diff --git a/agithub.py b/agithub.py
new file mode 100644
index 0000000..f1f8fbd
--- /dev/null
+++ b/agithub.py
@@ -0,0 +1,202 @@
+from __future__ import annotations
+
+import datetime
+import json
+import weakref
+import asyncio
+import logging
+import time
+from typing import (
+  AsyncGenerator, Tuple, Any, Dict, Optional, List, Union,
+)
+
+from aiohttp.client import ClientResponse
+import aiohttputils
+
+logger = logging.getLogger(__name__)
+
+JsonDict = Dict[str, Any]
+Json = Union[List[JsonDict], JsonDict]
+
+def parse_datetime(s):
+  dt = datetime.datetime.strptime(s, '%Y-%m-%dT%H:%M:%SZ')
+  return dt.replace(tzinfo=datetime.timezone.utc)
+
+class GitHubError(Exception):
+  def __init__(self, message, documentation, code):
+    self.message = message
+    self.documentation = documentation
+    self.code = code
+
+class GitHub(aiohttputils.ClientBase):
+  baseurl = 'https://api.github.com/'
+
+  def __init__(self, token, session=None):
+    self.token = f'token {token}'
+    super().__init__(session = session)
+
+  async def api_request(
+    self, path: str, method: str = 'get',
+    data: Optional[JsonDict] = None, **kwargs,
+  ) -> Tuple[Json, ClientResponse]:
+    h = kwargs.get('headers', None)
+    if not h:
+      h = kwargs['headers'] = {}
+    h.setdefault('Accept', 'application/vnd.github.v3+json')
+    h.setdefault('Authorization', self.token)
+
+    if data:
+      binary_data = json.dumps(data, ensure_ascii=False).encode('utf-8')
+      if method == 'get':
+        method = 'post'
+      h.setdefault('Content-Type', 'application/json')
+      kwargs['data'] = binary_data
+
+    for _ in range(3):
+      res = await self.request(path, method=method, **kwargs)
+      j: JsonDict
+      if res.status == 204:
+        j = {}
+      else:
+        j = await res.json()
+        if 'message' in j:
+          if res.status == 403 and int(res.headers.get('X-RateLimit-Remaining', -1)) == 0:
+            reset = int(res.headers['X-RateLimit-Reset']) - time.time() + 1
+            logger.warn('rate limited; sleeping for %ds: %s', reset, j['message'])
+            await asyncio.sleep(reset)
+            continue
+          raise GitHubError(j['message'], j['documentation_url'], res.status)
+      return j, res
+
+    raise Exception('unreachable')
+
+  async def get_repo_issues(
+    self, repo: str, *, state: str = 'open', labels: str = '',
+  ) -> AsyncGenerator[Issue, None]:
+    params = {'state': state}
+    if labels:
+      params['labels'] = labels
+    j, r = await self.api_request(f'/repos/{repo}/issues', params = params)
+    assert isinstance(j, list)
+
+    for x in j:
+      yield Issue(x, self)
+
+    while 'next' in r.links:
+      url = str(r.links['next']['url'])
+      j, r = await self.api_request(url)
+      assert isinstance(j, list)
+      for x in j:
+        yield Issue(x, self)
+
+  async def get_issue(self, repo: str, issue_nr: int) -> 'Issue':
+    j, _ = await self.api_request(f'/repos/{repo}/issues/{issue_nr}')
+    assert isinstance(j, dict)
+    return Issue(j, self)
+
+  async def get_issue_comments(
+    self, repo: str, issue_nr: int,
+  ) -> AsyncGenerator[Comment, None]:
+    j, r = await self.api_request(f'/repos/{repo}/issues/{issue_nr}/comments')
+
+    assert isinstance(j, list)
+    for x in j:
+      yield Comment(x, self)
+
+    while 'next' in r.links:
+      url = str(r.links['next']['url'])
+      j, r = await self.api_request(url)
+      assert isinstance(j, list)
+      for x in j:
+        yield Comment(x, self)
+
+  async def create_issue(
+    self, repo: str, title: str, body: Optional[str] = None,
+    labels: List[str] = [],
+  ) -> 'Issue':
+    data: JsonDict = {
+      'title': title,
+    }
+    if body:
+      data['body'] = body
+    if labels:
+      data['labels'] = labels
+
+    issue, _ = await self.api_request(f'/repos/{repo}/issues', data = data)
+    assert isinstance(issue, dict)
+    return Issue(issue, self)
+
+  async def find_login_by_email(self, email: str) -> str:
+    j, _ = await self.api_request(f'/search/users?q={email}')
+    assert isinstance(j, dict)
+    try:
+      return j['items'][0]['login']
+    except IndexError:
+      raise LookupError(email)
+
+class Issue:
+  def __init__(self, data: JsonDict, gh: GitHub) -> None:
+    self.gh = weakref.proxy(gh)
+    self._data = data
+    self.body = data['body']
+    self.number = data['number']
+    self.title = data['title']
+    self.labels = [x['name'] for x in data['labels']]
+    self.updated_at = parse_datetime(data['updated_at'])
+    self._api_url = f"{data['repository_url']}/issues/{data['number']}"
+    self.closed = data['state'] == 'closed'
+    self.author = data['user']['login']
+    self.closed_by = data.get('closed_by') and data['closed_by']['login'] or None
+
+  async def comment(self, comment: str) -> JsonDict:
+    j, _ = await self.gh.api_request(f'{self._api_url}/comments', data = {'body': comment})
+    return j
+
+  async def add_labels(self, labels: List[str]) -> JsonDict:
+    j, _ = await self.gh.api_request(f'{self._api_url}/labels', data = labels)
+    return j
+
+  async def assign(self, assignees: List[str]) -> JsonDict:
+    payload = {'assignees': assignees}
+    j, _ = await self.gh.api_request(f'{self._api_url}/assignees', data = payload)
+    return j
+
+  async def close(self) -> None:
+    data, _ = await self.gh.api_request(
+      f'{self._api_url}', method = 'patch', data = {'state': 'closed'})
+    self._data = data
+    self.closed = data['state'] == 'closed'
+
+  async def reopen(self) -> None:
+    data, _ = await self.gh.api_request(
+      f'{self._api_url}', method = 'patch', data = {'state': 'open'})
+    self._data = data
+    self.closed = data['state'] == 'closed'
+
+  def __repr__(self):
+    return f'<Issue {self.number}: {self.title!r}>'
+
+class Comment:
+  def __init__(self, data: JsonDict, gh: GitHub) -> None:
+    self.gh = weakref.proxy(gh)
+    self._update(data)
+
+  def _update(self, data: JsonDict) -> None:
+    self._data = data
+    self.author = data['user']['login']
+    self.html_url = data['html_url']
+    self.url = data['url']
+    self.body = data['body']
+
+  async def delete(self) -> None:
+    await self.gh.api_request(self.url, method = 'DELETE')
+
+  async def edit(self, body: str) -> None:
+    data, _ = await self.gh.api_request(
+      self.url, method = 'PATCH',
+      data = {'body': body},
+    )
+    self._update(data)
+
+  def __repr__(self) -> str:
+    return f'<Comment by {self.author}: {self.html_url}>'
diff --git a/aiohttputils.py b/aiohttputils.py
new file mode 100644
index 0000000..5f9654b
--- /dev/null
+++ b/aiohttputils.py
@@ -0,0 +1,87 @@
+import os
+from http.cookiejar import MozillaCookieJar
+from urllib.parse import urljoin
+from typing import Optional
+import asyncio
+
+import aiohttp
+from aiohttp.client import ClientResponse
+
+class ClientBase:
+  session = None
+  userAgent = None
+  lasturl = None
+  auto_referer = False
+  baseurl: Optional[str] = None
+  cookiefile: Optional[os.PathLike] = None
+  __our_session: bool = False
+
+  def __init__(self, *, baseurl=None, cookiefile=None, session=None):
+    if baseurl is not None:
+      self.baseurl = baseurl
+    self.session = session
+    self.cookiefile = cookiefile
+
+  async def async_init(self) -> None:
+    if not self.session:
+      s = aiohttp.ClientSession()
+      self.__our_session = True
+      self.session = s
+
+      if self.cookiefile:
+        s.cookies = MozillaCookieJar(self.cookiefile)
+        if os.path.exists(self.cookiefile):
+          s.cookies.load() # type: ignore
+
+  def __del__(self):
+    if self.cookiefile:
+      self.session.cookies.save()
+    if self.__our_session:
+      loop = asyncio.get_event_loop()
+      closer = self.session.close()
+      if loop.is_running():
+        asyncio.ensure_future(closer)
+      else:
+        asyncio.run(closer)
+
+  async def request(
+    self, url: str, method: Optional[str] = None, **kwargs,
+  ) -> ClientResponse:
+    if not self.session:
+      await self.async_init()
+
+    if self.baseurl:
+      url = urljoin(self.baseurl, url)
+
+    if self.auto_referer and self.lasturl:
+      h = kwargs.get('headers', None)
+      if not h:
+        h = kwargs['headers'] = {}
+      h.setdefault('Referer', self.lasturl)
+
+    if self.userAgent:
+      h = kwargs.get('headers', None)
+      if not h:
+        h = kwargs['headers'] = {}
+      h.setdefault('User-Agent', self.userAgent)
+
+    if method is None:
+      if 'data' in kwargs:
+        method = 'post'
+      else:
+        method = 'get'
+
+    response = await self.session.request(method, url, **kwargs) # type: ignore
+    # url may have been changed due to redirection
+    self.lasturl = str(response.url)
+    return response
+
+async def test():
+  client = ClientBase(baseurl='https://www.baidu.com/', cookiefile='test')
+  res = await client.request('/')
+  res = await client.request('/404')
+  print(res, client.lasturl)
+
+if __name__ == '__main__':
+  loop = asyncio.get_event_loop()
+  loop.run_until_complete(test())
diff --git a/archpkg.py b/archpkg.py
new file mode 100644
index 0000000..9510e59
--- /dev/null
+++ b/archpkg.py
@@ -0,0 +1,89 @@
+from __future__ import annotations
+
+import os
+from collections import namedtuple
+import subprocess
+import re
+from typing import Tuple, List, Dict
+
+from pkg_resources import parse_version as _parse_version
+
+def parse_arch_version(v: str) -> Tuple[int, Tuple[str, ...]]:
+  if ':' in v:
+    epoch = int(v.split(':', 1)[0])
+  else:
+    epoch = 0
+  return epoch, _parse_version(v)
+
+class PkgNameInfo(namedtuple('PkgNameInfo', 'name, version, release, arch')):
+  def __lt__(self, other) -> bool:
+    if self.name != other.name or self.arch != other.arch:
+      return NotImplemented
+    if self.version != other.version:
+      return parse_arch_version(self.version) < parse_arch_version(other.version)
+    return float(self.release) < float(other.release)
+
+  def __gt__(self, other) -> bool:
+    # No, try the other side please.
+    return NotImplemented
+
+  @property
+  def fullversion(self) -> str:
+    return '%s-%s' % (self.version, self.release)
+
+  @classmethod
+  def parseFilename(cls, filename: str) -> 'PkgNameInfo':
+    return cls(*trimext(filename, 3).rsplit('-', 3))
+
+def trimext(name: str, num: int = 1) -> str:
+  for i in range(num):
+    name = os.path.splitext(name)[0]
+  return name
+
+def get_pkgname_with_bash(PKGBUILD: str) -> List[str]:
+  script = '''\
+. '%s'
+echo ${pkgname[*]}''' % PKGBUILD
+  # Python 3.4 has 'input' arg for check_output
+  p = subprocess.Popen(
+    ['bwrap', '--unshare-all', '--ro-bind', '/', '/', '--tmpfs', '/home',
+     '--tmpfs', '/run', '--die-with-parent',
+     '--tmpfs', '/tmp', '--proc', '/proc', '--dev', '/dev', '/bin/bash'],
+    stdin=subprocess.PIPE, stdout=subprocess.PIPE,
+  )
+  output = p.communicate(script.encode())[0].decode()
+  ret = p.wait()
+  if ret != 0:
+    raise subprocess.CalledProcessError(
+      ret, ['bash'], output)
+  return output.split()
+
+pkgfile_pat = re.compile(r'(?:^|/).+-[^-]+-[\d.]+-(?:\w+)\.pkg\.tar\.(?:xz|zst)$')
+
+def _strip_ver(s: str) -> str:
+  return re.sub(r'[<>=].*', '', s)
+
+def get_package_info(name: str, local: bool = False) -> Dict[str, str]:
+  old_lang = os.environ['LANG']
+  os.environ['LANG'] = 'C'
+  args = '-Qi' if local else '-Si'
+  try:
+    outb = subprocess.check_output(["pacman", args, name])
+    out = outb.decode('latin1')
+  finally:
+    os.environ['LANG'] = old_lang
+
+  ret = {}
+  for l in out.splitlines():
+    if not l:
+      continue
+    if l[0] not in ' \t':
+      key, value = l.split(':', 1)
+      key = key.strip()
+      value = value.strip()
+      ret[key] = value
+    else:
+      ret[key] += ' ' + l.strip()
+  return ret
+
+
diff --git a/htmlutils.py b/htmlutils.py
new file mode 100644
index 0000000..47cc3e3
--- /dev/null
+++ b/htmlutils.py
@@ -0,0 +1,82 @@
+from __future__ import annotations
+
+import re
+import copy
+from html.entities import entitydefs
+
+from lxml import html # type: ignore
+
+def _br2span_inplace(el):
+  for br in el.iterchildren(tag='br'):
+    sp = html.Element('span')
+    sp.text = '\n'
+    sp.tail = br.tail
+    el.replace(br, sp)
+
+def extractText(el):
+  el = copy.copy(el)
+  _br2span_inplace(el)
+  return el.text_content()
+
+def iter_text_and_br(el):
+  if el.text:
+    yield el.text
+  for i in el.iterchildren():
+    if i.tag == 'br':
+      yield '\n'
+    if i.tail:
+      yield i.tail
+
+def un_jsescape(s):
+    '''%xx & %uxxxx -> char, opposite of Javascript's escape()'''
+    return re.sub(
+        r'%u([0-9a-fA-F]{4})|%([0-9a-fA-F]{2})',
+        lambda m: chr(int(m.group(1) or m.group(2), 16)),
+        s
+    )
+
+def entityunescape(string):
+  '''HTML entity decode'''
+  string = re.sub(r'&#[^;]+;', _sharp2uni, string)
+  string = re.sub(r'&[^;]+;', lambda m: entitydefs[m.group(0)[1:-1]], string)
+  return string
+
+def entityunescape_loose(string):
+  '''HTML entity decode. losse version.'''
+  string = re.sub(r'&#[0-9a-fA-F]+[;；]?', _sharp2uni, string)
+  string = re.sub(r'&\w+[;；]?', lambda m: entitydefs[m.group(0)[1:].rstrip(';；')], string)
+  return string
+
+def _sharp2uni(m):
+  '''&#...; ==> unicode'''
+  s = m.group(0)[2:].rstrip(';；')
+  if s.startswith('x'):
+    return chr(int('0'+s, 16))
+  else:
+    return chr(int(s))
+
+def parse_document_from_requests(response, session=None, *, encoding=None):
+  '''
+  ``response``: requests ``Response`` object, or URL
+  ``encoding``: override detected encoding
+  '''
+  if isinstance(response, str):
+    if session is None:
+      raise ValueError('URL given but no session')
+    r = session.get(response)
+  else:
+    r = response
+  if encoding:
+    r.encoding = encoding
+
+  # fromstring handles bytes well
+  # https://stackoverflow.com/a/15305248/296473
+  parser = html.HTMLParser(encoding=encoding or r.encoding)
+  doc = html.fromstring(r.content, base_url=r.url, parser=parser)
+  doc.make_links_absolute()
+
+  return doc
+
+def parse_html_with_encoding(data, encoding='utf-8'):
+  parser = html.HTMLParser(encoding=encoding)
+  return html.fromstring(data, parser=parser)
diff --git a/myutils.py b/myutils.py
new file mode 100644
index 0000000..59f786d
--- /dev/null
+++ b/myutils.py
@@ -0,0 +1,350 @@
+from __future__ import annotations
+
+import os, sys
+import re
+import datetime
+import time
+from functools import lru_cache, wraps
+import logging
+import contextlib
+import signal
+import hashlib
+import base64
+import fcntl
+from typing import Tuple, Union, Optional, Dict, Any, Generator
+
+logger = logging.getLogger(__name__)
+
+def safe_overwrite(fname: str, data: Union[bytes, str], *,
+                   method: str = 'write', mode: str = 'w', encoding: Optional[str] = None) -> None:
+  # FIXME: directory has no read perm
+  # FIXME: symlinks and hard links
+  tmpname = fname + '.tmp'
+  # if not using "with", write can fail without exception
+  with open(tmpname, mode, encoding=encoding) as f:
+    getattr(f, method)(data)
+    # see also: https://thunk.org/tytso/blog/2009/03/15/dont-fear-the-fsync/
+    f.flush()
+    os.fsync(f.fileno())
+  # if the above write failed (because disk is full etc), the old data should be kept
+  os.rename(tmpname, fname)
+
+UNITS = 'KMGTPEZY'
+
+def filesize(size: int) -> str:
+  amt, unit = filesize_ex(size)
+  if unit:
+    return '%.1f%siB' % (amt, unit)
+  else:
+    return '%dB' % amt
+
+def filesize_ex(size: int) -> Tuple[Union[float, int], str]:
+  left: Union[int, float] = abs(size)
+  unit = -1
+  n = len(UNITS)
+  while left > 1100 and unit < n:
+    left = left / 1024
+    unit += 1
+  if unit == -1:
+    return size, ''
+  else:
+    if size < 0:
+      left = -left
+    return left, UNITS[unit]
+
+class FileSize(int):
+  def __str__(self) -> str:
+    return filesize(self).rstrip('iB')
+
+def parse_filesize(s: str) -> int:
+  s1 = s.rstrip('iB')
+  if not s1:
+    raise ValueError(s)
+
+  last = s1[-1]
+  try:
+    idx = UNITS.index(last)
+  except ValueError:
+    return int(float(s1))
+
+  v = float(s1[:-1]) * 1024 ** (idx+1)
+  return int(v)
+
+def humantime(t: int) -> str:
+  '''seconds -> XhYmZs'''
+  if t < 0:
+    sign = '-'
+    t = -t
+  else:
+    sign = ''
+
+  m, s = divmod(t, 60)
+  h, m = divmod(m, 60)
+  d, h = divmod(h, 24)
+  ret = ''
+  if d:
+    ret += '%dd' % d
+  if h:
+    ret += '%dh' % h
+  if m:
+    ret += '%dm' % m
+  if s:
+    ret += '%ds' % s
+  if not ret:
+    ret = '0s'
+  return sign + ret
+
+def dehumantime(s: str) -> int:
+  '''XhYmZs -> seconds'''
+  m = re.match(r'(?:(?P<d>\d+)d)?(?:(?P<h>\d+)h)?(?:(?P<m>\d+)m)?(?:(?P<s>\d+)s)?$', s)
+  if m:
+    return (
+      int(m.group('d') or 0) * 3600 * 24 +
+      int(m.group('h') or 0) * 3600 +
+      int(m.group('m') or 0) * 60 +
+      int(m.group('s') or 0)
+    )
+  else:
+    raise ValueError(s)
+
+def _timed_read(file, timeout):
+  from select import select
+  if select([file], [], [], timeout)[0]:
+    return file.read(1)
+
+def getchar(prompt, hidden=False, end='\n', timeout=None):
+  '''读取一个字符'''
+  import termios
+  sys.stdout.write(prompt)
+  sys.stdout.flush()
+  fd = sys.stdin.fileno()
+  ch: Optional[str]
+
+  def _read() -> Optional[str]:
+    ch: Optional[str]
+    if timeout is None:
+      ch = sys.stdin.read(1)
+    else:
+      ch = _timed_read(sys.stdin, timeout)
+    return ch
+
+  if os.isatty(fd):
+    old = termios.tcgetattr(fd)
+    new = termios.tcgetattr(fd)
+    if hidden:
+      new[3] = new[3] & ~termios.ICANON & ~termios.ECHO
+    else:
+      new[3] = new[3] & ~termios.ICANON
+    new[6][termios.VMIN] = 1
+    new[6][termios.VTIME] = 0
+    try:
+      termios.tcsetattr(fd, termios.TCSANOW, new)
+      termios.tcsendbreak(fd, 0)
+      ch = _read()
+    finally:
+      termios.tcsetattr(fd, termios.TCSAFLUSH, old)
+  else:
+    ch = _read()
+
+  sys.stdout.write(end)
+  return ch
+
+def loadso(fname):
+  '''ctypes.CDLL 的 wrapper，从 sys.path 中搜索文件'''
+  from ctypes import CDLL
+
+  for d in sys.path:
+    p = os.path.join(d, fname)
+    if os.path.exists(p):
+      return CDLL(p)
+  raise ImportError('%s not found' % fname)
+
+def dofile(path):
+  G = {}
+  with open(path) as f:
+    exec(f.read(), G)
+  return G
+
+def restart_if_failed(func, max_tries, args=(), kwargs={}, secs=60, sleep=None):
+  '''
+  re-run when some exception happens, until `max_tries` in `secs`
+  '''
+  import traceback
+  from collections import deque
+
+  dq = deque(maxlen=max_tries)
+  while True:
+    dq.append(time.time())
+    try:
+      return func(*args, **kwargs)
+    except Exception:
+      traceback.print_exc()
+      if len(dq) == max_tries and time.time() - dq[0] < secs:
+        break
+      if sleep is not None:
+        time.sleep(sleep)
+    else:
+      break
+
+def daterange(start, stop=datetime.date.today(), step=datetime.timedelta(days=1)):
+  d = start
+  while d < stop:
+    yield d
+    d += step
+
+@lru_cache()
+def findfont(fontname):
+  from subprocess import check_output
+  out = check_output(['fc-match', '-v', fontname]).decode()
+  for l in out.split('\n'):
+    if l.lstrip().startswith('file:'):
+      return l.split('"', 2)[1]
+
+def debugfunc(logger=logging, *, _id=[0]):
+  def w(func):
+    @wraps(func)
+    def wrapper(*args, **kwargs):
+      myid = _id[0]
+      _id[0] += 1
+      logger.debug('[func %d] %s(%r, %r)', myid, func.__name__, args, kwargs)
+      ret = func(*args, **kwargs)
+      logger.debug('[func %d] return: %r', myid, ret)
+      return ret
+    return wrapper
+  return w
+
+@contextlib.contextmanager
+def execution_timeout(timeout):
+  def timed_out(signum, sigframe):
+    raise TimeoutError
+
+  delay, interval = signal.setitimer(signal.ITIMER_REAL, timeout, 0)
+  old_hdl = signal.signal(signal.SIGALRM, timed_out)
+  now = time.time()
+  try:
+    yield
+  finally:
+    # inner timeout must be smaller, or the timer event will be delayed
+    if delay:
+      elapsed = time.time() - now
+      delay = max(delay - elapsed, 0.000001)
+    else:
+      delay = 0
+    signal.setitimer(signal.ITIMER_REAL, delay, interval)
+    signal.signal(signal.SIGALRM, old_hdl)
+
+def find_executables(name, path=None):
+  '''find all matching executables with specific name in path'''
+  if path is None:
+    path = os.environ['PATH'].split(os.pathsep)
+  elif isinstance(path, str):
+    path = path.split(os.pathsep)
+  path = [p for p in path if os.path.isdir(p)]
+
+  return [os.path.join(p, f) for p in path for f in os.listdir(p) if f == name]
+
+# The following three are learnt from makepkg
+def user_choose(prompt, timeout=None):
+  # XXX: hard-coded term characters are ok?
+  prompt = '\x1b[1;34m::\x1b[1;37m %s\x1b[0m ' % prompt
+  return getchar(prompt, timeout=timeout)
+
+def msg(msg):
+  # XXX: hard-coded term characters are ok?
+  print('\x1b[1;32m==>\x1b[1;37m %s\x1b[0m' % msg)
+
+def msg2(msg):
+  # XXX: hard-coded term characters are ok?
+  print('\x1b[1;34m  ->\x1b[1;37m %s\x1b[0m' % msg)
+
+def is_internal_ip(ip):
+  import ipaddress
+  ip = ipaddress.ip_address(ip)
+  return ip.is_loopback or ip.is_private or ip.is_reserved or ip.is_link_local
+
+@contextlib.contextmanager
+def at_dir(d: os.PathLike) -> Generator[None, None, None]:
+  old_dir = os.getcwd()
+  os.chdir(d)
+  try:
+    yield
+  finally:
+    os.chdir(old_dir)
+
+def firstExistentPath(paths):
+  for p in paths:
+    if os.path.exists(p):
+      return p
+
+def md5sum_of_file(file):
+  with open(file, 'rb') as f:
+    m = hashlib.md5()
+    while True:
+      d = f.read(81920)
+      if not d:
+        break
+      m.update(d)
+  return m.hexdigest()
+
+def md5(s, encoding='utf-8'):
+  m = hashlib.md5()
+  m.update(s.encode(encoding))
+  return m.hexdigest()
+
+def base64_encode(s):
+  if isinstance(s, str):
+    s = s.encode()
+  return base64.b64encode(s).decode('ascii')
+
+def lock_file(path: os.PathLike) -> None:
+  lock = os.open(path, os.O_WRONLY | os.O_CREAT, 0o600)
+  try:
+    fcntl.flock(lock, fcntl.LOCK_EX|fcntl.LOCK_NB)
+  except BlockingIOError:
+    logger.warning('Waiting for lock to release...')
+    fcntl.flock(lock, fcntl.LOCK_EX)
+
+@contextlib.contextmanager
+def file_lock(file: os.PathLike) -> Generator[None, None, None]:
+  lock = os.open(file, os.O_WRONLY | os.O_CREAT, 0o600)
+  try:
+    fcntl.flock(lock, fcntl.LOCK_EX)
+    yield
+  finally:
+    os.close(lock)
+
+def dict_bytes_to_str(d: Dict[Any, Any]) -> Dict[Any, Any]:
+  ret = {}
+  for k, v in d.items():
+    if isinstance(k, bytes):
+      try:
+         k = k.decode()
+      except UnicodeDecodeError:
+        pass
+
+    if isinstance(v, bytes):
+      try:
+         v = v.decode()
+      except UnicodeDecodeError:
+        pass
+    elif isinstance(v, dict):
+      v = dict_bytes_to_str(v)
+    elif isinstance(v, list):
+      try:
+         v = [x.decode() for x in v]
+      except UnicodeDecodeError:
+        pass
+
+    ret[k] = v
+
+  return ret
+
+def xsel(input=None):
+  import subprocess
+
+  if input is None:
+    return subprocess.getoutput('uniclip')
+  else:
+    p = subprocess.Popen(['uniclip', '-i'], stdin=subprocess.PIPE)
+    p.communicate(input.encode())
+    return p.wait()
diff --git a/nicelogger.py b/nicelogger.py
new file mode 100644
index 0000000..326728a
--- /dev/null
+++ b/nicelogger.py
@@ -0,0 +1,95 @@
+'''
+A Tornado-inspired logging formatter, with displayed time with millisecond accuracy
+
+FYI: pyftpdlib also has a Tornado-style logger.
+'''
+
+from __future__ import annotations
+
+import sys
+import time
+import logging
+
+class TornadoLogFormatter(logging.Formatter):
+  def __init__(self, color, *args, **kwargs):
+    super().__init__(*args, **kwargs)
+    self._color = color
+    if color:
+      import curses
+      curses.setupterm()
+      if sys.hexversion < 0x30203f0:
+        fg_color = str(curses.tigetstr("setaf") or
+                   curses.tigetstr("setf") or "", "ascii")
+      else:
+        fg_color = curses.tigetstr("setaf") or curses.tigetstr("setf") or b""
+      self._colors = {
+        logging.DEBUG: str(curses.tparm(fg_color, 4), # Blue
+                     "ascii"),
+        logging.INFO: str(curses.tparm(fg_color, 2), # Green
+                    "ascii"),
+        logging.WARNING: str(curses.tparm(fg_color, 3), # Yellow
+                     "ascii"),
+        logging.ERROR: str(curses.tparm(fg_color, 1), # Red
+                     "ascii"),
+        logging.CRITICAL: str(curses.tparm(fg_color, 9), # Bright Red
+                     "ascii"),
+      }
+      self._normal = str(curses.tigetstr("sgr0"), "ascii")
+
+  def format(self, record):
+    try:
+      record.message = record.getMessage()
+    except Exception as e:
+      record.message = "Bad message (%r): %r" % (e, record.__dict__)
+    record.asctime = time.strftime(
+      "%m-%d %H:%M:%S", self.converter(record.created))
+    prefix = '[%(levelname)1.1s %(asctime)s.%(msecs)03d %(module)s:%(lineno)d]' % \
+      record.__dict__
+    if self._color:
+      prefix = (self._colors.get(record.levelno, self._normal) +
+            prefix + self._normal)
+    formatted = prefix + " " + record.message
+
+    formatted += ''.join(
+      ' %s=%s' % (k, v) for k, v in record.__dict__.items()
+      if k not in {
+        'levelname', 'asctime', 'module', 'lineno', 'args', 'message',
+        'filename', 'exc_info', 'exc_text', 'created', 'funcName',
+        'processName', 'process', 'msecs', 'relativeCreated', 'thread',
+        'threadName', 'name', 'levelno', 'msg', 'pathname', 'stack_info',
+      })
+
+    if record.exc_info:
+      if not record.exc_text:
+        record.exc_text = self.formatException(record.exc_info)
+    if record.exc_text:
+      formatted = formatted.rstrip() + "\n" + record.exc_text
+    return formatted.replace("\n", "\n    ")
+
+def enable_pretty_logging(level=logging.DEBUG, handler=None, color=None):
+  '''
+  handler: specify a handler instead of default StreamHandler
+  color:   boolean, force color to be on / off. Default to be on only when
+           ``handler`` isn't specified and the term supports color
+  '''
+  logger = logging.getLogger()
+  if handler is None:
+    h = logging.StreamHandler()
+  else:
+    h = handler
+  if color is None:
+    color = False
+    if handler is None and sys.stderr.isatty():
+      try:
+        import curses
+        curses.setupterm()
+        if curses.tigetnum("colors") > 0:
+          color = True
+      except:
+        import traceback
+        traceback.print_exc()
+  formatter = TornadoLogFormatter(color=color)
+  h.setLevel(level)
+  h.setFormatter(formatter)
+  logger.setLevel(level)
+  logger.addHandler(h)