.execute(q) + + def on_get(self, req, resp): # noqa + resp.status = falcon.HTTP_200 + resp.content_type = falcon.MEDIA_TEXT + resp.text = "We don't serve GETs here\n" + + def send_error(self, resp, message): + resp.status = falcon.HTTP_500 + resp.text = json.dumps({'result': 'error', 'message': message}) + + def send_success(self, resp, message): + resp.status = falcon.HTTP_200 + resp.text = json.dumps({'result': 'success', 'message': message}) + + def auth_new(self, jdata, resp): + # Is it already authorized? + conn = self._engine.connect() + md = sa.MetaData() + t_auth = sa.Table('auth', md, autoload=True, autoload_with=self._engine) + email = jdata.get('email') + pubkey = jdata.get('key') + q = sa.select([t_auth.c.auth_id]).where(t_auth.c.email == email, t_auth.c.pubkey == pubkey) + rp = conn.execute(q) + if len(rp.fetchall()): + self.send_error(resp, message='%s:%s is already authorized' % (email, pubkey)) + return + # delete any existing challenges for this and create a new one + t_challenge = sa.Table('challenge', md, autoload=True, autoload_with=self._engine) + q = sa.delete(t_challenge).where(t_challenge.c.email == email, t_challenge.c.pubkey == pubkey) + conn.execute(q) + # create new challenge + import uuid + cstr = str(uuid.uuid4()) + q = sa.insert(t_challenge).values(pubkey=pubkey, email=email, challenge=cstr) + conn.execute(q) + # TODO: Actual mail sending + logger.info('Challenge: %s', cstr) + self.send_success(resp, message='Challenge generated') + + def auth_verify(self, jdata, resp): + # Do we have a record for this email/challenge? + conn = self._engine.connect() + md = sa.MetaData() + t_challenge = sa.Table('challenge', md, autoload=True, autoload_with=self._engine) + email = jdata.get('email', '') + challenge = jdata.get('challenge', '') + sigdata = jdata.get('sigdata', '') + q = sa.select([t_challenge.c.pubkey]).where(t_challenge.c.email == email, t_challenge.c.challenge == challenge) + rp = conn.execute(q) + qres = rp.fetchall() + if not len(qres): + self.send_error(resp, message='No such challenge for %s' % email) + return + pubkey = qres[0][0] + vk = VerifyKey(pubkey.encode(), encoder=Base64Encoder) + try: + vk.verify(sigdata.encode(), encoder=Base64Encoder) + except BadSignatureError: + self.send_error(resp, message='Could not validate signature for %s' % email) + return + # validated at this point, so record this as valid auth + name = jdata.get('name') + t_auth = sa.Table('auth', md, autoload=True, autoload_with=self._engine) + q = sa.insert(t_auth).values(pubkey=pubkey, name=name, email=email) + conn.execute(q) + q = sa.delete(t_challenge).where(t_challenge.c.email == email, t_challenge.c.challenge == challenge) + conn.execute(q) + self.send_success(resp, message='Challenge verified') + + def on_post(self, req, resp): + if not req.content_length: + resp.status = falcon.HTTP_500 + resp.content_type = falcon.MEDIA_TEXT + resp.text = 'Payload required\n' + return + raw = req.bounded_stream.read() + try: + jdata = json.loads(raw) + except: + resp.status = falcon.HTTP_500 + resp.content_type = falcon.MEDIA_TEXT + resp.text = 'Failed to parse the request\n' + return + logger.info(jdata) + action = jdata.get('action') + if action == 'auth-new': + self.auth_new(jdata, resp) + if action == 'auth-verify': + self.auth_verify(jdata, resp) + else: + resp.status = falcon.HTTP_500 + resp.content_type = falcon.MEDIA_TEXT + resp.text = 'Unknown action: %s\n' % action + return + + +app = falcon.App() +dburl = os.getenv('DB_URL', 'sqlite:///:memory:') +engine = sa.create_engine(dburl) +srl = SendReceiveListener(engine) +mp = os.getenv('MOUNTPOINT', '/_b4_submit') +app.add_route(mp, srl) + + +if __name__ == '__main__': + from wsgiref.simple_server import make_server + logger.setLevel(logging.DEBUG) + ch = logging.StreamHandler() + formatter = logging.Formatter('%(message)s') + ch.setFormatter(formatter) + ch.setLevel(logging.DEBUG) + logger.addHandler(ch) + + with make_server('', 8000, app) as httpd: + logger.info('Serving on port 8000...') + + # Serve until process is killed + httpd.serve_forever() diff --git a/misc/test.sqlite b/misc/test.sqlite new file mode 100644 index 0000000..fc7a99b Binary files /dev/null and b/misc/test.sqlite differ diff --git a/patatt b/patatt index d1279c4..0eb41be 160000 --- a/patatt +++ b/patatt @@ -1 +1 @@ -Subproject commit d1279c4ca7f29b6f98d784c7bf053c3b2ff11578 +Subproject commit 0eb41be65707a1e156a59fd25ea9824c1a9e95ce diff --git a/requirements.txt b/requirements.txt index 166914b..c52cbed 100644 --- a/requirements.txt +++ b/requirements.txt @@ -3,3 +3,4 @@ requests>=2.24,<3.0 dnspython>=2.1,<3.0 dkimpy>=1.0,<2.0 patatt>=0.5,<2.0 +git-filter-repo>=2.30,<3.0 -- cgit v1.2.3